Rituximab and hypogammaglobulinemia in the setting of ABO-incompatible kidney transplantation

Background: ABO-incompatible (ABOi) kidney transplantation can be achieved by desensitizing the recipient using apheresis plus rituximab-based immunosuppression. Objectives: We sought to ascertain the factors that contributed to low immunoglobulin levels at post-ABOi kidney transplantation. Patients and Methods: This single-center study included 43 ABO-i kidney-transplant recipients desensitized with rituximab-based therapy. Posttransplant immunoglobulin levels (IgG, IgA, and IgM) were prospectively monitored within 2 years. If severe hypogammaglobulinemia occurred, i.e., IgG levels <4 g/L, patients received polyvalent immunoglobulin (IVIg substitution). Results: Within 1-year posttransplantation, 25% of patients experienced at least once severe hypogammaglobulinemia. On D –30 (pre-transplantation), IgG, IgA, and IgM levels were within normal ranges: 10 ± 4.4, 1.9 ± 1.2, and 0.8± 0.5 g/L, respectively. IgG levels were significantly decreased at D0 (4.2 ± 3.8 g/L) compared to D–30. At D15, IgG levels did not significantly differ from those on D0 or D –30. Conversely, beyond month-1 posttransplant IgG levels were within normal ranges and were significantly higher than levels measured on D0. Within three months posttransplantation, 11 patients required IVIg because IgG levels were <4 g/L (IVIg+ group). When these patients were compared with those that did not receive IVIg within 3 months posttransplantation (IVIg– group), IgG levels were similar at D –30 in both groups. Conversely, at D0, IgG levels were significantly lower in the Ig+ group (2.4 ± 2 vs. 5.5± 4.2 g/L; P = 0.009); t he d ifference remained significant until D15 posttransplantation (Ig+: 3.4 ± 1.7, Ig–: 6.6 ± 2 g/L; P = 0.0002). There was no statistical difference between the two groups after D15. Infectious complications did not significantly vary between patients with or without hypogammaglobulinemia. Conclusions: We conclude that hypogammaglobulinemia occurred frequently after ABOincompatible kidney transplantation but did not cause more infectious complications.


Background
Many patients worldwide develop end-stage renal disease (ESRD). The best possible treatment option for ESRD is kidney transplantation (1,2). However, due to the shortage of deceased kidney donors, live-kidney transplantation is now common (3). In this setting, ABO-incompatible (ABOi) kidney transplantation can occur where the potential recipient has isoagglutinins directed against the potential donor. Transplantation can be done provided the recipient undergoes pretransplant desensitization, which includes rituximab therapy (that has now replaced splenectomy) and conventional immunosuppression (tacrolimus, steroids, mycophenolic acid) (4,5).
Rituximab causes rapid depletion of CD20-expressing B-cell precursors and mature B-cells, which then remain at very low or undetectable levels for 6--9 months before slowly returning to pretreatment levels (10). Rituximab therapy has also been associated with late-onset neutropenia (11) and hypogammaglobulinemia (12).
Patients with rheumatoid arthritis and that receive rituximab have greater infection rates if they develop hypogammaglobulinemia (12). Similarly, ANCAassociated vasculitis that is treated with rituximab, and where subsequent immunoglobulin-G (IgG) levels are <4 g/L, is an independent factor predictive for severe infection (13).
Following standard kidney transplantation, hypogammaglobulinemia is rarely a matter of concern as most immunosuppressive drugs target T-cells, and not B-cells. Nonetheless, a meta-analysis reported posttransplant IgG levels of <7 g/L in 40% of patients; in the same study, when severe hypogammaglobulinemia was observed, i.e., a level of <4 g/L, this was significantly associated with mortality (14). Conversely, a recent single-center study did not find a high prevalence of hypogammaglobulinemia or that a level of <4 g/L was associated with mortality (15).
In the setting of pretransplant desensitization, where rituximab is widely used, it is important to monitor immunoglobulin levels so that, if they become low (e.g., <4 g/L), polyvalent immunoglobulins can be substituted.
As of today, no study has examined immunoglobulin levels in the setting of ABO-incompatible kidney transplantation.
In this study, we prospectively monitored pre-and post-ABOi kidney transplantation immunoglobulin levels: if hypogammaglobulinemia occurred, i.e., if IgG levels decreased to <4 g/L, patients were given immunoglobulin (IVIg) substitution.

