James E. Allison

A Comparison of Fecal Occult-Blood Tests for Colorectal-Cancer Screening

BACKGROUND Hemoccult II, a widely used guaiac test for fecal occult blood, has a low sensitivity for detecting colorectal neoplasms in asymptomatic patients at average risk. In such patients, the performance characteristics of screening tests developed to improve on Hemoccult II are not known. METHODS A set of three fecal occult-blood tests--Hemoccult II; Hemoccult II Sensa, a more sensitive guaiac test; and HemeSelect, an immunochemical test for human hemoglobin--was mailed to all patients 50 years of age or older who were scheduled for personal health appraisals at the Kaiser Permanente Medical Center in Oakland, California. The performance of each test and of a combination test (HemeSelect to confirm positive Hemoccult II Sensa results) was evaluated by identifying screened patients who had colorectal neoplasma (carcinoma or a polyp > or = 1 cm in diameter) in the two years after screening. RESULTS Of the 10,702 eligible patients, 8104 (75.7 percent) had at least one interpretable sample and were screened on the basis of at least one test; 96 percent of these patients had complete two-year follow-up. The sensitivity of the tests for detecting carcinoma was lowest with Hemoccult II (37.1 percent; 95 percent confidence interval, 19.7 to 54.6 percent), intermediate with the combination test (65.6 percent; 95 percent confidence interval, 47.6 to 83.6 percent) and with HemeSelect (68.8 percent; 95 percent confidence interval, 51.1 to 86.4 percent), and highest with Hemoccult II Sensa (79.4 percent; 95 percent confidence interval, 64.3 to 94.5 percent). The specificity for detecting carcinoma was 86.7 percent with Hemoccult II Sensa, 94.4 percent with HemeSelect, 97.3 percent with the combination test, and 97.7 percent with Hemoccult II. HemeSelect and the combination test detected more colorectal carcinomas and polyps than Hemoccult II, with only slight increases in the number of colonoscopies needed. CONCLUSIONS HemeSelect and a combination test in which HemeSelect is used to confirm positive Hemoccult II Sensa results improve on Hemoccult II in screening patients for colorectal carcinoma.

Incidence and Mortality of Colorectal Adenocarcinoma in Persons With Inflammatory Bowel Disease From 1998 to 2010

BACKGROUND & AIMS The relationship between inflammatory bowel disease (IBD) and the incidence and mortality of colorectal adenocarcinoma (CRC) has not been evaluated recently. METHODS We calculated the incidence and standardized incidence and mortality rate ratios of CRC among adult individuals with intact colons using Kaiser Permanente of Northern Californias database of members with IBD and general membership data for the period of 1998 to June 2010 (data through 2008 were used to calculate mortality). We also evaluated trends in medication use and rates of cancer detection over time. RESULTS We identified 29 cancers among persons with Crohns disease (CD) and 53 among persons with ulcerative colitis (UC). Overall, the incidence rates of cancer among individuals with CD, UC, or in the general membership were 75.0, 76.0, and 47.1, respectively, per 100,000 person-years. In the general population, the incidence of CRC was 21% higher in 2007-2010 than in 1998-2001 (P for trend, <.0001), coincident with the growth of CRC screening programs. The incidence of CRC among individuals with CD or UC was 60% higher than in the general population (95% confidence interval [CI] for CD, 20%-200%; 95% CI for UC, 30%-200%) and was stable over time (P for trend was as follows: CD, .98; UC, .40). During 1998-2008, the standardized mortality ratio for CRC in individuals with CD was 2.3 (95% CI, 1.6-3.0) and 2.0 in those with UC (95% CI, 1.3-2.7). Over the study period, anti-tumor necrosis factor agents replaced other therapies for CD and UC; the rate of colonoscopy increased by 33% among patients with CD and decreased by 9% in those with UC. CONCLUSIONS From 1998 to 2010, the incidence of CRC in patients with IBD was 60% higher than in the general population and essentially stable over time.

