The Effect of Hyperosmolar Water-Soluble Contrast for the Management of Adhesive Small Bowel Obstruction: A Systematic Review and Meta-Analysis : Annals of Surgery

Journal Logo

Meta-Analysis

The Effect of Hyperosmolar Water-Soluble Contrast for the Management of Adhesive Small Bowel Obstruction

A Systematic Review and Meta-Analysis

Klingbeil, Kyle D. MD, MS*; Wu, James X. MD*; Osuna-Garcia, Antonia MLIS; Livingston, Edward H. MD*

Author Information
doi: 10.1097/SLA.0000000000005573

Abstract

Of the ~330,000 patients who are hospitalized for small bowel obstruction (SBO) each year in the United States, about 1 in 4 will require surgery.1 Currently, there are no reliable methods to predict whether a patient will resolve their obstruction spontaneously. The optimal duration of observation in a stable patient prior to proceeding with an operation is also not known. Prolonged hospital length of stay (HLOS) may occur as a result, particularly for SBOs that fail initial observation and ultimately require an operation.

Water-soluble contrast (WSC) administration through a nasogastric tube (NGT) may have both diagnostic and therapeutic benefits in the management of adhesive small bowel obstruction (aSBO). As a diagnostic tool, passage of WSC into the large intestine within a specified time period predicts success with nonoperative management. Failure to do so predicts a need for an operation. Additionally, WSC may provide a therapeutic benefit as a hyperosmolar agent, drawing edema into the intestinal lumen. This effect may hasten the resolution of aSBO to reduce HLOS in nonsurgical patients and reduce the rate of surgical intervention.

Both randomized controlled trials (RCTs) and observational studies have suggested that WSC is beneficial in the management of aSBO. Yet, these studies lack sufficient sample size and contain substantial heterogeneity. Systematic reviews on the use of WSC for treating aSBO have been performed but, to date, the reviews have not summarized all of the currently available literature. Guidelines have been published, but were based on few articles on the topic.

To better understand the utility of WSC in managing aSBO, we performed a systematic review and meta-analysis of the current literature. We hypothesized that WSC improves the management of aSBO by reducing both HLOS and the rate of surgery.

METHODS

Search Strategy and Study Eligibility

The search strategies were developed by a health sciences librarian (A.O.G.) who translated the search concepts using each database platform’s syntax, including search fields and field tags. The following databases were searched using the aforementioned strategies: PubMed (includes Medline), Embase, Web of Science Core Collection, and Cochrane Reviews and Trials.

For the search terms, MeSH, Emtree, and keywords were used for the concepts of “intestinal obstruction” and “water-soluble contrast.” All concepts were combined with the “AND” Boolean operator (Supplement Digital Content: Search Strategy for detailed search strategies, https://links.lww.com/SLA/E22). A date limit was applied to each search strategy to obtain articles published beginning January 1, 1980 to November 1, 2021. The references were downloaded for deduplication, screening, and appraisal.

Studies were included if they reported the effect of WSC on SBO for HLOS and/or operative rates. Patients in these studies were required to have uncomplicated SBO thought to be secondary to adhesions manifested by abdominal pain and distention, nausea with or without emesis, and not passing flatus or stool. aSBO also required conformation by imaging. To be included, studies were required to have radiologic confirmation of SBO from presumptive adhesive disease. Studies were excluded if patients had evidence of an acute abdomen, malignant bowel obstruction, prior abdominal radiation, prior abdominal surgery within 6 weeks or abdominal (internal or abdominal wall) hernia causing an SBO. Studies without control groups were excluded. One group of patients who have aSBO treated with WSC compared with another not treated with WSC. Studies reporting outcomes from barium sulfate as the contrast agent were excluded.

Primary outcomes assessed were HLOS and rates of surgery. Secondary outcomes included time to surgery, time to resolution (first stool or first meal), rates of bowel resection, mortality rates, complication rates, and quality-of-life assessment. Objective measures used to resume feeding and measure its progress were recorded. Time from resolution of SBO to hospital discharge was also assessed as were any protocols used to determine hospital discharge eligibility.

Data Extraction and Quality Assessment

The quality of the reviewed literature was assessed by the Cochrane Risk of Bias tool for RCTs and the Newcastle-Ottawa scale for observational research.2,3 RCTs were also evaluated by the CONSORT criteria, observational studies using STROBE and each of these supplemented by additional features thought to be important when assessing studies of WSC and aSBO (Supplement Digital Content: Study Quality Assessment, https://links.lww.com/SLA/E23).

