To the Editor:—
We read with interest the prospective evaluation by Hansdottir et al. 1of patient-controlled thoracic epidural analgesia versus intravenous patient-controlled analgesia. Given the important implications of their data showing no analgesic or other benefit and questioning the risk–benefit of patient-controlled thoracic epidural analgesia, we have several comments about the study.
Figure 1 indicates that 7 and 6 patients, respectively, were excluded from each group for analysis (leaving 48 and 49), although tables 1 and 2 list 55 patients in each group, and other figures and tables do not indicate “n.” Of the 55 patients in the patient-controlled thoracic epidural analgesia group, 1 died intraoperatively and 1 had surgery postponed, making intraoperative data for 55 patients unlikely.
Although 7 patients with malfunctioning catheters were analyzed as intention to treat, 3 had catheters replaced postoperatively (with inherent absence of epidural activity for a definitive period), and 4 did not have the catheter replaced. Therefore, given the study protocol for extended postoperative infusion, the absence of epidural effect for this 7–12% of the group would likely have prejudiced results of the patient-controlled thoracic epidural analgesia group.
Similarly, the 2 and 4 patients in each group who experienced confusion or stroke are included in the tabulated complications, but the authors do not discuss their ability or inability to meaningfully complete quality of recovery or analgesia scores.
Given the potentially significant findings of this study, it is important to clarify the conduct of the study for practicing anesthesiologists. Unfortunately, the reality of elective cardiac surgical practice in the United States for this relatively healthy cohort will preclude either epidural placement 1 day before surgery or a postoperative stay of almost 10 days (which would alter the power of the study as originally applied to this Swedish environment).
*University of Texas Health Science Center, Houston, Texas. firstname.lastname@example.org