Mortality in Barrett's esophagus: three decades of experience at a single center.

BACKGROUND AND STUDY AIMS: There is a view that the majority of deaths in patients with Barrett's esophagus are from causes other than esophageal adenocarcinoma (EAC). The aim of this analysis was to establish the pattern of mortality for a number of causes in patients with Barrett's esophagus. PATIENTS AND METHODS: This was a single-center prospective cohort study of patients from Rotherham District General Hospital, which is a secondary referral center. The cohort consisted of 1239 patients who were diagnosed with Barrett's esophagus between April 1978 and March 2009.  Follow-up for mortality was undertaken by "flagging" the patients with the NHS Information Center. Causes of death were compared with UK Office of National Statistics age- and sex-specific mortality data for 1999, the median year of diagnosis. Analysis was by a "person - years at risk" calculation from date of diagnosis. RESULTS: The ratio of observed deaths from EAC compared with those expected in this cohort was 25.02 - a very large excess. There was no difference in mortality from colorectal cancer or circulatory disease and there were fewer deaths from cancers other than esophageal adenocarcinoma and colon cancer compared with national statistics. There was a small statistically significant difference in mortality from all causes but this disappeared completely when deaths from esophageal adenocarcinoma were excluded. CONCLUSIONS: Overall, mortality in Barrett's esophagus is increased significantly but only as a result of the large excess of deaths from EAC. This strengthens the case for endoscopic surveillance if successful interventions can be undertaken in patients with Barrett's esophagus to prevent development of esophageal adenocarcinoma.

The influence of symptom type and duration on the fate of the metaplastic columnar-lined Barrett's oesophagus.

BACKGROUND: Prolonged gastro-oesophageal reflux resulting in columnar metaplasia of the oesophagus is the main risk factor for oesophageal adenocarcinoma. AIM: To examine the duration of symptoms and associations of different symptoms with the development of columnar-lined oesophagus, dysplasia and adenocarcinoma. METHODS: UK multicentre cohort study of patients with columnar-lined oesophagus whose date of symptom onset (1082 patients) and/or types of symptoms reported (1681 patients) were documented. Follow-up was examined by analysis of histological reports from the registering centers. RESULTS: Symptoms of dysphagia/odynophagia and nausea/vomiting were associated with development of dysplasia. High-grade dysplasia and adenocarcinoma were associated with dysphagia/odynophagia and weight loss. Median duration from symptom onset to detection of columnar-lined oesophagus without intestinal metaplasia: 2.6 years, columnar-lined oesophagus with intestinal metaplasia: 5.0 years, indefinite changes for dysplasia: 19.3 years and low-grade dysplasia: 30.0 years. One tenth of patients had developed high-grade dysplasia at 9.6 years and one tenth had developed adenocarcinoma at 13.8 years from symptom onset. CONCLUSIONS: In patients with columnar-lined oesophagus, symptoms of dysphagia/odynophagia and nausea/vomiting were associated with a higher risk of development of dysplasia and adenocarcinoma. There is a trend for longer duration of symptoms to the detection of dysplasia.

Barrett's columnar-lined oesophagus: demographic and lifestyle associations and adenocarcinoma risk.

OBJECTIVES: Lifestyle and demographic risk factors for the development of oesophageal adenocarcinoma developing from columnar-lined oesophagus are not well defined. METHODS: Demographic and lifestyle factors, endoscopy and histology reports were extracted from 1,761 subjects from seven UK centres. The associations of columnar-lined oesophagus with demographic and lifestyle factors and the development of adenocarcinoma were examined. RESULTS: 5.5% of patients had prevalent adenocarcinoma (more common in males, older patients, patients diagnosed earlier in the cohort and current or recent smokers). Adenocarcinoma incidence was 23 patients in 3,912 years or 0.59% per annum. Only increased age at diagnosis correlated with an increased risk of incident adenocarcinoma. There was no association with obesity or alcohol history. CONCLUSIONS: Oesophageal adenocarcinoma occurs more commonly in older patients and is more frequent in males than females. Once columnar-lined oesophagus had been diagnosed, there were no other demographic or lifestyle factors which were predictive of the development of incident adenocarcinoma in this cohort.

