Immunoglobulin deficiency as an indicator of disease severity in patients with COVID-19

Abstract

Despite the pandemic status of COVID-19, there is limited information about host risk factors and treatment beyond supportive care. Immunoglobulin G (IgG) could be a potential treatment target. Our aim was to determine the incidence of IgG deficiency and associated risk factors in a cohort of 62 critically ill patients with COVID-19 admitted to two German ICUs (72.6% male, median age: 61 yr). Thirteen (21.0%) of the patients displayed IgG deficiency (IgG < 7 g/L) at baseline (predominant for the IgG1, IgG2, and IgG4 subclasses). Patients who were IgG-deficient had worse measures of clinical disease severity than those with normal IgG levels (shorter duration from disease onset to ICU admission, lower ratio of $\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$ to ${\text{FI}}_{{\text{O}}_{\text{2}}}$, higher Sequential Organ Failure Assessment score, and higher levels of ferritin, neutrophil-to-lymphocyte ratio, and serum creatinine). Patients who were IgG-deficient were also more likely to have sustained lower levels of lymphocyte counts and higher levels of ferritin throughout the hospital stay. Furthermore, patients who were IgG-deficient compared with those with normal IgG levels displayed higher rates of acute kidney injury (76.9% vs. 26.5%; P = 0.001) and death (46.2% vs. 14.3%; P = 0.012), longer ICU [28 (6–48) vs. 12 (3–18) days; P = 0.012] and hospital length of stay [30 (22–50) vs. 18 (9–24) days; P = 0.004]. Univariable logistic regression showed increasing odds of 90-day overall mortality associated with IgG-deficiency (odds ratio 5.14, 95% confidence interval 1.3–19.9; P = 0.018). IgG deficiency might be common in patients with COVID-19 who are critically ill, and warrants investigation as both a marker of disease severity as well as a potential therapeutic target.

INTRODUCTION

Severe acute respiratory syndrome coronavirus type 2 (SARS-CoV-2) is a pathogen that has been confirmed in more than 32 million cases worldwide and is associated with a mortality rate of ∼3% [as of September 27, 2020, according to a World Health Organization (WHO) Report] (1). Although there is rapidly growing knowledge on the clinical course of SARS-CoV-2-induced coronavirus disease (termed COVID-19), data on host risk factors and treatment options beyond supportive care are limited. Importantly, as a result of this lack of evidence, there is uncertainty around the best clinical practices.

Respiratory failure from acute respiratory distress syndrome (ARDS) is the main cause of death in COVID-19 cases (2). Evidence suggests that a subgroup of patients with COVID-19 may develop ARDS as a result of a maladaptive, detrimental host inflammatory response to SARS-CoV-2 secondary to macrophage activation syndrome (MAS)/acquired hemophagocytic lymphohistocytosis (HLH) (3, 4). This is mirrored systematically by pronounced increases in inflammatory cytokines and lymphopenia, which are considered markers of disease severity (3, 5, 6). Thus, the abnormal immune response could be a target for the treatment of patients with COVID-19, and one of the potential treatment strategies to dampen this immune response is the administration of intravenous immunoglobulin (IVIg) (2, 6, 7). In addition, administration of IVIg can mitigate virus-induced immunosuppression and provide passive immune protection against a broad range of pathogens (2, 6, 7). However, administration of IVIg in the absence of clinical indicators (such as hypogammaglobulinemia secondary to immunodeficiencies) may be associated with adverse events (8). Importantly, the immunoglobulin levels in patients with COVID-19 who received IVIg have not been systematically reported (2, 6, 9). Therefore, our aim was to determine the incidence of immunoglobulin G (IgG) deficiency in hospitalized and critically ill patients with COVID-19.

METHODS

Study Design and Participants

This prospective observational study was approved by the institutional ethics board of Justus Liebig University Giessen (AZ 58/20) and was in compliance with the tenets of the Declaration of Helsinki. The study enrolled consecutive patients aged ≥ 18 yr with confirmed COVID-19 admitted to the intensive care units (ICUs) of the University Hospital of Giessen and Marburg and Hannover Medical School, Hannover, between March 16 and May 28, 2020. The participants or their legal representatives provided written informed consent. All patients with COVID-19 enrolled in this study were diagnosed according to the interim guidelines of the WHO (10).

Data Collection

We reviewed the clinical electronic medical records, nursing records, and laboratory findings for all patients. Clinical data were collected and entered into an electronic database. Any missing or unclear records were collected and clarified through direct communication with the involved healthcare providers and patients’ families. The date of disease onset was defined as the day when the first symptom was noticed. ARDS was diagnosed according to the 2012 Berlin definition (11). Lung-protective ventilation strategies were the standard of care in patients requiring invasive mechanical ventilation (12). Acute kidney injury (AKI) was diagnosed and staged using full 2012 Kidney Disease: Improving Global Outcomes (KDIGO) consensus criteria (13). Use of renal replacement therapy (RRT) was at the discretion of the attending physician rather than by predefined biochemical or clinical criteria. Bacterial pneumonia was defined as an acute infection diagnosed by positive bacterial cultures with known lung pathogenicity of sputum or bronchoalveolar lavage. Sepsis was defined according to the 2016 Third International Consensus Definition for Sepsis and Septic Shock (14). Urinary tract infection was defined as leukocyturia and the presence of positive bacterial urine cultures showing a bacterial colony count of ≥10⁵ colony-forming units/mL. Acute Physiology and Chronic Health Disease Classification System II (APACHE II) was determined upon hospital admission. The Sequential Organ Failure Assessment (SOFA) score was determined throughout hospital stay. Data on ventilator-free days, AKI, bacterial pneumonia, urinary tract infection, sepsis, ICU length of stay, and hospital length of stay were collected up to 28 days after admission, and data on overall mortality were collected up to 90 days after admission. In case the patient was discharged before 28 days postadmission, the patient was followed by telephone interviews.

