Key Points
Hypogammaglobulinemia, infections/severe infections, and antimicrobial use were decreased in patients with CLL or NHL treated with IgRT.
Increased IgG testing was associated with increased hypogammaglobulinemia detection and lower rates of severe infection.
Visual Abstract
Patients with chronic lymphocytic leukemia (CLL) and non-Hodgkin lymphoma (NHL) can develop hypogammaglobulinemia, a form of secondary immune deficiency (SID), from the disease and treatments. Patients with hypogammaglobulinemia with recurrent infections may benefit from immunoglobulin replacement therapy (IgRT). This study evaluated patterns of immunoglobulin G (IgG) testing and the effectiveness of IgRT in real-world patients with CLL or NHL. A retrospective, longitudinal study was conducted among adult patients diagnosed with CLL or NHL. Clinical data from the Massachusetts General Brigham Research Patient Data Registry were used. IgG testing, infections, and antimicrobial use were compared before vs 3, 6, and 12 months after IgRT initiation. Generalized estimating equation logistic regression models were used to estimate odds ratios, 95% confidence intervals, and P values. The study population included 17 192 patients (CLL: n = 3960; median age, 68 years; NHL: n = 13 232; median age, 64 years). In the CLL and NHL cohorts, 67% and 51.2% had IgG testing, and 6.5% and 4.7% received IgRT, respectively. After IgRT initiation, the proportion of patients with hypogammaglobulinemia, the odds of infections or severe infections, and associated antimicrobial use, decreased significantly. Increased frequency of IgG testing was associated with a significantly lower likelihood of severe infection. In conclusion, in real-world patients with CLL or NHL, IgRT was associated with significant reductions in hypogammaglobulinemia, infections, severe infections, and associated antimicrobials. Optimizing IgG testing and IgRT are warranted for the comprehensive management of SID in patients with CLL or NHL.
Introduction
B-cell lymphoma/leukemia such as chronic lymphocytic leukemia (CLL) and non-Hodgkin lymphomas (NHL) are associated with secondary immune deficiencies (SIDs).1 Many lymphoid malignancies are themselves associated with SIDs. This is amplified by therapies such as chemotherapy, monoclonal antibodies (mAbs), immunomodulatory drugs, targeted therapies, and chimeric antigen receptor (CAR) T-cell therapies, which have improved lymphoma outcomes but also contribute to SIDs.2 Hypogammaglobulinemia is a common form of SID and contributes to infection risk in patients with CLL and NHL.2,3 Infections are major drivers of mortality, accounting for up to 50% of cases in CLL and up to 33% of cases in NHL.2,4 Despite this, a gap in efforts to address SIDs and reduce infections persists.2,5-7
Current guidelines (eg, National Comprehensive Cancer Network, European Society for Medical Oncology, and European Medicines Agency) advise routine immunoglobulin G (IgG) testing for CLL and NHL, although specific recommendations vary.8,9 The National Comprehensive Cancer Network CLL guidelines recommend evaluating IgG levels in patients with infections requiring hospitalizations or intravenous antibiotics, and starting monthly immunoglobulin replacement therapy (IgRT) if IgG is <500 mg/dL, whereas the European Medicines Agency recommends IgRT use in patients with severe or recurrent infections if IgG is <400 mg/dL. In contrast, the American Academy of Allergy, Asthma and Immunology recommends monitoring immunoglobulins and specific antibody responses every 6 months before treating with B-cell targeted therapy, anti-CD20 mAb, or CAR T-cell therapy, or as needed on the basis of patient’s infection history in patients with CLL.10 For hematologists, recognition of hypogammaglobulinemia is typically defined based on IgG levels alone, whereas immunologists may perform more comprehensive assessment including vaccine response testing to assess functional deficiencies.
IgRT is used to increase IgG levels and improve patients’ ability to overcome infection.11 Despite existing guidelines, many patients do not undergo regular IgG monitoring at diagnosis, before treatments associated with hypogammaglobulinemia (eg, rituximab and obinutuzumab), or after infections, and few patients undergo routine monitoring.12 The lack of harmonization among current consensus guidelines with respect to optimal frequency/timing of IgG testing, the definition of what constitutes hypogammaglobulinemia, and when to initiate IgRT, represents an unmet need and a potential opportunity to improve patient outcomes.10,13 This study was undertaken in order to examine IgG testing patterns and assess the effectiveness of IgRT for treating hypogammaglobulinemia and reducing infection rates in real-world patients with CLL and NHL.
Methods
Data source
This study used clinical data (inpatient and outpatient) from the Massachusetts General Brigham (MGB) Research Patient Data Registry (RPDR).14 The registry gathers data from 8 affiliated hospitals (Massachusetts General Hospital, Brigham and Women’s Hospital, Brigham and Women’s Faulkner Hospital, Massachusetts Eye and Ear Hospital, McLean Hospital, Newton-Wellesley Hospital, North Shore Medical Center, and Spaulding Rehabilitation Hospital) and the Dana-Farber Cancer Institute in Massachusetts. The data presented herein are the results of a preplanned study of this database. The MGB institutional review board reviewed the study protocol (no. 2022P000465) and waived the requirement for documentation of informed consent.
Study design
This was a retrospective, longitudinal, observational study of adult patients (aged ≥18 years) in the MGB RPDR with a diagnosis of CLL and/or NHL identified using International Classification of Diseases, ninth (ICD-9) and International Classification of Diseases, tenth revision (ICD-10) revision codes (supplemental Table 1) between 1 January 2010 and 15 February 2023 (ie, study period). The follow-up spanned from CLL/NHL diagnosis to the end of clinical information, data availability (cutoff: 15 February 2023), or death, whichever was first (supplemental Figure 1).
Study population
Patients in the CLL and NHL cohorts were identified separately. Eligible patients were aged ≥18 years at the time of initial CLL/NHL diagnosis and had ≥12 months of clinical data after diagnosis, unless the patient passed away. Additionally, to ensure that patients received continuous care within the MGB network, patients were required to have ≥3 annual visits (outpatient, inpatient, or emergency department visits) during the study period. Patients were excluded if they were diagnosed with another primary malignancy requiring systemic therapy, a primary immunodeficiency disease or HIV/AIDS, or had a history of solid organ transplantation.
IgRT subcohorts were identified in the CLL and NHL cohorts. Patients in the subcohorts had ≥1 record of IgRT during the follow-up period (with the date of first receipt of IgRT defined as the index date) and ≥3 months of follow-up before and after the index date (supplemental Figure 2).
Outcomes
IgG levels and the proportion of patients with hypogammaglobulinemia (defined as IgG level of <500 mg/dL) in patients with ≥1 record of IgG testing, as well as the rate of infections and associated antimicrobial use were compared 3, 6, and 12 months before vs after the index date. IgG testing was identified using Logical Observation Identifiers Names and Codes code 2465-3 (IgG [mass/volume] in serum or plasma) for the CLL and NHL cohorts. Infections were identified using ICD-9 and ICD-10 diagnosis codes (supplemental Table 1) and antimicrobial use (ie, antibiotics, antivirals, and antifungals) associated with infections were defined as a prescription for an antimicrobial within 30 days of infection diagnosis. Other evaluated outcomes were severe infections (defined as infection leading to hospitalization or treatment with any intravenous antibiotic, antiviral, or antifungal medication) and antimicrobial use for severe infections, defined as a prescription for an antimicrobial (ie, antibiotics, antivirals, or antifungals) within 30 days of a severe infection diagnosis. For the CLL and NHL IgRT subcohorts, all outcomes were evaluated at 3, 6, and 12 months before and after the index date to determine the short- and long-term effectiveness of IgRT.
Statistical analysis
Patient demographics, clinical characteristics (eg, body mass index, diabetes, neutropenia, sinusitis, and bronchitis) and treatment characteristics for the CLL and NHL cohorts were summarized with descriptive statistics including mean (±standard deviation), and median with interquartile range (IQR) for continuous variables, and frequency (proportion) for categorical variables. Generalized estimating-equation logistic regression models were used to estimate the odds ratios (ORs), 95% confidence intervals (CIs), and P values for infections, severe infections, and antimicrobial use associated with infections and with severe infections. A multivariable logistic regression model was used to examine the factors associated with the occurrence of the first severe infection, controlling for key confounders (eg, age and sex). The final model was chosen via Akaike Information Criterion; results were reported as OR with 95% CI. In addition, as a sensitivity analysis, hypogammaglobulinemia was included as an additional covariate in the multivariable logistic regression model. All analyses were conducted using SAS Enterprise Guide, version 9.4 (SAS institute, Cary, NC).
Subgroup analyses
A subgroup analysis was performed for patients in the NHL cohort who were treated with CAR T-cell therapy. Additionally, subgroup analyses were also performed for patients in each cohort who received IgRT with immune globulin infusion (human) 10% (Gammagard Liquid). Statistical methods described in the main analysis were applied to subgroup analyses.
