Here, we report on the results of a single-center retrospective study investigating the efficacy of Rituximab in treating CVID-associated GLILD. Based on clinical and radiological data from lung CT scans, a diagnosis of GLILD was achieved in 6.0% of our CVID cohort. GLILD prevalence is reported to vary between 8 and 20% depending on the study setting, and our data is in line with the literature [4]; however, we could possibly have underestimated the actual prevalence of GLILD in our cohort of CVID patients as 22 of them have been excluded from this study because lung CT scans data were not available. Common radiological features that led to GLILD diagnosis were ground-glass opacities and nodules/micronodules, which were present in all of our GLILD-affected patients; additional recurrent features such as airspace consolidation, interlobular septal thickening, hilar/mediastinal lymphadenopathies, and splenomegaly were also present in two-thirds of the patients. Taken together, these radiological findings have already shown to be highly specific for GLILD diagnosis, in particular when comprehensively evaluated using a scoring system such as the Baumann score [14]. Considering that lung biopsy carries an increased risk of morbidity and mortality in interstitial lung disease patients [15], in selected situations such as tertiary referral centers where experienced physicians deal with IEI-affected patients, invasive procedures could be overcome when clinical, radiological, and PFTs data are strongly suggesting the presence of GLILD.

Many patients in our cohort presented CVID-related symptoms since pediatric age, but a diagnosis of CVID was done in the vast majority of patients in adult age. This might have played a role in exposing patients to persistent microbiological stimuli that together with impaired antigen clearance may at least partially contribute to the development of immune dysregulatory manifestations [16]. Prior to GLILD diagnosis, all 6 CVID patients had already developed a heterogenous and complex spectrum of complications, in particular autoimmune phenomena and chronic lymphoproliferation. Hypersplenism and autoimmune manifestations have already been reported as strong risk factors for GLILD diagnosis in CVID patients [17, 18], and this must be kept in mind when evaluating CVID patients with complex clinical history, as CVID patients with non-infectious complications seem to be more prone to develop interstitial lung disease and therefore must be screened accordingly.

Increased IgM serum levels have been reported in GLILD-affected CVID patients, and an expansion of IgM memory B cells has been proposed as part of the pathogenetic mechanism beyond granulomata and lymphocytic infiltrates in the lung [19]; in contrast with these findings, only one of our patients displayed raised IgM serum levels. We also observed an expansion of CD19hiCD21lo B cells in 66.6% of GLILD-affected patients, although the reduced size of our cohort could not lead to statistical significance. Nevertheless, CD19hiCD21lo expansion has already been reported in GLILD-affected CVID patients and therefore should be highly considered as a potentially warning sign for GLILD development risk [20].

A recent publication by Fraz and colleagues analyzed various biomarkers in GLILD patients [21]: we confirmed increased sTIM-3 levels when compared to healthy donors and CVID patients without GLILD diagnosis. Conversely, in our cohort, sCD25 serum levels, a marker of T-cell activation, did not differ between CVID patients without or with GLILD diagnosis, but this might be due to the reduced size of our cohort.

In contrast with other published cohorts of GLILD patients [7], genetic analysis in our cohort did not reveal monogenic defects linked to immune dysregulation in CVID.

All our patients were treated with Rituximab monotherapy as first-line treatment, without the needs of systemic steroids nor additional chemotherapy. Anti-CD20 monoclonal antibody was administered at 375 mg/m2/months for 6 months (induction) and later continued at the same dosage every 3 months (maintenance). Excluding a case report already published by our group [22], Rituximab monotherapy for GLILD has been reported in a total number of 19 patients [6, 23,24,25,26,27]. Comparing the therapeutical scheme, all the other reported patients have been treated with Rituximab 375 mg/m2 weekly infusions for 4 weeks, with a 4- to 6-month interval between each cycle. While a comparison between the different therapeutical schemes should require a different study design, it is interesting to observe that the patient reported by Zdziarsky and colleagues that was initially treated with low dosage Rituximab monotherapy (150 mg/m2 weekly for 6 weeks) presented GLILD relapse 6 months after the end of the treatment but achieved complete remission after increasing the dosage to the standard dose of 375 mg/m2 [23]. We therefore strongly suggest that Rituximab-based treatment for GLILD should follow the standard dosage of 375 mg/m2, despite the chosen interval of administration.