Laboratory Methods

Blood was sampled for extended analyses within the first 24 h of ICU admission including inflammatory parameters, immunoglobulins, and lymphocyte subset count. Ferritin was measured with a chemiluminescence immunoassay assay on the Centaur XPT analyzer (Siemens Healthineers, Erlangen, Germany). Interleukin (IL)-6 was determined on the Siemens Immulite 1000 system with Siemens reagents. The Bio-Plex ProTM Human Isotyping Panel, 6-plex kit (Bio-Rad Laboratories, Inc., Hercules, CA) was used to measure the concentrations of immunoglobulins, including IgM, IgG1, IgG2, IgG3, and IgG4. CD3⁺/CD4⁺/CD8⁺ T cell and CD19⁺ B cell counts (cells/μL) were measured by multiple-color flow cytometry with human monoclonal anti-CD3-fluorescein isothiocyanate (FITC), anti-CD4-phycoerythrin (PE), anti- CD8-allophycocyanin (APC), and anti-CD19-PE antibodies (BD Multitest) according to the manufacturer’s instructions. The cells were analyzed on a BD FACS Canto II flow cytometry system (BD Biosciences).

Nasopharyngeal swabs and bronchoalveolar lavage fluid specimens were collected from patients with suspected and confirmed SARS-CoV-2 infection (eSwab, Bruker, Germany) and processed by the local Institutes of Medical Virology (Justus-Liebig University Giessen and Hannover Medical School). Viral nucleic acids were extracted from 400 µL from each sample using an EZ1 Advanced XL system and the EZ1 Mini Virus Kit v2.0 (Qiagen, Germany). Later, 10 µL of the resulting 60 µL eluate were used in a one-step real-time PCR according to the previously described protocols (15). Assessment of SARS-CoV-2 antibodies was performed with a chemiluminescent microparticle immunoassay for qualitative detection of IgG against SARS-CoV-2 nucleoprotein (SARS-CoV-2 IgG Reagent kit for use with ARCHITECT; Abbott Laboratories, Abbott Park, IL; Reference 6R86-22). Briefly, the amount of IgG antibodies to SARS-CoV-2 in each sample is determined by comparing its chemiluminescent relative light unit (RLU) to the calibrator RLU (index S/C). Using an index S/C threshold of 1.4, the manufacturer reported a sensitivity of 86.4% after 7 days from symptom onset and 100% after 14 days, and a specificity of 99.6%, using reverse transcription polymerase chain reaction as the gold standard (16).

Statistical Analyses

The aim of this study was to estimate the rate of immunoglobulin deficiency (IgG < 7 g/L) in critically ill patients during the study period. Therefore, no formal hypotheses were implemented to drive the sample size calculation, and we included the maximum number of patients who could be enrolled.

Numerical data are summarized by median [interquartile range (IQR)], nominal data are summarized by absolute and relative frequencies. To compare the patients with IgG < 7 g/L to those with IgG >7 g/L, we used the nonparametric Mann–Whitney U test and the χ² test for continuous and categorical variables, respectively. Correlation between IgG and other continuous variables was tested by the Spearman test. Correlation coefficient values > 0.3 were considered relevant. Time trends were analyzed using R 4.0.0 (R Center for Statistical Computing, Vienna, Austria) (17). The models were fit using the function lme4::lmer (18). The time variable was fit to a natural spline basis with 1 − 3 degrees of freedom (df), depending on the sparsity of the data. The spline basis allows to fit curved relationships between the response and time. Splines with 3 df can model time-courses with a maximum and a minimum, much like a polynomial with 3 degrees of freedom, but are less prone to exaggerated slopes at the limits of the range of predictor values (19). The spline basis was created using the function splines::ns(time, df) (17). The model contained the interaction of the group (categorical: high or low initial IgG) and the spline of the time. The random part of the model always refers to all coefficients of a spline with 1 df less than the fixed-effects spline. For a fixed-effects spline with 2 or 1 df, this represents a random slope and intercept, or a random intercept, respectively. All variables except APACHE II and SOFA were log-transformed in the models. P values are from likelihood-ratio tests. Model predictions and confidence bands were calculated using the package effects (20). Ninety-day overall mortality was calculated using a univariable logistic regression model. All other statistical analyses were performed using SPSS 23.0 software (IBM Corp., Armonk, NY), and P values < 0.05 were considered to indicate statistical significance.