Data were deidentified and comply with the patient requirements of the Health Insurance Portability and Accountability Act of 1996. Therefore, the institutional review board determined that the study met the criteria for exempt research per Title 45 of Code of Federal Regulations, Part 46.101(d)(#).
Results
Patient characteristics
The study population included 17 192 patients with CLL (n = 3960) or an NHL subtype (n = 13 232; supplemental Figure 3). In the CLL cohort (n = 3960), the median age was 68 years, 61.2% were male, 92.2% were White, and the median follow-up was 4.8 years (Table 1). The most frequent comorbidities were diabetes (15.9%), neutropenia (10.7%), sinusitis (9.2%), and bronchitis (5.3%). Previous therapies included targeted therapy (30.4%), mAbs (29.2%), cytotoxic chemotherapy (19.9%), hematopoietic stem cell transplantation (5.5%), and CAR T-cell therapy (2.0%). Demographic, clinical, and treatment characteristics for patients receiving ≥1 IgG testing (n = 2652), having hypogammaglobulinemia (n = 917), and receiving IgRT (n = 259) are summarized in supplemental Table 2.
Characteristics . | CLL (n = 3960) . | NHL (n = 13 232) . |
---|---|---|
Demographic characteristics | ||
Year of first diagnosis, n (%) | ||
Between 2010 and 2015 | 1319 (33.3) | 3692 (27.9) |
Between 2016 and 2023 | 2641 (66.7) | 9540 (72.1) |
Age at diagnosis, y | ||
Mean ± SD | 67.7 ± 11.9 | 62.3 ± 15.4 |
Median (IQR) | 68.0 (60.0-76.0) | 64.0 (54.0-73.0) |
18-44, n (%) | 121 (3.1) | 1739 (13.1) |
45-64, n (%) | 1351 (34.1) | 4882 (36.9) |
≥65, n (%) | 2488 (62.8) | 6611 (50.0) |
Sex, n (%) | ||
Male | 2422 (61.2) | 7317 (55.3) |
Female | 1538 (38.8) | 5915 (44.7) |
Race, n (%) | ||
White | 3653 (92.2) | 9042 (87.3) |
Black or African American | 88 (2.2) | 335 (3.2) |
Asian | 45 (1.1) | 300 (2.9) |
American Indian or Alaska Native | 4 (0.1) | 12 (0.1) |
Native Hawaiian and other Pacific Islander | 1 (0.0) | 2 (0.0) |
Multiracial | 26 (0.7) | 47 (0.5) |
Other | 55 (1.4) | 304 (2.9) |
Unknown | 88 (2.2) | 320 (3.1) |
BMI,∗ kg/m2 | ||
Mean ± SD | 28.3 ± 5.7 | 28.6 ± 56.9 |
Median (IQR) | 27.5 (24.4-31.2) | 27.3 (24.0-31.1) |
Unknown | 571 (14.4) | 2253 (17.0) |
Clinical conditions,† n (%) | ||
Diabetes | 630 (15.9) | 2041 (15.4) |
Neutropenia | 425 (10.7) | 2831 (21.4) |
Sinusitis | 365 (9.2) | 1083 (8.2) |
Bronchitis | 211 (5.3) | 712 (5.4) |
Characteristics . | CLL (n = 3960) . | NHL (n = 13 232) . |
---|---|---|
Demographic characteristics | ||
Year of first diagnosis, n (%) | ||
Between 2010 and 2015 | 1319 (33.3) | 3692 (27.9) |
Between 2016 and 2023 | 2641 (66.7) | 9540 (72.1) |
Age at diagnosis, y | ||
Mean ± SD | 67.7 ± 11.9 | 62.3 ± 15.4 |
Median (IQR) | 68.0 (60.0-76.0) | 64.0 (54.0-73.0) |
18-44, n (%) | 121 (3.1) | 1739 (13.1) |
45-64, n (%) | 1351 (34.1) | 4882 (36.9) |
≥65, n (%) | 2488 (62.8) | 6611 (50.0) |
Sex, n (%) | ||
Male | 2422 (61.2) | 7317 (55.3) |
Female | 1538 (38.8) | 5915 (44.7) |
Race, n (%) | ||
White | 3653 (92.2) | 9042 (87.3) |
Black or African American | 88 (2.2) | 335 (3.2) |
Asian | 45 (1.1) | 300 (2.9) |
American Indian or Alaska Native | 4 (0.1) | 12 (0.1) |
Native Hawaiian and other Pacific Islander | 1 (0.0) | 2 (0.0) |
Multiracial | 26 (0.7) | 47 (0.5) |
Other | 55 (1.4) | 304 (2.9) |
Unknown | 88 (2.2) | 320 (3.1) |
BMI,∗ kg/m2 | ||
Mean ± SD | 28.3 ± 5.7 | 28.6 ± 56.9 |
Median (IQR) | 27.5 (24.4-31.2) | 27.3 (24.0-31.1) |
Unknown | 571 (14.4) | 2253 (17.0) |
Clinical conditions,† n (%) | ||
Diabetes | 630 (15.9) | 2041 (15.4) |
Neutropenia | 425 (10.7) | 2831 (21.4) |
Sinusitis | 365 (9.2) | 1083 (8.2) |
Bronchitis | 211 (5.3) | 712 (5.4) |
BMI, body mass index; SD, standard deviation.
BMI value closest to the date of CLL diagnosis was measured.
Clinical conditions were assessed throughout the study period.
In the NHL cohort (n = 13 232), the median age was 64 years, 55.3% were male, 87.3% were White, and the median follow-up was 4.4 years (Table 1). Histological subtypes included B-cell lymphoma (72.3%), T-cell/natural killer–cell lymphomas (11.6%), and unspecified (20.5%). The histology subtypes were not mutually exclusive, because patients could present with >1 histology subtype, highlighting the heterogeneous nature of NHL. The most frequent comorbidities included neutropenia (21.4%), diabetes (15.4%), sinusitis (8.2%), and bronchitis (5.4%). Previous therapies included anti-CD20 mAbs (38.9%), hematopoietic stem cell transplantation (15.2%), and CAR T-cell therapy (6.9%). Demographic, clinical, and treatment characteristics for patients receiving ≥1 IgG testing (n = 6773), having hypogammaglobulinemia (n = 2411), and receiving IgRT (n = 623) are summarized in supplemental Table 2.
IgG assessment during follow-up
In the CLL cohort, 67.0% (n = 2652) and 47.2% (n = 1743) had ≥1 or ≥2 IgG tests during the follow-up period, respectively, and the median number of IgG tests was 2 (IQR, 1-7; Table 2). Among those with ≥1 IgG test (n = 2652), 34.6% (n = 917) patients had hypogammaglobulinemia. IgG testing was performed at the time of initial CLL diagnosis in 29% of patients, and the median time from diagnosis to first IgG test was 3 months (IQR, 0.6-14.8).
