Failing recruitement and more exclusions than anticipated.

Patients with acute abdominal pain and suspicion of appendicitis are common. The management of these patients is controversial with large variations between hospitals. The clinical diagnosis is regarded as insufficient necessitating the use of diagnostic techniques like ultrasound, computerized tomography and diagnostic laparoscopy. However, the role of these diagnostic techniques is not clear.

This aim of this prospective interventional study is to analyze the impact of the implementation of a clinical score to improve the clinical diagnosis and to serve as a basis for a structured management of these patients by comparison of the results after the implementation of the score with that of a baseline period. In the group of patients with indeterminate diagnosis according to the clinical score, the value of diagnostic imaging will be evaluated and compared with a period of in-hospital observation by randomization. The hypothesis is that a clinical score will decrease the use of unnecessary diagnostic imaging and unnecessary admissions to hospital for observation.

Patients with suspected appendicitis are common. Evidence based and efficient guidelines for the management of these patients are lacking. Small changes in the management of these patients may have important consequences for the health system.

The diagnosis is difficult and in up to 25% of operations a non-inflamed appendix is found. The management of patients with an equivocal diagnosis is controversial. Some advocate early surgical exploration to avoid increased morbidity following perforation, with an associated high frequency of negative explorations as an acceptable trade off. Others propose active observation, which gives an improved diagnostic accuracy without increasing the number of perforations. Early detection and treatment is more important in advanced appendicitis, which is associated with a higher morbidity and mortality, whereas spontaneous resolution is a possibility in phlegmonous appendicitis.

The management is based on the disease history, the clinical presentation and laboratory examinations. We have shown that the inflammatory variables have higher discriminating capacity than is previously thought, especially in advanced appendicitis. There are also new inflammatory markers with unknown diagnostic value. The clinical diagnosis is a subjective art, that can be made more objective by the use of a clinical scoring system. This suggests that the clinical diagnosis can be improved.

Imaging techniques, such as ultrasound (US) and Computerised Tomography (CT) have shown high sensitivity and specificity in prospective studies in specialised centres but the results in everyday use are less encouraging. Some results suggests that unselective use of imaging will lead to an increased number of appendectomies because of an increased detection of appendicitis cases that would otherwise resolve untreated. This result is in line with reports which have shown 3 to 5 times more cases of appendicitis among patients that were randomised to early diagnostic laparoscopy compared to traditional management. The impact of US on the diagnosis of appendicitis compared with traditional management has been evaluated in only one randomised trial which found no benefit from US. For CT there are three randomised small trials with conflicting results. These techniques are expensive, have limited availability and suffer from other limitations like potentially harmful ionizing radiation (CT) and examiner dependent efficacy (US). The exact role of imaging in the management of patients with suspicion of appendicitis therefore still remains to be defined.

The clinical diagnosis involves a subjective synthesis of a large amount of complex information, which relies on the surgeon's knowledge and previous experience of similar cases. This process could be improved by using a clinical scoring system, which can objectively determine the prognosis for the current patient from that of similar patients from which the scoring was constructed. This can be used as a basis for a structured algorithm for the management of the patients depending on the probability of appendicitis.

Previous scoring systems, of which the "Alvarado" score is the most well known, have not gained wide acceptance partly due to lack of performance in validation studies. Possible reasons for this are weaknesses in the construction. A scoring system should ideally be constructed from the diagnostic variables that have the strongest independent discriminating capacity in a group of patients that is similar to that which it is intended to be applied on, ie patients with a suspicion of appendicitis. Few scoring systems include inflammatory markers, which have been shown to have high discriminating power. All previous scores loose discriminating capacity because of dichotomisation of the variables. Many scoring systems also fail to use an appropriate mathematical model for the construction.

Most scoring systems have only defined one cut off point with a high sensitivity for appendicitis but an insufficient specificity to decide on operation. A more realistic approach is to define three diagnostic test zones, one with a high sensitivity for appendicitis to identify the patients that can be safely discharged with an outpatient follow up within 24 hours, another with a high specificity for appendicitis to identify patients that can be operated without further examinations, and an indeterminate group of patients which need additional diagnostic workup.

