In Reply:—
Dr. Roboubi's comments reflect a number of common misconceptions regarding clinical trials of new therapeutics. Our study was a placebo-controlled, randomized, double-blind study, which is considered the gold standard for assessing the safety and efficacy of a new single-therapeutic agent such as extended-release epidural morphine (EREM; DepoDur™; Endo Pharmaceuticals Inc., Chadds Ford, PA). Several other features of the trial are also intrinsic to the drug development process: the dose-finding nature of the study (including doses that may never be approved) and the requirement that EREM demonstrate efficacy as a single agent.
Therefore, Dr. Roboubi's comment regarding differences in general or regional anesthesia reflects a lack of understanding regarding patient randomization. The patient randomization was stratified such that the number of patients receiving general and regional anesthesia was evenly distributed across treatment groups. Had there been differences between patients receiving these two techniques, the effects of randomization would control for such differences, as would be true for other patient variables (e.g. , sex, age). However, there were no meaningful differences in the amounts of intravenous patient-controlled analgesia (PCA) used by patients receiving general or regional anesthesia: The amounts were 971.9 versus 836.1 μg, respectively (a 16% difference). Moreover, the amounts were also similar between the subsets of placebo patients who received general or regional anesthesia: 2,027.4 and 2,176.8 μg, respectively. It is precisely because of randomization and the robust nature of the data that the conclusions regarding median time to first fentanyl and total opioid use are valid, as stated in the article.
I also disagree with Dr. Roboubi's statements on the uses of fentanyl in pain management and whether patients had access to adequate analgesia. Intraoperative intravenous fentanyl administered at 250 μg is a reasonable dose for pain control in hip arthroplasty, and PCA fentanyl is widely used for postoperative pain management. At my institution, which is one of the highest volume joint replacement centers in the United States, intravenous PCA fentanyl has been the standard for many years because of its favorable metabolite profile. At equipotent analgesic doses, fentanyl produces analgesia similar to that of other opioids. Although the clinical study protocol provided guidelines for the use of PCA fentanyl, it was at the discretion of the physician to ascertain the individual needs of the patient. Patients could have been given bolus doses of fentanyl, and/or the PCA lockout interval could have been modified to address the needs of each patient.
Although the data were not presented, I also note that the study included a pharmacokinetic analysis of EREM. This would not have been possible if morphine were used for PCA instead of fentanyl; PCA morphine would have interfered with the pharmacokinetic analysis of EREM and its metabolites, whereas fentanyl did not.
Dr. Roboubi also takes issue with the number of patients receiving opioid antagonists (12.5% [17 EREM-treated patients]), but this again reflects a failure to understand the nature of a dose-finding clinical study and the inherent limitation of a study designed to focus on the safety and efficacy of a single agent. Dose-finding studies are designed knowing that some of the doses used may fall outside the range that is ultimately approved by the U.S. Food and Drug Administration, and this proved to be true in our study for the 25-mg dose and for the 20-mg dose in older patients. Therefore, it is critical for physicians to read the product label to understand the approved doses and uses.
It is also important to note that although the randomized nature of the clinical trial provides scientific validity, the environment is artificial and represents a worst-case scenario because patients are randomly assigned to the treatment drug and dosed without benefit of clinical judgment. In clinical practice, patient needs, comorbid medical conditions, and/or overall health as well as clinical practice guidelines are considered in determining an appropriate drug treatment. For example, in our study among the patients who received the Food and Drug Administration–approved doses (15 and 20 mg), three patients (< 4%) were treated for respiratory depression. Of these three patients, one was aged 75 yr and another was morbidly obese. In clinical practice, the overall health and preexisting medical conditions of these two patients would have been considered before selecting a drug treatment and dose.
It is also important to recognize that the clinical trial of EREM did not permit the use of multimodal analgesia. In clinical practice, the multimodal approach for pain management in postoperative patients is common practice and has proven quite effective for pain management. As shown by a recent meta-analysis, the use of nonsteroidal antiinflammatory drugs can reduce morphine consumption substantially and significantly decrease the rates of adverse events. In my clinical practice, patients generally receive lower doses of EREM (≤ 10 mg) in conjunction with other modalities. Under these conditions, patients frequently transition to oral medications without intravenous PCA. Regular monitoring for adverse events is performed, but the time required for monitoring is more than compensated for by the time saved on PCA setup, monitoring, and maintenance.
In summary, I thank Dr. Roboubi for providing me with an opportunity to dispel a number of common misconceptions regarding randomized clinical trials. Although such studies have limitations in describing how a drug may ultimately be used for effect in real clinical settings, the rigor of randomized studies provides us with greater certainty regarding the study outcomes. No single study can capture the many important nuances of drug performance, and clinicians should always read the product label to ensure that they understand the approved drug doses and indications.
Jefferson Medical College, Thomas Jefferson University, Philadelphia, Pennsylvania. eugene.viscusi@jefferson.edu
