Can we improve diagnosis of acute appendicitis?

“Diagnosis of appendicitis is usually easy”—thus wrote Sir Zachary Cope, but with the rider: “but there are difficulties which need to be discussed.”¹ The essential features of appendicitis are well known to most clinicians; there is gradual onset of central abdominal pain, often followed by vomiting, with localisation of the pain to the right iliac fossa. Localised tenderness and evidence of peritoneal inflammation (guarding and percussion tenderness) make the diagnosis probable. Clinical diagnosis is based on showing that movement between adjacent inflamed peritoneal surfaces causes pain.² Laboratory investigations usually contribute little and can be misleading. For example, the proportion of gangrenous and perforated appendixes in patients with a normal white count is the same as in those with an raised count.³ The diagnosis is essentially a clinical one—or so it would seem.

The “difficulty” alluded to by Cope relates to our inability to reliably diagnose appendicitis on clinical grounds. The vagaries of presentation and the variability of signs are such that even the most experienced surgeons may remove normal appendixes or “sit on” those that have perforated. The sequelae of delayed diagnosis may result from late presentation by the patient but are sometimes due to the initial failure of the clinician to make the correct diagnosis.⁴ The sequelae of delayed treatment include a higher incidence of postoperative sepsis and longer hospital stay. Against this, it is generally accepted that unnecessary surgery should be avoided, and this aspect of care is usually measured by the proportion of appendixes that are normal on histology. The Australian Council of Healthcare Standards has chosen this criterion as one of its clinical indicators of outcome in appendicitis.⁵

Can we improve our clinical performance? Over the years various clinical scoring systems (some computer assisted) have been used, and, although their clinical benefit has varied, most reports describe some improvement in clinical performance with their use—at least for the duration of the study. The greatest beneficiaries may be junior staff, whose diagnostic accuracy increases from 58 % to 71%.⁶ In some reports perforation rates have dropped by 50% (in one study from 27% to 12.5%), but in others no reduction has been shown.⁶,⁷ A prospective study of 118 children found that current clinical practice was more accurate than the modified Alvarado score (that measures the likelihood of appendicitis by producing a score based on various clinical and other parameters) in the diagnosis of acute appendicitis.⁸ The main value of computer aided diagnosis may be as an ongoing stimulus to good clinical practice.⁶,⁷ Despite initial optimism, it has become apparent that in most units the normal appendix rate remains 15-30%.

Can graded compression ultrasonography improve our diagnostic accuracy? In the study reported in this issue of the BMJ (p 919) the use of a diagnostic protocol incorporating both the Alvarado score and graded compression ultrasonography failed to produce better outcomes than unaided clinical diagnosis.^7a The proportion of patients in each group who had an adverse outcome (either a non-therapeutic operation or delayed treatment in patients with appendiceal perforation) was nearly identical—about 12%. Graded compression ultrasonography performed by experienced ultrasonographers still produced a 5% false negative result.

Given the frequency of both false positives and false negatives with ultrasonography, should it be allowed to override clinical judgment? Could it cause too many patients to be subjected to non-therapeutic operations (arguably unnecessary surgery) where clinical judgment might have avoided this, or could it have resulted in surgery where observation alone would have led to resolution of symptoms? In contrast, a positive result on graded compression ultrasonography may enable earlier operation in some patients with equivocal clinical signs and facilitate prompt and appropriate surgical intervention, thus reducing morbidity.

Current evidence, mostly from series of patients and retrospective studies, suggests there is probably no role for ultrasonography where clinical evidence of appendicitis is convincing, given the known false negative rate of graded compression ultrasonography and the knowledge that it may delay appropriate surgery.⁹ Moreover, the low false positive rate (6%) in clinically obvious cases of appendicitis does not warrant routine ultrasonography.¹⁰ One prospective observational multicentre study of 2280 patients found no clinical benefit when routine ultrasonography was performed in all patients.¹¹

The main role for ultrasonography may be for the equivocal case, where a combination of repeated clinical assessment and graded compression ultrasonography may provide the additional information required to determine whether surgery is necessary.¹² Finally, we should heed the advice offered by the authors in this issue that patients should not be sent home after negative results on ultrasonography unless there are also clinical grounds for their discharge. The hands of clinicians are not yet superfluous.

Papers p 919