RESULTS

Clinical Data and Treatment

This study included a total of 62 patients, and IgG deficiency (<7 g/L IgG) was identified in 13 cases or 21.0% of the total patients [11/49 (22.4%) in Giessen, and 2/13 (15.4%) in Hannover]. Supplemental Table S1 (all Supplemental Material is available at https://doi.org/10.6084/m9.figshare.13231796) summarizes the key clinical and laboratory findings for individual patients with IgG deficiency, and Table 1 (21, 22) summarizes the median values for the comparison of both groups (IgG < 7 g/L vs. IgG > 7 g/L). Six patients received IVIg on the day of admission after blood sampling, whereas four patients required additional administration of IVIg during the ICU course due to a further drop in IgG levels (Supplemental Table S1). None of the patients received an antiviral therapy. Patient 8 was administered tocilizumab 2 days before laboratory analysis. Patient 7 received maintenance immunosuppressive regimen comprising low-dose prednisolone, tacrolimus, and mycophenolate mofetil for lung transplantation. Only one patient who was among those with normal baseline IgG levels received systemic corticosteroids for the treatment of ARDS during the course of ICU. There was no recognizable association between IgG deficiency and patients’ history, except for two patients who were IgG-deficient who had received immunosuppression treatment for chronic lymphocytic leukemia and lung transplantation in the past.

Table 1.

Comparison of demographic, clinical, and laboratory data between patients with COVID-19 with reduced and normal IgG levels at baseline

