To the Editor:
Pneumocystis jirovecii pneumonia (PCP) is an infectious fungal disease caused by P. jirovecii, which has attracted the attention of physicians treating patients with human immunodeficiency virus infection1, as well as those with connective tissue diseases, malignancies, and organ transplantation2.
In patients with rheumatoid arthritis (RA), PCP used to be an uncommon infectious disease, but the number of case reports of PCP in patients with RA has increased since the introduction of low-dose methotrexate as an anchor drug for RA in the 1980s3. Moreover, with the introduction of anti-tumor necrosis factor (TNF) therapy, a further increase of incidence of PCP in patients with RA has been noticed, especially in Japan where strict postmarketing surveillance (PMS) programs have been conducted for patients with RA treated with TNF inhibitors4,5,6. The numbers of patients with RA with PCP in these PMS programs who were treated with infliximab (IFX), etanercept (ETN), or adalimumab (ADA) were 22 (0.4%) out of 50004, 25 (0.18%) out of 13,8945, and 25 (0.33%) out of 7469 patients6, respectively. Notably, these incidence rates of PCP in Japan are higher than the corresponding figure (0.01%) reported from the United States. Therefore, we previously analyzed the clinical characteristics and risk factors for PCP in patients with RA in Japan treated with these 3 TNF inhibitors7,8,9,10. In most cases, PCP developed rapidly with respiratory failure, and P. jirovecii organisms could not be detected microscopically, requiring for diagnosis the PCR test for P. jirovecii DNA or the measurement of plasma or serum 1, 3-β-D-glucan levels. Only some of the cases were in an immunocompromised state, showing a remarkable decrease in concentrations of serum immunoglobulins and the number of peripheral blood lymphocytes. Some of the risk factors for PCP were common to IFX and ETN, but others differed among the 3 TNF inhibitors8,9,10. It is possible that small sample size, patient background, the launch time of each drug, distinct mechanism of action among TNF inhibitors, or other unmeasured factors resulted in these differences in risk factors for PCP among the drugs. We therefore merged and analyzed the data of our previous studies to identify risk factors for development of PCP common to these 3 TNF inhibitors. Details of the designs and the methods of our previous studies were published elsewhere8,9,10.
In our previous studies, we accumulated a total of 53 patients with RA who developed PCP under treatment with 1 of the 3 TNF inhibitors: IFX in 21 cases, ETN in 15 cases, and ADA in 17 cases. Of these, 51 patients who developed PCP within 12 months after commencement of a TNF inhibitor (the PCP group) were analyzed. For a control group, 265 patients with RA who did not develop PCP within 12 months after the beginning of a TNF inhibitor (the non-PCP group) were randomly selected from a consecutive series of patients with RA in the hospitals that participated in these studies.
To characterize the PCP group, we compared demographics, comorbidities, laboratory data, and concomitant drugs between the PCP and the non-PCP groups at the time of initiation of treatment with TNF inhibitors (Table 1). Compared with patients of the non-PCP group, patients of the PCP group were significantly older, had a lower percentage of women, had a higher prevalence of comorbid pulmonary disease and diabetes mellitus, and were treated with a higher daily dose of prednisolone. None of the 51 patients were receiving chemoprophylaxis for PCP at the time of PCP diagnosis. All of the patients received therapeutic doses of trimetho-prim/sulfamethoxazole, except for 1 who received pentamidine isethionate. One case in an ETN study9 and 3 cases in an ADA study10 died after developing PCP.
A multivariate Cox proportional hazard analysis yielded HR and 95% CI for the following risk factors: older age (per 10-yr increment), 1.8 (1.4–2.5); presence of comorbid pulmonary disease, 3.1 (1.8–5.3); comorbid diabetes mellitus, 2.9 (1.5–5.8); and daily dose of prednisolone ≥ 5 mg, 2.7 (1.5–5.0). The cumulative probability for developing PCP was calculated using the Kaplan-Meier method (Figure 1). Patients with 3 or more of these risk factors had a significantly higher risk for PCP than those with 2, 1, or no risk factors (p < 0.001 for all comparisons by the log-rank test). Patients with 2 risk factors had a significantly higher risk for PCP than those with 1 or no risk factors (p = 0.045 and p < 0.001, respectively). Patients with 1 risk factor had a significantly higher risk for PCP than those with no risk factors (p = 0.008).
We were able to identify common and more robust risk factors for the development of PCP by combining the data of PCP developed during treatment with any 1 of the 3 TNF inhibitors. Although prophylaxis success must be demonstrated in randomized controlled trials, we consider that it is important to screen for these risk factors based on this analysis and consider diligent prophylaxis before starting a TNF inhibitor.
Acknowledgment
We thank all patients who participated in the clinical studies and all doctors who provided clinical data of their patients. We express special thanks for providing the clinical data of the patients in the control group to Prof. Yoshiya Tanaka (the First Department of Internal Medicine, School of Medicine, University of Occupational and Environmental Health, Japan), Prof. Tsutomu Takeuchi (Division of Rheumatology, Department of International Medicine, Keio University School of Medicine), Prof. Koichi Amano (Department of Rheumatology and Clinical Immunology, Saitama Medical Center, Saitama Medical University), Associate Professor Yuji Akiyama (Department of Rheumatology and Applied Immunology, Saitama Medical University), Dr. Kei Ikeda (Department of Allergy and Clinical Immunology, Chiba University Hospital), and Dr. Mitsuhiro Iwahashi (Higashi-Hiroshima Hospital). We also thank Drs. Kaori Watanabe-Imai, Yukiko Komano, Toshihiro Nanki, and Nobuyuki Miyasaka (Department of Rheumatology, Graduate School of Medical and Dental Sciences, Tokyo Medical Dental University) for their critical reading and helpful comments for the manuscript.
Footnotes
Supported by a grant-in-aid for scientific research (KAKENHI) from the Japan Society for the Promotion of Science to RK (#19590530), MH (#20390158), and MT (#23590171), and by a grant-in-aid from the Ministry of Health, Labor and Welfare, Japan (H19-meneki-ippan-009 to NM and MH, and 22-meneki-ippan-001 to MH).
REFERENCES
- 1.
- 2.
- 3.
- 4.
- 5.
- 6.
- 7.
- 8.
- 9.
- 10.