Characteristics . | CLL (n = 2652) . | NHL (n = 6773) . |
---|---|---|
Number of IgG testing per patient during the follow-up period | ||
Mean ± SD | 6.3 ± 10.0 | 7.8 ± 13.2 |
Median (IQR) | 2.0 (1.0-7.0) | 2.0 (1.0-8.0) |
Number of IgG testing per patient during the follow-up period, n (%) | ||
1 | 909 (34.3) | 2467 (36.4) |
2 | 422 (15.9) | 1042 (15.4) |
3 | 269 (10.1) | 502 (7.4) |
4 | 163 (6.1) | 337 (5.0) |
5 | 120 (4.5) | 233 (3.4) |
6 | 88 (3.3) | 204 (3.0) |
7 | 83 (3.1) | 166 (2.5) |
≥8 | 598 (22.5) | 1822 (26.9) |
Number of IgG testing per patient per year | ||
Mean ± SD | 1.5 ± 4.2 | 2.5 ± 15.7 |
Median (IQR) | 0.6 (0.3-1.4) | 0.7 (0.3-2.2) |
Number of patients with IgG testing on the date of first diagnosis, n (%) | 768 (29.0) | 1127 (16.6) |
Number of patients with IgG testing after the date of first diagnosis, n (%) | 1884 (71.0) | 5646 (83.4) |
Time from diagnosis to first IgG testing, mo | ||
Mean ± SD | 13.0 ± 21.6 | 12.8 ± 21.8 |
Median (IQR) | 3.0 (0.6-14.8) | 2.8 (0.5-14.6) |
Time from first IgG testing to end of follow-up, y | ||
Mean ± SD | 4.9 ± 3.2 | 4.6 ± 3.3 |
Median (IQR) | 4.5 (2.1-7.2) | 4.0 (1.8-6.8) |
Proportion of patients with hypogammaglobulinemia, n (%) | ||
Any IgG test level of <500 mg/dL | 917 (34.6) | 2411 (35.6) |
First IgG test level of <500 mg/dL | 593 (22.4) | 1615 (23.8) |
Lowest IgG test result (mg/dL)∗ , n (%) | ||
<200 | 250 (9.4) | 627 (9.3) |
≥200 to <400 | 430 (16.2) | 1115 (16.5) |
≥400 to <500 | 237 (8.9) | 669 (9.9) |
≥500 to <600 | 317 (12.0) | 688 (10.2) |
≥600 | 1418 (53.5) | 3674 (54.2) |
Characteristics . | CLL (n = 2652) . | NHL (n = 6773) . |
---|---|---|
Number of IgG testing per patient during the follow-up period | ||
Mean ± SD | 6.3 ± 10.0 | 7.8 ± 13.2 |
Median (IQR) | 2.0 (1.0-7.0) | 2.0 (1.0-8.0) |
Number of IgG testing per patient during the follow-up period, n (%) | ||
1 | 909 (34.3) | 2467 (36.4) |
2 | 422 (15.9) | 1042 (15.4) |
3 | 269 (10.1) | 502 (7.4) |
4 | 163 (6.1) | 337 (5.0) |
5 | 120 (4.5) | 233 (3.4) |
6 | 88 (3.3) | 204 (3.0) |
7 | 83 (3.1) | 166 (2.5) |
≥8 | 598 (22.5) | 1822 (26.9) |
Number of IgG testing per patient per year | ||
Mean ± SD | 1.5 ± 4.2 | 2.5 ± 15.7 |
Median (IQR) | 0.6 (0.3-1.4) | 0.7 (0.3-2.2) |
Number of patients with IgG testing on the date of first diagnosis, n (%) | 768 (29.0) | 1127 (16.6) |
Number of patients with IgG testing after the date of first diagnosis, n (%) | 1884 (71.0) | 5646 (83.4) |
Time from diagnosis to first IgG testing, mo | ||
Mean ± SD | 13.0 ± 21.6 | 12.8 ± 21.8 |
Median (IQR) | 3.0 (0.6-14.8) | 2.8 (0.5-14.6) |
Time from first IgG testing to end of follow-up, y | ||
Mean ± SD | 4.9 ± 3.2 | 4.6 ± 3.3 |
Median (IQR) | 4.5 (2.1-7.2) | 4.0 (1.8-6.8) |
Proportion of patients with hypogammaglobulinemia, n (%) | ||
Any IgG test level of <500 mg/dL | 917 (34.6) | 2411 (35.6) |
First IgG test level of <500 mg/dL | 593 (22.4) | 1615 (23.8) |
Lowest IgG test result (mg/dL)∗ , n (%) | ||
<200 | 250 (9.4) | 627 (9.3) |
≥200 to <400 | 430 (16.2) | 1115 (16.5) |
≥400 to <500 | 237 (8.9) | 669 (9.9) |
≥500 to <600 | 317 (12.0) | 688 (10.2) |
≥600 | 1418 (53.5) | 3674 (54.2) |
SD, standard deviation.
For patients with multiple IgG tests, the lowest test result during the follow-up was used.
In the NHL cohort, 51.2% (n = 6773) and 32.5% (n = 4306) had ≥1 or ≥2 IgG test during the follow-up period, respectively, and the median number of IgG tests was 2 (IQR, 1-8; Table 2). Among patients with ≥1 IgG test (n = 6773), 35.6% (n = 2411) patients had hypogammaglobulinemia. IgG testing was performed at initial NHL diagnosis in 16.6% of patients, and the median time from diagnosis to first IgG test was 2.8 months (IQR, 0.5-14.6).
IgRT assessment during follow-up
Of 3960 patients with CLL, 6.5% (n = 259) received IgRT (CLL IgRT subcohort; Table 3). The median number of IgRT administrations was 2 (IQR, 1-4), and the median time from last IgG test to IgRT initiation was 0.6 months (IQR, 0.1-3.5). Overall, 56.8% (n = 147) received ≥2 IgRT administrations, and the median time from the first administration of IgRT to subsequent IgG testing was 1.6 months (IQR, 0.7-7.0).
Characteristics . | CLL (n = 259) . | NHL (n = 623) . |
---|---|---|
Year of first IgRT, n (%) | ||
2010-2015 | 14 (12.5) | 18 (6.1) |
2016-2020 | 59 (52.7) | 166 (56.7) |
2021-2023 | 39 (34.8) | 109 (37.2) |
Number of IgRT | ||
Mean ± SD | 5.5 ± 8.8 | 4.9 ± 10.0 |
Median (IQR) | 2.0 (1.0-4.0) | 2.0 (1.0-4.0) |
Number of IgRT, n (%) | (1, 47) | (1, 101) |
1 | 112 (43.2) | 293 (47.0) |
2 | 40 (15.4) | 112 (18.0) |
3 | 25 (9.7) | 47 (7.5) |
4 | 21 (8.1) | 40 (6.4) |
5 | 8 (3.1) | 16 (2.6) |
6 | 8 (3.1) | 17 (2.7) |
7 | 1 (0.4) | 13 (2.1) |
≥8 | 44 (17.0) | 85 (13.6) |
Patients with at least 1 IgG test, n (%) | 244 (94.2) | 584 (93.7) |
Serum IgG level closest to IgRT initiation (mg/dL)∗, n (%) | ||
<200 | 39 (16.0) | 93 (15.9) |
≥200 to <400 | 80 (32.8) | 238 (40.8) |
≥400 to <500 | 29 (11.9) | 54 (9.2) |
≥500 to <600 | 34 (13.9) | 37 (6.3) |
≥600 | 62 (25.4) | 162 (27.7) |
Patients with at least 1 IgG test before IgRT initiation, n (%) | 215 (83.0) | 516 (82.8) |
Time from last IgG test to IgRT initiation†, mo | ||
Mean ± SD | 6.5 ± 16.3 | 2.5 ± 9.9 |
Median (IQR) | 0.6 (0.1-3.5) | 0.2 (0.1-1.0) |
Serum IgG level, mg/dL | ||
Mean ± SD | 462.0 ± 359.0 | 455.2 ± 415.0 |
Median (IQR) | 369.0 (255.0-545.0) | 356.5 (248.0-502.5) |
Patients with ≥2 IgRT, n (%) | 147 (56.8) | 330 (53.0) |
Time from first IgRT to subsequent IgG testing, mo | ||
Mean ± SD | 6.3 ± 10.6 | 3.7 ± 7.8 |
Median (IQR) | 1.6 (0.7-7.0) | 1.0 (0.3-3.0) |
Characteristics . | CLL (n = 259) . | NHL (n = 623) . |
---|---|---|
Year of first IgRT, n (%) | ||
2010-2015 | 14 (12.5) | 18 (6.1) |
2016-2020 | 59 (52.7) | 166 (56.7) |
2021-2023 | 39 (34.8) | 109 (37.2) |
Number of IgRT | ||
Mean ± SD | 5.5 ± 8.8 | 4.9 ± 10.0 |
Median (IQR) | 2.0 (1.0-4.0) | 2.0 (1.0-4.0) |
Number of IgRT, n (%) | (1, 47) | (1, 101) |
1 | 112 (43.2) | 293 (47.0) |
2 | 40 (15.4) | 112 (18.0) |
3 | 25 (9.7) | 47 (7.5) |
4 | 21 (8.1) | 40 (6.4) |
5 | 8 (3.1) | 16 (2.6) |
6 | 8 (3.1) | 17 (2.7) |
7 | 1 (0.4) | 13 (2.1) |
≥8 | 44 (17.0) | 85 (13.6) |
Patients with at least 1 IgG test, n (%) | 244 (94.2) | 584 (93.7) |
Serum IgG level closest to IgRT initiation (mg/dL)∗, n (%) | ||
<200 | 39 (16.0) | 93 (15.9) |
≥200 to <400 | 80 (32.8) | 238 (40.8) |
≥400 to <500 | 29 (11.9) | 54 (9.2) |
≥500 to <600 | 34 (13.9) | 37 (6.3) |
≥600 | 62 (25.4) | 162 (27.7) |
Patients with at least 1 IgG test before IgRT initiation, n (%) | 215 (83.0) | 516 (82.8) |
Time from last IgG test to IgRT initiation†, mo | ||
Mean ± SD | 6.5 ± 16.3 | 2.5 ± 9.9 |
Median (IQR) | 0.6 (0.1-3.5) | 0.2 (0.1-1.0) |
Serum IgG level, mg/dL | ||
Mean ± SD | 462.0 ± 359.0 | 455.2 ± 415.0 |
Median (IQR) | 369.0 (255.0-545.0) | 356.5 (248.0-502.5) |
Patients with ≥2 IgRT, n (%) | 147 (56.8) | 330 (53.0) |
Time from first IgRT to subsequent IgG testing, mo | ||
Mean ± SD | 6.3 ± 10.6 | 3.7 ± 7.8 |
Median (IQR) | 1.6 (0.7-7.0) | 1.0 (0.3-3.0) |
SD, standard deviation.
The closest IgG test to IgRT initiation was used, either before or after IgRT initiation.
The last IgG test before IgRT imitation was used.
