Diagnostic Accuracy of Abdominal Ultrasonographic Sliding Sign in the Evaluation of Severe Intra-Abdominal Adhesions Involving the Uterus in Women Undergoing Repeat Caesarean Delivery: Systematic review and meta-analysis.

One of the most common obstetric procedures is Caesarean delivery (CD), representing approximately 30% of all births. 1 2 There has been a dramatic increase in CD rates; this can be attributed to rising multiple pregnancy rates, maturing mothers and medico-legal concerns. 3 4 5 Postoperative adhesions, a potential complication of any surgery including CDs, occur in 24–83% of cases. 6 7 Postoperative adhesions might result in small intestinal obstruction, infertility, challenging repeat surgeries and chronic abdominal pain. 8 9 10 Therefore, it is crucial to accurately diagnose the degree of pre-existing pelvic adhesions to plan subsequent operative procedures and forecast the likelihood of postoperative adhesion formation. 11 Intra-abdominal adhesions following CD are common and can be hazardous upon abdominal re-entry, often forming between the uterus and bladder or abdominal wall. The severity and scoring of intra-abdominal adhesions are usually higher with increasing CDs. In Tulandi et al .'s study involving 1,026 women, dense adhesions were significantly higher after ≥ 2 CDs (46.3% and 48.2%) than after 1 CD (29.8% and 25.6%). 12 Post-adhesion consequences include complicated repeat abdominal surgeries, bowel or bladder injury, haemorrhage, lengthier surgery, a higher chance of hysterectomy, infections and poor neonatal outcomes. 13 14 There is currently no dependable approach to predict the existence of intra-abdominal adhesions. 15 Intra-abdominal adhesions are primarily predicted by skin scar texture, degree of striae gravidarum, and uterine thickness on ultrasonography. 16 The sliding sign on real-time ultrasound may indicate severe pelvic endometriosis with high accuracy and repeatability. 17 Baron et al . extended this method for predicting substantial adhesions in women with repeat CDs. 18

No funding was received for this study.

Preferred Reporting Items for the Systematic Review and Meta-analysis of Diagnostic Test Accuracy Studies (PRISMA-DTA) guidelines were used in this systematic review and meta-analysis. 19 The search process used the following terms ‘c-section’, ‘C-Sections’, ‘Cesarean’, ‘Adhesions’, ‘Surgery Induced Tissue Adhesions’ and ‘sliding sign’. Studies were considered for eligibility if the population was gravida women with at least 1 prior CD, the diagnostic test was the absence of sliding sign in trans-abdominal ultrasonography, the reference test was surgical reports following CD and the outcome was existence of intra-abdominal adhesions involving the uterus. Intraabdominal adhesions in included studies involved any adhesions related to the access of the surgeons to the planned uterine incision (thin, filmy and easily separated adhesions by gentle, blunt or manual dissection with no vascular structures or adhesions between bowel or bladder and anterior uterine side making access to the lower uterine segment difficult and often require sharp dissection to release). Based on intraoperative findings, 