Definition and Prevalence

Gastroparesis is a condition in which patients experience symptoms of delayed gastric emptying in the absence of an actual physical blockage.¹ The most common symptoms are nausea, vomiting, early satiety, bloating, abdominal pain, and postprandial fullness.² Assessing gastric emptying delay is essential to diagnosing gastroparesis. In clinical research, the definition of gastroparesis is delayed gastric emptying as detected by clinical testing and the presence of symptoms of nausea and/or vomiting, postprandial fullness, early satiety, bloating, or epigastric pain for more than 3 months. Using this definition, the cumulative incidence of gastroparesis is 4.8 percent in people with type 1 diabetes, 1.0 percent in people with type 2 diabetes, and 0.1 percent in people without diabetes, who may have idiopathic gastroparesis or other etiologies.² A 2007 community-based study estimated the prevalence of gastroparesis to be 9.6 per 100,000 for men and 37.8 per 100,000 for women.² Newer estimates of prevalence report a higher rate of 24.2 per 100,000 inhabitants. Some experts estimate that more than 1.5 to 3 million Americans may have gastroparesis.^3,4

Etiology and Clinical Course

The etiologies of gastroparesis are most often idiopathic, diabetic, or postsurgical, but can also be autoimmune, paraneoplastic, or neurologic. The condition is generally assessed in the outpatient setting, but some patients become severely ill with intractable vomiting and dehydration and are hospitalized. Hospitalizations for gastroparesis increased by 158 percent between 1995 and 2004.⁵ In individuals with diabetes and gastroparesis, digestion of food is unpredictable, and wild swings in blood glucose can increase morbidity and necessitate medical care.

Evaluation of Possible Gastroparesis

A standard assessment for patients with typical symptoms (e.g., nausea, vomiting, bloating, abdominal pain, early satiety) of gastroparesis starts in the office of a physician, who takes a careful medical history and performs a physical examination.⁶ First, the physician must rule out mechanical or medication-related dysfunction. Medications that commonly cause gastric emptying delay are opiates or glucagon-like peptide agonists. Second, the physician needs to test for gastric emptying. Methods of testing include gastric emptying scintigraphy, antroduodenal manometry, and now wireless motility capsule (WMC) technology. Motility disorders are difficult to diagnose. Multiple contributing factors make pathophysiology more complex, and physicians can have difficulty gathering a unifying diagnosis from a single test. In addition, most of the available tests have some inconsistency in performance, which can make their interpretation difficult.

Use of Gastric Emptying Testing To Guide Treatment

Effective gastric-emptying-delay testing guides physicians in their recommendations for nutrition, medication, and surgical therapies. Testing informs physicians about the length and severity of delay, and this information can guide changes in diet to accommodate better gastric emptying. Recommended changes in diet may include a lowfat diet, a low-residue diet (i.e., low fiber, easy to empty from the stomach), a liquid diet, or changing one's consumption pattern to multiple small meals per day. Testing can also inform physicians about the use of prokinetic medicines like metoclopramide or erythromycin, which are often used to treat gastroparesis. This is important because of the FDA black box warning about the side effects of using metoclopramide for more than 3 months. Both metoclopramide and erythromycin can cause profound tachyphylaxis, limiting any intended benefit. Similarly, domperidone (Motilium^®) is not FDA-approved but is available in many countries outside the United States and is used in clinical care and research in the United States through an Investigational New Drug Application. Therefore, clear documentation of gastroparesis is important to physicians who are considering using a prokinetic. Patients with severe symptoms and severe emptying delay despite dietary changes may need feeding tubes, such as jejunostomy or gastrojejunostomy tubes, that bypass the stomach entirely. As patients undergo consideration for compassionate use of gastric stimulation therapy, one of the eligibility criteria is the presence of gastric emptying delay on testing. Thus, accurate diagnosis of gastroparesis is integral to decisions about management.

Outcomes

Major outcomes of interest are assessment of motility and diagnosis of gastric emptying delay. Other outcomes include the ability of testing to influence treatment decisions (e.g., changes in medications, nutrition), or to affect patient-centered outcomes (e.g., symptom improvement, need for surgery, quality of life, patient satisfaction). It is important to consider potential harms of testing such as capsule retention, radiation exposure, and mortality. Clinicians and policymakers may also be interested in the effects on resource utilization, such as the need for additional tests, physician services, or hospitalizations.

