Journal Logo

Clinical Investigations

Inflammatory Response and Phenotyping in Severe Acute Respiratory Infection From the Middle East Respiratory Syndrome Coronavirus and Other Etiologies

Arabi, Yaseen M. MD, FCCM1,2,3,4; Jawdat, Dunia PhD2,3,4,5; Hajeer, Ali H. PhD3,4,6; Sadat, Musharaf MBBS1,2,3,4; Jose, Jesna MSc2,7; Vishwakarma, Ramesh K. PhD2,3,4,8; Almashaqbeh, Walid BSc2,3,4,5; Al-Dawood, Abdulaziz MD1,2,3,4

Author Information
doi: 10.1097/CCM.0000000000004724

Abstract

In the last 2 decades, three major outbreaks have occurred with infections related to emerging coronaviruses. The clinical presentations of the severe acute respiratory syndrome (SARS), the Middle East respiratory syndrome (MERS), and coronavirus disease 2019 (COVID-19) have similarities as they all present with pneumonia, acute respiratory distress syndrome (ARDS), and multiple organ failure and are associated with high mortality. Studies have demonstrated marked proinflammatory response in patients with severe coronaviruses infections, often labeled as cytokine storm syndrome, that is associated with worse clinical outcomes (1,2). While proinflammatory response has been recognized in cohorts of patients with SARS, MERS, and COVID-19, it is unclear whether distinct inflammatory subphenotypes exist within these cohorts.

Data from unselected critically ill patients with ARDS, acute renal failure, and sepsis have demonstrated the presence of different inflammatory subphenotypes (3–9). Post hoc data from several clinical trials have demonstrated that approximately 30% of ARDS patients manifest a hyperinflammatory subphenotype and 70% of patients a hypoinflammatory subphenotype (3–6). These two subphenotypes differ in outcome, with a higher mortality in the hyperinflammatory subphenotype (5). Importantly, these subphenotypes may respond differently to different treatments (3,8,9). For example, the two subphenotypes of ARDS may differ in response to positive end-expiratory pressure (PEEP), fluid management, and simvastatin (5).

The objective of this study was to evaluate the inflammatory response in a cohort of patients with severe acute respiratory infection due to the MERS (MERS SARI) and patients with the non-MERS (non-MERS SARI) and to assess for the presence of inflammatory subphenotypes using latent class analysis (LCA).

MATERIALS AND METHODS

In this prospective cohort study, we included consecutive critically ill patients with laboratory-confirmed MERS SARI and community-acquired non-MERS SARI admitted to the ICU at King Abdulaziz Medical City, Riyadh, Saudi Arabia, between June 2014 and January 2017. The study was approved by Institutional Review Board of the Ministry of National Guard Health Affairs, Riyadh, Saudi Arabia, and informed consent was obtained for participation in the study.

Data Collection

Data were collected using the standardized International Severe Acute Respiratory and Emerging Infection Consortium case report forms (10). In addition to demographic data, underlying comorbidities and the durations from symptom onset to presentation to the emergency department, ICU admission, and intubation were documented. We also documented severity of illness using the Sequential Organ Failure Assessment score, as well as laboratory and ventilator parameters and main interventions during the ICU stay. We collected data on other bacterial and viral pathogens. In addition, we documented mortality at ICU and hospital discharge and at day 90. We calculated mechanical ventilation duration, length of stay in the ICU, and the hospital among survivors.

Blood Sampling and Measurement of Cytokines

Blood samples were collected prospectively on ICU days 1, 3, 7, and 14. Blood samples were also collected from 10 healthy individuals to serve as control. Serum was prepared from the clotted blood samples by centrifugation for 10 minutes at 1,000 × g at 4°C and stored at –80°C prior to assay in the serology laboratory. The cytokines and other signaling molecules were measured using Milliplex panel, human cytokine/chemokine magnetic bead panel kit (catalog code HCYTMAG-60K-PX29; Merck Millipore, Darmstadt, Germany) with the Luminex 3D platform (Luminex, Austin, TX), according to the manufacturer instructions. We measured the following cytokines: interleukin (IL)-1α, IL-1β, IL-1ra, IL-2, IL-3, IL-4, IL-5, IL-6, IL-7, IL-8, IL-10, IL-12p40, IL-12p70, IL-13, IL-15, IL-17A, inducible protein-10 (IP-10), monocyte chemoattractant protein (MCP-1), macrophage inflammatory protein (MIP)-1α, MIP-1β, tumor necrosis factor (TNF)-a, TNF-β, eotaxin, epidermal growth factor (EGF), granulocyte-colony stimulating factor (G-CSF), granulocyte-macrophage colony-stimulating factor (GM-CSF), interferon (IFN)-α, IFN-γ, and vascular endothelial growth factor (VEGF). All the cytokine levels were reported in pg/mL.

