| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
Departments of Surgery (A.G.R., J.E.P., S.T.O.) and Endocrinology (S.M.S.), Christie Hospital National Health Service Trust, and Department of Pathology, South Manchester University Hospitals (N.H.), Manchester, United Kingdom M20 4BX; and Departments of Surgery (P.B., J.V.) and Endocrinology (S.G.B.), University of Newcastle, Newcastle, United Kingdom NE2 4HH
Address all correspondence and requests for reprints to: Prof. S. M. Shalet, Department of Surgery, Christie Hospital National Health Service Trust, Wilmslow Road, Manchester, United Kingdom M20 4BX. E-mail: arenehan{at}picr.man.ac.uk
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
10 mm, 27%
vs. 13% and 9%), and of advanced histology (for
tubulovillous, 27% vs. 4% and 22%). No associations
were found between the presence of colonic neoplasia and the duration
of disease, total GH exposure, cure status, and serum insulin-like
growth factor I. This study has failed to demonstrate an increased
prevalence of neoplasia in acromegalic patients compared with the
expected prevalence in the general population and questions the need
for an aggressive colonoscopic screening policy. |
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Against this controversial background, this study aims to clarify the use of colonoscopy in the management of these patients by evaluating the prevalence and characteristics of colorectal neoplasia in a prospectively screened cohort of acromegalics from two centers in the United Kingdom. In the absence of ideal age-matched controls, we calculated the prevalence of occult adenocarcinomas, adenomas, and hyperplastic polyps in the general population from cumulative data acquired from a comprehensive review of the published literature. We also evaluated a number of parameters of disease activity with the presence of colonic polyps.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
A total of 122 patients (male, 69; female, 63) were recruited and underwent large bowel studies at 2 centers (Christie Hospital, Manchester, and Royal Victoria University Hospital, Newcastle). All patients were initially diagnosed with acromegaly between 1973 and 1998, with a median age at diagnosis of 41.9 (range, 17–73) yr. The median age at the time of large bowel examination was 54.8 (range, 25–82) yr. In 4 patients, colonoscopy was performed at the time of diagnosis; for the remainder, the median duration since first presentation was 10.4 (range, 1–25) yr. There were no differences between median ages at initial diagnosis and colonoscopy in either study center. No acromegalic patient had a history of a previous colonic neoplasm, and only 3 had a positive family history of colorectal cancer.
Endocrine evaluation
The initial diagnosis of acromegaly was made on clinical grounds
supported by failure to suppress plasma GH to less than 1mU/L during a
standard oral glucose (75-g) tolerance test (OGTT). Hyperprolactinemia
was documented in 18% of patients at initial diagnosis. Classification
of pituitary adenomas into macroadenomas (10 mm) or
microadenomas (<10 mm) was based on computerized tomography or
magnetic resonance scanning; macroadenoma was present in 75 (mean
± SE pre-op GH, 99.4 ± 8.9 mU/L), microadenoma was
present in 30 (mean pre-op GH, 38.2 ± 8.8 mU/L), and tumor size
was indeterminable in 17 patients (mean pre-op GH, 33.2 ± 8.9
mU/L). The main treatment modality was combined surgery and
radiotherapy in 57 (47%), surgery only in 43 (35%), radiotherapy only
in 13 (11%), and medical therapy alone in 9 patients.
A number of parameters of disease severity were determined using case note review, including age at diagnosis, total GH exposure, and clinical activity. The total GH exposure for each patient was determined by calculating the area under the curve using Biosoft FigP computer software (Biosoft, Cambridge, UK), of mean GH values at multiple time points: initial diagnosis, postsurgery, and annual or biannual values thereafter. Mean GH values were derived mainly from OGTTs, but GH profiles (eight samples taken at hourly intervals throughout the day) were also included, as this form of assessment shows a significant correlation with results obtained from an OGTT (16). GH exposure before presentation was estimated by backward extrapolation based on duration of symptoms. GH status at the time of colonoscopy was defined as active if GH concentrations were more than 5 mU/L and/or serum insulin-like growth factor I (IGF-I) values were greater than the normal age-matched range (17). Note was taken where acromegaly was complicated by diabetes mellitus, as this condition is associated with an increased risk of colorectal cancer in nonacromegalic populations (18). In addition, height and weight were measured at the time of colonoscopy, and body mass index was calculated.
