Original Investigation

Pruritus in Hemodialysis Patients: Longitudinal

Associations With Clinical and Patient-Reported
Outcomes
Nidhi Sukul, Junhui Zhao, Ronald L. Pisoni, Sebastian Walpen, Thilo Schaufler, Elham Asgari,
Fitsum Guebre-Egziabher, Li Zho, Mohammed Abdulrahman Al-Ghonaim, Kosaku Nitta, Bruce M. Robinson, and
Angelo Karaboyas

Rationale & Objective: Cross-sectional studies Results: 51% of patients had moderate to severe Complete author and article
have reported an association of chronic kidney CKD-aP symptoms at either assessment (22% at information provided before
references.
disease–associated pruritus (CKD-aP) with both). The prevalences of depression, restless
adverse clinical events and patient-reported sleep, and feeling drained increased over the Correspondence to
outcomes (PROs). We studied the longitudinal study period (+13%, +10%, and +14%, N. Sukul (nsukul@med.
umich.edu)
associations between changes in CKD-aP and respectively) among patients with incident
clinical outcomes among patients receiving pruritus and decreased (−5%, −8%, and −12%, Am J Kidney Dis.
maintenance hemodialysis. respectively) among patients with resolved 82(6):666-676. Published
online August 16, 2023.
Study Design: Prospective cohort study. pruritus. Minimal changes in PROs over time
were observed for the absent and persistent doi: 10.1053/
Setting & Participants: 7,976 hemodialysis re- groups. Changes over time in laboratory values j.ajkd.2023.04.008
cipients across 21 countries in phases 4-6 (2009- (phosphorus, Kt/V) were not detected for either © 2023 The Authors.
2018) of the Dialysis Outcomes and Practice of these groups. Compared with patients with Published by Elsevier Inc.
Patterns Study (DOPPS) who had 2 CKD-aP absent CKD-aP, the adjusted HRs for patients on behalf of the National
Kidney Foundation, Inc. This
assessments approximately 12 months apart. with persistent CKD-aP were 1.29 (95% CI,
is an open access article
1.09-1.53) for all-cause mortality, 1.17 (1.07- under the CC BY-NC-ND
Exposures: Exposure status was based on the
1.28) for all-cause hospitalization, and 1.48 license (http://
assessment of pruritis initially and again approxi-
(1.26-1.74) for cardiovascular events. creativecommons.org/
mately 1 year later. Four groups were identified, licenses/by-nc-nd/4.0/).
including those with moderate or more severe Limitations: No interim evaluation of CKD-aP
pruritis only at the initial assessment (resolved), symptoms between the 2 assessments;
only at the second assessment (incident), at potential selection bias from patients who died
neither assessment (absent), or at both assess- or were otherwise lost to follow-up before the
ments (persistent). second assessment.
Outcomes: Laboratory values and PROs ascer- Conclusions: CKD-aP symptoms are chronic,
tained at the initial assessment of pruritis and 1 and these findings highlight the potential value
year later. of repeated assessment of this symptom using
Analytical Approach: Linear mixed model to standardized approaches. Future research
investigate changes in laboratory values and should systematically investigate potential cau-
PROs over the 1-year study period across the 4 ses of CKD-aP and options for its effective
exposure groups. treatment.

C hronic kidney disease–associated pruritus (CKD-aP)—

generalized itching related to CKD—and its associa-
tions with clinical and patient-reported outcomes (PROs)
recovery time after a dialysis session, and higher likelihood
of skipping dialysis sessions or withdrawing altogether.1
PROs such as depressive symptoms, poor sleep quality,
have been extensively studied in the hemodialysis (HD) and decreased physical and mental measures of health-
population. In 2020, an analysis of HD recipients partici- related quality of life have been associated with CKD-aP,1-3
affecting not only patients receiving HD but also those
Editorial, p. 647 with nondialysis CKD.4 However, these studies were all
based on single assessments of CKD-aP at baseline, as are
pating in phases 4-6 (2009-2018) of the Dialysis Outcomes most studies in the literature.
and Practice Patterns Study (DOPPS) demonstrated that the Unfortunately, given the absence of a standard mea-
prevalence of moderate to extreme pruritus was unchanged surement method, there is a lack of consensus on the
across DOPPS phases 4, 5, and 6 at 37%,1 reflecting an assessment of pruritus, which likely leads to underdiag-
improvement from 45% in 1996-1999 and 42% in 2000- nosis. Indeed, it has been shown that medical directors
2003.2 The 2020 analysis also investigated the diverse array underestimated the prevalence of pruritus in nearly 70% of
of adverse outcomes associated with CKD-aP, including HD facilities.3 There is even more uncertainty regarding
higher rates of mortality and hospitalization, prolonged treatments, resulting in many practitioners often using

666 AJKD Vol 82 | Iss 6 | December 2023

Sukul et al

captured prospectively throughout follow-up, and PROs

PLAIN-LANGUAGE SUMMARY were captured on the annual self-administered patient
Previous research has studied itching and its conse- questionnaire (PQ). All data were obtained using uniform
quences in hemodialysis recipients only at a single time and standardized data-collection tools.
point. We surveyed 7,976 patients receiving mainte-
nance hemodialysis to assess itching over a period of 1 Variables and Study Sample
year. We found that, among those experiencing itching CKD-aP was operationalized based on a question from the
at the initial assessment, more than half had persistent Kidney Disease Quality of Life 36-item short form survey.7
symptoms 1 year later. Those in whom itching devel- Patients were asked, “During the past 4 weeks, to what
oped during follow-up were more likely to experience extent were you bothered by itchy skin?” Response op-
tions were “not at all,” “somewhat,” “moderately,” “very
depression, poor sleep, long recovery times after dial-
much,” and “extremely.” For parsimony, we grouped
ysis, and feeling faint or drained. These patients also
patients into 4 categories (2 × 2) based on whether they
rated their quality of life as poorer than those who did were at least moderately bothered by itchy skin at their
not experience itching. These findings emphasize the initial assessment (PQ1) and the assessment approximately
potential value of clinical detection of itching and the 1 year later (PQ2): persistent (yes/yes), resolved (yes/
pursuit of effective treatments for patients receiving no), incident (no/yes), and absent (no/no). This analysis
dialysis experiencing these symptoms. was thus restricted to 7,976 patients who responded to this
question regarding itchy skin on both PQ1 and PQ2
ineffective treatments such as antihistamines or phosphate- (approximately 12 months apart; range, 6-18 months).
lowering dietary advice. Indeed, the use of prescription Patients with fewer than 2 CKD-aP assessments were
medications was ranked as least important by 45% of excluded (Fig 1).
medical directors,3 and 18% of patients who were always Other PROs included as outcome variables were physical
or nearly always bothered by itching reported receiving no and mental component summary scores derived from the
treatment for pruritus.3 short form-12,7 with lower scores indicative of worse
The high prevalence and far-reaching negative effects of quality of life; burden of kidney disease score derived from
CKD-aP makes it prudent to enhance the knowledge sur- the Kidney Disease Quality of Life 36-item short form survey
rounding the course of pruritus and how the chronicity of (7), with lower scores indicative of higher burden; 2 items
this debilitating symptom affects varied aspects of patients’ from the symptoms of kidney disease scale (Kidney Disease
lives. However, many studies evaluate only cross-sectional Quality of Life 36-item short form survey) asking the extent
associations of pruritus with outcomes, and most do not to which patients were bothered by “faintness or dizziness”
focus on PROs specifically. Given the lack of data on the
longitudinal course of CKD-aP in the HD population, this
study seeks to investigate the prevalence of pruritus and
medication use over the course of 12 months and to evaluate
the associations between changes in CKD-aP symptoms and
(1) concurrent change in laboratory values and PROs and
(2) subsequent rates of mortality and hospitalization.