Two investigators (E.H.L. and K.D.K.) independently reviewed the studies identified for potential inclusion by the medical librarian. Disagreements were resolved by consensus or by additional review by a third investigator (J.X.W.). The following information was extracted from each study (if appropriate): trial design characteristics (randomization procedure and blinding), descriptions of the experimental and control groups, baseline characteristics of the patients, eligibility criteria for WSCS, and trial location. Sample size calculations and the rationale for selection of minimal clinically important differences (MCIDs) were reviewed. For each study, 2 reviewers (E.H.L. and K.D.K.) independently assessed the risk of bias for the outcomes of interest.

A spreadsheet was created containing data fields necessary for the Newcastle-Ottawa quality tool. One investigator filled in this spreadsheet (K.D.K.). The accuracy of the data extraction was then verified by EHL. An excel file was used to catalog data used for the Cochrane Risk of Bias 2 instrument (https://www.riskofbias.info/welcome/rob-2-0-tool/current-version-of-rob-2).

Analysis of Results

The primary analysis included only English-language studies published in peer-reviewed journals. Studies were analyzed as 2 major groups: RCTs and observational studies. For outcomes with available data (HLOS and operative rates), a random effects meta-analysis was performed to summarize treatment effects using absolute risk differences.4 A random effects model was selected because the effect of WSC on aSBO may differ from 1 center from another because of local differences in treatment protocols. Treatment effects across the RCTs were estimated and summarized on an absolute risk difference scale. Heterogeneity across the RCTs was assessed by using the I2 and τ metrics.5 HLOS and the proportion of patients undergoing surgery were displayed as forest plots. Where heterogeneity precluded meta-analysis, a qualitative synthesis of the studies was performed.

Studies were included if they had a water, saline or NGT only comparator group. Studies were analyzed if the patients had an NGT placed soon after the diagnosis of aSBO followed by the administration of WSC and abdominal radiographs. HLOS for patients not undergoing surgery who received WSC or placebo was analyzed to determine the effect of WSC as a bowel stimulant. Operative rates were similarly analyzed.

Missing data and data presented in differing formats (other than SD or interquartile range) were reconciled using Cochrane recommendations.6 If minor amounts of data were missing, it was imputed by averaging similar fields reported in studies having similar characteristics. If continuous data were reported in differing formats (eg, reporting dispersion as SEs instead of SDs), they were converted to means and SDs using the recommendations presented in the Cochrane handbook.6

All statistical tests were 2-sided and statistical significance was based on the 95% confidence intervals (CIs) excluding the null. All analyses were conducted using R version 4.1.0 and the meta and metafor packages (R Foundation for Statistical Computing, https://cran.r-project.org/web/packages/meta/meta.pdf).

Determination of the utility that further clinical trials will change the conclusions of our meta-analysis was performed using Trial Sequential Analysis (TSA).7 TSA treats each study in a meta-analysis similarly to an evaluation of data during an interim analysis of a clinical trial. In this way the addition of studies over time can be modeled to predict if performing trials in the future will change the conclusions of a meta-analysis.

The quality of guidelines was assessed by the Institutes of Medicine criteria.8

RESULTS

Study Characteristics

Of the 4879 articles obtained in the initial search, 28 were selected for full text review. After manual review of the full text articles, 8 studies were excluded (4 lacked control groups,9–12 2 used barium as contrast agent,13,14 1 used a low osmolarity WSC agent15 and 1 had overlap of patients receiving WSC between study groups16), resulting in 20 (11 RCTs and 9 observational studies) articles for the final analysis (Supplement Digital Content: Table 1, https://links.lww.com/SLA/E24). The included studies were published between 1994 and 2021 and included a total of 4761 patients and 4896 aSBO episodes. Five studies were from the United Stated, 3 from New Zealand, 2 from Iran, 2 from Israel, and 1 each from Egypt, France, Hong Kong, Italy, the Netherlands, Spain, Turkey, and Uganda. Eleven studies including 817 patients and 931 episodes were RCTs, 4 were prospective cohort studies including 860 patients and 881 episodes and the remaining 5 were retrospective cohort studies including 3084 patients.

Study Quality

Cochrane Risk of Bias Analysis for RCTs and Additional Quality Measures

None of the 11 published RCTs were accompanied by trial protocols or statistical analytic plans. Cochrane risk of bias assessment of the RCTs revealed few issues with randomization, deviation from the intended treatment and having all the outcome data available for analysis (Table 1).17–27 Most of the RCTs lacked protocols specifying exactly what treatments (such as when surgery should be performed) were rendered after WSC was administered or what criteria were used to determine hospital discharge.