Definition and Prevalence

Constipation is common, occurring in 15 to 20 percent of the U.S. population.^11,14,15 Multiple professional societies define constipation (with slight variation) as fewer than two bowel movements per week or a decrease in a person's normal frequency of stools accompanied by straining, difficulty passing stool, or passage of hard solid stools.¹¹ Physicians must assess patients with symptoms of constipation via their medical history and a physical examination to exclude malignant or organic causes of constipation. Clinicians should ask about warning signs such as new onset of symptoms, obstructive symptoms, rectal bleeding, unintentional weight loss, or family history of early colon cancer. A rectal examination can help to delineate rectal function and tone and exclude a low rectal cancer. Clinicians should perform a colonoscopy on all patients over 50 who have never received a screening colonoscopy, and those who have fecal occult blood, iron deficiency anemia, or any other warning signs.¹⁶ However, the yield of colonoscopy in patients with constipation with warning signs is low. Once a physician has eliminated all organic causes for constipation, a diagnosis of functional constipation is appropriate. Physicians do not need to test an individual less than 50 years old and without “red flag” symptoms in order to diagnose constipation if the patient meets the Rome III criteria.

The Rome III criteria define functional constipation as follows:¹⁷

1. Two or more of the following:
   1. Straining during at least 25 percent of defecations
   2. Lumpy or hard stools in at least 25 percent of defecations
   3. Sensation of incomplete evacuation for at least 25 percent of defecations
   4. Sensation of anorectal obstruction/blockage for at least 25 percent of defecations
   5. Manual maneuvers to facilitate at least 25 percent of defecations (e.g., digital evacuation, support of the pelvic floor)
   6. Fewer than three defecations per week
2. Loose stools rarely present without the use of laxatives
3. Insufficient criteria for irritable bowel syndrome

A patient must have two or more of the above criteria for the last 3 months, with symptom onset being at least 6 months prior to diagnosis.

Clinically, patients with slow-transit constipation, also known as colonic inertia, often have the most severe symptoms of those patients with constipation, with prolonged periods of time between bowel movements. Often, standard medical therapies have failed these patients. The definition of slow-transit constipation is retention of greater than six radiopaque markers after 5 days from ingestion.^11,18 The reported incidence of slow-transit constipation is 1 in 3,000 or 0.033 percent. Other studies list a prevalence of 0.17 percent.¹⁹ The true incidence is likely unknown.

Etiology and Clinical Course

There are several types of chronic constipation including slow-transit, normal-transit, and dyssynergic defecation. There is also constipation-predominant irritable bowel syndrome.¹¹ Physicians should recommend lifestyle changes and medical management for all patients with symptoms of constipation. Lifestyle changes include drinking appropriate quantities of liquid, removing all possible offending medications, and eating the U.S. Department of Agriculture's recommended amount of vegetables, fruit, and fiber. Medical management includes avoiding constipating medications and initiating bulking agents (e.g., fiber supplements), stool softeners (docusate, mineral oil), osmotic and stimulant laxatives (e.g., lactulose, milk of magnesia, magnesium citrate, polyethylene glycol [Miralax^®], PEG-3350, senna), or prokinetics (e.g., bisacodyl), and secretagogues/prokinetics (e.g., lubiprostone, linaclotide), or in other countries prucalopride (not yet FDA-approved), as indicated. Thus, the initial evaluation of constipation symptoms does not often involve colonic transit testing.

