LAUSR

To the Editor: The article by Chime et al (1), published in a recent issue of Pediatric Critical Care Medicine, which is a scoping review addressing the use of an epinephrine autoinjector (EAI) versus epinephrine drawn up for anaphylaxis management, raised some doubts and concerns about such an important subject that should deserve more attention. Thus, we have some remarks to make and issues to discuss. Although the relevant number of articles identified through the initial database search of the literature was high (n = 5,768), the resulting number of articles included in the review was small (n = 27), including four case reports/case series. In addition, only one is a randomized trial. We think that the small amount of included studies reduces the strength of the review. An important issue concerning EAI is the cost. Although the authors state that “the average price for EAIs is similar and range from $186 to $300 per single injector,” a recent medical letter from JAMA alerts for recent price increases in EAI, as well as the removal of some brands of the market (2). We also have to consider that in many settings, as in many countries, such as third world countries or even in developing countries, EAIs are not available. Though there is no dilemma on the way to administer epinephrine in these settings. The most important thing about the article is that we believe that the aim of the scope review is actually of secondary importance to anaphylaxis management. According to the World Allergy Organization, the first step in the management of anaphylaxis is to have a written protocol (3). In our opinion, if a written protocol were available in the setting, with a periodic rehearsal of healthcare providers, the issue of the use of EAI versus drawn up epinephrine would be a minor concern, since the errors of doses and route of administration would The author replies: We agree with Drs. Kumar Angurana and Kumar (1) that the issue of corticosteroids in pediatric septic shock needs to be studied in both developing and developed countries. We also agree that although including patients from both sets of countries in a single randomized controlled trial (RCT) would increase enrollment, we believe that this approach may have several limitations. As pointed out by Drs. Kumar Angurana and Kumar, empiric corticosteroid use is very frequent in the developed world (2, 3), but one cannot assume that practice patterns would be similar in resource-limited settings. Second, Drs. Kumar Angurana and Kumar also highlighted that septic shock patients in the developing world have a higher mortality rate than those in more developed countries (4). This has important implications when choosing the primary outcome measure for a septic shock RCT. Mortality may be a relevant and feasible primary outcome in developing countries, but the low mortality rate in developed countries may mandate alternative outcomes (5, 6). In addition, septic children may present late in their course in developing countries thus raising the question of the effect of timing of corticosteroid administration on outcome. Perhaps most importantly, in resource-limited settings the fundamental research question may change from “does adjunctive corticosteroids administration in septic shock improve outcomes in patients who have received antibiotics and been adequately fluid resuscitated” to “does corticosteroid administration improve outcomes independent of other therapies”? One could conduct a single RCT and stratify by country. However, we believe that stratification and the large number of potential confounders (which may not be evenly distributed post stratification and given high mortality rates) would increase the required sample size considerably and perhaps result in the loss of important region specific information. Another approach would be to conduct parallel trials in developing and developed countries. This approach would take into account the respective needs and specifications of differently resourced settings but still allow scientific collaboration, sharing of expertise and comparisons of healthcare delivery models. Finally, we would like to clarify one point raised by Drs. Angurana and Kumar with regard to our reported recruitment rate. Although the study ran for a total of 20 months, not all sites started enrolling patients at the same time. Thus, the actual recruitment rate was 0.54 patients/site/month when the numbers of sites recruiting per month was accounted for. Dr. Menon’s institution received funding from the Canadian Institutes of Health Research.