7143 Cjasn 2023 June
7143 Cjasn 2023 June
7143 Cjasn 2023 June
Authors: CHAN, KW; KWONG, A; TAN, K; LUI, S; CHAN, GC; IP, T; YIU, W; COWLING, B;
WONG, V; LAO, L; FENG, Y; LAI, K; TANG, S
Title: Add-on Rehmannia-6–Based Chinese Medicine in Type 2 Diabetes and CKD: A Multicenter
Randomized Controlled Trial
N/mZSmEnSZyTuvQ0mJAo= on 06/13/2023
Manuscript Category:
Funders:
Disclosures:
Author Contributions:
Background: Diabetes is the leading cause of chronic kidney disease (CKD) and kidney failure We
assessed the real-world effectiveness of Rehmannia-6-based Chinese medicine treatment, the most
used Chinese medicine formulation, on the change in estimated glomerular filtration rate (eGFR) and
albuminuria in diabetic CKD patients with severely increased albuminuria.
Methods: In this randomized, assessor-blind, standard care-controlled, parallel, multi-center trial, 148
adult patients from outpatient clinics with type 2 diabetes, estimated glomerular filtration rate (eGFR)
of 30-90 ml/min/1.73m2 and urinary albumin-to-creatinine ratio (UACR) of 300-5000 mg/g were
randomized 1:1 to a 48-week add-on protocolized Chinese medicine treatment program (using
Rehmannia-6-based formulations in granule form taken orally) or standard care alone. Primary
outcomes were the slope of change in eGFR and UACR between baseline and endpoint (48 weeks
after randomization) in the intention-to-treat population. Secondary outcomes included safety, and the
change in biochemistry, biomarkers and concomitant drug use.
Results: The mean age, eGFR and UACR were 65 years, 56.7 ml/min/1.73m2 and 753 mg/g,
respectively. 95% (n=141) of endpoint primary outcome measures were retrievable. For eGFR, the
estimated slope of change was -2.0 (95%CI: -0.1 to -3.9) and -4.7 (95%CI: -2.9 to -6.5)
ml/min/1.73m2 in participants treated with add-on Chinese medicine or standard care alone, resulting
a 2.7 ml/min/1.73m2 per year (95%CI: 0.1 to 5.3, p=0.04) less decline with Chinese medicine. For
UACR, the estimated proportion in the slope of change was 0.88 (0.75 to 1.02) and 0.99 (0.85 to 1.14)
in participants treated with add-on Chinese medicine or standard care alone. The inter-group
proportional difference (0.89, 11% slower increment in add-on Chinese medicine, 95%CI: 0.72 to
1.10, p=0.28) did not reach statistical significance. 85 adverse events were recorded from 50
participants (add-on Chinese medicine vs control: 22 (31%) vs 28 (36%)).
Kam Wa CHAN MSPH DLSHTM MD PhD1, Alfred Siu Kei KWONG MBBS FHKCP2, Kathryn Choon
Beng TAN MBBCh MD FRCP3, Sing Leung LUI MBBS MD PhD FRCP4, Gary Chi-wang CHAN MBBS PhD
5
MRCP , Tai Pang IP MBBS FRCP4, Wai Han YIU PhD1, Benjamin John COWLING PhD FFPH6, Vivian
Taam WONG MBBS FRCP FRCOG FFPH7, Lixing LAO MB PhD7,8, Yibin FENG MB PhD7, Kar Neng LAI
1
MBBS MD DSc FRCP FRCPath , Sydney Chi-wai TANG MBBS MD PhD FRCP FACP1
1
Division of Nephrology, Department of Medicine, The University of Hong Kong, Hong Kong SAR
2
Department of Family Medicine and Primary Healthcare, Hong Kong West Cluster, Hospital Authority, Hong
Kong SAR
3
Division of Endocrinology & Metabolism, Department of Medicine, The University of Hong Kong, Hong
Kong SAR
4
Department of Medicine, Tung Wah Hospital, Hong Kong SAR
5
Department of Medicine, Queen Mary Hospital, Hong Kong SAR
6
Division of Epidemiology and Biostatistics, School of Public Health, The University of Hong Kong, Hong
Kong SAR
7
School of Chinese Medicine, The University of Hong Kong, Hong Kong SAR
8
Virginia University of Integrative Medicine, Fairfax, Virginia, USA
Correspondence
Department of Medicine, 4/F Professorial Block, 102 Pokfulam Road, Hong Kong (Email:
[email protected], Tel: +852 2255 3603).
Chinese medicine was associated with reduced risk of kidney failure and mortality in
Taiwan CKD patients.4,5 Despite the increasing use gloablly,6-8 current evidence in diabetic
CKD is limited by non-representative cohorts with short follow-up (median: 24 weeks) and
high attrition.8-10 Previously, a diabetic CKD treatment protocol based on a classical Chinese
medicine formulation named Rehmannia-6 (also known as Liu-wei-di-huang-wan: Radix-
Rehmannia, Fructus-Corni, Rhizoma-Dioscoreae, Poria, Cortex-Moutan and Rhizoma-
Alismatis)11 was developed from expert consensus and literature review.12 Rehmannia-6 is the
most prescribed Chinese medicine formulation for diabetes and CKD through “replenishing
kidney” according to traditional theory.11 Meta-analyses,9 and multiple cohorts associated
Rehmannia-6-based treatment with stabilized kidney function.5,13 Ingredients of Rehmannia-6
could attenuate podocyte apoptosis/detachment, and reduce oxidative stress via TNF, NOD-
like receptor, HIF-1 and PI3K-Akt signaling, complement and coagulation cascades.8,11,14,15
The protocol personalized Rehmannia-6 treatment with prespecified minor adjustment to
phenotype-based subgroups according to Chinese medicine practice.8
Design
This study was approved by the Institutional Review Board of the HKU/Hospital
Authority HK West Cluster (Ref:UW14-301), adhered to the Declaration of Helsinki, and
monitored by a Data and Safety Monitoring Board of independent academics (supplemental
material S2). All participants provided written informed consent. Trial was registered at
ClinicaTrials.gov (NCT02488252).
Participants
Patients aged 35 to 80 with 1) type 2 diabetes for ≥5 years; 2) repeated eGFR of 30-90
mL/min/1.73m2 for ≥3 months; 3) persistent macro-albuminuria (UACR of 300-5000 mg/g
from ≥3-month-apart repeated consecutive first morning void urine samples); and 4) stable
anti-diabetic medications and angiotensin-converting-enzyme inhibitor (ACEi)/angiotensin II
receptor blocker (ARB) for ≥12 weeks were recruited.
Eligible patients were randomized 1:1 to the add-on Chinese medicine treatment program or
standard care alone. Randomization was stratified by field-tested phenotype-based
diagnosis18 with three symptoms as 1) phenotype predicts clinical outcomes,21,22 including
eGFR decline in diabetic patients,18,23 2) phenotype-based diagnosis is fundamental to
Chinese medicine real-world practice,8,10 and 3) treatment personalization reduces response
heterogeneity, increases power of trials24 and better mimic clinical application.25 The
phenotype-based diagnosis was adapted from a clinical practice guideline based on expert
consensus (supplemental material S5).12
The Chinese medicines used were licensed in Australia, Canada, China and US.
Fabrication of medication (water-extracted) complied with Good Manufacturing Practice
ensuring consistent quality, stability and safety (against heavy metal, microorganism and
pesticide contamination). Chemical constituents of the medications were determined by high-
performance liquid chromatography (supplemental material S5). Voucher specimens are
retained at provider and S.C.W.T. laboratory.
Outcomes
Prespecified primary outcomes were the slope of eGFR and UACR change over the
study period. Secondary outcomes included the endpoint eGFR, UACR, blood pressures,
biochemistry, biomarkers, and the rate of concomitant drug change, specialist referral and
adverse events. Endpoint was defined as 48 weeks after randomization. Definition and
detailed assessment methods are listed in supplemental material S7.
Outcomes were assessed at 6, 24 and 48 weeks after randomization. Blood and urine
were sampled after overnight (≥8 hours) fasting, freshly assessed by accredited hospital
laboratories, and stored at -80oC until further analysis. Creatinine was measured by
colorimetric assay, and urine albumin was measured by immunoturbidimetric assay. GFR
was estimated by the CKD-EPI formula: GFR = 141 × min(Scr/κ, 1)α × max(Scr/κ, 1)-1.209 ×
0.993Age × 1.018 [if female] ×1.159 [if black in ethnicity]. Inflammation- and fibrosis-related
biomarkers were measured at S.C.W.T. laboratory with enzyme-linked immunosorbent assay.
Post-hoc analysis of insulin resistance (TyG index) was estimated by using fasting blood
triglyceride and glucose. TyG index correlates well with HOMA-IR index,27 and is associated
with cardiovascular and all-cause mortality.28,29
Adverse events were obtained from all linked hospital emergency records, follow-up
consultations, and participants’ self-reporting with a questionnaire. Participants who
developed serious adverse event (SAE) were withdrawn from study medication for safety but
Statistical analysis
eGFR reflects kidney function and was used for sample size calculation. The calculation was
based on 1) the expected accuracy of obtained estimates as a definitive study, and 2) the
desired control of inflation factor to the sample size estimation for subsequent large-scale
territory-wide implementation studies as a pilot (supplemental material S3).17,30 Briefly, ≥80
patients in total were needed to generate estimates for the sample size estimation of
subsequent studies (to minimize nominal power forfeit to ≦10%) testing if add-on Chinese
medicine stabilizes GFR community-wide with 95% two-sided confidence. 125 patients were
needed to also have 80% (nominal power forfeit ≦20%) one-sided confidence that the effect
remains significant at subgroup level stratified by phenotype. 148 patients allowed 15%
attrition and provided 66% power with 95% two-sided confidence to detect a 5
min/ml/1.73m2 inter-group GFR difference (12 ml/min/1.73m2 as within-group standard
deviation) (G*Power 3.1.9.2).31 A 5 ml/min/1.73m2 annual eGFR decline is widely defined as
rapid progression.25,32
UACR was log-transformed for analysis. Rapid eGFR decline was predefined as
consecutive yearly eGFR drop of ≥5 ml/min/1.73m2 or cumulative 5-year drop of ≥25
ml/min/1.73m2.25,32 Means and proportions were compared by t-test and χ2 test. Intention-to-
treat population was used for primary analysis.
The adjusted mean (±SE) of quantitative outcomes at endpoint, and the estimated
difference in the change between groups were compared by linear mixed-effect model (and
linear growth model for slope estimation) with measures at all prespecified timepoints
(baseline, 6-, 24- and 48-week).31,33,34 Baseline value, smoking history (as baseline history
differed between groups), allocation, visit, and the interaction between allocation and visit
were included as covariates. Correlation between the percentage change of biomarkers and
eGFR from baseline was assessed by univariable linear regression to explore underpinning
mechanism. A sensitivity analysis was performed to exclude extreme percentage change of
TNF receptor 2 and eGFR.
RESULTS
Recruitment and demographics
Median follow-up for the primary analysis was 48 weeks (IQR: 47 to 49) for all
participants and both groups. Median Chinese medicine intake was 44 weeks (IQR: 34 to 48)
in the intervention group. Prospective endpoint primary outcome measurements of 141 (95%)
participants were available for analysis. Three (one deceased, one relocated geographically,
one self-withdrew for other Chinese medicines) and four (one deceased, two were busy, one
self-withdrew to start Chinese medicine) participants were lost to follow-up in the
intervention and control arm, respectively. 117 (79%) participants completed the full 48-
week study protocol. 17 participants of add-on Chinese medicine group (one deceased, ten
other developed SAEs, two self-withdrew after non-serious adverse events, one relocated
geographically and three were busy) and 14 participants of standard care group (one deceased,
ten other developed SAEs, one self-initiated Chinese medicine and two were busy) dropped
out during observation.
Primary outcomes
The estimated difference in the slope of eGFR decline was 2.7 ml/min/1.73m2 per year
(95%CI: 0.1 to 5.3, p=0.04) less in Chinese medicine-treated participants (within-group
change: -2.0±1.0 vs -4.7±0.9) (table 2, figure 2). For UACR, the estimated proportional
difference in the slope of change (0.89, 11% slower increment in add-on Chinese medicine,
95%CI: 0.72 to 1.10, p=0.28) between add-on Chinese medicine (0.88±1.08) and standard
care (0.99±1.08) group was statistical insignificant (table 2, figure 2).
Secondary outcomes
Chinese medicine-treated patients had 3.1 ml/min/1.73m2 (95%CI: 1.0 to 5.3, p=0.004) higher
endpoint least-squares mean eGFR (±SE) when compared to standard care control (54.8±0.8
vs 51.7±0.8). For UACR, the estimated proportional difference at endpoint (0.88, 12% lower
in add-on Chinese medicine, 95%CI: 0.73 to 1.06, p=0.18) between groups (geometric mean
(±SE): 659±9 vs 748±9 mg/g) groups was insignificant.
At the first routine follow-up visit after the study period (median: 59 weeks; IQR: 55
to 63.2), 19 (27%) Chinese medicine-treated and 33 (43%) control participants had their anti-
diabetic medications increased from baseline (difference: -16%, 95%CI: -1 to -32, p(χ2)=0.04)
The percentage change of TNF receptor 2 level was inversely associated with that of
eGFR in either all participants (n=141, β=-0.1%, p=0.007) or the intervention arm alone, and
remained robust after excluding datapoints (supplemental material S10). The triglyceride-
glucose was significantly reduced with add-on Chinese medicine (estimated difference in
slope: -0.21, 95%CI: -0.36 to -0.06, p=0.006), regardless of the baseline use of insulin
(supplemental material S11).
Adverse events
The rate of SAEs, all-cause mortality, major adverse cardiac events, moderate to severity to
severe hyperkalemia and acute kidney injury was comparable between groups (table 4).
There were 11 SAEs from 11 Chinese medicine-treated participants, leading to protocolized
discontinuation of intervention. There were 21 SAEs from 11 participants who received
standard care. Most common SAEs (add-on Chinese medicine vs control) were severe
hyperkalemia (serum potassium ≥6.0 mmol/L or hospitalization) (two vs three participants),
hospitalized dyspnea (zero vs three) and neoplasms (one vs two). One participant died in each
arm (triple vessel disease vs sudden cardiac arrest). Detailed SAEs are listed in supplemental
material S12.
There was no significant interaction between treatment effect and subgroups (figure 3,
supplemental material S14). The results remained robust in sensitivity analyses (supplemental
material S14).
DISCUSSION
In this trial, diabetic patients with CKD and macroalbuminuria receiving add-on protocolized
Rehmannia-6-based Chinese medicine treatment had less 1) GFR decline, 2) increase in anti-
diabetic medication use and 3) risk of hypoglycemia after 48 weeks.
The reduction of hypoglycemia risk was independent to the reduced anti-diabetic drug
use, and the treatment effect remained robust after adjusting blood glucose and pressure
control. Post-hoc analysis on the TyG index28 shows a reduced insulin resistance in Chinese
medicine-treated patients (supplemental material S11) which could explain the stabilized
anti-diabetic drug use. The difference in estimated albuminuria change was insignificant,
likely due to the high within-person biological variance of albuminuria40 and the relatively
small sample size. Besides, the kidney-protective effect of Rehmannia-6 could be
independent to albuminuria reduction,41,42 similar to astrasentan43 and to some extent the
SGLT2i.44
The demographics of our cohort is comparable to other major trials (e.g. SONAR,25
DAPA-CKD,31 CREDENCE45) with comparable annual eGFR decline (4.1 to 4.7
ml/min/1.73m2 in sensitivity analyses) among these placebo-controlled groups (DAPA-
CKD,31: -4; CREDENCE45: -5 ml/min/1.73m2). The eGFR decline after add-on Chinese
Our trial has a pragmatic design based on converging qualitative, epidemiological and
mechanistic evidence. Although assessing a single herb or an extract could better assess the
efficacy of each herbal component, the translation potential would be limited as mixed
formulations are most used in practice. Most existing trials assessed single Chinese medicine
formulation either 1) in all participants without personalization, or 2) in a phenotype-based
subset with limited representation of the diabetic CKD population.10,16 Assessing Chinese
medicine as a program intervention20 better mimics clinical practice. Phenotype-stratified
randomization facilitates the analysis of effectiveness at both the program level (in the whole
diabetic CKD population) and formulation level (in each phenotype subgroup) without
excluding patients by phenotypes. Also, we integrated the trial to conventional healthcare
workflow with linked registry tracking to assess real-world effectiveness with surveillance on
adverse events and concomitant drug change.
Quality of life was not measured as CKD remains silent except with acute kidney
injury or kidney failure, and patients prefer using eGFR for monitoring.10 There was no
Chinese medicine-associated acute kidney injury and kidney failure in our trial, and a reduced
rate of mortality and kidney failure was associated with Chinese medicine use in long-term in
retrospective cohorts.4 Phenotype-based diagnosis requires training to implement. We
protocolized the diagnosis to stratify patients by only three symptoms and were able to
propagate to six outpatient clinics in this trial. Implementation could be further assessed by
community-wide trials with process evaluation.
This investigator-initiated trial was made possible in part through the support of the Health
and Medical Research Fund (Ref: 12133341, 14151731), philanthropic donation from Mr
Winston Leung, Mr Abraham Chan and the Croucher Senior Medical Research Fellowship
awarded to SC Tang. Purapharm provided discount for study medications. KW Chan was
supported by the Sir Edward Youde Memorial Fellowship, HKU Postgraduate Fellowships in
Integrative Medicine, and an endowment established for the Yu Professorship in Nephrology
at the University of Hong Kong awarded to SC Tang.
The funding organizations had no role in the design and conduct of the study; collection,
management, analysis, and interpretation of the data; preparation, review, or approval of the
manuscript; and decision to submit the manuscript for publication.
We thank the contribution of all clinical staff and scientists for their participation in the
SCHEMATIC initiative and the trial. We especially thank Wenjun ZOU for generating and
concealing the randomization sequence, Ms Pearl YUN and nurses for the support of patient
care, Ms Kam Yan YU for independent verification of clinical data, and Mr William CHUI,
Mr Edmond CHENG, Ms Tammy CHAN and pharmacy staff for dispensing study
medications. Clinical biochemistry and pathology investigations were provided
independently as routine clinical service by the Department of Pathology & Clinical
Biochemistry of Queen Mary Hospital. We thank Purapharm for providing discounted
medications for this study. We thank Prof Vincent CHUNG, Prof Wendy WONG and Dr
Jerry YEUNG for the supervision of the trial as Data and Safety Monitoring Board members.
We also thank Prof Jurgen FLOEGE (Aachen), and Prof Zhao-xiang BIAN (Hong Kong) for
their constructive comments on the study design and the final manuscript.
Abbreviations
The primary outcomes were the change of estimated glomerular filtration rate (eGFR) and
urine albumin-to-creatinine ratio (UACR). Same data with enlarged y-axis shown by inset. CI:
confidence interval.
A. The estimated difference in the slope of eGFR decline was 2.7 ml/min/1.73m2 per year
(95%CI: 0.1 to 5.3, p=0.04) less after receiving add-on Chinese medicine. Participants
who received add-on Chinese medicine had 3.1 ml/min/1.73m2 (95%CI: 1.0 to 5.3,
p=0.004) higher endpoint least-squares mean eGFR (±SE) when compared to standard
B. For UACR, the estimated proportional difference at endpoint (0.88, 12% lower in add-on
Chinese medicine, 95%CI: 0.73 to 1.06, p=0.18) between participants who received add-
on Chinese medicine and standard care control (geometric mean (±SE): 659±9 vs 748±9
mg/g) did not reach statistical significance. The estimated proportional difference in the
slope of change (0.89, 11% slower increment in add-on Chinese medicine, 95%CI: 0.72
Forest plots of the estimated difference in the slope of change of primary outcomes between
A. Forest plot of the estimated difference in the slope of estimated glomerular filtration rate
B. Forest plot of the estimated proportional difference in the slope of urine albumin-to-
albumin-to-creatinine ratio.
