88 Full
88 Full
88 Full
Ann Rheum Dis: first published as 10.1136/annrheumdis-2020-218323 on 25 September 2020. Downloaded from http://ard.bmj.com/ on August 29, 2021 by guest. Protected by copyright.
Preliminary predictive criteria for COVID-19
cytokine storm
Roberto Caricchio ,1 Marcello Gallucci ,2 Chandra Dass,3 Xinyan Zhang,1
Stefania Gallucci ,4 David Fleece,5 Michael Bromberg,6 Gerard J Criner,7 Temple
University COVID-19 Research Group
Ann Rheum Dis: first published as 10.1136/annrheumdis-2020-218323 on 25 September 2020. Downloaded from http://ard.bmj.com/ on August 29, 2021 by guest. Protected by copyright.
on admission and for at least the first 7 days of hospitalisation.
Table 1 Demographics and comorbidities in the cohort of patients
Eighty-two patients were enrolled in clinical trials with biologics
with COVID-19
and their laboratory results were initially included in the anal-
yses. The rest of the patients were clinically followed up and 64 Clinical consensus
Patients with COVID-19 P value
were considered in CS by a consensus between the pulmonolo- All No storm Storm
gists and rheumatologists. The initial consensus was based on the Numbers 513 449 64 <0.001
application of both MAS and HLH criteria; however, among the % 100 88 12
first few patients, very few met these criteria despite worsening Sex (%)
clinical status and elevation of inflammatory markers. Hence, a Females 43 45 33 <0.001
newly devised consensus was based on (1) worsening respiratory
Males 57 55 66
status defined as increased oxygen supplementation required to
Age (years) 58.3 57.7 62.2 0.041
maintain SpO2>93% and (2) elevation above threefold the upper
Race/ethnicity (%)
normal level of at least two of the following markers: C reactive
AA 53 54 53 n.s.
protein (CRP), ferritin, D-dimer, lactate dehydrogenase (LDH)
EA 11 11 14 n.s.
and cardiac troponin. Patients in this group were retrospectively
selected as the basis for the following statistical analyses. Hispanic 23 22 19 n.s.
To validate the results, an additional cohort of 258 patients Other 9 9 9 n.s.
was collected from all the patients admitted to Temple University Unknown 4 4 5 n.s.
Hospital from 18 April 2020 to 30 April 2020. Inclusion criteria Comorbidities (%)
were the same of the first cohort. Lung disease 26 24 36 0.077
Hypertension 69 68 72 n.s.
Obesity 52 51 55 n.s.
Statistical analyses Heart disease 25 25 25 n.s.
A series of univariate logistic regressions were used to assess the Smoking history 42 43 36 n.s.
association between each laboratory variable and the presence of Diabetes 48 50 36 0.057
the CS, with the criterion of the clinical consensus of the medical
Patients with diagnosis of COVID-19 infection and chest high-resolution CT with
group indicating the presence of the storm (see previous discus- ground-glass opacity were divided according to a consensus of clinicians for a
sion). The predictors were the laboratory variables, aggregated diagnosis of cytokine storm. P values were calculated using χ2 test for frequencies
by using each patient average up to the day when the clinicians and t-test for age.
made the consensus of CS or the first 7 days of hospitalisation Italics indicate significant p values.
for patients not diagnosed in storm. Due to the presence of AA, African–American; EA, European–American; n.s., not significant at α≥0.05.
missing values, only predictors obtained in at least 300 patients
were considered.
Predictors showing a significant odds ratio, at alpha-level 0.05, RESULTS
were then analysed by principal component analysis (PCA) with Description of the cohort
promax rotation to cluster them in coherent groups. We retained In this retrospective study, we investigated 513 patients who
the components (clusters) having eigenvalues larger than 1. Each presented GGOs by chest HRCT. Ninety-five per cent of the
laboratory variable was associated with the cluster in which it patients were also COVID-19 positive by RT-PCR. Of these 513,
featured the highest factor loading. 64 patients were eventually determined to be in CS and treated
Cut-off values for each individual laboratory variable were with biologics, such as monoclonal antibodies against IL-6R and
estimated using a genetic algorithm22 as implemented by recombinant IL-1R antagonist (table 1). Table 1 shows the demo-
Scrucca23 (for more details, see online supplemental methods). graphics of the cohort, comparing patients who reached or not a
In the algorithm, a population of 500 sets of cut-off values was clinical consensus for CS. As previously reported in COVID-19
defined, with mutation probability of 0.1 and crossover proba- pneumonia, more patients were male, and the average age was
bility of 0.8. In each generation, 5% of the sets of cut-off values 58.3. Mirroring the population that our hospital serves, most
were selected based on their fitness. The fitness function maxi- patients were African–American and Hispanic. Frequent comor-
mised the geometric mean of the sensitivity and specificity of the bidities included hypertension, obesity, diabetes and smoking
classification (confusion table) predicting the (COVID-19 cyto- history. We did not find any statistically significant difference in
kine storm (COVID-CS)) groups obtained with a given set of the distribution of race and comorbidities between the patients
cut-off values. The stopping rule was set to 200 generations with in storm or not, while older male patients were slightly more
no improvement in fitness. In order to develop cut-offs that can likely to develop CS, suggesting that sex and age, but not race
be feasibly used in the clinic, daily laboratory data were used in and specific comorbidities, increase the risk of developing CS
the genetic algorithm fitness function. A patient was classified as during COVID-19 infection.
COVID-CS positive when the criteria were met at least for 1 day.