Methods
Data Source
The DOPPS is a prospective cohort study of adult chronic in-
center HD recipients in 21 countries that has been ongoing
since 1996.5,6 HD recipients were randomly selected from
HD facilities in each country; detailed information is
included at http://www.dopps.org. Study approval and
Figure 1. Study flow chart. Note that the group of 16,647 pa-
patient consent were obtained as required by national and tients without pruritus information at patient questionnaire (PQ)
local ethics committee regulations. The DOPPS was 2 was a catch-all for patients who responded to the pruritus
approved by an independent institutional review board (E&I assessment at PQ1 but then had no pruritus information at
study no. 98004-2, latest). This analysis included data from PQ2 for any reason. This includes not only those who were alive
all participating countries in DOPPS phases 4-6 (2009- and enrolled in the Dialysis Outcomes and Practice Patterns
2018). Data on patient demographic characteristics and Study 12 months later and did not respond, but also patients
comorbid conditions were abstracted from medical records who died, transferred to another hemodialysis facility, switched
at DOPPS enrollment, laboratory values and prescriptions modality to peritoneal dialysis or home hemodialysis, received a
were abstracted from medical records at 4-month intervals, transplant, received dialysis in a facility that reached administra-
hospitalizations and mortality (including cause) were tive study end, or were lost to follow-up for any other reason.

AJKD Vol 82 | Iss 6 | December 2023 667

 Sukul et al

and feeling “washed out or drained”; Center for Epidemi- All models were adjusted for patient age, sex, and
ological Studies–Depression score of 10 or greater (indica- baseline covariates of body mass index, dialysis vintage, 13
tive of depressive symptoms)8; a single question from the classes of comorbidity (Table 1), serum albumin level,
Center for Epidemiological Studies–Depression question- hemoglobin level, serum phosphorus level, single-pool
naire asking how many days per week that “my sleep was Kt/V, and catheter as dialysis access. We used multiple
restless”; and a single question asking patients how long it imputation, assuming data were missing at random, to
typically takes to recover from a dialysis session.9 impute missing covariate values using the Sequential
Regression Multiple Imputation Method by IVEware.10
Statistical Analysis Results from 20 such imputed data sets were combined
We reported the distribution of self-reported CKD-aP at for the final analysis using Rubin’s formula.11 The pro-
PQ2 (approximately 12 months after baseline) by baseline portion of missing data was <10% for all covariates with
CKD-aP. We summarized the primary exposure variable, the exceptions of albumin (13%), body mass index (13%),
whether patients were at least moderately bothered by and single-pool Kt/V (25%). All analyses were conducted
itchy skin at PQ1 and/or PQ2 (4 categories: persistent using SAS software, version 9.4 (SAS Institute).
[yes/yes], resolved [yes/no], incident [no/yes], absent
[no/no]), overall and by country. Because the PQ was not Results
administered at exact 12-month intervals, we also sum-
marized the exposure variable by time between PQ as- Within-Patient Pruritus Trajectories
sessments. Baseline patient characteristics were described Among 7,976 eligible patients who responded to the CKD-
for the 4 categories of this longitudinal exposure variable. aP assessments at PQ1 and PQ2, the proportions of patients
We reported the prevalence and incidence of prescrip- at least moderately bothered by itchy skin were 36% at PQ1
tion medications often used to treat CKD-aP: gabapentin, and 36% at PQ2. CKD-aP severity at PQ2 stratified by CKD-aP
pregabalin, nalfurafine, and antihistamines. The indication severity at PQ1 is shown in Fig 2. Among patients not at all
for prescription (pruritus or another reason) was not or somewhat bothered by itchy skin at PQ1, 22% were at
recorded. Among patients at least moderately bothered by least moderately bothered at PQ2, and, among those at least
itchy skin at PQ1, we reported the baseline prevalences of moderately bothered at PQ1, 61% remained at least
these medication prescriptions by country. Among patients moderately bothered at PQ2 (Fig S1). Overall, 44% reported
at least moderately bothered by itchy skin but untreated at the same CKD-aP severity at PQ1 and PQ2, 28% reported
PQ1, we reported the proportion of patients who initiated improvement, and 28% reported worsening (Fig S2).
treatment at some point during the next 12 months. Approximately half of the patients (51%) were at least
To investigate the degree to which laboratory values moderately bothered by itchy skin at PQ1 or PQ2,
(most recently captured within 3 months before CKD-aP including 22% at both PQ1 and PQ2 (yes/yes). We
assessments) and PROs tracked longitudinally with CKD- observed minimal variation in this 4-category exposure
aP symptoms, we presented these outcomes (single mea- variable by country (Fig 3) and by months elapsed be-
surements) at PQ1 and PQ2 stratified by the 4-category tween PQ1 and PQ2 (Fig S3). No striking differences in
exposure variable. Linear mixed-effect models were used enrollment characteristics (eg, demographic data, comor-
to analyze the association between the 4-category exposure bidities) were observed across these 4 groups (Table 1).
variable and outcomes of changes in laboratory values as
well as the change in PROs between PQ1 and PQ2. The Use of Pruritus Treatment
REPEATED statement in SAS software (version 9.4; SAS Among patients at least moderately bothered by itchy skin
Institute) was used to include random intercept accounting at PQ1, 5% were treated with gabapentin or pregabalin; an
for clustering within a facility. additional 22% were prescribed an antihistamine. The
We investigated the 4-category exposure variable and proportions prescribed gabapentin or pregabalin were 2%
subsequent rate of clinical outcomes using Cox regression in Japan, 7% in Europe (ranging from 4% in Italy and
stratified by DOPPS phase and country and using a robust Belgium to 11% in the United Kingdom; Fig S4), and 18%
sandwich covariance estimator to account for facility in North America, and the proportions prescribed an
clustering. Outcomes included all-cause mortality, all- antihistamine were 20% in Europe (ranging from 4% in
cause hospitalization, cardiovascular events (composite Italy to 40% in Sweden; Fig S4), 23% in Japan, and 26%
of cardiovascular death and cardiovascular hospitaliza- in North America (Fig 4). Nalfurafine was available only in
tion), infection events (composite of infection death and Japan and was prescribed to 10% of patients at least
infection hospitalization), and withdrawal from dialysis. moderately bothered by itchy skin at PQ1. Among patients
Time at risk started at PQ2 and ended at the time of the at least moderately bothered by itchy skin at PQ1, 68%
event of interest, 7 days after leaving the facility to transfer were untreated; in this subgroup, 13% initiated treatment
or change renal replacement modality, loss to follow-up, during the next 12 months: 9% with gabapentin or pre-
or administrative end of study phase (whichever gabalin, 3% with an antihistamine, and 2% (4% in Japan)
occurred first). with nalfurafine.

