Low IgG trough and lymphocyte subset counts are associated with hospitalization for COVID-19 in patients with primary antibody deficiency

Clinical Implications.

• Higher IgG trough in patients with primary antibody deficiency on IgG replacement therapy may reduce the need for hospitalization from coronavirus disease 2019. Monitoring lymphocyte counts in these patients may identify at-risk patients for hospitalization for coronavirus disease 2019.

High-dose IgG therapy has been used in coronavirus disease 2019 (COVID-19) to modulate inflammatory responses.¹ However, studies on the effect of IgG replacement therapy (IgGRT) on COVID-19 with primary antibody deficiency (PAD) are limited. IgGRT has been shown to modulate T-cell immunity and diminish proinflammatory responses of monocytes in common variable immune deficiency (CVID).^{2 , 3} A case report on clinical outcomes of a CVID patient with COVID-19 postulated benefit of both high-dose intravenous immunoglobulin treatment and/or compliance with IgGRT in reducing the severity of COVID-19.⁴ A study on cellular and humoral immune responses of 2 patients with CVID on IgGRT who presented with mild to asymptomatic COVID-19 showed robust CD4⁺ T-cell responses to severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), but 1 of the patients failed to mount SARS-CoV-2–specific antibody response.⁵ Well-established risk factors for COVID-19, including age and comorbidities, could also influence the outcome of COVID-19 in these patients.⁶ However, the function of IgGRT on modulation of cellular immunity and inflammation suggests a possible role for IgGRT in the outcomes of COVID-19 for these patients with PAD.^{2 , 3} Here, we investigated the clinical and immunologic characteristics of patients with PAD on IgGRT, in relation to their clinical outcomes of COVID-19, assessed by the hospitalization.

We performed a retrospective chart review of patients with COVID-19 from March 2020 to August 2021 at the Yale-New Haven Health System, Yale-affiliated community practices, and University of Virginia (UVA) Allergy and Immunology clinics. We collected demographics, comorbidities, home medications including IgGRT, laboratory data, and clinical outcomes of COVID-19 including hospitalization and mortality. We identified a total of 23 patients with PAD on IgGRT with SARS-CoV-2 infection based on positive nucleic-acid assay result (Yale-New Haven Health System [n = 10], Yale-affiliated community practices [n = 4], and UVA [n = 9]). Two of these patients were planned for, but not yet initiated on, IgGRT. This study was reviewed and approved by the institutional review boards of Yale University and UVA.

Of the total 23 PAD patients with COVID-19 identified in this cohort, 26% (n = 6) of patients were hospitalized whereas 74% (n = 17) were treated as outpatients (Table I ). None required intensive care unit management, and all patients survived. From this cohort, 56% (n = 13) of patients met the criteria for CVID and 43% (n = 10) of patients met the criteria for other PADs (Table I).⁷ Protective titer pneumococcal polysaccharide vaccine was defined as titer greater than or equal to 1.3 μg/mL in this study (see Table E1 in this article’s Online Repository at www.jaci-inpractice.org).⁷ However, given the lack of consensus on defining protective antibody response, we also included a lower threshold considered protective against invasive pneumococcal disease defined as titer greater than or equal to 0.35 μg/mL (Table E1).⁸ Criteria for diagnoses of CVID and other PADs are described in this article’s Online Repository’s Methods section at www.jaci-inpractice.org.⁷

Table I.

