Article Text



Statistics from

When is a cyst a tumour? ▸

Conventional wisdom states that 90% of cysts of the pancreas are secondary to pancreatitis and only 10% are neoplastic. In this study, the prevalence of pancreatic cysts was evaluated in patients undergoing axial imaging of the abdomen. Of more than 24 000 magnetic resonance imaging or computed tomography scan reports which were searched, 290 patients (1.2%) had pancreatic cysts. Of these, 168 patients had no documentation of pancreatitis in their medical records, suggesting a prevalence of neoplastic cysts of 0.7%. The authors also assessed growth of cystic tumours by comparing the size of the cysts in those patients who had undergone follow up scans. Of 79 patients who had undergone follow up scans, the cyst had increased in size in 15 (19%), had not changed in size in 47 (59%), and had decreased in size in 17 (22%) patients. The mean interval between scans was 16 months. Neither sex, cyst size, nor cyst location predicted neoplasia but symptomatic patients were more likely to have premalignant or malignant pathology (mainly mucinous lesions). Age over 70 years was also a predictor of malignancy. Only 41% of patients with cysts detected in this study had a history of pancreatitis, which is consistent with other recently published studies. Cysts in the pancreas are much more likely to be neoplastic cysts than was previously realised. Clinicians should not be fooled by a small pancreatic cyst—it could be premalignant or even malignant.

All are normal; BUT some are more normal ▸

The normal range of serum aminotransferase levels is based on distributions in apparently healthy populations, with two SD above the mean being considered the upper limit of normal. But, significant chronic liver disease can exist with minimal or no elevation in liver enzymes. In a large cohort (representing 11% of the Korean population), Kim et al have investigated the relationship between serum aminotransferase concentrations within the normal range (<40 IU/l) and mortality from liver disease. After excluding those who had any known diseases at baseline and those who died within the first year of follow up, 94 533 men and 47 522 women, aged 35–59 years, were enrolled in the analysis. The average of two enzyme measurements at two year intervals was used, and mortality from liver disease during eight years of follow up was estimated including only those who had a previous hospital admission for liver disease. Compared with the lowest concentration of aminotransferase (<20 IU/l), the adjusted relative risks (RR) for mortality from liver disease for levels of 20–29 IU/l were 2.5 (2.0–3.0) in men and 3.3 (1.7–6.4) in women, while for levels of 30–39 IU/l, RR were 8.0 (6.6–9.8) in men and 18.2 (8.1–40.4) in women. Based on receiver operating characteristic curves, a serum alanine transaminase level of 30 IU/l and an aspartate transaminase level of 31 IU/l were the best cut off values to identify those at risk of death from liver disease.

These findings challenge the standard method of determining the normal range. The redefined cut off for serum aminotransferase based on clinical outcome is arguably more relevant. Using the lowered cut off, more cases of chronic liver disease could be detected but with a lower positive predictive value. However, these effects would be less marked in selected groups (such as subjects with metabolic syndrome) with a higher likelihood of liver disease.

Good news? ▸

Alcohol consumption has long been regarded as a risk factor for colorectal cancer, with some studies finding a fivefold increase in risk for heavy drinkers compared with non-drinkers. What has not been clear is the overall magnitude of the increase and how much it is due to an increased risk of rectal cancer in beer drinkers. The pooled analysis of Cho et al used data from eight large cohort studies (five North American and three European) involving nearly 5000 colorectal cancers occurring in half a million subjects. One limitation was the need to rely on a single baseline self report of alcohol intake. Reassuringly, no increase in risk was found with reported alcohol intakes of <30 g/day (1 unit = 8 g alcohol). Drinking more than this was associated with a 25% increase in risk (relative risk 1.23 (95% confidence interval (CI) 1.07–1.42)) and risks were slightly higher for wine drinking (1.82) rather than beer (1.37) or spirits (1.21). Beer drinkers did have the greatest increase in risk of rectal cancer (1.59) but only slightly more so than wine drinkers (1.55). On the basis that 4% of women and 13% of men reported intakes of >30 g/day, the population attributable risk (risk avoided if the high risk group reduced intakes to that of low risk groups) was estimated to be 0.9% for women and 5.0% for men.

Clearly, this powerful analysis suggests that the positive association with colorectal cancer is due to alcohol itself rather than a specific beverage. Of several possible mechanisms for this association, recent attention has focussed on the contribution of alcohol to DNA hypomethylation which is exacerbated by low intakes of folate and/or methionine. Remarkably, the risk of colorectal cancer in multivitamin users and those with the highest methionine intakes showed no relationship with high alcohol intakes. So, taking multivitamins might be an alternative for those who will not reduce their alcohol consumption.

View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.