Evaluation of Possible Slow-Transit Constipation

For certain individuals with suspected slow-transit constipation, colon transit testing can provide valuable insight into the etiology of the constipation. Testing can explain why a patient fails basic therapy and can help identify or exclude patients as surgical candidates.¹¹ However, a single test may not reflect the full complexity of a patient's motility disturbances. For example, anorectal dysfunction can impact colonic transit, but must be assessed by anorectal manometry separate from other transit testing. Furthermore, most of the available tests have some inconsistency in performance, which makes their interpretation difficult in some cases. Transit disorders include slow colonic transit or colonic inertia, a hypomotile disorder of the colon where transit in the proximal colon is slow without evidence of retropulsion of the markers from the left colon and without evidence of anorectal dysfunction. Defecatory dysfunction (or functional outlet dysfunction) is the presence of uncoordinated motion of the anorectum muscles causing ineffective or weak expulsion of stool. Idiopathic megacolon (primary or secondary), a pathological enlargement of the colon, can also be present and may occur in conjunction with longstanding neurological diseases or Hirschsprung's disease, a failure of the development of the nerve cells within the colon wall.²⁰ The main diagnostic methods used to test for colonic motility are radiopaque marker (ROM) examination, colonic scintigraphy, colonic and anorectal manometry, and WMC testing.^21,22 The nonreference standard is ROM.

Use of Colon Transit Testing To Guide Treatment

Most patients with chronic constipation see symptom improvement with medical therapy and/or lifestyle changes. For some patients, all measures fail and physicians must use colon transit testing to better understand the motility disorders. Physicians use anorectal manometry to identify anorectal or outlet dysfunction, and treat with biofeedback therapy. Evidence of Hirschsprung's disease is an indication for surgical segmental resection. Megacolon requires medical therapy tailored to reducing gas formation, and reduction of fiber intake may paradoxically relieve symptoms. If these conservative measures fail, megacolon may require segmental or total colectomy. If testing confirms the presence of slow-transit constipation (colonic inertia) without the use of laxatives, then the next step in evaluation in some centers is transit testing with use of laxatives. Physicians should only consider surgery as a potential therapy after they have demonstrated colonic inertia.²⁶ Clear demonstration of severe total or segmental slow-transit constipation is an indication for colectomy; however, most clinicians reserve colectomy for patients with the most terminal or untreatable conditions.

Outcomes

A major outcome of interest to clinicians is the ability to characterize transit time and to diagnose slow-transit constipation. Other outcomes include the ability of testing to influence treatment decisions (e.g., change in medications, change in nutrition) or to affect patient-centered outcomes (e.g., symptom improvement, need for surgery, quality of life, patient satisfaction). It is important to consider potential harms such as capsule retention, radiation exposure, and mortality. Clinicians and policymakers may also be interested in the effects on resource utilization such as the need for additional tests, physician services, and hospitalizations.

Literature Search Strategy

We searched the following databases for primary studies for the periods in parentheses: MEDLINE^® (1966 to July 1, 2012) and Embase^® (1974 to July 1, 2012). We developed a search strategy for MEDLINE, accessed via PubMed, based on an analysis of the medical subject headings (MeSH) terms and text words of key articles identified a priori. Additionally, we reviewed the reference lists of included articles and any relevant review articles. We asked the manufacturer of WMC about any published or unpublished randomized controlled trials or observational studies that evaluated WMC. The manufacturer submitted comments on the draft report but did not submit any new materials. We searched ClinicalTrials.gov to identify any relevant trials.

Study Selection

Two independent reviewers evaluated each title, abstract, and full article. We included studies that compared WMC with other diagnostic tests among patients with suspected gastroparesis or slow-transit constipation, in terms of diagnostic accuracy, accuracy of motility transit time assessment, effect on treatment decisions, effect on patient-centered outcomes, effect on resource utilization, or harms. Other diagnostic tests were gastric scintigraphy, antroduodenal manometry, and endoscopy for the evaluation of gastroparesis, and scintigraphy and ROM for slow-transit constipation. There were no language restrictions. We resolved differences between investigators regarding eligibility through consensus adjudication.

Data Abstraction

We created and pilot tested standardized spreadsheets for data extraction. The study investigators performed double data abstraction on each article. The second reviewer confirmed the first reviewer's abstracted data for completeness and accuracy. We formed reviewer pairs that included personnel with both clinical and methodological expertise.