Objectives
We want to ascertain the factors that contributed to low immunoglobulin levels at post-ABOi kidney transplantation.

Study population
This is single-center study was conducted between March 2011 and May 2015. Forty-three patients underwent ABOi live-kidney transplantation; of these, 10 also had one or more donor-specific anti-HLA alloantibody (ies) (DSA). The desensitization protocol has already been published (16). Briefly, in the setting of ABOi/HLAi, patients received 40 days of pretransplant IV-Ig at 1 g/kg, followed by rituximab 375 mg/m² at days (D) -30 and -15 pretransplantation. Tacrolimus, mycophenolic acid, and steroids were started at D -10 (before transplantation). In addition, apheresis sessions For the purpose of this study, we prospectively monitored IgG, IgA, and IgM levels, as well as total lymphocyte counts, and CD4, CD8, and CD19 subsets at Hypogammaglobulinemia and ABOi transplantation 3

Ethical considerations
The research followed the tenets of the Declaration of Helsinki; the study protocol was approved by the ethics committee of CHU Toulouse (France) and all participants gave written informed consent before enrolling into the study.

Statistical analysis
Results are expressed as means (±standard deviation) or median (ranges) where appropriate. Comparisons between groups were done by using student t test. A P value of <0.05 was considered statistically significant.

Results
Of the 43 patients, two had early graft failure: one patient presented with immediate posttransplant allograft-vein thrombosis and underwent allograft removal, the other patient developed acute renal failure at three months posttransplant, which was related to secondary oxalosis: he lost his graft despite intensive hemodialysis therapy.

Overall population
Within the first year posttransplant 25% of patients experienced at least one severe case of hypogammaglobulinemia, i.e., IgG levels <4 g/L. On D --30, IgG, IgA, and IgM levels were within normal ranges at 10 ± 4.4, 1-9 ± 1.2, and 0.8 ± 0.5 g/L, respectively (Table 1). IgG levels were significantly decreased at D0 (4.2 ± 3.8 g/L) compared to D --30. At D15, IgG levels did not significantly differ from those on D0 or D --30. Conversely, IgG levels at posttransplant M1, M3, M6, M12, M18, and M24 were within normal ranges and were significantly higher than levels measured on D0 (Table 1). IgA levels were significantly decreased on D0 (1.1 ± 1.1 g/L) compared to those on D --30 (1.9 ± 1.2 g/L). After transplantation, they progressively increased and, at the last follow-up (M24), they were similar to those on D -30: i.e., 1.7 ± 0.9 g/L. In addition, IgM levels on D0 were significantly decreased compared to those on D --30 (0.3 ± 0.3 vs. 0.8 ± 0.5 g/L; P < 0.001). Thereafter, they progressively increased to reach pre-desensitization levels at M24 (0.7 ± 0.4 vs. 0.8 ± 0.5 g/L) ( Table 2). Within the first three months posttransplantation, 11 patients required IVIg therapy because their immunoglobulin levels were <4 g/L (IVIg + group). When we compared these patients with those that did not receive IVIg within the first three months posttransplantation (IVIggroup), IgG levels were similar at D --30 in both groups. Conversely, at D0, IgG levels were significantly lower in the Ig + group (2.4 ± 2.0 g/L vs. 5.5 ± 4.2 g/L; P = 0.009); the difference remained statistically significant until D15 posttransplantation (Ig + : 3.4 ± 1.7 g/L and Ig --: 6.6 ± 2 g/L; P = 0.0002), whereas there was no statistical difference between the two groups afterwards, except at M6 (Ig + 6.2 ± 2.3 vs. Ig -9.2 ± 2 g/L; P = 0.001). Table 3  HLAi (in addition to being ABOi) received two doses of rituximab (375 mg/m 2 ) at pretransplant instead of one dose in the ABOI group, we assessed immunoglobulin subclass levels and lymphocyte subsets, excluding the 10 HLA/ABOi patients ( Table 4). The levels of IgG, IgA, IgM and lymphocyte subsets were very similar to those of the entire population (Tables 1 and 3) for each timepoint; values were always non-significant. There was no significant difference between patients that had hypogammaglobulinemia and those that did not regarding the risk of posttransplant infections, e.g. bacterial infections, cytomegalovirus, BK-virus infections, or parasitic/fungal infections. (Table 5). Table 6 summarizes the studies reporting on prevalence of hypogammaglobulinemia in rituximab-rerated nontransplant patients.