	Normal Range	All Patients (n = 62)	IgG < 7 g/L (n = 13)	IgG > 7 g/L (n = 49)	P Value
Demographics
Sex–male	NA	45 (72.6%)	10 (76.9%)	35 (77.8%)	0.694
Age, yr	NA	61.0 [53.0–67.0]	65.0 [54.5–76.5]	59.0 [50.5–67.0]	0.132
Body mass index, kg/m2	18.5–24.9	25.5 [21.9–27.2]	26.8 [23.9–29.0]	25.3 [22.3–26.9]	0.132
Body surface area, m2	1.7–1.9	1.95 [1.81–2.07]	1.95 [1.81–2.06]	2.01 [1.80–2.07]	0.632
Duration from onset of disease to ICU admission, days	NA	9.0 [6.8–10.0]	5.0 [4.0–8.5]	9.0 [7.5–10.0]	0.007
Comorbidities
Hypertension	NA	36 (58.1%)	8 (61.5%)	28 (57.1%)	0.775
Diabetes	NA	12 (19.4%)	3 (23.1%)	9 (18.4%)	0.702
Chronic kidney disease	NA	4 (6.5%)	3 (23.1%)	1 (2.0%)	0.006
Clinical characteristics
APACHE II	0	9.0 [6.0–12.0]	12.0 [11.0–13.0]	7.0 [5.0–10.5]	0.003
SOFA	0	4.0 [3.0–9.0]	8.0 [4.5–10.5]	4.0 [3.0–7.3]	0.006
Norepinephrine, μg/kg/min	0	0.00 [0.00–0.07]	0.01 [0.00–0.42]	0.00 [0.00–0.03]	0.099
$\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$/${\text{FI}}_{{\text{O}}_{\text{2}}}$ ratio	400–500	207.0 [132.5–260.0]	115.0 [86.0–250.0]	214.0 [142.0–260.0]	0.038
Invasive ventilation	NA	34 (54.8%)	8 (61.5%)	26 (53.1%)	0.585
VT, mL/kg PBW	NA	7.1 [6.7–8.7]	7.8 [6.8–8.9]	6.9 [5.8–7.0]	0.040
VE, L/min	NA	10.0 [8.7–11.1]	10.0 [8.8–11.3]	9.2 [8.5–10.2]	0.182
PPLAT, cmH2O	NA	25 [20–28]	28 [22–29]	24 [20–28]	0.882
PEEP, cmH2O	NA	10 [8–13]	10 [9–13]	10 [8–13]	0.347
Driving pressure, cmH2O	NA	14 [11–16]	17 [13–18]	14 [11–16]	0.148
Compliance, mL/mbar	NA	39.2 [33.0–48.7]	36.3 [30.5–56.4]	40.5 [38.1–43.0]	0.456
ARDS adjuvant therapy
Prone positioning	NA	16 (25.8%)	8 (61.5%)	8 (16.3%)	0.001
Inhaled nitric oxide	NA	6 (9.7%)	1 (7.7%)	5 (10.2%)	0.785
ECMO	NA	5 (8.1%)	3 (23.1%)	2 (4.1%)	0.025
RRT	NA	6 (9.7%)	2 (15.4%)	4 (8.2%)	0.434
Antibiotic treatment
Penicillin derivate	NA	29 (46.8%)	6 (46.2%)	23 (46.9%)	0.960
Gyrase inhibitor	NA	14 (22.6%)	8 (16.3%)	6 (46.2%)	0.022
Cephalosporinderivate	NA	5 (8.1%)	3 (23.1%)	2 (4.1%)	0.025
Clindamycin	NA	3 (4.8%)	2 (15.4%)	1 (2.0%)	0.046
Carbapenem	NA	5 (8.1%)	1 (7.7%)	4 (8.2%)	0.956
Vancomycin	NA	5 (8.1%)	2 (15.4%)	3 (6.1%)	0.276
Macrolide	NA	5 (8.1%)	2 (15.4%)	3 (6.1%)	0.276
Laboratory findings
Leukocyte count, g/L	3.5–9.5	6.8 [4.8–10.4]	8.2 [5.3–15.0]a	6.3 [4.8–9.8]	0.358
Differential count, g/L
Total neutrophils	1.8–6.3	5.0 [3.0–7.4]	6.0 [3.1–11.2]	4.7 [3.0–7.1]	0.510
Total lymphocytes	1.1–3.2	0.9 [0.6–1.3]	0.7 [0.3–1.1]a	0.9 [0.7–1.3]	0.156
CD4+ T cells, cells/µLb	400–1400	405.0 [256.5–791.0]	148.5 [40.3–359.3]	631.0 [365.0–847.0]	0.004
CD8+ T cells, cells/µLb	200–900	162.5 [91.5–299.3]	110.0 [42.3–184.0]	241.8 [120.5–375.8]	0.047
CD4+/CD8+ ratiob	1.0–3.6	2.1 [1.5–4.1]	1.4 [1.0–2.4]	2.3 [1.9–5.7]	0.018
CD19+ B cells, cells/µLb	100–500	120.5 [67.0–287.5]	54.0 [10.1–112.8]a	190.5 [106.3–322.0]	0.018
Neutrophil/lymphocyte ratio	1.6–2.0	4.5 [3.7–9.6]	10.5 [10.4–21.6]	4.5 [3.7–4.6]	0.001
Total monocytes	0.1–0.6	0.5 [0.3–0.7]	0.5 [0.3–0.7]	0.5 [0.3–0.6]	0.910
Hematocrit, %	40–50	37.0 [32.7–41.0]	32.0 [28.5–38.0]	39.0 [35.5–41.0]	0.019
Platelet count, ×109/L	125–350	229.0 [149.5–284.5]	178.0 [130.0–271.5]	234.5 [170.0–286.0]	0.167
Creatinine, mg/dLc	0.5–1.2	0.90 [0.70–1.20]	1.20 [0.80–2.65]	0.80 [0.60–1.02]	0.031
Creatinine-based eGFR, mL/min/1.73 m2 d	>60	94.0 [57.0–106.0]	57.0 [24.5–96.0]	97.5 [73.8–106.3]	0.026
Creatinine-cystatin C based eGFR, mL/min/1.73 m2 e	>60	77.5 [42.0–97.5]	42.0 [18.0–86.5]	84.0 [57.0–101.5]	0.056
Urea, mg/dLf	10–50	30.0 [24.0–55.5]	49.0 [30.0–128.0]	25.5 [23.0–41.3]	0.003
Uric acid, mg/dL	2.4–5.7	4.1 [2.9–6.1]	6.2 [3.9–9.6]	3.9 [2.9–5.7]	0.044
Alanine aminotransferase, U/L	10–50	37 [26–59]	54 [33–65]	30 [24–46]	0.097
Aspartate aminotransferase, U/L	10–50	29 [17–57]	50 [20–103]	26 [16–52]	0.190
Lactate dehydrogenase, U/L	125–243	324 [254–447]	412.0 [345.0–557.5]	298.5 [250.5–426.0]	0.054
Creatine kinase, U/L	12–140	121.0 [52.0–241.0]	114.5 [41.5–229.0]	121.0 [53.0–300.5]	0.760
B-type natriuretic peptide, pg/mL	0–50	44.0 [14.5–94.0]	93.0 [31.0–129.0]	27.0 [12.0–91.0]	0.160
Albumin, g/L	32–48	35.7 [31.8–41.0]	32.8 [28.4–35.8]	36.3 [32.3–42.3]	0.141
Total IgG, g/L	8–17	9.0 [7.6–10.8]	5.3 [4.0–6.3]a	9.1 [8.5–11.3]	NA
IgG1, g/Lg	2.8–8	4.1 [2.6–5.5]	2.4 [2.0–3.1]a	4.4 [4.0–6.7]	<0.0001
IgG2, g/Lg	1.2–5.7	2.5 [1.9–3.7]	2.1 [1.5–2.1]a	3.5 [2.0–4.6]	0.008
IgG3, g/Lg	0.2–1.3	0.7 [0.5–0.9]	0.6 [0.2–0.7]a	0.8 [0.6–1.0]	0.056
IgG4, g/Lg	0.05–1.3	0.2 [0.1–0.3]	0.1 [0.04–0.2]a	0.3 [0.1–0.4]	0.015
IgM, g/L	0.4–2.3	0.8 [0.6–1.2]	0.6 [0.4–1.7]a	0.9 [0.7–1.2]	0.269
C-reactive protein, mg/L	<5	127.2 [59.7–193.3]	164.0 [112.2–266.2]	121.0 [49.5–177.7]	0.159
Procalcitonin, μg/L	<0.05	0.8 [0.5–2.0]	1.6 [0.5–36.5]	0.6 [0.5–1.6]	0.216
Interleukin-6, μg/L	<7	71.9 [18.4–247.9]	108.0 [61.9–1086.0]	25.8 [14.8–205.2]	0.099
Ferritin, μg/L	34–310	733.0 [269.5–1786.5]	1786.5 [341.0–5013.5]	657.0 [198.0–1242.3]	0.049
d-Dimer, mg/L	0–0.5	1.3 [0.6–3.8]	1.4 [0.8–8.9]	1.1 [0.6–3.3]	0.251

Values are expressed as median [interquartile range] or n (%). P values provided in italics indicate statistical significance (< 0.05).

^aPatient 2 (see Supplemental Table S1), who received two cycles of intravenous rituximab for the treatment of chronic lymphocytic leukemia at 3 mo before admission, was excluded.

^bLymphocyte subset count was performed in 21/62 (33.9%) patients.