For all articles, the reviewers extracted information on study characteristics (e.g., study design, country, location of recruitment, start year of recruitment, multicenter vs. single center, length of followup, length of time in between diagnostic tests), characteristics of study participants (e.g., condition; age; gender; race; weight; prior diagnostic tests; blood sugar; smoking status; diabetes status; defecatory dysfunction status; and the use of prokinetics, opiates, antidepressants, proton pump inhibitors, and laxatives), eligibility criteria, characteristics of WMC testing (e.g., was the pill swallowed or placed; did the study provide a standardized meal; did the study provide Ensure^® shakes, and if so, when?^a), characteristics of the other diagnostic tests, outcome measures, definitions, and the results of each outcome, including measures of variability. For each of the diagnostic tests, we collected information on the criteria used to make a diagnosis of gastroparesis or slow-transit constipation, and on whether the study instructed patients to abstain from tobacco, prokinetics, opiates, antidepressants, proton pump inhibitors, or laxatives at the time of the test.

Quality Assessment

Two reviewers independently assessed article quality. We selected and modified the questions from the QUADAS-2 quality assessment tool.²⁷ We supplemented this tool with quality-assessment questions (i.e., to assess spectrum bias) based on recommendations in the Methods Guide for Medical Test Reviews.²⁸ Our quality assessment included items on: (1) whether the study excluded healthy subjects from the diagnostic accuracy comparison, (2) whether the study excluded severely affected patients, (3) whether the study enrolled a random sample of patients, (4) whether all patients received the same reference standard, (5) whether the study included all patients in the analysis, (6) whether the study interpreted results of the test independently, (7) whether the time period between tests was reasonably short (within 3 months) to ensure that the condition did not change, (8) whether the study established cut-off values for test positivity before the study started, (9) whether a stated aim of the study was to compare diagnostic accuracy between WMC testing and other diagnostic tests, (10) whether the study reported on conflicts of interest, (11) whether a commercial source related to motility testing funded the study, and (12) whether a commercial source related to motility testing employed or gave funding or fees to any of the authors. The two reviewers resolved differences in quality assessment.

Applicability

We assessed the applicability of studies in terms of the degree to which the characteristics of the study population (e.g., age, etiology, comorbidities, prior surgery or gastric pacer), diagnostic test procedures (e.g., use of opiates during testing, use of bowel motility-altering agents such as laxatives or prokinetic agents), outcomes, and settings (e.g., referral center) were typical for the treatment of individuals with suspected gastroparesis or slow-transit constipation.

Data Analysis and Synthesis

We had planned to conduct meta-analyses if sufficient data were available (at least five studies for hierarchical summary receiver operator characteristic curves for diagnostic accuracy and at least three studies for other outcomes) and if studies were sufficiently homogenous with respect to key variables (e.g., population characteristics, study duration, diagnostic test procedures). We qualitatively summarized studies not amenable to pooling.

We considered gastric scintigraphy and clinical symptoms to be reference standards and ROM to be a nonreference standard. For measures of diagnostic accuracy when there was a reference standard, we summarized the results in terms of sensitivity, specificity, and test concordance. For measures of diagnostic accuracy when there was a nonreference standard, we summarized the results in terms of positive percent agreement, negative test agreement, and test concordance.²⁹ When the reference standard was a clinical diagnosis, we chose a 10 percent difference between tests in sensitivity or specificity as a potentially important difference because key studies were powered to detect a 10 percent difference.²⁵ When the reference/nonreference standard was another diagnostic test, we considered it similar if WMC had a test concordance of at least 80 percent.

We conducted a sensitivity analysis where we included data that was reported only in a conference abstract.

Rating the Body of Evidence

At the completion of our review, we graded the strength of the available evidence addressing the KQs by adapting an evidence grading scheme recommended in the “Methods Guide for Medical Test Reviews”²⁸ and in the “Methods Guide for Effectiveness and Comparative Effectiveness Reviews.”^30,31 Both of these evidence grading schemes are based on recommendations of the GRADE Working Group.³² We applied evidence grades to the bodies of evidence about each diagnostic test comparison for each outcome. We assessed the strength of the available evidence by assessing the risk of bias, consistency, directness, and precision.

