Impact of chronic gastrointestinal symptoms in diabetes mellitus on health-related quality of life

THE AMERICAN JOURNAL OF GASTROENTEROLOGY © 2001 by Am. Coll. of Gastroenterology Published by Elsevier Science Inc.

Vol. 96, No. 1, 2001 ISSN 0002-9270/01/$20.00 PII S0002-9270(00)02143-2

Impact of Chronic Gastrointestinal Symptoms in Diabetes Mellitus on Health-Related Quality of Life Nicholas J. Talley, M.D., Ph.D., F.A.C.G., Lisa Young, B.Sc., Peter Bytzer, M.D., Johann Hammer, M.D., Melanie Leemon, M.Nutr.Diet, Michael Jones, Ph.D., and Michael Horowitz, Ph.D., F.R.A.C.P. Department of Medicine, University of Sydney, Nepean Hospital, Kingswood, New South Wales, Australia

OBJECTIVES: Morbidity from GI symptoms in diabetes is considered to be high, but no studies have quantified the impact of GI symptoms in diabetes on health-related quality of life. We hypothesized that diabetics reporting increased GI symptoms would experience more impaired quality of life. METHODS: Subjects from the community with diabetes (n ⫽ 892) and outpatients with diabetes (n ⫽ 209) were recruited for this study. Subjects were divided into type 1 (diabetes diagnosed at age ⬍30 yr and requiring insulin) and type 2. A validated questionnaire measuring GI symptoms and diabetes status and the Short Form-36 were completed. The results were compared with Australian normal data. GI symptom groups measured were frequent abdominal pain, bowel-related abdominal pain, reflux, dyspepsia, constipation, diarrhea, and fecal incontinence. RESULTS: There was a clinically significant decrease in quality-of-life scores in diabetics compared with population norms across all subscales. The impact on quality of life in diabetes was predominantly observed in type 2 diabetics. The quality-of-life scores in all subscales decreased markedly with increasing numbers of distinct GI symptom groups, and this was similar in community and outpatient diabetics. For all the Short Form-36 subscales, GI symptom groups were significantly (all p ⬍ 0.0001) associated with poorer quality of life in diabetes, independent of age, gender, smoking, alcohol use, and type of diabetes. CONCLUSIONS: GI symptoms impact negatively on healthrelated quality of life in diabetes mellitus. (Am J Gastroenterol 2001;96:71–76. © 2001 by Am. Coll. of Gastroenterology)

INTRODUCTION The prevalence of GI symptoms seems to be increased in people with diabetes mellitus compared with the general population (1–3), although not all studies have agreed (4). In a recent population-based study involving 8185 nondiabetics and 423 diabetics randomly selected from the population living in Western Sydney, Australia, all upper and lower GI symptoms evaluated were more common in diabetics than in controls (3). However, the impact of these GI complaints on

health is unclear. If they merely represent a minor annoyance, not requiring therapy or driving health care seeking, then the GI complaints should be ignored. On the other hand, if they substantially impair quality of life, management becomes important. The impact of GI symptoms in diabetes has been a largely neglected field of research. The term health-related quality of life (HRQOL) refers to those aspects of health that impact on an individual’s physical, emotional and social functions, and also includes perceptions about the relative importance of those features of life affected by infirmity (5). There is good evidence that diabetes per se affects not only physical health but also the social, mental, and functional aspects of day-to-day living (6 –10). Although diabetes mellitus impairs quality of life, exactly which aspects of the condition are causally involved remains unclear. In diabetes mellitus, quality of life may potentially influence treatment compliance, which in turn may impact on the progression of the disease for better or worse. There is very little available information regarding the nature of the relationship between quality of life and GI symptoms among people with diabetes mellitus. In the present study, we aimed to determine the impact of GI symptomatology on HRQOL in persons with type 1 and type 2 diabetes, as compared with the general population. It was hypothesized that those diabetics who reported more GI symptoms would experience the greatest impairment in quality of life, relative to diabetics without GI symptoms or to nondiabetics.