^cTo convert the values for serum creatinine to μmol/L, multiply by 88.4.

^deGFR was calculated with the creatinine Chronic Kidney Disease Epidemiology Collaboration equation based on serum creatinine (21).

^eeGFR was calculated with the creatinine Chronic Kidney Disease Epidemiology Collaboration equation based on serum creatinine and cystatin C (22).

^fTo convert the values for urea to blood urea nitrogen, multiply by 0.467.

^gIgG subclasses analysis was performed in 32 patients.

APACHE II, Acute Physiology and Chronic Health Evaluation II; ARDS, acute respiratory distress syndrome; COVID-19, coronavirus disease 2019; ECMO, extracorporeal membrane oxygenation; eGFR, estimated glomerular filtration rate; ${\text{FI}}_{{\text{O}}_2}$, fraction of inspired oxygen; ICU, intensive care unit; IgG, immunoglobulin G; IgM, immunoglobulin M; $\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$, partial pressure of arterial oxygen; PBW, predicted body weight; PEEP, positive end-expiratory pressure; P_PLAT, plateau pressure; RRT, renal replacement therapy; SOFA, Sequential Organ Failure Assessment; NA, not available/not applicable; V_E, minute volume; V_T, tidal volume.

Comparison of Patients with COVID-19 with IgG Deficiency to Those without IgG Deficiency

Median IgG levels in patients with IgG deficiency versus those with normal IgG levels were 5.3 (IQR, 4.0–6.3) versus 9.1 (IQR, 8.5–11.3) g/L, respectively. Furthermore, IgG subclass analyses of both groups showed predominant IgG1, IgG2, and IgG4 deficiency (Fig. 1). Compared with patients with normal IgG levels, those with IgG deficiency had worse median measures of clinical disease severity, including a shorter duration from disease onset to ICU admission [5.0 (4.0–8.5) vs. 9.0 (7.5–10.0) days; P = 0.007], a lower ratio of partial pressure of arterial oxygen to the fraction of inspired oxygen [$\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$/${\text{FI}}_{{\text{O}}_2}$; 115.0 (86.0–250.0) vs. 214.0 (142.0–260.0); P = 0.038], both higher APACHE II [12.0 (11.0–13.0) vs. 7.0 (5.0–10.5); P = 0.003] and SOFA scores [8.0 (4.5–10.5) vs. 4.0 (3.0–7.3); P = 0.006] as well as higher levels of ferritin [1786.5 (341.0–5013.5) vs. 657.0 (198.0–1242.3) μg/L; P = 0.049], neutrophil-to-lymphocyte ratio (P = 0.001) and serum creatinine [1.20 (0.80–2.65) vs. 0.80 (0.60–1.02) mg/dL; P = 0.031] (Fig. 2). Furthermore, patients with IgG deficiency also had lower median levels of CD4⁺ than those with normal IgG levels [148.5 (40.3–359.3) vs. 631.0 (365.0–847.0) cells/µL; P = 0.004] and CD8⁺ T cells [110.0 (42.3–184.0) vs. 241.8 (120.5–375.8) cells/µL; P = 0.047] and CD19⁺ B cells [54.0 (10.1–112.8) vs. 190.5 (106.3–322.0) cells/µL; P = 0.018]; a finding that was not apparent when only comparing the leucocyte or lymphocyte counts between both groups (Fig. 3).

Figure 1.

Comparison of total IgG and IgG subclass levels in patients with COVID-19 with reduced vs. normal IgG levels at baseline. Comparison of levels of total IgG (A) and IgG1-4 subclasses (B–E) between patients with COVID-19 with reduced (<7 g/L) and normal (>7 g/L) IgG levels at baseline. Boxes show medians and IQR; whiskers show the 1.5 IQR of the 25th quartile or 1.5 IQR of the 75th quartile; and dots show individual observations. Extreme values of IgG4 were clipped. COVID-19, coronavirus disease 2019; IgG, immunoglobulin G; IQR; interquartile range.

Figure 2.

Comparison of measures of clinical disease severity in patients with COVID-19 with reduced vs. normal IgG levels at baseline. Comparison of duration of onset of disease to ICU admission (A), $\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$/${\text{FI}}_{{\text{O}}_2}$ (B), APACHE II (C), SOFA (D), neutrophil-to-lymphocyte ratio (E), ferritin (F), and serum creatinine (G) between patients with COVID-19 with reduced (<7 g/L) and normal (>7 g/L) IgG levels at baseline. Boxes show medians and IQR; whiskers show the 1.5 IQR of the 25th quartile or 1.5 IQR of the 75th quartile; and dots show individual observations. APACHE II, Acute Physiology and Chronic Health Evaluation II; COVID-19, coronavirus disease 2019; ${\text{FI}}_{{\text{O}}_2}$, fraction of inspired oxygen; ICU, intensive care unit; IgG, immunoglobulin G; IQR; interquartile range; $\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$, partial pressure of arterial oxygen; SOFA, Sequential Organ Failure Assessment.

Figure 3.