Statistical Analysis

We reported categorical variables as frequencies with percentages and continuous variables as medians with quartiles 1 and 3 (Q1–Q3). We performed two-group comparisons between the MERS SARI and non-MERS SARI groups using t test, Mann-Whitney U test, chi-square test, or Fisher exact test, as appropriate. We also compared day 1 cytokine levels in MERS SARI and non-MERS SARI patients to cytokine levels in healthy controls. In addition, we compared cytokine levels over time between MERS SARI and non-MERS SARI groups using repeated measures mixed linear models.

Latent Class Analysis

We performed LCA to identify the presence of subphenotypes (3). First, we included in the model day 1 cytokine data in addition to baseline clinical variables that were considered relevant and had missing values of less than 5% (Tables S1 and S2, Supplemental Digital Content 1, http://links.lww.com/CCM/F991). Outcome data were not included in the model. Next, we assessed candidate variables for multicollinearity, and we found a high level of multicollinearity between IL-13 and TNF-β; therefore, we retained TNF-β and excluded IL-13. We imputed clinical variables with less than 5% of missing values using mean imputation technique (11). We assessed the distributional characteristics graphically (i.e., by histograms and Q-Q plots), and we applied square root transformation for IL-15, IL-17A, MIP-1β, EGF, GM-CSF, VEGF, hematocrit, and heart rate and log transformation for the rest of highly skewed variables. All noncategorical variables were rescaled to a common z scale where the mean was set to 0 and the sd to 1 because the variances of different variables varied widely (12). To avoid a local maximum likelihood solution for LCA, we used 2,000 random starting values, of which the best 50 were then optimized. Those solutions were checked to ensure the same maximum likelihood was found.

Next, we fitted a series of LCAs with two classes to five classes. Criteria for model selection were based on the Bayesian information criteria, the Vuong-Lo-Mendell-Rubin likelihood ratio test, and the size of the smallest class. LCA estimation was based on full-information maximum likelihood method (Table S3, Supplemental Digital Content 1, http://links.lww.com/CCM/F991), which allows for the use of all data for all patients. This method incorporates the degree of uncertainty of class membership. Once we established the number of classes, we tested the associations between class and baseline characteristics and clinical outcomes.

We used Mplus (Version 8.3; Muthen and Muthen, Los Angeles, CA) for LCA and SAS (Version 9.4; SAS Institute, Cary, NC) for all other analyses. A two-tailed p value of less than or equal to 0.05 was considered statistically significant. For all cytokine comparisons, we considered p value of less than 0.0017 to be statistically significant accounting for multiple testing using Bonferroni correction.

RESULTS

Clinical Data

During the study period, 116 patients were included, 40 patients with MERS SARI and 76 patients with non-MERS SARI. Patients with MERS SARI were younger (median age [Q1–Q3], 56 yr [39–67 yr] vs 69 yr [48–81 yr]; p = 0.005) than patients with non-MERS SARI, more likely to be males (73% vs 50%; p = 0.02), and had fewer comorbidities (60% vs 84%; p = 0.004), with lower platelet counts (147 × 109/L [65–240 × 109/L] vs 226 × 109/L [158–326 × 109/L; p = 0.001]), and lower lactate levels (1.4 mmol/L [0.9–2.0 mmol/L] vs 2.2 mmol/L [1.3–4.0 mmol/L; p = 0.009]) (Table 1; and Table S4, Supplemental Digital Content 1, http://links.lww.com/CCM/F991). Viral and bacterial pathogens identified in the two groups are shown in Table S5 (Supplemental Digital Content 1, http://links.lww.com/CCM/F991). During the ICU stay, patients with MERS SARI received more renal replacement therapy (53% vs 26%; p = 0.005), neuromuscular blockade (75% vs 25%; p ≤ 0.0001), and prone positioning (23% vs 1%; p = 0.0003) than patients with non-MERS SARI (Table 2). Patients with MERS SARI had higher ICU and 90-day mortality than non-MERS SARI (ICU mortality: 45% vs 22%; p = 0.01 and 90-d mortality: 45% vs 26%; p = 0.04) (Table 2).