Clinical characteristics, outcome, and treatments of PAD patients with COVID-19

Patient	Age (y)	Sex	Ethnicity	Diagnosis	Hospitalized	CHF	Autoimmune disorder	Lung disease	Last CT chest (no. of years ago)	Malignancy history	Is therapy as outpatient?	COVID treatment
1	21	M	White	CVID	No	No	No	No	5	No	No	Casirivimab
2	25	F	White	CVID	No	No	No	No	6	No	No	No
3	35	F	White	CVID	No	No	No	Asthma	NA	No	No	No
4	64	F	White	CVID	No	No	Immune thrombocytopenic purpura	Bronchiectasis	6	Breast cancer	No	No
5	65	F	White	CVID	No	No	No	No	NA	No	No	No
6	65	F	White	CVID	No	No	No	No	5	No	No	No
7	69	F	White	CVID	No	No	No	Bronchiectasis	2	No	No	Bamlanivimab
8	71	M	White	CVID	No	No	No	No	7	No	No	No
9	32	F	White	CVID due to NFKB2 mutation	No	No	Alopecia areata	No	1	No	No	Bamlanivimab
10	39	M	White	Bruton’s agammaglobulinemia	No	No	No	No	4	No	No	Casirivimab/ imdevimab, prednisone
11	42	F	White	IgG deficiency	No	No	Hashimoto’s thyroiditis	No	1	No	No	No
12	65	F	White	IgG deficiency	No	No	No	Asthma	<1	Cervical cancer	No	No
13	69	M	White	IgG deficiency	No	No	No	Asthma	NA	No	No	No
14	71	M	White	IgG deficiency	No	No	No	COPD	1	Prostate cancer	No	Bamlanivimab
15	57	F	Hispanic	sAbD, IgG2 deficiency	No	No	No	Asthma, COPD, bronchiectasis	<1	No	No	No
16	42	F	White	sAbD	No	No	No	No	NA	No	No	Bamlanivimab
17	72	M	White	sAbD	No	No	No	COPD	1	No	No	No
18	58	F	White	CVID	Yes	No	No	Asthma	<1	No	No	Dexamethasone, remdesivir, convalescent plasma
19	68	M	White	CVID	Yes	No	Hashimoto’s thyroiditis	COPD, bronchiectasis	1	No	No	Dexamethasone, remdesivir
20	70	M	White	CVID	Yes	Yes	Sarcoidosis	Pulmonary sarcoidosis	2	No	Methotrexate	Dexamethasone, remdesivir
21	77	F	White	CVID	Yes	Yes	No	Asthma	3	No	No	Dexamethasone, remdesivir
22	18	F	White	Agammaglobulinemia due to TCF3 mutation	Yes	No	No	No	1	No	No	No
23	74	F	White	IgG deficiency	Yes	Yes	SLE, connective tissue overlap syndrome	Asthma	1	No	Leuflonimide, prednisone	Dexamethasone, remdesivir

COPD, Chronic obstructive pulmonary disorder; F, female; M, male; NA, information not applicable or not available; sAbD, specific antibody deficiency; SLE, systemic lupus erythematous.

Available laboratory data were compared between patients who were hospitalized and those managed as outpatients. Most recent IgG trough level (791 mg/dL vs 1090 mg/dL; P = .0033) and baseline absolute counts of CD3⁺ (683 cell/μL vs 1290 cell/μL; P = .0082), CD3⁺CD4⁺ (341 cell/μL vs 784 cell/μL; P = .0029), and CD19⁺ cells (24.3 cell/μL vs 101 cell/μL; P = .0081) were significantly lower in the hospitalized patients compared with the outpatients (Figure 1 ). There was no significant difference in age, sex, baseline IgG level before IgGRT, IgG monthly dosing, baseline absolute counts of CD3⁺CD8⁺ T cells and CD3⁻CD16⁺CD56⁺ natural killer cells, or switched memory B-cell percentages between these 2 groups (Table E1). Within patients with CVID only, hospitalization was associated with IgG trough level (806 mg/dL vs 1112 mg/dL; P = .0028) and baseline absolute CD3⁺CD4⁺ T-cell counts (276 cell/μL vs 652 cell/μL; P = .0485) (see Figure E1 in this article’s Online Repository at www.jaci-inpractice.org). Patients with CVID compared with all other patients with PAD had significantly higher IgGRT dose and lower levels of baseline IgA and IgM, but there were no significant differences in hospitalization, age, baseline IgG level, IgG trough, and other immune parameters.

