Design of Studies for Antibiotic Treatment of Acute Appendicitis in Children: In Support of RCTs
We thank Dr. Minneci and Dr. Deans for their interest in our article1 and agree that the patient and parents are at the center of whether nonoperative treatment for children with acute appendicitis is a viable treatment option. We read with interest their article on a patient preference feasibility study of nonoperative management for acute appendicitis in children,2 welcome the valuable information that it brings to the literature, and look forward to their reporting of 1-year outcomes. When we designed our pilot randomized controlled trial (RCT),1 there was no published literature on conservative, nonoperative treatment of acute appendicitis in children, except for the single article of Abes3 in which the diagnostic criteria used for acute appendicitis were unclear.
Pediatric surgeons are often faced with the question “Does my child really need an operation?” and we felt unable to answer this question for a child with acute appendicitis with any degree of confidence. At the Karolinska University Hospital in Stockholm, some children who were diagnosed with acute appendicitis and whose operation was delayed seemed to improve with antibiotics and were discharged without the need for operation. We initially considered reporting on this experience as a retrospective case note review; however, analysis of these data left considerable doubt regarding whether these children were comparable with those who were operated. In light of this, we decided to undertake a pilot RCT to investigate the potential treatment of acute appendicitis in children with antibiotics. We felt that it was important to undertake a prospective pilot RCT for this to provide a comparable operative group that was prospectively evaluated. As stated in our article, the specific objectives of this pilot study were to (1) evaluate the feasibility of recruiting children with acute appendicitis to an RCT comparing nonoperative treatment with appendectomy, (2) evaluate the safety of nonoperative treatment with antibiotics of acute nonperforated appendicitis in children, and (3) generate pilot data to inform our future planned efficacy trial.1 We emphasize that our study was a pilot trial. It was not designed to provide a definitive answer to our clinical question, moreover to provide the necessary baseline data that will underpin a future, large efficacy study.
Despite this, Minneci and Deans raise some interesting points which we welcome the opportunity to discuss. The balance between optimizing recruitment and achieving acceptability of a RCT to participants (and in the case of a pediatric RCT, their parents) is a challenge for researchers. To have high external validity, the proportion of the study population of interest who are actually recruited into the RCT should be high and representative. We agree that patient, family, and surgeon preference may have all hindered recruitment into our pilot trial. The benefit of having performed our pilot study is that we are now in a position to optimize recruitment technique to our future study with the knowledge gained from our pilot study. But, to have high internal validity it is also important to avoid selection bias, and this is best done with a RCT.
The influence of preference on trial outcome is more complex. Researchers should use outcomes that can be objectively measured in a standardized manner to avoid the subjective influences on outcome that are brought by a number of factors including preference. In our pilot study, 6 of 24 children in the nonoperative treatment group had an appendectomy “at surgeon and parental discretion” with no evidence of acute appendicitis on pathology. Had our study been a definitive efficacy study, this would clearly raise a significant problem for its interpretation.