Age – years 65 ± 9 66 ± 9 64 ± 9
Body mass index – kg/m2 27.7 ± 4.5 27.1 ± 3.8 28.3 ± 5.1
Source of recruitment
Smoking history
Diastolic 76 ± 11 76 ± 11 77 ± 11
Cholesterol – mg/dL
Triglyceride 70 ± 35 70 ± 31 70 ± 39
High-density lipoprotein 43 ± 12 43 ± 12 43 ± 12
Low-density lipoprotein 77 ± 23 77 ± 23 77 ± 23
3-year pre-trial eGFR decline – ml/min/1.73m2 10.9 ± 14.1 10.7 ± 14.9 11.1 ± 13.3
Data are shown for the intention-to-treat population. Plus-minus values are means±standard deviation.
† Urine albumin-to-creatinine ratio (UACR) was log-transformed for statistical analysis and presented as
geometric mean.
‡ Rapid estimated glomerular filtration rate (eGFR) decline was pre-defined as yearly eGFR drop of ≥5
ml/min/1.73m2 or cumulative 5-year eGFR drop of ≥25 ml/min/1.73m2.
¶ No participants were on glucagon-like peptide-1 receptor agonists as not available at the time of recruitment.
Primary outcomes Adjusted mean (95% CI) Adjusted mean (95% CI) Mean difference (95% CI) p
Slope of eGFR change – ml/min/1.73m2 per year -2.0 (-0.1 to -3.9) -4.7 (-2.9 to -6.5) 2.7 (0.1 to 5.3) 0.04
Slope of UACR change 0.88 (0.75 to 1.02) 0.99 (0.85 to 1.14) 0.89 (0.72 to 1.10) 0.28
Secondary outcomes Adjusted mean (95% CI) Adjusted mean (95% CI) Mean difference (95% CI) P
eGFR – ml/min/1.73m2 54.8 (53.3 to 56.3) 51.7 (50.2 to 53.2) 3.1 (1.0 to 5.3) 0.004
UACR – mg/g 659 (577 to 754) 748 (659 to 852) 0.88 (0.73 to 1.06) 0.18
(proportional difference)
Systolic blood pressure – mmHg 141 (137 to 145) 140 (137 to 144) 0 (-4 to 5) 0.87
Diastolic blood pressure – mmHg 72 (69 to 74) 74 (72 to 76) -2 (-5 to 1) 0.15
Serum potassium – mmol/L 4.3 (4.2 to 4.4) 4.4 (4.3 to 4.4) -0.0 (-0.2 to 0.1) 0.53
Hemoglobin – g/dL 12.9 (12.6 to 13.1) 12.9 (12.7 to 13.1) -0.0 (-0.3 to 0.3) 0.84
Hemoglobin A1c – % 7.3 (7.1 to 7.5) 7.2 (7.0 to 7.3) 0.2 (-0.1 to 0.4) 0.21
Serum alkaline phosphatase – U/L 72 (67 to 76) 71 (68 to 75) 0 (-5 to 5) 0.90
Serum alanine aminotransferase – U/L 24 (21 to 26) 24 (22 to 26) -1 (-4 to 2) 0.57
Serum aspartate aminotransferase – U/L 24 (22 to 26) 25 (23 to 27) -1 (-3 to 2) 0.46
Serum gamma-glutamyl transferase – U/L 40 (35 to 45) 41 (37 to 45) -1 (-6 to 4) 0.71
Serum albumin – g/dL 4.3 (4.2 to 4.5) 4.3 (4.2 to 4.4) 0.0 (-0.1 to 0.2) 0.60
Serum urate – mg/dL 7.2 (6.8 to 7.6) 7.5 (7.2 to 7.8) -0.3 (-0.7 to 0.1) 0.15
Serum TNF receptor 1 – pg/ml 726.2 (663.6 to 788.9) 780.8 (729.3 to 832.2) -54.5 (-125.8 to 16.7) 0.13
Serum TNF receptor 2 – pg/ml 5463.6 (5012.7 to 5914.6) 6222.3 (5851.6 to 6593) -758.7 (-1268.4 to -248.9) 0.004
Serum TNF-alpha – pg/ml 3.4 (2.6 to 4.1) 3.8 (3.2 to 4.5) -0.5 (-1.3 to 0.4) 0.26
Serum VEGF – pg/ml 564.8 (495.3 to 634.3) 541.5 (484.8 to 598.3) 23.2 (-55.5 to 101.9) 0.56
Serum MCP-1 – pg/ml 100.6 (91.2 to 110.0) 103.9 (96.3 to 111.4) -3.4 (-14.0 to 7.2) 0.53
Serum TGF-beta1 – ng/ml 42.0 (39.6 to 44.4) 42.1 (40.1 to 44.1) -0.1 (-2.8 to 2.7) 0.96
Serum FGF-23 – pg/ml 156.2 (71.0 to 241.5) 140.9 (71.1 to 210.7) 15.3 (-80.9 to 111.5) 0.75
Mixed-effect regression model was used to analyze all repeated outcome measures at prespecified timepoints, including baseline for baseline values of dependent variables,
allocation, smoking history and the interaction between allocation and visit as covariates. Endpoint is defined as 48 weeks after randomization. Adjusted endpoint levels and
the within group difference with 95% confidence interval are shown. Urine albumin-to-creatinine ratio (UACR) was log-transformed and proportional difference is presented.
FGF: fibroblast growth factor; MCP: monocyte chemoattractant protein; TGF: transforming growth factor; TNF: tumor necrosis factor; VEGF: vascular endothelial growth
factor.
Add-on Chinese Standard care Difference – % (95%CI) p (χ2) Hazard ratio – (95%CI) p (Cox)
medicine control
n / N (%) n / N (%)
Increased anti-diabetic agents 19 / 71 (27) 33 / 77 (43) -16 (-1 to -32) 0.04 0.63 (0.36 to 1.10) 0.11
Increased SGLT2i/DPP4i 10 / 71 (14) 8 / 77 (10) 4 (-7 to 14) 0.50 1.04 (0.39 to 2.77) 0.94
Increased SGLT2i 2 / 71 (3) 4 / 77 (5) -2.4 (-9 to 4) 0.11 0.24 (0.04 to 1.40) 0.47
Increased DPP4i 8 / 71 (11) 5 / 77 (6) 5 (-5 to 14) 0.31 1.57 (0.49 to 5.05) 0.45
Reduced ACEi/ARB due to hyperkalemia 1 / 71 (1) 5 /77 (6) -5 (-12 to 1) 0.12 0.35 (0.07 to 1.75) 0.20
Increased anti-hypertensive agents 21 / 71 (30) 16 / 77 (21) 9 (-5 to 23) 0.22 1.06 (0.53 to 2.12) 0.86
Increased diuretics 6 / 71 (9) 6 / 77 (8) 1 (-8 to 10) 0.88 0.44 (0.13 to 1.44) 0.17
Increased lipid-lowering drugs 2 / 71 (3) 7 / 77 (9) -6 (-14 to 2) 0.11 0.41 (0.10 to 1.64) 0.21
The rate of concomitant drug change at endpoint were compared between group by Chi-square test. The hazard ratios of concomitant drug change were further estimated by
Cox proportional-hazards model adjusting baseline estimated glomerular filtration rate, urine albumin-to-creatinine ratio and smoking history.
ACEi: angiotensin-converting-enzyme inhibitor; ARB: angiotensin II receptor blocker; DPP-4i: dipeptidyl peptidase-4 inhibitor; SGLT2i: sodium–glucose cotransporter 2
inhibitor;
Major adverse events n / N (%) n / N (%) Difference – % (95%CI) p (χ2) Hazard ratio – (95%CI) p (Cox)
Any serious adverse events 11 / 71 (16) 11 / 77 (14) 1 (-10 to 13) 0.84 0.78 (0.32 to 1.91) 0.59
Death 1 / 71 (1) 1 / 77 (1) 0.1 (-4 to 4) 0.95 0.18 (0.003 to 9.19) 0.38
Major adverse cardiac events 2 / 71 (3) 2 / 77 (3) 0.2 (-5 to 6) 0.94 0.44 (0.06 to 3.41) 0.43
Moderate to severe hyperkalemia (10020647) 3 / 71 (4) 6 / 77 (8) -4 (-11 to 4) 0.37 0.26 (0.05 to 1.28) 0.10
Acute kidney injury (10069339) 0 / 71 (0) 1 / 77 (1) -1 (-4 to 1) 0.34 0.00 (. to 0) 1.00
Hypoglycemia (10021005) 4 / 71 (6) 11 / 77 (14) -9 (-18 to 1) 0.08 0.21 (0.06 to 0.71) 0.01
¶Adverse events with more than 1 event are shown. Full list of adverse events is listed in supplemental material S12/S13. Common Terminology Criteria for Adverse Events
(CTCAE) 4.03 code is included for all adverse events.
Hardly a day passes without the announcement of AI/ML applications for dialysis is proliferating,
1
new successes delivered by artificial intelligence mostly on research studies and much less on real- Renal Research
(AI) and machine learning (ML). The list of AI/ML world routine use. Broadly speaking, reports on Institute, New York,
New York
N/mZSmEnSZyTuvQ0mJAo= on 06/13/2023
applications that directly or indirectly touch our lives AI/ML in dialysis focus on diagnosis, prognosis, 2
Icahn School of
is growing rapidly, including health care. But, has AI/ and treatment recommendations. For example, Chan Medicine at Mount
ML reached the dialysis space, and if so, where and to et al.1 applied natural language processing to EHRs to Sinai, New York, New
what extent? identify symptom burden in dialysis patients; their York
3
Department of
Although there exists no universally accepted defi- study revealed that natural language processing had
Statistics & Applied
nition, AI can be described as computer-based intelli- higher sensitivity compared with International Clas- Probability, University
gence that encompasses the ability to learn from data. sification of Diseases codes for identification of of California at Santa
AI combines concepts from several fields, such as common hemodialysis-related symptoms. Zhang Barbara, Santa
computer science, mathematics, statistics, and data et al.2 successfully used cloud-based AI/ML-driven Barbara, California
science. ML is a subdiscipline of AI that aims to identify image analysis to classify vascular access aneurysms
Correspondence:
associations or patterns in data by a computer using as nonadvanced or advanced with an area under the Dr. Peter Kotanko,
mathematical algorithms. Broadly speaking, the ML receiver operating characteristic curve of 0.96. Prog- Renal Research
process can be categorized as either supervised or nosis is the most widely used application of AI/ML Institute, 315 East
unsupervised learning. Supervised learning seeks as- in dialysis. Using data from over 260,000 hemodial- 62nd Street, 3rd Floor,
sociations in situations where both input and output ysis sessions, Lee et al.3 developed a recurrent neural New York, NY 10065.
Email: peter.kotanko@
data are known. By contrast, unsupervised learning network model that predicted intradialytic hypoten- rriny.com
aims to identify patterns in input data without a priori sion with an area under the receiver operating
knowledge of a target. characteristic curve of 0.94. Other examples are
Irrespective of methodological nuances, AI/ML prognosis of hospitalization,4 mortality,5 coronavirus
models usually require large amounts of training disease 2019,6 and abnormal oxygen saturation pat-
data. This makes dialysis a particularly attractive tern.7 For treatment recommendation, Barbieri et al.8
field for AI/ML applications because dialysis care developed an anemia control model that suggests
is a highly standardized process that generates a iron and erythropoiesis-stimulating agent doses. As
large volume of longitudinal patient-level data. Di- pointed out by Correa and Mc Causland,9 AI/ML
alysis prescription (e.g., treatment time, ultrafiltration can help improve dialysis patient safety.
volume, flow rates) and medical treatment adminis- However, as noted earlier, routine use of AI/ML
tered on site (e.g., erythropoiesis-stimulating agents) applications in dialysis is rare. The previously
are usually documented in electronic health records mentioned anemia control model8 and ML-driven
(EHRs). These data are complemented by informa- interventions to reduce hospitalization rates4 are
tion on patient demographics, comorbidities, and notable exceptions. The hesitancy on the side of
laboratory results. In many US facilities, intradialytic dialysis organizations and health care providers to
biosignals are automatically recorded (e.g., BP, heart implement AI/ML models in their daily workflow is
rate). In addition, electronically stored free-text notes understandable. Development and implementation
are accessible to AI/ML-based natural language of AI/ML methods in dialysis shares common chal-
processing. In some use cases, there is potential lenges as in other areas, such as data privacy
value of incorporating other types of information, considerations and lack of a clear cause-and-effect
such as social determinants of health data, which interpretation of model output. AI/ML models only
may have just as large or even larger effect on rarely allow the user to understand how a specific
outcomes than standard medical information. In the outcome was reached; their inherent black-box nature
future, data from wearable devices may provide can be frustrating to users and may cause commu-
insights into the interdialytic period (Figure 1). nication issues with health care providers and pa-
Given the availability of large databases and tients. Other problems arise regarding the portability
multimodal data, such as text, images, and waveform of AI/ML models between patient populations. Even
data, it is not surprising that literature describing if a successful use case is well-described in the
www.cjasn.org Vol 18 June, 2023 Copyright © 2023 by the American Society of Nephrology 803
804 CJASN
eAX6oP97/t+reHY6iBuDuEG8j/KLK4+srUZdWUQbtcBUzqkxhgmmXz8HwSY/clF6gRoZeOUloCHsnrNgIHFbIB8VBAS7/AWU3tQV
Downloaded from http://journals.lww.com/cjasn by ovbQKPXYfBFIRkaiUq5PuVRWaEDA1iTVOAm/wWdC7XRdvs6oSNs
N/mZSmEnSZyTuvQ0mJAo= on 06/13/2023
Figure 1. Examples of data sources and applications of AI/ML in dialysis. The icons and arrows indicate the various components and flow
pathways. On the input side, examples above the arrows indicate strengths, whereas those below the arrows (in italic) indicate challenges. On
the output side, some examples of use cases are indicated. AI, artificial intelligence; ECG, electrocardiogram; EHR, electronic health records;
IT, information technology; ML, machine learning.
literature and the code made publicly available, it is likely compounded by the lack of clear regulatory requirements
that the respective AI/ML model needs to be adjusted and pathways because AI/ML applications may be
when applied to a different patient population. Experience classified as medical devices or clinical decision support
has shown that applying the same model (e.g., a model systems.10 Given these challenges, it is not surprising that
with the same neural network topology and neuron the real-world adoption of AI/ML models in the dialysis
transfer functions and weights) to a different population field is slow and that only a few AI/ML applications have
may result in inferior performance. Therefore, if model left the realm of research and found their way into routine
performance is poor, retraining is a sensible step. Real- patient care. Despite the pervasive AI/ML hype, it is
time prediction of intradialytic hypotension, for example, important to manage expectations and understand that
requires fast and reliable connectivity; as such, applica- AI/ML models cannot provide general solutions but are
tions are mostly web-based. In addition, because most AI/ tools to serve very specific and narrowly defined tasks.
ML models draw on data from multiple sources, such as AI/ML algorithms are meant to assist and not to sub-
EHRs, machine data, etc., extracting and merging of such stitute clinicians’ work. Any successful implementation of
multimodal data can be challenging. A frequently en- AI models in real practice will depend on the user
countered problem when developing AI/ML applications experience part, on how smoothly they are integrated
for dialysis is the different time scales of data recordings, in the real-life point-of-care processes.
ranging from seconds (e.g., some machine data) to minutes The US Food and Drug Administration (FDA), Health
(e.g., BP) to days (e.g., administration of drugs) and to Canada, and the United Kingdom’s Medicines and
weeks or months (e.g., laboratory results). AI/ML Healthcare Products Regulatory Agency have jointly
methods to address this specific complication are under identified ten guiding principles that inform the devel-
development. Another practical challenge is that opment of good ML practice.11 They address issues
AI/ML models require a mature technical infrastruc- including good software engineering practices, data
ture and experts to maintain it. The situation is further collection protocols to ensure that the relevant
CJASN 18: 803–805, June, 2023 Artificial Intelligence and Machine Learning in Dialysis, Kotanko et al. 805
References
cally important.
Downloaded from http://journals.lww.com/cjasn by ovbQKPXYfBFIRkaiUq5PuVRWaEDA1iTVOAm/wWdC7XRdvs6oSNs
measures will help to better define the place for AI/ML interventions associate with decreased hospitalization rates in
models in dialysis care. hemodialysis patients. Int J Med Inform. 2021;153:104541. doi:
10.1016/j.ijmedinf.2021.104541
5. Yang CH, Chen YS, Moi SH, Chen JB, Wang L, Chuang LY.
Disclosures Machine learning approaches for the mortality risk assessment of
P. Kotanko reports employment with Renal Research patients undergoing hemodialysis. Ther Adv Chronic Dis.
Institute, a wholly owned subsidiary of Fresenius Medical Care; 2022;13:20406223221119617. doi:10.1177/
holds stock in Fresenius Medical Care; reports research funding 20406223221119617
from Fresenius Medical Care, KidneyX, and NIH; is an inventor 6. Monaghan CK, Larkin JW, Chaudhuri S, et al. Machine learning
for prediction of patients on hemodialysis with an undetected
on patent(s) in the field of dialysis; receives author royalties from SARS-CoV-2 infection. Kidney360. 2021;2(3):456–468. doi:10.
HS Talks and UpToDate; and serves on the Editorial Boards of 34067/kid.0003802020
Blood Purification, Frontiers in Nephrology, and Kidney and Blood 7. Galuzio PP, Cherif A, Tao X, et al. Identification of arterial oxygen
Pressure Research. H. Zhang reports employment and patents intermittency in oximetry data. Scientific Rep. 2022;12(1):16023.
doi:10.1038/s41598-022-20493-0
with Renal Research Institute, a wholly owned subsidiary of
8. Barbieri C, Molina M, Ponce P, et al. An international observa-
Fresenius Medical Care. The remaining author has nothing tional study suggests that artificial intelligence for clinical de-
to disclose. cision support optimizes anemia management in hemodialysis
patients. Kidney Int. 2016;90(2):422–429. doi:10.1016/j.kint.
Funding 2016.03.036
National Institute of Diabetes and Digestive and Kidney Diseases 9. Correa S, Mc Causland FR. Leveraging deep learning to improve
safety of outpatient hemodialysis. Clin J Am Soc Nephrol. 2021;
of the National Institutes of Health (Grant/Award Number: 16(3):343–344. doi:10.2215/CJN.00450121
“R01DK130067”). 10. Barbieri C, Neri L, Chermisi M, et al. How to assess the risks
associated with the usage of a medical device based on predictive
Acknowledgments modeling: the case of an anemia control model certified as
This article is part of the Artificial Intelligence and Machine medical device. Expert Rev Med Devices. 2021;18(11):1117–
Learning in Nephrology series, led by series editor 1121. doi:10.1080/17434440.2021.1990037
Girish N. Nadkarni. 11. US Food and Drug Administration. Good machine learning
practice for medical device development: guiding principles.
The content of this article reflects the personal experience and
Published October 27, 2021. Accessed February 2, 2023. https://
views of the author(s) and should not be considered medical www.fda.gov/medical-devices/software-medical-device-samd/
advice or recommendation. The content does not reflect the good-machine-learning-practice-medical-device-development-
views or opinions of the American Society of Nephrology (ASN) guiding-principles
or CJASN.