TABLE 1. - Comparison of Baseline Characteristics and Physiologic Parameters Between Patients With the Middle East Respiratory Syndrome Severe Acute Respiratory Infection and Non-Middle East Respiratory Syndrome Severe Acute Respiratory Infection and Between Patients With Subphenotype 1 and Subphenotype 2
Variables Middle East Respiratory Syndrome Severe Acute Respiratory Infection, n = 40 Non-Middle East Respiratory Syndrome Severe Acute Respiratory Infection, n = 76 p Subphenotype 1, n = 74 Subphenotype 2, n = 42 p
Demographics
 Age (yr), median  (Q1–Q3) 56 (39–67) 69 (48–81) 0.005 63 (48–78) 65 (43–74) 0.66
 Body mass index (kg/m2), median (Q1–Q3) 28 (25–33) 28 (24–34) 0.76 28 (25–33) 29 (24–31) 0.68
 Male sex, n (%) 29 (73) 38 (50) 0.02 42 (57) 25 (60) 0.77
 Days from onset of symptoms to the emergency department, median (Q1–Q3) 3 (2–5) 3 (2–5) 0.59 3 (2–6) 4 (2–5) 0.96
 Days from onset of symptoms to ICU admission, median (Q1–Q3) 7 (4–12) 5 (3–9) 0.04 6 (3–10) 6 (4–12) 0.65
 Days from onset of symptoms to intubation, median (Q1–Q3) 9 (4–16) 6 (3–10) 0.11 6 (3–10) 7 (4–17) 0.16
Comorbidities, n (%)
 Any comorbidity 24 (60) 64 (84) 0.004 54 (73) 34 (81) 0.33
 Diabetes with chronic complications 4 (10) 40 (53) < 0.0001 27 (37) 17 (41) 0.67
 Chronic pulmonary disease (including asthma) 7 (18) 20 (26) 0.29 20 (27) 7 (17) 0.20
 Chronic liver disease 1 (3) 3 (4) > 0.99 4 (5) - 0.30
 Chronic renal disease 6 (15) 27 (36) 0.02 20 (27) 13 (31) 0.65
 Chronic cardiac disease 13 (33) 41 (54) 0.03 34 (46) 20 (48) 0.86
 Any malignancy 8 (20) 6 (8) 0.07 8 (11) 6 (14) 0.58
 Immunosuppressant use 6 (15) 11 (15) 0.94 12 (16) 5 (12) 0.53
Physiologic parameters on ICU day 1, median (Q1–Q3)
 Mechanical ventilation,  n (%) 27 (68) 63 (83) 0.06 62 (84) 28 (67) 0.03
 Vasopressor use, n (%) 15 (38) 38 (50) 0.19 32 (43) 21 (50) 0.48
 Pao 2/Fio 2 ratio 157 (126–215) 157 (104–217) 0.69 167 (126–215) 135 (90–217) 0.17
 Respiratory rate (breaths/min) 28 (20–35) 27 (22–35) 0.85 26 (20–30) 24 (20–28) 0.39
 Tidal volume (mL) 404 (310–422) 400 (362–450) 0.24 405 (360–450) 390 (330–421) 0.29
 Positive end-expiratory pressure (cm H2O) 12 (10–14) 8 (5–10) < 0.0001 10 (8–12) 8 (5–10) 0.04
 Plateau pressure (cm H2O) 25 (22–31) 23 (20–28) 0.17 24 (20–30) 25 (16–28) 0.45
 Systolic blood pressure (mm Hg) 113 (99–132) 130 (114–154) 0.004 129 (113–148) 119 (96–134) 0.02
 Heart rate (beats/min) 103 (91–122) 103 (83–127) 0.80 99 (86–115) 115 (97–138) 0.01
 Lactate (mmol/L) 1.4 (0.9–2.0) 2.2 (1.3–4.0) 0.009 1.8 (1.1–3.9) 1.9 (1.1–3.4) 0.89
 Creatinine (µmol/L) 70 (54–194) 106 (70–200) 0.08 84 (62–192) 134 (68–258) 0.18
 Hematocrit (%) 36 (28–40) 34 (29–38) 0.66 36 (30–41) 33 (28–37) 0.03
 Platelets (× 109/L) 147 (65–240) 226 (158–326) 0.001 219 (146–320) 178 (89–221) 0.01
 WBC (× 109/L) 6.8 (4.9–10.0) 11.4 (7.9–15.2) 0.0002 10.0 (6.7–14.5) 8.9 (5.2–12.5) 0.17
 Sequential Organ Failure Assessment score 11 (5–12) 7 (5–10) 0.11 7 (5–11) 8 (4–12) 0.83
Continuous variables were compared using Mann-Whitney U test, and categorical variables were compared using χ2 test.
See additional baseline characteristics in Table S4 (Supplemental Digital Content 1, http://links.lww.com/CCM/F991).