When a laboratory value was not present for a patient 1 day, the COVID-CS does not meet the 2004 HLH criteria and HScore
most recent available value was used. To understand the type of CS occurring during COVID-19
To evaluate the stability of the cut- off values, a bootstrap infection, we determined the number of patients in our cohort
procedure was employed to compute the CI classification statis- who fulfilled the HLH criteria and had an HScore ≥169 (online
tics (accuracy, sensitivity and specificity). For each statistic, supplemental tables S1 and S2)4 24 using the averages of labora-
a distribution of bootstrap estimates was created across 5000 tory tests performed during the first 7 days of hospitalisation.
bootstrap samples, and the 95% CIs were obtained by setting the We found that only 10 out of 513 patients fit the 2004 HLH
2.5th and 97.5th percentiles of the bootstrap distribution as the criteria (table 2), and most patients (8/10) did not fulfil the clin-
interval boundaries. Finally, the criteria and cut-off values were ical consensus of COVID-19 storm. We also found that 43 out
also validated on the second cohort of patients. of 513 patients had an HScore of >169, but only 12 also met the
Caricchio R, et al. Ann Rheum Dis 2021;80:88–95. doi:10.1136/annrheumdis-2020-218323 89
Criteria
Ann Rheum Dis: first published as 10.1136/annrheumdis-2020-218323 on 25 September 2020. Downloaded from http://ard.bmj.com/ on August 29, 2021 by guest. Protected by copyright.
Table 2 HLH, HScore and MAS criteria applied to the COVID-19 cohort
Clinical consensus storm H Score Clinical consensus storm MAS Clinical consensus storm
HLH No Yes No Yes No Yes
No 441 62 No 418 52 No 443 63
Yes 8 2 Yes 31 12 Yes 6 1
For HLH criteria, the specificity was 0.98 and the sensitivity was 0.2. For HScore, the specificity was 0.93 and the sensitivity was 0.28. For MAS, the specificity was 0.98 and the
sensitivity was 0.14.
HLH, haemophagocytic lymphohistiocytosis; MAS, macrophage activation syndrome.
clinical consensus of COVID-19 storm (table 2). In our analyses developed a CS did not meet the HLH criteria and the HScore24
of the HLH criteria and HScore, most patients admitted with performed poorly as previously suggested (table 2).12
COVID-19 infection did not have splenomegaly (not shown)
nor cytopenias affecting at least two cell lineages in the periph-
COVID-CS does not meet the 2016 MAS criteria
eral blood. On the contrary, they had normal absolute numbers
We analysed whether our cohort fulfilled the MAS criteria,
of monocytes and increased numbers of neutrophils. They also
reported in online supplemental table S3,5 and found that only
had normal or increased levels of fibrinogen, typically low in
7/513 did (table 2). Six out of seven of these patients were not
HLH, and mostly normal triglycerides, which are frequently
clinically found in storm, therefore fulfilling neither the labora-
increased in HLH4 (online supplemental table S4). All patients
tory nor the clinical judgement of MAS. These patients did not
with COVID-19 had high levels of serum ferritin. In addition,
fit the MAS criteria due to the uncommon presence of throm-
CRP, which is not included in the HLH 2004 criteria (online
bocytopenia, increased levels of fibrinogen and the relatively
supplemental table S1), was also elevated. We could not eval-
normal levels of triglycerides in the COVID-19-infected patients
uate natural killer (NK)-cell activity and the level of soluble
(online supplemental table S4).
interleukin-2 receptor (sIL-2R), since the results of these tests
were not rapidly available at our medical centre or in most
hospitals. Nevertheless, reports of other cohorts of patients with Criteria to predict the COVID-CS
COVID-19 showed sIL-2R levels below those considered in the Since most patients with COVID-19 in CS did not meet the clas-
HLH criteria.25 The analysis of haemophagocytosis in the bone sification criteria of HLH, HScore or MAS, we next followed
marrow or in secondary lymphoid organs was deemed unneces- the strategy depicted in figure 1 and analysed the predictive
sary, considering its invasiveness. Even if the tests that we did not power of 62 laboratory tests available in our hospital (table 3
perform were hypothetically positive, the majority of patients and online supplemental table S4). We aimed to find novel
did not meet the five out of eight criteria of HLH because the criteria to identify patients in CS. In order to reach a predictive
majority fulfilled only two—fever and hyperferritinemia. There- power that can be clinically useful to diagnose a COVID-C S,
fore, our results suggest that most patients in our cohort who we used the mean values of laboratory results of the first 7
th th
Data Collection March 10 -April 17 , 2020
Figure 1 Research strategy. Flowchart of the experimental strategy followed in the generation of the new criteria aimed to recognise the
cytokine storm in patients with COVID-19. COVID-CS, COVID-19 cytokine storm; GGO, ground-glass opacity; HRCT, high-resolution CT; PCA, principal
component analysis; RT-PCR, reverse transcriptase PCR.
90 Caricchio R, et al. Ann Rheum Dis 2021;80:88–95. doi:10.1136/annrheumdis-2020-218323
Criteria
Ann Rheum Dis: first published as 10.1136/annrheumdis-2020-218323 on 25 September 2020. Downloaded from http://ard.bmj.com/ on August 29, 2021 by guest. Protected by copyright.
Table 3 Laboratory parameters in the cohort of patients with COVID-19
Normal range All Clinical consensus OR P value
No Storm Storm
Albumin 3.2–4.6 g/dL 2.9±0.6 2.9±0.6 2.7±0.5 0.637 0.001
ALT 16–61 U/L 45±46 43±37 58±86 1.254 0.04
Anion gap 6–16 mmol/L 7.6±3.0 7.8±3.0 7±3.1 0.734 0.04
AST 15–37 U/L 54±92 50±80 82±145 1.249 0.028
BUN:creatinine ratio 10–20 ratio 18.9±8.3 18.5±8 21±10 1.295 0.03
Chloride 101–111 mmol/L 104±5 104±5 106±5 1.316 0.032
CRP 0–0.4 mg/dL 7.2±6.4 6.9±6.4 9.1±6.1 1.341 0.016
D-dimers 0–500 ng/mL 3,227±11,306 2,396±7,851 8,817±23,356 1.41 0.002
LDH 84–246 U/L 323±169 305±153 447±212 1.892 <0.001
Lymphocytes Abs 1–4.8 K/mm3 1.23±2.16 1.28±2.30 0.86±0.41 0.058 <0.001
Lymphocytes (%) 20%–40% 18±10 19±11 12±7 0.389 <0.001
Neutrophil Abs 1.8–7.8 K/mm3 6±3.6 5.8±3.6 7.23±3.5 1.4 0.004
Potassium 3.5–5.2 mmol/L 4.09±0.5 4.07±0.51 4.23±0.59 1.392 0.019
Troponin I 0.045–0.1 ng/mL 0.23±2.29 0.1±0.38 1.07±6.1 2.727 0.045
Average and SD of the laboratory parameters collected up to the 24 hours within reaching the clinical consensus of CS or in the first 7 days of hospitalisation in patients with
COVID-19 who never reached the clinical consensus of CS. ORs and p values were calculated by logistic regression. Normal range of values is shown for our laboratory as
reference.