668 AJKD Vol 82 | Iss 6 | December 2023

Sukul et al

Table 1. Baseline Patient Characteristics, by Whether Patients Were at Least Moderately Bothered by Itchy Skin at Initial
Assessment and Approximately 1 Year Later
Bothered by Itchy Skin (Initial/1 y Later)
No/No No/Yes Yes/No Yes/Yes
Characteristic (n = 3,940) (n = 1,130) (n = 1,137) (n = 1,769)
Demographic
Age, y 64 ± 14 65 ± 14 65 ± 14 66 ± 14
Male sex 62% 60% 61% 65%
Black race 2% 3% 3% 2%
Dialysis vintage, y 4.3 (2.2-8.8) 4.1 (2.0-8.2) 4.0 (2.0-7.7) 4.3 (2.1-8.2)
Dialysis access: catheter use 9% 13% 12% 12%
Comorbidity history
Diabetes 35% 41% 41% 42%
Hypertension 85% 87% 86% 86%
Coronary artery disease 27% 32% 29% 33%
Heart failure 15% 19% 18% 21%
Cerebrovascular 11% 15% 13% 16%
disease
Peripheral vascular disease 17% 20% 19% 19%
Other cardiovascular 23% 25% 21% 26%
disease
Gastrointestinal 3% 3% 3% 5%
bleeding
Lung disease 6% 9% 8% 11%
Neurologic disease 5% 9% 5% 7%
Psychiatric disorder 8% 11% 12% 12%
Cancer 13% 12% 11% 13%
Recurrent cellulitis/ 5% 6% 6% 5%
gangrene
HIV 0.7% 0.6% 0.4% 0.5%
Hepatitis C 5% 6% 5% 7%
Data presented as mean ± standard deviation or median (interquartile range) as applicable. Patients were grouped into 4 categories based on whether they were at least
moderately bothered by itchy skin at their initial assessment and another assessment approximately 1 year later.

Pruritus Changes and Laboratory Values were at least moderately bothered by itchy skin at PQ1
Mean laboratory values at PQ1 and PQ2 (and Δ from PQ2 and/or PQ2 (4-category exposure variable). We did not
to PQ1) are summarized in Table 2 by whether patients observe a difference in the Δ laboratory value for patients

Figure 2. Extent to which Dialysis Outcomes and Practice Patterns Study patients were bothered by itchy skin: symptoms at initial
assessment and approximately 1 year later. Note that the proportions of patients who completed the initial assessment but not the
one approximately 1 year later differed minimally by baseline pruritus severity.

AJKD Vol 82 | Iss 6 | December 2023 669

 Sukul et al

Figure 3. Proportion of Dialysis Outcomes and Practice Patterns Study patients at least moderately bothered by itchy skin at the
initial assessment and/or the assessment approximately 1 year later overall and by region. The Europe region included Belgium,
France, Germany, Italy, Spain, Sweden, and the United Kingdom; North America region included Canada and the United States;
and “Other” region included Australia, China, the Gulf Cooperation Council (Bahrain, Kuwait, Oman, Qatar, Saudi Arabia, and
the United Arab Emirates), New Zealand, Russia, and Turkey.

whose CKD-aP symptoms worsened versus improved per week of restless sleep increased from 38% to 48% in
(P > 0.05). For example, mean serum calcium levels patients with incident CKD-aP, remained unchanged at
increased by 0.03 mg/dL among patients with incident 27% in patients with absent CKD-aP, decreased from 48%
CKD-aP and increased by 0.04 among patients with to 40% in patients with resolved CKD-aP, and remained
resolved CKD-aP, suggesting that the prevalence of pruritus unchanged (54%-55%) in patients with persistent CKD-
symptoms was independent from these laboratory values. aP. The associations between the change in pruritus
symptoms and the Δ in PROs were all strong (P < 0.05),
Pruritus Changes and PROs but effect sizes ranged from clinically meaningful (eg,
The mean and prevalence of PROs at PQ1 and PQ2 (and Δ, substantial differences in the proportions of patients
ie, PQ2 – PQ1) are summarized in Table 3 by whether affected by depressive symptoms, feeling faint, and feeling
patients were at least moderately bothered by itchy skin at drained) to nominal (eg, 2- to 3-point difference in
PQ1 and/or PQ2 (4-category exposure variable). In physical and mental component summary scores;
contrast to the analysis of laboratory values, we found that Table 3).
PROs tracked closely with changes in CKD-aP. For The difference between how changes in pruritus
instance, the proportion of patients with at least 3 nights symptoms were associated with changes in laboratory

Figure 4. Pruritus treatment immediately before the initial assessment among patients at least moderately bothered by itchy skin at
the initial assessment stratified by region. Nalfurafine was available only in Japan. The Europe region included Belgium, France, Ger-
many, Italy, Spain, Sweden, and the United Kingdom; North America region included Canada and the United States; and “Other”
region included Australia, China, the Gulf Cooperation Council (Bahrain, Kuwait, Oman, Qatar, Saudi Arabia, and the United
Arab Emirates), New Zealand, Russia, and Turkey. PQ, patient questionnaire.