TABLE 2. - Comparisons of Interventions During ICU Stay and Outcomes Between Patients With the Middle East Respiratory Syndrome Severe Acute Respiratory Infection Versus Non-Middle East Respiratory Syndrome Severe Acute Respiratory Infection and Between Patients With Subphenotype 1 and Subphenotype 2
Variables Middle East Respiratory Syndrome Severe Acute Respiratory Infection, n = 40 Non-Middle East Respiratory Syndrome Severe Acute Respiratory Infection, n = 76 p Subphenotype 1, n = 74 Subphenotype 2, n = 42 p
Mechanical ventilation, n (%) 36 (90) 66 (87) 0.77 63 (85) 39 (93) 0.25
Vasopressor therapy, n (%) 32 (80) 50 (66) 0.11 48 (65) 34 (81) 0.07
Renal replacement therapy, n (%) 21 (53) 20 (26) 0.005 22 (30) 19 (45) 0.09
Corticosteroids, n (%) 17 (43) 45 (59) 0.09 39 (53) 23 (55) 0.83
Neuromuscular blockade, n (%) 30 (75) 19 (25) < 0.0001 31 (42) 18 (43) 0.91
Prone positioning, n (%) 9 (23) 1 (1) 0.0003 6 (8) 4 (10) >0.99
Clinical outcomes, n (%)
 Hospital mortality 18 (45) 24 (32) 0.15 23 (31) 19 (45) 0.13
 90-d mortality 18 (45) 20 (26) 0.04 20 (27) 18 (43) 0.08
 ICU mortality 18 (45) 17 (22) 0.01 17 (23) 18 (43) 0.03
Clinical outcomes among survivors, median (Q1–Q3)
 ICU length of stay, d 20.5 (15.0–36.0) 12.0 (7.0–24.0) 0.03 16.0 (7.0–23.0) 17.0 (7.0–32.0) 0.46
 Hospital length of stay, d 43.0 (29.0–63.0) 28.0 (15.0–95.0) 0.21 30.0 (17.0–63.0) 45.0 (19.0–139.0) 0.27
 Mechanical ventilation duration 18.0 (14.0–35.0) 8.0 (5.0–15.0) 0.007 12.50 (6.0–26.0) 7.0 (5.0–19.0) 0.29
Denominator of the percentage is the total number of subjects in the group.
Continuous variables were compared using Mann-Whitney U test, and categorical variables were compared using χ2 test.

Cytokine Data

On ICU day 1, both MERS SARI and non-MERS SARI patients had higher levels of IL-3, IL-4, IL-6, IL-8, IL-17A, eotaxin, and EGF compared with healthy control (p < 0.0017) (Fig. 1; and Fig. S1, Supplemental Digital Content 1, http://links.lww.com/CCM/F991). Accounting for multiple testing, there were no differences in cytokines on day 1 and over the ICU course between MERS SARI group and non-MERS SARI patients.

Figure 1.
Figure 1.:
Comparison of serial measurements for interleukin (IL)-2, IL-1RA, IL-6, and IL-8 in patients with Middle East respiratory syndrome severe acute respiratory infection (MERS SARI), non-Middle East respiratory syndrome severe acute respiratory infection (non-MERS SARI), and healthy control (HC) and in patients with subphenotype 1, subphenotype 2, and HC. Day 1 levels in MERS SARI, non-MERS SARI, subphenotype 1, and subphenotype 2 were compared with HCs using Mann-Whitney U test. The differences between groups with time (group × time) were tested by repeated measures mixed linear models. Box plots are displayed with medians and quartiles 1 and 3. The distribution of patients in each group over time for all the cytokines is as follows: MERS SARI (D1 = 36, D3 = 27, D7 = 12, and D14 = 9), non-MERS SARI (D1 = 73, D3 = 59, D7 = 35, and D14 = 9), subphenotype 1 (D1 = 69, D3 = 57, D7 = 30, and D14 = 18), subphenotype 2 (D1 = 40, D3 = 29, D7 = 17, and D14 = 11), and HC (n = 10). For all cytokine comparisons, we considered p < 0.0017 to be statistically significant accounting for multiple testing using Bonferroni correction.

