A computerized search was conducted to identify relevant articles in PubMed on June 15, 2019. We limited our search to articles published within the last 20 years (published from 1999 onward). The following combination of medical subject headings and relevant key words in the title or abstract was used: “intussusception or invagination” and “child or pediatric or paediatric” and “ultrasound or radiography or enema or sonography or CT or US or computed or tomography.” The search was repeated on February 15, 2020. Every title or abstract was reviewed by the first author to identify key original research and review articles that addressed radiologic management of children with ileocolic intussusception. As an added measure to help confirm that all relevant articles were included, this search strategy was combined with a manual search of reference lists from identified articles. Our search resulted in 149 articles that were then reviewed in more depth. Among these articles, we extracted the articles that reported the success rates of fluoroscopy-guided hydrostatic or pneumatic enema and US-guided hydrostatic or pneumatic enema and performed a meta-analysis.

Criteria for inclusion were articles that were published during the past 20 years and reported the success rate of either fluoroscopy-guided pneumatic or hydrostatic enema or US-guided pneumatic or hydrostatic enema in children with a confirmed diagnosis of intussusception. The type of imaging guidance and the type of enema had to be clearly stated.

Criteria for exclusion were articles that did not report the type of enema used, did not report the type of imaging guidance, did not report enema reduction success rate, were in languages other than English, or were reviews. A total of 90 articles were eligible for inclusion into our meta-analysis.

Quality assessment of all studies included in the meta-analysis was done by a single reviewer. Type of study, recruitment, definition of the diagnosis, protocol of the reduction procedure (pressure, sedation, and delayed repeated attempts), definition of a successful reduction, complications, recurrences, follow-up, and surgical correlation were assessed.

CIs for individual studies were obtained via separate logistic models with an intercept and no predictors. Combined CIs were obtained via a single logistic model with a random intercept [15, 16]. This allowed each study to vary about the typical value according to a random effect, a_i, specific to the i^th study. The random effect was assumed to follow a normal distribution on the log odds scale with between-study variance v_a. Given the value of the random effect, the number of events for the i^th study, y_i | a_i, was assumed to follow a binomial distribution with n_i trials and probability p_i. Thus the two-parameter random-effects model was:

All statistical analyses outlined above were performed using SAS/STAT version 14.1 software (SAS Institute) [17]. Specific SAS procedures included GENMOD for individual studies, GLIMMIX based on numerical quadrature methods for meta-analyses, and SGPLOT for plots of individual and combined CIs. All CIs were two-sided with 95% nominal coverage. To avoid numerical problems for studies with zero events, one success and one failure were added to the data of each study before all analyses [18].

Radiography—Radiography can be useful in the diagnosis of ileocolic intussusception. The intussusception can be diagnosed by the pathognomonic crescent sign, which is the presence of a curvilinear mass within the course of the colon on radiography (Fig. 2). Additionally, there may be signs of bowel obstruction at the level of the ileocecal valve with bowel dilation and air-fluid levels within the bowel proximal to it and the absence of bowel gas distal to it [19]. However, the diagnostic accuracy of radiography for detecting ileocolic intussusception is only approximately 25% [12], and for this reason radiography is not recommended for the diagnosis of intussusception. Nevertheless, it can still be used as an initial screening procedure to identify other acute abdominal emergencies [9].

Sonography—The modality of choice to diagnose or exclude an ileocolic intussusception is typically US. Multiple studies have shown that US is a rapid and dependable modality to diagnose an ileocolic intussusception with 92–100% sensitivity and specificity [3, 7, 9, 10, 12, 20]. With US, the intussusception is identified as a mass with the appearance of a target, doughnut, or bull's eye with multiple rings of differing echogenicity in the transverse view and as a pseudokidney, sandwich, or a hayfork sign in the longitudinal view [21] (Fig. 3).