T1
TABLE 1:
Cochrane Risk-of-Bias Tool for the RCTs Investigating the Effect of WSC on SBO

A review of CONSORT criteria for the RCTs is presented in Supplement Table 2, https://links.lww.com/SLA/E24. Most of the studies specified how and when WSC was administered but these treatment protocols varied. Few of the trials had clearly identified primary and/or secondary outcomes, a clearly stated a study hypothesis or MCID. The basis for sample size estimates were often vague. None of the RCTs expressed the differences between groups as absolute or relative differences with 95% CIs. Most of the articles lacked a section describing limitations and few were generalizable.

Aside from CONSORT criteria, the RCTs lacked other important study quality characteristics. Only 1 of the 11 trials specified a basic design (eg, noninferior, equivalence, or superior). Of the RCTs, only q explicitly stated a hypothesis,22 6 reported a sample size calculation and, of these, only 3 included an MCID to justify the sample size. None of the study’s outcomes were in range of the predicted event rates. Only 1 RCT was double blinded,22 with the remainder being nonblinded. Two RCTs described concealment methods,22,23 and 5 listed study dropouts and withdrawals.17,19,21–23

Newcastle-Ottawa Quality Assessment Tool for Observational Research and Additional Quality Measures

The Newcastle-Ottawa quality assessment scored 5 of the nine observational studies as having a very high risk of bias and 4 as having a high risk of bias (Supplement Digital Content: Table 3, https://links.lww.com/SLA/E24). The studies were downgraded for lack of generalizability (mostly single-center studies) and lack of precision in defining the main endpoint of length of stay and uncertainty regarding administration of the exposure intervention of WSC. Only one of the studies attempted to balance the patient baseline characteristics of the exposure and control groups. Lack of an independent blind assessment of outcomes was a major source for downgrading study quality.

Review of the enhanced STROBE instrument showed only 4 of 9 observational studies declared a study hypothesis (Supplement Digital Content: Table 4, https://links.lww.com/SLA/E24). One of these asked a discrete research question (if WSC would decrease HLOS).28 Two of the others asked multiple research questions and 1 investigated the effect of WSC administration on aSBO recurrence. All studies outlined complete eligibility and exclusion criteria but fell short in defining outcomes, confounders, and effect modifiers. Most of the studies did not address potential bias and none of them had sample size calculations.

The diagnosis of aSBO was made by CT in only one of the RCTs17 and 6 of the 9 observational studies. The remaining studies established a diagnosis of aSBO using a combination of physical examination and plain XR films. All studies examined patients who had a NGT placed immediately after the diagnosis of adhesive SBO was made.

Guidelines

Two guidelines were identified and reviewed (Supplement Table 5, https://links.lww.com/SLA/E24).29,30 Neither were based on a complete review of the available literature and lacked critical analysis of the quality of studies reviewed.

Intervention/Exposure

Gastrografin was used as the WSC agent in all RCTs and observational studies other than those by Yagci (Urografin Guerbet, Princeton, NJ)31 and Dombert (MD-Gastroview).28 The timing of WSC was inconsistent across studies. In some, it was given immediately following placement of an NGT, in others it was delayed to several days after hospital admission and, in some studies, WSC was given “as soon as practical.”

In 6 of the RCTs, WSC was given within a few hours of NGT placement. One of the RCTs reported administering the WSC after 24 hours of resuscitation23 and another after 48 hours if there was neither clinical nor radiographic improvement.18 In the remaining 4 RCTs, it was unclear when WSC was administered.

There was little consistency between the studies regarding the timing of abdominal imaging following administration of WSC. Two RCTs and 1 observational study did not specify when the images were obtained. No repeat abdominal radiographs were obtained following WSC administration in 1 RCT to maintain double blinding.22 In 7 RCTs and 4 observational studies, imaging guided the decision to perform surgery. Nine RCTs and 5 observational studies had protocols determining when surgery would be performed. Four RCTs and 2 observational studies had protocols determining when patients could be discharged from the hospital. A clearly defined primary outcome variable was defined in 5 RCTs and 3 observational studies.