Discussion
This is the first large series to report on posttransplant hypogammaglobulinemia in the setting of ABOi kidney transplantation where rituximab induction therapy was given at pretransplant to desensitize recipients. We found that within the first year posttransplant, 25% of recipients presented with at least one severe case of Table 3. Immunoglobulin G levels after ABOi kidney transplantation according to the need for IVIg at posttransplantation.    95% CI: 1.11--12.33; p = 0.03) for patients with IgG <4 g/L were higher than for patients with IgG >4 g/L. In addition, the odds ratio for 1-year all-cause mortality for those with severe hypogammaglobulinemia was 21.91 times greater than those with IgG >4 g/L (95% CI: 2.49--192.55; P = 0.005). Severe hypogammaglobulinemia during the first year posttransplantation significantly increased the risk of CMV, and fungal and respiratory infections, and was associated with a higher rate of 1-year all-cause mortality (14). Augusto et al, reported on a single-center retrospective study that included 318 kidney-transplant recipients. Immunoglobulins were measured prospectively on D15, M6, M12, and M24 posttransplant (15). The prevalence of IgG hypogammaglobulinemia was 56% and 36.8% at D15 and M6, respectively. Age was the sole identified risk factors for IgG hypogammaglobulinemia on D15 (OR: 1.02, P = 0.019). Risk factors at M6 for IgG hypogammaglobulinemia were the presence of IgG hypogammaglobulinemia on D15 (OR 6.41, P <0.001) and treatment of acute rejection (OR: 2.63, P = 0.014). Most infections occurred between D15 and M6 posttransplant. However, times of survival-free infection did not differ significantly between patients with or without IgG hypogammaglobulinemia on D15. Only septicemia that occurred between M6 and M12 posttransplant was more frequently observed in patients with hypogammaglobulinemia. The low prevalence of severe hypogammaglobulinemia (<4 g/d) did not allow conclusions on the risk of infections within this subgroup (15). In our series, severe hypogammaglobulinemia was relatively frequent after ABOi kidney transplantation; however, when replacement therapy was implemented (when IgG levels were <4 g/L), the risk of an infectious complication was no greater than in patients that did not have (severe) hypogammaglobulinemia.  (21). They found two subsets of patients, one of which recovers and one with persistent hypogammaglobulinemia with long-lasting low or absent memory B cells. They suggest that the treatment length in patients could be assessed by the trend of B-cell subsets by flow cytometry in addition to immunoglobulin levels.

Conclusions
In conclusion, our study is the first one to address hypogammaglobulinemia following ABO incompatible kidney transplantation. It was found to be frequent; however, when severe hypogammaglobulinemia was substituted with IVIg there was not more infectious complications as compared to that observed in patients with mild/moderate hypogammaglobulinemia that was not substituted. We recommend to monitor immunoglobulin levels following ABO incompatible kidney transplantation when rituximab therapy had been used for desensitization.

Study limitations
This is a single-center study with a limited number of patients. Some of them where only ABO incompatible, whereas others were in addition HLA incompatible. Therefore, the total dose of pretransplant rituximab was twice as much in the ABO incompatible/ HLA incompatible group as compared to the ABO incompatible group.