Fluoroscopy—Contrast enema has been a reference standard for the diagnosis of intussusceptions for many years. The classic signs of intussusception at the enema examination are the crescent and coiled spring signs. The crescent sign at enema examination is analogous to the crescent sign at radiography. The coiled spring sign is the result of the contrast material in the lumen of the colon filling the space between the wall of the colon and the invaginated ileum [21]. The advantage of diagnosing intussusception by enema is the possibility for immediate conversion to a therapeutic procedure. However, the excellent performance of US in the diagnosis of intussusception, lack of ionizing radiation with US, the ability to recognize alternative diagnoses with US, and a higher level of patient comfort have led many radiologists to reserve enemas for inconclusive cases of US and for therapeutic purposes [21].

CT—CT is not usually used for the diagnosis of ileocolic intussusception. However, it is an efficient imaging tool for many abdominal disorders, and unsuspected intussusceptions can be diagnosed when CT is used in pediatric patients with atypical abdominal complaints. CT appearance of intussusception includes a target sign, a sausage-shaped mass of different layers of attenuation, or a less-defined kidney-like mass. In addition to having 100% sensitivity for the detection of intussusception, CT can also be useful to diagnose pathologic lead points with higher accuracy compared with US or contrast enema [22, 23]. Additionally, CT can be used when there are concerns about peritoneal signs or in emergency situations.

After the diagnosis of ileocolic intussusception is confirmed, the child should have a surgical consultation to decide on the method of treatment. Treatment of most intussusceptions is noninvasive radiologic reduction. If noninvasive reduction is not successful, surgery is required. Children with evidence of peritonitis, shock, or perforation are not candidates for noninvasive reduction and require immediate surgical treatment.

The basic principle of noninvasive radiologic reduction of ileocolic intussusception is to increase the intraluminal pressure within the colon by enema to push out the invaginated small bowel back through the ileocecal valve. Different radiologic methods of reduction differ by the imaging modality used to guide the reduction and by the medium used for the enema. Imaging modalities to dynamically observe the status of the intussusception are fluoroscopy and US. When enema therapy is guided with fluoroscopy, a barium, water-soluble, or air enema is used (Fig. 4). When US guidance is used, a water, saline, or gas enema is used (Fig. 5). Major advantages and disadvantages of each method are summarized in Table 1.

Quality assessment—Our literature search yielded 90 articles eligible for inclusion into our meta-analysis (reporting success rates of reduction procedures). Eighty-five (94%) included studies were observational, and five (6%) were randomized controlled trials. Overall, 20 studies (22%) had consecutive recruitment, and 65 of the studies (72%) were retrospective. Almost all studies (86/90; 96%) clearly defined how they made the diagnosis of intussusception. As inclusion criteria demanded, all the studies explained the enema guidance and the type of enema used. Some studies explained the protocol in greater detail. For example, 61 studies (68%) reported the air pressure or liquid bag height pressure used, and 61 studies (68%) also clearly defined the criteria for a successful reduction. Reflux of air or fluid from the cecum into the ileum was a criterion in all of the studies that reported criteria of a successful reduction, except the studies involving US-guided pneumatic reduction, which relied mostly on the disappearance of the visualization of the intussusception. A majority of the studies (51/90; 57%) did not report on the use of sedation or general anesthesia. The use of sedation was reported by 23 studies (26%) and the use of general anesthesia by only five studies (6%). Only 25 studies (28%) reported the use of delayed repeated attempts. Reporting of results also differed between the studies. A majority of the studies reported the occurrence of perfo-ration during the procedure (75/90; 83%), recurrence of intussusception after successful initial reduction (63/90; 70%), and correlation with pathology findings in surgical cases (73/90; 81%). Finally, only 10 studies (11%) reported performing a routine follow-up examination to confirm successful reduction in the absence of clinical signs of recurrence.

Our meta-analysis had the following limitations. We included studies from all over the world, meaning that the health systems and the resources available differed greatly among different regions, which could lead to variations in the time of presentation, time of diagnosis, and time to treatment of the child with intussusception. The age groups of children included in the studies were not uniform. The use of sedation varied among the studies and its influence is not yet completely understood. Other details of each reduction procedure vary between the studies, including patient positioning, the pressure used for the procedure, and the duration of reduction attempts. Experience of the performers of the reduction procedure also varied between the studies. Additionally, most of the studies (81/90; 90%) evaluated only one of the enema reduction techniques.

The results of our meta-analysis for the probability of success of reduction of intussusception and the probability of perforation during the procedure for different radiologic methods are shown in Figures 6 and 7.