Outcomes

Effect of WSC on HLOS

HLOS was used as an outcome variable in all but 1 RCT and 3 observational studies. Nine RCTs and 6 observational studies included all patients (those who did and did not undergo surgery) in the HLOS analysis. Most of the observational studies did not report dispersion data such as SDs or interquartile ranges for HLOS, precluding the ability to perform a meta-analysis. Time from resolution of SBO to hospital discharge was reported in 1 RCT17 and none of the observational studies.

Eight of the RCTs reported the HLOS for patients who did not undergo surgery. These all had similar protocols involving the administration of WSC within hours of NGT placement and follow-up abdominal radiographs within 6–72 hours. They also used similar amounts of WSC, most commonly 100 ml in total volume. These similarities justified performance of a random effects meta-analysis. The R program used for the meta-analysis and results are presented in the Supplement (Supplement Digital Content: R Coding for Statistical Analysis, https://links.lww.com/SLA/E25). Random effects meta-analysis (assuming the effect of WSC might differ from one hospital to another) showed that there was a 1.95 day (95% CI: 0.56–3.3) reduction in HLOS attributable to WSC (Fig. 1).17,19,20,22–25,27,32–34 However, there is uncertainty of this result because of the substantial heterogeneity between the RCTs (I2=80%). A funnel plot showed the possibility of publication bias but there were too few studies to be conclusive (Supplement Digital Content: R Coding for Statistical Analysis, https://links.lww.com/SLA/E25).

F1
FIGURE 1:
Forest plot demonstrating the results of the meta-analysis for HLOS in patients not undergoing surgery.

TSA showing the cumulative effect of RCTs on knowledge regarding how WSC influences HLOS for patients who do not require surgery is shown in Figure 2 and Supplement Figure 1, https://links.lww.com/SLA/E26. The Z curve crosses the O’Brien-Fleming boundary suggesting that no further studies are needed to show the use of WSC reduces HLOS for the treatment of aSBO.

F2
FIGURE 2:
Trial sequential analysis of RCTs on HLOS for patients who do not require surgery for aSBO. The curved red line with diamond-shaped icons is the O’Brien-Fleming boundary for statistical significance with the triangular component centered around the zero-line being the boundary for futility. The blue line with square icons represents the cumulative Z value and crosses the O’Brien-Fleming boundary suggesting that no further studies are needed to determine the beneficial effect of WSC on reducing HLOS for aSBO. The brown line with circular icons is the statistical significance boundary (Z=1.96) if alpha-inflation is not accounted for when performing multiple comparisons (see Supplement for more detail).

Three observational studies reported HLOS for patients who did not undergo surgery, however, dispersion data was not provided, precluding the ability to perform a meta-analysis.

Effect of WSC on Operative Rates

The proportion of patients who underwent surgery was reported in all but one RCT and in all of the observational studies. In the remaining ten RCTs, there were 429 episodes of aSBO occurring in 420 patients treated with WSC, 85 (19.8%) requiring surgery. There were 415 aSBO episodes in 409 control patients, of whom, 89 (21.4%) underwent surgery. Random effects meta-analysis showed no association between treatment with WSC or control for the number of patients who ultimately underwent surgery (risk ratio 0.90, 95% CI: 0.53–1.55) (see Supplement Digital Content: Supplement Fig. 2, https://links.lww.com/SLA/E26).

There were 9 observational studies reporting 1423 episodes of aSBO in 1408 patients who received WSC (Supplement Digital Content: Supplement Table 1, https://links.lww.com/SLA/E24). Of those treated with WSC, there were 161 (11.3%) patients who underwent surgery. There were 2542 episodes of aSBO in 2536 control patients of whom 414 (16.2%) underwent surgery. Random effects meta-analysis yielded a significant reduction in the number of operations for patients treated with WSC (risk ratio 0.56, 95% CI: 0.39–0.82) (Supplement Digital Content: Fig. 3, https://links.lww.com/SLA/E26).

Operative findings were described in 7 RCTs and 5 observational studies. Findings from the surgeries were often limited or omitted altogether. Rates of bowel resection were described in 6 RCTs and 3 observational studies. Given that rates of bowel resection are assumed to be independent of randomization, RCTs and observational studies were combined. Random effects meta-analysis yielded no significant difference in rates of bowel resection (Supplement Digital Content: Fig. 4, https://links.lww.com/SLA/E26). Negative laparotomy rates were not reported in any of the RCTs and in only 1 of the observational studies.35

Other Outcomes

Time to resolution of symptoms, defined either by first flatus/stool or first meal, was reported in 8 RCTs and three observational studies (Supplement Digital Content: Table 1, https://links.lww.com/SLA/E24). Time to surgery was reported in 3 RCTs and 5 observational studies. The definitions for initial timing varied, but included timing of admission,28 CT imaging,17 and in the remaining studies, no clear definition was given.