Incidence and Prevalence of Inflammatory Bowel Disease in a Northern California Managed Care Organization, 1996–2002

OBJECTIVE:There are few estimates of the incidence and prevalence of inflammatory bowel disease in North American communities. We sought to estimate the incidence and prevalence of inflammatory bowel disease (IBD), including Crohns disease (CD), and ulcerative colitis (UC), among 3.2 million members of Kaiser Permanente, Northern California, for the period 1996–2002.METHODS:All health plan members who had one or more diagnoses of CD (ICD-9 code 555) or UC (ICD-9 code 556) on computerized records during the period 1996–2002 and with at least 12 months of membership were identified as possible IBD cases (N = 12,059). We randomly sampled 24% of these for chart review to confirm the diagnosis and obtain the initial diagnosis date. Incidence rates and the point prevalence on December 31, 2002 were standardized to the 2000 U. S. Census.RESULTS:The annual incidence rate per 100,000 persons was 6.3 for CD (95% confidence interval [CI], 5.6–7.0) and 12.0 for UC (CI, 11.0–13.0). The point prevalence per 100,000 on December 31, 2002 was 96.3 for CD (95% CI, 89.6–103.0) and 155.8 for UC (95% CI, 146.6–164.9), increasing to 100.3 and 205.8 per 100,000, respectively, when hospital discharge data from 1985 to 1995 were included. The age-specific incidence of CD was bimodal, while UC incidence rose in early adulthood and remained elevated with advancing age.CONCLUSIONS:The incidence we estimated for CD was similar to the previous U. S. estimate. Our incidence estimate for UC was much higher than the previous U.S. estimate, but similar to that of recent Canadian and European studies. The prevalence we estimated for CD was somewhat lower than previous estimates.

Role of thiopurine and anti-TNF therapy in lymphoma in inflammatory bowel disease.

OBJECTIVES:The objective of this study was to assess inflammatory bowel disease (IBD) medications in relation to lymphoma risk.METHODS:Information on IBD and relevant medications and other utilization was obtained from the Kaiser Permanente IBD Registry, 1996–2009. Lymphoma cases were ascertained from the Kaiser Permanente Cancer Registry. Lymphoma incidence was compared between the IBD cohort and the general Kaiser Permanente population.RESULTS:Of the 16,023 IBD patients without human immunodeficiency virus followed an average 5.8 years, 43 developed lymphoma. IBD patients with and without lymphoma did not differ with respect to past IBD-related visits, procedures, or tests. The standardized incidence rate ratio (SIRR) for lymphoma among IBD patients with no dispensing of thiopurine or anti-tumor necrosis factor (TNF) was 1.0 (95% confidence interval (CI): 0.96–1.1). Of the 21,282 person-years involving exposure to thiopurine or anti-TNF, 81% involved thiopurine alone; 3%, anti-TNF alone; and 16%, combination therapy. Among patients with thiopurine but not anti-TNF dispensings, the SIRR was 0.3 (95% CI: 0.2–0.4) for past use and 1.4 for current use (95% CI: 1.2–2.7). Among patients with dispensing of anti-TNF (with and without thiopurine), the SIRR was 5.5 for past use (95% CI: 4.5–6.6) and 4.4 for current use (95% CI: 3.4–5.4). The most common lymphoma subtypes were diffuse large B-cell lymphoma (44%), follicular lymphoma (14%), and Hodgkins disease (12%).CONCLUSIONS:Our study provides evidence that IBD alone is not associated with the risk of lymphoma. Use of anti-TNF with thiopurine and current use of thiopurine alone were associated with increased risk, although the effect of disease severity merits further evaluation.

Hemoccult Screening in Detecting Colorectal Neoplasm: Sensitivity, Specificity, and Predictive Value: Long-Term Follow-up in a Large Group Practice Setting

STUDY OBJECTIVE To determine the sensitivity, specificity, and predictive value of Hemoccult II tests for detecting colorectal neoplasm (colorectal carcinoma or polyp or both). STUDY DESIGN Prospective analyses of asymptomatic patients (greater than or equal to 45 years) followed for 4 years after screening with Hemoccult II testing and retrospective analyses of patients, with known colorectal carcinoma or polyps or both who had Hemoccult II testing within 2 years of diagnosis. SETTING A large, health maintenance organization practice. MEASUREMENTS AND MAIN RESULTS In the prospective analysis, the sensitivity of Hemoccult II was 50% for colorectal carcinoma diagnosed within 1 year of testing, 43% within 2 years, and 25% within 4 years. For polyps, sensitivity was 36% at 1 year, 28% at 2 years, and 17% at 4 years. Specificity was 99%. The predictive value of a positive test for colorectal carcinoma was 8% at 1 year, 10% at 2 years, and 11% at 4 years. On the basis of the retrospective analyses, the sensitivity of Hemoccult II for colorectal carcinoma diagnosed within 1 year of testing was 66% and was 61% within 2 years. Many of these patients had symptoms when tested. CONCLUSIONS An asymptomatic patient age 45 or older with a positive Hemoccult II test has about a chance of 1 in 10 for having colorectal carcinoma and a 1-in-3 chance of having either a colorectal carcinoma or polyp: The same patient with a negative Hemoccult test has a 0.2% chance of having a colorectal carcinoma diagnosed within 2 years of testing and a 0.7% chance of having a polyp. Within 4 years of testing the chance increases to 0.5% for colorectal carcinoma and 1.5% for polyps. If Hemoccult II slides are the only screening method used for detecting asymptomatic colorectal neoplasms, 50% to 60% of lesions will remain undetected. Clinical interpretation of Hemoccult screening requires appreciation of its limits as well as its benefits.

