Glucocorticoid use in rheumatoid arthritis patients and the onset of pneumonia: a systematic review and meta-analysis

Methods

Study design

A systematic review was conducted and completed on August 1, 2021, to evaluate the risk of GC (prednisone, methylprednisolone, and prednisolone) use in RA patients following the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) (Figure 1). There was no funding provided, and Institutional Review Board (IRB) and ethical approvals were not required because this was a review of previously published literature that did not include participants’ information.

Figure 1:

Preferred reporting items for systematic reviews and meta-analyses (PRISMA) flow chart.

Results

Meta-analysis

Examining model statistics, τ²=4.56 (95% CI: 1.39–44.91], I2=90% (95% CI: 78.8–95.0), H=3.12 (95% CI: 2.17–4.47), and Q=38.82 (p<0.01), demonstrated high heterogeneity and supported the use of a random-effects model. The Egger’s index showed a nonsignificant publication bias (p=0.414). Statistical results reveal that patients treated with DMARDs are 95% (95% CI: 0.65–0.99) less likely to develop pneumonia compared to patients treated with a DMARD and a GC (p=0.002) (Figure 2).

Figure 2:

Forrest plot of study outcome 1.

Discussion

Results from our meta-analysis supports our hypothesis that patients with RA who are treated with a DMARD (such as MTX, HCQ, and SSZ) alone are significantly less likely to develop pneumonia compared to patients treated with a DMARD and a GC (such as prednisone, methylprednisolone, and prednisolone). This is noteworthy because ILDs are only second to cardiac disease as a cause of mortality in RA patients [26, 27]. Among the ILDs, which can increase mortality, pneumonia seems to be a common etiology [26]. Despite usual interstitial pneumonia (UIP) being the most common pattern of pneumonia among RA patients with ILD [28], the onset of UIP carries a worse prognosis for RA patients compared to other patterns of RA-ILD [28]. Indeed, some research supports the notion that the outcomes of RA-UIP are similar to interstitial pulmonary fibrosis [28, 29]. While research has traditionally pointed the finger at the use of DMARDs in patients with RA as a cause of the onset of respiratory disease, our findings shift the focus.

Housden et al. [23] suggests that oral steroid therapy may be a major contributory factor in the development of LRTIs. In 22 RA patients hospitalized for pneumonia, 54% were utilizing GCs for the treatment of RA for at least 1 year prior to their admission [23]. The prevalence of oral steroid consumption among admissions was significantly higher than in the RA group as a whole (n=1822; 14%, p=0.021) [23]. Over half of the patients admitted in the study were taking oral steroids, compared to 14% of the RA population in general (p=0.021) [23]. However, the results of Holland-Fischer et al. [24] suggest that rather than DMARD use increasing the risk for pneumonia, the risk lies in disease activity. Patients with RA with at least one prescription for prednisolone within 3 months prior to hospitalization had a more than 40% increased 90-day all-cause mortality compared with patients with RA treated with MTX [24]. While treating RA patients with biologics has been associated with the onset of infection [30], results show that biologics do not significantly increase the mortality rate [24]. This could be due to a multitude of factors. Biologics fall into the class of DMARDs [31]; however, their mechanism of action works by preventing specific aspects of an immune response further along in the process [31], compared to conventional treatment (e.g., MTX), which acts by nonspecifically inhibiting cell-mediated immunity [31]. Biologics are an alternative treatment for RA [31] and have been shown to play a smaller role in the development of ILDs in RA due to this increased specificity [32], in contrast with conventional DMARD agents [32]. Nonetheless, future studies would benefit from analyzing whether the similar pattern of pneumonia onset seen with DMARD combination therapy with GC also occurs with biologic combination therapy with GC.

Grijalva et al. [25] established that the initiation of LEF, SSZ, or HCQ did not increase serious infections, compared with MTX [25]. However, both the initiation of a DMARD and concurrent GC use were associated with a dose-dependent increase in serious infections such as pneumonia [25]. Although these studies were conducted in different parts of the world, they share a common theme: the risk for developing pneumonia in RA patients has consistently been higher in those on combination therapy (DMARD + GC) or GC monotherapy compared to DMARD monotherapy [23], [24], [25]. Additionally, their results suggest that the progression of pneumonia, which can cause irreparable damage to RA patients [28, 29], can be associated with high disease activity. This emphasizes the importance of maintaining RA patients on DMARDs to ensure proper treatment and prevention of disease progression. Common practice tends to stop the use of the DMARD in RA patients at risk for infection, while our findings suggest that this measure could result in a worse prognosis compared to maintaining DMARD treatment [24].

Currently, there is no consensus on how to manage DMARD therapy in the perioperative setting [33]. Although the recent guidelines from the American College of Rheumatology/American Association of Hip and Knee Surgeons recommend continuing conventional DMARDs and holding biologics for one dosing interval before surgery [34], our findings support continuing DMARD therapy and withholding GCs. Ultimately, the cessation of DMARDs may set patients’ progress back by allowing the disease process to continue unchecked and the use of GCs to increase the risk of infection, regardless of the use of a DMARD [24, 34, 35]. Our results should encourage practitioners to be more mindful of discontinuing DMARD therapy and maintaining GCs in RA patients because of the possibility of pneumonia, which carries a significantly high burden of fatality in these patients [28, 29]. Our data do not definitively suggest the use of GCs as an etiology behind the onset of respiratory disease among RA patients, and our data contain natural limitations such as inevitable exhaustion bias from screeners while reading through articles. However, our findings do raise the important question of how to treat the underlying disease without simultaneously triggering the onset of a life-threatening infection. The current literature is limited on the specific adverse effects of treatment for RA. Although a significant amount of research addresses the idea that the effects of disease-managing treatment courses can increase the risk of infection and complicate the course of RA, limited work adequately correlates the onset of pneumonia with the use of GCs with patients already on DMARDs. Due to the specificity of this analysis, we were limited by the number and types of studies we included and by the high heterogeneity (I2=90%) exhibited in included studies. This review and analysis only utilized six search engines, which could have limited the search and possibly skew our statistical findings. Additionally, the performed searches only focused on three DMARDs and GCs commonly utilized for RA. Future studies should analyze the correlation of pneumonia or ILDs with additional DMARDs and GCs, which are also utilized in RA. Furthermore, the exclusion of articles written in English, as done by our review, could downplay the amount of universally present research covering this topic. We believe these factors to be the greatest limitations of this study. However, the data from these analyses appear to provide an interesting linkage between RA treatment and pneumonia. To provide a more salient relationship, future studies should not only focus on analyzing the same factors (RA patients being treated with only a DMARD vs patients treated with a DMARD plus a GC) and the occurrence of pneumonia, but also through an experimental randomized control study within a regulated environment, in order to eliminate heterogeneity.

Conclusions

Our data suggest a concern for pneumonia development in GCs plus DMARD-treated patients; however, our data do not allow us to determine the causality of the relationship. Although there are many published articles suggesting a relationship between GC treatment in rheumatic individuals and the development of infections, we believe that our study is unique in its approach and findings for (1) pneumonia infections specifically, and (2) with the precision that a statistical analysis provides. We believe that our statistical analysis will allow physicians to better devise a holistic treatment plan for rheumatic patients. This approach is central to osteopathic medicine, and will guide physicians to weigh the potential benefits with the statistically significant pneumonic complications.