Overall complication rates were provided in 9 RCTs and 4 observational studies. Mortality rates were reported in 12 RCTs and in 2 observational studies. Random effects meta-analysis yielded no significant difference in both overall complications and mortality between the 2 groups (Supplement Digital Content: Figs. 5 and 6, https://links.lww.com/SLA/E26). The rate of WSC-specific complications, including hypovolemia, electrolyte imbalance and allergic reaction, were described in 8 RCTs and 3 observational studies. Among these studies, no WSC-specific complications occurred. Patient satisfaction was measured in only a single observational study.36

We could not summarize when WSC was administered relative to hospital admission because of extreme variability in the relationship between these events in the reviewed studies. Similarly, we were unable to determine the time it took for WSC to pass through to the colon because the studies lacked consistent protocols for when imaging occurred after WSC was administered (Supplemental Digital Content: Table 6, https://links.lww.com/SLA/E24).

DISCUSSION

HLOS for aSBO was associated with a reduction of nearly 2 days for nonsurgical patients treated with WSC in this meta-analysis of 11 RCTs examining outcomes for 817 patients and 9 observational studies of 3944 patients. The risk of surgery was lower in WSC-treated patients in the observational studies, an effect not observed in the RCTs. The main beneficial effect of WSC may be a reduction in HLOS resulting from stimulation of bowel motility for patients who would eventually resolve their aSBO spontaneously. The use of WSC appears to be a safe means for stimulating the bowel given that none of the studies reported major complications attributable to WSC. Although encouraging, these findings have limitations because there was substantial heterogeneity between studies, and most were intermediate to low quality. No high-quality studies have been published to definitively confirm these observations.

WSC stimulates bowel activity by 2 major mechanisms. Gastrografin, is a hyperosmolar (1900 mOsm/L) mixture of nonabsorbable sodium diatrizoate and meglumine diatrizoate that has a 6-fold osmolar gradient relative to serum (275–295 mOsm/L).18 This gradient drives water from the bowel wall into the intraluminal space, reducing edema, restoring normal blood flow, and enhancing smooth muscle contractility, thereby accelerating bowel recovery from aSBO.37 Water entering the intestinal lumen may also dilute its contents to create a pressure gradient, facilitating passage beyond an obstruction.20 Other WSC agents, including Urografin and MD-Gastroview are solutions of varying concentrations of the same compounds, demonstrating similar hyperosmotic properties.28,31

An important limitation of the WSC treatment of aSBO body of literature is the lack of standardization of protocols between studies. Some, but not all, of the studies involved the placement of an NGT shortly after hospital admission. Timing of follow-up radiographs after administration of WSC varied between studies as did how decisions were made to perform surgery for aSBO. The variation in how WSC was used in these studies contributed to the substantial heterogeneity of the meta-analysis and results in difficulties in applying the results of these studies to clinical practice.

Because the risk of bias is difficult to control in observational research, we analyzed RCTs and observational studies separately. The proportions of patients undergoing surgery differed between the study types and only in the observational studies did WSC have a statistically significant effect on reducing the proportion of patients undergoing surgery.

In the RCTs, ~20% of patients receiving WSC underwent surgery as compared to 11% in the observational studies. The operative rate in control patients in the RCTS was 21% compared with 16% in the observational studies. There are several possible explanations to resolve this resulting discrepancy. When not participating in an RCT, clinicians caring for a patient with an aSBO that is believed to fail conservative management based on clinical presentation alone, may avoid WSC administration and undergo surgery more often, an example of selection bias. Conversely, patients presenting with aSBO who are poor surgical candidates may undergo WSC administration and proceed to surgery only under emergent conditions. Taking into effect the therapeutic benefit of WSC, patients who receive WSC may have higher rates of resolution resulting in fewer operations.

There was no difference in operative rates in the RCTs, highlighting how outcomes can differ when selection bias is present within observational studies. In the RCTs, a patient’s clinical presentation would not influence the decision to perform a WSC study yielding a different outcome relative to observational studies where the clinical presentation drives the decision to pursue WSC administration.