Treatment efficacy was evaluated comparing SGRQ score results, immunological data, PFT results, and lung CT scan findings (using the Baumann score) at GLILD diagnosis and after 6 months of treatment. The use of the SGRQ has enabled us to quantify the impact of GLILD in terms of symptoms, impairment of daily activities, and quality of life. When used for assessing treatment response in asthma or chronic obstructive pulmonary disease, a mean change score of 12 units is associated with very efficacious treatment [9]. After 6 months from the beginning of Rituximab treatment, we observed a mean change of 50.6 units (+ 5.4); since we have also reported an important reduction of the Baumann score and an amelioration of the PFTs parameters, such a large decrease of the SGRQ score might be explained on one hand by a reversion of the lung parenchyma alterations (e.g., granulomata, lymphocytic infiltrates, ground-glass opacities) that does not occur in other types of chronic lung disease where treatment is not targeting the potential pathogenic mechanisms and thus is not capable of reverting the damages.

Lymphocyte subset analysis after Rituximab treatment showed only minor modifications. As expected when using B-cell depleting therapies, we observed absence of peripheral B cells in all treated patients, while no other significant differences on the remaining lymphocyte subpopulations were noted. On the contrary, serum sCD25 and sTIM-3 levels showed a decline upon Rituximab treatment, suggesting a possible indirect effect of anti-CD20 monoclonal antibodies on T-cell activation and exhaustion. To the best of our knowledge, this is the first time that sCD25 and sTIM-3 are evaluated both at GLILD diagnosis and after treatment. Although available only in research settings, sCD25 and sTIM-3 could be used as GLILD biomarkers not only as an aid in GLILD diagnosis but also in disease activity monitoring during follow-up. In addition, if reproduced in other Rituximab-treated GLILD patients, these results could become fundamental when deciding the therapeutical regimen, as additional chemotherapy or immunomodulators targeting T cells might not be needed: in the Verbsky’s study [7], combined immunosuppressive therapy with Rituximab and antimetabolite (i.e., azathioprine or mycophenolate mofetil) resulted in long-lived remission in the vast majority of the treated patients; however, leukopenia o lymphopenia, increased rate of infectious episodes, and septicemia occurred, suggesting that combined therapy regimen might cause (at least in some patients) excessive immunosuppression.

In our cohort, PFTs have shown an important role both in highlighting a restrictive pattern, which is typical of GLILD, and in monitoring disease remission after Rituximab treatment. PFTs are a reproducible and noninvasive procedure already recommended but not always performed for CVID patients’ follow-up [28]. Therefore, our results underline the importance of monitoring lung function in CVID patients, as early identification of CVID patients presenting a decline in lung function warrants further investigations such as lung CT scans. Furthermore, the Baumann score has already been used for evaluating the extent of GLILD with good reproducibility among trained radiologists [12]. A decrease of the Baumann score followed the same trend of clinical and PFTs improvement; besides its valuable application in research settings, our data suggest that the Baumann score may be applied in clinical practice as a tool for screening CVID patients for GLILD and monitoring GLILD activity.

Treatment approach for GLILD is variable and mainly based on the single-center own experience. Systemic corticosteroids are usually proposed as first-line treatment; the large cohort recently reported by Smits and colleagues [29] revealed that high-dose corticosteroids are needed to achieve remission and patients with relapse tend to show a worse disease response when retreated with corticosteroids. In addition, opportunistic infections and steroid-related skeletal complications were reported [29]. On the other hand, as reported by Verbsky [7], combined immunosuppressive treatment induced radiological remission in the majority of patients even though GLILD relapse occurred when immunosuppressive treatment was stopped, and patients might present increased infectious episodes caused by targeting both B and T cell. On the contrary, Rituximab monotherapy seems safer as no increased rate of infectious episodes was observed in our cohort as well as no drug-related adverse events, with all the treated patients experiencing complete disease remission after 6 months of treatment.

This study presents some limitations. First of all, the retrospective nature of our study and the use of a tertiary academic single center represent another obvious bias. In addition, anti-CD20 monoclonal antibody was the only prescribed treatment in our cohort, and therefore, we could not compare the outcomes with other therapeutical options; likewise, in our cohort, only adult GLILD-affected CVID patients have been studied and treated, and therefore, our results may not be replicated with pediatrics patients. Patients were evaluated for treatment response after 6 months of treatment, but no additional related evaluations were performed after this interval—all patients remained in stable conditions. Finally, increasing the size of the cohort of patients could lead to more robust results.

In conclusion, first-line monotherapy with Rituximab displayed high efficacy in disease remission in the entire treated cohort, with improvement of symptoms and amelioration of quality of life, as well as restoration of PFTs and resolution of lung CT scan findings. Both preclinical studies on GLILD pathogenesis and multicenter randomized clinical trials to ultimately assess the efficacy of Rituximab therapy over steroids or other immunosuppressive agents are warranted to further understand this rare yet fatal complication of CVID and develop strong evidence-based therapeutical guidelines for affected patients.