In this retrospective study we found that age, WBC and DS on admission were independently associated with complicated appendicitis, and developed a prediction model based on these three independent predictors, aiming to make the discrimination of simple and complicated appendicitis in children younger than five years of age. Regarding prediction, the prediction model could identify children at high risk for complicated appendicitis, better than that of ALVARADO score and PAS. This model might be used to aid the differentiation of acute simple and complicated appendicitis for the optimal treatment strategy.
AA remains a clinical diagnosis with laboratory and radiological test as an auxiliary diagnostic method. Accurate differentiation between simple and complicated appendicitis is emerging as a potentially key issue as the historical standard of care, that is prompt appendectomy, is increasingly questioned in pediatric patients[7, 16]. Since AA has a rate of been complicated of approximately 40%, different methods for predicting complicated appendicitis have been tested with inconsistent results. Radiological tests and ultrasonography prove to have an approximately 20% of false negative complicated appendicitis. Both clinical and laboratory variables have been reported to be of value in diagnosing complicated appendicitis, but the results are equivocal in children younger than five years of age[7, 13, 17–19].
This study not only describe the independent risk factors for complicated appendicitis, but establish early identification of risk factors in order to predict complicated appendicitis. Thus, we included only those factors available in clinical database that were simple and easy to obtain. Based on the multivariate regression analysis results, we referred the Enter method to establish the prediction model. Even though DS were discussed in previous studies as well as in ours, we should notice that the factor is of subjective nature and its reproducibility is low. Objective variables obtained from blood sample usually better reproducible and therefore of higher value. Among the variables included in our prediction model, DS is the only modifiable risk factor. Several studies have shown that longer DS of AA, the more likely it was to develop perforated [20–23]. Bickell et al. reported the link between the duration of the symptoms and the probability of appendiceal perforation. They concluded that the chance of perforation is low in the first 36 hours of the disease and increases by 5% every 12 hours thereafter. We found a notable difference in the DS between the simple appendicitis and complicated appendicitis, which is why concluded that one of the reasons for high rates of complicated appendicitis in this age group could be a delayed visit to the doctor. Similar to our results, Bansal et al. revealed notable differences in the DS between the groups of perforated and non-perforated appendicitis. However, we thought that due to the lack of intestinal barrier and underdeveloped omentum in children younger than five years of age, the DS had a more obvious effect on the appearance of gangrene and perforation in AA. This reminded us that shortening the DS may effectively avoid the probability of complicated appendicitis.
According to the requirements of the international transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD) list and elaboration documents, the new prediction model needs to be verified by validation samples of the center or other centers in order to truly reflect the prediction performance of the model. We collected clinical data of 156 cases for external verification, the discrimination was evaluated by calculating the AUC of ROC. When the cutoff point was 0.62, the AUC for the prediction model in validation sample was 0.830 (95% CI: 0.762–0.885). Our prediction model was shown to have a sensitivity of 77.8%, a specificity of 89.2%, a PPV of 88.7%, and an NPV of 77.6%. The diagnostic accuracy of our model in this cohort was high. In the 2 most commonly cited scores (ALVARADO score and PAS), the authors assign point values to patient history, physical examination, and laboratory findings. In several studies, PAS and ALVARADO score could effectively diagnose complicated appendicitis[7, 25–27], but no research reported in patients younger than five years of age. We compare the predictive model with PAS and ALVARADO score for the differentiation of simple and complicated appendicitis. The prediction model had an AUC greater than that for the ALVARADO score or PAS in validation sample (P < 0.05). This may suggest that the ALVARADO score and PAS were not accurate enough to differentiate the type of AA in patients younger than five years of age. Therefore, the prediction model we made was a simple and efficient method that aids the differentiation of acute simple and complicated appendicitis.
Perforation in this age group often leads to diffuse peritonitis, and the most important thing in the management is to establish the accurate diagnosis and perform surgical treatment, assisted by broad-spectrum antimicrobial therapy[2, 21, 28]. Recently, several trials have focused on the non-operative treatment for AA[10, 29–31]. Studies suggested that different treatment strategies should be selected according to the type of AA: simple appendicitis should be the preferred antibiotic conservative treatment, while complicated appendicitis requires appendectomy in most cases[15, 32]. Children appendix is not a non-functional organ left in the body. The appendix is not only a “storage pool” for the gut microbiota to balance the steady state of the proinflammatory and anti-inflammatory activities of the intestine; and the high content of lymphoid tissue (mainly lymphocyte CD8 + T cells) in the appendix plays an important role in the immune function of the body[33, 34]. The age of 5 years and younger is an important period for children's immune function to gradually mature and the balance of intestinal flora to establish. Conservative treatment for simple appendicitis can preserve the appendix, which not only helps maintain intestinal flora homeostasis and immune system development, but also reduces medical costs[16, 35].
Furthermore, discrimination between simple and complicated appendicitis is important as it may guide appropriate intravenous fluid and antibiotic resuscitation prior to surgical intervention. The prediction model could guide preoperative (or postoperative) antibiotic selection and predict prognosis, referred the optimal cutoff point of 0.62. Although appendicitis protocols vary widely among centres, children with simple appendicitis typically receive a single antibiotic preoperatively and may even not receive postoperative treatment and get discharge home relatively soon. Conversely, children with a complicated appendicitis recognised on admission typically receive a combination of more antibiotics, undergo operative treatment, and continue antibiotic therapy postoperatively. Hence, identification of predictive indicators for the complicated appendicitis is essential.
It should be borne in mind that the present study was limited by its retrospective design and based on experiences within a single unit, further research with a larger prospective cohort study is necessary to validate the usefulness of the prediction model for predicting complicated appendicitis in children younger than five years of age. Furthermore, the definitions of simple and complicated appendicitis are based on the intraoperative findings and postoperative pathological results, and nonoperatively were excluded.