Comparison of leukocyte and lymphocyte subclasses count in patients with COVID-19 with reduced vs. normal IgG levels at baseline. Comparison of levels of leucocyte count (A), lymphocyte count (B), and lymphocyte subset count (C–F) between patients with COVID-19 with reduced (<7 g/L) and normal (>7 g/L) IgG levels at baseline. Boxes show medians and IQR; whiskers show the 1.5 IQR of the 25th quartile or 1.5 IQR of the 75th quartile; and dots show individual observations. Extreme values of CD4⁺/CD8⁺ ratio were clipped. *The lymphocyte count of Patient 2 (see Supplemental Table S1), who received two cycles of intravenous rituximab for the treatment of chronic lymphocytic leukemia at 3 mo before admission, was excluded. †Lymphocyte subset count was performed in 21/62 (33.9%) patients. COVID-19, coronavirus disease 2019; IgG, immunoglobulin G; IQR; interquartile range.

To investigate whether the observed IgG deficiency is specific to COVID-19, we examined data from the Giessen and Hannover ARDS Registry of 15 patients with influenza-associated ARDS who were comparable with the COVID-19 cohort in respect to clinical characteristics and were recruited between February 2018 and January 2020 (Supplemental Table S2). Briefly, IgG deficiency was detected in 4 (26.7%) patients, and median IgG levels in those with IgG deficiency versus those with normal IgG levels were 5.7 (IQR, 4.7–6.4) versus 8.7 (IQR, 8.0–12.1) g/L, respectively. None of the patients received IVIg before sampling.

Time Course of Clinical Characteristics and Laboratory Findings after Hospital Admission

Figure 4 illustrates the time course of key laboratory findings within the first 14 days after admission for patients hospitalized in Giessen with IgG deficiency and those without IgG deficiency at baseline. Patients with IgG deficiency compared with those with normal IgG levels at baseline had sustained lower levels of lymphocyte counts and IgG (despite IVIg) and higher levels of ferritin during the hospital stay. No difference was observed in respect to the time course of leucocyte count (P = 0.549), neutrophil/lymphocyte ratio (P = 0.056), lactate dehydrogenase (P = 0.763), albumin (P = 0.651), C-reactive protein (P = 0.188), procalcitonin (P = 0.937), IL-6 (P = 0.261), d-dimer (P = 0.780), SOFA score (P = 0.401), $\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$/${\text{FI}}_{{\text{O}}_2}$ ratio (P = 0.443), norepinephrine dose (P = 0.412), and SARS-CoV-2 viral load (P = 0.534) between both groups. Of note, among the patients with normal IgG levels at baseline, one patient received systemic corticosteroids for the treatment of ARDS during the ICU course and a mild IgG deficiency was observed during that treatment (lowest IgG level 6.9 g/L).

Figure 4.

Time course of laboratory findings and clinical characteristics in patients with COVID-19 with reduced and normal IgG levels at baseline. *6/13 (46.2%) patients with IgG deficiency at baseline received IVIg on the day of admission, whereas four required additional IVIg during the ICU course due to a further drop in IgG levels. The levels of lymphocyte count (A), IgG (B), and ferritin (C) in patients with COVID-19 with IgG-deficiency (<7 g/L IgG; red line) and normal IgG levels (>7 g/L IgG; blue line) at baseline were analyzed at different time points after hospital admission. The upper dotted lines show the upper normal limit of each parameter, and the lower dotted lines show the lower normal limit of each parameter. The graphs show the predictions from the generalized linear models for both groups (lines) with their 95% confidence bands (colored areas). The P values show the empirical significance of a difference in time trends between groups (group × time interaction, likelihood ratio test). COVID-19, coronavirus disease 2019; ICU, intensive care unit; IgG, immunoglobulin G; IVIg, intravenous immunoglobulin.

Clinical Outcomes

During the 28-day observational period, patients who were IgG-deficient at baseline compared with those with normal IgG levels were more likely to develop AKI (76.9% vs. 26.5%; P = 0.001), which was more frequently diagnosed at admission (80.0% vs. 38.5%; P = 0.046) (Table 2). Furthermore, the patients who were IgG-deficient versus those with normal IgG levels exhibited a longer ICU [median 28 (IQR, 6–48) vs. 12 (IQR, 3–18) days; P = 0.012] and hospital length of stay [median 30 (IQR, 22–50) vs. 18 (IQR, 9–24) days; P = 0.004] as well as a higher 90-day overall mortality rate (46.2% vs. 14.3%; P = 0.012). The odds ratio with univariable logistic regression for 90-day mortality with IgG-deficiency was 5.14 (95% confidence interval, 1.3–19.9; P = 0.018).

Table 2.

Outcome data of patients with COVID-19 with reduced and normal IgG levels at baseline

	IgG < 7 g/L (n = 13)	IgG > 7 g/L (n = 49)	P Value
Outcome
Ventilator-free days at 28 days	10 [0–28]	16 [12–28]	0.105
AKI	10 (76.9%)	13 (26.5%)	0.001
Stage 1	5 (50.0%)	6 (46.2%)
Stage 2	2 (20.0%)	1 (7.7%)
Stage 3	3 (30.0%)	6 (46.2%)
AKI at admission	8 (80.0%)	5 (38.5%)	0.046
Bacterial pneumonia*	2 (15.4%)	14 (28.6%)	0.334
Urinary tract infection*	0 (0%)	3 (6.1%)	0.360
Sepsis*	2 (15.4%)	10 (20.4%)	0.684
ICU length of stay, days	28 [6–48]	12 [3–18]	0.012
Hospital length of stay, days	30 [22–50]	18 [9–24]	0.004
90-day overall mortality	6 (46.2%)	7 (14.3%)	0.012
SARS-CoV-2 antibodies, index S/C†	5.8 [1.4–7.6]	7.4 [4.9–8.1]	0.181

Values are expressed as median [interquartile range] or n (%). P values provided in italics indicate statistical significance (< 0.05).