Operative rates have been used in previous studies as an outcome to assess the therapeutic effect of WSC administration. Yet, the rate of surgery does not imply surgical intervention was actually necessary. Therefore, to truly understand the therapeutic effect of WSC for aSBO, reporting of operative findings is imperative to ensure the necessity of surgical intervention. Only 7 of the RCTs and 5 observational studies described operative findings, however, these data were often incomplete. Negative laparotomy rates should be described, in addition to rates of adhesiolysis, missed strangulation (ie, necrotic bowel) requiring bowel resection and findings that elicit an etiology other than adhesive disease (eg, obstructing malignancy or abdominal hernia). Bowel resection performed for reasons other than missed strangulation should also be clearly stated (eg, enterotomy during adhesiolysis). The rate of negative laparotomy may be a more precise outcome to assess this benefit, hypothesizing WSC would reduce the rate of negative laparotomy compared to control settings.

In order to determine WSC’s ability to predict the need for surgery, the results of the WSC study in all patients is essential. Of the RCTs, few clearly stated the failure rates for WSC studies. In 4 of the RCTs, failure rates can be inferred from operative rates as patients who failed to pass WSC within a specified time period were treated with surgical intervention by protocol. This was similarly found in 3 observational studies. In the remaining studies, the decision to operate was based on the combination of clinical and radiographic findings, with detailed reasoning omitted.

Two guidelines recommend using WSC in the management of aSBO.29,30 The World Society of Emergency Surgery developed the 2017 Bologna guideline based on 3 systematic reviews32–34 stating that nonoperative management will probably fail if contrast is not visualized in the colon within 24 hours on abdominal radiographs. The guideline mentioned, but did not explore, WSC’s ability to stimulate the bowel. A panel of 10 acute care surgeons of The Eastern Association for the Surgery of Trauma reviewed a series of articles published between 2007 and 201130 concluding that WSC testing should be considered for partial SBO that has not resolved in 48 hours because it improves bowel function, decreases HLOS, and is both therapeutic and diagnostic. Neither guideline for management of aSBO fulfilled all the Institute of Medicine (IOM) criteria for guideline quality (Supplement Digital Content: Table 5, https://links.lww.com/SLA/E24).8 Additionally, both were based on an incomplete assessment of the literature and lacked critical analysis of the quality of studies reviewed.

The Cochrane risk-of-bias and Newcastle-Ottawa are the most commonly used tools to assess study quality in systematic reviews. Both have limitations and do not account for all the potentially important aspects of study quality that should be considered. The commonly used Cochrane risk-of-bias tool for parallel group RCTs assesses five major sources of uncertainty in the results of clinical trials. They are biases that arise from the randomization process itself, deviations problems related to missing outcome data, uncertainties in the measurement of the outcome, and bias related to selecting which outcome to report. Clinical trials may perform well on these measures but may still have significant limitations resulting in uncertainty of the results. These include a clear articulation of a study hypothesis, a well thought out and justified study design (eg, noninferiority, equivalence) appropriate sample size calculation justification for why any particular MCID was selected and emphasis on reporting results that objectively summarize the findings rather than promote findings the investigator is interested in. There is also no measure of potential conflicts of interest and sources of funding for clinical trials. Thus, the Cochrane risk-of-bias instrument may miss several important aspects of study design and execution that influence study quality.

The Newcastle-Ottawa instrument for examining observational studies emphasizes patient selection, how comparable the groups are, the assessment of outcomes for cohort studies and the exposure for case-control studies. These factors provide an incomplete picture of the totality of factors that should be considered when assessing the quality of observational data. The STROBE checklist reviews more detail about the study design than does the Newcastle-Ottawa scale. Like the Cochrane risk-of-bias instrument, the Newcastle-Ottawa survey for observational studies can miss a number of important factors influencing study quality assessment. We found the granularity of the CONSORT and STROBE checklists resulted in a more complete assessment of study quality.

There are limitations of this analysis and of the body of literature describing the effects of WSC for the treatment of SBO. (1) There was substantial heterogeneity between the studies resulting, in part, from a lack of standardization of when and how much WSC was given after aSBO was diagnosed, when radiographs were obtained after WSC was administered and how decisions were made to perform surgery. Another source of heterogeneity could be the inclusion of older studies that reflect practices that were employed in the past but have changed with time. (2) There was potential bias because in the RCTs the treating clinicians were not blinded. It is conceivable that clinicians treated WSC and control patients differently because they knew the test results. (3) The RCTs summarized in this analysis had design limitations such as the lack of a clear study hypothesis, inadequate MCID estimates resulting in deficient sample size calculations, study results that differed from those predicted in the sample size calculation resulting in potentially underpowered studies, inadequate presentation of results in a form enabling comparison between studies (eg, reporting risk differences and 95% CIs), reporting both absolute and relative risks and lack of reporting of potential conflicts of interest. (4) The observational studies also had significant design limitations including lacking precise study objectives, incomplete reporting of cohort characteristics, inadequate consideration of potential sources of bias, use of odds ratios for common outcome events and inconsistent reporting of potential conflicts of interest. (5) Some patients with aSBO may be treated without NGTs. It is not known how NGT placement and WSC administration should fit into the overall scheme of aSBO management given the various options available to the patient.