Responsibility for the information and views expressed herein lies
entirely with the author(s). Published Online Ahead of Print: January 27, 2023
Editorial
dangerously or not at all—if their doctors had their ated with a higher preeclampsia risk. Women with of Chicago,
way.”1 This quote by an anonymous author in a 1975 preeclampsia had a higher median preconception Chicago, Illinois
editorial described how physicians in the past felt serum creatinine and a lower 3-month postpartum
Correspondence:
about pregnancies in women with CKD. In line with eGFR. Another intriguing finding was that the
Dr. Michelle A.
this perspective, a 2002 survey of kidney transplant diagnosis of preeclampsia during pregnancy after Josephson, University
program directors found that approximately 15% of kidney transplantation increased dramatically over of Chicago 5841 S.
respondents counseled their kidney transplant patients time and is higher than the rates previously cited. Maryland Avenue,
against pregnancy.2 Despite this advice, there have The reported frequency of a diagnosis of preeclamp- MC5100 Chicago, IL
60637 . Email:
been numerous successful pregnancies in kidney trans- sia increased from 27% between 2000 and 2004 to mjosephs@bsd.
plant recipients starting with Edith Helm, the first 48% between 2018 and 2021. uchicago.edu
female transplant recipient and the first to become These findings are striking for several reasons.
pregnant, delivering a healthy baby in 1958.3 First, they differ from published findings in the
This long history does not mean that pregnancies general population.8 This leaves us asking why we
in kidney transplant patients are risk-free. Pregnant are not seeing a similar outcome in kidney transplant
kidney transplant recipients have a six-time patients with preeclamptic pregnancies. One expla-
higher risk of developing preeclampsia compared nation for the lack of detected effect of preeclampsia
with the general population. 4 This greater suscep- on graft outcome is that the effect of preeclampsia
tibility to preeclampsia in transplant patients is was minor compared with other factors such as
alarming not only for its potential to evolve into alloimmune responses affecting transplant recipi-
life-threatening eclampsia but also for its strong ents in both groups. Or perhaps the follow-up
association with several chronic complications, 5 time was not sufficiently long. Another explanation
including maternal cardiovascular diseases, ve- may be related to limitations in our ability to de-
nous thromboembolism, and metabolic syndrome. 6 finitively diagnose preeclampsia. We must consider
These sequelae make it all the more critical that we that the incidence of preeclampsia might have been
understand the associated risks and outcomes of either overdiagnosed or underdiagnosed, leading to
preeclampsia in kidney transplant patients to en- false positives or false negatives undermining our
able informed shared decision making and opti- ability to detect graft survival outcome differences
mal management. between the groups. The more plausible possibility of
In this issue of CJASN, Lu et al.7 attempted to do just these two is that preeclampsia was overdiagnosed.
that by examining the effect of a diagnosis of pre- After all, transplant recipients have a single kidney
eclampsia on kidney transplant graft function and that is already hyperfiltering, and many also have
survival. Analyzing retrospective data from January background hypertension and low-level proteinuria.
2000 to December 2021, obtained from the Australia With the additional hyperfiltration of pregnancy, pro-
and New Zealand Dialysis and Transplant Registry, teinuria may increase. These factors may confound the
they considered graft outcomes of 390 pregnancies diagnosis of preeclampsia, which is subjective and
in women who had a functioning kidney transplant susceptible to error, resulting in overdiagnosis. In-
at conception and pregnancy $20 week’s gestation. creased awareness around preeclampsia recognition
They found that preeclampsia was not associated and management may have also played a role. The
with worse graft survival or function. This study efforts of health care professionals and public health
also uncovered that among the observed maternal campaigns over the past decades have contributed to
factors of age, body mass index, primary kidney better outcomes for mothers and babies as well as
disease, transplant-pregnancy interval, preconcep- potentially increased the incidence of diagnosis such
tion serum creatinine concentration, as well as ta- as the one we are seeing reported in this study.
crolimus or cyclosporine exposure, preconception Furthermore, kidney transplant patients have the
www.cjasn.org Vol 18 July, 2023 Copyright © 2023 by the American Society of Nephrology 1
2 CJASN
additional complicating factor of immunosuppression. The it could help us anticipate which of our patients is at a
calcineurin inhibitor levels tend to drop during pregnancy, higher risk of preeclampsia and differentiate preeclamptic
leading to frequent dose increases for these patients and pregnancies from other mimics in the subset of pregnant
possible calcineurin inhibitor toxicity,8 which could lead to transplant patients. As the authors noted, this and other
findings that are misinterpreted as a manifestation of biomarkers were not checked because they were not clin-
eAX6oP97/t+reHY6iBuDuEG8j/KLK4+srUZdWUQbtcBUzqkxhgmmXz8HwSY/clF6gRoZeOUloCHsnrNgIHFbIB8VBAS7/AWU3tQV
Another registry that has tracked preeclampsia in In conclusion, pregnancy in women with kidney trans-
women with kidney transplants is the Transplant Preg- plantation confers a higher risk of complications includ-
nancy Registry International (TPRI, formerly known as ing hypertension, premature birth, reduced birth weight
the National Transplant Pregnancy Registry). Their 2022 of the newborn, and most importantly a higher risk of
annual report9 included 1417 female kidney transplant preeclampsia when compared with the general popula-
recipients with pregnancies in 2021–2022. Like the find- tion. These patients need close follow-up and manage-
ings by Lu et al., the preeclampsia rate was quite high, at ment by an expert multidisciplinary team to ensure the
30%, but not as high as the 2018–2021 rates found in best outcomes. The findings from Lu et al. provide much
the Australia and New Zealand Dialysis and Trans- needed data on this subject because it alerts us to the
plant Registry. higher incidence of preeclampsia diagnosis and demon-
N/mZSmEnSZyTuvQ0mJAo= on 06/13/2023
One of the notable findings of Lu et al. was the median strates that the risk was higher at creatinine levels that
preconception serum creatinine, at which preeclampsia many would have considered good function for a trans-
risk was significantly higher ($1.24 mg/dl [odds plant recipient. The findings surprise us with the indica-
ratio, 2.48; 95% confidence interval, 1.19 to 5.18]).7 tion that whatever other associated outcomes occur with
This observation illustrates that many women whom preeclampsia, worsening kidney transplant outcomes is
we might consider as having good kidney transplant not one of them, potentially making pregnancy in trans-
function are at higher risk of preeclampsia, including plant recipients a double jeopardy, not triple jeopardy
recipients of living donor kidneys—considered the best situation. However, this should not be interpreted to
quality kidneys. mean that preeclampsia in transplant recipients is a be-
We must also recognize the negative effects a preeclamp- nign process. The findings highlight the importance of
tic pregnancy has on the offspring. Preeclamptic pregnan- counseling for women who are considering pregnancy
cies have a higher risk of preterm delivery and lower birth after a kidney transplant. The study provides data to
weight of the infant. Children born after a pregnancy help us better understand which of our patients is at risk
complicated by preeclampsia have an average of 5% lower and to guide our clinical management decisions. This study
birth weight as compared with children born after an un- also underscores the need for further research in pregnancy
complicated pregnancy. This reduction in the birth weight complications after kidney transplant to advance our un-
of the newborn is even more prominent in women with derstanding of preeclampsia and improve patient care. A
pregnancies complicated by early-onset preeclampsia, promising direction for future research is exploring the role
who have on average a 23% lower birth weight than of predictive biomarkers to identify transplant recipients at
expected based on gestational age.6 Lu et al. found com- elevated risk of developing preeclampsia to optimize our
parable results because the median birth weight of babies management of this vulnerable subset of the population and
born to moms with preeclampsia was statistically lo- improve the outcomes for mother, baby, and the kidney
wer at 2330 g compared with 2800 g in babies without transplant in the short and long term.
preeclampsia.7 Sixty-four percent of babies diagnosed with
preeclampsia had birth weights ,2500 g compared with 37%
Disclosures
in babies born without a diagnosis of preeclampsia. The M.A. Josephson reports consultancy for Exosome Diagnostics,
median gestational age was 34 weeks in those with pre- IMMUCOR, Labcorp, Otsuka, UBC Pharmaceutical Support, and
eclampsia compared with 36 weeks in those without. These Vera Therapeutics services for the mycophenolate pregnancy
differences in gestational age and birth weight can have registry; ownership interest in Seagen; research funding from
profound effects on the offspring’s development for years to Bucksbaum Institute and Gift of Hope; honoraria from ASN:
come and consequently on the mother. Highlights and ASN Board Review Course; and an advisory or
One interesting avenue for future research into pre- leadership role for American Society of Nephrology. The re-
eclampsia and pregnancy in transplant patients is looking maining author has nothing to disclose.
at the placental soluble fms-like tyrosine kinase 1, which is
an antagonist of vascular endothelial growth factor and Funding
placental growth factor that gets upregulated in pre- None.
eclampsia and falls after delivery of the baby. The soluble
fms-like tyrosine kinase 1/placental growth factor ratio is
Acknowledgments
elevated in pregnant women 4–5 weeks before the clinical
The content of this article reflects the personal experience and
onset of preeclampsia, and it has been shown to be a views of the author(s) and should not be considered medical
reliable tool for predicting preeclampsia and discriminat- advice or recommendation. The content does not reflect the views
ing between different types of pregnancy-related hyper- or opinions of the American Society of Nephrology (ASN) or
tensive disorders.10 This biomarker has the potential to CJASN. Responsibility for the information and views expressed
be a valuable tool for clinicians in the future because herein lies entirely with the author(s).
CJASN 18: , July, 2023 The Complicated Interplay of Preeclampsia Risks, Al Sayyab and Josephson 3
References
plantation. 2013;95(7):908–915. doi:10.1097/tp.
1. Pregnancy and renal disease. Lancet. 1975;2(7939):801–802.
Downloaded from http://journals.lww.com/cjasn by ovbQKPXYfBFIRkaiUq5PuVRWaEDA1iTVOAm/wWdC7XRdvs6oSNs
0b013e318278d367
doi:10.1016/S0140-6736(75)80015-8
9. Moritz M, Constantinescu S, Coscia L. Annual report, Transplant
2. McKay DB, Adams P, Bumgardner G, et al. Reproduction and
Pregnancy Registry International (TPRI). 2023. Available at:
pregnancy in transplant recipients: current practices. Prog
https://www.transplantpregnancyregistry.org/publications-
Transplant. 2006;16(2):127–132. doi:10.7182/prtr.16.2.
collaborations/. Accessed May 8, 2023
j3324t64t6517u76
10. Ohkuchi A, Saito S, Yamamoto T, et al. Short-term prediction of
3. Shah S, Verma P. Overview of pregnancy in renal transplant pa-
preeclampsia using the sFlt-1/PlGF ratio: a subanalysis of
tients. Int J Nephrol. 2016;2016:1–7. doi:10.1155/2016/4539342
pregnant Japanese women from the PROGNOSIS Asia study.
4. Bramham K, Nelson-Piercy C, Gao H, et al. Pregnancy in renal
Hypertens Res. 2021;44(7):813–821. doi:10.1038/s41440-021-
transplant recipients: a UK national cohort study. Clin J Am Soc
00629-x
Nephrol. 2013;8(2):290–298. doi:10.2215/CJN.06170612
5. Kristensen JH. Pre-eclampsia and risk of later kidney disease:
nationwide cohort study. BMJ. 2019;365:l1516. doi:10. Published Online Ahead of Print: June 9, 2023
N/mZSmEnSZyTuvQ0mJAo= on 06/13/2023
1136/bmj.l1516
6. Bokslag A. Preeclampsia; short and long-term consequences for See related article, “Preeclampsia After Kidney Transplantation:
mother and neonate. Early Hum Dev. 2016;102:47–50. doi:10. Rates and Association with Graft Survival and Function,” on pages
1016/j.earlhumdev.2016.09.007 920–929.
Original Article
United States
Calyani Ganesan,1 Malorie Holmes,1 Sai Liu,1 Maria Montez-Rath,1 Simon Conti,2 Timothy C. Chang,2,3
Colin R. Lenihan,1 Xingxing S. Cheng ,1 Glenn M. Chertow ,1 John T. Leppert ,1,2,3 and Alan C. Pao1,2,3
Abstract
Background Kidney stone disease is common and can lead to complications such as AKI, urinary tract obstruction, 1
Division of
and urosepsis. In kidney transplant recipients, complications from kidney stone events can also lead to rejection Nephrology,
and allograft failure. There is limited information on the incidence of kidney stone events in transplant recipients. Department of
N/mZSmEnS6empR5LyiT0= on 06/13/2023
Medicine, Stanford
Methods We identified 83,535 patients from the United States Renal Data System who received their first kidney University, Stanford,
California
transplant between January 1, 2007, and December 31, 2018. We examined the incidence of kidney stone events 2
Department of
and identified risk factors associated with a kidney stone event in the first 3 years after transplantation. Urology, Stanford
University, Stanford,
Results We found 1436 patients (1.7%) who were diagnosed with a kidney stone in the 3 years after kidney California
3
transplant. The unadjusted incidence rate for a kidney stone event was 7.8 per 1000 person-years. The median Veterans Affairs Palo
Alto Health Care
time from transplant to a kidney stone diagnosis was 0.61 (25%–75% range 0.19–1.46) years. Patients with a System, Palo Alto,
history of kidney stones were at greatest risk of a kidney stone event after transplant (hazard ratio [HR], California
4.65; 95% confidence interval [CI], 3.82 to 5.65). Other notable risk factors included a diagnosis of gout
(HR, 1.53; 95% CI, 1.31 to 1.80), hypertension (HR, 1.29; 95% CI, 1.00 to 1.66), and a dialysis of vintage of $9 years Correspondence:
(HR, 1.48; 95% CI, 1.18 to 1.86; ref vintage #2.5 years). Dr. Calyani Ganesan,
Stanford University,
Division of
Conclusions Approximately 2% of kidney transplant recipients were diagnosed with a kidney stone in the 3 years Nephrology, 3180
after kidney transplant. Risk factors of a kidney stone event include a history of kidney stones and longer dialysis Porter Drive, Palo Alto,
vintage. CA 94304. Email:
CJASN 18: 777–784, 2023. doi: https://doi.org/10.2215/CJN.0000000000000176 [email protected]
www.cjasn.org Vol 18 June, 2023 Copyright © 2023 by the American Society of Nephrology 777
778 CJASN
interpretation of the US government. The USRDS contains counted once, at the time of his or her first qualification for
demographic, clinical, treatment, and survival information kidney stone events during the observation period. Patients
on patients who receive dialysis or undergo kidney trans- were censored at the end of the study (December 31, 2018),
plantation in the United States.6 These data are linked to loss of Medicare Parts A and B coverage, or 3 years after
Medicare Parts A and B records, which allow abstraction of transplant. In sensitivity analysis, graft failure and death
eAX6oP97/t+reHY6iBuDuEG8j/KLK4+srUZdWUQbtcBUzqkxhgmmXz8HwSY/clF6gRoZeOUloCHsnrNgIHFbIB8VBAS7/AWU3tQV
detailed inpatient and outpatient health care claims data for were treated as competing risks (see below). If a patient
Downloaded from http://journals.lww.com/cjasn by ovbQKPXYfBFIRkaiUq5PuVRWaEDA1iTVOAm/wWdC7XRdvs6oSNs
patients with Medicare coverage. Most US patients with received a diagnosis of a kidney stone during the same
kidney failure qualify for Medicare coverage after starting hospitalization as the kidney transplant, we considered the
dialysis or receiving kidney transplantation. Kidney trans- diagnosis of kidney stone to be a preexisting stone and not
plant recipients receive 3 years of Medicare coverage after an outcome.
successful transplantation and longer if they meet other
qualifications for Medicare coverage. Patient Characteristics
We assembled data on age, sex, race/ethnicity, cause of
Study Population kidney failure, body mass index (BMI), dialysis modality,
We used USRDS data to identify adult patients who dialysis vintage (time since initiation of dialysis in non-
received their first kidney transplant between January 1, preemptive transplantation), patient blood type, donor
N/mZSmEnS6empR5LyiT0= on 06/13/2023
2007, and December 31, 2018. Patients were included if they type (living and deceased), donor age, sex, HLA mismatch,
had uninterrupted Medicare Parts A and B for at least 6 calculated panel reactive antibody (at the time of trans-
months before kidney transplant and if they did not receive plant), cold ischemia time, induction immunosuppression,
multiorgan transplant (liver, lung, heart, or pancreas) within and maintenance immunosuppression at the time of trans-
2 years before kidney transplant. Figure 1 shows a flow plant. We identified comorbid illness with the inpatient
chart of patient cohort selection and exclusion. and outpatient ICD-9/ICD-10 claims-based algorithm as
described by Elixhauser et al.7 Comorbid illness was eval-
Kidney Stone after Kidney Transplantation uated within 2 years before kidney transplantation and
We identified kidney transplant recipients with a kidney included hyperparathyroidism (primary, secondary, and
stone event in the 3 years after kidney transplantation. We tertiary), gout, diabetes mellitus, hypertension, smoking,
defined patients with a kidney stone as those with one or coronary artery disease, cerebrovascular disease, obesity,
more inpatient encounters that included International Clas- alcohol dependence, lung disease, cancer, valvular disease,
sification of Diseases, Ninth Revision and Tenth Revision peripheral arterial disease, liver disease, and arrhythmias
(ICD-9 and ICD-10) codes for a kidney or ureteral stone, two (Supplemental Table 1). In addition, we determined health
or more outpatient encounters for a kidney or ureteral stone, care utilization in the 6 months before kidney transplant
or one or more kidney or ureteral stone procedure within 1 using the following metrics: skilled nursing facilities stay,
year using current procedural terminology (CPT), ICD-9 days spent in hospital, and number of non-nephrology out-
procedure, and ICD-10 procedure codes. Each patient was patient visits. We defined history of stones in patients if they
met one of the following criteria in the 2 years before kidney kidney transplant, we performed a sensitivity analysis in the
transplant or during the transplant hospitalization: (1) one subset of patients who had uninterrupted Medicare Part D
or more inpatient encounters that included a ICD-9 and coverage for at least 6 months before kidney transplant
ICD-10 code for a kidney or ureteral stone; (2) two or more (N546,488). For this analysis, patients were further
outpatient encounters for a kidney or ureteral stone; or (3) censored for loss of Medicare Part D. Given the relatively
eAX6oP97/t+reHY6iBuDuEG8j/KLK4+srUZdWUQbtcBUzqkxhgmmXz8HwSY/clF6gRoZeOUloCHsnrNgIHFbIB8VBAS7/AWU3tQV
one or more kidney or ureteral stone procedure within 1 high percentage of death and graft failure in comparison
Downloaded from http://journals.lww.com/cjasn by ovbQKPXYfBFIRkaiUq5PuVRWaEDA1iTVOAm/wWdC7XRdvs6oSNs
year using CPT, ICD-9 procedure, and ICD-10 proce- with stone events, we reanalyzed the data using subdistri-
dure codes. bution models to account for the competing risks of
death and graft failure.15,16 Final results were similar be-
tween the two analyses.
Missing Data
Of a total of 83,535 patients who received their first We conducted statistical analyses using SAS software,
kidney transplant between 2007 and 2018, 14% of patients version 9.4 (SAS institute, Inc., Cary, NC), and Stata MP,
(N512,067) had at least one covariate missing. We assumed version 17 (StataCorp, College Station, TX).
the data to be missing at random and used multiple impu-
tation with chained equations to impute 15 datasets.8,9 We
Results
included the exposure and all covariates in the analysis
N/mZSmEnS6empR5LyiT0= on 06/13/2023
Table 1. Baseline characteristics of first-time kidney transplant recipients in the United States, 2007–2018
Patient Characteristics Total Cohort (N583,535) Deceased Donor (n565,257) Living Donor (n513,212)
Race, n (%)
White 47,194 (56) 34,300 (53) 9748 (74)
Black 29,509 (35) 25,252 (39) 2585 (20)
Other 6832 (8) 5705 (9) 879 (7)
BMI at transplant (kg/m26SD), mean (SD) 28.3 (5) 28.4 (5) 27.9 (5)
Comorbidities, n (%)
Hyperparathyroidism 21,859 (26) 17,099 (26) 3449 (26)
Gout 6960 (8) 5443 (8) 1289 (10)
Kidney stones 1780 (2) 1333 (2) 334 (3)
Diabetes mellitus 39,168 (47) 29,639 (45) 5498 (42)
Hypertension 76,050 (91) 59,432 (91) 11,982 (91)
Coronary artery disease 27,401 (33) 21,516 (33) 4221 (32)
Valvular disease 10,415 (12) 8125 (12) 1681 (13)
N/mZSmEnS6empR5LyiT0= on 06/13/2023
BMI, body mass index; IL2 RA, IL-2 receptor antagonists; mTOR, mammalian target of rapamycin.