^*Detailed data on bacterial and fungal coinfection of the study cohort are provided in Supplemental Table S3.

^†SARS-CoV-2 antibodies were measured in 49/62 (79.0%) patients. One patient in the IgG-deficient group did not exhibit seroconversion for SARS-CoV-2 antibodies.

AKI, acute kidney injury; COVID-19, coronavirus disease 2019; ICU, intensive care unit; IgG, immunoglobulin G; SARS-CoV-2, severe acute respiratory syndrome coronavirus type 2.

Serum samples of survivors [49/62 (79.0%) patients] were tested for antibodies against SARS-CoV-2 after a median of 60 (IQR, 45–75) days after disease onset. All patients were tested positive for antiviral IgG, except one patient from the IgG-deficient group who showed a negative result (0.1 index S/O). Overall, there was no detectable difference in antibody titer between both groups. Furthermore, no difference was observed in the rate of bacterial or fungal coinfection between patients who were IgG-deficient compared with those with normal IgG levels.

DISCUSSION

In the present cohort, which included patients with COVID-19 who were critically ill admitted to two German ICUs, the incidence of IgG deficiency was common (21%) and IgG deficiency was associated with several clinical and laboratory markers of disease severity both at baseline and throughout the further hospital course. IgG deficiency at baseline was also associated with increased morbidity as well as short- and long-term mortality. Thus, IgG levels could help clinicians to identify at an early stage those patients with COVID-19 who have poor prognosis. In addition, these findings are critical considering the absence of specific treatment modalities or vaccination for COVID-19, as IgG could be a potential treatment target in which modulation is both established and feasible.

The IgG levels at baseline showed an association with ferritin levels and patients with IgG deficiency were more likely to have sustained lower levels of lymphocyte counts and higher levels of ferritin throughout the hospital stay. Accordingly, it has been reported that hyperferritinemia is essential to the diagnosis of associated MAS/HLH (23, 24) and ferritin is found elevated in many severe COVID-19 cases (6, 25). In addition, patients with IgG-deficiency in the present study also exhibited significantly lower levels of lymphocyte counts, in particular lower concentrations of CD4⁺ T cells, CD8⁺ T cells, and CD19⁺ B cells. Similarly, dysregulation of lymphocytes characterized by CD4⁺ and CD8⁺ lymphopenia and B cell lymphopenia has been described in severe SARS-CoV-2 infection and has been reported to be the result of the virally induced immunosuppression or direct viral cytopathic effects (26, 27). Thus, immunoglobulin deficiency may be common in patients with COVID-19 and particularly in patients with a severe clinical course.

IVIg has been used as a therapeutic agent in immunodeficiencies and as an immunomodulatory agent, based on the findings of clinical trials on Kawasaki disease, immune thrombocytopenia, Guillan–Barré syndrome, and chronic inflammatory demyelinating polyradiculoneuropathy (28). However, only limited data are available to support the routine use of standard IVIg in severe pneumonia or sepsis (29). It is likely that the hypogammaglobulinemia observed in our study is not unique to COVID-19. Indeed, IgG deficiency was similar between our COVID-19 and influenza cohort [13/62 (21.0%) versus 4/15 (26.7%) patients, respectively]. A recent study demonstrated that low IgG levels are frequently (∼40%) observed at the diagnosis of community-acquired pneumonia in those patients who require hospital admission (30). Interestingly, approximately half of the patients with low serum immunoglobulin levels in the acute phase had also persistent low levels in the convalescent phase (30). However, in that study, no differences related to outcome variables, such as length of hospital stay or mortality were observed. The multicenter, phase II CIGMA trial investigated a human polyclonal antibodies preparation (trimodulin) containing IgM, IgA, and IgG as add-on therapy to standard care in patients with severe community-acquired pneumonia (29). The primary end point of the study, the reduction of ventilator-free days, was not achieved and the mortality in the overall cohort was not different in patients who received trimodulin compared with those who did not (22.2% vs. 27.8%; P = 0.465) (29). However, post hoc subset analysis, which included the majority of the patients (58%–78%), showed significant reductions in all-cause mortality in patients with high CRP, low IgM, or both high CRP/low IgM (29). Thus, IVIg may be considered as a treatment option, but data from ongoing trials on COVID-19 (Trial Nos. NCT04350580 and NCT04261426) about immunoglobulin efficacy and potential side effects (8) must be acquired and analyzed. In line with this reasoning, the guidelines of the recent Surviving Sepsis Campaign on the management of critically ill adults with COVID-19 cautions against the routine use of IVIg (31). However, IVIg may be considered as a treatment option in patients with COVID-19 with documented IgG deficiency on a case-by-case basis. In addition, IVIg might be indicated in special cases in which immunodysregulation is evident, inflammatory parameters or cytokines are elevated, and other supportive therapies fail or are insufficient. Apart from these critically ill patients, in patients with chronic hypogammaglobulinemia, it is generally recommended that immunoglobulins be administered when the IgG levels are <4 g/L (8). Thus, once the presence of IgG deficiency is established in a patient with COVID-19, IVIg can be started on a case-by-case basis, according to the existing guidelines.