LCA Results

Based on the LCA, patients were classified into two classes as the best fit, subphenotype 1 (n = 74 [64%]) and subphenotype 2 (n = 42 [36%]) (Fig. 2). Subphenotype 2 had higher levels of IL-1β, IL-1ra, IL-2, IL-6, IL-7, IL-8, IL-10, IL-12p70, IL-15, IL-17A, IP-10, MCP-1, MIP-1α, MIP-1β, TNF-α, GM-CSF, G-CSF, IFN-α, and IFN-γ compared with subphenotype 1 (p < 0.0017) (Table 3). Baseline characteristics for the patients between the two subphenotypes were generally not different (Table 1; and Table S1, Supplemental Digital Content 1, http://links.lww.com/CCM/F991). There was no significant difference between the two subphenotypes in demographics including age, gender, body mass index, comorbidities, or interventions. However, patients in the subphenotype 2 had higher ICU mortality in comparison to the subphenotype 1 (18 [43%] vs 17 [23%]; p = 0.03). Among survivors, there was no difference in ICU LOS, hospital LOS, and mechanical ventilation duration between the two phenotypes.

TABLE 3. - Comparison of Day 1 Cytokine Levels Between Patients With the Middle East Respiratory Syndrome Severe Acute Respiratory Infection Versus Non-Middle East Respiratory Syndrome Severe Acute Respiratory Infection and Between Patients With Subphenotype 1 and Subphenotype 2 Using Mann-Whitney U Test
Cytokines (pg/mL), Median (Q1–Q3) Middle East Respiratory Syndrome Severe Acute Respiratory Infection, n = 40 Non-Middle East Respiratory Syndrome Severe Acute Respiratory Infection, n = 76 p Subphenotype 1, n = 74 Subphenotype 2, n = 42 p
IL-1α 157.2 (38.6–445.2) 20.0 (1.3–158.2) < 0.0001 25.8 (5.4–156.1) 150.9 (1.4–591.9) 0.01
IL-1β 1.3 (0.2–3.2) 1.7 (0.5–4.7) 0.33 0.7 (0.2–1.7) 4.8 (3.0–9.1) < 0.0001
IL-1ra 29.3 (8.1–158.8) 12.7 (3.1–83.1) 0.09 10.1 (1.5–29.0) 89.0 (10.2–433.4) < 0.0001
IL-2 3.8 (1.8–4.3) 6.3 (2.4–12.0) 0.02 3.8 (1.1–6.2) 10.1 (4.2–24.3) < 0.0001
IL-3) 0.3 (0.1–4.0) 1.3 (0.5–2.8) 0.02 0.9 (0.2–4.3) 1.0 (0.3–2.3) 0.83
IL-4 25.0 (7.9–73.7) 45.4 (17.7–123.9) 0.06 33.9 (10.1–93.9) 65.0 (27.8–136.0) 0.03
IL-5 0.3 (0.03–3.5) 1.4 (0.5–8.6) 0.002 0.7 (0.1–4.8) 2.2 (0.5–13.9) 0.03
IL-6 81.7(35.8–294.1) 103.3 (20.5–296.6) 0.80 41.0 (12.0–123.8) 360.1 (119.5–1,098.0) < 0.0001
IL-7 11.5 (3.8–29.8) 8.7 (3.0–28.6) 0.73 4.9 (2.0–10.6) 40.6 (17.7–63.9) < 0.0001
IL-8 71.9 (30.4–120.7) 39.1 (18.4–75.8) 0.09 29.7 (11.7–55.8) 105.2 (68.2–262.7) < 0.0001
IL-10 1.4 (0.04–15.0) 7.3 (0.8–75.2) 0.01 1.6 (0.1–10.8) 57.8 (3.8–329.5) < 0.0001
IL-12p40 0.1 (0.1–3.7) 0.04 (0.01–3.9) 0.11 0.1 (0.02–2.4) 0.1 (0.01–7.5) 0.85
IL-12p70 0.5 (0.1–3.2) 1.0 (0.3–2.2) 0.27 0.5 (0.1–1.3) 2.0 (0.5–4.0) 0.0003
IL-13 1.5 (0.01–215.6) 98.3 (6.8–307.3) 0.004 27.3 (0.1–214.2) 105.3 (9.9–429.0) 0.03
IL-15 8.2 (3.7–23.2) 8.6 (3.5–22.2) 0.67 4.8 (1.8–10.5) 28.3 (16.4–46.0) < 0.0001
IL-17A 0.5 (0.1–2.3) 0.7 (0.1–2.5) 0.88 0.3 (0.1–1.0) 2.4 (0.5–4.1) < 0.0001
Inducible protein-10 793.2 (385.4–3,478.0) 309.2 (151.8–959.9) 0.003 309.2 (139.4–847.9) 1,197.5 (385.1–4,951.0) < 0.0001
Monocyte chemoat tractant protein-1 1,754.0 (1,076.5–4,528.0) 1,510.5 (658.0–2,655.5) 0.12 1,119.5 (592.2–1,811.0) 3,637.5 (1,834.0–6,238.0) < 0.0001
MIP-1α 7.8 (2.7–23.5) 4.3 (1.3–37.0) 0.52 3.0 (1.2–11.9) 22.3 (3.5–39.2) 0.001
MIP-1β 21.4 (11.1–30.0) 40.8 (20.8–72.5) 0.005 24.3 (9.4–43.3) 49.2 (23.2–104.3) 0.0002
TNF-α 3.2 (1.2–8.6) 6.8 (1.8–15.3) 0.06 2.5 (1.0–6.5) 13.1 (7.0–23.0) < 0.0001
TNF-β 0.6 (0.01–95.6) 28.5 (0.1–145.1) 0.04 1.7 (0.01–98.1) 51.8 (0.6–159.2) 0.02
Eotaxin 5.9 (1.4–18.9) 14.3 (4.6–33.8) 0.03 7.9 (2.0–23.1) 20.5 (6.1–67.6) 0.003
Epidermal growth factor 370.9 (135.6–983.4) 426.8 (232.6–969.2) 0.56 489.1 (232.6–1,223.0) 333.5 (186.5–517.6) 0.03
Granulocyte-macrophage colony-stimulating factor 20.4 (7.4–41.3) 17.9 (3.8–56.5) 0.88 8.4 (2.5–19.0) 60.2 (32.3–97.1) < 0.0001
Granulocyte-colony stimulating factor 18.8 (0.1–126.9) 17.3 (3.7–169.2) 0.27 7.1 (0.5–28.3) 203.9 (28.3–816.1) < 0.0001
IFN-α 36.8 (11.0–82.9) 18.9 (4.5–47.1) 0.03 9.4 (3.8–28.3) 58.3 (27.6–141.2) < 0.0001
IFN-γ 3.6 (1.5–22.3) 2.7 (0.7–9.1) 0.08 1.6 (0.5–3.6) 10.8 (4.2–22.9) < 0.0001
Vascular endothelial growth factor 105.1 (36.7–292.2) 119.6 (32.4–264.1) 0.96 76.6 (10.5–300.0) 132.8 (68.8–233.9) 0.06
IFN = interferon, IL = interleukin, MIP = macrophage inflammatory protein, TNF = tumor necrosis factor.All cytokine levels were reported in pg/mL.For all cytokine comparisons, we considered p < 0.0017 to be statistically significant accounting for multiple testing using Bonferroni correction.

