Colorectal neoplasia in acromegaly *

The issue of whether patients with acromegaly are at increased risk of colorectal neoplasia was first raised some 20 years ago with a report detailing this association (Klein et al., 1982). Despite numerous subsequent studies confirming these observations, some authors have countered these findings by using early epidemiological surveys of morbidity and mortality in acromegaly which failed to document any increased prevalence of colorectal neoplasia (Wright et al., 1970; Alexander et al., 1980). However, the use of these studies is inappropriate as they were performed prior to the advent of modern neurosurgical and medical therapy and approximately half of the patients had died by the age of 60 years from other causes, principally cardiovascular and metabolic. It is only with current improved practice that patients are surviving long enough to develop the age-associated complication of colorectal neoplasia. In order to clarify this controversy we have performed ongoing studies at St Bartholomew's Hospital into the prevalence of colorectal neoplasia in acromegaly. Including those patients in our original report (Jenkins et al., 1997a), 220 acromegalic patients without colonic symptoms have undergone full colonoscopy: 10 patients (4 ́5%) had an adrenocarcinoma and 45 patients (21%) had a pre-malignant tubulovillous adenoma. In common with a number of other series, our study of the prevalence of colorectal neoplasia in acromegaly is disadvantaged by the lack of an age-matched simultaneous asymptomatic control group. However, an estimate of the relative risk can be gained by comparison with published series of almost 6000 screened asymptomatic non-acromegalic patients (Rex et al., 1993; Imperiale et al., 2000; Lieberman et al., 2000). Using this comparator, the risk of colorectal cancer in patients with acromegaly aged over 50 years is increased between 7and 18-fold (P , 0 ́00001 2 P , 10). A similar risk can be estimated by grouping the data from the numerous other published prospective screening studies of acromegalic patients, involving a total of almost 700 patients (Klein et al., 1982; Ituarte et al., 1984; Brunner et al., 1990; Ezzat et al., 1991; Ladas et al., 1994; Terzolo et al., 1994; Vasen et al., 1994; Delhougne et al., 1995; Colao et al., 1997; Archambeaud-Mouveroux et al., 1998; Renehan et al., 2000; Tzoiti et al., 2000). Despite wide variations in patient age, operator experience and colonic visualization, the overall prevalence of colorectal cancer was 3 ́7%, which represents an increase of between 4and 10-fold above that of the general population. Furthermore, comparison with those series that have used concurrent age-matched control patients gives a relative risk of 13 ́4 (P , 0 ́00001). These latter studies overcome the limitations of using either historical control patients or autopsy data; the latter will greatly exaggerate the prevalence of adenomas given the optimal conditions under which the resected colons are examined and the very small size of adenomas that can be recorded. Thus, we believe the current data clearly demonstrate that colorectal neoplasia is increased in acromegaly and, in common with other authors, we also believe that these patients should be considered as being at high risk for developing this malignancy (Klein, 1984; Ezzat & Melmed, 1991; Melmed et al., 1998; Jenkins et al., 1999; Giustina et al., 2000). Analysis of all these studies of acromegalic patients has revealed several characteristics of this increased prevalence of colonic neoplasia, of which the most important is age. In the Bart's series, the median age of patients with a carcinoma was 66 years. The proportion of those with one or more adenomas increased with each decade so that 40% of those aged over 60 years had such a lesion. This influence of age not only confirms that early epidemiological reviews are obsolete in determining the prevalence of neoplasia, but may also explain the lower prevalence rates in those colonoscopic screening series which have studied younger patients. In addition to agedependency, other characteristics of colorectal neoplasia in patients with acromegaly differ from the general population. The adenomas are often larger, multiple and show increased epithelial cell dysplasia, all of which are bad prognostic signs in terms of malignant potential (Vasen et al., 1994; Colao et al., 1997; Terzolo et al., 1994; Renehan et al., 2000). Furthermore, there is a tendency for the adenomas and cancer to be situated in the proximal colon (Delhougne et al., 1995; Renehan et al., 2000), with two-thirds of lesions being right-sided in one study (Colao et al., 1997). Thus, total colonoscopy is essential. This requirement is made more difficult by the increased length and capacity of the acromegalic colon (Jenkins et al., 1997b; Renehan et al., 2000). To overcome this and the prolonged colonic transit time, acromegalic patients require rigorous bowel preparation, in excess of that usually used. We have found that the PEG-electrolyte bowel preparation `KleanPrep' (Norgine, Harefield, UK) gives good cleansing if two litres are given at each of 6, 4 and 2 hours prior to the