4 levels of adhesions were identified: absent, mild (little or filmy adhesions), moderate (moderate to thick adhesions, require sharp dissection but do not involve bladder or bowel), and severe (absence of free space between uterine and abdominal walls or adhesions between uterus and bladder or bowel). A freely moving uterus indicated a low chance of adhesions (positive sliding sign). No uterine movement under the fascia of abdominal muscles suggested severe adhesions (negative sliding sign). US findings were compared to surgical reports following CD surgery as the reference standard test. All studies published from the inception of each database until October 2022 were eligible for inclusion. In addition, eligible study designs were retrospective and prospective observational studies. Title and abstract screening were independently performed by 2 authors (WM and TA), then full texts were downloaded and tested for eligibility by the same authors independently. A third senior author (AT) resolved any discrepancies in screening decisions. From the included studies, data was extracted and the quality was assessed [Supplementary Table 1 ]; 2 authors extracted the data independently using Excel (Microsoft Corp., Redmond, Washington, USA). The extracted data from the included studies was classified into 3 separate domains: (1) summary study characteristics; (2) baseline characteristics; and (3) diagnostic accuracy results. A third author resolved disagreements in study selection or data collection processes. Summary of included studies. 18,25-30 CD = Caesarean delivery; BMI = body mass index. Quality Assessment of Diagnosis Accuracy Studies-2 (QUADAS-2) was utilised to evaluate the risk of bias. 20 Selection of patients, the index test, the reference standard and flow and timing are 4 components of QUADAS-2 tool. The included studies were classified into low, unclear or high risk of bias; 2 co-authors independently assessed studies’ quality and discussions solved disagreements. The Grading of Recommendations Assessment, Development, and Evaluation (GRADE) method was also used to rate the strength of the evidence. 21 We performed the analysis using STATA, Version 17 (Stata Corp, College Station, TX, USA) and Comprehensive Meta-Analysis, Version 4 (Biostat, Englewood, New Jersey, USA). Meta-analytical Integration of Diagnostic Accuracy Studies and Metadta commands were used for the statistical analyses. 23 24 Flowchart of the study selection process. Primary outcomes were sensitivity, specificity, positive likelihood ratio (LR+), negative likelihood ratio (LR-) and diagnostic odds ratio (OR). In the meta-analysis, studies that reported these metrics were included. Values of I 2 ranging from 0–40%, 30–60%, 50–90% and 75–100% suggested that heterogeneity was rather insignificant, moderate, substantial and considerable, respectively. 22 A summary receiver operating characteristic (sROC) curve and area under ROC curve evaluated test performance. Post-test probabilities were shown on a Fagan nomogram. The LR test showed a P value of <0.0001 for the fitted random effects, compared to a fixed-effects model, demonstrating increased data fit by random effect model.