670 AJKD Vol 82 | Iss 6 | December 2023

Sukul et al

Table 2. Change in Laboratory Measures Between Initial Assessment and Assessment Approximately 1 Year Later Stratified by
Change in Pruritus Symptoms
Bothered by Itchy Skin (Initial/1 y Later) P Value for
No/No No/Yes Yes/No Yes/Yes Adjusted
Measure (n = 3,940) (n = 1,130) (n = 1,137) (n = 1,769) Model
Serum Ca, mg/dL
Initial assessment 8.88 ± 0.73 8.87 ± 0.73 8.88 ± 0.76 8.88 ± 0.78 0.9
1 y later 8.93 ± 0.72 8.90 ± 0.75 8.92 ± 0.75 8.91 ± 0.75
Δ 0.04 ± 0.71 0.03 ± 0.73 0.04 ± 0.72 0.03 ± 0.70
Serum P, mg/dL
Initial assessment 5.17 ± 1.42 5.11 ± 1.49 5.19 ± 1.52 5.28 ± 1.52 0.09
1 y later 5.10 ± 1.41 5.13 ± 1.46 5.13 ± 1.53 5.21 ± 1.49
Δ −0.07 ± 1.43 0.02 ± 1.51 −0.05 ± 1.49 −0.07 ± 1.48
PTH, pg/mL
Initial assessment 184 (99 to 326) 193 (103 to 354) 204 (104 to 354) 186 (94 to 338) 0.5
1 y later 191 (103 to 324) 197 (98 to 363) 194 (101 to 361) 185 (94 to 339)
Δ 5 (−68 to 89) 10 (−78 to 102) 1 (−83 to 87) 5 (−73 to 83)
Albumin, mg/dL
Initial assessment 3.74 ± 0.41 3.70 ± 0.43 3.69 ± 0.44 3.66 ± 0.41 0.2
1 y later 3.74 ± 0.41 3.69 ± 0.43 3.71 ± 0.44 3.64 ± 0.42
Δ −0.01 ± 0.32 −0.01 ± 0.34 0.01 ± 0.33 −0.01 ± 0.34
Hemoglobin, g/dL
Initial assessment 11.06 ± 1.28 10.94 ± 1.33 10.97 ± 1.34 10.96 ± 1.34 0.5
1 y later 11.13 ± 1.21 11.03 ± 1.27 11.04 ± 1.26 11.03 ± 1.31
Δ 0.07 ± 1.36 0.09 ± 1.39 0.07 ± 1.42 0.07 ± 1.46
Single pool Kt/V
Initial assessment 1.48 ± 0.31 1.45 ± 0.31 1.44 ± 0.30 1.44 ± 0.30 0.5
1 y later 1.53 ± 0.29 1.50 ± 0.29 1.49 ± 0.29 1.49 ± 0.29
Δ 0.05 ± 0.20 0.05 ± 0.21 0.05 ± 0.21 0.05 ± 0.21
C-reactive protein, mg/L
Initial assessment 2.0 (0.7 to 5.3) 3.0 (0.9 to 7.0) 2.2 (0.8 to 7.6) 2.3 (0.9 to 6.7) 0.4
1 y later 2.0 (0.7 to 6.0) 3.0 (1.0 to 8.1) 2.2 (0.8 to 7.0) 2.8 (1.0 to 7.0)
Δ 0.0 (−1.0 to 1.5) 0.0 (−1.4 to 2.3) 0.0 (−1.1 to 1.7) 0.0 (−1.0 to 2.0)
Data presented as mean ± standard deviation or median (interquartile range) shown for laboratory values measured concurrently with the pruritus assessment at initial
assessment and assessment approximately 1 year later, along with the difference (Δ) defined as measurement approximately 1 year after the initial measurement minus the
initial measurement. Exposure variable grouped into 4 categories based on whether patients were at least moderately bothered by itchy skin at their initial assessment and
assessment approximately 1 year later. Linear mixed-effect models were used to analyze the change in laboratory measurements between initial assessment and
assessment approximately 1 year later accounting for facility clustering and adjusted for Dialysis Outcomes and Practice Patterns Study phase, country, age, sex, body
mass index, dialysis vintage, 13 comorbidities, albumin, hemoglobin, phosphorus, single-pool Kt/V, and catheter use. Type 3 P values represent the overall effect of the 4-
category exposure variable.

values and changes in PROs is illustrated in Fig 5. hospitalization, 1.48 (1.26-1.74) for cardiovascular
Figure 5A shows no distinction in the patterns of serum events, 1.01 (0.80-1.29) for infection events, and 1.32
calcium level across the 4 exposure groups; in contrast, Fig (0.74-2.35) for withdrawal from dialysis. Results were
5B shows how the proportion of patients with poor sleep qualitatively similar for patients with incident CKD-aP but
quality was consistently high in the yes/yes group, mixed for patients with resolved CKD-aP (Fig 6).
consistently low in the no/no group, and changed with
pruritus symptoms in the no/yes and yes/no groups.
Detailed results for other laboratory examinations and Discussion
PROs are shown in Tables 2 and 3. The present international cohort study is the largest of its
kind to evaluate the longitudinal course of pruritus and its
Pruritus Changes and Clinical Outcomes associated outcomes and adds incrementally to prior
Median follow-up time from PQ2 was 15 (interquartile literature with several major findings. First, 51% of pa-
range, 9-21) months, and the mortality rate was 8.7 per tients were at least moderately bothered by CKD-aP at the
100 patient-years. Compared with patients with absent initial assessment or the assessment approximately 1 year
CKD-aP, the adjusted hazard ratios for patients with later, with nearly a quarter bothered both at baseline and
persistent CKD-aP were 1.29 (95% CI, 1.09-1.53) for all- 12 months later. Second, only 5% of patients bothered by
cause mortality, 1.17 (1.07-1.28) for all-cause CKD-aP at baseline were treated with gabapentin or