Since September 2020, the WHO recommends systemic corticosteroid therapy (6 mg dexamethasone daily for up to 10 days) for patients with severe and critical COVID-19 (32). However, corticosteroids may suppress humoral immunity and thus lead to hypogammaglobulinemia, although the dose and duration of corticosteroid therapy that might be associated with a suppression of humoral immunity are not completely understood. For example, one study (30) described that in steroid-dependent patients with asthma, hypogammaglobulinemia correlates with the daily dose of oral prednisolone and can be observed after 1 yr at a dose of 12.5 mg/day. Other chronic pulmonary diseases are also occasionally associated with low IgG levels. For example, in a retrospective study, patients with severe chronic obstructive pulmonary disease awaiting lung transplantation were found to have moderately low IgG levels, independently of corticosteroid therapy (33). Thus, currently it remains unclear whether dexamethasone at the dose and duration recommended by the WHO may alter serum IgG levels in patients with COVID-19 and warrants further investigation.

Recent studies have described a rapid decay of anti-SARS-CoV-2 antibodies after recovery from COVID-19 (34, 35). Although the protective role of antibodies against SARS-CoV-2 is currently unknown, these antibodies are usually a reasonable correlate of antiviral immunity, and antireceptor-binding domain antibody levels correspond to plasma viral neutralizing activity. In our study, SARS-CoV-2 antibodies after a median of 60 days after disease onset were similar between patients with IgG deficiency versus those with normal IgG levels but a longer duration of follow-up may be necessary to investigate differences in antibody levels decline between both groups, and whether decline in these antibodies increases risk of reinfection.

The limitations of the present study are the small sample size, dichotomization of IgG levels, and the higher variability of IgG levels in patients with IgG deficiency. Also, since there was no randomization and many patients with detected IgG deficiency at baseline received IVIg, only limited informational value can be drawn as regard to prognostic value of IgG deficiency and its influence on the time course of key laboratory findings and clinical characteristics in critically ill COVID-19 patients. Furthermore, data on kinetics of lymphocyte subset counts and anti-SARS-CoV-2 antibodies were not available.

In summary, IgG deficiency might be common in patients with COVID-19 who are and might serve as an indicator of both disease severity and potential inferior outcome. Prospective trials are warranted to clarify the role of IVIg as a potential treatment option in a targeted COVID-19 patient population. Furthermore, patients with documented IgG deficiency during COVID-19 should receive closer medical follow-up to evaluate future host adaption to SARS-CoV-2.

DATA AVAILABILITY

The data sets used and/or analyzed during the study are available from the corresponding author upon reasonable request.

SUPPLEMENTAL DATA

Supplemental Tables S1, S2, and S3: https://doi.org/10.6084/m9.figshare.13231796.

Supplemental References: https://doi.org/10.6084/m9.figshare.13231796.

GRANTS

I.V., S.H., W.S., and R.E.M. were supported by the German Center for Lung Research [Deutsches Zentrum für Lungenforschung (DZL)] and the German Research Foundation [Deutsche Forschungsgemeinschaft (DFG)] through EXC 2026 [390649896] and CRU 309 [284237345].

DISCLOSURES

W. Seeger discloses the receipt of personal fees for consultations from Bayer Pharma AG, Liquidia Technologies Inc., and United Therapeutics Corporation, but these consultations were not related to the submitted work. None of the other authors has any conflicts of interest, financial or otherwise, to disclose.

AUTHOR CONTRIBUTIONS

F.H.-S., I.V., W.S., T.W., and S.D. conceived and designed research; H.S., C.G.S., and S.D. performed experiments; F.H.-S, I.V., J.W., H.-D.W., W.S., H.-W.B., B.J., H.D., S.H., J.T., K.T., M.S., R.E.M., H.S., J.Z., S.K., C.R., F.F., K.W., K.S., B.S., M.M.H., T.W., and S.D. analyzed data; I.V., J.W., H.-D.W., W.S., H.-W.B., B.J., H.D., S.H., J.T., K.T., M.S., R.E.M., H.S., C.G.S., J.Z., S.K., C.R., F.F., K.W., K.S., B.S., M.M.H., T.W., and S.D. interpreted results of experiments; F.H.-S, I.V., J.W., H.-D.W., W.S., H.-W.B., S.H., M.S., R.E.M., C.R., T.W., and S.D. prepared figures; I.V., J.W., H.-D.W., W.S., H.-W.B., B.J., H.D., S.H., J.T., K.T., M.S., R.E.M., H.S., C.G.S., J.Z., S.K., C.R., F.F., K.W., K.S., B.S., M.M.H., T.W., and S.D. drafted manuscript; F.H.-S, I.V., J.W., H.-D.W., W.S., H.-W.B., B.J., H.D., S.H., J.T., K.T., M.S., R.E.M., H.S., C.G.S., J.Z., C.R., F.F., K.W., K.S., B.S., M.M.H., T.W., and S.D. edited and revised manuscript; F.H.-S, I.V., J.W., H.-D.W., W.S., H.-W.B., B.J., H.D., S.H., J.T., K.T., M.S., R.E.M., H.S., C.G.S., J.Z., S.K., C.R., F.F., K.W., K.S., B.S., M.M.H., T.W., and S.D. approved final version of manuscript.