A total of 249 articles were identified during the initial database search. After removing 11 duplicate records, the remaining 238 articles were evaluated by title and abstract screening; 220 were excluded because they did not match the methodology-based inclusion criteria, therefore 18 remained. After reading their full texts, the meta-analysis finally included 7 articles [ Fig. 1 ]. Of the included 7 articles (from March 2018 to October 2022), 6 studies examined the sonographic prediction of intra-abdominal adhesions involving the uterus in patients having a repeat CD. 18 25 26 27 28 29 The sliding sign, stria gravidarum and Caesarean scar was used by 1 study to predict intraperitoneal adhesion in repeat CDs. 30 The study sample sizes ranged from 59–380 women who underwent at least 1 prior CD and were scheduled to undergo repeat CD. All studies were prospective observational studies. An experienced surgeon conducted surgery in 5 studies. 18 25 26 28 29 Surgeons in all included studies were blinded to the procedure [ Table 1 ]. Each study contained information about the patients’ backgrounds [ Table 2 ]. Baseline characteristics of included studies. SD = standard deviation; BMI = body mass index; CD = Caesarean delivery; NM = not mentioned. The sliding sign's specificity varied from 80–97%, and its sensitivity in detecting severe intra-abdominal adhesion ranged from 25–76%. The PPV ranged from 30–84%. However, the NPV was between 85–98%. LR+ ranged from 3–22 and LR- ranged from 0.2–0.6 [Supplementary Table 2 ]. According to QUADAS-2, all studies enrolled pregnant women with 1 or more prior CDs who underwent abdominal ultrasonographic examinations during the third trimester. In 5 included studies, the method by which patients were assigned to receive each index test was not adequately described; 18 25 26 27 30 this posed a potential bias and low bias risk in the remaining 2 studies. 28 29 For the index-test domain, ultrasound was judged unclear in three studies because the index-test results were unclear when the threshold was used. 27 28 29 The reference standard was likely to accurately classify the target condition in all trials. With respect to the time interval between the index test and the reference standard, all examined studies showed a minimal probability of bias in the flow and timing domain. The time taken between the CD and US was found to have no significant impact on the desired outcome [Supplementary Table 1 ]. Using the GRADE system, the evidence's overall quality was moderate. The quality of evidence was downgraded by 1 level because the method of patient selection in most studies was unclear. Also, 2 studies did not prescribe the threshold used for interpretation [Supplementary Table 3 ]. Regarding their relevance, all studies were found to have included patients who are pertinent to the review topic matter. For the index test, all but 1 study showed low concerns about their applicability. 26 Regarding reference-standard domains, all research exhibited minor applicability problems [ Fig. 2 ]. Risk of bias graph for included studies according to quality assessment of diagnostic accuracy studies-2 tool. For detecting severe intra-abdominal adhesions involving the uterus, the transabdominal ultrasound sliding sign had a combined pooled sensitivity (64%, 95% confidence interval [CI]: 55–71%), specificity (93%, 95% CI: 89–96%), positive likelihood ratio (9.5, 95% CI: 5.7–16), negative likelihood ratio (0.39, 95% CI: 0.31–0.49) and diagnostic odds ratio (24, 95% CI: 13–46) [Supplementary Figure 1 ]. Heterogeneity was not important for sensitivity (Cochran's Q = 7.191; P = 0.304, I 2 = 17%) and substantial for specificity (Cochran's Q = 26.418; P < 0.05, I 2 = 77%). The area under the sROC curve was 0.67 (95% CI: 0.62–0.71) [Supplementary Figure 2A ]. Prediction interval for the sensitivity was (0.444–0.786) and for the specificity was (0.711–0.985). As demonstrated in Fagan nomogram, a negative sliding sign (positive test) in women undergoing repeat CD with suspected intra-abdominal adhesions involving the uterus raised the pretest likelihood of adhesions on CD from 48% to 90%, whereas a positive sliding sign (negative test) considerably lowered it from 48% to 27% [Supplementary Figure 3A ]. After the leave-1-out test, Shu was excluded to solve the heterogeneity. 26 Pooled sensitivity (64%, 95% CI: 54–73%), specificity (94%, 95% CI: 92–96%), positive LR (10.7, 95% CI: 7.7–14.9), negative LR (0.38, 95% CI: 0.29–0.5), and diagnostic OR (28, 95% CI: 18–45) [Supplementary Figure 4 ]. Heterogeneity was not important in either sensitivity (Cochran's Q = 6.519; P = 0.259, I 2 = 23%) or specificity (Cochran's Q = 7.701; P = 0.174, I 2 = 35%). The area under sROC curve was 0.91 (95% CI: 0.88–0.93) [Supplementary Figure 2B ]. Prediction interval for the sensitivity was 0.407–0.829 and the prediction interval for the specificity was 0.865–0.972. As demonstrated in the Fagan nomogram, a negative sliding sign (positive test) in women undergoing repeat CD with suspected intra-abdominal adhesions involving the uterus raised the pretest likelihood of adhesions on CD from 48% to 91%, whereas a positive sliding sign (negative test) considerably lowered it from 48% to 26% [Supplementary Figure 3B ].

