Water-assisted colonoscopy: an international modified Delphi review on definitions and practice recommendations.

BACKGROUND AND AIMS: Since 2008, a plethora of research studies has compared the efficacy of water-assisted (aided) colonoscopy (WAC) and underwater resection (UWR) of colorectal lesions with standard colonoscopy. We reviewed and graded the research evidence with potential clinical application. We conducted a modified Delphi consensus among experienced colonoscopists on definitions and practice of water immersion (WI), water exchange (WE), and UWR. METHODS: Major databases were searched to obtain research reports that could potentially shape clinical practice related to WAC and UWR. Pertinent references were graded (Grading of Recommendations, Assessment, Development and Evaluation). Extracted data supporting evidence-based statements were tabulated and provided to respondents. We received responses from 55 (85% surveyed) experienced colonoscopists (37 experts and 18 nonexperts in WAC) from 16 countries in 3 rounds. Voting was conducted anonymously in the second and third round, with ≥80% agreement defined as consensus. We aimed to obtain consensus in all statements. RESULTS: In the first and the second modified Delphi rounds, 20 proposed statements were decreased to 14 and then 11 statements. After the third round, the combined responses from all respondents depicted the consensus in 11 statements (S): definitions of WI (S1) and WE (S2), procedural features (S3-S5), impact on bowel cleanliness (S6), adenoma detection (S7), pain score (S8), and UWR (S9-S11). CONCLUSIONS: The most important consensus statements are that WI and WE are not the same in implementation and outcomes. Because studies that could potentially shape clinical practice of WAC and UWR were chosen for review, this modified Delphi consensus supports recommendations for the use of WAC in clinical practice.

Adenoma detection rate is necessary but insufficient for distinguishing high versus low endoscopist performance.

BACKGROUND: Endoscopist quality is benchmarked by the adenoma detection rate (ADR)-the proportion of cases with 1 or more adenomas removed. However, the ADR rewards the same credit for 1 versus more than 1 adenoma. OBJECTIVE: We evaluated whether 2 endoscopist groups could have a similar ADR but detect significantly different total adenomas. DESIGN: We retrospectively measured the ADR and multiple measures of total adenoma yield, including a metric called ADR-Plus, the mean number of incremental adenomas after the first. We plotted ADR versus ADR-Plus to create 4 adenoma detection patterns: (1) optimal (↑ADR/↑ADR-Plus); (2) one and done (↑ADR/↓ADR-Plus); (3) all or none (↓ADR/↑ADR-Plus); (4) none and done (↓ADR/↓ADR-Plus). SETTING: Tertiary-care teaching hospital and 3 nonteaching facilities servicing the same patient pool. PATIENTS: A total of 3318 VA patients who underwent screening between 2005 and 2009. MAIN OUTCOME MEASUREMENTS: ADR, mean total adenomas detected, advanced adenomas detected, ADR-Plus. RESULTS: The ADR was 28.8% and 25.7% in the teaching (n = 1218) and nonteaching groups (n = 2100), respectively (P = .052). Although ADRs were relatively similar, the teaching site achieved 23.5%, 28.7%, and 29.5% higher mean total adenomas, advanced adenomas, and ADR-Plus versus nonteaching sites (P < .001). By coupling ADR with ADR-Plus, we identified more teaching endoscopists as optimal (57.1% vs 8.3%; P = .02), and more nonteaching endoscopists in the none and done category (42% vs 0%; P = .047). LIMITATIONS: External generalizability, nonrandomized study. CONCLUSION: We found minimal ADR differences between the 2 endoscopist groups, but substantial differences in total adenomas; the ADR missed this difference. Coupling the ADR with other total adenoma metrics (eg, ADR-Plus) provides a more comprehensive assessment of adenoma clearance; implementing both would better distinguish high- from low-performing endoscopists.

Development and validation of a novel patient educational booklet to enhance colonoscopy preparation.

OBJECTIVES: The success of colonoscopy depends on high-quality bowel preparation by patients; yet inadequate preparation is common. We developed and tested an educational booklet to improve bowel preparation quality. METHODS: We conducted patient cognitive interviews to identify knowledge and belief barriers to colonoscopy preparation. We used these interviews to create an educational booklet to enhance preparatory behaviors. We then prospectively randomized patients scheduled for outpatient colonoscopy at a VA Medical Center to receive usual instructions vs. the booklet before colonoscopy. Patients in both groups received standard pharmacy instructions for single-dose bowel preparation; the protocol did not specify which purgatives to prescribe. The primary outcome was preparation quality based on blinded ratings using the validated Ottawa score. We performed bivariate analyses to compare mean scores between groups using a t-test, and logistic regression to measure the booklet effect on preparation quality, adjusting for potential confounders. RESULTS: A total of 436 patients were randomized between arms. In an intention-to-treat analysis of the primary outcome, mean Ottawa scores were superior in patients allocated to booklet vs. controls (P=0.03). An intention-to-treat analysis of the secondary outcome revealed a "good" preparation in 68 vs. 46% of booklet and control patients, respectively (P=0.054). In a per-protocol analysis limited to patients who actually received the booklet, preparation was good in 76 vs. 46% patients, respectively (P<0.00001). Regression analysis revealed that booklet receipt increased the odds of good preparation by 3.7 times (95% confidence interval=2.3-5.8). CONCLUSIONS: Provision of a novel educational booklet considerably improves preparation quality in patients receiving single-dose purgatives. The effect of the booklet on split-dose purgatives remains untested and will be evaluated in future research.

