A prospective, observational, single-center study was conducted from February to June 2021 in the pediatric emergency department (ED) of our University Hospital (NCT04885335, first trial registration date 13/03/2021). Ethics approval for the study was obtained through our Institutional Ethics Committee of Limoges University Hospital (identification number 447-2021-103; first registration date 10/01/2021).
Patients aged from 1 to 17 years whose primary complaint was abdominal pain, a digestive functional sign, or a fever were recruited. Of these, patients with clinical findings suggestive of AA after an initial examination by a senior doctor or a registrar in the ED were enrolled. Patients referred to the ED for further investigation after an initial examination by a primary care physician < 7 days prior were enrolled (despite ED examination). Patients with abdominal pain for more than 7 days or with a history of prior appendicectomy were excluded.
The enrolled children were allocated into three groups according to the pre-hospital management:
-
Family group (FG): brought in by the family with no previous consultation
-
Physician group (PG): referred by a primary care physician for further investigations
-
Non-referred group (NG): admitted to the ED after a previous consultation in primary care without having been referred by the primary care physician for further investigations
The data collected in the ED were the PAS score items and its final score, characteristics such as age, gender, a record of a previous consultation in primary care < 7 days, WBC counts and CRP values determined other than at the ED < 24 hours prior or during ED hospitalization, duration of pain (< 12, 12–24, 24–48, and > 48 hours), and the presence of abdominal guarding.
With a “do no harm” principle, blood tests were prescribed only if deemed necessary. According to the PAS score, the WBC count was considered positive for levels > 10 000/mm3 and the neutrophil count for levels > 7500/mm3. For CRP, the cut-off positive value was 5 mg/L. Ultrasonography was the first-performed imaging examination. CT was restricted to second-line if ultrasonography was not contributive. It was complementary if a complicated AA like an abdominal abscess was suspected. All mesenteric lymphadenitis (LM) were defined by positive imaging. AA were confirmed perioperatively by macroscopic observation and assigned to one of two pathological groups (1. uncomplicated appendicitis: inflammatory appendix and phlegmonous appendix; 2. complicated appendicitis: perforated appendix and appendicular peritonitis).
Definitive diagnoses were obtained from the ED, radiological, and pathology reports for AA. Patients and characteristics were compared in groups according to whether or not they had a previous consultation in primary care and depending on the final diagnosis. A clinical PAS score (cPAS) was calculated without WBC and neutrophil counts, scoring a total of 8 points to represent the primary care conditions as previously described [20].
Statistical analyses were undertaken with SAS Enterprise Guide V7.1 software. A threshold of 0.05 was used to define statistical significance. For the categorical variables, the results of descriptive analyses are presented as numbers and percentages. For the quantitative variables, they are presented as means and standard deviation (SD). The PPV and NPV of PAS scores in diagnosing AA were also estimated with their 95% confidence intervals. Comparisons of quantitative variables such as PAS scores and CRP levels according to binary variables were analyzed using a Mann-Whitney test. The analysis of these variables according to the diagnosis (“AA”, “ML” or gastroenteritis (“GE”) was by an analysis of variance (ANOVA) test. Two-by-two comparisons were computed using a Dunnett adjustment. The correlation between the PAS score and the CRP level was analyzed using a Spearman coefficient correlation test.
Comparisons of qualitative variables according to groups (“FG”, PG”, or “NG”) or diagnoses were analyzed using Chi² tests or Fisher’s tests according to the applied conditions. The analysis of the change in the PAS score and the CRP level according to the duration of pain and the diagnostic group was computed with a general linear model were the explicative variables were: the diagnoses (“AA”, “GE”, or “ML”), the duration of pain (“< 12 h”, “12–24 h”, “24–48 h, “and “> 48 h”), and the interaction between them.