Further observational studies of the effect of WSC on aSBO outcomes will not add meaningfully to decision making for the treatment of aSBO. Future studies will only be useful if they report outcomes from patients managed using protocols specifying who receives WSC, when WSC is administered and radiographs taken in relation to establishing the diagnosis of aSBO. There should also be protocols specifying how decisions are made to operate on patients and discharge them from the hospital. Studies should also report on operative findings to determine if the decision to operate was correct. There is a need for high quality RCTs comparing the effects of WSC versus control for the management of aSBO that account for all the factors outlined above for observational studies.

REFERENCES

1. HCUPnet HCaUP. Available at: https://hcupnet.ahrq.gov/ Accessed December 27, 2021.
2. Sterne JAC, Savović J, Page MJ, et al. RoB 2: a revised tool for assessing risk of bias in randomised trials. BMJ. 2019;366:l4898.
3. Wells GA, Savović B, O’Connell D, et al. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. Available at: http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp. Accessed February 3, 2022.
4. Balduzzi S, Rücker G, Schwarzer G. How to perform a meta-analysis with R: a practical tutorial. Evidence Based Mental Health. 2019;22:153–160.
5. Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med. 2002;21:1539–1558.
6. Collaboration C Cochrane Handbook for Systematic Reviews of Interventions: Data extraction for continuous outcomes. 2022. Available at: https://handbook-5-1.cochrane.org/chapter_7/7_7_3_data_extraction_for_continuous_outcomes.htm. Accessed January 19, 2022.
7. Wetterslev J, Jakobsen JC, Gluud C. Trial sequential analysis in systematic reviews with meta-analysis. BMC Med Res Methodol. 2017;17:39.
8. Cifu AS, Davis AM, Livingston EH. Introducing JAMA Clinical Guidelines Synopsis. JAMA. 2014;312:1208.
9. Chung CC, Meng WCS, Yu SCH, et al. A prospective study on the use of water-soluble contrast follow-through radiology in the management of small bowel obstruction. Aust Nz J Surg. 1996;66:598–601.
10. Lopez LM, Serra-Aracil X, Bayo HL, et al. Use of Gastrografin (R) in the management of adhesion intestinal obstruction. Cir Espan. 2013;91:384–388.
11. Nangare N, Nagur B, Biradar S, et al. A prospective study on the gastrografin in the management of adhesive small bowel obstruction. J Evol Med Dent Sci-. 2016;5:2989–2991.
12. Kapoor S, Jain G, Sewkani A, et al. Prospective evaluation of oral gastrografin in postoperative small bowel obstruction. J Surg Res. 2006;131:256–260.
13. Fevang BT, Jensen D, Fevang J, et al. Upper gastrointestinal contrast study in the management of small bowel obstruction—a prospective randomised study. Eur J Surg. 2000;166:39–43.
14. Ali M, Slack DR, Feinn R, et al. Early use of small bowel follow through reduces stay and cost in small bowel obstructions. Cureus. 2021;13:e15023 10.7759/cureus.15023
15. Brochwicz-Lewinski MJ, Paterson-Brown S, Murchison JT. Small bowel obstruction—the water-soluble follow-through revisited. Clin Radiol. 2003;58:393–397.
16. Weiss A, Sood D, Greenway SE, et al. Value of gastrografin in adhesive small bowel obstruction. Langenbecks Arch Surg. 2017;402:1233–1239.
17. Scotte M, Mauvais F, Bubenheim M, et al. Use of water-soluble contrast medium (gastrografin) does not decrease the need for operative intervention nor the duration of hospital stay in uncomplicated acute adhesive small bowel obstruction? A multicenter, randomized, clinical trial (Adhesive Small Bowel Obstruction Study) and systematic review. Surgery. 2017;161:1315–1325.
18. Choi HK, Chu KW, Law WL. Therapeutic value of gastrografin in adhesive small bowel obstruction after unsuccessful conservative treatment: a prospective randomized trial. Ann Surg. 2002;236:1–6.