Figure 1.

Serum IgG trough and absolute counts of peripheral lymphocyte subsets are lower in hospitalized patients with PAD on IgGRT than in outpatients. Serum IgG trough levels and baseline peripheral absolute counts of CD3⁺ and CD3⁺CD4⁺ T cells, and CD19⁺ B cells, were compared between hospitalized and nonhospitalized patients with PAD on IgGRT who were diagnosed with COVID-19.

Figure E1.

CVID-only, hospitalized vs outpatient lab comparisons.

Next, we compared the well-established comorbidities associated with poor outcome of COVID-19 between hospitalized and nonhospitalized PAD patients with COVID-19.⁶ Frequencies of lung disease, autoimmune disorder, malignancy, use of immunosuppressive therapy, hypertension, coronary artery disease, diabetes mellitus, obesity, chronic kidney disease, smoking, and congestive heart failure (CHF) were compared.⁶ Of these, hospitalized patients presented with a significantly higher frequency of CHF than nonhospitalized patients (50% vs 0%; P = .0113; Table I). Although lung disease did not show a significant difference between hospitalized patients and outpatients in all patients with PAD, an increased percentage of lung disease was observed in hospitalized patients with CVID compared with outpatients (100% vs 33%; P = .069; Table I). Many of our patients did not have recent chest imaging (Table I); therefore, assessment of current lung disease was limited.

This study showed that hospitalized patients with PAD on IgGRT due to COVID-19 presented with lower IgG trough level and absolute counts of CD3⁺, CD3⁺CD4⁺ T cells, and CD19⁺ B cells, as well as higher frequency of CHF compared with nonhospitalized patients. Lymphocytopenia and exhaustion of T lymphocytes have been associated with worse COVID-19 outcomes.⁹ CHF is also one of the well-described comorbidities associated with poor outcome of COVID-19.⁶ However, to our knowledge, this is the first study demonstrating an association of lower IgG trough level with hospitalization for COVID-19 in patients with PAD receiving IgGRT. High-dose immunoglobulin therapy (1 g/kg) is frequently used as an immune-modulating therapy for autoimmune and inflammatory disorders.¹ Although IgGRT is typically lower in dose (400-600 mg/kg) in PAD, IgGRT was also shown to modulate T-cell immunity by increasing CD4⁺ T-cell counts and suppressing CD8⁺ T-cell activation,² and diminishing proinflammatory responses of monocytes.³ Therefore, it is tempting to speculate that maintaining higher IgG trough levels may positively modulate the cellular immune response and inflammation, leading to favorable outcome of COVID-19.

Limitations of the study include the small sample size. Inherent to the retrospective chart review design, some subjects had incomplete clinical and laboratory values available for analysis. Six nonhospitalized subjects were treated with COVID-19 monoclonal antibody therapy as outpatients. Likely, their initial presentation was not severe for hospitalization; however, it is also feasible that this therapy may have later protected them from hospitalization. Further investigations are warranted to better characterize the effect of IgGRT on immune modulation and protection from severe disease outcomes of COVID-19 in patients with PAD.

Online Repository

Methods

Defining diagnoses of PADs

CVID was defined with low IgG, IgA, and/or IgM levels (2 SDs below the mean for age) and decreased immune response to a pneumococcal polysaccharide vaccine (PNA) vaccine.^E1 For PNA vaccine response, we used either 14 or 23 serotype panel that was available for both prevaccination and postvaccination. Protective titer was defined as titer greater than or equal to 1.3 μg/mL, and patients with more than 70% of serotypes with protective titers after vaccination were considered to have protective PNA vaccine response.^E1 IgG deficiency was assigned if IgG levels were lower than 2 SD below the mean for age without meeting the criteria for CVID.^E1 IgG subclass deficiency was assigned if IgG subclass 1, 2, 3, or 4 levels were lower than 2 SD below the mean of age, with normal levels of total IgG and other IgG subclasses.^E1 Specific antibody deficiency was assigned if a patient had a low immune response to a PNA vaccine but normal levels of IgG, IgA, and IgM.^E1 Patients were excluded from this study if chart review revealed an alternative cause for hypogammaglobulinemia such as intestinal loss, chronic corticosteroid use, preexisting diagnoses of hematopoietic malignancies before the diagnosis of PADs, or use of B-cell–depleting therapies.

