Impact of emergency physician performed ultrasound in the evaluation of adult patients with acute abdominal pain: a prospective randomized bicentric trial

Non-traumatic abdominal pain is one of the most common complaints in patients visiting the emergency department (ED) [1]. It represented 6.5% of patients admitted in a US ED in 2007 [1]. Many etiologies may be involved in adult patients presenting with acute abdominal pain: surgical (appendicitis, bowel obstruction…), medical (diverticulitis, cholelithiasis, gastritis, renal colic, urinary tract infections …) but also non-specific abdominal pains [2]. It therefore remains a daily diagnostic challenge for emergency physicians (EP). Diagnosis accuracy relies on improvements in the advanced imaging tools such as computed tomography (CT) or ultrasound (US) [1, 3]. Currently, the diagnostic workup in a patient presenting with acute abdominal pain, is based on clinical examination and if necessary, laboratory tests and, in many cases, imaging procedures.

Beside ultrasound performed by the radiology department, point-of-care ultrasound (POCUS) is increasingly used to assess patients in emergency medicine (EM) for many years [4]. It has become an integrated part of the EM curriculum [5]. US is particularly suited to assess patients with acute abdominal pain as many organs are easily explored [6]. Diagnostic performances, both sensitivity and specificity, are higher for cholelithiasis, renal colic and small bowel obstruction when compared with other illnesses such as diverticulitis. Lindelius demonstrated that a surgeon-performed US was able to increase diagnosis accuracy in patients with acute abdominal pain [7]. Furthermore, in this same study, POCUS decreased short-term complementary examinations [8] and increased patient’s satisfaction. It was demonstrated that POCUS could decrease the hospitalization rate ( − 7%), imaging prescriptions ( − 18%) and increase the direct orientation toward surgery by 18% without increasing rates of rehospitalization or death [9]. Another study showed that POCUS increased the diagnostic accuracy and planned diagnosis workup by 45% [36–54%] in a population of 128 patients admitted to the ED with abdominal pain without previous diagnostic orientation [10]. This improvement was achieved by comparing diagnosis before and after POCUS realization. POCUS can also be used as a first-line imaging procedure followed by a computed tomography if necessary. This strategy has been considered as the most accurate according to sensitivity and exposure to radiation [2]. POCUS could therefore be considered as an extension of the clinical examination and can increase the whole diagnostic performance, in particular in some illnesses as cholelithiasis or renal colic.

However, despite these few studies, the added value of POCUS performed by an emergency physician on patients with non-selected abdominal pain remains controversial. Actually, (i) it has rarely been evaluated by randomized controlled trial (RCT), (ii) US performances have only been demonstrated for some conditions such as renal colic, cholelithiasis, appendicitis or bowel obstruction and (iii), previous studies were performed by highly trained Emergency Physicians.

Here, we conducted a RCT to investigate the added value of early POCUS on the accuracy of the preliminary diagnosis made by the emergency physician before any radiologist-performed imaging study in patients presenting to the ED with acute abdominal pain [11].

Between June 2021 11th and June 2022 23th, 256 patients were randomized, but five were not included in the primary analysis (lack of valid consent), leaving 125 patients in the POCUS group and 126 patients in the usual care group (Fig. 1). Baseline characteristics are displayed in Table 2. Briefly, there was 130 women and 121 men, median [Q1-Q3] age: 42 [30;57]. Baseline characteristics were similar between the 2 groups. Pain visual analogic scale was 7 [5‐8] in POCUS group and 7 [5‐8] in the control group (median [Q1-Q3]). According to the adjudication committee, the most frequent diagnoses were: non-specific abdominal pain (69 patients, 28%), renal colic (30 patients, 12%), diverticulitis (18 patients, 7%), gastroenteritis (18 patients, 7%), cholecystitis (17 patients, 7%), appendicitis (16 patients, 6%), pyelonephritis (15 patients, 6%) and cholelithiasis (14 patients, 5%).

