Risk of Pneumonia in Inflammatory Bowel Diseases: Reading between the lines!

Safety of pharmacotherapy is a key consideration in positioning therapies in patients with inflammatory bowel diseases (IBD). Several disease- and treatment-related factors modify the risk of serious infections, including older age, severe disease, corticosteroid and/or opiate use, and IBD-directed immunosuppressive therapy.^{1, 2} Pneumonia is one of the most common extra-intestinal infections observed in patients with IBD.^{3, 4}

In this issue of Alimentary Pharmacology and Therapeutics, Khan and colleagues compared the risk of pneumonia in 56,410 patients with the IBD treated with different medications, from the Veterans’ Affairs Healthcare System. Over ~600,000 person-year follow-up, they observed incidence rates of pneumonia, pneumonia-related hospitalization, and pneumonia-related mortality of 6.47, 2.52, and 0.43 per 1000 patient years, respectively. When comparing different medications, the overall incidence of pneumonia was comparable between different biologics (tumor necrosis factor [TNF]-α antagonists, vedolizumab, ustekinumab), and higher compared with immunomodulators and 5-aminosalicylates (Figure 1). After adjusting for covariates, they confirmed increased risk of pneumonia in patients treated with TNFα antagonists with or without thiopurines (vs. 5-aminosalicylates as reference), and patients experiencing IBD flare, besides conventional risk factors such as corticosteroid and/or narcotic use, older age, etc. Risk estimates for ustekinumab and vedolizumab were not significantly increased, but the number of patients exposed to these agents was small and overall point estimates were comparable to risk observed with TNFα antagonists.

Figure 1.

Incidence rate of pneumonia with different medications in 56,410 patients with IBD

The investigators followed rigorous methodology for observational comparative effectiveness research, with thorough adjudication for pneumonia and related hospitalization including review of medical records and chest imaging, use of time-varying exposures, and accounting for IBD disease activity using surrogate measures of recent and cumulative corticosteroid exposure. There are inherent limitations in a Veterans Affairs’ cohort comprising mostly of older males with high prevalence of comorbidities and lack of a closed hospital system such that patients may have sought care outside the system which is not captured. The investigators followed patients till the first diagnosis of pneumonia regardless of exposure, which may bias findings – patients who do not develop pneumonia with exposure to potent immunosuppressives such as corticosteroids and TNFα antagonists may be inherently less likely to develop pneumonia with other biologics.

These findings highlight the importance of examining safety of treatment strategies in totality, rather than specific medications in isolation. IBD disease activity is a significant risk factor for serious infections, highlighting that both the systemic immune suppression potential of medications, and their ability to control disease and achieve corticosteroid-free remission modifies the risk of serious infections. This is also supported by recent studies in which vedolizumab was associated with lower risk of serious infections compared with TNFα antagonists only in patients with ulcerative colitis (where these medications may have comparable efficacy), but not in patients with Crohn’s disease (where TNFα antagonists may be more efficacious).^5–8 Future studies identifying modifiable risk factors for serious infections, including the impact of treatment effectiveness on short- and long-term risk of serious infections are warranted, to help inform positioning of therapies as our therapeutic armamentarium expands.⁹