AJKD Vol 82 | Iss 6 | December 2023 671

 Sukul et al

Table 3. Change in PROs Between Initial Assessment and Assessment Approximately 1 Year Later Stratified by Change in Pruritus
Symptoms
Bothered by Itchy Skin (Initial/1 y Later) P Value for
No/No No/Yes Yes/No Yes/Yes Adjusted
PRO (n = 3,940) (n = 1,130) (n = 1,137) (n = 1,769) Model
PCSa
Initial assessment 41.7 ± 10.3 38.8 ± 9.9 37.9 ± 10.5 35.6 ± 9.7 <0.001
1 y later 41.4 ± 10.4 36.2 ± 10.0 38.3 ± 10.5 34.9 ± 9.9
Δ −0.4 ± 8.3 −2.5 ± 8.6 0.4 ± 9.5 −0.6 ± 8.1
MCSa
Initial assessment 48.7 ± 10.7 45.1 ± 10.8 44.3 ± 10.9 42.2 ± 10.6 0.004
1 y later 48.1 ± 10.6 43.5 ± 10.8 45.0 ± 10.9 41.5 ± 10.7
Δ −0.6 ± 9.9 −1.6 ± 11.3 0.7 ± 11.5 −0.7 ± 10.5
Effects of kidney
diseaseb
Initial assessment 73.1 ± 18.8 65.0 ± 20.4 61.8 ± 22.7 56.4 ± 21.9 < 0.001
1 y year later 72.7 ± 19.4 58.3 ± 21.7 67.1 ± 20.8 54.8 ± 22.6
Δ −0.4 ± 16.0 −6.7 ± 21.5 5.4 ± 21.0 −1.7 ± 19.1
CES-D score≥10
(depression)c
Initial assessment 31% 43% 54% 61% < 0.001
1 y year later 33% 56% 49% 63%
Δ 2% 13% -5% 3%
≥3 nights per week
restless sleepd
Initial assessment 27% 38% 48% 54% < 0.001
1 y later 27% 48% 40% 55%
Δ 0% 10% -8% 1%
>6h to recover from
HD sessione
Initial assessment 16% 21% 23% 27% < 0.001
1 y later 15% 27% 22% 28%
Δ −1% 6% −1% 1%
Feeling faint (at least
moderately bothered)f
Initial assessment 14% 20% 32% 37% < 0.001
1 y later 14% 36% 22% 39%
Δ 0% 16% −10% 3%
Feeling drained (at least
moderately bothered)f
Initial assessment 26% 40% 47% 62% < 0.001
1 y later 27% 54% 35% 65%
Δ 1% 14% −12% 3%
Exposure variable grouped into 4 categories based on whether patients were at least moderately bothered by itchy skin at their initial assessment and assessment
approximately 1 year later. Prevalence (%) or mean ± standard deviation shown for PROs measured concurrently with the pruritus assessment at initial assessment and
assessment approximately 1 year later, along with the difference (Δ) defined as assessment approximately 1 year later minus initial assessment. Linear mixed-effect models
were used to analyze the change in PROs between initial assessment and assessment approximately 1 year later, accounting for facility clustering, and adjusted for Dialysis
Outcomes and Practice Patterns Study phase, country, age, sex, body mass index, dialysis vintage, 13 comorbidities, albumin, hemoglobin, phosphorus, single-pool Kt/V,
and catheter use. Type 3 P values represent the overall effect of the 4-category exposure variable. Abbreviations: CES-D, Center for Epidemiological Studies–Depression;
HD, hemodialysis; MCS, mental component summary; PCS, physical component summary; PRO, patient-reported outcome.
a
PCS and MCS are derived from the short form-12, with lower scores indicative of poorer quality of life.
b
Burden of kidney disease score was derived from the Kidney Disease Quality of Life 36-item short form survey, with lower scores indicative of greater disease burden.
c
A CES-D score ≥10 is indicative of depressive symptoms.
d
A single question from the CES-D asks how many days per week “my sleep was restless.”
e
A single question asking patients how long it typically takes to recover from a dialysis session.
f
The extent to which patients were bothered by “faintness or dizziness” and feeling “washing out or drained” were derived from individual items from the Kidney Disease
Quality of Life 36-item short form survey.

pregabalin; among patients bothered but untreated at approximately 1 year later between patients whose CKD-aP
baseline, only 9% initiated treatment during the next 12 symptoms improved versus worsened. In contrast, those
months. Third, there were minimal differences in labora- with incident CKD-aP experienced worsening symptoms
tory values between initial assessment and assessment for other PROs, including physical and mental quality of

672 AJKD Vol 82 | Iss 6 | December 2023

Sukul et al

Figure 5. Graphic display of selected outcomes from Table 2 (serum calcium) and Table 3 (sleep quality); results for other outcomes
are detailed in Table 2 (laboratory tests) and Table 3 (patient-reported outcomes). The exposure variable was grouped into 4 cate-
gories based on whether patients were at least moderately bothered by itchy skin at their initial assessment and an assessment
approximately 1 year later; linear mixed-effect models were used to analyze the change between initial assessment and the assess-
ment approximately 1 year later in (A) serum calcium and (B) proportion of patients with poor sleep quality accounting for facility
clustering and adjusted for Dialysis Outcomes and Practice Patterns Study phase, country, age, sex, body mass index, dialysis vin-
tage, 13 comorbidities, albumin, hemoglobin, phosphorus, single-pool Kt/V, and catheter use.

life, depression, restless sleep, prolonged recovery from a evaluated the prevalence of pruritus over a 12-week
dialysis session, and feeling faint and drained. period, 84% reported having itching daily or nearly daily
Among patients at least moderately bothered by pruri- and nearly 60% had reported daily or nearly daily itching
tus at baseline, 61% remained at least moderately bothered for more than 1 year at baseline.12 In another longitudinal
1 year later, emphasizing the chronicity of the condition. study of pruritus that used quarterly assessments within
In one of the first longitudinal studies of CKD-aP, which the Dutch renal registry between 2018 and 2020, pruritus

Figure 6. Association between change in pruritus symptoms and subsequent rates of clinical outcomes. Follow-up for events began
after the pruritus assessment approximately 1 year after the initial assessment. Cox regression models stratified by Dialysis Outcomes
and Practice Patterns Study phase and country and using a robust sandwich covariance estimator to account for facility clustering
were adjusted for age, sex, body mass index, dialysis vintage, 13 comorbidities, albumin, hemoglobin, phosphorus, single-pool Kt/V,
and catheter use. Cardiovascular and infection events were a composite of cause-specific mortality and cause-specific hospitaliza-
tions. Abbreviation: HD, hemodialysis.

