• regional anesthesia
• outcomes
• complications

We greatly appreciate the interest in our study and the comments brought forward by Deflandre and Lacroix1. Indeed, our population-based analysis demonstrates that even in a regional anesthesia setting patients with obstructive sleep apnea (OSA) experience a higher risk of postoperative complications than patients without OSA. However, it was not the aim of our study to compare patients with OSA having surgery under neuraxial anesthesia to patients with OSA having surgery under general anesthesia, and therefore our study does not provide data that support the conclusion that regional anesthesia is not effective in mitigating complications in this patient group. In fact, we have previously demonstrated that patients with OSA are at higher risk of complications compared to patients without OSA2 and that regional anesthesia may represent an intervention that can effectively reduce complication risk in this population.3

The purpose of this study was to investigate if patients with OSA would have the same (low) risk of postoperative complications as patients without OSA in the setting of almost universal use of regional anesthesia. In the current study, we showed that the risk of cardiac, renal, thromboembolic complications as well as delirium remained unaltered by OSA status. However, the risk of pulmonary and gastrointestinal complications was higher in the OSA population, as were the odds for prolonged length of stay. Therefore, we concluded that the use of regional anesthesia may not per se be sufficient to eliminate excess risk to patients with OSA after joint arthroplasty, cautioning readers not to let their guard down.

For the outcome of pulmonary complications, we used a variety of ICD-9 codes, among them codes for dyspnea, apnea, and respiratory failure. These can be found in appendix I. We agree that the use of benzodiazepines or opioids may increase the risk for these events and that reducing the use of these substances would result in a lower risk of pulmonary complications. Indeed, we found evidence for an association between opioid prescription levels and complication risk in the past.4 While in the current study we did not compare the risk of pulmonary and other complications between patients with and without benzodiazepine or opioid utilization as part of regional anesthesia, we identified the use of multimodal analgesia reducing opioid use in the OSA population to be indeed beneficial and able to reduce a number of complications.5 We concur, however, that future research regarding this topic would provide further interesting information.

Furthermore, we agree that the severity of OSA is an important factor. Using ICD-9 codes for the definition of OSA, it was not possible to distinguish between mild, moderate, and severe OSA. However, with only 7.7% of patients having an ICD-9 code for OSA, we expect underreporting of this diagnosis. As patients with mild OSA may be less likely to receive a diagnosis, we expect most of the patients carrying a diagnosis of OSA in our study to have more severe symptoms. Identifying patients with OSA based on polysomnography would provide more reliable information, but this was not feasible in the setting of this retrospective cohort study.

In conclusion, we thank Deflandre and Lacroix for their valuable comments and encourage further research defining the attributable benefit that regional anesthesia can provide in order to reduce complications in patients with OSA while identifying residual risk and opportunities for improvement.