Abs, absolute numbers; ALT, alanine aminotransferase; AST, aspartate aminotransferase; BUN, blood urea nitrogen; CRP, C reactive protein; CS, cytokine storm; LDH, lactate
dehydrogenase.
days of hospitalisation or up to the 24 hours within reaching be indicators of the same condition or mechanism. We consid-
the clinical consensus of CS. Using the logistic regression, we ered parameters belonging to the different clusters instead as
found that 12 laboratory parameters predict development of necessary indicators (AND rule) because they represent condi-
CS and by PCA, we determined that these 12 variables could tions or mechanisms that should be met. Our analyses high-
be included in three coherent clusters (table 4). Based on factor lighted three clusters of laboratory results, and the alteration
analysis, we considered the parameters belonging to the same of one parameter for each cluster predicts the development of
cluster as alternative indicators (OR rule), with the rationale COVID-C S (table 4).
that parameters of the same cluster highly correlate and may The first cluster included decreased levels of albumin and
percentage of lymphocytes, along with increased absolute
numbers of neutrophils in patients in storm compared with
Table 4 Predictive criteria for COVID-19 cytokine storm patients who did not develop a storm (tables 3 and 4). The abso-
Entry criteria (must be all met) Cut-off values lute number of lymphocytes formed a separated component and
correlated with the first cluster, and we excluded it from the
+Signs/symptoms of COVID-19
criteria because of its close correlation and redundancy with
±RT-PCR positive for COVID-19
the percentage of lymphocytes. The second cluster included the
+GGO by HRCT (or chest X-ray)
increased levels of alanine aminotransaminase (ALT), aspartate
Ferritin >250 ng/mL
aminotransferase (AST), D-dimers, LDH and troponin I. The
C reactive protein >4.6 mg/dL third cluster included the decreased anion gap and increased
AND (one variable from each cluster) levels of chloride, potassium and blood urea nitrogen (BUN):cre-
Cluster I
atinine ratio (tables 3 and 4). These results highlight an important
Albumin <2.8 g/dL component of tissue damage occurring during the COVID-CS.
Lymphocytes (%)
<10.2 In order to develop cut-offs that can be used in clinical prac-
Neutrophil Abs >11.4 K/mm3 tice, we used daily laboratory parameters and estimated the
Cluster II
cut-off for each individual laboratory parameter using a genetic
ALT >60 U/L algorithm.22 The predictive requirement for the first cluster
AST >87 U/L consisted of an albumin<2.87 mg/mL OR lymphocytes<10.2%,
D-dimers >4,930 ng/mL OR neutrophil absolute number>11.4×103/mL. For the
LDH >416 U/L second cluster, ALT>60 IU/L, OR AST>87 IU/L, OR D- di-
Troponin I >1.09 ng/mL mers>4930 ng/mL, OR LDH>416 U/L OR troponin I>1.09 ng/
Cluster III
mL were required. For the third cluster, anion gap<6.8 mmol/L,
Anion gap <6.8 mmol/L
OR chloride>106 mmol/L, OR potassium>4.9 mmol/L OR
BUN:creatinine ratio>29 were required (table 4).
Chloride >106 mmol/L
Interestingly, ferritin and CRP had the widest ranges and had
Potassium >4.9 mmol/L
discriminatory power only if transformed by logarithmic scale.
BUN:creatinine ratio >29 ratio
They were therefore added as such in the analyses. Although
Criteria are met when patients fulfil all the entry criteria and at least one criterion
the performance of their predictive algorithm did not add any
per each cluster. Cut-off values were calculated using a genetic algorithm.
Abs, absolute numbers; ALT, alanine aminotransferase; AST, aspartate power, for clinical reassurance of an ongoing systemic inflam-
aminotransferase; BUN, blood urea nitrogen; GGO, ground-glass opacity; HTCT, mation, we propose to add them to the predictive criteria of
high-resolution CT; LDH, lactate dehydrogenase; RT-PCR, reverse transcriptase PCR. COVID-CS (table 4).
Caricchio R, et al. Ann Rheum Dis 2021;80:88–95. doi:10.1136/annrheumdis-2020-218323 91
Criteria
Ann Rheum Dis: first published as 10.1136/annrheumdis-2020-218323 on 25 September 2020. Downloaded from http://ard.bmj.com/ on August 29, 2021 by guest. Protected by copyright.
Table 5 Validation of the novel criteria of COVID-CS
COVID-CS new criteria Consensus storm
LoS Mortality
N % (days)* (%)* No Yes
All group
No 340 66 5.7±6.7 6.6 330 10
Yes 173 34 15.1±13 28.8 119 54
SP=0.73 SE=0.84
ACC=0.75
No trials group
No 308 71.5 5.3±6.7 6.4 300 8
Yes 123 28.5 15.3±13.7 28.1 78 45
SP=0.79 SE=0.85
ACC=0.80
The new COVID-CS criteria were applied to the cohort (left) and to the same
patients divided according to the clinical consensus of cytokine storm (right). All
group includes all the COVID-19 cohort (513, 64 patients who reached and 449 who
did not reach the clinical consensus of COVID-19 cytokine storm). No trials group
includes 431 patients of the cohort because it excludes 82 patients with COVID-19
who were recruited in clinical trials testing biologic therapies. The new criteria
identified patients with significantly greater LoS and mortality.
*P<0.0001.
ACC, accuracy; COVID-CS, COVID-19 cytokine storm; LoS, length of stay; SE,
sensitivity; SP, specificity.