Fluoroscopy-guided hydrostatic enema—Fluoroscopy-guided hydrostatic enema was the first method of intussusception reduction. Barium was the first contrast medium used for the procedure, but its use has been steadily decreasing because of a higher risk for peritonitis in the event of perforation [24]. A recent survey revealed that for fluoroscopy-guided hydrostatic enema, radiologists in the United States currently predominantly use a water-soluble contrast medium such as cystografin (59%) or gastrografin (25%). The same survey also revealed that the method is currently used only by approximately 20% of radiologists in the country, whereas around 78% use the fluoroscopy-guided pneumatic method [14]. A great number of studies have compared the efficacy and safety of these methods over the years of use, including an extensive meta-analysis from 2015 of more than 32,000 cases that concluded that the pneumatic method was superior for reduction, with no difference for perforation [25]. Our analysis showed that the combined probability of successful reduction for this method was the lowest of all methods at 67% (95% CI, 60–73%), with a perforation rate of 2% (95% CI, 1–3%).

Fluoroscopy-guided pneumatic enema—As mentioned, fluoroscopy-guided pneumatic enema is the most commonly used method for intussusception reduction in the United States and in many institutions around the world. The method of inflating air into the colon to reduce intussusception is not a new method. Rather, it was first described in the 19th century but gained more attention in the Western world in the 1990s [26]. The major advantage of this method compared with the hydrostatic method is the higher intraluminal pressure that can be achieved, which should in theory lead to a higher reduction rate [27]. Other advantages include good visualization of successful reduction, and easy detection and a low chance of complications in case of perforation (Fig. 8). Our analysis showed that the combined probability of successful reduction for this method was 81% (95% CI, 78–84%), with a perforation rate of 1% (95% CI, 1–2%).

Ultrasound-guided hydrostatic enema—US-guided hydrostatic enema is a newer approach to the enema procedure, first described in the early 1980s [28]. It is increasingly used in the countries in which radiologists themselves perform the majority of US examinations and in which US is available throughout the day. The method's greatest advantages are a lack of ionizing radiation, better delineation of potential pathologic lead points, and possible recognition of other differential diagnoses. Our analysis showed that the combined probability of successful reduction for this method was 82% (95% CI, 79–85%), with a perforation rate of 1% (95% CI, 1–2%).

Ultrasound-guided pneumatic enema—US-guided pneumatic enema in theory represents the best of both worlds. It is a method that does not require ionizing radiation and also allows higher intraluminal pressure. The procedure is clean and easy to perform [29]. It was first described in 1990 [30] but has yet to gain widespread use. Our literature review revealed only three articles evaluating the success of the reduction of intussusception for this method within the last 20 years [29, 31, 32]. These studies report that a disadvantage of this method is a more difficult visualization of the invaginated bowel during the procedure and consequently also a more difficult confirmation of a successful reduction. This is because the confirmation of a successful reduction relies on the disappearance of the doughnut sign and the clinical sign of borborygmus, which is the sound of air passage through the small bowel that can be detected after a successful reduction. Another disadvantage of this method is the greater difficulty of detecting bowel perforation (a complication of the procedure) compared with other techniques [29–32]. We believe these disadvantages may be the reasons for the reluctance of a more widespread use of this technique and the low number of studies using it. Our analysis showed that the combined probability of successful reduction for this method was 93% (95% CI, 90–95%), with a perfo-ration rate of 1% (95% CI, 0–5%).

Another technique of reduction that has been described in the literature, although only in a single article involving 15 patients, is a technique of manual reduction by hand [33]. This technique uses US to diagnose and intermittently observe the status of the intussusception as it is reduced by massaging the abdomen in the counterclockwise direction at the point just distally of the intussusception. This technique does not use ionizing radiation and is more pleasant because it does not fill the bowel through the rectum. The disadvantage of the technique is its dependency on sedation and requirement for an anesthesiologist to be present. This technique showed a success rate of 80%. This method is not well studied and warrants further investigation [33].