Colonoscopic and pathological examination
Colonoscopies were performed by 2 endoscopists (J.E.P. and P.B.) using either an Olympus Corp. (Southend-on-Sea, UK) CF200HL or CF 230L videoendoscope (both 150 cm in length). Vigorous bowel preparation was used with either Klean-prep (Norgine, Hartford, UK; 4 sachets) or Picolax (Ferring Pharmaceuticals Ltd., Langley, UK; 2 sachets) with Senokot (1 sachet). In many cases, colonoscopy was technically challenging, and real-time three-dimensional magnetic scanning (19) was used to assist negotiation of difficult bowel loops and to confirm that the cecum had been reached. Additionally, using this instrument, total bowel length was measured in 25 acromegalics and compared with that in 42 age-matched nonacromegalics undergoing screening colonoscopy. The time from endoscope insertion to final withdrawal was also measured in 38 acromegalics and compared with that for 90 control colonoscopies performed by the same endoscopist (J.E.P.) in the same endoscopy suite.
All lesions were removed and/or biopsied endoscopically, examined histologically, and classified as neoplastic (adenomas and adenocarcinomas) or nonneoplastic (i.e. hyperplastic polyps). Adenomas were further classified as tubular, tubulovillous/villous, and serrated, and epithelial dysplasia was graded as mild, moderate, or severe (20, 21). Tumor sites were classified as right colon (cecum, ascending and transverse colon), left colon (descending and sigmoid colon), and rectum (distal 18 cm).
Controls
We carried out a comprehensive review of the literature searching for data on age prevalence of occult carcinomas, adenomas, and hyperplastic polyps in the general population. We included published studies after 1960, as before this date the criteria for the histological diagnosis of adenomatous and hyperplastic polyps were ill defined. Two groups of studies emerged. The first group comprised 8 autopsy studies (22, 23, 24, 25, 26, 27, 28, 29) totaling 3559 bowel examinations (male/female ratio, 2073:1486; age range, 20–100 yr). All studies were from populations with a similar colorectal cancer incidence as that in the United Kingdom and excluded autopsies for colorectal cancer-related deaths. The second group comprised 4 screening colonoscopy studies (30, 31, 32, 33) of asymptomatic subjects with average risk of colorectal cancer (i.e. without a family history of colorectal cancer or other known predispositions). In total, there were 810 bowel examinations (male/female ratio, 595:215; age range, 50–82 yr).
Assays
Plasma GH data were obtained from the clinical notes. Early determinations used in-house double antibody RIAs; more recently, a two-site immunoradiometric assay (NETRIA, London, UK) has been employed in both centers. Assay sensitivities were always less than 0.5 mU/L, and within- and between-batch coefficients of variation were less than 15% at all measurable analyte concentrations. During the majority of the study period, both centers participated in the United Kingdom External Quality Assessment Scheme for GH. Before 1990, levels were reported using the First International Reference Preparation of Human GH for immunoassay (66/217), and thereafter, using the First International Standard of Human Pituitary GH (80/505). Results were reported in milliunits per L and adjusted appropriately for differences between standards (34).
Serum IGF-I, IGF-II, IGF-binding protein-2 (IGFBP-2), and IGFBP-3 levels were determined in 60 patients from fasting blood samples taken at the time of colonoscopy. Serum was stored at -80 C before assaying. IGF-I was measured, after acid/alcohol extraction, by an in-house RIA using a polyclonal rabbit antiserum (R557A) raised against purified human IGF-I (35). IGF-II was determined using a commercially available extraction immunoradiometric assay kit (Diagnostic Systems Laboratories, Inc., Webster, TX), and IGFBP-2 and IGFBP-3 were measured using RIA and immunoradiometric assay reagents (Diagnostic Systems Laboratories, Inc.). The intraassay coefficients of variation on all four assays were less than 10% at low, medium, and high measurable analyte concentrations. The sensitivities of the assays are 14, 15, and 5 ng/mL and 0.5 µg/mL, respectively for IGF-I, IGF-II, IGFBP-2, and IGFBP-3.
Statistical analysis
Estimates of age prevalence of adenomas and hyperplastic polyps
in the general population were generated from linear regression models
for the two groups of control studies (model 1, autopsy studies; model
2, screening colonoscopy studies). As data on age bands were expressed
differently between studies, midpoints of age bands were plotted
against prevalence. Comparisons of prevalences were performed using
2 statistics with Yate’s corrections as
appropriate. In addition, multiple logistic regressions were
undertaken, testing for differences in prevalences between all studies
and the present study, adjusting for age, gender, and weight of
study.