AMT, TJA and YAS handled data collection. WAM, SAAS and TJA performed the screening process and AMT resolved any conflicts. AMT performed the meta-analysis. WAM and SAAS performed quality assessment, while AMT resolved any conflicts. YAM assessed the quality of evidence using the GRADE system. WAM, AMT, SAAS, TJA and YAS drafted the manuscript. JLA, NAR, AK and AHS critically reviewed the manuscript. AMT, MA-E and YAM edited the manuscript. AHS supervised the work. All authors approved the final version of the manuscript.

Transabdominal ultrasound showed low sensitivity and high specificity in diagnosing severe intra-abdominal adhesions involving the uterus after repeat CD compared to surgical reports following CD surgery as the reference standard test. The present evidence is insufficient to determine the efficacy of transabdominal ultrasound; however, women with a low risk of adhesions may find reassurance in the existence of a sliding sign. The authors recommend to use ultrasonography to rule-out intra-abdominal adhesions affecting the uterus before CD, as it is a simple, non-invasive, practical and easily accessible technique in most clinical settings.

In this meta-analysis, the uterine sliding sign in transabdominal ultrasonography (TAS) had an acceptable sensitivity of 64% (95% CI: 55–71%) and a high specificity of 93% (95% CI: 89–96%) in detecting severe intra-abdominal adhesions involving the uterus in women with at least 1 prior CD. Adhesion prediction is based mostly on the clinical assessment of past surgeries and the number of prior CDs. Preoperative transabdominal ultrasonography, while simple, may result in proper patient counselling for complications and careful planning for safer operations. 18 Prolonged operating timeframes (from cutaneous incision to delivery and total duration from skin incision to closure of skin) and a haemoglobin decrease of more than 3 g/dL are examples of surgical complications. 29 Recent studies have shown a correlation between the negative sliding sign, time between the skin incision and delivery, and the capability to predict bleeding, indicating more difficult surgery. 27 29 Ultrasound's sensitivity and specificity in identifying uterine intra-abdominal adhesions across studies varied from 54–70% and from 90–97%, respectively. This may be related to the number of previous CDs, parity, experience of the operators and sample size. Baron et al . showed the highest sensitivity and specificity (76% and 97%, respectively). 18 This could be attributed to parity, which was highest in this study compared to the other included studies; approximately half of the sample size in the study had more than 3 previous CDs. In contrast, Shu demonstrated the lowest diagnostic performance with a sensitivity of 53% and specificity of 80%. 26 These disparities may be ascribed to variations in baseline factors such as ethnicity, body mass index and the number of prior CDs. Combining the sliding sign with the existence of a depressed scar, severe striae or both, might improve predictive accuracy. Mokhtari et al . recommended the evaluation of adhesions by incorporating the sliding sign alongside a depressed scar, which had the highest positive predictive value (92%). 30 Drukker et al . also suggested combining a negative sliding sign with a history of adhesions after CD to predict severe intra-abdominal adhesions. 29 In a recently published meta-analysis, the use of ultrasonographic visceral sliding evaluation as a rule-out assessment test was validated with an adhesion rate of 14.4% and an NPV of 99.4% with slight variation across observations. 31 This finding is in line with the current review's results that suggest the benefit of the sliding sign in excluding the existence of severe intra-abdominal adhesions without proven evidence in the diagnosis. According to the literature, evaluating the sliding sign might only require a brief training period and be repeatable by skilled operators. 32 However, the capability to perform diagnostic methods depends on expertise and not every trainee will become competent. Thus, further research is required with established standardisation of exploratory methods for clear visualisation and better sonographic performance. Additionally, reproducibility should be assessed along with the evaluation of the learning curve of trainees through a systematic training programme. To the best of the Authors' knowledge, this is the first meta-analysis to analyse TAS sliding sign diagnostic performance in identifying intra-abdominal adhesions involving the uterus PRISMA-DTA was used as were validated tools (QUADAS-2) to assess study quality. Additionally, the included studies are recent and reflect the current implementation of ultrasonographic technology advancement. Furthermore, the GRADE system was used in this review; although all included studies were observational, the GRADE body of evidence did not have a low-certainty rating at the beginning because a good quality assessment tool (QUADAS-2) was used; most of the studies assessing test diagnostic accuracy are observational in nature. However, this review was subject to certain limitations, including limited number of studies covered and their heterogeneity. Although most included trials reported that skilled surgeons performed the procedure, it was not mentioned how thoroughly the intra-abdominal adhesions were assessed. Additionally, the GRADE approach is primarily designed for interventions that might affect the quality of evidence for diagnostic testing.

The authors declare no conflict of interests.

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Supplementary data to this article can be found online at https://doi.org/10.18295/squmj.10.2024.063 .