AJKD Vol 82 | Iss 6 | December 2023 673

 Sukul et al

was prevalent in nearly 50% of patients at baseline and a decrease in the prevalence of depression, and those with
persisted over time in nearly 70% of these patients.13 persistent CKD-aP had the highest depression prevalence.
The majority of patients bothered by CKD-aP were not This pattern was also seen for restless sleep, prolonged
taking any medications traditionally prescribed for pruri- recovery from a dialysis session, feeling faint, and feeling
tus, which may be for multiple reasons. First, providers are drained. Similar trends were observed for more general
not always aware of the severity of pruritus afflicting their measures of quality of life, including physical and mental
patients, likely as a result of underestimation by the pro- component summary scores and recovery time from a
vider and underreporting by the patient.3 Second, there is a dialysis session, although the magnitudes of those associ-
lack of clarity as to what constitutes effective treatment, as ations were smaller than for the more targeted PROs
phosphorus control was ranked as most important in every described above. Prior studies have demonstrated cross-
DOPPS country and overall by 60% of medical directors sectional associations between poorer PROs and pruri-
when ranking therapeutic options for patients with severe tus,1,2,21-23 but only a few have shown these associations
pruritus.3 Third, the use of prescription medications was in longitudinal assessments,12,13 and our study extends
ranked least important by nearly half of medical directors.3 these results to confirm PRO changes in parallel fashion in
Here we present the prevalence and initiation of med- patients whose pruritus resolved or worsened over the
ications often used to treat pruritus. Gabapentin, pre- course of 12 months.
gabalin, and nalfurafine have all been shown to reduce itch We also examined clinical outcomes and observed
in clinical trials,14 even though evidence remains limited. higher mortality rates among patients bothered by pruritus
However, consistent with findings from Rayner et al,3 very at either time point—not only in those with persistent or
few (10%-15%) patients bothered by CKD-aP in the pre- incident CKD-aP, but also those with resolved CKD-
sent study were prescribed gabapentin, pregabalin, or aP—than among patients with absent CKD-aP. Prior
nalfurafine. studies have shown higher rates of all-cause and cause-
Rayner et al demonstrated that the vast majority (93%) specific mortality with worse pruritus severity based on a
of nephrologists prescribed antihistamines for CKD-aP, single baseline CKD-aP assessment.1,2,22,24 Specifically, it
often as first-line therapy3; however, not a single has been theorized that increased cardiovascular mortality
placebo-controlled study was found to be of sufficient could be related to higher levels of inflammation or
quality to be included in the recent 2020 Cochrane review prevalence of heart failure, and infection-related mortality
of treatments for CKD-aP, and, when compared with other could be related to suppressed immune systems or the
drugs, antihistamines typically were found to be less presence of a catheter.1 The association between pruritus
effective, although results were highly variable.15 Difeli- and hospitalization has been previously shown1; we
kefalin, which was recently approved by the US Food and expanded on these prior results to show that hospitaliza-
Drug Administration and European Medicines Agency, has tion rates were highest among patients with persistent
gained traction as a new peripherally specific, highly se- CKD-aP and cardiovascular-related event rates were highest
lective κ-opioid agonist, demonstrating significant reduc- in those affected by CKD-aP at the assessment approxi-
tion in itch intensity and improved itch-related quality of mately 1 year later regardless of whether they were
life compared with a placebo.16 affected at the initial assessment.
In the present study, the longitudinal change in labo- Adjustment to pruritus as a new symptom may contrast
ratory values varied minimally by change in pruritus with patient expectations of dialysis or could simply add to
symptoms. Indeed, despite the previously held beliefs of a the already-high symptom burden. In our analysis of
relationship between CKD-aP and bone mineral disease withdrawal from dialysis, patients with incident CKD-aP
markers, most studies have found no relationship between had the highest withdrawal rate, even higher than those
pruritus and phosphorus levels.3,12,17-20 The most recent with persistent CKD-aP, who may have been forced to
DOPPS analysis of HD recipients also showed no notable adjust to this symptom over time. Although there are many
differences in parathyroid hormone or calcium levels factors related to dialysis withdrawal, this reinforces the
across pruritus severity.1 Results from this longitudinal need to screen for pruritus on a regular basis and focus on
analysis demonstrate that, although some patient charac- timely and effective relief for patients newly impacted by
teristics may be cross-sectionally associated with CKD-aP CKD-aP. The need to interpret these results with caution is
symptoms, this does not necessarily translate into labora- evident considering the wide and overlapping CIs seen in
tory values tracking with longitudinal changes in CKD-aP the results (Fig 4).
symptoms. There are a few limitations to the present study that are
In contrast with laboratory values, we observed clear worth noting. First, it is unclear if the CKD-aP symptoms
longitudinal changes in measures of depression (Center for patients experienced were continuous throughout the
Epidemiological Studies–Depression score), sleep quality, duration of the year or intermittent between the initial
feeling faint, and feeling drained that were directionally assessment and the assessment approximately 1 year later.
consistent with changes in pruritus symptoms. Whereas Further research is needed to prospectively capture more
patients with incident CKD-aP had an increase in the frequent assessments of CKD-aP. Second, the possibility of
prevalence of depression, those with resolved CKD-aP had recall or misclassification bias may have affected the