Does the length of the columnar-lined esophagus change with time?

A number of previous studies have reported patients with Barrett's columnar metaplasia who have an increase or decrease in segment length over time. It is not clear whether patients who have an apparent shortening of the metaplastic segment are subsequently at a lower neoplastic risk and those whose segment length appears to increase are at a higher risk of adenocarcinoma development. The aim of this study was to investigate these issues by studying a large cohort of patients from the UK National Barrett's Oesophagus Registry. Medical records of 1533 patients registered with the UK National Barrett's Oesophagus Registry were examined from seven UK centers. Data were extracted on metaplastic segment length at surveillance endoscopies and histological findings on biopsy. Overall changes in segment length, variability in measurement and probability of the development of dysplasia and neoplasia over time were examined. At least two segment lengths were measured in 763 patients. The median change from measured diagnostic length to most outlying measured segment length was 3.0 cm, but overall there was no tendency for segment length to increase or decrease in the majority of patients with a follow up of up to 20 years. Most patients were treated with proton pump inhibitors. One hundred and eighty-six patients had three or more segment lengths over the first 10 years of follow up. No change in risk was demonstrated in these patients where length appeared to consistently increase with time or when it appeared to decrease. Overall, metaplastic columnar-lined esophagus segment length does not change over time, and when an apparent change is observed, this does not influence a risk of dysplasia or adenocarcinoma.

A single centre's 20 years' experience of columnar-lined (Barrett's) oesophagus diagnosis.

BACKGROUND: The pattern of oesophageal carcinoma type has been changing for some time in a number of countries, with adenocarcinoma becoming more frequent OBJECTIVE: To investigate the prevalence of columnar-lined (Barrett's) oesophagus and oesophageal adenocarcinoma in Barrett's oesophagus during a 20-year period in a single centre. METHODS: All upper gastrointestinal endoscopy and histology reports for the period January 1977 to December 1996 inclusive were reviewed. Data were analysed from patients who had histologically proven Barrett's oesophagus. The data were analysed as a single cohort and in five-year bands according to the date of diagnosis. RESULTS: Of 44,721 endoscopies, 636 Barrett's oesophagus cases were diagnosed; 508 (323 males 185 females; M:F ratio 1.7) were histologically proven. The frequency of Barrett's oesophagus detection increased steadily from 0.2% to 1.6% of all endoscopies per five-year band. The M:F ratio and the mean ages at diagnosis (61 years, range 60-63 for males and 69 years, range 68-79 for females) remained constant throughout. Barrett's oesophagus was diagnosed at a younger age in males (peak 60-69 years) compared to females (peak 70-79 years). The male oesophageal adenocarcinoma incidence (11.1%) was almost twice that in females (6.5%). In the majority (81%), the initial diagnosis of oesophageal adenocarcinoma and Barrett's oesophagus was made concurrently. CONCLUSIONS: The increasing Barrett's oesophagus frequency may reflect an increasing incidence or recognition of this condition or both. Barrett's oesophagus males are more likely to develop oesophageal adenocarcinoma than females.

An initial comparison of nine centres registering patients with the UK National Barrett's Oesophagus Registry (UKBOR).

