Population – based studies of gastrointestinal symptoms in diabetes mellitus
Population-based studies of gastrointestinal symptoms in diabetic patients have been relatively few and the results conflicting. To date, a total of nine population-based studies have been undertaken evaluating gastrointestinal symptoms in subjects with diabetes mellitus. Dyck et al. studied 102 patients with type 1 and 278 patients with type 2 diabetes by interview.
They were selected randomly from a cohort of individuals who were diagnosed with diabetes mellitus (1.3% of the total population) in the community of Rochester, Minnesota, USA (n = 870, 23% with type 1 and 77% with type 2 diabetes). This represents an underestimate because of the relatively high frequency of undiagnosed type 2 diabetes. Symptoms of “gastroparesis” were reported by none of the subjects with type 1 diabetes and by only 1% of subjects with type 2 diabetes. Nocturnal diarrhoea was reported by just 1% of these with type 1 diabetes and 0.5% with type 2 diabetes. The diagnostic criteria for gastroparesis and nocturnal diarrhoea were not stated and no control group was included.
Among a population of 125 subjects who were first diagnosed as having type 1 diabetes between 1960 and 1969 in the Swedish county of Orebro, Schvarcz et al. surveyed 110 eligible subjects using a questionnaire that was previously validated for use in the general population. The prevalence of gastrointestinal symptoms was significantly higher among diabetic patients than among age- and sex-matched controls who were selected from a taxation register.
In particular, anorexia (17.8% vs. 3.6%), vomiting (12.2% vs. 3.0%) and abdominal distension (42.3% vs. 24.4%) were more frequent amongst subjects with diabetes. However, the population studied was small, and only middle-aged patients who had long-standing type 1 diabetes were enrolled.
In a survey of 624 subjects with diabetes who were on a drug reimbursement register and 648 controls from the population register of Kuopio, a Finnish community, Janatuinen et al. studied both subjects with type 1 (n = 87; mean age: men, 53 years; women, 56 years) and type 2 diabetes (n = 451; mean age: men, 56 years; women, 58 years). Subjects with type 1 diabetes had a mean disease duration of 17 years, while for those with type 2 diabetes the mean disease duration was 9 years. No differences were observed with respect to the prevalence of dysphagia, nausea, vomiting, abdominal pain, diarrhoea or constipation, and overall the prevalence of gastrointestinal symptoms was low. Frequent vomiting (once a week or more often) was experienced by 5% of patients, frequent abdominal pain (>. once a week) by 26%, constipation ‘usually or always’ by 16% and frequent diarrhoea (>. once a week] by 5%. However, the questionnaire used had not been validated, and patients with non-insulin-dependent diabetes mellitus who were on diet therapy only were not studied.
In another Scandinavian study, Spangeus et al. investigated subjects with diabetes aged 24 – 59 years and sex- and age-matched controls living in the Swedish county of Umea. Patients were identified by checking the registration forms of 14 primary care centres within the county. The healthy controls were medical students and hospital staff. All were mailed a validated questionnaire that was previously used by Schvarcz et al.. The response rate among the diabetics was 59% and among the controls was 53%. Half of the patients were female and most of the responders were identified as type 1 diabetics (200 vs. 61 type 2 diabetics). The medical records of the responders were checked for glucose control, body mass index, medications and diabetes-specific complications. Patients with both type 1 and type 2 diabetes reported gastrointestinal symptoms more often than the control group. Patients with type 1 diabetes had an increased frequency of constipation (19.5% vs. 6.5% in controls); nocturnal urgency, feelings of incomplete rectal evacuation and straining were also more frequent compared to controls. In contrast, patients with type 2 diabetes had a higher frequency of abdominal pain (28.3% vs. 14.3%) and faecal incontinence (4.9% vs. 0%); they also had a higher prevalence of a nocturnal urgency, feelings of incomplete evacuation at defecation and a need to strain at defecation. Diarrhoea was not more frequent in patients with diabetes compared to controls. Patients with signs of neuropathy had a higher frequency of gastrointestinal symptoms compared to patients who had no signs of neuropathy. Other diabetic complications, such as retinopathy and nephropathy, were not associated with a higher frequency of gastrointestinal symptoms. However, the results of this study are hard to interpret, since an inadequate response rate was achieved, the patients and control subjects were not randomly selected, the proportion with type 1 diabetes was inappropriately high and the methodology to identify diabetes complications was not standardised.
Ricci et al. reported on the frequency of upper gastrointestinal symptoms in a US national sample of patients with diabetes mellitus and controls who were identified by a telephone survey. Of the 874 patients who identified themselves as diabetes sufferers, 483 completed a structured interview evaluating the presence of gastrointestinal symptoms within the past month. Two-thirds of the participants were women and the age range was 18 – 70+ years. The type of diabetes was not determined. Among the patients with diabetes, 50% reported an upper gastrointestinal symptom in the past month compared with 38% in the control group. Bloating and early satiety were more frequent in diabetics than in controls. The frequently of abdominal pain and nausea and vomiting, however, were similar in both of the groups.
A small population-based study from Olmsted County, Minnesota, evaluating the prevalence of gastrointestinal symptoms, was performed by Maleki et al.. The authors detected no differences in the prevalence of most gastrointestinal symptoms between type 1 and type 2 diabetes and controls. A slightly increased prevalence of constipation and laxative use in type 1 patients (27% vs. 19% in controls) was related to calcium channel blocker use, but not to autonomic neuropathy.
Another study was performed inWestern Sydney, Australia. These investigators assessed the frequency of gastrointestinal symptoms in 113 diabetics from an outpatient clinic, 400 diabetics that were selected at random from a diabetes support group, and a random sample of the general population (n = 1000) using a validated questionnaire; the response rates were 100%, 71% and 63%, respectively. After adjusting the results for age, sex and body mass index, none of the gastrointestinal symptoms reported was more frequent in the random diabetes population than in the control population. However, dysphagia, bloating, abdominal pain, constipation and diarrhoea were more frequent in outpatients with diabetes compared to the random diabetes population and controls. The authors also concluded that gastrointestinal symptoms may be related to glycaemic control, since the prevalence of nausea and dysphagia was greater in outpatients with glycated haemoglobin levels >. 10 mg%. Other data support this conclusion.
In a large study from Australia, Bytzer et al. mailed a short questionnaire containing questions on the frequency of troublesome gastrointestinal symptoms and diabetic status to a random sample of 15 000 randomly chosen adults; 60% responded. Overall, 4.9% of the responders reported diabetes (95% of whom were type 2), who were generally older than controls. The authors evaluated the frequency of five symptom complexes, i.e. oesophageal (heartburn and/or dysphagia), upper gut dysmotility, any bowel symptom, diarrhoea and constipation.
After adjusting for age and gender, all symptom complexes were more frequent in diabetics than in controls, and the symptoms nausea, diarrhoea or constipation and faecal incontinence were independently associated with diabetes.
In conclusion, there is evidence that gastrointestinal symptoms are linked with diabetes mellitus, but the prevalence over and above the general population is at most only modestly increased. Some studies have failed to detect an association between diabetes and gastrointestinal symptoms, but several confounders may have obscured the findings. For example, it is well documented that chronic gastrointestinal symptoms are common in non-diabetics in the community, presumably due to functional gastrointestinal disorders such as the irritable bowel syndrome. Moreover, the presence of diabetic complications and possibly long-term glycaemic control appear to be important factors in symptom onset. This may explain the difficulty in establishing a firm link between diabetes and chronic gastrointestinal complaints in population-based studies.