Figure 2 Rapid progression for patients with COVID-19 towards
meeting the COVID-CS criteria. The cohort of 431 patients with
COVID-19 (no trials) was plotted for the accumulation of the laboratory
Preliminary validation of the novel criteria of COVID-CS parameters fulfilling the COVID-CS criteria during hospitalisation.
We validated the ability of the proposed criteria to identify The blue line represents the percentage of patients who received the
COVID-C S. The upper rows of table 5 show the initial vali- clinical diagnosis of CS and met the COVID-CS criteria. The orange line
dation where all the patients were included. The criteria clas- represents the percentage of patients who did not receive the clinical
sified 34% of the patients as in COVID-C S (173/513). Next, diagnosis of CS and met the COVID-CS criteria. COVID-CS, COVID-19
we applied the new criteria to the 64 patients originally consid- cytokine storm; CS, cytokine storm.
ered in CS by clinical consensus, and 84% of them (54/64)
were correctly classified as in COVID- C S. The new criteria
had a specificity of 0.73 (CI 0.70 to 0.77) and a sensitivity of Markers of inflammation and tissue damage in COVID-CS
0.84 (CI 0.78 to 0.92). We then performed the validation after We analysed the laboratory results in our cohort of patients
excluding the 82 patients who were enrolled in clinical trials now divided as fitting or not the COVID-CS criteria (table 6
(table 5, bottom rows). In this subpopulation of 431 patients, and online supplemental figure S1). The COVID-CS group had
the criteria showed an even higher specificity of 0.79 (CI 0.76 significantly higher levels of ferritin, CRP and triglycerides, and
to 0.83) and a sensitivity of 0.85 (CI 0.78 to 0.93), suggesting decreased levels of albumin, all signs of systemic inflammation.
that these criteria have a strong predictive power in our popula- Ferritin showed an OR of 14, indicating an important role in
tion of patients with COVID-19. COVID-CS. Strong inflammation was confirmed by the level of
When we analysed the disaggregated laboratory parameters IL-6, which was elevated in most patients with COVID-19 but
to determine the length of time patients required to meet the significantly higher in COVID-CS (35 vs 96 pg/mL). The white
criteria of COVID-CS, we found that among the patients with blood cells, and especially neutrophils and monocytes, were
the clinical consensus of CS, 43% met the criteria on hospital significantly increased in the COVID-CS group, suggesting an
admission, and the rest reached the asymptote by 10 days of active role of the innate immunity in the storm. The lymphocytes
hospitalisation (figure 2, blue line). Among the patients who did instead were decreased, with averages half of the normal lower
not reach the clinical consensus of CS, 20% met the COVID-CS limit, suggesting a functional depletion of the adaptive immunity
criteria with a similar timeline (figure 2 orange line). These (table 6 and online supplemental figure S1).
results suggest an early and rapid progression in those patients We also found that five markers of tissue damage were signifi-
bound to develop COVID-CS, as well as the low likelihood of cantly higher in patients with COVID-CS than in the rest of the
developing the condition 10 or more days into the admission. patients with COVID-19. The liver enzymes ALT and AST had
levels twice as high, indicative of liver damage, while D-dimers
had levels more than six times higher, suggesting endothelial
COVID-CS criteria identify severely ill patients damage. The increase in LDH is a sign of cell death, while the
To determine whether our criteria could predict clinical severity, moderately elevated levels of troponin I suggest damage to the
we analysed the hospital length of stay (LoS) and mortality. We cardiovascular system (table 6 and online supplemental figure
found that the group of patients who met COVID-CS criteria S1).
had a significantly higher LoS (15.1±13 vs 5.7±6.7) and impor- Laboratory parameters pertaining to the electrolyte metab-
tantly higher mortality (28.8% vs 6.6%) (table 5). For both LoS olism, namely, chloride, potassium and sodium, the first two
and mortality, the p value was <0.0001. Excluding the patients predictive of COVID-CS, were still in the normal range, while
in trials yielded similar results. creatinine, BUN and their ratio were all increased compared
92 Caricchio R, et al. Ann Rheum Dis 2021;80:88–95. doi:10.1136/annrheumdis-2020-218323
Criteria
Ann Rheum Dis: first published as 10.1136/annrheumdis-2020-218323 on 25 September 2020. Downloaded from http://ard.bmj.com/ on August 29, 2021 by guest. Protected by copyright.
Table 6 Laboratory parameters in the cohort of patients with COVID-19 at Temple University associated with the new criteria of COVID-CS
Parameters Normal range All No COVID-CS COVID-CS OR P value N
Albumin 3.2–4.6 g/dL 2.9±0.6 3.1±0.6 2.6±0.4 0.292 <0.001 495
Ferritin 8–388 ng/mL 947±2,754 502±738 1,701±4,319 14.725 <0.001 444
C reactive protein 0–0.4 mg/dL 7.0±6.3 5.8±5.8 9.3±6.5 1.781 <0.001 457
Triglycerides <150 mg/dL 178±205 138±72 234±300 3.120 <0.001 330
Interleukin-6 <5 pg/mL 69±126 35±35 96±162 3.799 <0.001 75
Lymphocytes (%) 20%–40% 18±10 21±10 11±7 0.217 <0.001 509
Monocytes Abs 0–0.8 K/mm3 0.59±0.28 0.56±0.28 0.63±0.27 1.260 0.01 509
Neutrophil Abs 1.8–7.8 K/mm3 6.00±3.6 4.98±3.1 8.01±3.7 2.602 <0.001 509
WBC Abs 4–11 K/mm3 7.88±4.5 6.83±3.1 9.95±5.8 2.818 <0.001 508
ALT 16–61 U/L 45±46 35±25 65±66 2.830 <0.001 495
AST 15–37 U/L 53±88 40±76 79±101 2.883 <0.001 495
D-d imers 0–500 ng/mL 3,119±10,443 1,017±2,138 6,933±16,648 10.409 <0.001 456
LDH 84–246 U/L 323±166 249±87 456±191 10.545 <0.001 462
Troponin I 0.045–0.1 ng/mL 0.23±2.29 0.05±0.12 0.55±3.8 438.236 <0.001 416
Chloride 101–111 mmol/L 104±5.4 103±4.8 105±6.1 1.627 <0.001 509
Potassium 3.5–5.2 mmol/L 4.09±0.5 3.99±0.5 4.29±0.5 1.815 <0.001 509
Sodium 136–145 mmol/L 137±4.01 136±3.14 138±5.12 1.607 <0.001 509
Creatinine 0.6–1.10 mg/dL 1.89±2.8 1.61±2.5 2.43±3.3 1.317 0.003 509
BUN 8–20 mg/dL 27±25 21±19 39±32 2.266 <0.001 509
BUN:creatinine ratio 10–20 ratio 18.9±8.4 17.1±6.4 22.6±10.5 2.076 <0.001 509
Average and SD of the laboratory parameters collected in the first 7 days of hospitalisation in patients with COVID-19 divided according to the new COVID-CS criteria. ORs and p
values were calculated by univariate logistic regression. Normal range of values is shown for our laboratory as reference.