We classified evidence pertaining to the KQs into four basic categories: (1) “high” strength of evidence or SOE (indicating high confidence that the evidence reflects the true effect and that further research is very unlikely to change our confidence in the estimate of the effect); (2) “moderate” SOE (indicating moderate confidence that the evidence reflects the true effect and that further research may change our confidence in the estimate of the effect and may change the estimate); (3) “low” SOE (indicating low confidence that the evidence reflects the true effect and that further research is likely to change our confidence in the estimate of the effect and is likely to change the estimate); and (4) “insufficient” SOE (indicating that evidence is unavailable or does not permit a conclusion).³²

Search Results

Figure B summarizes the results of our literature search. Our search retrieved 2,028 unique records. After reviewing the titles and abstracts, we considered 142 articles as potentially relevant and we reviewed the full text of the article for eligibility. We included a total of 12 studies (in 18 publications) in this review.^11,25,33-42 Seven studies (10 publications) evaluated WMC among patients with gastroparesis^33-39 and nine studies (14 publications) evaluated WMC among patients with slow-transit constipation.^{11,25,33,34,36,38,40-42}

Figure B

Summary of literature search, with numbers of articles involved in each search step. * Total may exceed number in corresponding box, as articles could be excluded for more than one reason at this level.

Study Design Characteristics

Seven of the 12 studies were prospective,^{10,11,25,35,37,41,42} 4 studies were retrospective,^33,34,36,38 and 1 did not specify a study design.⁴⁰ All prospective studies applied the tests concurrently. Six studies appeared in meeting abstracts,^35-38,40,41 the remainder were in peer-reviewed publications.

All studies that reported the study location occurred in the United States.^{10,11,25,33-35,37,38} One study took place in multiple countries including the United States.²⁵ All studies that reported the location of recruitment occurred in tertiary centers.^11,33-38

Length of followup for the prospective studies and those with unspecified designs included the day of the testing only,^{35,37,38,40,43} 3 days,¹⁰ 5 days,^41,42 14 days,²⁵ and 21 days.¹¹

Prospective studies included patients with known gastroparesis^10,35,37 or constipation.^11,25,42 Four retrospective studies included patients with suspected gastroparesis or constipation^33,34,36,38 and one included patients with known constipation exclusively.⁴⁰ Six of the prospective studies also included patients without gastroparesis or constipation,^{10,11,35,37,41,42} whereas one study included only patients with known constipation.²⁵ Three studies that included patients with constipation used the Rome III criteria as inclusion criteria.^11,25,42 Three studies reported age restrictions. One allowed patients 18 to 80 years of age²⁵ and two others included patients older than 65 years of age.^41,42

Study Population Characteristics

No gender restrictions were made in the inclusion criteria, although most of participants with gastroparesis or constipation were female. The mean age was 40 or greater in all studies that reported an average.^{11,25,33,34,40,41} Three studies reported on race or ethnicity.^10,25,34 More than 80 percent of the participants were white in these studies. No study reported a measure of weight, blood sugar, or smoking status at baseline. Two studies reported on the percent of patients with diabetes,^33,39 reporting 15 and 37 percent with the disease, respectively. Two studies reported on defecatory dysfunction.^33,40 In one study, 20 of 32 subjects had defecatory dysfunction,⁴⁰ and in another study 64 percent of patients had this dysfunction.³³ Studies rarely reported on prior or concurrent use of medications, including prokinetics, opiates, antidepressants, proton pump inhibitors, and laxatives. Diagnostic testing prior to the study included scintigraphy^10,33,34,37 and ROM.^33,34

Characteristics of Diagnostic Tests

We summarized the characteristics of the tests used in the studies, taking into consideration how the evaluation of gastrointestinal motility is dependent on multiple factors, including not only the types of test but also the specific protocols the studies employed, which were often not standardized. Our criteria for study assessment suggested that “best practice” studies would report on smoking, use of prokinetics, use of selective serotonin reuptake inhibitors, use of antacids, and the specific timing of ingestion of test meals. However, only a few of the studies with larger populations specified a predetermined meal and meal schedule for patients undergoing WMC testing. Several of the studies also specified that participants did not use prokinetics within the immediate timeframe of WMC testing. Clinicians most frequently performed gastric scintigraphy using the consensus protocol.⁸ The community referral practice coordinated the ROM studies as per their local standards or the study made reference to a variation of the Metcalf protocol, wherein patients ingest ROMs and then receive an interval x ray and assessment of the marker location and number.^11,44-46 Few articles gave more specific test characteristics for ROM testing. Most abstracts did not report on any of these characteristics.

