Discussion
This study is the first to estimate and emphasize on the frequency (12.5%) and the clinical utility of L-TGL determination at DLBCL diagnosis using systematic SEP. We demonstrated that L-TGL has a strong negative impact on the overall and infectious-related mortality rates. The higher AG ratio in L-TGL than in H-TGL patients suggests that the pronounced decrease in TGL levels and the associated clinical outcomes in the L-TGL subgroup may be 1) first, independent of the other possible associated unspecific causes and 2) second, mostly and directly related to one of our two hypotheses: either a pre-existent but previously undiagnosed PID on which DLBCL developed or an SID directly induced by the DLBCL condition. Finally, for both hypotheses, considered alone or in combination, the ability of a chemotherapy-induced SID to worsen, or reveal a greater susceptibility to infections in L-TGL patients is probable but remains unquantifiable in the context of the design of the present study. Our population was close to what has been described in France during this period of time, with a median age of 69 years, representing 35 to 40% of patients with lymphoma.(23,24)
At DLBCL diagnosis SEP is not unanimously recommended according to the international workgroup guidelines for malignant lymphoma(13–15), however from the results in our study we propose SEP to be performed at diagnosis. In our study, there were no significant differences between the two subgroups regarding inflammatory parameters, as assessed using CRP level, and albuminemia, suggesting that the decrease in TGL may not be secondary to metabolic/inflammatory causes or to increased loss of immunoglobulins but is probably directly related to low gamma-globulin production. In the same manner, since neither the Ann Arbor staging classification, IPI, nor the chemotherapy regimen were significantly different between the two groups, the increase in the overall and specific infection-related death rates may be attributed to the decrease in TGL in L-TGL subgroup. We didn’t exclude neither Richter syndrome nor associated CNS involvement DLBCL subtypes that could exhibited worst malignant prognosis, because the first were identified only in H-TGL subgroup and both subtypes didn’t show overall highest mortality rate in this study. Therefore, we may consider that this study drew up an overall picture of all subtypes of DLBCL. Moreover, showing that L-TGL patients exhibited a lower survival rate than all other possible subgroups of DLBCL, including patients who did not get a SEP at diagnosis, we further confirmed the importance of TGL assessment.
The main limitations of this study include its retrospective design with some missing data, including mainly immunoglobulin isotype levels, absence of systematic and reliable assessment of previous infections, and monitoring of TGL in the follow-up for the entire cohort. Consequently, two main secondary objectives could not be reached in this study: determination of an approach for eventual CVID diagnosis criteria and a differentiation scheme for the two causes of L-TGL considered herein, namely, primary or DLBCL-related secondary immunodeficiencies. We assumed that combining analyses of both DLBCL and TGL outcomes under chemotherapy should partially help to differentiate these two causes of decrease in TGL for each patient. Indeed, complete or partial remission of DLBCL and concomitant increase in TGL under chemotherapy would support a DLBCL-related SID. Conversely, in cases of a lack of response or early death, no conclusion could be drawn, since chemotherapy would induce a further decrease in TGL for an unknown duration.
To conclude, while epidemiological data on lymphoma treatment strategies are optimistic and have seen improvements in survival throughout the years. This study, using a simple, widely and easily available biological test, seems to show that L-TGL patient subgroup exhibits worse prognosis with higher overall and infection-related mortality risks than H-TGL subgroup. Therefore, this study argues for the need to systematically perform at least SEP at lymphoma and especially at DLBCL diagnosis and emphasizes on the need of further studies to confirm these results especially in prospective studies: 1) to determine nosological and prognostic distinctions between possible previously undiagnosed PIDs and DLBCL-related SIDs using, for example, next-generation sequencing methods to uncover molecular defects of humoral PIDs(22,25) and to assess the timing of a possible TGL recovery after chemotherapy in relation to lymphoma outcomes; 2) to assess the possible clinical benefit of immunoglobulin supplementation for SIDs, or prophylactic antibiotics, taking into account advances in chemotherapy and management strategies in Onco-Haematology.