Abs, absolute numbers; ALT, alanine aminotransferase; AST, aspartate aminotransferase; BUN, blood urea nitrogen; COVID-CS, COVID-19 cytokine storm; LDH, lactate
dehydrogenase; WBC, white blood cell.
with the upper normal limits in patients fitting the criteria of Despite the limitations of missing three HLH criteria and
COVID-CS. These results suggest a prerenal imbalance and renal one for the HScore, namely, the hemophagocytosis, the NK-cell
damage (table 6 and online supplemental figure S1). Together, activity and the sIL-2R, we propose that the lack of cytopenias,
these results highlight systemic tissue damage affecting many the normal levels of fibrinogen and the only mildly elevated
organs in the COVID-CS. levels of triglycerides indicate that the COVID- CS is very
different from HLH and the HScore is not useful.12
It was recently reported that LDH, CRP and low lymphocytes
Second validation of the COVID-CS criteria
are associated with higher mortality in patients with COVID-
Finally, we further validated the novel criteria by applying them
19.18 26 Our results are in agreement with these studies. Indeed,
to a second cohort of 258 patients, 128 women and 130 men,
our COVID-CS criteria identify a group of patients with longer
with a mean age of 59 years. Out of the 258 patients, 39 (15%)
LoS and increased mortality. Therefore, our criteria predict not
were considered in CS by the same clinical consensus used in
only the development of the storm but also clinical severity. Both
the first cohort. In the new cohort, the novel criteria correctly
CD4+ and CD8+ T cells were initially reported to be decreased
classified 69% of the patients, with a specificity of 0.73 (CI 0.69
in severe cases of COVID-1927 28 and more recently were shown
to 0.78) and a sensitivity of 0.69 (CI 0.58 to 0.81), indicating
to recover during disease resolution.29 T cells are pivotal in the
that the criteria can be successfully applied to new cohorts.
elimination of viral infected cells.30 Moreover, it has been shown
Similar to the first cohort, patients who met the criteria (33%)
in other CSs that the excess of cytokines can be due to a defi-
had significantly higher LoS (15.5±10.1 vs 4.7±3.7, p<0.001)
cient elimination of cytokine-producing innate immune cells,
and mortality (33.7% vs 4.2%, p<0.0001) (online supplemental
such as inflammatory monocytes and macrophages, by CD8+ T
table S5).
cells.31 Therefore, low lymphocytes as criterion for COVID-CS
highlight the role of deficient T-cell functions in COVID-CS
DISCUSSION pathogenesis, allowing innate immunity overactivation and
Our analyses highlight the unique features of COVID-CS. We uncontrolled viral infection.32
found that laboratory parameters indicative of a strong proin- The increased levels of cell death markers such as liver enzymes,
flammatory status, systemic cell death and multiorgan tissue LDH, D-dimers and troponin I indicate that COVID-CS is char-
damage, and prerenal electrolyte imbalance are predictive of acterised by significant systemic tissue damage that in different
this hyperimmune condition. We found clear differences with patients may target the liver, the cardiovascular system and the
other CSs, such as MAS, from which COVID- CS is distin- kidney, as suggested by recent autopsy results.13–16 High levels
guished for the uncommon thrombocytopenia and the increased of D-dimers have been reported in several cohorts of patients
number of neutrophils, suggestive of an active innate immune with COVID-19 and correlate with increased mortality.33 The
system. Other distinct differences were the increased levels of elevated levels of LDH, D- dimers and troponin, especially
fibrinogen and the relatively normal levels of triglycerides in early on, could also indicate pulmonary immunothrombosis
the COVID-CS, which, together with the low levels of albumin, and secondary pulmonary arterial hypertension, both impli-
suggest a different type of inflammation. cated in the devastating lung damage that COVID-19 inflicts.34
Caricchio R, et al. Ann Rheum Dis 2021;80:88–95. doi:10.1136/annrheumdis-2020-218323 93
Criteria
Ann Rheum Dis: first published as 10.1136/annrheumdis-2020-218323 on 25 September 2020. Downloaded from http://ard.bmj.com/ on August 29, 2021 by guest. Protected by copyright.
Therefore, anticoagulant therapy has been recommended in Acknowledgements We thank Dr Randy Cron for critically reading the manuscript
those with high levels,35 and our therapeutic approach has and the numerous suggestions.
changed as well. Compared with the initial cohort, the valida- Collaborators Temple University COVID-19 Research Group: Aaron Mishkin;
tion cohort received higher and earlier doses of steroids, and a Abbas Abbas; Abhijit S Pathak; Abhinav Rastogi; Adam Diamond; Aditi Satti; Adria
Simon; Ahmed Soliman; Alan Braveman; Albert J Mamary; Aloknath Pandya; Amy
larger percentage received anticoagulants. These changes might Goldberg; Amy Kambo; Andrew Gangemi; Anjali Vaidya; Ann Davison; Anuj Basil;
explain the lower sensitivity of the criteria in the validation Arthur Lau; Arundathi Jayatilleke; Bakhos, Charles T; Bill Cornwell; Brent Lawrence;
cohort; nevertheless, they remain very valuable as there is not Brianna Sanguily; Brittany Corso; Carla Grabianowski; Carly Sedlock; Catherine
yet a standard to aggressively or conservatively treat patients Myers; Charles Bakhos; Chenna Kesava; Reddy Mandapati; Cherie Erkmen; Chethan
Gangireddy; Chih-ru Lin; Christopher T Burks; Claire Raab; Crabbe, Deborah;
with COVID-CS around the world.
