Article Text


  1. G Aithal,
  2. P Moayyedi,
  3. I D Penman,
  4. S P L Travis

Statistics from

COX-2 inhibitors: complications and costs ▸

Cyclooxygenase-2 (COX-2) inhibitors have a lower risk of GI bleeding with a similar analgesia efficacy but are more expensive than NSAIDs. Spiegel et al report a systematic review on the adverse events of COX-2 compared with NSAIDs and constructed a Markov model comparing the cost-effectiveness of these two alternatives. Their meta-analysis suggested that COX-2 inhibitors had a lower risk of dyspeptic symptoms (relative risk (RR) 0.77; 95% confidence intervals (CI) 0.74 to 0.81) and clinically significant ulcer complications (RR=0.50; 95% CI 0.40 to 0.63) compared to NSAIDs. The authors assumed that all NSAID patients with dyspepsia would be investigated with endoscopy, be prescribed long term proton pump inhibitor therapy and switch to a COX-2 inhibitor. They also assumed that there would be no difference in cardiovascular events between the two options. Despite these assumptions in favour of COX-2 inhibitors the model suggested this strategy would cost $275 809/QALY gain compared with prescribing Naproxen in 60 year old patients. If patients had a previous bleeding ulcer the cost would reduce to $55 803/QALY gained. This was cost effective for US budgets but would be too expensive for the UK and many European health care systems. COX-2 inhibitors are currently too expensive to recommend as alternatives to NSAIDs for average risk patients Modelling from the perspective of other health care systems are needed to assess the cost effectiveness of COX-2 inhibitors in high-risk groups.

Old, new, borrowed, and blue? ▸

Many of us practice surveillance colonoscopy despite the lack of evidence demonstrating efficacy in preventing cancer and the need to take possibly at least 40 to 50 biopsies to be confident of detecting dysplasia. Adapting techniques used for detecting dysplasia in Barrett’s oesophagus, Kiesslich et al randomised 174 patients with longstanding UC to either conventional colonoscopy (n=81) or magnification colonoscopy using a zoom colonoscope combined with 0.1% methylene blue dye staining (n=84). Five biopsies were taken every 10 cm and all visible lesions were also targeted for biopsy and assessed by blinded pathologists. The authors hypothesised that chromoendoscopy would detect threefold more neoplastic lesions and this was indeed the case (35 lesions in 13 of 84 patients v 11 in 6 of 81 undergoing conventional colonoscopy, p=0.003). Significantly more flat or depressed lesions were detected by chromendoscopy (21) compared to colonoscopy (4). Of the 121 macroscopically visible lesions in chromoendoscopy patients, only 35 were neoplastic on biopsy, the remainder being hyperplastic or inflammatory. These lesions were all classified using the Japanese “pit pattern diagnosis” method before biopsy and this correlated well with the results of pathology as 80 of the 86 non-neoplastic lesions were pit pattern type I or II. Rather than committing ourselves to more biopsies, if corroborated by others, magnification and dye staining together might allow us to take fewer, targeted biopsies with a greater chance of detecting neoplasia. Once again, maybe the answer is in front of us, if we will only look carefully enough and design good quality randomised endoscopic trials like this one?

Not thinning but sieving blood for colitis? ▸

Treatment of active ulcerative colitis by the physical removal of circulating granulocytes and monocytes through adsorptive apheresis created much interest at this year’s Digestive Disease Week (DDW). Activated granulocytes and macrophages in the mucosa are the main source of pro-inflammatory cytokines in active UC. Two types of cellulose-acetate bead column are under evaluation: the Adacolumn for granulocyte/monocyte apheresis (GMA) (Japan Immunoresearch Laboratories, Japan) and the Cellsorba FX for leucocytapheresis (LCA) (Asahi Medical, Japan).

In this uncontrolled study 31 patients who had not responded to oral or intra venous prednisolone (up to 80mg/day) received GMA. The 31 came from a cohort of 146 patients with severe UC, all having three or more Truelove & Witts’ criteria. A further eight patients, who had deteriorated after responding to mesalazine or sulphasalazine also received GMA. GMA was given once a week for up to 11 weeks. No information is given about concomitant medical therapy. At week 12, 82% of the 39 patients had responded or entered remission. The CRP declined from a mean 85 mg/L to 14 mg/L at 12 weeks. Only four out of 146 came to surgery.

Something is happening, but quite what is unclear. It beggars belief that 146 patients with severe colitis coped with an undefined period of oral 5-ASA therapy, followed by an undefined period of oral or intravenous steroids and then a further 12 weeks of GMA, after which only four came to colectomy. In the UK the colectomy rate in 116 patients with severe colitis defined according to the same criteria was 28% (

). Nevertheless, apheresis looks interesting. At DDW 2003, there were seven abstracts on GMA for UC and four on LCA, reporting case series of 5–60 patients. The only controlled study reported response or remission in 8/10 for LCA and 3/9 for sham apheresis (

). Larger controlled trials are essential and clinicians involved in recruiting patients must insist on this.

NAFLD if you do and NAFLD if you don’t! ▸

Non-alcoholic fatty liver disease (NAFLD) is now recognised as a leading cause of chronic liver disease. Clinical and histological spectrum of NAFLD has been mostly described in subjects being investigated for elevated liver enzymes. Mofrad and colleagues identified 51 subjects with normal alanine aminotransaminase (ALT) values (study group) from a total of 386 cases of NAFLD. The major indications for liver biopsy in this group were unexplained hepatomegaly (21/51) and living donor for transplantation (16/51). This group was compared with 50 consecutive cases of NAFLD with elevated ALT values (control group). There were no differences in the demographic and clinical parameters between the two groups with 80% of subjects in both groups manifesting one or more features of metabolic syndrome. More worryingly, there was no difference in the degree of histological changes between the two groups. Twelve (24%) subjects with normal ALT had bridging fibrosis and six (12%) had cirrhosis. Diabetes mellitus was the only independent predictor of advanced fibrosis.

This study emphasises yet again that ALT values do not correlate with the severity of liver injury in patients with chronic liver disease and normal ALT does not guarantee freedom from advanced fibrosis. The findings raise challenging questions as to whether one could predict or identify advanced fibrosis in patients with NAFLD, how to monitor the progress of liver disease, and how to evaluate the efficacy of any treatment without resorting to liver biopsy on each occasion.

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.