Colonoscopy yields fewer polyps as the day progresses despite using social influence theory to reverse the trend.

BACKGROUND: We previously reported that fewer polyps are detected by colonoscopy as the day progresses, a phenomenon that could be modified with "social influence theory" by using auditing and feedback. OBJECTIVE: To measure the impact of a social influence informational poster on the relationship between time of day and colonoscopy yield. DESIGN: Controlled before-and-after study comparing the polyp yield and time of day relationship in a historical cohort versus a 3-month intervention period. SETTING: University-based Veterans Affairs medical center. PATIENTS: Patients undergoing outpatient screening, surveillance, or diagnostic colonoscopies. INTERVENTION: Placement of informational posters in endoscopy rooms within view of operators and nurses. The poster depicted a bar graph of the previously documented hour-by-hour decreases in polyp yield coupled with prominent text: "What Time Is It Now?" MAIN OUTCOME MEASUREMENT: Polyp yield, including secondary end point limited to adenoma detection. We performed regression to measure the effect of start time on polyp yield. RESULTS: There were 477 and 301 patients in the control and intervention periods, respectively. There was a negative relationship between start time and polyp yield, including adenoma detection, for both periods (P = .001). Start time remained negatively predictive of polyp and adenoma yield after adjusting for poster exposure and confounders (P = .01). LIMITATIONS: Nonrandomized study design. CONCLUSION: An informational poster did not alter the relationship between colonoscopy start time and polyp yield. This strengthens the previous finding that start time may affect polyp yield and suggests that passive use of social influence theory is inadequate to modify this effect. Shortening endoscopy shifts and active auditing with feedback may be necessary.

Fewer polyps detected by colonoscopy as the day progresses at a Veteran's Administration teaching hospital.

BACKGROUND & AIMS: One objective of colonoscopy is to identify and remove polyps-this process requires attention to detail and prolonged concentration. Providers are predisposed to cognitive errors because the procedure is often performed repetitively throughout the day. We measured the adjusted relationship between colonoscopy start time and polyp yield. METHODS: We performed a prospective study of 477 patients that received screening, surveillance, or diagnostic colonoscopies at a Veteran's Administration (VA) teaching hospital. The primary outcome measure was polyp yield. We collected data on colonoscopy start times, which were analyzed both as a dichotomous time period ("early-morning case" vs "later case") and as a continuous variable (start time). We identified significant risk factors using univariate analysis and performed Poisson multivariable regression to measure the independent effect of colonoscopy start time on polyp yield. We evaluated evidence of decreasing polyp yield as the day progressed throughout pre-specified time intervals. RESULTS: In univariate analysis, early-morning cases yielded 27% more polyps per patient than later cases (95% confidence interval, 11%-45%; P < .001). The total numbers of, hyperplastic and adenomatous polyps found decreased hour-by-hour as the day progressed. Multivariable analysis demonstrated that early-morning cases yielded 20% more polyps per patient than later cases (95% confidence interval, 5%-38%; P = .007). CONCLUSIONS: At a VA medical center, more polyps were detected in patients that received colonoscopies early in the morning compared with later in the day. Moreover, adenoma detection reduced as the day progressed. Providers might be most adept at detecting polyps at the beginning of the day; further validation in other practice settings is required.

Biliary disease in pregnancy with an emphasis on the role of ERCP.

The incidence of gallstones during pregnancy is estimated to be between 3% and 12%. About one-third of pregnant patients with cholelithiasis become symptomatic and may require surgical intervention. Choledocholithiasis during pregnancy although infrequent usually requires therapeutic intervention. Abdominal ultrasonography is insensitive for the detection of common bile duct stones. Magnetic resonance imaging is not being associated with known adverse effects and seems to be an excellent diagnostic modality in this context. Paramagnetic contrast agents have been associated with increased spontaneous abortion rates and other abnormalities in animals and should only be used when absolutely necessary. Endoscopic ultrasonography is highly accurate for the detection of common bile duct stones and may be useful before consideration of endoscopic retrograde cholangiopancreatography (ERCP) in select patients. The second trimester seems to be the safest time to perform surgery, as organogenesis is complete and the incidence of spontaneous abortion lower. ERCP followed by sphincterotomy and stone extraction is very effective and can be performed safely during all trimesters of pregnancy with a premature delivery rate less than 5%. All efforts to minimize radiation exposure should be undertaken. These include lead shielding and avoiding hard copy radiographs. When possible, category B (such as meperidine) or C drugs only should be used for sedation during pregnancy. Therapeutic ERCP is now the standard of care for treating choledocholithiasis during pregnancy. Endoscopic sphincterotomy for symptomatic patients with normal cholangiograms is controversial. Consideration of ERCP demands a judicious approach, paying careful attention to risks and benefits of intervention.