Figure 2.
Figure 2.:
Differences in standardized values of each continuous variable by subphenotype among patients with the Middle East respiratory syndrome severe acute respiratory infection (MERS SARI) and patients with the non-Middle East respiratory syndrome severe acute respiratory infection (non-MERS SARI). Variables with maximum positive separation (i.e., values are higher in subphenotype 2 compared with subphenotype 1) are on the left, and the variables with maximum negative separation (i.e., values are lower in subphenotype 2 compared with subphenotype 1) are on the right. All means are scaled to zero and sds to one. BMI = body mass index, EGF = epidermal growth factor, G-CSF = granulocyte-colony stimulating factor, GM-CSF = granulocyte-macrophage colony-stimulating factor, HCT = hematocrit, HR = heart rate, IFN-gamma = interferon-γ, IL = interleukin, INF-alpha = interferon-α, IP = inducible protein, MCP-1 = monocyte chemoattractant protein-1, MIP = macrophage inflammatory protein, PLT = platelets, SBP = systolic blood pressure, TNF = tumor necrosis factor, VEGF = vascular endothelial growth factor.

DISCUSSION

In our study, we found that critically ill patients with MERS SARI and non-MERS SARI demonstrated a proinflammatory response characterized by significant elevation of several cytokines compared with healthy control (IL-3, IL-4, IL-6, IL-8, IL-17A, eotaxin, and EGF). Interestingly, there were no differences in cytokine levels between MERS SARI and non-MERS SARI. Using cytokine and clinical data, patients were classified using LCA into two classes; subphenotype 1 (64% of patients) and subphenotype 2 (36% of patients). Subphenotype 2 was characterized by higher IL-1β, IL-1ra, IL-2, IL-6, IL-7, IL-8, IL-10, IL-12p70, IL-15, IL-17A, IP-10, MCP-1, MIP-1α, MIP-1β, TNF-α, GM-CSF, G-CSF, IFN-α, and IFN-γ compared with subphenotype 1 (p < 0.0017). For example, the subphenotype 2 had around seven-fold higher IL-1β and nine-fold higher IL-6 than the subphenotype 1. The baseline characteristics of subphenotype 1 and subphenotype 2 were not different. Yet, patients with the subphenotype 2 had higher ICU mortality than the subphenotype 1, suggesting that the difference in outcomes might be related to differences in inflammatory pattern.

