In part 1 of this journal club, we reviewed the utility of a clinical decision rule for the inconclusive ultrasound in appendicitis. Today we look at a meta-analysis of the Alvarado and Pediatric Appendicitis Score (PAS).
Onto the article.
Ann Emerg Med. 2014 Oct;64(4):365-372.e2. doi: 10.1016/j.annemergmed.2014.02.025. Epub 2014 Apr 14.
What are the most clinically useful cutoffs for the Alvarado and Pediatric Appendicitis Scores? A systematic review.
This systematic review included articles from a PubMed search and a search of reference lists of previous systematic reviews. Only prospective studies were included. Likelihood ratios were calculated for each of low, moderate and high-risk score groups for each score from the set of articles.
Total number of articles included for analysis: 26, with n = 3193 in adult studies and n = 4776 in pediatric studies.
The outcome measure was likelihood ratios calculated for various levels of Alvarado and PAS score cutoffs, deemed low (<4), moderate (5-7) and high-risk scores (>7) at various levels of pre-test probability for appendicitis.
The results of this systematic review found that in adults with a pre-test probability of >60%, and Alvarado score >7 suggests a very high likelihood (>85%) of appendicitis, while a pre-test probability of <50%, with an Alvarado score <4 suggests a very low likelihood of appendicitis (<3%).
In children with a pre-test probability of appendicitis >60%, an Alvarado score <4 suggests a very low likelihood of appendicitis (<3%), while a pre-test probability of >50% with an Alvarado score of >7 suggests a very high likelihood of appendicitis (>85%). The Pediatric Appendicitis Score was not as predictable at various pre-test probabilities for rule-out or rule-in of appendicitis.
Critical appraisal and journal club consensus:
The literature search in this systematic review was incomplete, searching only one database for full-text articles, whereas a more comprehensive strategy would have searched the grey literature (published abstracts, conference abstract presentations) and there should have been a more systematic approach to searching the reference lists of published reviews. Furthermore, there were many non-English articles included and there is risk of misinterpretation of results in all of these articles. Articles included had a high degree of heterogeneity of results and should not have been pooled for analysis.
While the overall objectives of the review aim to identify clinical score cutoffs that reliably predict important clinical outcomes (i.e. >85% likelihood of appendicitis and <3% likelihood appendicitis), their results may not be easily applied to ED patient care. The results are presented in an algorithm of estimation of pre-test probability combined with either Alvarado Score or Pediatric Appendicitis Score to lead to “rule-out” or “rule-in” decisions. Estimating pre-test probability is often not so straightforward in EM practice. Estimating pre-test probability is often not so straightforward in EM practice. Additionally, we are not simply “ruling-in” appendicitis at a >85% likelihood threshold and sending patients off to surgery, as the article suggests may be done, as this would result in an unacceptable 15% negative laparotomy rate. This study does not address imaging for appendicitis as an adjunct to clinical evaluation for appendicitis, as is the common practice in the ED.
Discussion at journal club emphasized that “pre-test probability” is a difficult to quantify value, and many clinicians will rely on clinical gestalt and components of the Alvarado (+/- Pediatric Appendicitis Score) when determining pre-test probability. It follows that if forming a pre-test probability of appendicitis based on components of the scoring systems, continuing on to use the same scoring system to reach a different post-test probability is not logical or clinically sound. Furthermore, the findings rely heavily on a structured clinical scoring system to “rule out” appendicitis, which is a clinical condition that can present atypically as often as it presents typically.
Thanks to Kathryn Crowder (CCFP-EM resident) and Shawn Dowling for this summary.