Opioid Analgesic Use and Risk for Invasive Pneumococcal Diseases

IN RESPONSE: The possibility of confounding by indication is a concern when conducting pharmacoepidemiologic studies, as discussed in our article. Although evidence from animal studies suggests that acute pain is associated with immune modulation, whether this relationship also exists among humans is not well understood (1, 2). In some studies, acute pain was linked to immunosuppression in patients who had surgery; however, whether pain or another factor related to surgery (for example, surgical trauma, hyperglycemia, or hypothermia) was the primary driver of immunosuppression remains to be determined (2, 3). Furthermore, whether pain is associated with a clinically important increased risk for infection among healthier or nonsurgical populations is unclear (3). Nevertheless, the results of our sensitivity analysis assessing whether unmeasured confounding might affect the observed results indicate that a potential unmeasured confounder (that is, pain intensity) would need to be a strong, independent risk factor for invasive pneumococcal disease and also have a large absolute difference in prevalence between opioid users and nonusers to account for the observed association. We focused on laboratory-confirmed invasive pneumococcal disease as a highly specific outcome that helped minimize misclassification. Doing so also allowed us to examine a prototypical community-acquired infection unlikely to be associated with other factors related to prescription opioid use, including recent hospitalization or injection drug use. As our article and Dublin and Von Korff's (4) editorial discussed, our results complement those of existing studies that have reported that opioid analgesic use is associated with an increased risk for all-cause pneumonia and hospitalizations for other serious infections (5).