Study Quality

We reported study quality separately for the full-length publications and the abstracts, because the abstracts had limited information about study methods. Overall, study quality was fair among the 11 full-length publications we assessed.^{10,11,25,33,34,39,42,47-50} Half of them used a uniform reference standard.^{10,11,25,47,48} Only three studies interpreted the WMC results independently from the reference standard.^11,25,34 In another three studies that did not report blinding, we were able to confirm, after contacting the authors, that the studies interpreted results independently.^10,39,47

KQ 1. Evaluation of Gastric Dysmotility: WMC Alone Versus Other Diagnostic Tests; and KQ 2. Evaluation of Gastric Dysmotility: WMC in Combination With Other Diagnostic Tests Versus Other Diagnostic Tests Alone

We summarized the results for KQ 1 and KQ 2 in Table A.

Table A

Summary of the strength of evidence (SOE) and main findings of studies comparing WMC alone (KQ 1) or in combination (KQ 2) with other diagnostic tests for the evaluation of gastroparesis.

KQ 3. Evaluation of Colonic Dysmotility: WMC Alone Versus Other Diagnostic Tests; and KQ 4. Evaluation of Colonic Dysmotility: WMC in Combination With Other Diagnostic Tests Versus Other Diagnostic Tests Alone

We summarized the results from KQ 3 and KQ 4 in Table B.

Table B

Summary of the SOE and main findings of studies comparing WMC alone (KQ 3) or in combination (KQ 4) with other diagnostic tests for the evaluation of slow-transit constipation.

Potential Niche for WMC

WMC is a potential improvement over previous testing modalities for patients with possible gastroparesis or slow-transit constipation because it is small and can be transported to patients wherever they live. Also, the capsule does not contain any radioactive material or entail x-ray exposure, and can record information about pressure, transit, and location simultaneously. Other testing modalities for gastric emptying and colonic motility assessment do not share these characteristics. Certain academic centers use scintigraphy to assess gastric transit abnormalities and evaluate whole gut motility; however, this procedure involves radiation exposure, significant patient time, and significant cost. Antroduodenal manometry assesses gastric pressure parameters but has limited availability and is more invasive than other testing modalities; thus, physicians commonly use it as an investigative tool rather than as a clinical test. ROMs are portable and small, but require radiation exposure, access to fluoroscopy, and radiology interpretation. In addition, all other methods for evaluating either gastric or colonic motility evaluate either transit or pressure, but not both; yet both are involved in disease pathogenesis. Since WMC can evaluate both transit and pressure simultaneously, it could allow more optimal assessment of motility than evaluation of either parameter independently. Likewise, by recording both parameters, WMC has the potential to replace a combination of modalities and provide more accurate diagnosis with less resource utilization and enhanced patient convenience.

In light of this potential niche, WMC is becoming much more readily available in both academic and community centers. However, questions remain about the position of WMC in the diagnostic algorithm for suspected motility disorders such as gastroparesis and slow-transit constipation. Is WMC equivalent to conventional testing? Is it superior? Is it more likely to establish a concrete diagnosis or guide medical therapy than conventional motility testing? Should it be used as a stand-alone test? What should be done when WMC results are normal but clinical suspicion remains? Recommendations from the ANMS practice guidelines suggest that WMC can be useful in the diagnostic work up of patients with suspected gastroparesis and slow-transit constipation as well as those with more generalized motility disorders, but these are consensus guidelines. There is no specific or clear information about when or how physicians should utilize a WMC.

We must also consider potential limitations of WMC. The manufacturer lists severe gastroparesis as a contraindication to capsule placement due to fear of capsule retention. In addition, by definition, WMC evaluates motility at only a single point, as opposed to antroduodenal manometry, which has multiple recording points, or scintigraphy, which looks at transit of an entire meal. One assumes that the single point of measurement is representative of motility parameters as a whole; however, this is an assumption only and is not clearly established in the literature. When assessing constipation, one cannot distinguish patients with slow-transit constipation from those with defecatory dysfunction based on only colonic transit time, so we need further motility testing with anorectal manometry and clinical judgment to evaluate defecation. Finally, parameters of motility for a nondigestible solid are different from those for either liquids or a meal—so that patients can have abnormalities that would be detected with one modality but that would not be seen with another. In short, while the potential of WMC testing is exciting, many questions remain as to its appropriate place in the diagnostic algorithm.