Of 13 232 patients with NHL, 4.7% (n = 623) received IgRT (NHL IgRT subcohort; Table 3). The median number of IgRT administrations was 2 (IQR, 1-4), and the median time from last IgG test to IgRT initiation was 0.2 months (IQR, 0.1-1.0). Overall, 53.0% (n = 330) received ≥2 IgRT administrations, and the median time from first administration of IgRT to subsequent IgG testing was 1 month (IQR, 0.3-3.0).
Of patients with hypogammaglobulinemia and 2 infections, 26% (139/535) and 23.4% (361/1545) received IgRT at any time in the CLL and NHL cohorts, respectively. In addition, among patients identified with hypogammaglobulinemia and either ≥2 infections within 6 months or 1 severe infection, 27.2% (147/541) and 23.4% (392/1675) received IgRT at any time in the CLL and NHL cohort, respectively.
IgG levels and hypogammaglobulinemia before and after IgRT initiation
In the IgRT subcohorts, IgG levels and hypogammaglobulinemia was assessed in the 3, 6 and 12 months before vs after IgRT initiation. The median IgG levels were significantly higher corresponding to a lower proportion of patients with hypogammaglobulinemia in the 3 months after vs 3 months before the index date (Table 4). In the CLL IgRT subcohort, the median IgG level at 3, 6, and 12 months after index was 494, 491, and 499 mg/dL compared with preindex levels of 342.5, 355, and 359 mg/dL, respectively (P < .0001). Accordingly, the prevalence of hypogammaglobulinemia was lower at all 3 follow-up time points (3 months: 65.6% vs 85.4%, P < .0001; 6 months: 62.7% vs 84.7%, P < .0001; and 12 months: 64.0% vs 81.6%, P < .0001). Similarly, in the NHL IgRT subcohort, the median IgG level at 3, 6, and 12 months after index was 521, 507, and 507 mg/dL, respectively, compared with the preindex levels of 337.5, 346.5, and 352 mg/dL, respectively (P < .0001). The prevalence of hypogammaglobulinemia was also lower after IgRT initiation (3 months: 61.9% vs 85.4%, P < .0001; 6 months: 68.0% vs 85.2%, P < .0001; and 12 months: 70.7% vs 84.0%, P < .0001).
. | CLL . | P value . | NHL . | P value . |
---|---|---|---|---|
3 mo | (N = 96) | (N = 294) | ||
IgG level before IgRT initiation,∗ mg/dL | ||||
Mean ± SD | 361.6 ± 226.8 | - | 394.3 ± 325.0 | - |
Median (IQR) | 342.5 (214.0-409.5) | 337.5 (234.0-405.0) | ||
IgG level within 3 mo IgRT initiation,† mg/dL | ||||
Mean ± SD | 506.1 ± 188.9 | <.0001‡ | 634.6 ± 518.3 | <.0001‡ |
Median (IQR) | 494.0 (385.0-586.0) | 521.0 (412.0-665.0) | ||
Proportion of patients with hypogammaglobulinemia 3 mo before IgRT initiation§ | 82 (85.4) | <.0001‡ | 251 (85.4) | <.0001‡ |
Proportion of patients with hypogammaglobulinemia 3 mo after IgRT initiation§ | 63 (65.6) | 182 (69.1) | ||
6 mo | (N = 118) | (N = 344) | ||
IgG level before IgRT initiation,∗ mg/dL | ||||
Mean ± SD | 376.1 ± 230.2 | - | 407.5 ± 378.9 | - |
Median (IQR) | 355.0 (220.0-427.0) | 346.5 (235.5-418.5) | ||
IgG level within 6 mo IgRT initiation,† mg/dL | ||||
Mean ± SD | 511.8 ± 192.1 | <.0001‡ | 614.1 ± 494.7 | <.0001‡ |
Median (IQR) | 491.0 (382.0-600.0) | 507.0 (393.0-651.0) | ||
Proportion of patients with hypogammaglobulinemia 6 mo before IgRT initiation§ | 100 (84.7) | <.0001‡ | 293 (85.2) | <.0001‡ |
Proportion of patients with hypogammaglobulinemia 6 mo after IgRT initiation§ | 74 (62.7) | 234 (68.0) | ||
12 mo | (N = 136) | (N = 368) | ||
IgG level before IgRT initiation,∗ mg/dL | ||||
Mean ± SD | 408.6 ± 297.1 | - | 424.4 ± 389.1 | - |
Median (IQR) | 359.0 (227.5-456.5) | 352.0 (245.0-446.5) | ||
IgG level within 12 mo IgRT initiation,† mg/dL | ||||
Mean ± SD | 533.8 ± 245.1 | <.0001‡ | 619.7 ± 510.3 | <.0001‡ |
Median (IQR) | 499.0 (378.0-632.0) | 507.0 (388.0-670.0) | ||
Proportion of patients with hypogammaglobulinemia 12 mo before IgRT initiation§ | 111 (81.6) | <.0001‡ | 309 (84.0) | <.0001‡ |
Proportion of patients with hypogammaglobulinemia 12 mo after IgRT initiation§ | 87 (64.0) | 260 (70.7) |
. | CLL . | P value . | NHL . | P value . |
---|---|---|---|---|
3 mo | (N = 96) | (N = 294) | ||
IgG level before IgRT initiation,∗ mg/dL | ||||
Mean ± SD | 361.6 ± 226.8 | - | 394.3 ± 325.0 | - |
Median (IQR) | 342.5 (214.0-409.5) | 337.5 (234.0-405.0) | ||
IgG level within 3 mo IgRT initiation,† mg/dL | ||||
Mean ± SD | 506.1 ± 188.9 | <.0001‡ | 634.6 ± 518.3 | <.0001‡ |
Median (IQR) | 494.0 (385.0-586.0) | 521.0 (412.0-665.0) | ||
Proportion of patients with hypogammaglobulinemia 3 mo before IgRT initiation§ | 82 (85.4) | <.0001‡ | 251 (85.4) | <.0001‡ |
Proportion of patients with hypogammaglobulinemia 3 mo after IgRT initiation§ | 63 (65.6) | 182 (69.1) | ||
6 mo | (N = 118) | (N = 344) | ||
IgG level before IgRT initiation,∗ mg/dL | ||||
Mean ± SD | 376.1 ± 230.2 | - | 407.5 ± 378.9 | - |
Median (IQR) | 355.0 (220.0-427.0) | 346.5 (235.5-418.5) | ||
IgG level within 6 mo IgRT initiation,† mg/dL | ||||
Mean ± SD | 511.8 ± 192.1 | <.0001‡ | 614.1 ± 494.7 | <.0001‡ |
Median (IQR) | 491.0 (382.0-600.0) | 507.0 (393.0-651.0) | ||
Proportion of patients with hypogammaglobulinemia 6 mo before IgRT initiation§ | 100 (84.7) | <.0001‡ | 293 (85.2) | <.0001‡ |
Proportion of patients with hypogammaglobulinemia 6 mo after IgRT initiation§ | 74 (62.7) | 234 (68.0) | ||
12 mo | (N = 136) | (N = 368) | ||
IgG level before IgRT initiation,∗ mg/dL | ||||
Mean ± SD | 408.6 ± 297.1 | - | 424.4 ± 389.1 | - |
Median (IQR) | 359.0 (227.5-456.5) | 352.0 (245.0-446.5) | ||
IgG level within 12 mo IgRT initiation,† mg/dL | ||||
Mean ± SD | 533.8 ± 245.1 | <.0001‡ | 619.7 ± 510.3 | <.0001‡ |
Median (IQR) | 499.0 (378.0-632.0) | 507.0 (388.0-670.0) | ||
Proportion of patients with hypogammaglobulinemia 12 mo before IgRT initiation§ | 111 (81.6) | <.0001‡ | 309 (84.0) | <.0001‡ |
Proportion of patients with hypogammaglobulinemia 12 mo after IgRT initiation§ | 87 (64.0) | 260 (70.7) |
SD, standard deviation.
Serum IgG laboratory test result closest to the index date was used. If multiple tests were taken on the same day, then the average test result was calculated.
Serum IgG laboratory test results were summarized based on average level. Tests taken on the date of IgRT initiation (ie, index date) were not included.
P < .05 were considered to be significant. P values were calculated by McNemar test comparing proportion of patients with hypogammaglobulinemia, and by Wilcoxon signed-rank test comparing IgG level after IgRT initiation vs before IgRT initiation.
Hypogammaglobulinemia was defined as serum IgG level of <500 mg/dL.
