• 1805: Are DMARDs Associated with Increased Risk of Infection in RA?
• 1804: Congestive Heart Failure: Contributor of Increased Mortality?
• ACR Special Abstract: Vioxx Cardiovascular Safety Data
• Abstract 1764: TNF Inhibitor Therapy and Postoperative Orthopedic Infection in RA

Abstract 1805 Are DMARDs Associated with Increased Risk of Infection in RA?
Lacaille, Anis, Esdaile, Guh

Because RA disease itself predisposes to an increased risk of infection, determining if additional infection risk can be ascribed to DMARDs can be difficult. Further, though potentially immunosuppressing, the ability of DMARDs to suppress RA disease activity may result in an overall reduction in infection risk in some patients. Here, Lacaille et al explore infection risk associated with non-biologic DMARD in a large population-based longitudinal cohort.

Methods: Centralized medical coding data was used to identify all RA patients treated in British Columbia from January 1996 to December 2000. Data on infections in these patients was obtained from ICD-9 billing codes from outpatient visits, hospitalizations, and from antibiotics dispensed. Data on DMARD exposure (methotrexate, leflunomide, sulfasalazine, azathioprine, cyclosporine, and cyclophosphamide) and exposure to corticosteroids were obtained from centralized pharmacy records. Infection rates on and off DMARDs were calculated and compared among RA patients.

Results: 27,104 RA patients were identified (corresponding to an RA prevalence of 0.76% of the population of British Columbia) with 103, 588 person years of available follow-up. 67% of patients were women with a mean age of 62 years. 43% of patients were prescribed at least one included DMARD. 38% were prescribed oral corticosteroids (with a statistically significant imbalance in oral glucocorticoid use between the DMARD treated and untreated groups, with more patients in the DMARD group treated with glucocorticoids). 83% were prescribed NSAIDs.

There was no statistically significant difference observed in the number of infections requiring hospitalization between the DMARD treated and DMARD untreated groups. In a multivariate analysis, a decreased risk of infection was observed in DMARD treated compared to DMARD untreated patients:

Conclusions: Non-biologic DMARD therapy was not associated with an increased risk of infection compared to patients not treated with DMARDs.

Editorial Comments: These results support the premise that suppression of RA disease activity may serve to counterbalance, or perhaps supercede, the potential immunosuppressive effects of DMARDs that predispose to infection. While this investigation looked at several DMARDs as a group, there are important differences in the individual DMARDs in terms of their immunosuppressive capabilities (with cyclophosphamide and azathioprine more likely than methotrexate or leflunomide to be associated with serious infections) that would merit sub-analyses of each DMARD alone (or when DMARDs are used in combination). However, as the likely proportion of patients receiving cyclophosphamide or azathioprine for RA in any population is low, most of these results can be extrapolated to apply primarily to the other included DMARDs.

As with all population based studies based on medical coding for reimbursement, some uncertainties regarding diagnosis and patient characteristics are well recognized. On the other hand, this lack of detail is somewhat counterbalanced by the large sample size - one that would be impossible to attain for a clinic-based cohort study or controlled clinical trial. Finally, these results cannot account for confounding by indication, in which DMARD prescription (or avoidance) may have an unaccounted-for relationship to infection risk in this population.

(top of page)

Abstract 1804 Is Congestive Heart Failure the Major Contributor of the Increased Mortality in Rheumatoid Arthritis?
Nicola, Crowson, Maradit-Kremers, Gabriel

A number of population-based studies have identified premature mortality in patients with RA, largely secondary to cardiovascular causes (myocardial infarction (MI), stroke, and congestive heart failure (CHF). However, the relative contributions of comorbid cardiovascular disease to mortality have not been explored. Here, Nicola et al examine the contribution of comorbid ischemic heart disease (IHD) and CHF to overall cardiovascular mortality in a comprehensive cohort of RA patients.

Methods: All patients in the Olmstead County, Minnesota database with RA from 1/1/1955 to 1/1/1995 were identified, paired with age- and gender-matched non-RA controls, and followed until 1/1/2001. Medical records were scrutinized for the development of clinical IHD (defined as clinical and silent MI, angina, and revascularization), CHF (by Framingham criteria), and cause of death.

Results: 603 incident cases of RA were identified (mean age 58 years, 73% women) and matched to 603 non-RA controls from the same general population based on age and gender.

The corresponding difference in CHF deaths between RA and non-RA patients was 9 excess deaths per 1000 patient-years. The corresponding difference in IHD deaths between RA and non-RA patients was 5 excess deaths per 1000 patient-years. Therefore, of the total excess CVD deaths, 63% were due to CHF and 37% were due to IHD.

(top of page)

ACR Special Abstract Vioxx Cardiovascular Safety Data from the APPROVe Study
Bresalier, Lanas, Morton, Riddell, Sandler, et al

There has been concern regarding the cardiovascular safety of rofecoxib (Vioxx^®) ever since the VIGOR study comparing rofecoxib to Naproxen for the treatment of OA, in which patients receiving Vioxx had a higher incident rate of myocardial infarctions than naproxen treated patients. Furthermore, reports indicating that Vioxx causes edema and blood pressure elevation has contributed to this concern. Just before the ACR meetings, Merck voluntarily decided to remove rofecoxib from the commercial market.

This decision was triggered by an interim safety analysis of the APPROVe (Adenomatous Polyp Prevention on Vioxx) study . This large, multicenter, randomized, double-blind, placebo-controlled trial begun in 2000 was designed to examine the effect of three years of therapy with rofecoxib 25mg per day on the recurrence of neoplastic colon polyps in patients with a history of colorectal adenomas. While the efficacy data for colonic polyp recurrence remains blinded, the cardiovascular safety data was presented for the first time at a special session of the 2004 ACR annual scientific meeting.

