Prospective validation of the APPEND clinical prediction rule within a pathway dedicated to right iliac fossa pain

Abstract

Background

Right iliac fossa (RIF) pain is one of the most common reasons for acute presentation to general surgical services, and acute appendicitis is one of the most common underlying diagnoses. The clinical diagnosis of appendicitis continues to challenge clinicians and this is reflected in negative appendicectomy rates of up to 20%. Clinical predication rules (CPRs) are one method used to improve diagnostic accuracy of diagnose appendicitis and reduce negative appendicectomy rates. The APPEND score is a novel CPR that was developed at Middlemore Hospital.

Aim

The aims of this thesis were firstly to undertake a systematic literature review of published CPRs, and secondly to prospectively evaluate the performance of the APPEND CPR within a pathway dedicated to the management of RIF pain.

Methods

A systematic review of the published CPRs for the diagnosis of appendicitis in adults was undertaken. All studies that derived or validated a CPR were included and their performance was assessed using sensitivity, specificity and area under curve (AUC) values. The second part of the thesis was a comparative cohort study of the APPEND CPR within a clinical pathway dedicated to managing adults presenting with RIF pain. The primary endpoint was negative appendicectomy rate and the study was powered to detect a 7% difference. Secondary outcomes were length of hospital stay and number of radiological investigations (ultrasound and CT) performed. Staff satisfaction with the APPEND CPR was assessed by a survey.

Results

Thirty-four papers fulfilled inclusion criteria for the systematic review; 12 derived a CPR and 22 validated these CPRs. Analysis was limited by the heterogeneity and quality of included studies. The overall best performer in terms of sensitivity (92%), specificity (63%), and AUC values (0.84 -0.97) was the Acute Inflammatory Response (AIR) score but only three studies validated this CPR. Prospective evaluation of the APPEND CPR was performed on 437 consecutive adult patients presenting acutely with RIF pain to Middlemore Hospital over a 6-month period. The negative appendicectomy rate in the prospective cohort was 9.2% (95% CI: 5.3%, 13.2%) compared to 19.8% (CI 16.2, 23.4%) in the retrospective cohort that did not use the APPEND CPR. After adjusting for multiple variables, the odds of a negative appendicectomy was 2.33 (95% CI; 1.26, 4.3, P value 0.007) in the retrospective cohort. An APPEND score of ≥ 5 was 87 % specific for ruling in appendicitis (PPV 94%) and a score of ≥ 1 was 100% sensitive in ruling out appendicitis (NPV 100%). There were more US scans but no significant difference in CT scans performed in the APPEND cohort. The length of stay was 0.9 days more in the APPEND cohort (p=< 0.0001). Survey respondents reported the APPEND CPR easy to use, but response rate was only 12%.

Conclusion

Twelve CPRs for the diagnosis of appendicitis in adults have been published. The AIR score appeared to perform best but further validation is required. In a comparative cohort study incorporating the APPEND CPR the rate of negative appendicectomy was reduced by more than 50%. Further validation of the APPEND CPR, including a comparison with the AIR, would be beneficial.