Infections and associated antimicrobial use before and after IgRT
In total, 196 patients with CLL and 417 with NHL in the IgRT subcohorts had ≥3 months of follow-up after IgRT initiation. The proportions of patients with infections and infections requiring antimicrobials are shown in supplemental Table 3. In the CLL IgRT subcohort, IgRT initiation was associated with significantly lower odds of infection and severe infection at 3 months after vs 3 months before IgRT initiation (infection: OR, 0.52; 95% CI, 0.38-0.71; P < .0001; severe infection: OR, 0.48; 95% CI, 0.35-0.67; P < .0001; Table 5). In the NHL IgRT subcohort, IgRT initiation was associated with significantly lower odds of infection and severe infection at 3 months after vs 3 months before IgRT initiation (infection: OR, 0.60; 95% CI, 0.49-0.75; P < .0001; severe infection: OR, 0.40; 95% CI, 0.32-0.50; P < .0001). In both subcohorts, the odds of sinopulmonary and skin or soft tissue infections, infection-associated antimicrobial use, and severe infection–associated antimicrobial use were also lower in the 3 months after vs 3 months before the index date. The findings were consistent at the 6- and 12-month follow-up time points. The same trends were observed in the CAR T-cell therapy subgroup of the NHL IgRT subcohort (supplemental Table 4) and in patients with CLL and NHL who received Gammagard Liquid as IgRT (supplemental Table 5). That is, rates of infection (including severe infection) and antimicrobial use associated with infections and severe infections were lower 3 months after, compared with 3 months before, IgRT initiation.
. | CLL . | NHL . | ||||
---|---|---|---|---|---|---|
3 mo (n = 196)∗ . | 6 mo (n = 175)∗ . | 12 mo (n = 137)∗ . | 3 mo (n = 417)† . | 6 mo (n = 343)† . | 12 mo (n = 250)† . | |
OR (95% CI)∗∗ . | OR (95% CI)∗∗ . | OR (95% CI)∗∗ . | OR (95% CI)∗∗ . | OR (95% CI)∗∗ . | OR (95% CI)∗∗ . | |
Infections | ||||||
All infections,‡,§ n (%) | 0.52 (0.38-0.71) a,∗ | 0.66 (0.47-0.92) b,∗ | 0.62 (0.42-0.90) g,∗ | 0.60 (0.49-0.75) a,∗ | 0.70 (0.55-0.90) l,∗ | 0.86 (0.63-1.18) m |
Individual infection, n (%) | ||||||
Sinopulmonary and skin or soft tissue infections|| | 0.41 (0.29-0.59) a,∗ | 0.61 (0.44-0.87) c,∗ | 0.68 (0.46-1.00) h | 0.55 (0.44-0.69) a,∗ | 0.71 (0.56-0.90) j,∗ | 0.95 (0.71-1.28) n |
Pneumonia | 0.51 (0.38-0.70) | 0.73 (0.54-0.99) | 0.82 (0.55-1.20) | 0.54 (0.43-0.66) | 0.65 (0.52-0.82) | 0.75 (0.57-0.99) |
Ear-nose-throat infection | 0.57 (0.34-0.96) | 0.71 (0.44-1.15) | 0.56 (0.33-0.98) | 0.63 (0.41-0.96) | 0.86 (0.57-1.31) | 1.06 (0.69-1.63) |
Bronchitis | 0.44 (0.18-1.09) | 0.42 (0.19-0.91) | 0.73 (0.35-1.52) | 0.77 (0.44-1.33) | 0.97 (0.61-1.53) | 1.15 (0.72-1.82) |
Skin/soft tissue/joint/bone infection | 0.55 (0.23-1.33) | 0.76 (0.33-1.75) | 0.77 (0.36-1.63) | 0.71 (0.45-1.13) | 0.82 (0.48-1.38) | 0.91 (0.55-1.49) |
Other infections, n (%) | 0.49 (0.34-0.71) a,∗ | 0.55 (0.38-0.79) d,∗ | 0.53 (0.36-0.80) i,∗ | 0.45 (0.36-0.57) a,∗ | 0.52 (0.41-0.67) a,∗ | 0.67 (0.50-0.90) o,∗ |
Severe infections, n (%) | ||||||
All severe infections,‡,§,¶ n (%) | 0.48 (0.35-0.67) a,∗ | 0.52 (0.37-0.74) a,∗ | 0.46 (0.30-0.70) a,∗ | 0.40 (0.32-0.50) a,∗ | 0.42 (0.33-0.54) a,∗ | 0.44 (0.33-0.58) |
Individual infection, n (%) | ||||||
Sinopulmonary and skin or soft tissue infections|| | 0.38 (0.26-0.57) a,∗ | 0.46 (0.31-0.68) a,∗ | 0.48 (0.29-0.80) j,∗ | 0.38 (0.29-0.49) a,∗ | 0.40 (0.30-0.53) a,∗ | 0.50 (0.36-0.69) |
Pneumonia | 0.40 (0.27-0.59) | 0.50 (0.34-0.74) | 0.57 (0.34-0.95) | 0.37 (0.28-0.47) | 0.37 (0.28-0.50) | 0.44 (0.31-0.63) |
Ear-nose-throat infection | 0.87 (0.35-2.15) | 1.13 (0.47-2.72) | 0.87 (0.35-2.18) | 0.57 (0.28-1.16) | 0.62 (0.30-1.28) | 0.61 (0.27-1.35) |
Bronchitis | 0.49 (0.12-2.03) | 0.59 (0.14-2.56) | 0.66 (0.11-4.09) | 0.52 (0.23-1.22) | 0.61 (0.30-1.26) | 0.74 (0.31-1.75) |
Skin/soft tissue/joint/bone infection | 0.66 (0.24-1.78) | 0.61 (0.21-1.78) | 0.56 (0.18-1.77) | 0.67 (0.37-1.24) | 0.85 (0.42-1.72) | 0.66 (0.29-1.51) |
Other infections, n (%) | 0.33 (0.21-0.52) a,∗ | 0.40 (0.26-0.63) a,∗ | 0.45 (0.28-0.71) d,∗ | 0.28 (0.22-0.37) a,∗ | 0.33 (0.25-0.44) a,∗ | 0.36 (0.26-0.50) |
Antimicrobial use associated with infections,# n (%) | ||||||
Infection episodes requiring antibiotic use | 0.38 (0.27-0.54) a,∗ | 0.47 (0.32-0.69) a∗ | 0.52 (0.34-0.79) a,∗ | 0.46 (0.36-0.58) a,∗ | 0.47 (0.37-0.60) a,∗ | 0.61 (0.45-0.82)d,∗ |
Infection episodes requiring antiviral use | 0.36 (0.25-0.52) a,∗ | 0.40 (0.28-0.58) a,∗ | 0.39 (0.25-0.62) a,∗ | 0.46 (0.37-0.57) a,∗ | 0.53 (0.43-0.67) a,∗ | 0.61 (0.47-0.79) a,∗ |
Infection episodes requiring antifungal use | 0.45 (0.30-0.67) a,∗ | 0.57 (0.38-0.86)e,∗ | 0.54 (0.32-0.90)k,∗ | 0.54 (0.43-0.68) a,∗ | 0.53 (0.42-0.69) a,∗ | 0.56 (0.41-0.76) a,∗ |
Antimicrobial use associated with severe infections,# n (%) | ||||||
Severe infection episodes requiring antibiotic use | 0.42 (0.29-0.60) a,∗ | 0.50 (0.34-0.73)f,∗ | 0.46 (0.29-0.73) d,∗ | 0.36 (0.28-0.45) a,∗ | 0.38 (0.29-0.48) a,∗ | 0.42 (0.31-0.56) a,∗ |
Severe infection episodes requiring antiviral use | 0.36 (0.24-0.53) a,∗ | 0.37 (0.25-0.55)a,∗ | 0.29 (0.17-0.48) a,∗ | 0.37 (0.29-0.47) a,∗ | 0.42 (0.33-0.54) a,∗ | 0.46 (0.34-0.61) a,∗ |
Severe infection episodes requiring antifungal use | 0.38 (0.24-0.59) a,∗ | 0.47 (0.30-0.73)d,∗ | 0.44 (0.25-0.77)l,∗ | 0.48 (0.38-0.61) a,∗ | 0.51 (0.39-0.66) a,∗ | 0.55 (0.40-0.76) a,∗ |
. | CLL . | NHL . | ||||
---|---|---|---|---|---|---|
3 mo (n = 196)∗ . | 6 mo (n = 175)∗ . | 12 mo (n = 137)∗ . | 3 mo (n = 417)† . | 6 mo (n = 343)† . | 12 mo (n = 250)† . | |
OR (95% CI)∗∗ . | OR (95% CI)∗∗ . | OR (95% CI)∗∗ . | OR (95% CI)∗∗ . | OR (95% CI)∗∗ . | OR (95% CI)∗∗ . | |
Infections | ||||||
All infections,‡,§ n (%) | 0.52 (0.38-0.71) a,∗ | 0.66 (0.47-0.92) b,∗ | 0.62 (0.42-0.90) g,∗ | 0.60 (0.49-0.75) a,∗ | 0.70 (0.55-0.90) l,∗ | 0.86 (0.63-1.18) m |