Age-dependent parameters in acromegalics, such as serum IGF-I, were compared to age-matched values generated from an in-house dataset of 295 healthy individuals (age range, 20–90 yr: 162 males and 133 females) using the above assays. Distributions of continuous variables were compared using Student’s t tests (paired and unpaired) and Mann-Whitney U tests as appropriate. P < 0.05 was considered to indicate statistical significance. SPSS 7.0 (SPSS, Inc., Chicago, IL) and SPlus (MathSoft, Seattle, WA) were used for computations.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Three patients (2.6%) had previously undiagnosed adenocarcinomas. An additional 11 patients had at least 1 adenoma, providing an overall prevalence rate for colonic neoplasia of 12% (14 of 115). There were no neoplastic lesions in the 3 acromegalics with a positive family history of colorectal cancer. Eighteen acromegalics (16%) had hyperplastic polyps, of whom 5 also had adenomas.
Adenocarcinomas
All three adenocarcinomas were located in the rectum and
rectosigmoid areas, occurring in a 49-yr-old male (Dukes C), a
79-yr-old female (Dukes A), and an 81-yr-old female (Dukes B). For the
autopsy and screening colonoscopy studies, the detection rates of
occult colorectal cancers were 2.3% and 0.9%, respectively. There
were no statistically significant differences between these rates and
that for the acromegalics (Table 1
).
|
Prevalence rates of adenomas varied considerably between studies
in both autopsy and screening colonoscopy study groups, but general
trends were apparent. The prevalence of adenomas increased with age in
both groups, although the incremental increase was steeper in the
screening colonoscopy studies (Fig. 1A).
Similar to adenomas, there were considerable variations in age
prevalences of hyperplastic polyps between studies, but again some
general trends emerged. The prevalence of hyperplastic polyps increased
steadily below the age of 50 yr and thereafter plateaued (Fig. 1B).
|
For acromegalics, as in the control models, the prevalence of
adenomas increased with age; for age bands less than 40, 40–49,
50–59, 60–69, and 70+ yr, the rates were 0%, 8%, 12%, 20%, and
21%, respectively. No increase in prevalence rates was seen in any age
band compared with the control models (Table 2). In addition, the multiple regression
analysis revealed no instance where age/sex/study weight-adjusted
prevalence estimates were lower than the acromegalic series. The
prevalence of hyperplastic polyps in acromegalics showed a pattern of
change similar to that in the autopsy study model; for age bands less
than 40, 40–49, 50–59, 60–69, and 70+ yr, the rates were 11%, 16%,
12%, 20%, and 16%, respectively. Again, there was no increase in
prevalence of hyperplastic polyps in acromegalics compared with the
control models.
|
There were 22 adenomas in 11 patients; 5 patients had 2 or more
adenomatous polyps. The histological types were 12 tubular, 6
tubulovillous, and 4 serrated adenomas. The pathological
characteristics of the adenomas arising in acromegalics were compared
with those in the 2 control groups. Although category percentages
varied considerably between studies, some differences between controls
and acromegalics were noted; namely, adenomas in acromegalics tended to
be right sided, larger, and of more advanced histology, but
paradoxically, a less advanced degree of dysplasia was identified in
acromegalics (Table 3). By contrast,
there were 37 hyperplastic polyps in 18 patients, but accepting
variations between control studies, the site and size distributions for
hyperplastic polyps in acromegalics were broadly similar to those in
the controls.
|
For acromegalics, the sigmoid loop length was almost twice that
for matched controls (mean ± SD, 63 ± 22
vs. 39 ± 15 cm; P < 0.001; Fig. 2, A and B). Total bowel length,
determined at maximum endoscope insertion, was also significantly
increased in acromeglics (132 ± 22 vs. 114 ± 16
cm; P < 0.001; Fig. 2, C and D). In addition,
endoscopic procedure times were longer in acromegalic patients compared
with controls; for nonpolypectomy cases, the total procedure times
(mean ± SD) were 38.6 ± 14.6
vs. 26.3 ± 8.1 min (P < 0.001), and
for polypectomy cases, they were 48.6 ± 18.0 vs.
30.7 ± 11.8 min (P = 0.001), respectively.
|
A number of parameters of disease activity and the presence of
colonic polyps in acromegalics were evaluated (Table 4). There were no differences in terms of
median age at diagnosis, median age at colonoscopy, or median time
since diagnosis. Consistent with the literature studies (data not
shown), adenomas in acromegalics were more frequent among males, but
the difference was not statistically significant (14% vs.