674 AJKD Vol 82 | Iss 6 | December 2023

Sukul et al

results. Third, the indication for prescription medications Thilo Schaufler, MD, Elham Asgari, MD, Fitsum Guebre-Egziabher,
was unknown; medicines commonly used to treat pruritus MD, Li Zho, MD, Mohammed Abdulrahman Al-Ghonaim, MD,
Kosaku Nitta, MD, Bruce M. Robinson, MD, and Angelo
may have been prescribed for an indication other than Karaboyas, PhD.
pruritus, and some patients may have received treatment
Authors’ Affiliations: Division of Nephrology, Department of Internal
before study enrollment or between the 4-month intervals Medicine, University of Michigan (NS, BMR), Division of Nephrology,
but discontinued it because of side effects. Fourth, the Veterans Affairs Ann Arbor Health System (NS), and Arbor Research
relationships with outcomes in this study are only associ- Collaborative for Health (JZ, RLP, BMR, AK), Ann Arbor, MI; CSL-
ations; causal statements about these associations cannot be Vifor, Glattbrugg, Switzerland (SW, TS); Department of
made. Fifth, prescription information was abstracted from Nephrology, Guy’s St Thomas Hospital, London, United Kingdom ˇ
(EA); Department of Nephrology Dialysis Hypertension, Hopital
patients’ medical records at the dialysis unit, so it is Edouard Herriot, Hospices Civils de Lyon, Laboratoire de
possible that the proportion of patients initiating treat- Recherche en Cardiovasculaire, M
etabolisme, Diab
etologie et
ments for pruritus was underreported in the study. Finally, Nutrition, Institut National de la Sant
e et de la Recherche
patients who died or withdrew from HD within 12 M
edicale 1060, University Lyon-1, Lyon, France (FG-E);
months after completion of PQ1 will not have completed Department of Nephrology, Peking University People’s Hospital,
Beijing, China (LZ); Department of Medicine, King Khalid
PQ2 and, by definition, would have been excluded from University Hospital (MAA-G) and College of Medicine, King Saud
the analysis. Excluding patients who did not survive to PQ2 University (MAA-G), Riyadh, Saudi Arabia; and Department of
could potentially create a selection bias to the extent that Nephrology, Tokyo Women’s Medical University, Tokyo, Japan (KN).
CKD-aP at PQ1 was associated with mortality. Address for Correspondence: Nidhi Sukul, MD, Division of
In conclusion, at least half of patients receiving chronic Nephrology, Department of Internal Medicine, University of
HD were affected by CKD-aP over the course of 1 year, and Michigan, 1500 E Medical Center Dr, SPC 5364, Ann Arbor, MI
48109-5364. Email: [email protected]
symptoms remained unresolved 12 months later for the
majority of HD recipients bothered by itchy skin at baseline. Authors’ Contributions: Study or design/or design, data
acquisition, and interpretation of results: NS, JZ, RLP, SW, TS,
Concurrent worsening of PROs for those with incident CKD- BMR, AK; supervision or mentorship: EA, FG-E, LZ, MAA-G, KN.
aP highlights a possible effect of patients’ significant Each author contributed important intellectual content during
adjustment to this new, chronic symptom, emphasizing the manuscript drafting or revision and agrees to be personally
need for a better consensus among providers to standardize accountable for the individual’s own contributions and to ensure
frequent assessment of CKD-aP symptoms so the condition that questions pertaining to the accuracy or integrity of any portion
of the work, even one in which the author was not directly
can be identified earlier. Early patient education about CKD- involved, are appropriately investigated and resolved, including
aP and setting/managing expectations about CKD-aP may with documentation in the literature if appropriate.
increase the likelihood of patients to report and more Support: This manuscript was directly supported by CSL-Vifor.
effectively cope with this symptom. More work is needed to Global support for the ongoing DOPPS Programs is provided
elucidate the etiology of pruritus, potentially uncovering without restriction on publications by a variety of funders. For
whether its physiologic effect has a causal effect on increased details, see https://www.dopps.org/AboutUs/Support.aspx. As of
April 21, 2023, the DOPPS program is supported by Amgen Inc
mortality. Future studies should also focus on the impact of (since 1996, founding sponsor); Akebia Therapeutics Inc; Astellas
treatments on not only CKD-aP severity, but relevant adverse Pharma Inc; Bard Peripheral Vascular Inc; Baxter Healthcare Corp;
outcomes as well. Bayer AG & Bayer Yakuhin Ltd; Cara Therapeutics Inc; Chugai
Pharmaceutical Co, Ltd; GlaxoSmithKline LLC; Japanese Society
for Peritoneal Dialysis; JMS Co, Ltd; Kidney Foundation Japan;
Supplementary Material Kissei Pharmaceutical Co, Ltd; Kyowa Kirin Co, Ltd (since 1999
Supplementary File (PDF) for Japan DOPPS); Merck Sharp & Dohme Corp; Nikkiso Co, Ltd;
Figure S1: Extent to which HD recipients were bothered by itchy ONO Pharmaceutical Co, Ltd; Terumo Corporation; Torii
Pharmaceutical Co, Ltd; and CSL-Vifor Ltd; along with public
skin in an assessment approximately 1 year after the first assess-
funding by agencies in France, Thailand, the United Kingdom, and
ment stratified by initial assessment (dichotomized).
the United States. Jennifer McCready-Maynes, an employee of
Figure S2: Extent to which HD recipients were bothered by itchy Arbor Research Collaborative for Health, provided editorial
skin in an assessment approximately 1 year later versus the initial assistance on this paper. Arbor Research Collaborative for Health
assessment: better, the same, or worse. receives funding from CSL-Vifor to support the costs of
Figure S3: Proportion of HD recipients at least moderately bothered conducting this research, including salary support for Arbor
Research personnel performing data management and analysis,
by itchy skin in the initial assessment and/or assessment approxi-
manuscript development, and submission support, data licensing,
mately 1 year later overall and by time elapsed between the initial
and administrative fees, and others. CSL-Vifor did not have a role
assessment and the assessment approximately 1 year later. in study design, data collection, analysis, reporting, or the decision
Figure S4: Pruritus treatment immediately before initial pruritus to submit for publication.
assessment among patients at least moderately bothered by itchy Financial Disclosure: Dr Sukul has received fees from Medscape
skin in the initial assessment by country in Europe. for a one-time lecture on pruritus and from Karger Publishing for
cowriting a handbook on pruritus and developing an online
learning module, all paid directly to her. Dr Asgari is the chair of
Article Information the British Transplant Society Standards and Guidelines
Authors’ Full Names and Academic Degrees: Nidhi Sukul, MD, Committee and Co-Country investigator for DOPPS in the United
Junhui Zhao, PhD, Ronald L. Pisoni, PhD, Sebastian Walpen, MD, Kingdom. Drs Walpen and Schaufler are employees of CSL-Vifor.