CJASN 18: 777–784, June, 2023 Kidney Stone Events after Kidney Transplant, Ganesan et al. 781
We next identified risk factors associated with a kidney and thiazide-type diuretics, or allopurinol and risk of a
stone event after transplant (Table 2). We found that a kidney stone event (Supplemental Table 3).
history of kidney stones in the recipient was strongly
associated with a kidney stone event after transplant (haz-
ard ratio [HR], 4.65; 95% confidence interval [CI], 3.82 to Discussion
eAX6oP97/t+reHY6iBuDuEG8j/KLK4+srUZdWUQbtcBUzqkxhgmmXz8HwSY/clF6gRoZeOUloCHsnrNgIHFbIB8VBAS7/AWU3tQV
5.65). We also found that patients with more than 9 years of In this cohort study, the incidence of kidney stone events
Downloaded from http://journals.lww.com/cjasn by ovbQKPXYfBFIRkaiUq5PuVRWaEDA1iTVOAm/wWdC7XRdvs6oSNs
dialysis vintage had nearly double the risk of a kidney in patients who received their first kidney transplant was
stone event compared with patients with vintage ,2.5 just under 2% in the 3 years after transplantation. The
years (HR, 1.48; 95% CI, 1.18 to 1.86). Other notable risk median time from transplant to a kidney stone diagnosis
factors for a kidney stone event after transplant included a was just over 7 months. Post-transplant kidney stone dis-
diagnosis of gout (HR, 1.53; 95% CI, 1.31 to 1.80) and ease was associated with a history of kidney stones, longer
hypertension (HR, 1.29; 95% CI, 1.00 to 1.66). In a com- dialysis vintage, gout, and hypertension.
panion analysis in which we examined only patients who The IR for kidney stone events in kidney transplant
were covered by Medicare Part D insurance, we found no patients from our study (7.8 per 1000 person-years) is higher
association between prescription of cinacalcet, thiazide than the IR for kidney stone events reported in an earlier
N/mZSmEnS6empR5LyiT0= on 06/13/2023
Table 2. Multivariable adjusted hazard ratios for incident kidney stone diagnosis in first-time kidney transplant patientsa
Variable Incident Kidney Stone, n (%) Adjusted Hazard Ratio (95% CI) P Value
and is associated with a history of stones, longer dialysis Methodology: Timothy C. Chang, Glenn M. Chertow, Calyani
vintage, and diagnosis of gout or hypertension. We propose Ganesan, Malorie Holmes, John T. Leppert, Sai Liu, Maria Montez-
that new kidney transplant recipients, particularly those Rath, Alan C. Pao.
with a history of kidney stones, should receive evaluation Project administration: John T. Leppert.
for risk factors of kidney stones, particularly in the first year Resources: John T. Leppert, Alan C. Pao.
eAX6oP97/t+reHY6iBuDuEG8j/KLK4+srUZdWUQbtcBUzqkxhgmmXz8HwSY/clF6gRoZeOUloCHsnrNgIHFbIB8VBAS7/AWU3tQV
after kidney transplantation. This approach could take the Writing – original draft: Glenn M. Chertow, Calyani Ganesan,
Downloaded from http://journals.lww.com/cjasn by ovbQKPXYfBFIRkaiUq5PuVRWaEDA1iTVOAm/wWdC7XRdvs6oSNs
form of metabolic testing with 24-hour urine collection, Maria Montez-Rath, Alan C. Pao.
dietary counseling, prescription of stone-related medica- Writing – review & editing: Xingxing S. Cheng, Glenn M. Chertow,
tions, and periodic imaging to lower the risk of kidney Simon Conti, Calyani Ganesan, Colin R. Lenihan, John T. Leppert,
stone disease and to prevent or ameliorate complications Maria Montez-Rath, Alan C. Pao.
thereof. We also contend that more focus should be directed
at medical or surgical treatment of post-transplant hyper- Data Sharing Statement
parathyroidism to prevent ensuing kidney stone and bone Data available through archived NIDDK Repository USRDS.
disease in this population.
Supplemental Material
This article contains the following supplemental material online
Disclosures
N/mZSmEnS6empR5LyiT0= on 06/13/2023
at http://links.lww.com/CJN/B762.
T.C. Chang reports employment with Veterans Affairs Palo Alto Supplemental Table 1. Claims-based algorithms and CPT codes
Health Care and stocks in Amazon, AMD, American Express, for kidney stone disease.
Apple, Disney, Facebook, Google, Intel, Microsoft, Nvidia, Tesla, Supplemental Table 2. Number of patients with a diagnosis of
Visa, and Walmart. T.C. Chang’s spouse reports employment with primary, secondary, and tertiary hyperparathyroidism.
TheMiilk. X.S. Cheng reports an advisory or leadership role for the Supplemental Table 3. Multivariable adjusted hazard ratios for
National Living Donor Assistance Center and research funding incident kidney stone diagnosis in kidney transplant patients with
from the American Heart Association and the National Institute of Medicare Part D.
Health. G.M. Chertow reports consultancy for Akebia, Ardelyx,
AstraZeneca, Gilead, Miromatrix, Reata, Sanifit, Unicycive, and
Vertex; ownership interest in Ardelyx, CloudCath, Durect, DxNow, References
Eliaz Therapeutics, Outset, Physiowave, PuraCath, Renibus, and 1. Scales CD Jr., Smith AC, Hanley JM, Saigal CS. Prevalence of
kidney stones in the United States. Eur Urol. 2012;62(1):160–
Unicycive; research funding from CSL Behring, NIAID, and
165. doi:10.1016/j.eururo.2012.03.052
NIDDK; advisory or leadership role for Board of Directors of Sat- 2. Tatapudi VS, Goldfarb DS. Differences in national and in-
ellite Healthcare and as the Co-Editor of Brenner & Rector’s The ternational guidelines regarding use of kidney stone formers as
Kidney (Elsevier); and DSMB service for Bayer, Gilead, Mineralys, living kidney donors. Curr Opin Nephrol Hypertens. 2019;28(2):
NIDDK, Palladio, and ReCor. C.R. Lenihan reports research 140–147. doi:10.1097/mnh.0000000000000480
3. Palmisano A, Gandolfini I, Delsante M, Cantarelli C, Fiaccadori
funding from Astellas and CareDx and honoraria from Veloxis. J.T. E, Cravedi P, et al. Acute kidney injury (AKI) before and after
Leppert reports employment with the VA Palo Alto Health Care kidney transplantation: causes, medical approach, and impli-
System and research funding from NIH/NCI and VA HSR&D. A.C. cations for the long-term outcomes. J Clin Med. 2021;10(7):1484.
Pao reports employment with the Veterans Affairs Palo Alto Health doi:10.3390/jcm10071484
4. Abbott KC, Schenkman N, Swanson SJ, Agodoa LY. Hospitalized
Care System, consultancy for Novome Biotechnologies, ownership
nephrolithiasis after renal transplantation in the United States.
interest in Novome Biotechnologies, and advisory or leadership Am J Transplant. 2003;3(4):465–470. doi:10.1034/j.1600-6143.
roles for the American Journal of Physiology-Renal Physiology, Frontiers 2003.00080.x
Physiology, and Physiological Reports. All remaining authors have 5. Wu DA, Robb ML, Forsythe JLR, Bradley C, Cairns J, Draper H, et al.
nothing to disclose. Recipient comorbidity and survival outcomes after kidney trans-
plantation: a UK-wide prospective cohort study. Transplantation.
2020;104(6):1246–1255. doi:10.1097/tp.0000000000002931
Funding 6. Johansen KL, Chertow GM, Foley RN, Weinhandl ED, Winkelmayer
This work was supported by the Department of Veterans Affairs WC, Wetmore JB. US renal data System 2020 annual data report:
Health Services Research and Development (I01 HX003091 to epidemiology of kidney disease in the United States. Am J Kidney
A.C.P. and J.T.L.) and the National Institutes of Diabetes and Di- Dis. 2021;77(4):A7–A8. doi:10.1053/j.ajkd.2021.01.002
7. Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity
gestive and Kidney Diseases (K23DK128651 to C.G.).
measures for use with administrative data. Med Care. 1998;36(1):
8–27. doi:10.1097/00005650-199801000-00004
Acknowledgments 8. van Buuren S, Boshuizen HC, Knook DL. Multiple imputation of
The contents do not represent the views of the US Department of missing blood pressure covariates in survival analysis. Stat Med.
Veterans Affairs or the US Government. 1999;18(6):681–694. doi:10.1002/(sici)1097-0258(19990330)
18:6,681::aid-sim71.3.0.co;2-r
Author Contributions 9. Van Buuren S, Brand JPL, Groothuis-Oudshoorn CGM, Rubin
DB. Fully conditional specification in multivariate imputation. J
Conceptualization: Xingxing S. Cheng, Calyani Ganesan, Malorie
Stat Comput Simul. 2006;76(12):1049–1064. doi:10.1080/
Holmes, Colin R. Lenihan, John T. Leppert, Alan C. Pao. 10629360600810434
Data curation: John T. Leppert, Sai Liu, Maria Montez-Rath, Alan 10. White IR, Royston P. Imputing missing covariate values for the Cox
C. Pao. model. Stat Med. 2009;28(15):1982–1998. doi:10.1002/sim.3618
Formal analysis: Calyani Ganesan, Malorie Holmes, John T. Lep- 11. Mislevy RJ, Little RJA, Rubin DB. Statistical analysis with missing
data. J Educ Stat. 1991;16(2):150–155. doi:10.2307/1165119
pert, Sai Liu, Maria Montez-Rath, Alan C. Pao. 12. Royston P, Parmar MK. Flexible parametric proportional-hazards and
Funding acquisition: John T. Leppert, Alan C. Pao. proportional-odds models for censored survival data, with applica-
Investigation: Calyani Ganesan, Malorie Holmes, John T. Leppert, tion to prognostic modelling and estimation of treatment effects. Stat
Alan C. Pao. Med. 2002;21(15):2175–2197. doi:10.1002/sim.1203
784 CJASN
13. Lambert PC, Royston P. Further development of flexible para- risk factors and clinical implications. Nephrol Dial Transplant.
metric models for survival analysis. Stata J. 2009;9(2):265–290. 2007;22(3):906–910. doi:10.1093/ndt/gfl714
doi:10.1177/1536867x0900900206 22. Zuckerman JM, Assimos DG. Hypocitraturia: pathophysiology
14. Heinze G, Wallisch C, Dunkler D. Variable selection - a review and medical management. Rev Urol. 2009;11(3):134–144.
and recommendations for the practicing statistician. Biometrical 23. Amin T, Coates PT, Barbara J, Hakendorf P, Karim N. Prevalence
J. 2018;60(3):431–449. doi:10.1002/bimj.201700067 of hypercalcaemia in a renal transplant population: a single
eAX6oP97/t+reHY6iBuDuEG8j/KLK4+srUZdWUQbtcBUzqkxhgmmXz8HwSY/clF6gRoZeOUloCHsnrNgIHFbIB8VBAS7/AWU3tQV
15. Lau B, Cole SR, Gange SJ. Competing risk regression models for centre study. Int J Nephrol. 2016;2016:7126290. doi:10.1155/
Downloaded from http://journals.lww.com/cjasn by ovbQKPXYfBFIRkaiUq5PuVRWaEDA1iTVOAm/wWdC7XRdvs6oSNs
the literature. J Endourol. 2012;26(1):38–44. doi:10.1089/end. calcium metabolism after kidney transplantation: a single-centre
2011.0049 study. Nephrol Dial Transplant. 2004;19(5):1281–1287. doi:10.
19. Kim IK, Tan JC, Lapasia J, Elihu A, Busque S, Melcher ML. In- 1093/ndt/gfh128
cidental kidney stones: a single center experience with kidney 27. Torres A, Lorenzo V, Salido E. Calcium metabolism and skeletal
donor selection. Clin Transplant. 2012;26(4):558–563. doi:10. problems after transplantation. J Am Soc Nephrol. 2002;13(2):
1111/j.1399-0012.2011.01567.x 551–558. doi:10.1681/ASN.V132551
20. Bolen E, Stern K, Humphreys M, Brady A, Leavitt T, Zhang N,
et al. Urine metabolic risk factors and outcomes of patients Received: October 24, 2022 Accepted: April 5, 2023
with kidney transplant nephrolithiasis. Clin Kidney J. 2021; Published Online Ahead of Print: April 18, 2023
15(3):500–506. doi:10.1093/ckj/sfab208
21. Keven K, Ozturk R, Sengul S, Kutlay S, Ergun I, Erturk S, et al. C.G. and M.H. share equal first authorship.
Renal tubular acidosis after kidney transplantation–incidence, J.T.L. and A.C.P. share equal senior authorship.
Editorial
1,2 2
Joyita Bharati and Kenar D. Jhaveri
CJASN 18: 699–701, 2023. doi: https://doi.org/10.2215/CJN.0000000000000171
Progression of IgA nephropathy varies on the basis of proteinuria at baseline; most were treated with immu-
clinical and histological variables at onset, response to nosuppressive agents. The eGFR decline (22.6 ml/min
treatment, and ethnicity. Evaluation of the prognosis of per 1.73 m2) in this group was statistically similar to the 1
Department of
IgA nephropathy has been recently simplified by an eGFR decline (22.1 ml/min per 1.73 m2) in the group Nephrology, Post
international risk-prediction tool that uses clinical and with subnephrotic proteinuria having a time-average Graduate Institute of
L2Xvf+Qaq0H8BCwjO6r8= on 06/13/2023
histological parameters to predict a 50% decline in eGFR proteinuria of 0.5 g/day. In a Chinese registry da- Medical Education and
or kidney failure.1 While this tool has been validated in tabase8 of 1155 patients with IgA nephropathy with a Research, Chandigarh,
India
multiethnic cohorts across the globe and is easily acces- median follow-up of 5.4 years, approximately 80% 2
Glomerular Center at
sible (www.qxmd.com), it cannot reliably predict kid- had CKD stage 1–2 at baseline, baseline proteinuria Northwell Health,
ney failure risk after 5–7 years from kidney biopsy was 0.89 g/d, and 45% had time-average protein- Division of Kidney
because of the limited follow-up duration of the deri- uria ,0.5 g/d during follow-up. The rate of kidney Diseases and
Hypertension,
vation cohorts. Given that IgA nephropathy predomi- function decline was 21.7 ml/min per 1.73 m2, and
Department of
nantly affects young adults, information on its course 20-year kidney survival was 67%. Patients with time- Medicine, Donald and
over the expected lifetimes of patients is crucial to un- average proteinuria ,0.5 g/d had the best kidney out- Barbara Zucker School
derstand the real story. Furthermore, after the success of comes, followed by 0.5–1 g/d and .1 g/d. The group of Medicine at Hofstra/
early phases of clinical trials testing novel therapies for with time-average proteinuria ,0.5 g/d had better out- Northwell, Great
Neck, New York
IgA nephropathy,2 more than a dozen randomized comes than that with 0.5–1 g/d.8
clinical trials (RCTs) are in progress or are in the pipeline Pitcher and colleagues9 reported on the long-term
Correspondence:
for the near future. Because RCTs are planned for the (over expected lifetimes) outcomes of IgA nephropathy Dr. Kenar D. Jhaveri,
short term, evaluating the ability of short-term surrogate in a predominantly White cohort in this issue of CJASN. Glomerular Center at
end points to predict disease progression over one’s Both adults and children (5.7%) were included, and the Northwell Health,
lifetime seems an interesting unmet need. median follow-up was 5.9 years. The median eGFR was Division of Kidney
Diseases and
Proteinuria has been frequently used as a surrogate lower (47.6 ml/min per 1.73 m2) than in previous similar Hypertension,
end point in observational studies and clinical trials of studies; most patients (approximately 65%) had CKD Department of
IgA nephropathy. Our current perception of IgA ne- stage 3 or below at baseline. Baseline proteinuria was Medicine, Donald and
phropathy is that in most patients, this is a relatively 1.72 g/d. Kidney failure or death occurred in approx- Barbara Zucker School
of Medicine at Hofstra/
slowly progressive disease. A slow and steady ap- imately 50% of patients. The mean eGFR decline during
Northwell, Great
proach with a liberal target of proteinuria reduction follow-up was 23.6 ml/min per 1.73 m2. CKD stage Neck, New York, NY.
to ,1 g/d3 is also perceived to be associated with was the strongest baseline predictor of eGFR slope and Email: kjhaveri@
favorable outcomes. However, we also know that 10-year kidney survival. Interestingly, time-average northwell.edu
higher baseline proteinuria is known to be associated proteinuria over the first 2 years and the median
with a higher risk of adverse kidney outcomes during follow-up, respectively, were noted to be a significant
the long term.4–6 Time-average proteinuria during predictor of kidney failure, irrespective of baseline pro-
follow-up is increasingly recognized to be more rele- teinuria and after adjusting for baseline CKD stage.
vant in predicting disease progression than baseline Patients with time-average proteinuria .0.88 g/g
proteinuria. In a single-center study of 130 patients in had a higher risk of progression to kidney failure or
Malaysia7 having a baseline CKD stage 1–2 in 64% death than those with ,0.88 g/g. Of note, even a time-
and a baseline proteinuria of 2.4 g/d (approximately average proteinuria of ,0.88 g/g (i.e., protein excretion
44% having nephrotic-range proteinuria), time- of ,1 g/d), which is considered a low risk for disease
average proteinuria was the only factor associated progression,3 was associated with significantly lowered
with a 50% decline in eGFR or kidney failure during kidney survival (almost reduced by 50%) when fol-
follow-up. Time-average proteinuria ,0.5 g/d was lowed over 15 years. Furthermore, time-average pro-
noted in only 28% of patients. The difference in kidney teinuria ,0.44 g/g was observed in 24% of patients, and
survival between 0.5 g/d and 0.5–1 g/d groups was this cutoff of time-average proteinuria was associated
noted after 20 years of follow-up on Kaplan–Meier with a slower eGFR decline (20.3 ml/min per 1.73 m2)
curves. The annual decline of eGFR was –2 ml/min than that of 0.44 to ,0.88 g/g (21.6 ml/min per
per 1.73 m2 in this cohort, and 30-year kidney survival 1.73 m2). What is striking to note is that almost all
was 65%. Of note, time-average proteinuria was re- patients were observed to be at risk of kidney failure
duced to 1.26 g/d in patients with nephrotic-range within their expected lifetime unless eGFR decline was
www.cjasn.org Vol 18 June, 2023 Copyright © 2023 by the American Society of Nephrology 699
700 CJASN
eAX6oP97/t+reHY6iBuDuEG8j/KLK4+srUZdWUQbtcBUzqkxhgmmXz8HwSY/clF6gRoZeOUloCHsnrNgIHFbIB8VBAS7/AWU3tVc
Downloaded from http://journals.lww.com/cjasn by ovbQKPXYfBFIRkaiUq5PuVRWaEDA1iTVOAm/wWdC7XRdvs6oSNs
L2Xvf+Qaq0H8BCwjO6r8= on 06/13/2023
Figure 1. Choosing a novel aggressive approach over the conventional liberal approach in treating IgA nephropathy might avoid kidney failure in
patient’s expected lifetime. *In clinical trials, not approved yet. ACEi, angiotensin‐converting enzyme inhibitor; APRIL, a proliferation-inducing
ligand; ARB, angiotensin receptor blocker; MMF, mycophenolate mofetil; SGLT2i, sodium‐glucose cotransporter 2 inhibitor.