The increase in proinflammatory cytokines among patients with MERS SARI and non-MERS SARI is in line with what has been observed in other studies on MERS, SARS, and COVID-19 (13–19). Infection with SARS-CoV-1 has been shown to be associated with increase in IL-1, IL-6, IL-12, IP-10, MCP-1, and IFN-y (13, 14). Studies have demonstrated that MERS is associated with increased serum levels of IL-1α, IL-1β, IL-1ra, IL-6, IL-8, IL-12, IL-15, IL-17A, IP-10, MCP-1, TNF-α, and IFN-γ (15–18). A study from Wuhan, China found that ICU patients with COVID-19 compared with non-ICU patients had higher plasma levels of IL-2, IL-7, IP-10, MCP-1, MIP-1α, TNF-α, G-CSF, and IFN-γ (19). Another study (n = 452) found that patients with severe COVID-19 had higher inflammatory cytokines, including IL-2R, IL-6, IL-8, IL-10, and TNF-α compared with nonsevere cases (20). A study of 10 COVID-19 ICU patients found elevation of 17 inflammatory analytes compared with age- and sex-matched non-COVID-19 ICU patients (21). A study of 48 patients with COVID-19 found that IL-6 levels were increased by almost 10-fold in critically ill patients compared with other patients and that high IL-6 level closely correlated with the frequency of RNAaemia (22).

We identified two inflammatory subphenotypes among patients with MERS SARI and non-MERS SARI; subphenotype 2 was characterized by increase in proinflammatory cytokines and was associated with higher ICU mortality. We observed significant differences in clinical characteristics of MERS SARI and non-MERS SARI, but no significant differences in cytokine levels between these two groups. Conversely, the two subphenotypes did not differ in clinical characteristics despite substantial differences in cytokine levels. Although we included in the LCA model several relevant clinical variables, the resulting separation of the two subphenotypes by LCA was based largely on cytokine levels and not on clinical variables. These observations together suggest limited correlation between inflammatory subphenotypes and clinical presentation.

The importance of subphenotyping is that different inflammatory subphenotypes may respond differently to therapies. At present, this premise is supported mainly by post hoc observational data, suggesting that statins, corticosteroids, vasopressors, PEEP, and fluid management may have differential effect according to subphenotypes in patients with ARDS, sepsis, and acute kidney injury (3,5,8,9). The presence of inflammatory subphenotyping and its clinical implications in patients with coronavirus infections requires further study. It has been suggested that patients with severe COVID-19 related cytokine storm syndrome may benefit from immunomodulation (1).

Strengths of the study include being a prospective study, serial blood sampling, and the inclusion of MERS SARI and non-MERS SARI patients. Limitations include that our study was not designed to assess the effect of therapeutics on different subphenotypes. Considering our study sample size, we identified two subphenotypes, but we cannot exclude the presence of more subphenotypes with larger cohorts. Our study evaluated inflammatory cytokines but did not evaluate aspects of cellular immunity response, which has been shown to be important in MERS (16).

In conclusion, critically ill patients with MERS SARI and non-MERS SARI demonstrated a proinflammatory response characterized by significant elevation of several cytokines. Around one third of patients were characterized as subphenotype 2, which had higher levels of proinflammatory cytokines and had higher ICU mortality compared with subphenotype 1. Further research is needed to examine whether different therapeutics, such as immunomodulators, have differential effects based on the inflammatory subphenotypes.

ACKNOWLEDGMENTS

We wish to thank the following coordinators who have contributed to the study: Eman Alqasim, Sheryl Abdukahil, Felwa Bin Humaid, Lara Afesh, and Turki Almoammar.