A Proposal to Standardize Reporting Units for Fecal Immunochemical Tests for Hemoglobin

Fecal immunochemical tests for hemoglobin are replacing traditional guaiac fecal occult blood tests in population screening programs for many reasons. However, the many available fecal immunochemical test devices use a range of sampling methods, differ with regard to hemoglobin stability, and report hemoglobin concentrations in different ways. The methods for sampling, the mass of feces collected, and the volume and characteristics of the buffer used in the sampling device also vary among fecal immunochemical tests, making comparisons of test performance characteristics difficult. Fecal immunochemical test results may be expressed as the hemoglobin concentration in the sampling device buffer and, sometimes, albeit rarely, as the hemoglobin concentration per mass of feces. The current lack of consistency in units for reporting hemoglobin concentration is particularly problematic because apparently similar hemoglobin concentrations obtained with different devices can lead to very different clinical interpretations. Consistent adoption of an internationally accepted method for reporting results would facilitate comparisons of outcomes from these tests. We propose a simple strategy for reporting fecal hemoglobin concentration that will facilitate the comparison of results between fecal immunochemical test devices and across clinical studies. Such reporting is readily achieved by defining the mass of feces sampled and the volume of sample buffer (with confidence intervals) and expressing results as micrograms of hemoglobin per gram of feces. We propose that manufacturers of fecal immunochemical tests provide this information and that the authors of research articles, guidelines, and policy articles, as well as pathology services and regulatory bodies, adopt this metric when reporting fecal immunochemical test results.

Clustering of Inflammatory Bowel Disease With Immune Mediated Diseases Among Members of a Northern California-Managed Care Organization

BACKGROUND AND AIMS:Previous studies provide evidence that some immune-mediated diseases occur at greater frequency among inflammatory bowel disease (IBD) patients than in the general population. The present study examined the co-occurrence of IBD with common immune-mediated disorders including asthma, psoriasis, type 1 diabetes, rheumatoid arthritis, multiple sclerosis, systemic lupus erythematosus, vitiligo, autoimmune thyroiditis (Graves and Hashimotos), and chronic glomerulonephritis.METHODS:We conducted a cross-sectional study among members of the Kaiser Permanente Medical Care Program for the period 1996–2005. A total of 12,601 patients with at least two IBD diagnoses in computerized visit data were ascertained. Four persons without IBD were matched to each IBD patient on age, gender, and length of enrollment. Information on co-occurring diseases was obtained from computerized visit data for 1996–2005. Conditional logistic regression was used to estimate the odds ratio and 95% confidence interval for the association of IBD with immune-mediated disorders after adjusting for smoking.RESULTS:Seventeen percent of the IBD patients and 10% of the persons without IBD had a diagnosis for at least one immune-mediated disease. IBD patients were more likely to have asthma (1.5, 95% CI 1.4–1.6), psoriasis (1.7, 95% CI 1.5–2.0), rheumatoid arthritis (1.9, 95% CI 1.5–2.3), and multiple sclerosis (2.3, 95% CI 1.6–3.3).CONCLUSIONS:Among the immune-mediated diseases we studied, most were more common in IBD patients than in persons without IBD, suggesting that IBD shares common etiologic factors with other immune-mediated diseases.