maintained at ,1 ml/min per 1.73 m2. In the high-risk prognostication and promising novel agents to treat the
(proteinuria .1 g/d, eGFR 30 ml/min per m2) RCT rep- disease, we surely are in the right place at the right time
resentative population, even an 80% reduction in protein- to aim for zero kidney failures from IgA nephropathy. IgA
uria from baseline was not associated with an eGFR slope nephropathy may not be that slow progressive disease we
of ,1 ml/min per 1.73 m2, implying the need for a more once believed it was. Treating aggressively up front may
aggressive target of proteinuria reduction to achieve such matter in our patients. Further work on applying the com-
an eGFR slope target. Missing information on treatment, ponents of the tools, i.e., clinical and biopsy variables, to
BP control, and kidney biopsy findings limits drawing any decide on one or more agents from the therapeutic arma-
inferences from this study for guiding patient care. More- mentarium is necessary to lead us to the road toward
over, generalizability is limited because the study included our goal.
mainly a White population. Most importantly, it highlights
the long-term risks associated with the perceived ideal Disclosures
target (,1 g/d proteinuria). Stricter proteinuria control K.D. Jhaveri is a founder and co-president of the American So-
to ,0.5 g/d seemed to improve kidney survival. In addi- ciety of Onco-Nephrology, serves as the US co-national lead for the
tion, almost all patients are at risk of kidney failure within IgA Nephropathy trial—VISIONARY, and reports employment
expected lifetimes unless an eGFR slope of ,1 ml/min per with Northwell Health. K.D. Jhaveri reports consultancy agree-
1.73 m2 is sustained. While these observations are insight- ments with Calliditas, George Clinical, PMV Pharmaceuticals, and
ful, the target eGFR slope seems arduous to achieve with Secretome; reports honoraria from the American Society of Ne-
the current patient care approach. However, aiming to get phrology and UpToDate.com; reports serving on the editorial
closer to such a target eGFR slope with modifications in the boards of the American Journal of Kidney Diseases, CJASN, the Clinical
current approach, such as better proteinuria and BP control Kidney Journal, the Journal of Onconephrology, Kidney International,
with therapies targeting specific pathogenic pathways, can and Nephrology Dialysis Transplantation; and reports serving as
improve patient outcomes (Figure 1). Editor-in-Chief of ASN Kidney News and section editor for onco-
What modifications must be made in the management nephrology for Nephrology Dialysis Transplantation. The remaining
approach of IgA nephropathy to reach a stringent eGFR author has nothing to disclose.
slope target of approximately 1 ml/min per 1.73 m2? Will
treating the typical patient with IgA nephropathy with tar- Funding
geted therapies early on, instead of waiting 3–6 months to None.
optimize nonspecific supportive therapies,3 move us toward
meeting the target? This leads to the next question, “which Acknowledgments
factors should decide early treatment?” and the following The content of this article reflects the personal experience and
question, “which agent(s) should be used for a particular views of the author(s) and should not be considered medical advice
patient?” With the background of handy tools for or recommendation. The content does not reflect the views or
CJASN 18: 699–701, June, 2023 Long-Term Follow-Up of IgA Nephropathy, Bharati and Jhaveri 701
opinions of the American Society of Nephrology (ASN) or CJASN. review of the literature. Medicine. 1994;73(2):79–102. doi:10.
Responsibility for the information and views expressed herein lies 1097/00005792-199403000-00002
entirely with the author(s). 5. Maixnerova D, Bauerova L, Skibova J, et al. The retrospective
analysis of 343 Czech patients with IgA nephropathy–one centre
experience. Nephrol Dial Transplant. 2012;27(4):1492–1498. doi:
Author Contributions 10.1093/ndt/gfr482
eAX6oP97/t+reHY6iBuDuEG8j/KLK4+srUZdWUQbtcBUzqkxhgmmXz8HwSY/clF6gRoZeOUloCHsnrNgIHFbIB8VBAS7/AWU3tVc
Conceptualization: Kenar D. Jhaveri. 6. Moriyama T, Tanaka K, Iwasaki C, et al. Prognosis in IgA ne-
Downloaded from http://journals.lww.com/cjasn by ovbQKPXYfBFIRkaiUq5PuVRWaEDA1iTVOAm/wWdC7XRdvs6oSNs
Data curation: Joyita Bharati, Kenar D. Jhaveri. phropathy: 30-year analysis of 1,012 patients at a single center in
Formal analysis: Kenar D. Jhaveri. Japan. PLoS One. 2014;9(3):e91756. doi:10.1371/journal.
Investigation: Kenar D. Jhaveri. pone.0091756
7. Mohd R, Mohammad Kazmin NE, Abdul Cader R, et al. Long term
outcome of immunoglobulin A (IgA) nephropathy: a single center
References experience. PLoS One. 2021;16(4):e0249592. doi:10.1371/
1. Barbour SJ, Coppo R, Zhang H, et al. Evaluating a new international journal.pone.0249592
risk-prediction tool in IgA nephropathy. JAMA Intern Med. 2019; 8. Le W, Liang S, Hu Y, et al. Long-term renal survival and related risk
179(7):942–952. doi:10.1001/jamainternmed.2019.0600. factors in patients with IgA nephropathy: results from a cohort of
2. Barratt J, Hour B, Kooienga L, et al. POS-109 Interim results of phase 1155 cases in a Chinese adult population. Nephrol Dial Trans-
1 and 2 trials to investigate the safety, tolerability, pharmacokinetics, plant. 2012;27(4):1479–1485. doi:10.1093/ndt/gfr527
pharmacodynamics, and clinical activity of BION-1301 in patients 9. Pitcher D, Braddon F, Hendry B, et al. Long-term outcomes in IgA
nephropathy. Clin J Am Soc Nephrol. 2023;18(6):727–738. doi:
L2Xvf+Qaq0H8BCwjO6r8= on 06/13/2023
www.cjasn.org Vol 18 June, 2023 Copyright © 2023 by the American Society of Nephrology 813
814 CJASN
eAX6oP97/t+reHY6iBuDuEG8j/KLK4+srUZdWUQbtcBUzqkxhgmmXz8HwSY/clF6gRoZeOUloCHsnrNgIHFbIB8VBAS7/AWU3tQV
Downloaded from http://journals.lww.com/cjasn by ovbQKPXYfBFIRkaiUq5PuVRWaEDA1iTVOAm/wWdC7XRdvs6oSNs
N/mZSmEnS6empR5LyiT0= on 06/13/2023
Figure 1. Selected images from diagnostic testing. (A) Representative slice from computed tomography of the chest showing randomly
distributed cavitary nodules compatible with septic emboli (arrowheads). (B) Apical four-chamber view from transthoracic echocardiogram
demonstrating bulky tricuspid vegetations. (C) Glomerulus showing mild mesangial hypercellularity, endocapillary hypercellularity, and a
cellular crescent (Periodic acid-Schiff, original magnification, 3400). (D) Glomerulus showing large cellular crescent (Jones methenamine
silver; original magnification, 3400).
Author Contributions
in IRGN failed to show significant improvement in out-
Conceptualization: Pietro A. Canetta.
comes, with the caveats that the trial was underpowered
Writing – original draft: Pietro A. Canetta.
and did not include patients with endocarditis.11 It is dif-
Writing – review & editing: Pietro A. Canetta.
ficult to extrapolate from such data a clear picture of which
patients might benefit from immunosuppression.
References
1. Luque Y, Louis K, Jouanneau C, et al. Vancomycin-associated
cast nephropathy. J Am Soc Nephrol. 2017;28(6):1723–1728.
Case—Conclusion doi:10.1681/ASN.2016080867
Because of persistent worsening kidney function despite 2. McGrath MM, Isakova T, Rennke HG, Mottola AM, Laliberte KA,
clearance of blood cultures and continued antibiotics, and Niles JL. Contaminated cocaine and antineutrophil cytoplasmic
after interdisciplinary discussion with the cardiology and antibody-associated disease. Clin J Am Soc Nephrol. 2011;6(12):
2799–2805. doi:10.2215/CJN.03440411
infectious disease consultants, the patient was started on
3. Boils CL, Nasr SH, Walker PD, Couser WG, Larsen CP. Update on
high-dose glucocorticoids. Methylprednisolone 500 mg was endocarditis-associated glomerulonephritis. Kidney Int. 2015;
given daily for 3 days, followed by prednisone 60 mg daily 87(6):1241–1249. doi:10.1038/ki.2014.424
for 1 week, then tapering. Immediately after glucocorticoid 4. Majumdar A, Chowdhary S, Ferreira MA, et al. Renal patho-
initiation, the creatinine began decreasing. He was dis- logical findings in infective endocarditis. Nephrol Dial Trans-
plant. 2000;15(11):1782–1787. doi:10.1093/ndt/15.11.1782
charged on hospital day 51, nearly 2 weeks into the gluco- 5. Guo S, Kapp ME, Beltran DM, et al. Spectrum of kidney diseases
corticoid taper, with a creatinine of 1.22 mg/dl. He did not in patients with hepatitis C virus infection. Am J Clin Pathol.
follow up at a scheduled postdischarge visit with our ne- 2021;156(3):399–408. doi:10.1093/ajcp/aqaa238
phrology clinic, but he was briefly in the emergency de- 6. Cunha BA, Gill MV, Lazar JM. Acute infective endocarditis.
Diagnostic and therapeutic approach. Infect Dis Clin North Am.
partment for unrelated reasons 1 month later, and creatinine
1996;10(4):811–834. doi:10.1016/s0891-5520(05)70328-7
at that time was stable. 7. Satoskar AA, Parikh SV, Nadasdy T. Epidemiology, pathogenesis,
treatment and outcomes of infection-associated glomerulone-
phritis. Nat Rev Nephrol. 2020;16(1):32–50. doi:10.1038/
s41581-019-0178-8
Summary 8. Langlois V, Lesourd A, Girszyn N, et al. Antineutrophil cyto-
Glomerulonephritis associated with infective endocarditis plasmic antibodies associated with infective endocarditis.
is a subtype of IRGN characterized by distinct clinical features Medicine (Baltimore). 2016;95(3):e2564. doi:10.1097/md.
and patterns of kidney injury. Despite the frequently crescentic 0000000000002564
pathology, ANCA are only detected in a minority of patients. 9. Brunet A, Julien G, Cros A, et al. Vasculitides and glomerulone-
phritis associated with Staphylocococcus aureus infective endo-
Outcomes are generally poor; the role of additional immuno- carditis: cases reports and mini-review of the literature. Ann Med.
suppression remains undefined but may be considered in 2020;52(6):265–274. doi:10.1080/07853890.2020.1778778
select cases who worsen despite appropriate infection control. 10. Le Moing V, Lacassin F, Delahousse M, et al. Use of cortico-
steroids in glomerulonephritis related to infective endocarditis:
three cases and review. Clin Infect Dis. 1999;28(5):1057–1061.
Disclosures doi:10.1086/514734
P.A. Canetta reports consultancy for Chinook, Novartis, Otsuka, 11. Arivazhagan S, Lamech TM, Myvizhiselvi M, et al. Efficacy of
and Travere and research funding from Calliditas, Novartis, corticosteroids in infection-related glomerulonephritis—a ran-
and Travere. domized controlled trial. Kidney Int Rep. 2022;7(10):2160–2165.
doi:10.1016/j.ekir.2022.07.163
Funding
None. Published Online Ahead of Print: March 29, 2023
Critical Care Nephrology
and Acute Kidney Injury
Abstract
While the administration of intravenous fluids remains an important treatment, the negative consequences of
subsequent fluid overload have raised questions about when and how clinicians should pursue avenues of fluid
1
removal. Decisions regarding fluid removal during critical illness are complex even for patients with preserved Renal, Electrolyte and
kidney function. This article seeks to apply general concepts of fluid management to the care of patients who also Hypertension Division,
Perelman School of
require KRT. Because optimal fluid management for any specific patient is likely to change over the course of
L2Xvf+Qaq1MC/lwRu920= on 06/13/2023
Medicine, University
critical illness, conceptual models using phases of care have been developed. In this review, we will examine the of Pennsylvania,
implications of one such model on the use of ultrafiltration during KRT for volume removal in distributive shock. Philadelphia,
This will also provide a useful lens to re-examine published data of KRT during critical illness. We will highlight Pennsylvania
2
Center for Clinical
recent prospective trials of KRT as well as recent retrospective studies examining ultrafiltration rate and Epidemiology and
mortality, review the results, and discuss applications and shortcomings of these studies. We also emphasize that Biostatistics, Perelman
current data and techniques suggest that optimal guidelines will not consist of recommendations for or against School of Medicine,
absolute fluid removal rates but will instead require the development of dynamic protocols involving frequent University of
cycles of reassessment and adjustment of net fluid removal goals. If optimal fluid management is dynamic, then Pennsylvania,
Philadelphia,
frequent assessment of fluid responsiveness, fluid toxicity, and tolerance of fluid removal will be needed. Pennsylvania
Innovations in our ability to assess these parameters may improve our management of ultrafiltration in the future. 3
Department of
CJASN 18: 788–802, 2023. doi: https://doi.org/10.2215/CJN.0000000000000164 Anesthesiology and
Critical Care, Perelman
School of Medicine,
University of
Introduction conceptual framework to guide the administration Pennsylvania,
KRT plays a fundamental role in the management of and removal of fluid during critical illness.11 The Philadelphia,
fluid balance in the critically ill. However, management model consists of four phases: Resuscitation, Optimi- Pennsylvania
4
is complex and often particularly challenging in the zation, Stabilization, and De-escalation (ROS-D).11 The Pulmonary, Allergy,
setting of distributive shock. Intravenous fluids are a ROS-D paradigm seeks to move beyond labels of con-
and Critical Care
Division, Perelman
cornerstone of resuscitation.1,2 However, the hemody- servative versus liberal fluid management and recog- School of Medicine at
namic effect of volume expansion is often short-lived, nizes that the optimal strategy during one phase of the University of
and many clinicians continue to reflexively infuse fluids illness may be unhelpful or even dangerous during Pennsylvania,
in absence of evidence of fluid responsiveness.2,3 Critical another.12,13 This context-dependent approach has
Philadelphia,
illness itself contributes to retention of fluids through Pennsylvania
some important implications in the management of
multiple insults to the integrity of the vascular endothe-
patients receiving KRT (Figure 1). Correspondence:
lium.4 The resultant fluid overload has increasingly been
During the rescue phase, a large amount of intrave- Dr. Dan Negoianu,
recognized as a complication of critical illness and is
nous fluid is given immediately after the recognition of University of
associated with higher morbidity and mortality.4–9 How- Pennsylvania, Renal,
distributive or hypovolemic shock. Fluid is typically Electrolyte and
ever, studies that have examined this relationship are
given rapidly, often according to a weight-based pro- Hypertension Division,
highly susceptible to confounding even when adjusting
tocol such as the Surviving Sepsis Campaign recom- 3400 Spruce Street, 1
for severity of illness. Given that many patients who Founders Pavilion,
develop fluid overload also remain hypotensive, deci- mendation of administering 30 cc/kg of isotonic fluid
Philadelphia, PA
sions regarding optimal fluid removal remain complex.10 within the first 3 hours of resuscitation.14 The goal is to 19104. Email: dan.
In the context of such uncertainty, what paradigms give fluid quickly rather than to give incremental test negoianu@
doses. Fluid removal using KRT is unlikely to have a pennmedicine.upenn.
should we use to guide decisions regarding fluid re- edu
moval using KRT in the intensive care unit (ICU)? In role in this phase.
this review, we describe KRT fluid management mod- In the optimization phase, intravenous fluid is given
els that are broadly applicable to most ICU patients more judiciously. Volume management is guided by
(although management of patients with primarily car- evidence of improved tissue perfusion after fluid chal-
diogenic shock is outside the scope of this review). lenges. While fluid removal during a volume challenge
would clearly be counterproductive, these fluid chal-
lenges may contribute to fluid accumulation that will
Phases of Fluid Management eventually need to be addressed.
A working group of nephrologists and intensivists In the stabilization phase, tissue perfusion is no
in the Acute Dialysis Quality Initiative proposed a longer fluid-responsive, so a net positive state is not
788 Copyright © 2023 by the American Society of Nephrology www.cjasn.org Vol 18 June, 2023
CJASN 18: 788–802, June, 2023 Volume Management with KRT in Critical Illness, Wang et al. 789
eAX6oP97/t+reHY6iBuDuEG8j/KLK4+srUZdWUQbtcBUzqkxhgmmXz8HwSY/clF6gRoZeOUloCHsnrNgIHFbIB8VBAS7/AWU3tVc
Downloaded from http://journals.lww.com/cjasn by ovbQKPXYfBFIRkaiUq5PuVRWaEDA1iTVOAm/wWdC7XRdvs6oSNs
L2Xvf+Qaq1MC/lwRu920= on 06/13/2023
Figure 1. Summary of the different phases of resuscitation and patient volume status. Included are clinical features of each phase, general
fluid strategy, some potential metrics to guide fluid delivery, and considerations for fluid removal. Image adapted from ADQI (2013).
desirable. However, significant net fluid removal should resource-intensive KRT modalities such as frequent inter-
also be avoided because patients in this phase are not mittent treatments, prolonged intermittent therapy,16 or
sufficiently hemodynamically stable to tolerate it. A net continuous KRT (CKRT) may be beneficial in achieving
even goal is therefore typical. While this goal might seem volume management goals.
simple, it is often difficult to achieve. Even if no resuscitative Finally, during the de-escalation phase (also known as
fluid is given, ICU patients frequently receive significant deresuscitation), hemodynamics have improved suffi-
amounts of fluid through medications and nutrition. In one ciently to tolerate the removal of fluid that has accumu-
large single-center cohort of ICU patients, medications and lated during earlier phases. The goal of this phase is to
nutrition accounted for 66% of total fluid intake.15 Patients return the patient to a state of minimized tissue edema, a
in the stabilization phase may intermittently decompensate, process often referred to as decongestion. KRT becomes
leading to further intravenous fluid administration. As a essential for patients who cannot be decongested by other
result, fluid removal using KRT may become critical during means. While patients in this phase have improved suffi-
stabilization to prevent ever-worsening venous congestion. ciently to tolerate net fluid removal, they have not neces-
In patients with large amounts of daily fluid intake, more sarily returned to their baseline hemodynamic state. A
790 CJASN
potential pitfall in such patients is the temptation to fact, supplemental data from the IDEAL-ICU trial showed
transition away from resource-intensive therapies too that participants in the early arm had a median daily
early and thereby undermine the achievement of decon- fluid removal by KRT of 0 ml for each of the 7 days after
gestion.17 In short, clinicians may have to choose between randomization. An exception was the STARRT-AKI trial,
de-escalating fluid congestion versus de-escalating resource- where a prespecified secondary analysis of participants with
eAX6oP97/t+reHY6iBuDuEG8j/KLK4+srUZdWUQbtcBUzqkxhgmmXz8HwSY/clF6gRoZeOUloCHsnrNgIHFbIB8VBAS7/AWU3tVc
intensive KRT modalities. fluid balance data collected showed a 1.1 L (P 5 0.03) dif-
Downloaded from http://journals.lww.com/cjasn by ovbQKPXYfBFIRkaiUq5PuVRWaEDA1iTVOAm/wWdC7XRdvs6oSNs
methods are invasive and can serve as a portal for infection. between the treatment arms. In fact, only five of ten trials
Point-of-care ultrasound provides a noninvasive technique comparing CKRT with intermittent KRT even quantified net
to estimate stroke volume by measuring aortic valve area fluid balance.40–49 Of these five, four showed no between-
and flow20 but requires significant expertise. As a result, in group differences,42–44,46 and one reported greater net neg-
clinical settings, an increase in systolic BP, pulse pressure, ative fluid balance in the CKRT group at 72 hours but no
or mean arterial pressure (MAP) immediately after fluid fluid balance data after that.40 A study by Mehta et al. did
administration is often used as a metric fluid of responsive- not report net fluid balance but did note that 28.8% of
ness instead of change in cardiac output. However, the cor- intermittent KRT treatments fell short of ultrafiltration goals,
relation between these metrics and cardiac output after fluid compared with just 9% of treatments in the CKRT arm.41
challenge has been inconsistent across studies.21–23 Finding In the meantime, a lack of evidence is not evidence of
better predictors of fluid responsiveness has been the subject lack of effect. Although useful trial data are absent, en-
of several recent reviews,24 including from the CJASN Critical hanced fluid removal using CKRT as compared with in-
Care Nephrology and Acute Kidney Injury Series.25 Unfor- termittent KRT is both physiologically reasonable and has
tunately, many of these metrics require invasive measure- some support, albeit limited, from cross-sectional data.50
ments or ultrasound expertise or have only been validated in a
subset of patients undergoing mechanical ventilation with
specific respiratory system mechanics—typically deep seda- Optimizing Randomized Trial Design
tion and/or relatively normal lung compliance.4,26–28 In designing trials of fluid management in patients
Despite these limitations, assessing fluid status and re- not requiring KRT, intensivists have suggested trialing
sponsiveness through a multimodal approach means conservative fluid management or goal-directed fluid re-
remains a core part of the everyday care of critically ill moval during the stabilization and de-escalation phases,
patients.24,25 While it may be challenging to define the during which fluid removal may be both feasible and
management phase of individual patients with certainty, impactful.13,51 Future trials of fluid removal using KRT
fluid paradigms such as ROS-D may nevertheless provide might also focus on such later phases of management rather
a useful lens in evaluating current studies of KRT and their than the period immediately after AKI.
ability to inform optimal fluid removal in the critically ill. In addition, trials of volume management using
diuretics—such as one in patients with acute respiratory
distress syndrome52—used detailed protocols to achieve
Limitations of Trials of KRT Initiation differences in net volume status between treatment arms.
