Wednesday, February 15, 2006
January's Journal of Pain and Symptom Management has an interesting trial of adding midazolam to morphine to relieve dyspnea in dying cancer patients.
101 patients were randomized to one of three groups: scheduled morphine with midazolam for breakthrough; scheduled midazolam with morphine breathrough; or scheduled morphine AND midazolam with morphine for breakthrough. All drugs were given subcutaneously; doses and how they managed people who were previously on opioids etc. are spelled out in the abstract. Data was collected for 48hours, which is pretty reasonable given these were dying patients & the goal here was for rapid/immediate dyspnea relief. Results generally were solidly in favor of scheduled morphine plus midazolam and the scheduled midazolam only group fared the worst. More patients in the combo group had dyspnea relief at 24 hours compared to the others (92% vs. 69% for scheduled morphine vs. 46% for scheduled midazolam); "breakthrough" dyspnea occured less in the combo group (non-significant trend). All groups improved markedly on median dyspnea scores (from ~8/10 to ~2/10 at 48 hours). This suggests 1) regression to the mean aside, all interventions were helpful, and 2) the benefit from the addition of scheduled midazolam is probably mostly in obtaining more immediate relief in a subset of patients early on. Of note, the scheduled morphine group had significantly more 'clinically relevant toxicity' than the combo group.
Results aside, this is a remarkable study for several reasons, and it's great to see this sort of work being done. This was a randomized trial of dying patients on a palliative ward (all had advanced cancer, ECOG performance status of 4, and life expectancies of <1 week), and had a decent n. Unfortunately it was only single blinded although there didn't seem to be any methodologic reason why it had to be. Additionally, it would have made a lot of sense to have a morphine ONLY group since that is, I suspect, the usual practice for many of us (I, personally, don't immediately reached for benzodiazepines especially if patients are doing well on opioids alone). (Also, it is unclear to me how they used the data on those that died during the 48 hours of data collection--ignored it? used intention-to-treat analysis? used as much data as was collected?). That aside I want to restate that this was a decently designed and executed study, with clinically useful findings, that was performed on the sickest patients imaginable--it can be done.