10%). There was no association between the total GH exposure and the
presence of neoplasia, but at the time of colonoscopy, 62 acromegalics
(54%) had active disease, and in these patients, the prevalence of
neoplasia tended to be higher (18% vs. 6%;
P = 0.08). Acromegaly was complicated by diabetes
mellitus in 13 patients (11%) in whom there was no colonic neoplasia.
Similar analyses were performed, evaluating parameters of disease and
the presence of hyperplastic polyps, revealing no significant
associations (data not shown).
|
Serum IGF-I, IGF-II, IGFBP-2, and IGFBP-3 were measured in sera
from 60 patients taken at the time of colonoscopy. Not surprisingly,
IGF-I, IGFBP-3, and the IGF-I/IGFBP-3 molar ratio were significantly
elevated in acromegalics compared with age- and sex-matched controls
(Table 5). Unexpectedly, however, both
serum IGF-II and IGFBP-2 levels were elevated in acromegalics. Among
the acromegalic patients, there were no associations between serum
IGF-I, IGF-II, IGFBP-2, and IGFBP-3 and the presence of adenomas.
Similar analyses were undertaken evaluating the relationships between
serum IGF-I, IGF-II, IGFBP-2, and IGFBP-3 and the presence of
hyperplastic polyps, but no significant associations were found (data
not shown).
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The present study set out to test the hypothesis that the prevalence of colorectal adenomas was increased in acromegalic patients. In the absence of ideal controls, it is our belief that the generation of regression models based on autopsy and screening colonoscopy studies provided the best estimates for age-matched comparisons. Colonoscopic screening studies were disadvantaged by small sample sizes and a lack of data for persons under 50 yr, whereas the autopsy studies were more robust and inclusive of wider age ranges. Criticisms that autopsy studies may represent a magnifying glass examination of the large bowel, and hence overestimate prevalences may be exaggerated, as the proportions of lesions less than 5 mm in these studies were not excessive. The regression models generated from these data allowed us to compare our acromegalic data for all age ranges with the control data and also to test control data from other studies. Thus, for instance, Delhougne et al. (9) reported an increased prevalence of adenomatous polyps in acromegalics (23%) compared with control subjects referred for irritable bowel syndrome (8%); the latter percentage was clearly an underestimate of prevalence for all but the youngest age bands.
All colonoscopies in our study were performed by two experienced endoscopists, and only complete examinations were included in the analysis to prevent underestimations. The overall prevalence rate for colonic neoplasia of 12% in this study is similar to the range of 9–15% reported in some acromegalic series (4, 5, 12), but less than the range of 22–38% reported by others (3, 6, 7, 8, 9, 10). Differences in prevalence rates may reflect sex and age characteristics as well as inherent variations between studies. However, these apparent high prevalence rates are not increased compared with our control values; thus, for example, in a large series from St. Bartholomew’s Hospital, United Kingdom, Jenkins et al. (10) reported adenoma prevalence rates of 29%, 36%, and 39% for age bands 50–59, 60–69, and more than 69 yr; these rates are not increased compared to control values in our models.
At first glance, the detection of 3 occult adenocarcinomas may be a cause for concern. However, when this was compared with the 2 control groups, it did not represent an increased prevalence. There were 60 cancers (2.3%) of 2559 autopsies, and 7 cancers (0.9%) in 810 screening colonoscopies; the higher rates in autopsy studies (and our study) may reflect the inclusion of older patients.
Hyperplastic polyps are not premalignant precursors, although they may be biomarkers for the presence of adenomas, and their inclusion in the current analysis was to test the hypothesis that acromegaly may predispose to hyperplastic conditions such as that reported for prostate hyperplasia (36). Our results showed no increase in the prevalence of colonic hyperplastic polyps in acromegalics.
Previous colonoscopy studies in acromegalics have suggested that adenoma characteristics may differ from those of the general population; namely, that they are more proximal (right sided), larger in size, and of more advanced histology (7, 9, 10). To some extent, our data support these observations, although the sample sizes of the subgroups are too small to be conclusive. Nevertheless, a thesis evolves that suggests that although the prevalence of colonic neoplasia may not be increased, there are characteristics of colonic polyps in acromegaly that are distinct and merit further study.