Individual infection, n (%) | ||||||
Sinopulmonary and skin or soft tissue infections|| | 0.41 (0.29-0.59) a,∗ | 0.61 (0.44-0.87) c,∗ | 0.68 (0.46-1.00) h | 0.55 (0.44-0.69) a,∗ | 0.71 (0.56-0.90) j,∗ | 0.95 (0.71-1.28) n |
Pneumonia | 0.51 (0.38-0.70) | 0.73 (0.54-0.99) | 0.82 (0.55-1.20) | 0.54 (0.43-0.66) | 0.65 (0.52-0.82) | 0.75 (0.57-0.99) |
Ear-nose-throat infection | 0.57 (0.34-0.96) | 0.71 (0.44-1.15) | 0.56 (0.33-0.98) | 0.63 (0.41-0.96) | 0.86 (0.57-1.31) | 1.06 (0.69-1.63) |
Bronchitis | 0.44 (0.18-1.09) | 0.42 (0.19-0.91) | 0.73 (0.35-1.52) | 0.77 (0.44-1.33) | 0.97 (0.61-1.53) | 1.15 (0.72-1.82) |
Skin/soft tissue/joint/bone infection | 0.55 (0.23-1.33) | 0.76 (0.33-1.75) | 0.77 (0.36-1.63) | 0.71 (0.45-1.13) | 0.82 (0.48-1.38) | 0.91 (0.55-1.49) |
Other infections, n (%) | 0.49 (0.34-0.71) a,∗ | 0.55 (0.38-0.79) d,∗ | 0.53 (0.36-0.80) i,∗ | 0.45 (0.36-0.57) a,∗ | 0.52 (0.41-0.67) a,∗ | 0.67 (0.50-0.90) o,∗ |
Severe infections, n (%) | ||||||
All severe infections,‡,§,¶ n (%) | 0.48 (0.35-0.67) a,∗ | 0.52 (0.37-0.74) a,∗ | 0.46 (0.30-0.70) a,∗ | 0.40 (0.32-0.50) a,∗ | 0.42 (0.33-0.54) a,∗ | 0.44 (0.33-0.58) |
Individual infection, n (%) | ||||||
Sinopulmonary and skin or soft tissue infections|| | 0.38 (0.26-0.57) a,∗ | 0.46 (0.31-0.68) a,∗ | 0.48 (0.29-0.80) j,∗ | 0.38 (0.29-0.49) a,∗ | 0.40 (0.30-0.53) a,∗ | 0.50 (0.36-0.69) |
Pneumonia | 0.40 (0.27-0.59) | 0.50 (0.34-0.74) | 0.57 (0.34-0.95) | 0.37 (0.28-0.47) | 0.37 (0.28-0.50) | 0.44 (0.31-0.63) |
Ear-nose-throat infection | 0.87 (0.35-2.15) | 1.13 (0.47-2.72) | 0.87 (0.35-2.18) | 0.57 (0.28-1.16) | 0.62 (0.30-1.28) | 0.61 (0.27-1.35) |
Bronchitis | 0.49 (0.12-2.03) | 0.59 (0.14-2.56) | 0.66 (0.11-4.09) | 0.52 (0.23-1.22) | 0.61 (0.30-1.26) | 0.74 (0.31-1.75) |
Skin/soft tissue/joint/bone infection | 0.66 (0.24-1.78) | 0.61 (0.21-1.78) | 0.56 (0.18-1.77) | 0.67 (0.37-1.24) | 0.85 (0.42-1.72) | 0.66 (0.29-1.51) |
Other infections, n (%) | 0.33 (0.21-0.52) a,∗ | 0.40 (0.26-0.63) a,∗ | 0.45 (0.28-0.71) d,∗ | 0.28 (0.22-0.37) a,∗ | 0.33 (0.25-0.44) a,∗ | 0.36 (0.26-0.50) |
Antimicrobial use associated with infections,# n (%) | ||||||
Infection episodes requiring antibiotic use | 0.38 (0.27-0.54) a,∗ | 0.47 (0.32-0.69) a∗ | 0.52 (0.34-0.79) a,∗ | 0.46 (0.36-0.58) a,∗ | 0.47 (0.37-0.60) a,∗ | 0.61 (0.45-0.82)d,∗ |
Infection episodes requiring antiviral use | 0.36 (0.25-0.52) a,∗ | 0.40 (0.28-0.58) a,∗ | 0.39 (0.25-0.62) a,∗ | 0.46 (0.37-0.57) a,∗ | 0.53 (0.43-0.67) a,∗ | 0.61 (0.47-0.79) a,∗ |
Infection episodes requiring antifungal use | 0.45 (0.30-0.67) a,∗ | 0.57 (0.38-0.86)e,∗ | 0.54 (0.32-0.90)k,∗ | 0.54 (0.43-0.68) a,∗ | 0.53 (0.42-0.69) a,∗ | 0.56 (0.41-0.76) a,∗ |
Antimicrobial use associated with severe infections,# n (%) | ||||||
Severe infection episodes requiring antibiotic use | 0.42 (0.29-0.60) a,∗ | 0.50 (0.34-0.73)f,∗ | 0.46 (0.29-0.73) d,∗ | 0.36 (0.28-0.45) a,∗ | 0.38 (0.29-0.48) a,∗ | 0.42 (0.31-0.56) a,∗ |
Severe infection episodes requiring antiviral use | 0.36 (0.24-0.53) a,∗ | 0.37 (0.25-0.55)a,∗ | 0.29 (0.17-0.48) a,∗ | 0.37 (0.29-0.47) a,∗ | 0.42 (0.33-0.54) a,∗ | 0.46 (0.34-0.61) a,∗ |
Severe infection episodes requiring antifungal use | 0.38 (0.24-0.59) a,∗ | 0.47 (0.30-0.73)d,∗ | 0.44 (0.25-0.77)l,∗ | 0.48 (0.38-0.61) a,∗ | 0.51 (0.39-0.66) a,∗ | 0.55 (0.40-0.76) a,∗ |
aP < .0001; bP = .015; cP = .006; dP = .001; eP = .007; fP = .000; gP = .012; hP = .051; iP = .002; jP = .005; kP = .018; lP = .004; mP = .345; nP = .750; oP = .007.
GEE, generalized estimating equations.
196, 175, and 137 patients had at least 3-month, 6-month, 12-month follow-up time, respectively, before and after IgRT initiation in the CLL cohort.
417, 343, and 250 patients had at least 3-month, 6-month, 12-month follow-up time, respectively, before and after IgRT initiation in the NHL cohort.
Medical encounters with the same infection diagnosis within 7 days of each other will be attributed to the same infection episode.
All infections include the individual infection that listed below (ie, pneumonia, pyrexia unknown origin, ear-nose-throat infections, bronchitis, skin/soft tissue/joint/bone infections, sepsis, bacteremia, genitourinary infections, gastrointestinal infections, brain/spinal infections, COVID-19, blood infections other than sepsis, cardiac infections, and other [not otherwise specified]).
Sinopulmonary and skin or soft tissue infections include pneumonia, ear-nose-throat infections, bronchitis, and skin/soft tissue/joint/bone infections.
A severe infection was defined as an infection leading to a hospitalization or a treatment with any intravenous antibiotic, antiviral, or antifungal medication.
Patients with infections requiring antimicrobials was defined as medical encounters with a diagnosis of infection and a prescription of antibiotics/antiviral/antifungal for the same diagnosis within 30 days.
GEE logistic regression model with robust variance was fit to obtain OR, 95% CI, and P value for comparing infection rate in the post-IgRT period to that in the pre-IgRT period. P < .05 were considered to be significant and denoted with “∗.”
Factors associated with severe infection
In the multivariable logistic regression model, increased IgG testing was associated with a significantly lower likelihood of severe infection after adjusting for key variables, with the odds decreasing as the number of IgG tests increased. For CLL, OR for severe infections was 0.36 (95% CI, 0.29-0.44) with 1 preceding test, 0.23 (95% CI, 0.17-0.31) with 2 preceding tests, and 0.08 (95% CI, 0.06-0.10) with ≥3 preceding tests (all P < .0001; Table 6). For NHL, ORs were 0.44 (95% CI, 0.39-0.50), 0.38 (95% CI, 0.31-0.46), and 0.14 (95% CI, 0.12-0.16), respectively (all P < .0001; Table 7). Results remained consistent after including hypogammaglobulinemia as an additional covariate (supplemental Tables 6 and 7).