We have constructed a clinical scoring system that includes mainly objective inflammatory variables. In a validation study this score can correctly classify 63% of patients with suspicion of appendicitis to groups with high and low suspicion of appendicitis with an accuracy of 97%. This leaves 37% with an indeterminate result where the selected use of CT could play a role.

Aim of current study The aim of this study is to analyze the effects of the implementation of a clinical score as a basis for the structured management of patients with suspected appendicitis, especially for the selection of patients that may benefit from a CT examination. The study is an interventional study with a Base-line period where the current management of patients with suspicion of appendicitis is registered, and an Intervention-period where the management is based on the results of the clinical scoring system. In the Intervention phase of the study patients with an indeterminate scoring result are randomised to diagnostic imaging or repeat examination after a period of observation.

Patients All patients with age 10 years and over and less than 5 days duration of symptoms with suspicion of appendicitis at the emergency department are included in the study. In the Baseline phase the actual management of the management of the patients are registered (symptoms, laboratory values, admission to hospital, use of diagnostic imaging, surgical interventions and complications). The management of these patients is not changed. During the Intervention phase management is based on the results of the clinical score and the associated proposed algorihtm. For the groups with high and low probability of appendicitis the algorithm recommends surgical exploration or observation, respectively. Patients with a low scoring and unaltered general condition and no suspicion of other serious disease are recommended to be observed at home and followed up as outpatients within 24 hours. The algorithm is only an advice and the surgeon has the liberty to manage the patient according to his own choice or according to the local tradition, but he is asked to note the reasons for breaches in the proposed algorithm. The outcomes of the implementation of this algorithm is analysed by comparison with the results of a Baseline registration.

The outcomes to be evaluated are:

1. The incidence of patients that were admitted for observation with a discharge diagnosis of non-specific abdominal pain, compared with the Baseline phase. We expect that the intervention will lead to fewer such admissions.
2. The incidence of operations for phlegmonous and advanced appendicitis and negative explorations, compared with the Baseline phase. We expect a decrease in both the incidence of phlegmonous appendicitis and the negative explorations without an increase in the incidence of perforations.
3. The utilisation of CT, ultrasound and diagnostic laparoscopy in the groups with high and low scoring, compared with the Baseline phase. We expect that the use of imaging should decrease in these groups after the intervention.

Part B - Randomised trial Patients with an indeterminate scoring result are randomised to repeat examination after in-hospital observation or diagnostic imaging. The patients that are randomised to observation will be reassessed with the score after 6 hours of observation or earlier if the patients condition is deteriorating. At this second assessment the surgeon is free to decide on the management.

The following outcomes will be analysed:

1. The number of patients that were operated for non-perforated appendicitis. This may be higher in the imaging group as more patients with mild appendicitis that will resolve without treatment may be detected.
2. The number of negative appendectomies. This may be lower in the imaging group.
3. The number of perforations. We do not expect any difference between the groups.
4. The delay from admittance to surgery. This may be shorter in the imaging group.
5. The number of other diagnoses that need treatment. This may be higher in the imaging group.
6. Costs during the hospital stay. The costs of imaging will be balanced by an earlier discharge.

Power calculation According to the validation of the score 37% of the admitted patients are classified as indeterminate by the score. Of these 41% had appendicitis, 15% had negative appendectomies and 44% were not operated. Diagnostic imaging may decrease the number of negative appendectomies. A sample size of 686 randomised patients in each arm may detect a decrease from 15% to 10% (alfa error 0.05, beta error 0.80).

Ethical issues The proposed algorithm is close to the standard management of patients with suspected appendicitis. As an effect of the intervention this management is formalised and more evidence based than the current management. The algorithm is only an advice and the surgeon is free to diverge from this advice according to his judgment and local traditions. This means that the risk for adverse effects for the patient is small and is compensated for by the positive effects. The study protocol has been approved by the regional ethical committee (M15-09).

Inclusion Criteria:

• patients from 10 years of age with <5 days of abdominal pain and suspicion of appendicitis

Exclusion Criteria:

• pregnant women, children <10 years of age