Table E1.

Immunologic characteristics, IgGRT dose, and IgG trough levels of PAD patients with COVID-19

Patient	Age (y)	Sex	Baseline (mg/dL)			Baseline (cells/mm3)					Baseline (%)	Pneumococcal vaccine response			IgG trough (mg/dL)	IgGRT dose (mg/kg/mo)
			IgG	IgA	IgM	CD3	CD4	CD8	NK	CD19	smB	Clinically noted poor response?∗	Serotypes with titer ≥1.3 μg/mL	Serotypes with titer ≥0.35 μg/mL
1	21	M	490	0	20	1701	707	714	105	27	0.8	NA	8 of 23	17 of 23	1096	925
2	25	F	591	14	32	1448	766	590	38	84	1	Yes	8 of 23	22 of 23	1000	430
3	35	F	638	0	5	NA	NA	NA	NA	NA	NA	Yes	3 of 14	8 of 14	1350	820
4	64	F	74	0	67	994	783	192	104	71	0.1	Yes	0 of 14	1 of 14	1017	960
5	65	F	463	62	65	727	570	191	220	124	NA	Yes	8 of 23	18 of 23	986	230
6	65	F	453	52	55	993	741	234	227	122	5.6	NA	1 of 14	NA	1072	400
7	69	F	NA	35	22	1562	475	1025	152	19	NA	NA	NA	NA	1141	515
8	71	M	435	67	27	1218	1003	193	263	64	NA	Yes	5 of 14	9 of 14	1132	500
9	32	F	NA	0	0	360	170	180	37	106	NA	NA	NA	NA	1211	455
10	39	M	NA	0	17	2430	1010	792	273	0	NA	NA	NA	NA	526	200
11	42	F	NA	218	127	1235	793	402	233	166	NA	Yes	NA	NA	1146	500
12	65	F	454	155	122	975	654	321	71	119	12.8	Yes	7 of 23	13 of 23	1207	480
13	69	M	586	173	63	1034	740	262	NA	NA	NA	No	17 of 23	23 of 23	586	†
14	71	M	310	357	65	1745	1299	467	203	61	19.2	Yes	5 of 14	11 of 14	1062	410
15	57	F	1109	407	93	1623	1267	430	796	290	NA	Yes	6 of 23	17 of 23	1633	350
16	42	F	875	393	142	1047	645	399	NA	206	NA	NA	1 of 23	8 of 23	1330	380
17	72	M	692	314	NA	1479	922	556	192	52	NA	NA	NA	NA	1037	420
18	58	F	527	110	26	974	554	358	113	60	15.8	Yes	NA	NA	527	†
19	68	M	198	9	0	NA	NA	NA	NA	NA	0.1	Yes	0 of 14	0 of 14	900	600
20	70	M	486	61	10	220	145	64	173	12	4.8	Yes	9 of 23	17 of 23	896	350
21	77	F	NA	59	94	524	129	98	206	19	18.3	NA	NA	NA	901	540
22	18	F	0	0	0	1000	603	338	197	15	NA	NA	NA	NA	995	400
23	74	F	454	268	80	699	272	436	50	16	8.2	Yes	2 of 14	6 of 14	528	400

F, Female; M, male; NA, information not available; NK, natural killer; smB, switched memory B.

^∗Clinically noted in chart as poor pneumococcal vaccine responder in charting documentation, independent of available laboratory data in our electronic medical system because some patients had outside workup done before establishing care.

^†Not yet started on IgGRT.