AJKD Vol 82 | Iss 6 | December 2023 675

 Sukul et al

Drs Zhao, Pisoni, Robinson, and Karaboyas are employees of Arbor 11. Little RJA, Rubin DB. Statistical Analysis with Missing Data.
Research Collaborative for Health, which administers the DOPPS Wiley; 1987.
Programs. Dr Robinson has also received consultancy fees or 12. Mathur VS, Lindberg J, Germain M, et al; ITCH National Reg-
travel reimbursement since 2019 from AstraZeneca, istry Investigators. A longitudinal study of uremic pruritus in
GlaxoSmithKline, Kyowa Kirin Co, and Monogram Health, all paid hemodialysis patients. Clin J Am Soc Nephrol. 2010;5:1410-
directly to his institution of employment. The remaining authors 1419. doi:10.2215/CJN.00100110
declare that they have no relevant financial interests. 13. van der Willik EM, Lengton R, Hemmelder MH, et al. Itching
Peer Review: Received October 31, 2022. Evaluated by 2 external in dialysis patients: impact on health-related quality of life
peer reviewers, with direct editorial input from a Statistics/Methods and interactions with sleep problems and psychological
Editor, an Associate Editor, and the Editor-in-Chief. Accepted in symptoms - results from the RENINE/PROMs registry.
revised form April 16, 2023. Nephrol Dial Transplant. 2022;37(9):1731-1741. doi:10.
1093/ndt/gfac022
14. Hercz D, Jiang SH, Webster AC. Interventions for itch in people
with advanced chronic kidney disease. Cochrane Database
References Syst Rev. 2020;12:CD011393. doi:10.1002/14651858.
1. Sukul N, Karaboyas A, Csomor PA, et al. Self-reported pruritus CD011393.pub2
and clinical, dialysis-related, and patient-reported outcomes in 15. Rayner H, Sukul N. Fast Facts: Kidney Itch; CKD-Associated
hemodialysis patients. Kidney Med. 2020;3(1):42-53.e1. doi: Pruritus: Under-Recognized and Under-Treated. S. Karger;
10.1016/j.xkme.2020.08.011 2022. doi:10.1159/isbn.978-3-318-06513-8
2. Pisoni RL, Wikstr€ om B, Elder SJ, et al. Pruritus in haemodialysis 16. Fishbane S, Jamal A, Munera C, Wen W, Menzaghi F. A phase 3
patients: international results from the Dialysis Outcomes and trial of difelikefalin in hemodialysis patients with pruritus. N Engl
Practice Patterns Study (DOPPS). Nephrol Dial Transplant. J Med. 2020;382(3):222-232. doi:10.1056/NEJMoa1912770
2006;21:3495-3505. doi:10.1093/ndt/gfl461 17. Amatya B, Agrawal S, Dhali T, Sharma S, Pandey SS. Pattern
3. Rayner HC, Larkina M, Wang M, et al. International compari- of skin and nail changes in chronic renal failure in Nepal: a
sons of prevalence, awareness, and treatment of pruritus in hospital-based study. J Dermatol. 2008;35:140-145. doi:10.
people on hemodialysis. Clin J Am Soc Nephrol. 2017;12: 1111/j.1346-8138.2008.00433.x
2000-2007. doi:10.2215/CJN.03280317 18. Rayner H, Baharani J, Smith S, Suresh V, Dasgupta I. Uraemic
4. Sukul N, Speyer E, Tu C, et al. Pruritus and patient reported pruritus: relief of itching by gabapentin and pregabalin.
outcomes in non-dialysis CKD. Clin J Am Soc Nephrol. Nephron Clin Pract. 2012;122:75-79. doi:10.1159/
2019;14(5):673-681. doi:10.2215/CJN.09600818 000349943
5. Young EW, Goodkin DA, Mapes DL, et al. The Dialysis Out- 19. Khanna D, Singal A, Kalra OP. Comparison of cutaneous
comes and Practice Patterns Study (DOPPS): an international manifestations in chronic kidney disease with or without dial-
hemodialysis study. Kidney Int. 2000;57(suppl 74):S74-S81. ysis. Postgrad Med J. 2010;86:641-647. doi:10.1136/pgmj.
doi:10.1046/j.1523-1755.2000.07413.x 2009.095745
6. Pisoni RL, Gillespie BW, Dickinson DM, et al. The Dialysis 20. Shirazian S, Kline M, Sakhiya V, et al. Longitudinal predictors of
Outcomes and Practice Patterns Study (DOPPS): design, uremic pruritus. J Ren Nutr. 2013;23:428-431. doi:10.1053/j.
data elements, and methodology. Am J Kidney Dis. jrn.2013.08.002
2004;44(suppl 2):7-15. doi:10.1053/j.ajkd.2004.08.005 21. Ramakrishnan K, Bond TC, Claxton A, et al. Clinical charac-
7. Hays RD, Kallich JD, Mapes DL, et al. Development of the teristics and outcomes of end-stage renal disease patients with
Kidney Disease Quality of Life (KDQOL) instrument. Qual Life self-reported pruritus symptoms. Int J Nephrol Renovasc Dis.
Res. 1994;3:329-338. doi:10.1007/BF00451725 2014;7:1. doi:10.2147/IJNRD.S52985
8. Kohout FJ, Berkman LF, Evans DA, Cornoni-Huntley J. Two 22. Kimata N, Fuller DS, Saito A, et al. Pruritus in hemodialysis
shorter forms of the CES-D Depression Symptoms Index. patients: results from the Japanese Dialysis Outcomes and
J Aging Health. 1993;5:179. doi:10.1177/089826439300 Practice Patterns Study (JDOPPS). Hemodial Int. 2014;18(3):
500202 657-667. doi:10.1111/hdi.12158
9. Rayner HC, Zepel L, Fuller DS, et al. Recovery time, quality of 23. Lopes GB, Nogueira FC, de Souza MR, et al. Assessment of
life, and mortality in hemodialysis patients: the Dialysis Out- the psychological burden associated with pruritus in hemodi-
comes and Practice Patterns Study (DOPPS). Am J Kidney alysis patients using the kidney disease quality of life short
Dis. 2014;64(1):86-94. doi:10.1053/j.ajkd.2014.01.014 form. Qual Life Res. 2012;21:603-612. doi:10.1007/s11136-
10. Raghunathan TE, Solenberger PW, Van Hoewyk J. IVE- 011-9964-x
ware: Imputation and Variance Estimation Software: User 24. Narita I, Aichi B, Omori K, et al. The effect of poly-
Guide. Institute for Social Research, University of Michi- methylmethacrylate dialysis membranes on uremic pruritus.
gan; 2002. Kidney Int. 2006;69:1626-1632. doi:10.1038/sj.ki.5000251

676 AJKD Vol 82 | Iss 6 | December 2023