Written by Rohit Mohindra and Shelley McLeod, edited by Anton Helman, November, 2023
Appendicitis, “Surgery to see in the morning” – Safe or not?
It’s just after midnight when you get the radiologist’s report for a CT scan you had ordered on a 47 year-old male who presented with RLQ pain. Unsurprisingly, it confirms acute appendicitis. At your center, the practice is to hold these patients overnight to be seen by General Surgery in the morning. Of course, practically speaking, this can mean a wide range of wait times before patients actually undergo appendectomy. You wonder if a delay to appendectomy can lead to patient harm?
The PERFECT study – appendicitis delay to appendectomy
Role of preoperative in-hospital delay on appendiceal perforation while awaiting appendicectomy (PERFECT): a Nordic, pragmatic, open-label, multicentre, non-inferiority, randomised controlled trial – ScienceDirect
This study has a lot of “research jargon” in the title, so let’s look at each component:
- Pragmatic – the study was designed to reflect day-to-day practice, not ideal conditions
- Open-label – as it was hard to blind patients or clinicians to the randomization group
- Multi-center – to show that the results are generalizable to more than one site, they included multiple sites in two countries
- Non-inferiority – the objective was to show holding patients for 24 hours with appendicitis was no worse than having them go for surgery within 8 hours – a difference in the primary outcome of less than 5% between the groups would indicate non-inferiority
- Randomized controlled trial – patients were randomized to either surgery within 8 hours or surgery within 24 hours of enrollment in the trial. Patients were also randomized to receive antibiotics or placebo while awaiting surgery.
- Inclusion – adult patients who either had an Acute Appendicitis Score ≥ 16 or confirmed uncomplicated appendicitis on diagnostic imaging were included in the study
- Exclusion – pregnant, unable to consent, fever, signs of peritonitis were excluded from the study
The primary endpoint was complicated appendicitis (ie, perforated appendicitis) assessed by the surgeon intra-operatively.
Results of this RCT on appendicitis delay to appendectomy
2095 patients were screened, and 1803 patients had surgery (907 in the 8-hour group and 896 in the 24-hour group). Dropout rates were low. The median preoperative delay from randomization to incision was 6 h (IQR 3–10) in the 8-hour group and 14 h (IQR 8–20) in the 24 hour group. However, 214 (24%) patients in the 24-hour group underwent appendectomy within 8 h.
In terms of the primary outcome, perforated appendicitis occurred in 77 (8%) patients in the 8-hour group and 81 (9%) of the 24-hour group for an absolute difference of 0.6% (−2.1 to 3.2). This was within the prespecified non-inferiority margin of 5%.
Regarding secondary outcomes, the length of stay was shorter in patients randomized to the 8-hour group. All other outcomes, including surgical or medical complications and pain scores, were not significantly different at 30 days.
The authors also conducted post-hoc analysis to look at the results based on the actual wait time for surgery and saw no difference from the intention to treat results. There was a statistically insignificant increase in perforation rate in patients operated on in 8–24 h (absolute difference 3.1%, 95% CI: –0.2 to 6.4; p=0.07).
Unfortunately, they did not report the perforation rate between the groups who received antibiotics and those who did not.
Interpretation of this RCT on appendicitis delay to appendectomy
This RCT has some concerns to address. First, there was a lack of blinding. This was understandable because of how the study was designed. However, they could have relied on the blinded pathological diagnosis instead of the intraoperative diagnosis. Second, although perforated appendicitis seems to be an important outcome, is it the most clinically relevant diagnosis? Especially if the rate of complications is low at 30 days. I think a 30-day complication rate (or even longer term) would be more patient-oriented. Finally, there was a lot of cross-over in the study (24% of patients randomized to 24 hours got operated on under 8 hours), but their per protocol analysis demonstrated no change in the results. Although a trend of increased perforations was noted, it was not part of the primary outcome, so it does warrant a closer look in future trials.
Interestingly, there isn’t any prospective randomized data to support a delay to appendectomy in uncomplicated cases. Retrospective reviews have been contradictory, and guidelines vary widely in their recommendations. Although flawed, this pragmatic trial is really the best evidence we have. Hopefully, this will prompt future trials to help us answer the question of whether or not a delay to appendectomy is safe.
The only review I could find on the role of pre-operative antibiotics reported that 45 studies demonstrated a benefit for preventing wound infection and abscess, but data on antibiotics preventing perforation while awaiting surgery is lacking.
Conclusion for “Appendicitis delay to Appendectomy – Safe or Not?”
Given your patient is clinically well and has normal vital signs, with no signs of perforation on the CT scan, you hold them in the ED for general surgery to see in the morning.
Research Methodology Hot Take on appendicitis delay to appendectomy
by Shelley McLeod
This is a fairly large trial and the authors should be applauded for their efforts, but there are a few methodological issues to consider when interpreting the results. The trial recruiters were on-duty surgeons and surgical residents who decided whether participants should proceed with the appendicectomy, but they were not blinded to group assignment. Unblinded recruiters may unintentionally introduce bias in the selection of participants, and they might inadvertently provide biased information during the informed consent process, potentially influencing participants’ decisions to enroll in the trial. The outcome assessors were also not blinded to the group assignment, so there is also the possibility of outcome assessment bias which can be particularly problematic if subjective outcomes are involved.
The authors report more than a few multiple comparisons, including two pre-specified interim safety analyses, but there is no mention of alpha adjusting/correction. When conducting multiple statistical tests, the probability of making at least one Type I error (rejecting a null hypothesis) increases, and there is a higher likelihood of finding at least one significant result purely by chance. That being said, there were no significant differences in any of the outcomes so this really isn’t a concern when interpreting the findings of this trial but is something readers should be aware of when critically appraising the results of any research study.
Patients were simultaneously randomly assigned to receive antibiotics while awaiting surgery. If antibiotics reduce perforation risk, the effect of delay on perforations might be reduced. As Dr. Mohindra pointed out, 24% of the patients in the delayed surgical group were operated on within 8 hours, and when the authors removed these patients in a post-hoc, modified per-protocol analysis, they found a small insignificant increase in perforation rate (absolute difference 3.1%, 95% CI: -0.2 to 6.4). Readers should note the upper limit of the 95% CI is higher than the authors pre-specified margin of non-inferiority of 5%.
Jalava K, Sallinen V, Lampela H, Malmi H, Leppäniemi A, Mentula P. Role of delay and antibiotics on PERForation rate while waiting appendicECTomy (PERFECT): a protocol for a randomized non-inferiority trial. BJS Open. 2021 Sep 6;5(5):zrab089.
Daskalakis K, Juhlin C, Påhlman L. The use of pre- or postoperative antibiotics in surgery for appendicitis: a systematic review. Scand J Surg. 2014 Mar;103(1):14-20.