Methods: Subjects with a history of colorectal adenomas were randomized to receive rofecoxib 25 mg per day or placebo. Low-dose aspirin therapy was allowed in 20% of enrolled subjects. Follow-up colonoscopic examination was performed after one and three years of therapy. The primary efficacy endpoint was the incidence of pathologically proven colorectal adenoma. Safety was regularly monitored by an External Safety Monitoring Board (ESMB).

Results: 2586 patients were included in the interim safety analysis. 62% were male with a mean age of 59 years at entry.

Excess confirmed CV events in the rofecoxib group were primarily myocardial infarctions and strokes. The relative risk of a confirmed CV event over the entire study period was 1.96 (95% CI 1.20 - 3.19; p=0.007). However, this risk was not observed in a similar analysis of the first 18 months of the trial, suggesting that the elevated risk was incurred after the first 18 months of therapy. No differences in overall mortality were noted between the two groups.

Conclusion: Rofecoxib taken daily at a dose of 25 mg is associated with a two-fold higher risk of myocardial infarction or stroke after 18 months of therapy in patients with a previous history of colorectal adenoma.

Editorial Comment: In the VIGOR trial, patients were not allowed to take aspirin and some of the excess MIs were thought to be due to absence of aspirin prophylaxis in a cardiovascular vulnerable population. In addition, there was a lot of discussion after the VIGOR trial that naproxen may be cardioprotective, as opposed to Vioxx being cardio"toxic". However, the current data from the APPROVe trial again show substantial CV risk associated with a therpauetic dose of rofecoxib. Why the risk was not noted in the first 18 months is a matter of debate, but presumably the putative pro-atherogenic /pro-thrombotic effect of Vioxx exerts its effect slowly over time.

Although this trial involves a select population (patients with a previous history of colorectal adenoma), the similarity in results to the VIGOR trial suggest that it is generalizable to the broader population. At this time there is no evidence that this risk extends to celecoxib (Celebrex), although recent data on Bextra have also raised some concern. Additional close scrutiny for all COX-2 specific inhibitors is likely to follow - particularly as pertains to new COX-2 inhibitors entering the commercial market. An FDA meeting to address the safety of COX-2 specific inhibitors has been scheduled for early 2005.

(top of page)

Abstract 1764: Tumor Necrosis Factor Inhibitor Therapy Increases the Risk of Postoperative Orthopedic Infection in Patients with Rheumatoid Arthritis
J. Giles, A. Gelber, S. Nanda, S. Bartlett, J. Bathon

An increased risk of certain infections (particularly tuberculosis) is well recognized in conjunction with therapy with inhibitors of tumor necrosis factor alpha (TNF-a). Whether these agents are associated with other infections, particularly deep surgical infections after orthopedic surgery, is unknown.

Methods: Of the 546 regularly followed patients at the Johns Hopkins Arthritis Center, 271 were identified from routinely administered questionnaires as having been hospitalized, underwent any surgical procedure, or experienced a serious infection. Among these, 91 were confirmed to have undergone at least one orthopedic surgical procedure (involving instrumentation into a joint space or dissection into bone). Demographic characteristics, perioperative medications (including TNF inhibitors and directives given to discontinue TNF inhibitors prior to surgery), comorbidities, and the occurrence of post-operative infection (deep wound infection, joint sepsis, or osteomyelitis occurring within 30 days of surgery) were collected from patient records.

Results: Of the 91 patients who underwent orthopedic surgery, 35 (39%) were treated with TNF inhibitors (28 etanercept, 6 infliximab, 1 adalimumab). Demographic characteristics, clinical parameters, use of non-biologic DMARDs, type of orthopedic procedure performed, and measures of disease activity and severity were balanced between the TNF inhibitor treated and untreated groups.

10 (11%) patients experienced a deep postoperative infection (four septic arthritis, four osteomyelitis, two paraspinal abscess). A larger proportion of patients in the TNF inhibitor treated group experienced a postoperative infection compared to the TNF inhibitor untreated group (20% vs. 5%, respectively: p = 0.041) corresponding to a four-fold greater odds of postoperative infection associated with TNF inhibitor use (95% CI 1.1 18.4). This association was maintained in multivariate models adjusting for age, gender, and disease duration; use of glucocorticoids, diabetes, and RF status: or all six of these parameters.

Due to ambiguities in patient records concerning the date of last preoperative administration of TNF inhibitors, the investigators were unable to discern whether discontinuation of TNF inhibitors within a set interval of time pre- and post-surgery could ameliorate risk for infection.

Conclusion: TNF inhibitor therapy may increase the risk of early deep postoperative orthopedic infection. Whether this risk can be ameliorated by preoperative discontinuation of TNF inhibitors remains to be answered.

Editorial Comment: This investigation is the product of the Johns Hopkins Arthritis Center and the editors of this website. To our knowledge, it is the first investigation linking the prescription of TNF inhibitor therapy to an adverse surgical outcome. At the time of the first academic presentation of this abstract at the American College of Rheumatology Annual Scientific Meeting, preliminary confirmation of these results from a larger cohort of RA patients was publicly announced. These results suggest that TNF inhibitors should be discontinued prior to orthopedic surgery; though at this time, the exact period of discontinuation remains to be established. A prospective, controlled study is required to confirm these findings.

All information contained within the Johns Hopkins Arthritis Center website is intended for educational purposes only. Physicians and other health care professionals are encouraged to consult other sources and confirm the information contained within this site. Consumers should never disregard medical advice or delay in seeking it because of something they may have read on this website.