Variable . | OR . | 95% CI . | P value . |
---|---|---|---|
IgG testing∗ | |||
No IgG testing | Reference | ||
1 | 0.36 | (0.29-0.44) | <.0001† |
2 | 0.23 | (0.17-0.31) | <.0001† |
≥3 | 0.08 | (0.06-0.10) | <.0001† |
IgRT use∗ | 0.52 | (0.28-0.97) | .038† |
Chemotherapy‡,§ | 2.39 | (1.85-3.09) | <.0001† |
mAb‡ | 1.27 | (0.99-1.61) | .058 |
Targeted therapy‡ | 2.49 | (2.03-3.06) | <.0001† |
Immunomodulatory drug‡ | 3.76 | (1.82-7.80) | .0004† |
Corticosteroid‡ | 5.46 | (4.42-6.75) | <.0001† |
HSCT‡ | 4.12 | (2.73-6.22) | <.0001† |
CAR-T cell therapy‡ | 3.30 | (1.66-6.57) | .001† |
Age | 1.02 | (1.08-1.02) | .0001† |
Sex | |||
Male | Reference | ||
Female | 0.84 | (0.71-1.00) | .054† |
Variable . | OR . | 95% CI . | P value . |
---|---|---|---|
IgG testing∗ | |||
No IgG testing | Reference | ||
1 | 0.36 | (0.29-0.44) | <.0001† |
2 | 0.23 | (0.17-0.31) | <.0001† |
≥3 | 0.08 | (0.06-0.10) | <.0001† |
IgRT use∗ | 0.52 | (0.28-0.97) | .038† |
Chemotherapy‡,§ | 2.39 | (1.85-3.09) | <.0001† |
mAb‡ | 1.27 | (0.99-1.61) | .058 |
Targeted therapy‡ | 2.49 | (2.03-3.06) | <.0001† |
Immunomodulatory drug‡ | 3.76 | (1.82-7.80) | .0004† |
Corticosteroid‡ | 5.46 | (4.42-6.75) | <.0001† |
HSCT‡ | 4.12 | (2.73-6.22) | <.0001† |
CAR-T cell therapy‡ | 3.30 | (1.66-6.57) | .001† |
Age | 1.02 | (1.08-1.02) | .0001† |
Sex | |||
Male | Reference | ||
Female | 0.84 | (0.71-1.00) | .054† |
Patients diagnosed with CLLL: n = 3960.
HSCT, hematopoietic stem cell transplantation.
For patients having an infection, IgG testing and IgRT use were measured in the 12 months before the date of infection.
P < .05 were considered to be significant.
Assessed at any time point because dates are not available in the clinical notes.
In the univariate logistic regression model, patients receiving chemotherapy had significantly higher likelihood of hypogammaglobulinemia (OR, 3.98; 95% CI, 3.37-4.70; P < .0001) compared to patients who never received chemotherapy.
Variable . | OR . | 95% CI . | P value . |
---|---|---|---|
IgG testing∗ | |||
No IgG testing | Reference | ||
1 | 0.44 | (0.39-0.50) | <.0001† |
2 | 0.38 | (0.31-0.46) | <.0001† |
≥3 | 0.14 | (0.12-0.16) | <.0001† |
IgRT use∗ | 0.62 | (0.39-0.992) | .046† |
Rituximab-based regimens‡ | 2.61 | (2.39-2.85) | <.0001† |
Targeted therapy‡ | 3.35 | (2.88-3.89) | <.0001† |
Neutropenia§ | 8.0 | (7.11-8.90) | <.0001† |
Diabetes§ | 2.42 | (2.16-2.70) | <.0001† |
Sinusitis§ | 1.74 | (1.49-2.04) | <.0001† |
Bronchitis§ | 2.69 | (2.23-3.25) | <.0001† |
Age | <.0001† | ||
18-64 | Reference | ||
≥65 | 1.26 | (1.16-1.37) | <.0001† |
Sex | |||
Male | Reference | ||
Female | 0.87 | (0.80-0.95) | .002† |
Variable . | OR . | 95% CI . | P value . |
---|---|---|---|
IgG testing∗ | |||
No IgG testing | Reference | ||
1 | 0.44 | (0.39-0.50) | <.0001† |
2 | 0.38 | (0.31-0.46) | <.0001† |
≥3 | 0.14 | (0.12-0.16) | <.0001† |
IgRT use∗ | 0.62 | (0.39-0.992) | .046† |
Rituximab-based regimens‡ | 2.61 | (2.39-2.85) | <.0001† |
Targeted therapy‡ | 3.35 | (2.88-3.89) | <.0001† |
Neutropenia§ | 8.0 | (7.11-8.90) | <.0001† |
Diabetes§ | 2.42 | (2.16-2.70) | <.0001† |
Sinusitis§ | 1.74 | (1.49-2.04) | <.0001† |
Bronchitis§ | 2.69 | (2.23-3.25) | <.0001† |
Age | <.0001† | ||
18-64 | Reference | ||
≥65 | 1.26 | (1.16-1.37) | <.0001† |
Sex | |||
Male | Reference | ||
Female | 0.87 | (0.80-0.95) | .002† |
Patients diagnosed with NHL: n = 13 232.
For patients having an infection, IgG testing and IgRT use were measured in the 12 months before the date of infection.
P < .05 were considered to be significant.
Assessed at any time point because dates are not available in the clinical notes.
Clinical conditions were assessed at any time point.
An exploratory analysis was conducted to assess the relationship between the frequency of IgG testing, the proportions of patients with hypogammaglobulinemia, and the proportion of patients with hypogammaglobulinemia who receive IgRT (supplemental Table 8). Compared with 1 or 2 tests, having ≥3 IgG tests was associated with a higher likelihood of hypogammaglobulinemia in the CLL cohort (52.5% [693/1321] vs 16.8% [224/1331], P < .0001); and among patients with CLL with hypogammaglobulinemia and among those with hypogammaglobulinemia (n = 917), having ≥3 IgG tests was associated with a higher likelihood of receiving IgRT (26.1% [181/693] vs 7.6% [17/244], P < .0001). Similarly, in the NHL cohort, having ≥3 IgG tests was associated with a higher likelihood of hypogammaglobulinemia in the NHL cohort (55% [1769/3264] vs 17.5% [615/3509], P < .0001) and among patients with NHL with hypogammaglobulinemia (n = 2411) a higher likelihood of receiving IgRT (23.6% [423/1796] vs 8.1% [50/615], P < .0001).
Discussion
This retrospective, observational cohort study evaluated the real-world rates of IgG testing and effectiveness of IgRT in a large cohort of patients (>17 000) with CLL or NHL in the United States. We observed significant variability with ∼67% and ∼51.2% receiving IgG testing in the CLL and NHL cohorts, respectively. We also observed that among patients with CLL and NHL, IgRT increased IgG levels resulting in a lower proportion of patients with hypogammaglobulinemia, and significant reductions in infections, severe infections, and associated antimicrobial use.
Routine evaluation of IgG levels is a critical step in assessing infection risk for patients with CLL and NHL.8,10,15 Although most guidelines recommend routine IgG testing for patients with CLL and NHL, real-world evidence on the patterns of IgG testing is limited. In 1 study of patients with CLL exposed to an anti-CD20 mAb, only 33% had IgG testing within 6 months of starting treatment.16 In another study, physicians caring for patients with hematological malignancies reported IgG testing in ∼75% of patients.17 Although this study has important limitations (eg, potential for selection and reporting bias), at a minimum, these data illustrate a disconnect between guideline recommendations and real-world practice. Our results revealed variability in practice regarding IgG testing among hematologists caring for patients with CLL and NHL. In both cohorts, the number of IgG tests per patient varied widely, including a large proportion of patients without IgG testing throughout the study period. These findings underscore the urgency to establish consensus regarding best practices for IgG testing in patients with CLL and NHL to minimize the morbidity and mortality associated with infections. Given the evolving treatment landscape for CLL and NHL, it is crucial to formulate guidelines for routine IgG testing in individuals undergoing these treatments.
Increased IgG testing was associated with a significantly lower risk of severe infections among patients with CLL or NHL, with an incrementally lower risk of severe infection associated with increasing number of IgG preceding tests performed. Within each cohort, those with ≥3 IgG tests had an increased detection of hypogammaglobulinemia and were also more likely to receive IgRT. These findings suggest the hypothesis that increased IgG testing improves hypogammaglobulinemia detection and IgRT use. Put simply, patients known to have hypogammaglobulinemia might be more aware that they have a modifiable risk factor for infections, and our data suggest the hypothesis that these patients are more likely to communicate recurrent minor infections to their hematologists, leading to initiation of IgRT before development of severe infections. These data suggest that increased IgG testing would assist clinicians in identifying patients who are likely to benefit from IgRT and highlights the need to develop and disseminate standardized protocols for IgG testing and educate health care providers on the best practices for managing patients with hypogammaglobulinemia with/without infections.
Only 6.5% of patients with CLL and 4.7% of patients with NHL received IgRT in this real-world analysis, which hints at a more conservative approach to IgRT use in the United States. This may be driven by our reactive approach to hypogammaglobulinemia, because most clinicians only initiate IgRT in patients who experienced severe or recurrent infections. There are risks associated with overuse of IgRT, that is, many patients with hypogammaglobulinemia do not experience recurrent or severe infections and therefore physicians do not advocate for starting IgRT in all patients with hypogammaglobulinemia. However, our data suggest that a gap exists between the low proportion of patients with hypogammaglobulinemia who receive IgRT and the relatively high proportion of patients experiencing recurrent or severe infections. In our study, only ∼25% of patients with hypogammaglobulinemia and ≥2 infections received IgRT, and only ∼25% of patients with hypogammaglobulinemia and either 1 severe infection or ≥2 infections within 6 months received IgRT. These findings suggest that among patients with CLL and NHL, many patients do not receive IgRT despite having an indication per national guidelines. This implies the potential benefit of adhering to national guidelines regarding IgRT use for patients with CLL or NHL associated with hypogammaglobulinemia would result in improved infection outcomes.