Modified intend-to-treat analysis showed no difference for the exact diagnosis between the two groups (POCUS group: 70/125, (56% [95% CI 47–64%]) versus usual care group: 78/126 (62% [95% CI 53–70%]), RD 1.23 [95% CI 0.74–2.04]). NPV and PPV in the POCUS group were 79% [69%; 86%] and 62% [42%; 79%] respectively, and in the control group, 84% [75%; 90%] and 53.6% [33.9%; 72.5%] respectively. There was no difference in the accuracy for the diagnosis of non-specific abdominal pain between POCUS and control groups, sensitivity 47% versus 48% and specificity 87% versus 86%, respectively. There was no difference in numbers of laboratory tests, CT and US performed by the radiologists (Table 3). Time from door to diagnosis were 3.8 + 2.2 and 4.0 + 2.5 h in POCUS and control groups, respectively (p = 0.78). ED length of stay ED were 9.9 + 6.5 h and 10.0 + 6.3 h in POCUS and control groups, respectively (p = 0.57). Finally, there was no difference between the two groups for hospitalization rate (POCUS group: 38/123, (31% [95% CI 23–40%]) versus control group: 45/126 (36% [95% CI 28–44%]), absolute risk difference (RD) 0.8 [95% CI: 0.47–1.37]) nor for readmission to the ED at day-7 (POCUS group 3/113, 3% [95% CI 0.6–8%]) versus control (3/111, 3% [95% CI 0.6–8%]), RD 0.98 [95% CI 0.19–5.02]). Duration of POCUS was 8 + 4 min and the self-assessed difficulty on a Likert scale ranging from 1 (very easy) to 10 (impossible) was 2 + 2.

In this bicentric randomized controlled study, systematic POCUS performed by an EP did not improve the rate of exact diagnosis in adult patients presenting to ED with acute abdominal pain. This particular endpoint, rate of exact diagnosis in patients with unselected abdominal pain, has not been frequently addressed. Lindelius [7] showed a positive effect on diagnostic performance with an improvement of exact preliminary diagnosis from 57 to 65%. In a population with similar inclusion criteria, Durgun [13] found that POCUS was able to narrow the number of suspected diagnoses, to reduce the ED length of stay, but not the whole cost. When comparing population between our study and the one from Lindelius, there was no obvious differences. Distribution of final diagnosis and diagnostic accuracy in the control group, were similar. The lack of positive effect of POCUS on diagnostic accuracy could be explained by several factors: (I) insufficient POCUS skills of investigators. They were diverse even if a validated training session was required to be an investigator. However, duration and modalities of these training programs could be different” In the study from Lindelius, surgeons had a 4-week training session which is longer than the ones of some of our investigators. (ii) POCUS could be performed by an investigator who was not the EP in charge of the patient. Therefore, US findings were possibly not fully integrated with other clinical findings. In addition, we did not observe a decrease in US performed by a radiologist ordered by the emergency physician in charge of the patient in the POCUS group compared to the control group. This result contrasted with Lindelius study in which less radiological US were performed in the PoCUS group. It could be related to the lack of integration of PoCUS findings in the diagnostic process or to the lack of confidence of investigators.

Other studies investigated the diagnostic accuracy of POCUS in specific presentations such as suspected bowel obstruction [14, 15], acute flank pain [14], suspected appendicitis [15]. In all these studies, POCUS improved the diagnostic accuracy. Another approach was the diagnosis or management changes when POCUS was performed. Jang [10] showed that POCUS improved the decision making process by 45% [CI 95% 36–54%].

Our study had some limitations: (I) varied POCUS skills of investigators; (ii) Overestimation of the potential diagnostic improvement induced by PoCUS, (iii) The experts reviewed the patient’s file together which could introduce bias and (iv) US clips quality were not adjudicated since it should have require a full review process.

Despite these limitations, it was a prospective randomized control study without major deviation. Moreover, POCUS was harmless as there was no difference in safety criteria (same hospitalization and 7-day readmission rates between the two groups). Indeed, abdominal US, including POCUS, only provides useful insights in some etiologies (cholelithiasis, renal colic, appendicitis, small bowel obstruction). This could explain the modest but real POCUS effect in the study from Lindelius (7%) and its absence in the whole population of the current one. A potentially interesting strategy could be a clinical evaluation followed by a POCUS only in case of suspected diagnoses accessible to POCUS. A CT would be required in case of inconclusive POCUS or clinical situation requiring definite diagnosis.

Our study did not demonstrate a positive effect of POCUS in the diagnostic process of adult patients with abdominal pain. Other diagnostic strategies including PoCUS could be tested for patients consulting for acute abdominal pain. Further research should focus on the accuracy of strategies including PoCUS only if an etiology where it is accurate, are suspected. This approach was investigated with some success [2] but still need confirmation [16].