MATERIALS AND METHODS Subjects DIABETES MELLITUS. A total of 1101 subjects with diabetes (aged ⬎17 yr) participated in the study. They were recruited from two distinct populations, as described here: 1) Community diabetics. A random sample of n ⫽ 1800 was identified using the mailing lists of Diabetes Australia (NSW branch), a nonprofit lay organization with approximately 100,000 registered individuals. The original contact list was stratified into four equal-sized groups of 1) men using insulin, 2) men not using insulin, 3) women using insulin, and 4) women not using insulin.

72

Talley et al.

AJG – Vol. 96, No. 1, 2001

Each individual was mailed a copy of a validated questionnaire, along with a cover letter, a A$1 scratch lottery ticket (as an incentive), and a reply-paid envelope. Follow-up letters were mailed at 3 and 6 wk after the initial mailout. A total of 892 completed questionnaires was received, yielding a response rate of 50%. 2) Outpatient diabetics. A sample of 209 consecutive outpatients (51% women) attending either the Nepean Hospital Diabetes Service in Sydney or the Diabetes Center of the Royal Adelaide Hospital in Adelaide, Australia, were also asked to complete the questionnaire. A total of 70 patients using insulin and 139 diabetic patients not using insulin were studied, 59% from the Sydney clinic and 41% from the Adelaide clinic. Those who had had previous GI surgery, dementia, or major psychosis were excluded.

Table 1. Comparative Demographics and Observed Prevalence Rates for Gastrointestinal Symptom Groups in Subjects With Diabetes

Diabetic subjects were classified according to the World Health Organization criteria (1994), such that persons who were diagnosed with diabetes at age ⬍30 yr and had used insulin from that time were classified as type 1 diabetics, and all others were classified as type 2 diabetics (11). Across the outpatient and Diabetes Australia populations, 145 (13%) persons were classified as having type 1 diabetes, and 956 (87%) were classified as having type 2 diabetes mellitus.

1) Frequent abdominal pain. Abdominal pain (moderate, severe, or very severe) occurring once a week or more frequently. 2) Bowel-related abdominal pain. Abdominal pain relieved by defecation or related to changes in either stool form or stool frequency. 3) Reflux. Heartburn or acid regurgitation at least once a month. 4) Dyspepsia. Epigastric pain or discomfort for ⬎3 months. 5) Constipation. This was broadly defined as any of the following: usual frequency of bowel movements ⬍2 per week; need for manual evacuation; need for laxatives because of constipation; a feeling of blockage in the anus (at least “often”); hard or lumpy bowel movements (at least “often”); incomplete evacuation (at least “often”); or straining with bowel movements (at least “often”). 6) Painless diarrhea. Loose or watery stools (at least “very often”) without abdominal pain. 7) Fecal incontinence. Leakage of bowel movements once a month or more frequently.

AUSTRALIAN NORMAL DATA. The quality-of-life score from the diabetic populations were compared with Australian Normal data. All diabetics (mean age, 60 yr) were compared with a normative population group aged 55 to 64 yr (n ⫽ 2033), and type 1 diabetics (mean age 40 yr) were compared with a normal population aged 35 to 44 yr (n ⫽ 4110; see Ref. 12). Health-Related Quality of Life The Short-Form 36 (SF-36) was used as an HRQOL measure and presented as a subsection of the GI questionnaire (13). It contains 36 individual items that cluster into eight subscales: physical functioning, role—physical, bodily pain, general health, vitality, social functioning, role— emotional, and mental health. The SF-36 is a widely used disease– generic measure that has been established to be both valid and reliable for the Australian population (12–14). Higher scores indicate better quality of life, and differences of ⱖ5 points are generally considered clinically significant. GI Symptom Groups GI symptoms were assessed by a standardized questionnaire, based on the extensively validated Bowel Symptom Questionnaire (15, 16). The validity of the GI and diabetic items in the questionnaire has been established by applying a test-retest procedure for reliability plus concurrent validity measures via an independent physician interview (Talley et al., unpublished). The following chronic GI symptom clusters experienced over the past 12 months were defined a priori:

Cohort With Diabetes Mellitus Demographic Group Women (%) Mean age, yr (SD) Smokers, past or present (%) ⱕ2 drinks/wk (%) Number of GI symptom groups reported (%) None 1 2 3⫹