Key Findings and Implications

Few studies met our criteria for evaluation. The paucity of full-length articles with independent data limited our ability to answer the KQs definitively.

Applicability

Limiting the application of the literature is the fact that all studies occurred at referral centers and that all prospective studies involved patients with known disease (thereby providing no prospective testing of WMC as a diagnostic tool). When a study used a comparison group without constipation or gastroparesis, it included “healthy” controls instead of patients who may have similar presenting symptoms but who do not have constipation or gastroparesis. These controls tended to be college-age men compared with middle-age females with suspected disease. Additionally, it is unclear how previous treatments or comorbidity, including diabetes, affect test performance or how the test results ultimately affect management.

Limitations and Strengths of Our Review Process

Our review had three major limitations:

1. No standards exist in the field of motility assessment for determining the minimum improvement of diagnostic accuracy that will identify one test as superior to another test. There are also no standards to establish the equivalence of motility tests. We arbitrarily chose a 10 percent difference in sensitivity or specificity as a potentially important difference between tests.²⁵ We felt that this threshold was a conservative minimum improvement over a reference standard with moderate diagnostic accuracy (between 50 and 80 percent). If the reference standard had a larger diagnostic accuracy (90 percent or greater), a 10 percent absolute difference is too large to expect.
2. We excluded studies that included non-diseased participants exclusively, because our review focused on studies that compared the diagnostic accuracy of the tests for patients with gastroparesis or slow-transit constipation. We recognize that many of the most commonly cited studies in the field included non-diseased participants exclusively.^12,13,51-64 Thus, we excluded a number of studies that evaluated characteristics of WMC.
3. Experts in the field acknowledge that scintigraphy and ROM have imperfect diagnostic accuracy. There are several options to account for the imperfection of the reference standard.⁶⁵ We chose to incorporate two of these in our review: (1) We presented the results as if the reference standard had no measurement error and acknowledged this imperfection. (2) We presented concordance of the test results when available. We did not attempt to adjust the results to correct for the measurement error. This adjustment would have required assumptions that we did not have sufficient data to justify. Another option is to examine patient outcomes according to WMC. We had included patient outcomes (need for medications, additional tests) as outcomes in our review. Unfortunately, we found few studies evaluating these outcomes.

The major strength of our review process was its comprehensiveness. We included abstracts, contacted industry for unpublished studies, and contacted study authors for missing data.

Limitations of the Identified Literature

Our aim was to compare the diagnostic accuracy of WMC with other testing modalities to diagnose and manage gastroparesis and slow transit constipation. The identified literature limited our ability to answer our KQs for several reasons:

1. No study directly addressed the incremental value of using WMC in addition to ROM or scintigraphy in the evaluation of colonic dysmotility (KQ 4). Only limited data addressed the incremental value of using WMC in addition to gastric scintigraphy, antroduodenal manometry, or endoscopy in the evaluation of gastric dysmotility (KQ 2).
2. All study sites were referral centers that tend to have patients with more severe disease. The study results have limited generalizability to general gastroenterology or primary care clinics where there is a greater spectrum of disease severity. The sensitivity and specificity of WMC may be different in referral center settings than in other settings, and the positive and negative predictive values will be different when the prevalence of disease is different.
3. Many studies included nondiseased patients in the comparison of the diagnostic accuracy of WMC with other tests, using a clinical diagnosis of disease as the reference standard rather than the results of the other diagnostic tests.
4. The non-diseased participants had demographic characteristics very different from the gastroparesis and slow-transit-constipation patients. For example, the majority of the non-diseased participants were college-age males, whereas the gastroparesis and slow-transit-constipation patients were middle-age women. Using clinical diagnosis as the reference standard, it is difficult to determine if WMC and other tests are distinguishing disease from non-disease or measuring differences in motility by demographic differences such as age and sex.
5. Variability in the administration of the motility tests and outcome assessments may explain some of the heterogeneity in the study results. Many studies used similar protocols to perform WMC testing and other tests, but with slight modifications such as the contents of the meal. Frequently, the timing of the motility assessment differed for WMC and the alternative test within and between studies, which may explain differences in the test results and the diagnostic accuracy differences between studies.
6. The abstracts we included did not report enough data to allow us to fully understand the study population, answer our KQs, and assess the quality of the studies.
7. We were unable to compare the results of studies with and without industry or investigator conflicts of interest because the company that manufactures the WMC sponsored most of the studies. The other studies did not report on conflicts of interest. No study stated that it was performed independent of industry sponsorship with authors who had no previous or current financial relationships with the manufacturer of the WMC.
8. Many studies included patients with gastroparesis defined by clinical symptoms and a prior abnormal gastric scintigraphy via local standards; however, symptoms of gastroparesis can be non-specific and many local facilities do not follow a standardized gastric scintigraphy protocol. As such, it is difficult, based on the data, to separate patients with gastroparesis from those with functional dyspepsia or other functional gastrointestinal disorders. This may have, to some degree, affected data with regards to sensitivity, specificity, and device correlation.
9. We attempted to assess publication bias by contacting the manufacturer of the WMC and requesting any unpublished data, but received no response.
10. Not all studies reported sufficient numbers to describe all the combinations of test results; some only provided means or medians. This hampered our ability to perform analyses, especially when analyzing combinations of tests.
11. Very few studies reported on patient-centered outcomes, limiting our abilities to draw conclusions on these outcomes.