Similar to the findings of others (37), serum IGF-I, IGFBP-3, and the IGF-I/IGFBP-3 molar ratio were significantly elevated in our acromegalics compared with those in age- and sex-matched controls. By contrast, serum IGF-II, although it is minimally GH dependent, was also elevated in our patient cohort. There were no associations between the presence of colonic neoplasia and serum IGF-I, IGFBP-3, total GH exposure, and current disease status, which tends to negate hypotheses that implicate these growth factors in the pathogenesis of colonic neoplasia in acromegaly. We observed that both serum IGF-II and IGFBP-2 are elevated in individuals with sporadic adenomas (38), but in the presence of overall elevations of these peptides in our acromegalics, these associations were not replicated.
There is little doubt that the detection and removal of adenomas significantly reduce the incidence of subsequent colorectal cancer (39). However, due to the lack of randomized trials, there is no direct evidence on which to base clinical guidelines for colonoscopic screening, and herein lies the controversy. In the U.S., consensus guidelines recommend a single colonoscopy in average risk individuals at age 50 yr (39). There is an increased risk of colorectal cancer in those who have first degree relatives with colorectal cancer (relative risk, 2.76 for two or more affected first degree relatives) (40), and consequently, it is normal practice to offer such individuals early colonoscopy at age 40 yr, with regular surveillance thereafter even in the presence of a normal colonoscopic examination. Despite a more modest colonic cancer risk (Standard Incidence Ratio, 1.68; P = 0.06) reported for acromegalics (14), some researchers recommend a colonoscopic screening policy similar to that for persons with a strong family history of colorectal cancer (15). The present study failed to demonstrate an increased prevalence of colorectal neoplasia in acromegalic patients and, as such, does not support such an aggressive approach.
|
Acknowledgments |
|---|
Received February 29, 2000.
Revised May 10, 2000.
Accepted May 24, 2000.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  E. Resmini, A. Tagliafico, L. Bacigalupo, G. Giordano, E. Melani, A. Rebora, F. Minuto, G. A. Rollandi, and D. Ferone Computed Tomography Colonography in Acromegaly J. Clin. Endocrinol. Metab., January 1, 2009; 94(1): 218 - 222. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Bogazzi, C. Cosci, C. Sardella, A. Costa, L. Manetti, M. Gasperi, G. Rossi, L. Bartalena, and E. Martino Identification of Acromegalic Patients at Risk of Developing Colonic Adenomas J. Clin. Endocrinol. Metab., April 1, 2006; 91(4): 1351 - 1356. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Terzolo, G. Reimondo, M. Gasperi, R. Cozzi, R. Pivonello, G. Vitale, A. Scillitani, R. Attanasio, E. Cecconi, F. Daffara, et al. Colonoscopic Screening and Follow-Up in Patients with Acromegaly: A Multicenter Study in Italy J. Clin. Endocrinol. Metab., January 1, 2005; 90(1): 84 - 90. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Colao, D. Ferone, P. Marzullo, and G. Lombardi Systemic Complications of Acromegaly: Epidemiology, Pathogenesis, and Management Endocr. Rev., February 1, 2004; 25(1): 102 - 152. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Parkinson, W. M. Drake, M. E. Roberts, K. Meeran, G. M. Besser, and P. J. Trainer A Comparison of the Effects of Pegvisomant and Octreotide on Glucose, Insulin, Gastrin, Cholecystokinin, and Pancreatic Polypeptide Responses to Oral Glucose and a Standard Mixed Meal J. Clin. Endocrinol. Metab., April 1, 2002; 87(4): 1797 - 1804. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Melmed CLINICAL PERSPECTIVE: Acromegaly and Cancer: Not a Problem? J. Clin. Endocrinol. Metab., July 1, 2001; 86(7): 2929 - 2934. [Full Text] [PDF]
|
|
|
|
|
|
P. J. Jenkins and M. Besser CLINICAL PERSPECTIVE: Acromegaly and Cancer: A Problem J. Clin. Endocrinol. Metab., July 1, 2001; 86(7): 2935 - 2941. [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
, screening colonoscopy studies. Informative
data for age prevalences in adenomas were obtained from Refs. 22–33,
and data for hyperplastic polyps were obtained from Refs. 24–29 and
31. For adenomas, the regression equations for autopsy and screening
colonoscopy studies were: model 1, y =
0.82x - 15.6; and model 2, y =
1.30x - 54.3. For hyperplastic polyps, the
regression equations were: model 1, y =
1.34x - 37.4 for x > 50 yr,
and y = -0.01x + 28., for x 