Several randomized clinical trials (RCTs) have examined Future trials of volume management with KRT will likely
the timing of initiation of KRT at the onset of critical illness need such protocols to guide the initiation, management,
(Table 1).29–37 However, such studies may not be ideal to and discontinuation of therapy. Given the uncertainty
answer questions about fluid removal. By design, these trials around these aspects of study design, pilot studies may
focused on initiation of KRT shortly after the development of be needed before trials can be attempted on a larger scale.
AKI. In the critical care setting, the first 24 hours of AKI
development often coincides with hemodynamic instability.
Trials examining early initiation of KRT may therefore be Mixed Results from Observational Studies of
likely to include patients in phases of management where Ultrafiltration Rate
volume removal is not yet desirable. While this hypothesis is If currently available RCTs are not sufficient to guide
impossible to confirm, it is notable that across all trials where optimal KRT fluid strategies, what about cohort data?
the cause of KRT initiation was reported, only a minority of Four recent retrospective studies examined whether the
participants had KRT initiated for volume management. In rate of ultrafiltration in critically ill patients receiving
addition, in all but one trial (Standard versus Accelerated KRT is associated with mortality (Table 2).53–56 Because
Initiation of Renal-Replacement Therapy in Acute Kidney KRT requires some administration of intravenous fluid
Injury [STARRT-AKI]), there was no significant difference in (to prime the circuit, rinse back blood, and replace fluid if
fluid balance reported between the randomized arms. In hemofiltration is used), these studies refer to fluid removal
L2Xvf+Qaq1MC/lwRu920= on 06/13/2023
eAX6oP97/t+reHY6iBuDuEG8j/KLK4+srUZdWUQbtcBUzqkxhgmmXz8HwSY/clF6gRoZeOUloCHsnrNgIHFbIB8VBAS7/AWU3tVc
Downloaded from http://journals.lww.com/cjasn by ovbQKPXYfBFIRkaiUq5PuVRWaEDA1iTVOAm/wWdC7XRdvs6oSNs
Median Time
KRT Initiated
Mortality Difference from Randomization Cumulative Fluid
KRT Criteria for Delayed for Volume (%)
Trial Patients Indication for Trial Enrollment to KRT Balance during
Modality KRT Initiation
Trial Period
Follow-up (d) Benefit Early KRT Late KRT Late KRT
HEROICS29 224 CVVH High-dose catecholamine 1. AKIN stage III AKI 30 No Unk Unk 0 Unknown
infusion after cardiac 2. BUN .100 mg/dl
surgery or extracorporeal 3. Life-threatening
membrane oxygenation hyperkalemia
requirement
ELAIN30 231 CVVHDF 1. KDIGO stage II AKI 1. BUN .100 mg/dl 90 Yes 6h 25.5 h Unk No significant
2. NGAL .150 ng/ml 2. Potassium .6 mEq/L w/ difference in
3. Severe sepsis EKG changes cumulative fluid
4. Use of vasopressors 3. Magnesium .4 mmol/L balance between
5. PaO2:FiO2 ,300 mm Hg 4. Urine output ,200 ml/12 h groups 3 d after
6. Fluid balance .10% of ABW 5. Organ edema resistant to randomization
loop diuresis
AKIKI31 619 Any KDIGO stage III AKI and either 1. Oliguria .72 h after 28 No 2h 57 h 6 Unknown
VDRF or vasopressor use randomization
2. BUN .112 mg/dl
3. Potassium .5.5 mmol/L
despite treatment
4. pH ,7.15
5. Pulmonary edema due to
fluid overload, refractory to
diuretics, requiring .5 L/
min of supplemental oxygen
or .50% FiO2
IDEAL-ICU37 488 Any RIFLE stage F AKI 1. Potassium .6.5 mmol/L 90 No 7.6 h 51.5 h 2 No significant
2. pH ,7.15 difference in
791
L2Xvf+Qaq1MC/lwRu920= on 06/13/2023
eAX6oP97/t+reHY6iBuDuEG8j/KLK4+srUZdWUQbtcBUzqkxhgmmXz8HwSY/clF6gRoZeOUloCHsnrNgIHFbIB8VBAS7/AWU3tVc
Downloaded from http://journals.lww.com/cjasn by ovbQKPXYfBFIRkaiUq5PuVRWaEDA1iTVOAm/wWdC7XRdvs6oSNs
792
Table 1. (Continued)
CJASN
Median Time
KRT Initiated
Mortality Difference from Randomization Cumulative Fluid
KRT Criteria for Delayed for Volume (%)
Trial Patients Indication for Trial Enrollment to KRT Balance during
Modality KRT Initiation
Trial Period
Follow-up (d) Benefit Early KRT Late KRT Late KRT
STARRT-AKI33 3019 Any KDIGO stage II AKI or urine 1. Potassium .6.0 mmol/L 90 No 6.1 h 31.1 h 43.6 1.1 L difference in
output ,6 ml/kg for 12 h 2. pH #7.20 fluid balance at
3. Serum bicarbonate #12 14 d (P 5 0.03)
mmol/L
4. Fluid overload with PaO2/
FiO2 #200 mm Hg
5. Persistent AKI 72 h after
randomization
AKIKI-234 278 Any KDIGO stage III AKI and either 1. BUN .140 mg/dl 28 No 3h 33 h 10 No significant
ventilatory-dependent 2. Potassium .5.5 mmol/L difference in
respiratory failure or despite treatment cumulative fluid
vasopressor use with 3. pH ,7.15 balance between
1. Oligoanuria .72 h or 4. Pulmonary edema due to groups either 2
2. BUN .112 mg/dl fluid overload, refractory to or 7 d after
diuretics, requiring .5 L/ randomization
min of supplemental oxygen
or .50% FiO2
HEROICS, Hemofiltration to Rescue Severe Shock following Cardiac Surgery; CVVH, continuous venovenous hemofiltration; AKIN, Acute Kidney Injury Network; Unk, unknown; ELAIN, Early versus Late
Initiation of Renal Replacement Therapy in Critically Ill Patients with Acute Kidney Injury; CVVHDF, continuous venovenous hemodiafiltration; KDIGO, Kidney Disease Improving Global Outcomes; NGAL,
neutrophil gelatinase–associated lipocalin; PaO2, partial pressure of oxygen in arterial blood; FiO 2, fraction of inspired oxygen; ABW, admission body weight; EKG, electrocardiogram; AKIKI, Artificial Kidney
Initiation in Kidney Injury; VDRF, ventilator-dependent respiratory failure; RIFLE, Risk, Injury, Failure, Loss, and End-stage renal failure; FST, furosemide stress test; CKRT, continuous KRT; pNGAL, plasma
neutrophil gelatinase–associated lipocalin; STARRT-AKI, Standard versus Accelerated Initiation of Renal-Replacement Therapy in Acute Kidney Injury.
L2Xvf+Qaq1MC/lwRu920= on 06/13/2023
eAX6oP97/t+reHY6iBuDuEG8j/KLK4+srUZdWUQbtcBUzqkxhgmmXz8HwSY/clF6gRoZeOUloCHsnrNgIHFbIB8VBAS7/AWU3tVc
Downloaded from http://journals.lww.com/cjasn by ovbQKPXYfBFIRkaiUq5PuVRWaEDA1iTVOAm/wWdC7XRdvs6oSNs
Study Population Pre-KRT Fluid Status KRT Modality Exposure Outcome/Results Notable Features
Murugan Retrospective cohort Fluid overload Intermittent HD UF intensity High-intensity UF Sensitivity analyses
et al. Single-center $5% body weight and CKRT - Low (#20 ml/kg per day) (versus low- - Explored alternative UF
(2018)53,a Medical and surgical on admission - CVVHDF - Moderate (20–25 ml/kg intensity) associated thresholds
ICU - CVVHD per day) with improved 1-yr - Propensity score
Jul 2000–Oct 2008 - CVVH - High (.25 ml/kg per day) survival matching
Pittsburgh, USA - SCUF Subgroup CKRT only No difference - Quantitative bias analysis
N51075 - Low (,0.5 ml/kg per hour) between moderate- Demographic differences
AKI only - Moderate (0.1–1.0 ml/kg versus low-intensity - Liver disease more in
per hour) groups low-intensity group
- High (.1.0 ml/kg Secondary outcomes - Lower BP in low-
per hour) - Hospital LOS intensity group
- In-hospital - Higher vasopressor dose
mortality in low-intensity group
- Kidney recovery - Higher KDIGO stage in
high-intensity group
Tehranian Retrospective cohort No criteria CKRT (CVVH) UF intensity High-intensity UF Sensitivity analyses
et al. Single-center Subset analysis of - Low (,35 ml/kg per day) (versus low- - Explored alternative UF
(2019)54,a Medical and surgical fluid overload - High ($35 ml/kg per day) intensity) associated thresholds
ICU $10% body weight with improved 30-d - Subgroup analyses
Dec 2006–Nov 2015 on admission survival Demographic differences
N51398 Secondary outcomes - More early hypotension
Minnesota, USA - Hospital LOS in low-intensity group
AKI only - In-hospital mortality - More mechanical
- MAKE90 ventilation in high-
- Kidney recovery intensity group
- 90-d mortality
793
L2Xvf+Qaq1MC/lwRu920= on 06/13/2023
eAX6oP97/t+reHY6iBuDuEG8j/KLK4+srUZdWUQbtcBUzqkxhgmmXz8HwSY/clF6gRoZeOUloCHsnrNgIHFbIB8VBAS7/AWU3tVc
Downloaded from http://journals.lww.com/cjasn by ovbQKPXYfBFIRkaiUq5PuVRWaEDA1iTVOAm/wWdC7XRdvs6oSNs
794
Table 2. (Continued)
CJASN
Study Population Pre-KRT Fluid Status KRT Modality Exposure Outcome/Results Notable Features
Naorungroj Retrospective cohort No criteria CRKT UF rate (in first 48 h) High UF rate (versus Sensitivity analyses
et al. Single-center - CVVHDF - Low (,1.01 ml/kg per low UF rate) - Subgroup analyses
(2021)56,b Medical and surgical - CVVH hour) associated with - Varied time of mortality
ICU - Moderate (1.01–1.75 ml/kg higher assessment
2016–2018 per hour) 28-d mortality Demographic differences
Australia - High (.1.75 ml/kg - Higher vasopressor use
N5347 per hour) in low UF rate group
AKI and chronic - Longer duration of KRT
kidney failure in high UF rate group
This table provides a summary of the studies examining ultrafiltration and mortality in a critical care setting, with comparisons of study population, inclusion criteria, KRT modality,
definition of net UF groups, outcomes, and other notable features. Note that while the studies refer to a net UF to account for the administration of intravenous fluids needed to complete KRT
(in the case of these studies, this consisted of replacement fluid during hemofiltration) to avoid confusion, all fluid removal achieved using KRT is referred to simply as UF. ICU, intensive
care unit; HD, hemodialysis; CKRT, continuous KRT; CVVHDF, continuous venovenous hemodiafiltration; CVVHD, continuous venovenous hemodialysis; CVVH, continuous venovenous
hemofiltration; SCUF, slow continuous ultrafiltration; UF, ultrafiltration; LOS, length of stay; KDIGO, Kidney Disease Improving Global Outcomes; RENAL, Randomized Evaluation of
Normal versus Augmented Level (RENAL) Replacement Therapy Study.
a
The two studies that found an association between higher ultrafiltration rate and lower mortality.
b
The two studies that found an association between higher ultrafiltration rate and higher mortality.
CJASN 18: 788–802, June, 2023 Volume Management with KRT in Critical Illness, Wang et al. 795
as net ultrafiltration. However, the term net ultrafiltration result, except in Figure 2, we will refer to net ultrafiltration
can be misleading because it can be mistaken for net fluid as simply ultrafiltration. Ultrafiltration intensity is typically
balance. Figure 2 serves to illustrate the potential for con- obtained by dividing ultrafiltration by admission weight
fusion regarding metrics such as net ultrafiltration. As a and by unit of time.53 For patients receiving only CKRT, the
eAX6oP97/t+reHY6iBuDuEG8j/KLK4+srUZdWUQbtcBUzqkxhgmmXz8HwSY/clF6gRoZeOUloCHsnrNgIHFbIB8VBAS7/AWU3tVc
Downloaded from http://journals.lww.com/cjasn by ovbQKPXYfBFIRkaiUq5PuVRWaEDA1iTVOAm/wWdC7XRdvs6oSNs
L2Xvf+Qaq1MC/lwRu920= on 06/13/2023
Figure 2. Proposed approach to decisions that guide fluid removal and net fluid balance for patients on KRT. The diagram depicts several
descriptors of fluid removal and fluid balance used in studies of KRT. This serves to emphasize that the term net UF—which is used in multiple
studies of KRT—is substantially different from overall net fluid balance. Exact calculations used in individual studies may vary. Sources of fluid
input and output arise from both routine clinical care (non-KRT) and from KRT modalities, including intermittent hemodialysis, prolonged or
extended KRT, and continuous KRT such as CVVH and CVVHD. Included is an illustrative example of the concepts of gross UF, net UF, and net
UF intensity, as well as cumulative fluid balance in two scenarios: one with low daily non-KRT input and one with high daily non-KRT input.
a
KRT durations are typical examples and assume no interruptions or early cessation of treatment, which may not be the case in clinical practice.
b
Gross UF is the total amount of fluid removed during KRT, not accounting for any KRT-related fluid input. cFluid used to prime the circuit at
start of treatment and rinse blood back at end of treatment is often recorded in clinical practice but is frequently ignored in retrospective studies.
For continuous modalities, it is assumed that therapy is neither started nor stopped during the 24 hours in question, so no prime or rinse-back fluid
is given. dNet UF represents the net fluid removed when accounting for KRT fluid input, that is, net UF equals the gross UF minus KRT fluid input.
e
UF Intensity5(Total Net UF for all KRT)/(Days of KRT)/(Admission Weight). For CKRT, number of hours of CKRT is converted into days.53 For
intermittent KRT, each treatment contributes a day of KRT. For patients who receive both, these two are added together. Non-KRT days are not
counted. fFor patients receiving CKRT only, UF intensity can be expressed per hour instead of per day.53 CKRT, continuous KRT; CVVH,
continuous venovenous hemofiltration; CVVHD, continuous venovenous hemodialysis; UF, ultrafiltration.
796 CJASN
unit is per hour, while for participants receiving intermit- higher 30-day survival than those who received low-
tent KRT, the unit is per day—including only the days on intensity ultrafiltration (,35 ml/kg per day).53,54
which dialysis is received (therefore ignoring nondialy- By contrast, a separate study by Murugan et al. consisting
sis days). of a secondary analysis of the Randomized Evaluation of
In a single-center retrospective study by Murugan Normal versus Augmented Level (RENAL) Replacement
eAX6oP97/t+reHY6iBuDuEG8j/KLK4+srUZdWUQbtcBUzqkxhgmmXz8HwSY/clF6gRoZeOUloCHsnrNgIHFbIB8VBAS7/AWU3tVc
et al. of critically ill patients with AKI receiving either Therapy Study found that ultrafiltration .1.75 ml/kg per
Downloaded from http://journals.lww.com/cjasn by ovbQKPXYfBFIRkaiUq5PuVRWaEDA1iTVOAm/wWdC7XRdvs6oSNs
intermittent KRT or CKRT with pre-KRT positive fluid hour was associated with higher 90-day mortality compared
balance $5% above admission body weight, patients with ultrafiltration ,1.01 ml/kg per hour.55 Using the
who received high-intensity ultrafiltration (defined as an same ultrafiltration cutoffs in a single-center retrospective
average daily ultrafiltration .25 ml/kg per day) had lower cohort, Naorungroj et al. demonstrated similar associa-
risk-adjusted 1-year mortality compared with patients tions with 28-day mortality.56 Both studies found a dose-
who received low-intensity ultrafiltration (average daily dependent relationship between each 0.5 ml/kg per hour
ultrafiltration ,20 ml/kg per day).53 The association re- higher ultrafiltration and higher mortality.55,56
mained consistent in the CKRT-only subgroup, with better Why are there such different results among these studies?
survival in patients with ultrafiltration .1 ml/kg per hour As with all observational data, even with adjustment, these
compared with ultrafiltration ,0.5 ml/kg per hour.53 Simi- studies are subject to substantial confounding such as from
L2Xvf+Qaq1MC/lwRu920= on 06/13/2023
larly, in a retrospective single-center study of critically ill differences in provider practices, severity of illness, indication
patients with AKI initiating CKRT, patients who received for KRT, and definition of fluid overload. However, one
more intensive ultrafiltration ($35 ml/kg per day) had additional possible explanation is that patients in different
Table 3. Comparison of net fluid balance achieved across studies on net ultrafiltration rate
Murugan et al. Thresholds for UF #20 ml/kg per day 20–25 ml/kg per day .25 ml/kg per day
(2018)53 rate intensity
CKRT subset only Fluid balance before KRT 2.3 L (15.6%) 2.7 L (17.3%) 2.3 L (21.0%)
UF during KRT 19.5 L 27.9 L 26.6 L
(CKRT1intermittent HD)
Cumulative fluid balance 13.5 L 22.0 L 19.0 L
during KRT (excludes UF)
Cumulative fluid balance 26 L (inferreda) 25.9 L (inferreda) 27.6 L (inferreda)
during KRT (includes UF)
Thresholds for UF ,5 ml/kg per hour 0.5–1.0 ml/kg per hour .1 ml/kg per hour
rate intensity
Fluid balance before KRT Unknown Unknown Unknown
Net UF during CKRT 3.4 L 11.6 L 16.2 L
Cumulative fluid balance Unknown Unknown Unknown
during KRT (includes UF)
Tehranian Thresholds for UF ,35 ml/kg per day No moderate group $35 ml/kg per day
et al. (2019)54 rate intensity
Fluid balance before KRT 4.2 L — 5.8 L
UF during CKRT Unknown — Unknown
Cumulative fluid balance Unknown — Unknown
during KRT (includes UF)
Murugan Thresholds for UF ,1.01 ml/kg 1.01–1.75 ml/kg .1.75 ml/kg
et al. (2019)55 rate intensity per hour per hour per hour
Fluid balance before CKRT Unknown Unknown Unknown
UF during CKRT 1.7 L 8.5 L 16.5 L
Cumulative fluid balance 4.6 L 8.5 L 11.1 L
during CKRT (excludes UF)
Cumulative fluid balance 12.3 L 20.4 L 23.6 L
during CKRT (includes UF)
Naorungroj Thresholds for UF ,1.01 ml/kg 1.01–1.75 ml/kg .1.75 ml/kg
et al. (2021)56 rate intensity per hour per hour per hour
Fluid balance before KRT 0.22 L 0.31 L 0.68 L
UF during CKRT 1.0 L 3.6 L 5.3 L
Cumulative fluid balance 10.53 L 20.66 L 21.75 L
during KRT (includes UF)
This table provides a detailed comparison of the net fluid balance within each of the net ultrafiltration groups (low, moderate, and
high) across the four studies detailed in Table 2. Net fluid balance before initiation of KRT and during KRT is listed for each category
of ultrafiltration intensity, when available. CKRT, continuous KRT; UF, ultrafiltration; HD, hemodialysis.
a
An inferred value rather than a reported value based on the following two assumptions for comparison purposes only. If fluid
balance during KRT is not reported, then the net fluid balance after KRT including UF, the difference between the cumulative
fluid balance during KRT (excluding UF), and the UF during KRT was used to estimate the median cumulative fluid balance during
KRT (including UF). Note that while the studies refer to a net UF to account for the administration of intravenous fluids needed
to complete KRT (in the case of these studies, this consisted of replacement fluid during hemofiltration), to avoid confusion all
fluid removal achieved using KRT is referred to simply as UF.