Although IgRT has been shown to reduce recurrent and severe infections, the available data supporting its use in CLL and NHL are limited. In a US-based administrative claims database study, the authors demonstrated a 49% lower odds of severe bacterial infections (OR, 0.51; P = .0036) in patients with diagnosis of NHL and SID who received IgRT.18 In contrast, in the same study, there was no significant association between severe bacterial infections and receiving IgRT (OR, 0.93; P = .8243) in patients with CLL and SID. In another retrospective, observational study, patients with CLL or NHL, IgRT use was associated with a 59% reduction in severe infections (OR, 0.41; 95% CI, 0.24-0.270; P = .001), although separate estimates for the CLL and NHL cohorts were not reported.19 Our study supports use of IgRT to reduce hypogammaglobulinemia and prevent infections and severe infections in appropriately selected patients with CLL or NHL. We also demonstrated that IgRT led to significantly less antimicrobial use associated with infections and severe infections. This finding has potential to affect the risk of developing antimicrobial resistance.
Unexpectedly, the number of IgRT administrations per patient varied widely (median of 2 per patient, with ∼15% receiving ≥8 IgRT administrations). With current guidelines calling for continued IgRT in patients with hypogammaglobulinemia and recurrent/severe infections, these data suggest that a gap exists between current recommendations and real-world practice, raising important questions regarding IgRT use. For some patients, a limited course of IgRT is likely suboptimal, and these patients might benefit from continued therapy. However, sustained improvement in infection outcomes was observed in a cohort of patients with a median of 2 IgRT administrations, suggesting that a limited course of IgRT might be sufficient in some patients. More robust tools to identify which patients benefit from continuous vs limited course of IgRT are needed. There are additional variations in the clinic, for example, seasonal administration of IgRT for select patients. However, current guidelines recommend continued IgRT use in patients with hypogammaglobulinemia and recurrent or severe infections. Overall, these data highlight the need to disseminate harmonized guideline recommendations on IgG testing and IgRT use to mitigate infection risk.20,21
Although IgRT use was associated with an increase in IgG levels and less hypogammaglobulinemia in this study at 3, 6, and 12 months follow-up, it is crucial to emphasize that IgRT should be used to sustain the biologic IgG level specific to each individual patient, aiming to enhance clinical outcomes rather than adopting a generic "one-size-fits-all" approach of merely elevating IgG levels across all patients.22,23 This suggests the need to optimize IgRT more effectively in patients with hypogammaglobulinemia, highlighting the need for clearer guidance in tailoring treatment to the unique needs of each patient. Whether hematologists should engage immunologists to perform a more comprehensive assessment, for example, with vaccine testing, to assess distinguish between quantitative and qualitative IgG deficiencies and provide a more nuanced understanding of the patient’s immunological status, is unknown.24,25
Although anti-CD19 CAR T-cell therapies have improved survival outcomes for patients with select NHL subtypes who develop relapsed or refractory disease,26 CAR T cells result in B-cell aplasia leading to reduced antibody production, and these effects can be long-lasting and contribute to infection risk. In addition, the rates of hypogammaglobulinemia vary significantly among different CAR T-cell products, with reported occurrences ranging from 46% to 62% of patients.27,28 In our study, IgRT use was associated with significant reductions in infections and severe infections among patients diagnosed with NHL who are treated with CAR T-cell therapy. Notably, in the United States, there exists a broad spectrum of strategies for addressing hypogammaglobulinemia in patients treated with CAR T-cell therapies. This ranges from a universal application of prescribing IgRT for hypogammaglobulinemia to a more conservative “watch-and-wait” approach, wherein treatment is administered selectively to patients who develop infections. Significantly more research is necessary to elucidate the most effective approach to managing hypogammaglobulinemia in the context of CAR T-cell therapy.29
This study has several limitations. Patients in the MGB RPDR are more likely to be White and therefore may not be fully representative of the US population. There is potential for confounding by indication or unmeasured confounders in this retrospective observational analysis of IgRT administration, given that IgRT is usually initiated after infections. Although our primary analysis focused on the 3 months before and after IgRT initiation, and thus might be more prone to this bias, the apparent effect on infection outcomes were sustained at 6 and 12 months. These data are further limited in that data are restricted to the MGB network. However, although the MGB network might not be representative of the entire US population, and it is possible that IgG testing and IgRT use are different in other settings, MGB comprises multiple academic medical centers, community hospitals, and numerous outpatient practices. Furthermore, it is important to consider that some patients may not receive IgRT, potentially because of historical shortages of IgRT or lack of insurance coverage. Insurance coverage for IgRT can vary significantly across different plans, and there may be instances in which patients with IgG levels of <500 mg/dL are unable to obtain coverage for IgRT. Additional important limitations include the potential for missing data from patient care received outside of the MGB network, or because identification of variables is based on ICD-9/ICD-10 diagnosis/procedure codes, both of which might result in underestimates or misclassifications of conditions, outcomes (eg, infections), IgG testing, and IgRT use. As a point of reference, data for infection events managed out of the MGB network may not have been captured. Although treatments were identified using both structured data and clinical notes (unstructured data) to comprehensively capture the treatment characteristics of patients with CLL/NHL, it was challenging to determine dates of treatment receipt from the clinical notes. Finally, the generalizability of the study findings for hypogammaglobulinemia is limited to the cutoff used to define hypogammaglobulinemia (ie, IgG levels of <500 mg/dL), which was determined based on a combination of literature review and clinical insights.
Conclusions
We observed heterogeneity in IgG testing strategies and IgRT use in real-world patients with CLL or NHL in the United States. More frequent IgG testing was associated with a lower likelihood of severe infections, and IgRT administration was associated with lower rates of infection and severe infection and a reduction in antimicrobial use in this population. As the treatment landscape evolves for patients with CLL and NHL, a critical need exists to develop guidelines and consensus on SID monitoring and management.
Acknowledgments
Medical writing assistance was provided by a professional medical writer, Christopher Crotty, an employee of Analysis Group, Inc, a consulting company that provided paid consulting services to Takeda Pharmaceuticals USA, Inc, which funded the development and conduct of this study and preparation of the manuscript.
This study was funded by Takeda Pharmaceuticals USA. J.D.S. is supported by the National Cancer Institute at the National Institutes of Health under award number K08CA270202.
The study sponsor was involved in several aspects of the study, including study design, data interpretation, manuscript writing, and the decision to submit the manuscript for publication.
Authorship
Contribution: R.D., L.H., M.Y., A.B., L.C., M.P., and M.S.D. contributed to study conception and design and data acquisition, assembly, analysis, and interpretation; J.D.S., T.G., M.S., and Z.Y. contributed to study conception and design, and data analysis and interpretation; and all authors made substantial contributions to manuscript drafting and critical revision for important intellectual content, reviewed and approved the final version to be published, and agree to be accountable for all aspects of the work.
Conflict-of-interest disclosure: R.D., L.H., M.Y., A.B., L.C., M.P., and M.S.D. are employees of Analysis Group, Inc, a consulting company that provided paid consulting services to Takeda Pharmaceuticals USA, Inc, the study sponsor. J.D.S. received consulting fees from AbbVie, AstraZeneca, BeiGene, Bristol Myers Squibb, Roche, Seattle Genetics, TG Therapeutics, and Verastem; and received research funding from Adaptive Biotechnologies, BeiGene, BostonGene, Genentech/Roche, GlaxoSmithKline, Moderna, Takeda, and TG Therapeutics. T.G. and M.S. are employees of Takeda Pharmaceuticals USA, Inc; and are Takeda shareholders. Z.Y. is an employee the University of California San Diego and clinical consultant for Takeda Pharmaceuticals USA, Inc. G.B. declares no competing financial interests. The remaining authors declare no competing financial interests.
Correspondence: Jacob D. Soumerai, Massachusetts General Hospital Cancer Center: Hematology Oncology, 55 Fruit St, Boston, MA 02114; email: [email protected].
References
Author notes
M.S. and S.N.M. are joint senior authors.
Presented, in part, in poster form at the Clinical Immunological Society 2023 annual meeting, 18 to 21 May 2023, St. Louis, MO; and the Immunoglobulin National Society 2023 national conference, 1 November to 31 December 2023; Denver, CO.
The data supporting the findings of this study are available from the Massachusetts General Brigham Research Patient Data Registry. Restrictions apply to the availability of these data, which were used under license for this study.
The full-text version of this article contains a data supplement.