The gastric cardia: to be or not to be?

The origin and biologic significance of cardiac gastric mucosa are controversial. Traditionally, it has been considered native mucosa and part of normal foregut development. It has been recently suggested that cardiac mucosa is present only as a metaplastic response to gastroesophageal reflux disease and therefore always abnormal. We evaluated the esophagogastric junction in 100 pediatric autopsy samples to determine the existence, characteristics, and length of pure cardiac mucosa at different ages. No patient had a history of gastroesophageal reflux disease. Cardiac mucosa immediately distal and contiguous to the esophageal squamous mucosa was identified in all 100 samples, varying in length from 0.1 to 3 mm; the mean length was 1 mm. There was an inverse correlation between patient age and length of cardiac mucosa; gender had no influence on measured length. Three patients had mild to moderate histologic esophagitis; two had gastritis. No metaplastic features or Helicobacter pylori were identified. These findings support the concept that there is a normal, variably narrow developmental zone at the esophagogastric junction covered by cardiac mucosa and is present at birth. When cardiac type mucosa is found in biopsy material, it does not necessarily represent evidence of a mucosal metaplastic response to gastroesophageal reflux disease.

Common bile duct stenosis and pancreatitis in Kawasaki disease: two cases and a review of the literature.

We report 2 children with Kawasaki disease, who developed pancreatitis and common bile duct abnormalities. After reviewing the literature, we conclude that Kawasaki disease may lead to such changes as a result of the underlying vasculitis characterizing this disease, or of direct inflammation of the biliary tree.

Diagnostic value of detection of IgM antibodies to Helicobacter pylori.

This report describes the diagnostic value of anti-Helicobacter pylori IgM detection. Serum samples from 9043 symptomatic and asymptomatic individuals were evaluated with ELISA for the presence of anti-H. pylori IgG, IgM, and IgA. The specificity of detected IgM was confirmed by inhibition and cross-reactivity assays. Treatment of IgM-positive specimens with 1% 2-mercaptoethanol resulted in approximately 90% inhibition. Our data suggest a low level of cross-reactivity (5%) between H. pylori and four different enteropathogenic bacteria tested. The specificity of anti-H. pylori IgM was also demonstrated by Western blot and linearity studies. Data show that the detected IgM is highly specific. Western blot analysis revealed a variable IgM response to H. pylori antigens among patients, with the most reactive antigenic fractions being in the range of 55- to 100-kDa. Overall, the data confirm the diagnostic value of anti-H. pylori IgM detection. The prevalence of IgM antibodies to H. pylori in tested sera was significantly higher in symptomatic patients (10.4%) than in asymptomatic individuals (1.1%). Likewise, the percentage of sera positive for IgM alone was higher in symptomatic than in asymptomatic groups (3.8 vs 0.22%). About 5% of sera were positive only for IgA. We concluded that ELISA can be used for the detection of specific IgM to H. pylori and that the presence or absence of IgM antibodies to H. pylori may reflect whether or not an acute infection exists.

Placement of self-expanding metal stents for acute malignant large-bowel obstruction: a collective review.

BACKGROUND: The purpose of this study was to review our experience with self-expanding metal stents as the initial interventional approach in the management of acute malignant large-bowel obstruction. METHODS: Twenty-six patients who underwent placement of colonic stents at our institution between June 1994 and June 2000 were identified and reviewed. RESULTS: In 14 patients, the stents were placed for palliation, whereas in 12, they were placed as a bridge to surgery. In 22 patients (85%), stent placement was successful on the first occasion. In the remaining four individuals, one was successfully stented at the second occasion, and three required emergency surgery. Nine of the 12 patients (75%) in the bridge-to-surgery group underwent elective colon resection. In the palliative group, four patients (29%) had reobstruction of the stents, and in one (9%), the stent migrated. In the remaining nine patients (64%), the stent was patent until the patient died or until the time of last follow-up (median, 156 days). CONCLUSIONS: In our experience with 26 patients who developed a complete bowel obstruction as a consequence of a malignant tumor, placement of colonic stents to achieve immediate nonoperative decompression proved to be both safe and effective. Subsequent elective resection was accomplished in the majority of resectable cases.