To the Editor:
The issue of whether inflammation (and thus ibuprofen) plays a significant role in the pathophysiology of cerebral forms of altitude illness (including acute mountain sickness [AMS]) is somewhat controversial. The literature has been contradictory, and Lipman et al have expressed reasonable arguments based on previously published clinical data. 1
To address their points specifically, Lipman et al wonder whether differences in headache incidence between treatment and placebo groups in our study may be minor. Furthermore, they suggest that the effects of headache suppression (via simple analgesia) may not necessarily be a large contributor to the apparent efficacy of ibuprofen in reducing the Lake Louise Acute Mountain Sickness Questionnaire (LLQ) scores. 2 Although our tables demonstrated significant differences in headache prevalence (and severity) between groups in univariate as well as multivariate analyses, we did not present results comparing differences in other subcomponents of the LLQ score because of space restrictions. With regard to each subcomponent of the LLQ score, we found differences between treatment and control in headache incidence alone; there were no significant trends in incidence or severity in any other component of the LLQ score (data not shown). Thus, the data substantiate an isolated decrease in headache pain only—a pattern potentially consistent with both sides of the discussion (but notably leaning against the inflammatory component).
Although different definitions and applications of the term intent to treat exist, we agree with Lipman et al that a careless reader may have mistakenly concluded that those lost to follow-up were included in all table entries, rather than only in table entries labeled “lost to follow-up.” Labeling the table and section “intent to treat” was shortsighted. Conducting a randomized trial in a remote region of a developing country is difficult, and we are aware of the potential bias that could be introduced by lost to follow-up and noncompliant participants. Much of the foundation of altitude medicine has been established with field-based clinical trials, and the inherent weaknesses are well known. For this reason we took pains to analyze data on participants who deviated from the protocol to the fullest extent possible, and reported results as fully as space permitted.
No statistical leaps of faith are required to realize that bigger sample sizes are better, with the important caveat that all else is equal. As in previous studies we followed established CONSORT guidelines for best practice in clinical trials; the rigor and large sample size can be combined to produce valid results that should be considered for determining the efficacy of ibuprofen in preventing cerebral forms of altitude illness.
The salient question we and Lipman et al have struggled with remains the same: in a study using observational clinical end points (ie, the LLQ), is it even possible to determine with any reasonable confidence whether a nonsteroidal anti-inflammatory drug (NSAID) such as ibuprofen can reduce a constellation of clinical symptoms either by antagonizing the underlying pathophysiology or by a simple analgesic effect (ie, an epiphenomenon)? Lipman et al have studied the clinical effect of ibuprofen to a reasonable degree (as have we); it is a reliable and reproducible finding to varying degrees in various models. After completing a subanalysis of the data, we were able to reconfirm the prior clinical effect—but have recognized we could go no farther using symptomatology as a guide. Although we and others also recognize that objective markers of such a diffuse, nonspecific clinical syndrome as hypobaric hypoxemia will likely be elusive for some time, we can also be thankful that biomarkers for inflammation are better established. In fact, an elegant blood biomarker study performed by Julian et al concluded that “resistance to AMS was accompanied by a marked anti-inflammatory and/or anti-permeability response that may have prevented downstream pathophysiological events leading to AMS. Conversely, AMS susceptibility does not appear to be related to an exaggerated inflammatory response.” 3 Although the results are complex and open to interpretation, a “smoking gun” implicating inflammation was not evident.
Two questions should be considered for future discussion: 1) if ibuprofen’s efficacy is relatable to simple analgesia, is its use completely benign, or does headache pain suppression create greater downstream discomfort or risk via rebound effect?; and 2) is it possible that ibuprofen via simple analgesia narrows the window of safety between benign and malignant forms of altitude illness such as high altitude cerebral edema? Although it is doubtful that the use of NSAIDs at altitude represents a threat based on its ubiquitous availability and use, common sense suggests that it is not wise to disable the symptom of headache, our natural “hypoxic altimeter.”