Can Treatment of Chronic Inflammatory Diseases Reduce the Risk of Diabetes Mellitus?

Abstract

WHEN CHEMIST FELIX HOFFMANN REDISCOVered the formula of acetylsalicylic acid in 1899, he produced a small amount to treat his father’s arthritis pain. With good results, he convinced German company Bayer to market the medication as a potent painkiller. Unknown at the time was the pleiotropism of the new drug, especially its antithrombotic potential. It would take several decades before the observation of low cardiovascular event rates in patients treated with acetylsalicylic acid revealed its beneficial effects on reducing the risk of myocardial infarction and stroke. Since then, the discovery of new jobs for “old” drugs has been the rule, not the exception, and the common expansion of drug use beyond the initial indication is not surprising. Pharmacologic interventions interfere with complex systems and their targets are typically involved in a multitude of biological functions. While the pleiotropism of medications can be problematic because of unintended adverse effects, this property offers important opportunities if a single medication provides beneficial effects for more than one disease. There are several examples in which dual effects of single medications allow physicians to address 2 clinical problems with 1 agent. For instance, duloxetine may be the preferred agent in patients who are affected by neuropathic pain and major depression because it exerts beneficial effects on both conditions. Likewise, the antihypertensive drug losartan offers advantages for treatment of patients with hyperuricemia in addition to high blood pressure. By virtue of the uricosuric properties of losartan, treatment with a separate uric acid–lowering agent may become unnecessary. Selection of drugs with pleiotropic effects appears to hold particular promise for treatment of chronic diseases that have multifaceted manifestations and associated comorbidities. For example, a substantially increased risk of cardiovascular disease has emerged as an important factor of the excess mortality in patients with rheumatoid arthritis and other chronic inflammatory diseases. As a result, rheumatologists must not only focus on treating the inflammatory component of rheumatoid arthritis but also must aggressively manage the patient’s cardiovascular risk profile. This approach often results in multiple medications for the patient and a substantial increase in the overall pill burden. Treatment strategies that target the patient’s primary disease and also help manage important risk factors and comorbidities would therefore be highly attractive. In this context, the findings reported in this issue of JAMA by Solomon and colleagues are of particular interest. In a retrospective cohort study of 13 905 patients, the authors found that among patients with 1 of 2 common chronic inflammatory diseases—psoriasis or rheumatoid arthritis— treatment with hydroxychloroquine and tumor necrosis factor (TNF)–inhibiting agents was associated with a decreased incidence of diabetes. After a first prescription for a disease-modifying antirheumatic drug (DMARD), patients were followed up and assessed for a new diagnosis of diabetes mellitus. Compared with patients treated with other anti-inflammatory medications, and after accounting for possible confounding factors such as age, sex, and several clinical variables, the hazard ratios for incident diabetes were 0.54 (95% confidence interval [CI], 0.36-0.80) for patients treated with hydroxychloroquine, 0.62 (95% CI, 0.42-0.91) for those treated with TNF inhibitors, and 0.77 (95% CI, 0.53-1.13) for those treated with methotrexate. While caution is warranted when inferring causation between specific anti-inflammatory treatments and a low risk of developing diabetes based on a retrospective cohort study, 2 factors give substantial weight to these findings: the context of prior studies on the same subject and the accumulating evidence on the biological plausibility of these observations. A significant reduction in the risk of diabetes has been reported in a previous observational study in patients with rheumatoid arthritis treated with hydroxychloroquine. Furthermore, 2 randomized controlled trials in patients with poorly controlled type 2 diabetes reported significant improvements in glycated hemoglobin with hydroxychloroquine treatment. With TNF blockade, the situation is more controversial. Studies involving patients with diabetes or metabolic syn-