19. Di Saverio S, Catena F, Ansaloni L, et al. Water-soluble contrast medium (gastrografin) value in adhesive small intestine obstruction (ASIO): a prospective, randomized, controlled, clinical trial. World J Surg. 2008;32:2293–2304.
20. Rahmani N, Mohammadpour RA, Khoshnood P, et al. Prospective evaluation of oral gastrografin(®) in the management of postoperative adhesive small bowel obstruction. Indian J Surg. 2013;75:195–199.
21. Khorshidi HR, Majidi P, Pirdehghan A. Therapeutic effect of gastrografin and predictors of operative intervention in patients with adhesive small bowel obstruction: a randomized controlled study. Turk J Surg. 2019;35:131–135.
22. Burge J, Abbas SM, Roadley G, et al. Randomized controlled trial of Gastrografin in adhesive small bowel obstruction. ANZ J Surg. 2005;75:672–674.
23. Haule C, Ongom PA, Kimuli T. Efficacy of Gastrografin(®) compared with standard conservative treatment in management of adhesive small bowel obstruction at Mulago National Referral Hospital. J Clin Trials. 2013;3:1000144.
24. Farid M, Fikry A, El Nakeeb A, et al. Clinical impacts of oral gastrografin follow-through in adhesive small bowel obstruction (SBO). J Surg Res. 2010;162:170–176.
25. Assalia A, Schein M, Kopelman D, et al. Therapeutic effect of oral Gastrografin in adhesive, partial small-bowel obstruction: a prospective randomized trial. Surgery. 1994;115:433–437.
26. Biondo S, Pares D, Mora L, et al. Randomized clinical study of Gastrografin administration in patients with adhesive small bowel obstruction. Br J Surg. 2003;90:542–446.
27. Feigin E, Seror D, Szold A, et al. Water-soluble contrast material has no therapeutic effect on postoperative small-bowel obstruction: results of a prospective, randomized clinical trial. Am J Surg. 1996;171:227–229.
28. Dombert L, Hussain A, Bullock B, et al. Impact of protocol utilizing water-soluble contrast for adhesive small bowel obstruction. J Surg Res. 2021;259:487–492.
29. Ten Broek RPG, Krielen P, Di Saverio S, et al. Bologna guidelines for diagnosis and management of adhesive small bowel obstruction (ASBO): 2017 update of the evidence-based guidelines from the world society of emergency surgery ASBO working group. World J Emerg Surg. 2018;1–13.
30. Maung AA, Johnson DC, Piper GL, et al. Evaluation and management of small-bowel obstruction: an Eastern Association for the Surgery of Trauma practice management guideline. J Trauma Acute Care Surg. 2012;73(suppl 4):S362–S369.
31. Yagci G, Kaymakcioglu N, Can MF, et al. Comparison of Urografin versus standard therapy in postoperative small bowel obstruction. J Invest Surg. 2005;18:315–320.
32. Ceresoli M, Coccolini F, Catena F, et al. Water-soluble contrast agent in adhesive small bowel obstruction: a systematic review and meta-analysis of diagnostic and therapeutic value. Am J Surg. 2016;211:1114–1125.
33. Abbas S, Bissett IP, Parry BR. Oral water soluble contrast for the management of adhesive small bowel obstruction. Cochrane Database Syst Rev. 2005;1:Cd004651.
34. Branco BC, Barmparas G, Schnüriger B, et al. Systematic review and meta-analysis of the diagnostic and therapeutic role of water-soluble contrast agent in adhesive small bowel obstruction. Br J Surg. 2010;97:470–478.
35. Zielinski MD, Haddad NN, Cullinane DC, et al. Multi-institutional, prospective, observational study comparing the Gastrografin challenge versus standard treatment in adhesive small bowel obstruction. J Trauma Acute Care Surg. 2017;83:47–54.
36. Zhou C, Udwadia FR, Chen L, et al. Using Gastrografin to manage adhesive small bowel obstruction: a nonrandomized controlled study with historical controls. Br Columbia Med J. 2021;63:247–251.
37. Vather R, Josephson R, Jaung R, et al. Gastrografin in prolonged postoperative ileus: a double-blinded randomized controlled trial. Ann Surg. 2015;262:23–30.
Keywords:

intestinal obstruction; tissue adhesions; small bowel obstruction; contrast media

Supplemental Digital Content

Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.