REFERENCES

1. Mehta P, McAuley DF, Brown M, et al. COVID-19: Consider cytokine storm syndromes and immunosuppression. Lancet. 2020; 395:1033–1034
2. Tang Y, Liu J, Zhang D, Xu Z, et al. Cytokine storm in COVID-19: The current evidence and treatment strategies. Front Immunol. 2020; 11:1708
3. Calfee CS, Delucchi K, Parsons PE, et al.; NHLBI ARDS Network. Subphenotypes in acute respiratory distress syndrome: Latent class analysis of data from two randomised controlled trials. Lancet Respir Med. 2014; 2:611–620
4. Famous KR, Delucchi K, Ware LB, et al.; ARDS Network. Acute respiratory distress syndrome subphenotypes respond differently to randomized fluid management strategy. Am J Respir Crit Care Med. 2017; 195:331–338
5. Calfee CS, Delucchi KL, Sinha P, et al.; Irish Critical Care Trials Group. Acute respiratory distress syndrome subphenotypes and differential response to simvastatin: Secondary analysis of a randomised controlled trial. Lancet Respir Med. 2018; 6:691–698
6. Sinha P, Delucchi KL, Thompson BT, et al.; NHLBI ARDS Network. Latent class analysis of ARDS subphenotypes: A secondary analysis of the statins for acutely injured lungs from sepsis (SAILS) study. Intensive Care Med. 2018; 44:1859–1869
7. Bos LDJ, Scicluna BP, Ong DSY, et al. Understanding heterogeneity in biologic phenotypes of acute respiratory distress syndrome by leukocyte expression profiles. Am J Respir Crit Care Med. 2019; 200:42–50
8. Bhatraju PK, Zelnick LR, Herting J, et al. Identification of acute kidney injury subphenotypes with differing molecular signatures and responses to vasopressin therapy. Am J Respir Crit Care Med. 2019; 199:863–872
9. Antcliffe DB, Burnham KL, Al-Beidh F, et al. Transcriptomic signatures in sepsis and a differential response to steroids. From the VANISH randomized trial. Am J Respir Crit Care Med. 2019; 199:980–986
10. The Global Health Network. The International Severe Acute Respiratory and Emerging Infection Consortium (ISARIC). Available at: https://isaric.tghn.org/protocols/severe-acute-respiratory-infection-data-tools/. Accessed March 25, 2020
11. Jakobsen JC, Gluud C, Wetterslev J, et al. When and how should multiple imputation be used for handling missing data in randomised clinical trials - a practical guide with flowcharts. BMC Med Res Methodol. 2017; 17:162
12. Baxter MJ. Standardization and transformation in principal component analysis, with applications to archaeometry. J R Stat Soc Series C (Appl Stat). 1995; 44:513–527
13. Wong CK, Lam CW, Wu AK, et al. Plasma inflammatory cytokines and chemokines in severe acute respiratory syndrome. Clin Exp Immunol. 2004; 136:95–103
14. Cheung CY, Poon LL, Ng IH, et al. Cytokine responses in severe acute respiratory syndrome coronavirus-infected macrophages in vitro: Possible relevance to pathogenesis. J Virol. 2005; 79:7819–7826
15. Kim ES, Choe PG, Park WB, et al. Clinical progression and cytokine profiles of Middle East respiratory syndrome coronavirus infection. J Korean Med Sci. 2016; 31:1717–1725
16. Shin HS, Kim Y, Kim G, et al. Immune responses to Middle East respiratory syndrome coronavirus during the acute and convalescent phases of human infection. Clin Infect Dis. 2019; 68:984–992
17. Mahallawi WH, Khabour OF, Zhang Q, et al. MERS-CoV infection in humans is associated with a pro-inflammatory Th1 and Th17 cytokine profile. Cytokine. 2018; 104:8–13
18. Lau SKP, Lau CCY, Chan KH, et al. Delayed induction of proinflammatory cytokines and suppression of innate antiviral response by the novel Middle East respiratory syndrome coronavirus: Implications for pathogenesis and treatment. J Gen Virol. 2013; 94:2679–2690
19. Huang C, Wang Y, Li X, et al. Clinical features of patients infected with 2019 novel coronavirus in Wuhan, China. Lancet. 2020; 395:497–506
20. Qin C, Zhou L, Hu Z, et al. Dysregulation of immune response in patients with COVID-19 in Wuhan, China. Clin Infect Dis. 2020; 71:762–768
    21. Fraser DD, Cepinskas G, Slessarev M, et al. Inflammation profiling of critically ill coronavirus disease 2019 patients. Crit Care Explor. 2020; 2:e0144
      22. Chen X, Zhao B, Qu Y, et al. Detectable serum SARS-CoV-2 viral load (RNAaemia) is closely associated with drastically elevated interleukin 6 (IL-6) level in critically ill COVID-19 patients. Clin Infect Dis. 2020ciaa449
      Keywords:

      coronavirus; coronavirus disease 2019; cytokine; Middle East respiratory syndrome; severe acute respiratory syndrome coronavirus 2

      Supplemental Digital Content

      Copyright © 2020 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.