Crystal Chen; Daniel Edmundowicz; Daniel Sacher; Daniel Salerno; Daniele Simon;
Acute respiratory distress syndrome (ARDS) is undeniably one David Ambrose; David Ciccolella; Debra Gillman; Dolores Fehrle; Dominic Morano;
of the most lethal manifestations of COVID-19 infection.36 The Donnalynn Bassler; Edmund Cronin; Eduardo Dominguez; Ekam Randhawa;
abnormal laboratory work in our criteria could be explained by Ekamjeet Randhawa; Eman Hamad; Eneida Male; Erin Narewski; Francis Cordova;
ARDS in which both hypoxia and hyperaemia could drive eleva- Frederic Jaffe; Frederich Kueppers; Fusun Dikengil; Galli, Jonathan; Gangemi, Andrew;
Garfield, Jamie; Gayle Jones; Gennaro Calendo; Gerard Criner; Gilbert D’Alonzo;
tion of LDH, liver enzymes and renal dysfunction with albumin Ginny Marmolejos; Gordon, Matthew; Gregory Millio; Gupta, Rohit; Gustavo
levels as predictor of ARDS.37 38 Nevertheless, a significant Fernandez; Hannah Simborio; Harwood Scott; Heidi Shore-Brown; Hernan Alvarado;
number of immunoprofiling results point to a systemic inflam- Ho-Man Yeung; Ibraheem Yousef; Ifeoma Oriaku; Iris Jung-won Lee; Isaac Whitman;
matory response with the lung at the epicentre.32 38 39 James Brown; Jamie L. Garfield; Janpreet Mokha; Jason Gallagher; Jeffrey Stewart;
Jenna Murray; Jessica Tang; Jeyssa Gonzalez; Jichuan Wu; Jiji Thomas; Jim Murrett;
There are limitations to our work. First, in the absence of an Joanna Beros; John M. Travaline; Jolly Varghese; Jordan Senchak; Joseph Lambert;
established definition of COVID-CS in the literature, the clin- Joseph Ramzy; Joshua Cooper;Jun Song; Junad Chowdhury; Justin Levinson; Kaitlin
ical ‘gold standard’ was defined by the clinical judgement of CS Kennedy; Karim B Ahmed; Karim Loukmane; Karthik Shenoy; Kathleen Brennan; Keith
itself. Second, the vast majority of our patients received steroids Johnson; Kevin Carney; Kevin Lu; Kraftin Schreyer; Kristin Criner; Kumaran, Maruti;
Lauren Miller; Laurie Jameson; Laurie Johnson; Laurie Kilpatrick; Lawrence Brent;
as part of the early standard of care at Temple University.40
Lii-Yoong Criner; Lily Zhang; Lindsay K Mcgann; Llera A Samuels; Marc Diamond;
Third, our investigation was conducted in a single centre and Margaret Kerper; Maria Vega Sanchez; Mariola Marcinkienwicz; Maritza Pedlar; Mark
with a specific racial/ethnicity composition. These limitations Aksoy; Mark Weir; Marla R. Wolfson; Marla Wolfson; Marron, Robert; Martin Keane;
might make our cohort somewhat different from other centres. Massa Zantah; Mathew Zheng; Matthew Delfiner; Matthew Gordon; Maulin Patel;
Future validations with other cohorts from multiple centres and Megan Healy; Melinda Darnell; Melissa Navaro; Meredith A. Brisco-Bacik; Michael
Bromberg; Michael Gannon; Michael Jacobs; Mira Mandal; Nanzhou Gou; Narewski,
countries will resolve these limits. Erin; Nathaniel Marchetti; Nathaniel Xander; Navjot Kaur; Neil Nadpara; Nicole
The high levels of cell death markers shed light in COVID-CS Desai; Nicole Mills; Norihisa Shigemura; Ohoud Rehbini; Oisin O’Corragain; Omar
pathogenesis. Both necroptosis and pyroptosis can occur during Sheriff; Oneida Arosarena; Osheen Abramian; Paige Stanley; Parag Desai; Parth Rali;
viral infections and are mediated by proinflammatory cyto- Patrick Mulhal; lPravin Patil; Priju Varghese; Puja Dubal; Puja Patel; Rachael Blair;
Rajagopalan Rengan; Rami Alashram; Randol Hooper; Rebecca A Armbruster; Regina
kines such as interferon-gamma and IL-1-beta and by inflam- Sheriden; Robert Marron; Rogers Thomas; Rohit Gupta; Rohit Soans; Roman Petrov;
masome.32 41 A recent longitudinal immune analysis revealed a Roman Prosniak; Romulo Fajardo; Ruchi Bhutani; Ryan Townsend; Sabrina Islam;
subgroup of patients who eventually died of COVID-19 with a Samantha Pettigrew; Samantha Wallace; Sameep Sehgal; Samuel Krachman; Santosh
cytokine profile indicative of CS, inflammasome involvement and Dhungana; Sarah Hoang; Sean Duffy; Seema Ran; iShapiro William; Sheila Weaver;
Shelu Benny; Sheril George; Shuang Sun; Shubhra Srivastava-Malhotra; Stephanie
tissue damage.32 In this group, the levels of IL-6 were extremely
Brictson; Stephanie Spivack; Stephanie Tittaferrante; Stephanie Yerkes; Stephen
elevated.32 We also found elevated IL-6 levels in all patients and Priest; Steve Codella; Steven G Kelsen; Steven Houser; Steven Verga; Sudhir Bolla;
more in COVID-CS, although not as high as reported by others Sudhir Kotnala; Sunil Karhadkar; Sylvia Johnson; Tahseen Shariff; Tammy Jacobs;
32 39
possibly because we tested early during hospitalisation. Thomas Hooper; Tom Rogers; Tony S. Reed; Tse-Shuen Ku; Uma Sajjan; Victor Kim;
Together with the high levels of CRP, which is induced by IL-6,42 Whitney Cabey; Wissam Chatila; Wuyan Li; Zach Dorey-Stein; Zachariah Dorey-Stein;
Zachary D Repanshek.