Initial data from the first nine hospitals registering at least 50 patients each with UKBOR were analysed. This involved 2102 Barrett's oesophagus (BO) cases (M1261:F841), mean 234 patients per centre (range 73-636) and M:F ratio 1.5 (range 1.1-2.3). There was an equal geographical distribution of the hospitals, three each in the north of the country (N), Midlands (Mid) and the south of the country (S). The catchment populations varied from 145,000 to 450,000. The M:F ratio for N, Mid and S was 1.6, 1.3, 1.7, respectively. The mean age at diagnosis in males was 62.0 years (range 53.2-66.3) and in females 67.6 years (range 59.3-73.4), with little geographical variation. The age distribution varied somewhat between the centres; the peak age for males being 40-49 years in one northern hospital, 60-69 years in seven others and 70-79 years in one hospital. For females it was 60-69 years and 70-79 years in each of four hospitals, and 80-89 years in one. The BO diagnosis rate in the under 50s was fairly constant; F mean 14% (range 0-23%); M (eight centres) mean 23% (range 16-27%). However, in one northern centre it was much higher (43%). Information on patients with a diagnosis of oesophageal adenocarcinoma (AC) was available from seven centres. A total of 59 AC were diagnosed (M44:F15, ratio 2.9). The overall mean rate of AC in BO was 3.6% (range 0.5-7.5%). Minor variations in BO patient characteristics may have been due to the hospitals' different policies on diagnostic and reporting criteria. However, the much higher percentage of men under age 50 in the one N centre may reflect a genuine difference in diet and lifestyle, or possibly genetic susceptibility.

Relationship between the intake of high-fibre foods and energy and the risk of cancer of the large bowel and breast.

A recent analysis of Food and Agriculture Organization (FAO) data and mortality data has shown that not all fibre sources are equally protective against colorectal and breast cancers. We correlated the risk of cancers of the colon and breast with the intake of cereals, starchy roots, vegetables, fruits and total energy, either concurrently with the cancer mortality data, or from 20 years earlier. The patterns of the effects of cereals, starchy roots, vegetables and fruits were very different, with cereals and vegetables being protective against both cancers, fruit having no effect and starchy roots having a very weak and non-significant promoting effect. There is strong current interest in the protective effects of fruit and vegetables against cancers at a number of sites. Our analysis showed that only the current intake of vegetables was protective. Intake early in life seemed to offer no protection. The protective effect of cereals was manifest both early in life as well as for current intake for female breast and colorectal cancer, but only for the current period for male colorectal cancer. Calorie restriction, but only early in life, provides protection against all three cancers. Most advice on healthy eating, other than that for small children, is given to (and taken by) the senior age groups and these are the ones likely to benefit. In our study fruit intake was not correlated at all with the risk of either colorectal or breast cancers at either time period. Fruit is clearly more protective against cancers of the upper digestive tract and respiratory tract than against the cancers considered here.

Fat, fish, fish oil and cancer.

There is an ecological association between total and animal fat consumption and colorectal and breast cancer risk. Mortality data for breast and colorectal cancer for 24 European countries correlated, as expected, with the consumption of animal, but not vegetable, fat. There was an inverse correlation with fish and fish oil consumption, when expressed as a proportion of total or animal fat, and this correlation was significant for both male and female colorectal cancer and for female breast cancer, whether the intakes were in the current time period, or 10 years or 23 years before cancer mortality. These effects were only seen in countries with a high ( > 85 g caput-1 day-1) animal fat intake. This evidence suggests that fish oil consumption is associated with protection against the promotional effects of animal fat in colorectal and breast carcinogenesis.

Fish, n-3 fatty acids and human colorectal and breast cancer mortality.

Although there is a strong positive association between total fat consumption and colorectal and breast cancer risk, there is evidence that n-3 fatty acids, mainly found in fish oil, are protective. If this is so, we would expect to be able to detect an inverse correlation between fish consumption and colorectal cancer and breast cancer mortality. Mortality data for breast and colorectal cancer in 24 European countries were correlated with current fish and fish oil consumption and with consumption 10 and 23 years previously. In males there was an inverse correlation between colorectal cancer mortality and current intake of fish (P = 0.036), a weaker correlation with fish consumption 10 years earlier (P = 0.042) and none with consumption 23 years earlier (P = 0.12). The data were not statistically significant in females. There was no correlation at all between breast cancer mortality and fish or fish oil consumption at any time. It is concluded that fish consumption is associated with protection against the later promotional stages of colorectal carcinogenesis, but not with the early initiation stages. The results are discussed in terms of the role of prostaglandins in colorectal carcinogenesis.