Statistics from Altmetric.com
Editor,—We read with interest a recent paper by Jenkins et al(Gut1999;44:585–587). However, we are concerned with their assertion that acromegaly is a high risk condition for colorectal neoplasia, and their recommended advice on colonoscopic screening and surveillance. Jenkins and colleagues had found that 33 (26%) of 129 patients (updated to 155), treated for acromegaly at St Bartholomew's Hospital, had at least one adenoma and six (5%) had adenocarcinomas.1
We feel that the choice of controls in this study was inappropriate because although there is no ideal control population, the authors used comparative data on the incidence of adenomatous polyps from only two cohorts: a published study of left sided adenomas,1a and colonoscopic records of all patients without acromegaly that had been examined by one of the authors. Matched for age (and side), the relative risk of adenomas was higher in patients with acromegaly when compared with data from the first study, but not when compared with data from the second.
In an attempt to estimate more appropriately the prevalence of adenomas in the normal population, we have carried out a comprehensive review of the literature on adenoma prevalence per decade of life from which two groups of studies emerged. The first group comprised six necropsy studies (n=2914), and the second comprised three screening colonoscopy studies of asymptomatic average risk volunteers (n=720, table1).2-13 With the exception of those patients between 50–59 years old, the prevalence rates of adenomas in patients with acromegaly are remarkably similar to those for individuals in screening colonoscopy studies, and less than those from necropsy studies. We find no evidence that patients with acromegaly are at increased risk of developing adenomas.
For colorectal cancer, Jenkins and colleagues1 used comparative data from a regional cancer registry and estimated increased relative risks of 13–90. These figures are exaggerated in comparison with studies using standardised age and sex adjusted population data. Ron and colleagues14 reported 13 colonic cancers in 1041 male veteran acromegalics (standardised incidence ratio (SIR) 3.1; 95% CI 1.7 to 5.1) and, using uniform methods of ascertainment of cases and comparison groups, Orme and colleagues15 found 12 cases of colonic cancer in a larger study of 1362 patients with acromegaly (SIR 1.68; p=0.06).
There are approximately 1500 patients with acromegaly in the United Kingdom, and it would seem sensible for strategies for large bowel screening to be evidence based. The data given above suggest that the reported increased prevalence of colorectal neoplasia in patients with acromegaly is overestimated and thus, the recommendations given by these authors for early colonoscopic screening and subsequent regular surveillance above that of the normal population cannot be supported by the evidence currently available.
Editor,—Our conclusion that acromegaly may be a high risk condition for the development of colorectal cancer is based not only on our own data (we have now discovered 10 patients with cancer from approximately 210 patients who have had a colonoscopy), but also on those of several other studies. In the prospective studies by Archambeaud-Mouveroux and colleagues1-1 and Ituarteet al,1-2 cancers were discovered in 12.5% and 20% respectively of patients with acromegaly. Retrospective epidemiological surveys may have failed to show an increased risk of cancer because of differences in methodology—for example, in one study case ascertainment depended upon death certificate entries and cancer registrations which may have been incomplete.1-3 Furthermore, these studies did not discuss the relevance of the age of the patients. The mean and range of age during follow up were not stated,1-3 and in another study the mean age of the patients at diagnosis of acromegaly was 52 years old and that at follow up was only 61 years old.1-4 Our results clearly show that colorectal cancer is a late complication of acromegaly, as the mean age of our affected patients was 67 years old.
The situation for adenomas is less clear and we agree that there is a lack of proper control groups. However, many of the prevalence figures given by Renehan and colleagues were obtained from necropsy studies and therefore cannot provide a valid comparison because the resected bowel was thoroughly washed on up to three occasions, repeatedly examined under magnification in optimal lighting, and lesions as small as l mm were classified as adenomas. This gives an increased prevalence of the disease compared with incidences of neoplasia revealed by colonoscopic screening. Furthermore, these studies were of populations with very different demographic, socioeconomic and dietary influences, which are factors known to influence the incidence and prevalence of colonic adenomas. By contrast, our control groups were taken from similar populations to the patients with acromegaly, and in one group, the colonoscopies were performed by the same operator. A comparison between these groups showed a significant increase in the relative risk of adenomas in patients with acromegaly, although we accept that this risk is not as high as that for cancer, and this raises intriguing questions about the causes of colorectal cancer in patients with acromegaly. It is possible that the adenoma–carcinoma sequence in patients with acromegaly differs from that in the non-acromegalic population, or that the cancers arise de novo without an adenomatous stage.
We fully concur that strategies for large bowel screening should be evidence based. Our initial screening recommendations are based on current data and will be modified according to our continuing prospective studies. Subsequent strategies might take into account not only the age of the patients but also the activity of their acromegaly, because preliminary data suggest that those patients in whom the disease is more active (elevated serum IGF-I) are more likely to develop adenomas. Until more multicentre studies involving larger numbers of patients undergoing careful and total colonoscopy allow the risks to be better quantified, it seems prudent for patients to undergo colonoscopic screening every five years, or every three years if an adenoma is found.
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.