as COVID-CS criterion, these findings demonstrate the validity
of our criteria in capturing the impending storm. Several clinical Contributors RC, MC and SG designed, analysed and interpreted the data and
drafted the manuscript. RC, CD, DF and XZ acquired and analysed the data. MB,
trials testing cytokine inhibitors are presently ongoing in patients DF and GJC revised the manuscript critically for important intellectual content.
with COVID-19 and may soon provide evidence for a role of RC, MC, SG, RC, CD, XZ. MB, DF and GJC approved the final version of the
cytokine-mediated cell death in patients with COVID-CS. manuscript.
In summary, we provide new criteria to diagnose the Funding This work was supported by the NIH grant R56 AR072115-01 (to RC),
COVID-CS at an early stage, which predict longer hospitalisa- Lupus Research Alliance (RC).
tion and increased mortality, therefore requiring specific treat- Competing interests None declared.
ments. While the criteria need further validation, they represent Patient and public involvement Patients and/or the public were not involved in
a first step toward early diagnosis and intervention in this lethal the design, conduct, reporting or dissemination plans of this research.
pandemic. Patient consent for publication Not required.
Ethics approval This study was approved by Temple University local institutional
Author affiliations review board (IRB) and informed consent was waived based on retrospective
1
Medicine/Rheumatology, Lewis Katz School of Medicine at Temple University, evaluation of deidentified data (IRB protocol number 27120).
Philadelphia, Pennsylvania, USA
2
Department of Psychology, University of Milano-Bicocca, Milan, Italy Provenance and peer review Not commissioned; externally peer reviewed.
3
Radiology, Lewis Katz School of Medicine at Temple University, Philadelphia, This article is made freely available for use in accordance with BMJ’s website
Pennsylvania, USA terms and conditions for the duration of the covid-19 pandemic or until otherwise
4
Microbiology and Immunology, Lewis Katz School of Medicine at Temple University, determined by BMJ. You may use, download and print the article for any lawful,
Philadelphia, Pennsylvania, USA non-commercial purpose (including text and data mining) provided that all copyright
5
Pediatrics, Lewis Katz School of Medicine at Temple University, Philadelphia, notices and trade marks are retained.
Pennsylvania, USA
6
Medicine/Hematology, Lewis Katz School of Medicine at Temple University, ORCID iDs
Philadelphia, Pennsylvania, USA Roberto Caricchio http://orcid.org/0000-0002-1379-1118
7
Thoracic Medicine and Surgery, Lewis Katz School of Medicine at Temple University, Marcello Gallucci http://orcid.org/0000-0003-3546-0093
Philadelphia, Pennsylvania, USA Stefania Gallucci http://orcid.org/0000-0003-4737-8003
Ann Rheum Dis: first published as 10.1136/annrheumdis-2020-218323 on 25 September 2020. Downloaded from http://ard.bmj.com/ on August 29, 2021 by guest. Protected by copyright.
REFERENCES 22 Luque G, Alab E. Parallel genetic algorithms theory and real world application. Spain:
1 Guan W-J, Ni Z-Y, Hu Y, et al. Clinical characteristics of coronavirus disease 2019 in Springer-Verlag Berlin Heidelberg, 2011.
China. N Engl J Med 2020;382:1708–20. 23 Scrucca L. Ga: a package for genetic algorithms in R. Journal of Statistical Software
2 Henderson LA, Canna SW, Schulert GS, et al. On the alert for cytokine storm: 2013;53:37.
immunopathology in COVID-19. Arthritis Rheumatol 2020;72:1059–63. 24 Fardet L, Galicier L, Lambotte O, et al. Development and validation of the HScore, a
3 Petrosillo N, Viceconte G, Ergonul O, et al. COVID-19, SARS and MERS: are they score for the diagnosis of reactive hemophagocytic syndrome. Arthritis Rheumatol
closely related? Clin Microbiol Infect 2020;26:729–34. 2014;66:2613–20.
4 La Rosée P, Horne A, Hines M, et al. Recommendations for the management of 25 Qin C, Zhou L, Hu Z, et al. Dysregulation of immune response in patients with
coronavirus 2019 (COVID-19) in Wuhan, China. Clin Infect Dis 2020;71:762–8.
hemophagocytic lymphohistiocytosis in adults. Blood 2019;133:2465–77.
26 Liu X, Shi S, Xiao J, et al. Prediction of the severity of corona virus disease 2019 and
5 Ravelli A, Minoia F, Davì S, et al. 2016 classification criteria for macrophage
its adverse clinical outcomes. Jpn J Infect Dis 2020. doi:10.7883/yoken.JJID.2020.194.
activation syndrome complicating systemic juvenile idiopathic arthritis: a European
[Epub ahead of print: 29 May 2020].
League against Rheumatism/American College of Rheumatology/Paediatric
27 Sun Y, Dong Y, Wang L, et al. Characteristics and prognostic factors of disease severity
rheumatology international trials organisation collaborative initiative. Ann Rheum Dis
in patients with COVID-19: the Beijing experience. J Autoimmun 2020;112:102473.