CJASN 18: 788–802, June, 2023 Volume Management with KRT in Critical Illness, Wang et al. 797
studies may not have been in the same fluid management the number of days when a fluid balance goal was set
phase. In studies that found high ultrafiltration rate to be correlated positively with the likelihood of achieving a
associated with lower mortality,53,54 patients had significant lower cumulative fluid balance during CKRT, further sup-
positive fluid balance at KRT initiation (Table 3). By contrast, porting the importance of fluid balance goals to guide
participants in Naorungroj et al.56—in whom high ultrafil- management. The relevance of achieving these goals
eAX6oP97/t+reHY6iBuDuEG8j/KLK4+srUZdWUQbtcBUzqkxhgmmXz8HwSY/clF6gRoZeOUloCHsnrNgIHFbIB8VBAS7/AWU3tVc
tration rate was associated with higher mortality—had only was emphasized in a recent study showing that a larger
Downloaded from http://journals.lww.com/cjasn by ovbQKPXYfBFIRkaiUq5PuVRWaEDA1iTVOAm/wWdC7XRdvs6oSNs
minimally positive fluid balance at baseline. In addition, this gap between prescribed and achieved fluid balance was
study used only the ultrafiltration rate during the first 48 independently associated with higher mortality in a cohort
hours of KRT the analysis, reasoning that this is “a time of critically ill patients receiving CKRT.59 Of note, while
where illness severity may be most likely to be either posi- low ultrafiltration rate was also associated with higher
tively or negatively affected by [ultrafiltration] management.”56 mortality, this association was no longer present after
adjustment for the gap between prescribed and achieved
fluid balance as well as other clinical parameters. This
Ultrafiltration Rate versus Net Fluid Balance further suggests that targets based on fluid balance may
Both the ROS-D model and typical clinical practice focus be preferable to those based on ultrafiltration rate alone.
on net fluid balance rather than fluid removal alone. Just as Of course, as with all observational studies, RCTs will
L2Xvf+Qaq1MC/lwRu920= on 06/13/2023
it is difficult to infer net fluid balance during diuresis from be needed to confirm the benefit of achieving net fluid
urine output alone, net fluid balance in patients receiving balance on clinically relevant outcomes.
KRT cannot be extrapolated from only the daily ultrafiltra-
tion (Figure 2). A quick glance at the four cohort studies in
Table 3 shows that three of the studies seem to have different Defining Fluid Overload versus Fluid Intolerance
Many observational studies have defined fluid over-
amounts of cumulative fluid balance during KRT, and the
load based on percent net fluid gain divided by weight
fourth does not report fluid balance during KRT at all.54
on admission, often using an arbitrary cutoff such as a .10%
Moreover, the relationships among ultrafiltration rate, net
increase to define fluid overload.50,60–62 By this definition,
fluid balance during KRT, and mortality differ among the
fluid overload is clearly associated with higher mortality,
studies. In the two studies showing higher mortality asso-
AKI, and other negative outcomes.50,60–62 However, this
ciated with higher ultrafiltration rate, the low-intensity
metric is limited by confounding given that fluid gain is
groups had a net positive median fluid balance during
associated with severity of illness and degree of volume
KRT, and the high-intensity groups had a net negative
depletion on admission. In addition, there are practical
median balance.55,56 By contrast, in the single-center study
challenges to measuring net fluid balance precisely. Thus,
by Murugan et al.53—which found lower mortality with
the presence of fluid overload defined purely by net pos-
high ultrafiltration rate—all groups seemed to have a net
itive fluid balance compared with admission weight is
negative median balance. This is additional evidence that
not likely to justify fluid removal in an individual patient.
the latter study is drawn from a different patient population
The degree of net positive fluid balance must be placed in
than the former two. Focusing on ultrafiltration rate alone
context of both evidence of fluid responsiveness and signs of
may mask this important observation.
fluid toxicity. It is important to point out that the presence of
Ultrafiltration rate is therefore a problematic stand-alone
fluid responsiveness does not preclude fluid toxicity. Even
metric for either clinical investigation or clinical management
when cardiac output is augmented, additional fluid admin-
at the bedside, especially for critically ill patients who have
istration can still have negative consequences such as wors-
widely varying fluid intake. The same ultrafiltration rate may
ening hypoxemia or increasing intra-abdominal pressure.
lead to either volume overload or volume depletion if there
The degree to which a patient can receive fluid without
are different amounts of fluid input and/or non-KRT fluid
developing toxicity has been defined as fluid tolerance.63,64
loss (Figure 2). Even in the outpatient setting, implementing
When clinicians at the bedside discuss volume overload, they
ultrafiltration rate thresholds without consideration of
are often referring to evidence of fluid toxicity/intolerance,
overall volume status can lead to unintended fluid-related
rather than merely net positive fluid balance.
weight gain and failure to achieve target weights.57
Particularly in the critically ill population, multiple factors—
Net fluid balance is therefore a more clinically relevant
including cardiopulmonary function and the presence of
target during KRT. A recent retrospective study of critically
pathologic capillary leak—affect the balance between fluid
ill patients with AKI receiving CKRT found that achieving a
responsiveness and fluid tolerance.65 Point-of-care ultrasound
greater reduction in cumulative fluid balance was associ-
has emerged as an important bedside tool to rapidly and
ated with lower ICU and hospital mortality.58 There are
noninvasively assess parameters associated with both fluid
several other important findings from this study. First, fluid
responsiveness and fluid intolerance (e.g., repeated presence
balance before CKRT initiation was not associated with
of B lines on lung ultrasonography or evidence of right ven-
mortality after adjustment for severity of illness and need
tricular overload on echocardiography).20,66–68 As these tech-
for vasopressor support, suggesting that confounding fac-
niques continue to mature, they will likely need to be combined
tors are contributing to the higher mortality in unadjusted
with the metric of net fluid gain to guide optimal management.
analysis of baseline fluid overload. Second, while the
achievement of a lower cumulative fluid balance during
CKRT was independently associated with lower mortality, Avoiding Hypoperfusion due to Fluid Removal
the duration of time needed to reach it was not. This raises Using KRT
the possibility that the rate of decongestion CKRT is less Just as the presence of fluid responsiveness may not rule
important than achievement of decongestion itself. Third, out fluid toxicity, the toxicity from fluid does not ensure that
798 CJASN
its removal will be well tolerated. As recent reviews have Near infrared spectroscopy is a growing area of research
described,69,70 classic strategies to mitigate hemodynamic to noninvasively measure changes in regional microcir-
instability—such as higher dialysate sodium concentration, culation, particularly in cerebral blood flow and oxygen-
ultrafiltration profiling, isolated ultrafiltration, and cooled ation during KRT.81–83 However, it is not generally
dialysate—have limited data that are largely extrapolated available for clinical use.
eAX6oP97/t+reHY6iBuDuEG8j/KLK4+srUZdWUQbtcBUzqkxhgmmXz8HwSY/clF6gRoZeOUloCHsnrNgIHFbIB8VBAS7/AWU3tVc
from the outpatient setting. Among the various strategies, Continuous blood volume monitoring is a technique that
Downloaded from http://journals.lww.com/cjasn by ovbQKPXYfBFIRkaiUq5PuVRWaEDA1iTVOAm/wWdC7XRdvs6oSNs
cool dialysate temperature has the best evidence base and is is directly integrated into several currently available inter-
already widely used. mittent HD machines. By optically measuring changes in
Regardless, the most important element in avoiding is- hematocrit in real-time, relative changes in blood volume
chemic injury due to fluid removal during KRT is knowing can be measured, and the plasma refill rate can be esti-
when to stop removing fluid. Many clinicians use a decline mated.84 Unfortunately, studies of this technology in the
in MAP or escalating pressors as triggers to decrease fluid outpatient setting have had mixed results.85–87 One random-
removal. A MAP $65 mm Hg has been suggested as a target ized trial using it to guide ultrafiltration rate during inter-
in most forms of shock.71,72 With a more proactive ap- mittent HD in critically ill patients showed no difference in
proach, it may be possible to predict intradialytic hypoten- hemodynamics or intradialytic complications.88 While newer
sion during ultrafiltration using bedside maneuvers studies suggest novel methods to use blood volume moni-
L2Xvf+Qaq1MC/lwRu920= on 06/13/2023
normally used to predict fluid responsiveness.73–75 toring to inform decisions regarding ultrafiltration,89–92
Unfortunately, BP is a crude measure of hemodynamic these are at a preliminary phase and have only been studied
status and often an inadequate marker for adequate organ in outpatients. It therefore remains unclear whether this
perfusion. Myocardial stunning has been demonstrated technique will prove useful in the critical care setting.
during ultrafiltration—regardless of BP changes—in pa-
tients with kidney failure undergoing routine intermittent
hemodialysis (HD), patients starting intermittent HD for Conclusion
AKI, and even critically ill patients within hours of initiat- Current evidence and technologies are insufficient to
ing CKRT.76–78 Whether these observations are due to di- provide clarity regarding optimal fluid management strat-
alysis, ultrafiltration, or other factors pertaining to critical egies using KRT in the critically ill. However, we believe
illness remains unclear. there are reasonable inferences that can be drawn from the
Tissue perfusion has correlated better with microcircu- existing literature to guide both current management and
latory alterations than with changes in systemic BP.79,80 future investigation.
Figure 3. An approach to ultrafiltration targets in patients with oliguria and distributive shock. aExamples of fluid overload with organ
toxicity include the following: pulmonary edema, right ventricular volume overload, elevated intra-abdominal pressure, and elevated
central venous pressure with decreased organ perfusion pressures. bFluid responsiveness is commonly defined as a 10%–15% increase in
cardiac output after an intravenous fluid bolus and may be predicted with provocative maneuvers.24,25 cThere is no consensus definition
for refractory shock. Examples include the following: high doses of an intravenous vasoactive medication (e.g., .0.2 mg/kg per minute of
norepinephrine or equivalent), need for multiple vasoactive medications, or adjunctive glucocorticoid therapy. dAs part of the frequent
hemodynamic reassessment, discussions with the intensive care unit team should involve decisions on the priority of fluid removal, limits
of vasopressor titration, and ultimately optimal kidney replacement modality to achieve fluid targets.
CJASN 18: 788–802, June, 2023 Volume Management with KRT in Critical Illness, Wang et al. 799
First, both clinical management and research protocols Writing – original draft: Kevin Fay, Dan Negoianu, Christina
should focus, when possible, on net fluid balance rather H. Wang.
than simply the ultrafiltration rate. While the latter is easier Writing – review & editing: Kevin Fay, Dan Negoianu, Michael
to measure, the former is more clinically relevant. Conse- G.S. Shashaty, Christina H. Wang.
quently, we prescribe fluid removal on CKRT at our in-
eAX6oP97/t+reHY6iBuDuEG8j/KLK4+srUZdWUQbtcBUzqkxhgmmXz8HwSY/clF6gRoZeOUloCHsnrNgIHFbIB8VBAS7/AWU3tVc
KRT. Repeated assessments of fluid responsiveness and Pfortmueller CA. Fluid overload and mortality in adult critical
care patients—a systematic review and meta-analysis of obser-
fluid toxicity are needed to guide fluid balance and ultra- vational studies. Crit Care Med. 2020;48(12):1862–1870. doi:10.
filtration goals. Feasible protocols based on physiologic 1097/ccm.0000000000004617
rationales that also leverage available assessment methods 6. Malbrain MLNG, Van Regenmortel N, Saugel B, et al. Principles
(Figure 3) are needed both as a starting point for current of fluid management and stewardship in septic shock: it is time to
consider the four D’s and the four phases of fluid therapy. Ann
clinical management and to design trials to test effect on
Intensive Care. 2018;8(1):66. doi:10.1186/s13613-018-0402-x
patient-centered outcomes. 7. Pajewski R, Gipson P, Heung M. Predictors of post-
Third, the focus of volume management using KRT hospitalization recovery of renal function among patients with
should not be limited to the earliest stages of critical care. acute kidney injury requiring dialysis. Hemodialysis Int. 2018;
The stabilization and de-escalation phases may be particu- 22(1):66–73. doi:10.1111/hdi.12545
8. The RENAL Replacement Therapy Study Investigators, Bellomo
larly suited for future trials. For patients on CKRT, the R, Cass A, Cole L, et al. An observational study fluid balance and
transition to intermittent therapy can be a vulnerable period patient outcomes in the randomized evaluation of normal vs.
and may provide a useful study target. In one recent study augmented level of replacement therapy trial. Crit Care Med.
of patients transitioning from CKRT, intra-dialytic hypo- 2012;40(6):1753–1760. doi:10.1097/ccm.0b013e318246b9c6
9. Prowle JR, Kirwan CJ, Bellomo R. Fluid management for the
tension occurred during 50% of first intermittent HD treat-
prevention and attenuation of acute kidney injury. Nat Rev
ments and was independently associated with mortality.93 Nephrol. 2014;10(1):37–47. doi:10.1038/nrneph.2013.232
In addition, the critical care goals during later phases of 10. Douvris A, Zeid K, Hiremath S, et al. Mechanisms for hemo-
management—such as durable liberation from mechanical dynamic instability related to renal replacement therapy: a
ventilation and intravenous vasopressors—may serve as narrative review. Intensive Care Med. 2019;45(10):1333–1346.
doi:10.1007/s00134-019-05707-w
useful surrogate outcomes for smaller preliminary trials 11. Hoste EAA, Maitland K, Brudney CSS, et al. Four phases of in-
that would not be powered to examine mortality.94 travenous fluid therapy: a conceptual model. Br J Anaesth. 2014;
Awaiting more robust evidence, we suggest that such 113(5):740–747. doi:10.1093/bja/aeu300
critical care goals should still inform our decisions about 12. Cordemans C, De laet I, Van Regenmortel N, et al. Fluid man-
agement in critically ill patients: the role of extravascular lung water,
volume removal targets and the means to achieve them.
abdominal hypertension, capillary leak, and fluid balance. Ann
Should an edematous patient currently on CKRT be Intensive Care. 2012;2(suppl 1):S1. doi:10.1186/2110-5820-2-s1-s1
switched to intermittent HD as soon as intravenous pressors 13. Malbrain MLNG, Marik PE, Witters I, et al. Fluid overload, de-
are discontinued or should CKRT be continued until there is resuscitation, and outcomes in critically ill or injured patients: a
minimal evidence of pulmonary edema? While the answer systematic review with suggestions for clinical practice. Anaesthesiol
Intensive Ther. 2014;46(5):361–380. doi:10.5603/ait.2014.0060
remains uncertain, merely assessing whether the patient is 14. Alhazzani W, Møller MH, Arabi YM, et al. Surviving Sepsis
likely to tolerate intermittent HD seems unlikely to be Campaign: guidelines on the management of critically ill adults
sufficient for this decision. Comprehensive assessment of with Coronavirus Disease 2019 (COVID-19). Intensive Care
factors such as degree of volume overload, daily fluid in- Med. 2020;46(5):854–887. doi:10.1007/s00134-020-06022-5
15. Van Regenmortel N, Verbrugghe W, Roelant E, Van den
take, ventricular function, respiratory status, and functional Wyngaert T, Jorens PG. Maintenance fluid therapy and fluid
goals should inform the optimal approach. creep impose more significant fluid, sodium, and chloride bur-
dens than resuscitation fluids in critically ill patients: a retro-
Disclosures spective study in a tertiary mixed ICU population. Intensive Care
M.G.S. Shashaty reports support from the NIH (R01DK111638). Med. 2018;44(4):409–417. doi:10.1007/s00134-018-5147-3
C.H. Wang reports funding from the NIH (K23DK129770). All 16. Levine Z, Vijayan A. Prolonged intermittent kidney replacement
remaining authors have nothing to disclose. therapy. Clin J Am Soc Nephrol. 2023;18(3):383–391. doi:10.
2215/CJN.04310422
17. Kashani K, Mehta RL. We restrict CRRT to only the most he-
Funding modynamically unstable patients. Semin Dial. 2016;29(4):
None. 268–271. doi:10.1111/sdi.12507
18. Kattan E, Ospina-Tascón GA, Teboul J-L, et al. Systematic as-
Author Contributions sessment of fluid responsiveness during early septic shock re-
Conceptualization: Dan Negoianu. suscitation: secondary analysis of the ANDROMEDA-SHOCK
Supervision: Dan Negoianu. trial. Crit Care. 2020;24(1):23. doi:10.1186/s13054-020-2732-y
800 CJASN
19. Leisman DE, Doerfler ME, Schneider SM, Masick KD, D’Amore patients with acute kidney injury and sepsis. N Engl J Med. 2018;
JA, D’Angelo JK. Predictors, prevalence, and outcomes of early 379(15):1431–1442. doi:10.1056/NEJMoa1803213
crystalloid responsiveness among initially hypotensive patients 38. Wald R, Kirkham B, DaCosta BR, et al. Fluid balance and renal
with sepsis and septic shock. Crit Care Med. 2018;46(2): replacement therapy initiation strategy: a secondary analysis of
189–198. doi:10.1097/ccm.0000000000002834 the STARRT-AKI trial. Crit Care. 2022;26(1):360. doi:10.1186/
20. Beaubien-Souligny W, Trott T, Neyra JA. How to determine fluid s13054-022-04229-0
eAX6oP97/t+reHY6iBuDuEG8j/KLK4+srUZdWUQbtcBUzqkxhgmmXz8HwSY/clF6gRoZeOUloCHsnrNgIHFbIB8VBAS7/AWU3tVc
management goals during continuous kidney replacement 39. Nash DM, Przech S, Wald R, O’Reilly D. Systematic review and
Downloaded from http://journals.lww.com/cjasn by ovbQKPXYfBFIRkaiUq5PuVRWaEDA1iTVOAm/wWdC7XRdvs6oSNs
therapy in patients with AKI: focus on POCUS. Kidney360. 2022; meta-analysis of renal replacement therapy modalities for acute
3(10):1795–1806. doi:10.34067/KID.0002822022 kidney injury in the intensive care unit. J Crit Care. 2017;41:
21. Monnet X, Letierce A, Hamzaoui O, et al. Arterial pressure allows 138–144. doi:10.1016/j.jcrc.2017.05.002
monitoring the changes in cardiac output induced by volume 40. Augustine JJ, Sandy D, Seifert TH, Paganini EP. A randomized
expansion but not by norepinephrine. Crit Care Med. 2011;39(6): controlled trial comparing intermittent with continuous dialysis
1394–1399. doi:10.1097/ccm.0b013e31820edcf0 in patients with ARF. Am J Kidney Dis. 2004;44(6):1000–1007.