In cases when the initial reduction procedure is not successful, a delayed repeated attempt should be considered when partial reduction has been achieved with the first attempt and the patient is clinically stable. Sometimes additional consecutive repeated attempts are required to fully reduce the intussusception. The delayed repeated attempts can be successful in reducing the intussusception in up to 50% of cases [31, 34].

The benefit of sedation and general anesthesia (GA) for the reduction procedure is unclear and its use greatly varies between institutions. Some authors [35] believe that the use of sedation improves the reduction rate, possibly by smooth muscle relaxation, whereas other authors [36] report that the reduction rate has been found to be lower when parenteral sedation is used because it prevents the child from performing the Valsalva maneuver during straining. They believe that the Valsalva maneuver is helpful, because it increases the intraluminal pressure and decreases the colonic transmural pressure gradient, thus improving the effectiveness of enema therapy and protecting against perforation [24]. The benefit of GA for the procedure remains even more uncertain. Already in the 1980s some authors suggested the benefit of GA [37, 38]. Additionally, a more recent study in 2012 reported that GA was associated with an increased rate of success [39]. However, to the contrary, Suzuki et al. [40] did not find any benefit of GA in their study.

Delayed repeated attempts of intussusception reduction, which have been proven to be beneficial, present a drawback for the use of GA for the reduction procedure [31, 34]. When the reduction procedure is performed under GA, surgery usually follows an unsuccessful attempt, and thus delayed repeated attempts cannot be attempted. If surgery does not immediately follow an unsuccessful attempt and a delayed repeated attempt is performed, then GA is again used, which is less than ideal given the increasing awareness of the potential morbidity associated with GA in children [41]. Another important drawback of using GA is that an anesthesiologist needs to be readily available for radiologic reduction by enema, which may not always be possible and may delay the procedure.

Sometimes ascertaining whether the intussusception has been successfully reduced is difficult, particularly if there is substantial swelling at the ileocecal valve region. A persistent intraluminal filling defect seen in the cecum on fluoroscopy-guided reduction may pose a diagnostic dilemma in differentiating an edematous ileocecal valve from residual or recurrent intussusception, or a pathologic lead point. This dilemma can be solved with a US-after reduction [42]. However, assessment can be difficult, and occasionally the dilemma remains even with the use of US [43].

The recurrence of intussusception is relatively common regardless of the method of reduction [44, 45]. The reported recurrence rate varies between 6–16% according to different authors [44, 46–49]. Approximately half of the children who develop recurrent intussusception will present within 48 hours [50]. In the case of a recurrent intussusception, a repeat reduction attempt is both safe and effective [34]. It is reasonable to perform up to three repeated noninvasive reductions, but if intussusception again recurs, surgery should follow [51].

The reasons some children have recurrences is not known. Some have a focal pathologic lead point. Many authors report that the risk of a pathologic lead point in children with recurrent intussusception is higher than in children without recurrent intussusception [46, 49, 52]. However, others reported that the rate of pathologic lead points was similar among patients with one, two, or more episodes of intussusceptions [53]. In general, the incidence of pathologic lead points in intussusception is 5–6%, although it is much higher in children over 3 years old at 30% [34, 54, 55]. Thus, a suspicion for the presence of a pathologic lead point in children with recurrent intussusception, and especially in children older than 3 years presenting with intussusception, should be raised [56, 57].

The most common pathologic lead points are a Meckel diverticulum, duplication cyst, polyps, and lymphoma [34]. Imaging can be used to detect these lead points (Fig. 9). Because intussusceptions affect mostly young children, US as a method without ionizing radiation is the optimal choice. With US it is possible to identify 66% of pathologic lead points [58]. Although US does not detect all pathologic lead points, the risk of missing one without any other signs or symptoms to guide management is unlikely [56].

According to the previously published studies, 3–8% of pediatric intussusception cases require surgical resection [39, 59, 60]. This resection rate suggests that we cannot expect the enema procedure to be successful in all cases. Additionally, it suggests we will not be able to avoid perforation as a possible complication, because ischemic or necrotic bowel wall is more susceptible for such an event. However, even if perforation occurs during the reduction procedure, it does not influence further treatment, because surgery has to eventually follow an unresolved intussusception. The use of air enema for the reduction procedure presents a lower risk for complications in such an event [22].