2016;75:481–9. 28 Wang F, Nie J, Wang H, et al. Characteristics of peripheral lymphocyte subset
6 Zachariah P, Johnson CL, Halabi KC, et al. Epidemiology, clinical features, and disease alteration in COVID-19 pneumonia. J Infect Dis 2020;221:1762–9.
severity in patients with coronavirus disease 2019 (COVID-19) in a children’s hospital 29 Cavalli G, De Luca G, Campochiaro C, et al. Interleukin-1 blockade with high-dose
in New York City, New York. JAMA Pediatr 2020:e202430. anakinra in patients with COVID-19, acute respiratory distress syndrome, and
7 Adhikari SP, Meng S, Wu Y-J, et al. Epidemiology, causes, clinical manifestation and hyperinflammation: a retrospective cohort study. Lancet Rheumatol 2020;2:e325–31.
diagnosis, prevention and control of coronavirus disease (COVID-19) during the early 30 Vardhana SA, Wolchok JD. The many faces of the anti-COVID immune response. J Exp
outbreak period: a scoping review. Infect Dis Poverty 2020;9:29. Med 2020;217. doi:10.1084/jem.20200678
8 Mehta P, McAuley DF, Brown M, et al. COVID-19: consider cytokine storm syndromes 31 Grom AA, Horne A, De Benedetti F. Macrophage activation syndrome in the era of
and immunosuppression. Lancet 2020;395:1033–4. biologic therapy. Nat Rev Rheumatol 2016;12:259–68.
9 Ruscitti P, Berardicurti O, Iagnocco A, et al. Cytokine storm syndrome in severe 32 Lucas C, Wong P, Klein J, et al. Longitudinal analyses reveal immunological misfiring in
COVID-19. Autoimmun Rev 2020;19:102562. severe COVID-19. Nature 2020;584:463–9.
10 Alunno A, Carubbi F, Rodríguez-Carrio J. Storm, typhoon, cyclone or Hurricane in 33 Cummings MJ, Baldwin MR, Abrams D, et al. Epidemiology, clinical course, and
patients with COVID-19? beware of the same storm that has a different origin. RMD outcomes of critically ill adults with COVID-19 in New York City: a prospective cohort
Open 2020;6:e001295. study. medRxiv 2020.
11 McGonagle D, Sharif K, O’Regan A, et al. The role of cytokines including interleukin-6 34 Parry AH, Wani AH, Yaseen M, et al. Demystifying pulmonary vascular
in COVID-19 induced pneumonia and macrophage activation syndrome-like disease. complications in severe coronavirus disease-19 pneumonia (COVID-19) in the
Autoimmun Rev 2020;19:102537. light of clinico-radiologic-pathologic correlation. Thromb Res 2020. doi:10.1016/j.
12 Leverenz DL, Tarrant TK. Is the HScore useful in COVID-19? Lancet 2020;395:e83. thromres.2020.06.043. [Epub ahead of print: 27 Jun 2020].
13 Liu Q, Wang RS, Qu GQ, et al. Gross examination report of a COVID-19 death autopsy. 35 Al-Samkari H, Karp Leaf RS, Dzik WH, et al. COVID-19 and coagulation: bleeding and
Fa Yi Xue Za Zhi 2020;36:21–3. thrombotic manifestations of SARS-CoV-2 infection. Blood 2020;136:489–500.
14 Su H, Yang M, Wan C, et al. Renal histopathological analysis of 26 postmortem 36 Hasan SS, Capstick T, Ahmed R, et al. Mortality in COVID-19 patients with acute
findings of patients with COVID-19 in China. Kidney Int 2020;98:219–27. respiratory distress syndrome and corticosteroids use: a systematic review and meta-
15 Wichmann D, Sperhake J-P, Lütgehetmann M, et al. Autopsy findings and venous analysis. Expert Rev Respir Med 2020. doi:10.1080/17476348.2020.1804365. [Epub
ahead of print: 31 Jul 2020].
thromboembolism in patients with COVID-19: a prospective cohort study. Ann Intern
37 Maccora I, Calabri GB, Favilli S, et al. Long-Term follow-up of coronary artery lesions
Med 2020;173:268–77.
in children in Kawasaki syndrome. Eur J Pediatr 2020. doi:10.1007/s00431-020-
16 Xu Z, Shi L, Wang Y, et al. Pathological findings of COVID-19 associated with acute
03712-6. [Epub ahead of print: 15 Jun 2020].
respiratory distress syndrome. Lancet Respir Med 2020;8:420–2.
38 Hoeboer SH, Oudemans-van Straaten HM, Groeneveld ABJ. Albumin rather than C-
17 Gupta A, Madhavan MV, Sehgal K, et al. Extrapulmonary manifestations of COVID-19. reactive protein may be valuable in predicting and monitoring the severity and course
Nat Med 2020;26:1017–32. of acute respiratory distress syndrome in critically ill patients with or at risk for the
18 Yan L, Zhang H-T, Goncalves J, et al. An interpretable mortality prediction model for syndrome after new onset fever. BMC Pulm Med 2015;15:22.
COVID-19 patients. Nat Mach Intell 2020;2:283–8. 39 Hadjadj J, Yatim N, Barnabei L, et al. Impaired type I interferon activity and
19 Caricchio R, Criner GJ. Rheumatologists and Pulmonologists at temple university inflammatory responses in severe COVID-19 patients. Science 2020;369:718–24.
weather the COVID-19 storm together. J Rheumatol 2020. doi:10.3899/ 40 RECOVERY Collaborative Group, Horby P, Lim WS, et al. Dexamethasone in
jrheum.200740. [Epub ahead of print: 08 Jun 2020]. Hospitalized Patients with Covid-19 - Preliminary Report. N Engl J Med 2020.
20 Wang Z, Yang B, Li Q, et al. Clinical features of 69 cases with coronavirus disease doi:10.1056/NEJMoa2021436. [Epub ahead of print: 17 Jul 2020].
2019 in Wuhan, China. Clin Infect Dis 2020;71:769–77. 41 Jorgensen I, Rayamajhi M, Miao EA. Programmed cell death as a defence against
21 Kucirka LM, Lauer SA, Laeyendecker O, et al. Variation in false-negative rate of reverse infection. Nat Rev Immunol 2017;17:151–64.
transcriptase polymerase chain reaction-based SARS-CoV-2 tests by time since 42 Nehring SM, Goyal A, Bansal P. C reactive protein (CRP). Treasure Island, FL:
exposure. Ann Intern Med 2020;173:262–7. StatPearls, 2020.