Future Research Needs

Future research should ideally concentrate on finding a cure to these diseases that is nontoxic, cheap, easily available, and safe without major surgery or implanted devices. As far as diagnostic testing, the goal is always to find accurate, effective, and inexpensive tools to diagnose or exclude cases and qualify their severity in a reproducible way, especially when treatment is expensive, unavailable, or accompanied by great risks. Studies that compare the diagnostic modalities should have blinded interpretation of the results and make every attempt to classify patients by identical criteria and standardized protocols that other centers can repeat and verify. We recommend that research focus more on prospectively studied patients in larger numbers with an appropriate spectrum of symptoms and adequate followup to determine whether the diagnosis was accurate over time. Due to the difficulty enrolling patients, studies should carefully craft retrospective analyses.

We need research studies that evaluate how clinicians should use the WMC in combination with or instead of other testing modalities for evaluating slow-transit constipation. The studies we reviewed used alternative measures to assess anorectal function, such as anorectal manometry, as WMC does not capture data about this region reliably. Thus, clinicians will likely use WMC in combination with this test.

Eventually, we need outcomes studies to see if testing helps to improve quality of life or symptom control. It is unclear at present whether a more sensitive diagnostic test might just provide lead-time bias—or apparent superiority for an earlier diagnosis—but not actually change the outcomes or management steps overall for the patient. As we identify other targeted therapies, we will need to reassess the value of testing. We are aware that a new therapy is in Stage II trials for patients with diabetes and gastric emptying delay, which may increase the need for research into this area if it becomes available for use.⁶⁶ Currently, most patients with nausea- and vomiting-predominant symptoms of gastroparesis receive similar first-line treatment with antiemetics or prokinetics. As treatment options for gastroparesis expand (some at great expense), then more accurate detection of disease prior to initiation of therapy may play a more prominent role in disease management. The literature does not currently report resource utilization with and without WMC—we will need more studies evaluating these measures.

Little data is available to determine the optimal timing of WMC testing in the diagnostic and therapeutic approach to patients with symptoms of possible gastroparesis or slow-transit constipation. We need to do further work to classify the types of patients within subgroups of gastroparesis or slow-transit constipation in order to identify severe cases that may need more urgent evaluation. Finally, little is known about whether physicians should use testing to assess the effectiveness of treatment or if subsequent testing would offer any benefit in long-term management of patients. Currently, symptoms and symptom resolution guide therapeutic decisions, but these require careful interpretation.

Conclusions

Based on the current literature, WMC appears to be accurate in detection of gastroparesis and slow-transit constipation and may provide increased diagnostic gain as compared with standard motility testing. While the SOE is low, the data were relatively consistent and suggested that this modality is no less sensitive than conventional testing. The evidence is insufficient to determine whether use of WMC will improve outcomes of care.