Serious and Non-Serious Infections in Non-Splenectomized Adults Treated for Persistent or Chronic Primary Immune Thrombocytopenia: Effect of Treatments and Protection of Vaccines. Results of the French Nationwide Faith Study

Background: No study has specifically assessed the risk factors of infection in immune thrombocytopenia (ITP) patients in the era of rituximab therapy. Similarly, the protective effect of pneumococcal and influenza vaccines is unknown in real-life settings.

Aim: The objective of this study was to assess the risk factors of serious infection (SI) and non-serious infection (NSI) in non-splenectomized adults treated for persistent or chronic primary ITP, including baseline characteristics, ITP treatments as well as pneumococcal and influenza vaccines.

Methods: The population was the 2009-2012 FAITH cohort, which is the cohort of all incident primary ITP adults persistently treated in France built in the national health insurance database (SNIIRAM). The FAITH study is registered ENCePP n°4574. SIs were hospitalizations with infection as primary diagnosis code (predictive positive value: 97%). NSIs were identified using out-hospital antibiotic dispensing. Cox models were performed to assess separately the risk of SIs and of NSIs. The index date was the date of first ITP treatment exposure (date of entry in the FAITH cohort). Follow-up was censored at a first infection, splenectomy, death or end of follow-up (December, 31 2012). The following covariates were included: age, gender, mucosal bleeding at ITP onset, diabetes mellitus, chronic pulmonary, heart and kidney diseases. Exposure to ITP treatments was considered as time-varying covariates. Exposure to rituximab was defined as the 6 months period following dispensing and exposure to vaccine occurring in that period was considered as null.

Results: The cohort included 1805 patients. Mean age was 57.6 ± 21.3 years and 58.9% were females. With a mean follow-up of 18.5 ± 6.8 months, 681 patients were exposed to rituximab, 1035 to IVIg, 90 to other immunosuppressants (azathioprine, mycophenolate); 161 (9.1%) patients experienced SIs (39.7% respiratory) and 1227 (68.0%) NSIs. The respective incidences were 6.3/100 patients-years and 8.4/100 patients-months. In multivariate analyses, increasing age and chronic pulmonary disease were associated with both SI and NSI occurrence. Mucosal bleeding at ITP onset, male gender and chronic kidney disease were also associated with SI occurrence. The hazard ratios (HRs) for corticosteroids were 3.83, 95% confidence interval - 95% CI [2.76-5.31] (p<0.0001) for SI and 2.46, 95% CI [2.19-2.76] for NSI (p<0.0001). HRs were respectively 2.60, 95% CI [1.67-4.03] (p<0.0001) and 1.49, 95% CI [1.28-1.74] (p<0.0001) for rituximab. Pneumococcal vaccine showed a protective effect for both SIs and NSIs (respectively, 0.38, 95% CI [0.20-0.73], p=0.003 and 0.52, 95% CI [0.43-0.65], p<0.0001), as well as influenza vaccine (respectively, 0.42, 95% CI [0.27-0.64], p<0.0001 and 0.49, 95% CI [0.41-0.59], p<0.0001). IVIg was also associated with a higher risk of infection, probably due to channeling bias.

Conclusions: SIs are mainly respiratory infections in this population. Chronic pulmonary disease, corticosteroids and rituximab are the main risk factors for infections, while pneumococcal and influenza vaccines are protective against SIs and NSIs.