22. Pierrakos C, Velissaris D, Scolletta S, Heenen S, De Backer D, doi:10.1053/j.ajkd.2004.08.022
Vincent J-L. Can changes in arterial pressure be used to detect 41. Mehta RL, McDonald B, Gabbai FB, et al. A randomized clinical
changes in cardiac index during fluid challenge in patients with trial of continuous versus intermittent dialysis for acute renal
septic shock? Intensive Care Med. 2012;38(3):422–428. doi:10. failure. Kidney Int. 2001;60(3):1154–1163. doi:10.1046/j.1523-
1007/s00134-011-2457-0 1755.2001.0600031154.x
23. Ait-Hamou Z, Teboul J-L, Anguel N, Monnet X. How to detect a 42. Schefold JC, Haehling SV, Pschowski R, et al. The effect of
positive response to a fluid bolus when cardiac output is not continuous versus intermittent renal replacement therapy on the
L2Xvf+Qaq1MC/lwRu920= on 06/13/2023
measured? Ann Intensive Care. 2019;9(1):138. doi:10.1186/ outcome of critically ill patients with acute renal failure
s13613-019-0612-x (CONVINT): a prospective randomized controlled trial. Crit Care.
24. Vincent J-L, Joosten A, Saugel B. Hemodynamic monitoring and 2014;18(1):R11. doi: 10.1186/cc13188.
support. Crit Care Med. 2021;49(10):1638–1650. doi:10.1097/ 43. Uehlinger DE, Jakob SM, Ferrari P, et al. Comparison of con-
ccm.0000000000005213 tinuous and intermittent renal replacement therapy for acute
25. Kashani K, Omer T, Shaw AD. The intensivist’s perspective of shock, renal failure. Nephrol Dial Transplant. 2005;20(8):1630–1637.
volume management, and hemodynamic monitoring. Clin J Am Soc doi:10.1093/ndt/gfh880
Nephrol. 2022;17(5):706–716. doi:10.2215/CJN.14191021 44. Vinsonneau C, Camus C, Combes A, et al. Continuous veno-
26. Cecconi M, De Backer D, Antonelli M, et al. Consensus on venous haemodiafiltration versus intermittent haemodialysis for
circulatory shock and hemodynamic monitoring. Task force of acute renal failure in patients with multiple-organ dysfunction
the European Society of Intensive Care Medicine. Intensive Care syndrome: a multicentre randomised trial. Lancet. 2006;
Med. 2014;40(12):1795–1815. doi:10.1007/s00134-014-3525-z 368(9533):379–385. doi:10.1016/s0140-6736(06)69111-3
27. Pinsky MR, Brophy P, Padilla J, Paganini E, Pannu N. Fluid and 45. Lins RL, Elseviers MM, Van Der Niepen P, et al. Intermittent
volume monitoring. Int J Artif Organs. 2008;31(2):111–126. doi: versus continuous renal replacement therapy for acute kidney
10.1177/039139880803100205 injury patients admitted to the intensive care unit: results of a
28. Murugan R, Hoste E, Mehta RL, et al. Precision fluid management randomized clinical trial. Nephrol Dial Transplant. 2009;24(2):
in continuous renal replacement therapy. Blood Purif. 2016; 512–518. doi:10.1093/ndt/gfn560
42(3):266–278. doi:10.1159/000448528 46. John S, Griesbach D, Baumgärtel M, Weihprecht H, Schmieder
29. Combes A, Bréchot N, Amour J, et al. Early high-volume he- RE, Geiger H. Effects of continuous haemofiltration vs in-
mofiltration versus standard care for post-cardiac surgery shock termittent haemodialysis on systemic haemodynamics and
the HEROICS study. Am J Respir Crit Care Med. 2015;192(10): splanchnic regional perfusion in septic shock patients: a pro-
1179–1190. doi:10.1164/rccm.201503-0516oc spective, randomized clinical trial. Nephrol Dial Transplant.
30. Zarbock A, Kellum JA, Schmidt C, et al. Effect of early vs delayed 2001;16(2):320–327. doi:10.1093/ndt/16.2.320
initiation of renal replacement therapy on mortality in critically ill 47. Stephen CNJ, Simpson K, Allison MEM. Long-term quality of life
patients with acute kidney injury: the ELAIN randomized clinical and hospital mortality in patients treated with intermittent or
trial. JAMA. 2016;315(20):2190–2199. doi:10.1001/jama.2016. continuous hemodialysis for acute renal and respiratory failure.
5828 Ren Fail. 2006;28(4):323–330. doi:10.1080/
31. Gaudry S, Hajage D, Schortgen F, et al. Initiation strategies for 08860220600591487
renal-replacement therapy in the intensive care unit. N Engl J 48. Davenport A, Will EJ, Davison AM, et al. Changes in intracranial
Med. 2016;375(2):122–133. doi:10.1056/nejmoa1603017 pressure during haemofiltration in oliguric patients with grade IV
32. Lumlertgul N, Peerapornratana S, Trakarnvanich T, et al. Early hepatic encephalopathy. Nephron. 1989;53(2):142–146. doi:10.
versus standard initiation of renal replacement therapy in furose- 1159/000185727
mide stress test non-responsive acute kidney injury patients (the FST 49. Gašparovic V, Filipovic-Grcic I, Merkler M, Pišl Z. Continuous
trial). Crit Care. 2018;22(1):101. doi:10.1186/s13054-018-2021-1 renal replacement therapy (CRRT) or intermittent hemodialysis
33. The STARRT-AKI Investigators. Timing of initiation of renal- (IHD)—what is the procedure of choice in critically ill patients?
replacement therapy in acute kidney injury. N Engl J Med. 2020; Ren Fail. 2003;25(5):855–862. doi:10.1081/jdi-120024300
383(3):240–251. doi:10.1056/nejmoa2000741 50. Bouchard J, Soroko SB, Chertow GM, et al. Fluid accumulation,
34. Gaudry S, Hajage D, Martin-Lefevre L, et al. Comparison of two survival and recovery of kidney function in critically ill patients
delayed strategies for renal replacement therapy initiation for with acute kidney injury. Kidney Int. 2009;76(4):422–427. doi:
severe acute kidney injury (AKIKI 2): a multicentre, open-label, 10.1038/ki.2009.159
randomised, controlled trial. Lancet. 2021;397(10281):1293– 51. Silversides JA, Major E, Ferguson AJ, et al. Conservative fluid
1300. doi:10.1016/s0140-6736(21)00350-0 management or deresuscitation for patients with sepsis or
35. Srisawat N, Laoveeravat P, Limphunudom P, et al. The effect of acute respiratory distress syndrome following the resuscitation
early renal replacement therapy guided by plasma neutrophil phase of critical illness: a systematic review and meta-analysis.
gelatinase associated lipocalin on outcome of acute kidney in- Intensive Care Med. 2017;43(2):155–170. doi:10.1007/s00134-
jury: a feasibility study. J Crit Care. 2018;43:36–41. doi:10.1016/ 016-4573-3
j.jcrc.2017.08.029 52. National Heart Lung and Blood Institute Acute Respiratory
36. Geri G, Grimaldi D, Seguin T, et al. Hemodynamic efficiency of Distress Syndrome ARDS Clinical Trials Network, Wiedemann
hemodialysis treatment with high cut-off membrane during the HP, Wheeler AP, Bernard GR, et al. Comparison of two fluid-
early period of post-resuscitation shock: the HYPERDIA trial. management strategies in acute lung injury. N Engl J Med. 2006;
Resuscitation. 2019;140:170–177. doi:10.1016/j.resuscitation. 354(24):2564–2575. doi:10.1056/NEJMoa062200
2019.03.045 53. Murugan R, Balakumar V, Kerti SJ, et al. Net ultrafiltration in-
37. Saber Barbar D, Clere-Jehl R, Bourredjem A, et al; Jean-Phil for tensity and mortality in critically ill patients with fluid overload.
the I-ITI and the CTN. Timing of renal-replacement therapy in Crit Care. 2018;22(1):223. doi:10.1186/s13054-018-2163-1
CJASN 18: 788–802, June, 2023 Volume Management with KRT in Critical Illness, Wang et al. 801
54. Tehranian S, Kashani KB, Shawwa K, Kashani KB. Net ultrafil- international consensus conference, paris, France, 27–28 April
tration rate and its impact on mortality in patients with acute 2006. Intensive Care Med. 2007;33(4):575–590. doi:10.1007/
kidney injury receiving continuous renal replacement therapy. s00134-007-0531-4
Clin Kidney J. 2019;14(2):564–569. doi:10.1093/ckj/sfz179 72. Thooft A, Favory R, Salgado DR, et al. Effects of changes in ar-
55. Murugan R, Kerti SJ, Chang CCH, et al. Association of net ul- terial pressure on organ perfusion during septic shock. Crit Care.
trafiltration rate with mortality among critically ill adults with 2011;15(5):R222. doi:10.1186/cc10462
eAX6oP97/t+reHY6iBuDuEG8j/KLK4+srUZdWUQbtcBUzqkxhgmmXz8HwSY/clF6gRoZeOUloCHsnrNgIHFbIB8VBAS7/AWU3tVc
acute kidney injury receiving continuous venovenous hemo- 73. Kaptein MJ, Kaptein EM. Inferior vena cava collapsibility index:
Downloaded from http://journals.lww.com/cjasn by ovbQKPXYfBFIRkaiUq5PuVRWaEDA1iTVOAm/wWdC7XRdvs6oSNs
diafiltration: a secondary analysis of the randomized evaluation clinical validation and application for assessment of relative
of normal vs augmented level (RENAL) of renal. JAMA Netw intravascular volume. Adv Chronic Kidney Dis. 2021;28(3):
Open. 2019;2(6):e195418. doi:10.1001/jamanetworkopen. 218–226. doi:10.1053/j.ackd.2021.02.003
2019.5418 74. Monnet X, Cipriani F, Camous L, et al. The passive leg raising test
56. Naorungroj T, Neto AS, Zwakman-Hessels L, et al. Early net to guide fluid removal in critically ill patients. Ann Intensive Care.
ultrafiltration rate and mortality in critically ill patients receiving 2016;6(1):46. doi:10.1186/s13613-016-0149-1
continuous renal replacement therapy. Nephrol Dial Transplant. 75. Malbrain MLNG, Martin G, Ostermann M. Everything you need
2021;36(6):1112–1119. doi:10.1093/ndt/gfaa032 to know about deresuscitation. Intensive Care Med. 2022;48(12):
57. Flythe JE, Assimon MM, Overman RA. Target weight achievement 1781–1786. doi:10.1007/s00134-022-06761-7
and ultrafiltration rate thresholds: potential patient implications. 76. Mahmoud H, Forni LG, McIntyre CW, Selby NM. Myocardial
BMC Nephrol. 2017;18(1):185. doi:10.1186/s12882-017-0595-5 stunning occurs during intermittent haemodialysis for acute
58. Hall A, Crichton S, Dixon A, Skorniakov I, Kellum JA, Ostermann kidney injury. Intensive Care Med. 2017;43(6):942–944. doi:10.
M. Fluid removal associates with better outcomes in critically ill 1007/s00134-017-4768-2
L2Xvf+Qaq1MC/lwRu920= on 06/13/2023
patients receiving continuous renal replacement therapy: a co- 77. Slessarev M, Salerno F, Ball IM, McIntyre CW. Continuous renal
hort study. Crit Care. 2020;24(1):279. doi:10.1186/s13054-020- replacement therapy is associated with acute cardiac stunning in
02986-4 critically ill patients. Hemodial Int. 2019;23(3):325–332. doi:10.
59. Neyra JA, Lambert J, Ortiz-Soriano V, et al. Assessment of pre- 1111/hdi.12760
scribed vs. achieved fluid balance during continuous renal re- 78. Buchanan C, Mohammed A, Cox E, et al. Intradialytic cardiac
placement therapy and mortality outcome. PLoS One. 2022; magnetic resonance imaging to assess cardiovascular responses in a
17(8):e0272913. doi:10.1371/journal.pone.0272913 short-term trial of hemodiafiltration and hemodialysis. J Am Soc
60. Vaara ST, Korhonen A-M, Kaukonen K-M, et al. Fluid overload is Nephrol. 2017;28(4):1269–1277. doi:10.1681/ASN.2016060686
associated with an increased risk for 90-day mortality in critically 79. Lima A, Van Rooij T, Ergin B, et al. Dynamic contrast-enhanced
ill patients with renal replacement therapy: data from the pro- ultrasound identifies microcirculatory alterations in sepsis-
spective FINNAKI study. Crit Care. 2012;16(5):R197. doi:10. induced acute kidney injury. Crit Care Med. 2018;46(8):
1186/cc11682 1284–1292. doi:10.1097/ccm.0000000000003209
61. Payen D, de Pont A-CJM, Sakr Y, Spies C, Reinhart K, Vincent J-L. 80. Montomoli J, Donati A, Ince C. Acute kidney injury and fluid
A positive fluid balance is associated with a worse outcome in resuscitation in septic patients: are we protecting the kidney?
patients with acute renal failure. Crit Care. 2008;12(3):R74. doi: Nephron. 2019;143(3):170–173. doi:10.1159/000501748
10.1186/cc6916 81. Wolfgram DF. Intradialytic cerebral hypoperfusion as mechanism
62. Woodward CW, Lambert J, Ortiz-Soriano V, et al. Fluid overload for cognitive impairment in patients on hemodialysis. J Am Soc
associates with major adverse kidney events in critically ill pa- Nephrol. 2019;30(11):2052–2058. doi:10.1681/ASN.2019050461
tients with acute kidney injury requiring continuous renal re- 82. MacEwen C, Sutherland S, Daly J, Pugh C, Tarassenko L. Re-
placement therapy. Crit Care Med. 2019;47(9):E753–E760. doi: lationship between hypotension and cerebral ischemia during
10.1097/ccm.0000000000003862 hemodialysis. J Am Soc Nephrol. 2017;28(8):2511–2520. doi:10.
63. Kattan E, Castro R, Miralles-Aguiar F, Hernández G, Rola P. The 1681/ASN.2016060704
emerging concept of fluid tolerance: a position paper. J Crit Care. 83. Polinder-Bos HA, Elting JWJ, Aries MJH, et al. Changes in ce-
2022;71:154070. doi:10.1016/j.jcrc.2022.154070 rebral oxygenation and cerebral blood flow during hemodialysis
64. Argaiz ER, Rola P, Haycock KH, Verbrugge FH. Fluid management —a simultaneous near-infrared spectroscopy and positron
in acute kidney injury: from evaluating fluid responsiveness to- emission tomography study. J Cereb Blood Flow Metab. 2020;
wards assessment of fluid tolerance. Eur Hear J Acute Cardiovasc 40(2):328–340. doi:10.1177/0271678x18818652
Care. 2022;11(10):786–793. doi:10.1093/ehjacc/zuac104 84. Mahony S, Ward F. Blood volume monitoring: a clinical tool to
65. Grams ME, Estrella MM, Coresh J, Brower RG, Liu KD. Fluid guide ultrafiltration in volume control and optimisation of in-
balance, diuretic use, and mortality in acute kidney injury. Clin J tradialytic blood pressure. EMJ Nephrol. 2021;9(1):70–78.
Am Soc Nephrol. 2011;6(5):966–973. doi:10.2215/CJN. 85. Antlanger M, Josten P, Kammer M, et al. Blood volume-monitored
08781010 regulation of ultrafiltration to decrease the dry weight in fluid-
66. Argaiz ER, Koratala A, Reisinger N. Comprehensive assessment of overloaded hemodialysis patients: a randomized controlled trial.
fluid status by point-of-care ultrasonography. Kidney360. 2021; BMC Nephrol. 2017;18(1):238. doi:10.1186/s12882-017-0639-x
2(8):1326–1338. doi:10.34067/kid.0006482020 86. Gil HW, Bang K, Lee SY, et al. Efficacy of hemocontrol bio-
67. Reisinger N, Lohani S, Hagemeier J, Panebianco N, Baston C. feedback system in intradialytic hypotension-prone hemodialysis
Lung ultrasound to diagnose pulmonary congestion among pa- patients. J Korean Med Sci. 2014;29(6):805–810. doi:10.3346/
tients on hemodialysis: comparison of full versus abbreviated jkms.2014.29.6.805
scanning protocols. Am J Kidney Dis. 2022;79(2):193–201.e1. 87. Reddan DN, Szczech LA, Hasselblad V, et al. Intradialytic blood
doi:10.1053/j.ajkd.2021.04.007 volume monitoring in ambulatory hemodialysis patients: a ran-
68. Maw AM, Hassanin A, Ho PM, et al. Diagnostic accuracy of point- domized trial. J Am Soc Nephrol. 2005;16(7):2162–2169. doi:10.
of-care lung ultrasonography and chest radiography in adults with 1681/ASN.2004121053
symptoms suggestive of acute decompensated heart failure: a 88. Du Cheyron D, Terzi N, Seguin A, et al. Use of online blood
systematic review and meta-analysis. JAMA Netw Open. 2019; volume and blood temperature monitoring during haemodialysis
2(3):e190703. doi:10.1001/jamanetworkopen.2019.0703 in critically ill patients with acute kidney injury: a single-centre
69. Douvris A, Malhi G, Hiremath S, et al. Interventions to prevent randomized controlled trial. Nephrol Dial Transplant. 2013;
hemodynamic instability during renal replacement therapy in 28(2):430–437. doi:10.1093/ndt/gfs124
critically ill patients: a systematic review. Crit Care. 2018;22(1): 89. Preciado P, Zhang H, Thijssen S, Kooman JP, Van Der Sande FM,
41. doi:10.1186/s13054-018-1965-5 Kotanko P. All-cause mortality in relation to changes in relative
70. Chan RJ, Helmeczi W, Canney M, Clark EG. Management of blood volume during hemodialysis. Nephrol Dial Transplant.
intermittent hemodialysis in the critically ill patient. Clin J Am 2019;34(8):1401–1408. doi:10.1093/ndt/gfy286
Soc Nephrol. 2023;18(2):245–255. doi:10.2215/CJN. 90. Mitsides N, Pietribiasi M, Waniewski J, Brenchley P, Mitra S.
04000422 Transcapillary refilling rate and its determinants during haemo-
71. Antonelli M, Levy M, Andrews PJD, et al. Hemodynamic dialysis with standard and high ultrafiltration rates. Am J Nephrol.
monitoring in shock and implications for management: 2019;50(2):133–143. doi:10.1159/000501407
802 CJASN
91. Chaudhuri S, Han H, Monaghan C, et al. Real-time prediction of 93. Beaubien-Souligny W, Yang Y, Burns KEA, et al. Intra-dialytic
intradialytic relative blood volume: a proof-of-concept for in- hypotension following the transition from continuous to in-
tegrated cloud computing infrastructure. BMC Nephrol. 2021; termittent renal replacement therapy. Ann Intensive Care. 2021;
22(1):274. doi:10.1186/s12882-021-02481-0 11(1):96. doi:10.1186/s13613-021-00885-7
92. Wang CH, Negoianu D, Zhang H, et al. Dynamics of plasma refill 94. Veldhoen RA, Howes D, Maslove DM. Is mortality a useful
rate and intradialytic hypotension during hemodialysis: retro- primary end point for critical care trials? Chest. 2020;158(1):
eAX6oP97/t+reHY6iBuDuEG8j/KLK4+srUZdWUQbtcBUzqkxhgmmXz8HwSY/clF6gRoZeOUloCHsnrNgIHFbIB8VBAS7/AWU3tVc