AIHA and rituximab: time for Pneumocystis prophylaxis?

In this issue of Blood Advances, the study by Zadro et al¹ contributes new valuable information about the risk of infections after rituximab treatment of autoimmune hemolytic anemia (AIHA). Although corticosteroids as the preferred first-line therapy in AIHA often results in an initial response, relapse after tapering and steroid dependency are frequent. Therefore, a second-line treatment may be required, and rituximab has become the drug of choice in this setting for decades and is generally perceived as well tolerated and safe.^2,3 An increased risk of infections and related mortality among patients with AIHA have been reported previously, but the sequence between treatment and subsequent infection risk has not been described before.^2,4,5 

In the French AHEAD cohort, comprising all patients in France diagnosed with autoimmune hemolytic anemia, the authors identified all incident cases with rituximab exposures.⁴ They included 959 such patients on their first day of rituximab treatment to assess the risk of infections in the following 6 months.¹ The majority of included patients were refractory to first-line treatment and progressed quickly to second-line rituximab, often included shortly after initial diagnosis.¹ 

The results reveal a concerning cumulative mortality of 9% at the end of the 6-month follow-up. Although the included patients with their early requirement of second-line treatment (ie, rituximab) may not be representative of all patients with AIHA, these results warrant consideration in the routine care of comparable patients. The authors present a detailed analysis of outcomes after rituximab exposure, highlighting specific subgroups in which the association between disease, treatment, and adverse events are particularly high.

Beyond mortality, the cumulative proportion of hospitalizations from infections reached 17.6% after 6 months, and the time from first rituximab exposure to hospitalization was short, occurring after a median of 43 days. Additionally, hospitalization for infections were associated with a subsequent high 30-day mortality of 12.5% and with a median time to death of just 15 days. In a comprehensive analysis of risk factors associated with this, only advanced age and concomitant corticosteroid exposure remained clearly associated with hospitalization for infections.

Noteworthy, microbiological agents usually classified as opportunistic were frequently the cause of hospitalizations, accounting for 16.6% of all hospitalizations. Pneumocystis was, among these, the most frequently registered opportunist infection, followed by Aspergillus infections, with candidemia being the third most. Patients hospitalized for pneumocystosis (1.4%) were all concomitantly exposed to corticosteroids for >30 days and were older than the patient population in general, with a mean age of 77.5 years. Of note, none of these patients received concomitant Pneumocystis prophylaxis. The 30-day cumulative mortality was 28.6% among patients with pneumocystosis, with a median time to death of just 11 days.

The association between autoimmune hemolytic anemia, immunosuppressive treatment, and life-threatening pneumocystosis is not completely surprising, because this has been reported previously in other autoimmune diseases treated with rituximab.^6-8 However, these novel disease- and treatment-specific risk estimates call for translation into clinical care. In AIHA, it is hardly feasible to conduct dedicated randomized clinical trials on low-risk supportive care, such as sulfamethoxazole/trimethoprim, for all treatment exposures, and we must consider such large-scale observational data to guide our care.

Among the 959 patients treated with rituximab, only 14 were hospitalized with pneumocystosis. This cumulative risk of 1.4% is lower than the pooled overall risk described in the Cochrane review of Pneumocystis prophylaxis among non-HIV immunocompromised patients.⁹ However, when we consider that all but 1 of the patients hospitalized for pneumocystosis were aged >60 years, the cumulative risk in this more restricted population increases, approaching the baseline risk observed in the studies included in the Cochrane meta-analysis.

We as scientists and clinicians could wish for further studies to corroborate these risks and research comparing the risk of infections in patients with AIHA, with and without rituximab exposure. Nevertheless, given the scale and comprehensiveness of this epidemiological real-world study at hand, it is probably the best evidence we will get for a long time. We suggest that the results from Zadro et al are sufficient both to raise concern and suggest management. Pneumocystis infection among older patients with AIHA treated with rituximab and corticosteroids has a high mortality, and these data should guide a recommendation for pneumocystosis prophylaxis according to local guidelines to mitigate the risk.^6,9 

Despite the results from Zadro et al, rituximab remains a good and often well-tolerated second-line treatment in many patients with AIHA not responding or relapsing after first-line corticosteroids.^2,3 

The study reminds us that combining evidence from classical, sufficiently powered randomized trials to evaluate the best disease-modifying treatments with large-scale real-world data to assess the risks of complications and toxicity leads to the best-informed decision-making.

Conflict-of-interest disclosure: D.L.H. reports conference fees from EUSA Pharma and Alexion, and advisory board fees from Takeda and Janssen. H.F. reports project grants from Novartis, Alexion, Sanofi, and Gilead for research unrelated to this study, and honoraria from Sanofi and Alexion for lectures on thrombotic microangiopathies.