Time Trends in Therapies and Outcomes for Adult Inflammatory Bowel Disease, Northern California, 1998–2005

BACKGROUND & AIMS The management of inflammatory bowel disease (IBD) has become increasingly complicated, and it is unknown whether poor outcomes (prolonged steroid use, hospitalizations, and surgery) have declined in the general population. METHODS This multilevel study used computerized clinical data. The study comprised 2892 adults with Crohns disease (CD) and 5895 with ulcerative colitis (UC) who received care at 16 medical centers within an integrated care organization in Northern California between 1998 and 2005. RESULTS Time trends included (1) a shift in gastroenterology-related visits from the gastroenterology division to primary care; (2) increased use of IBD-related drugs, except for a 7% decline in use of 5-aminosalicylate in CD and no change in steroid use for CD; (3) for the prevalence of prolonged steroid exposure (120 days of continuous use), a 36% decline for CD with a 27% increase for UC; (4) declines in the hospitalization rates of 33% for CD and 29% for UC; and (5) for the surgery rate, no significant change for CD with a 50% decline for UC. CONCLUSIONS Declines in prolonged steroid exposure and the hospitalization rate for CD and in the hospitalization and surgery rate for UC are encouraging; however, the increase in prolonged steroid exposure for UC merits concern and further investigation. The variability in care patterns observed in this study suggests lack of standardization of care and the opportunity to identify targets for quality improvement. These findings should stimulate research to quantify the effect of current trends in IBD management.

Estimation of the period prevalence of inflammatory bowel disease among nine health plans using computerized diagnoses and outpatient pharmacy dispensings

Background: There are few contemporary estimates of prevalence rates for inflammatory bowel disease (IBD) in diverse North American communities. Methods: We estimated the period prevalence of IBD for January 1, 1999, through June 30, 2001, among 1.8 million randomly sampled members of nine integrated healthcare organizations in the US using computerized diagnoses and outpatient pharmaceutical dispensing. We also assessed the positive predictive value (PPV) and sensitivities of 1) the case‐finding algorithm, and 2) the 30‐month sampling period using medical chart review and linkage to a 78‐month dataset, respectively. Results: The PPV of the case‐finding algorithm was 81% (95% confidence interval [CI], 78–87) and 84% (95% CI, 79–89) in two different organizations. In both, the sensitivity of the optimal algorithm, compared with the most inclusive, exceeded 90%. The sensitivity of the 30‐month sampling period compared with 78 months was 61% (95% CI, 57–64) in one organization. Applying a slightly more sensitive case‐finding algorithm, the average period prevalence of IBD across the nine organizations, standardized to the age‐ and gender‐distribution of the US population, 2000 census, was 388 cases (95% CI, 378–397) per 100,000 persons (range 209–784 per 100,000; average follow‐up 26 months). The prevalence of Crohns disease, ulcerative colitis, and unspecified IBD was 129, 191, and 69 per 100,000, respectively. Conclusions: The observed average prevalence was similar to prevalence proportions reported for other North American populations (369–408 per 100,000). Additional research is needed to understand differences in the occurrence of IBD among diverse populations as well as practice variation in diagnosis and treatment of IBD.

Incidence, prevalence, and time trends of pediatric inflammatory bowel disease in Northern California, 1996 to 2006.

OBJECTIVE To examine the incidence and prevalence of pediatric inflammatory bowel disease (IBD) during 1996-2006 in a community-based health-care delivery system. STUDY DESIGN Members of Kaiser Permanente Northern California aged 0 to 17 years with IBD were identified by use of computerized medical information with confirmation obtained through review of the medical record. RESULTS The average annual incidence of IBD per 100000 was 2.7 (95% confidence interval [CI], 2.3-3.1) for Crohns disease (CD) and 3.2 (CI, 2.8-3.6) for ulcerative colitis (UC). During the 11-year study period, the annual incidence per 100000 increased from 2.2 to 4.3 for CD (P = .09) and from 1.8 to 4.9 for UC (P < .001). The ratio of incident CD cases to incident UC cases was 0.9 in non-Hispanic whites, 1.6 in African Americans (P = .12), 0.3 in Hispanics (P < .001) and 0.4 in Asians (P = .04). The average length of enrollment during the 11-year study period exceeded 8 years. The point prevalence on December 31, 2006, per 100000 was 12.0 for CD (CI, 9.6-14.4) and 19.5 (CI, 16.5-22.6) for UC. CONCLUSIONS In this population the incidence of UC increased significantly by 2.7-fold and CD increased 2.0-fold without reaching statistical significance. Hispanic and Asian children had development of UC more often than CD, suggesting possible etiologic differences across racial and ethnic groups.