All N ⫽ 1101

Type 1 N ⫽ 145

Type 2 N ⫽ 956

46.3 60 (14.7) 52.1 67.8

54.5 40 (12.5) 32.3 65.5

47.2 63 (12.1) 55.1 68.1

54.6 22.6 12.0 10.8

57.9 22.8 9.7 9.7

54.1 22.6 12.3 11.0

Statistical Analyses The prevalence of GI symptom groups was compared between type 1 and type 2 diabetes using the Pearson ␹2 test. Analysis of variance and covariance was used to assess the independence of the effects of diabetes and GI symptom groups (nil, single, multiple) on quality of life both before and after controlling for a number of potential confounding factors. All p values calculated were two tailed. The alpha level of significance was set at p ⬍ 0.05.

RESULTS Demographic characteristics of the diabetes patients are summarized in Table 1. Type 2 diabetics were older and more likely to be males, compared with type 1 diabetics and control subjects. These differences were controlled for in the multivariate analyses. GI symptoms were common in the population of diabetics

AJG – January, 2001

Diabetes Mellitus: Chronic Gastrointestinal Symptoms and HRQOL

73

Table 2. Number and Prevalence of Gastrointestinal Symptom Constellations Across All Subjects With Diabetes, and Type 1 and Type 2 Diabetes Group

Abdominal Pain

Bowel-Related Pain

Dyspepsia

Reflux

Diarrhea

Constipation

Incontinence

All Diabetics Type 1 Type 2

96 (8.7%) 11 (7.6%) 85 (8.9%)

149 (13.5%) 19 (13.1%) 130 (13.6%)

155 (14.1%) 21 (14.5%) 134 (14.1%)

158 (14.4%) 13 (9.0%) 145 (15.2%)

28 (2.5%) 3 (2.1%) 25 (2.6%)

270 (24.5%) 38 (26.2%) 232 (24.3%)

74 (6.7%) 5 (3.4%) 69 (7.2%)

(Table 2). The prevalence of GI symptom groups and the proportion of subjects having any given number of GI symptoms was similar across the type 1 and type 2 diabetics (Table 2). The population and symptom characteristics were similar in outpatients and community diabetics, and therefore, these groups were combined (data not shown). The quality-of-life scores across the populations evaluated (diabetes and Australian normal values) are shown in Figure 1. In all scales, there was a clinically significant decrease in quality of life in diabetes compared with established population normal values. The impact on quality of life of type 2 diabetics was greater than that observed in type 1 diabetics, compared with nondiabetics (Fig. 2). It is of note that the mean subscale scores for type 1 diabetics and controls were comparable for all domains, except general health, role—physical, vitality, and social functioning. Thus, the impact on quality of life in diabetes was predominantly a function of type 2 diabetics. The differences between type 1 and type 2 diabetics were marked for the physical functioning and role—physical subscales, exceeding 20 points in both cases. There were also substantial differences in general health, vitality, and role— emotional, where the difference was ⬎5 points. The quality-of-life scores in all subscales decreased

markedly with increasing numbers of distinct GI symptom groups in the population of diabetics (Fig. 3). Similar results were evident in the community and outpatient diabetics, and hence the results were combined (data not shown). The decrease in mean score varied somewhat among subscales but was typically around 10 points per GI symptom category (nil, one or two, multiple), a clinically significant decrease. The largest effects were seen in role—physical, bodily pain, and vitality, in which the smallest and largest means differed by 30 points or more. The least effect was observed for mental health, where the smallest and largest means differed by approximately 15 points. For all of the subscales, GI symptom groups were significantly (p ⬍ 0.0001) associated with poorer quality of life in diabetes, independently of age, sex, smoking, alcohol, and type of diabetes in a logistic regression model. When considering the effects of diabetes type (1 vs. 2) and GI symptoms simultaneously, GI symptoms were found to have an effect on all domains of quality of life, independent of diabetes type (all p ⬍ 0.001), whereas this was not always the case for the effect of diabetes type after controlling for GI symptoms. After controlling for the effect of GI symptoms, the effect of diabetes type did not reach statistical significance for role—physical (p ⫽ 0.7), pain (p ⬎

Figure 1. Effect of diabetes status on quality of life scores across SF-36 subscales.

74

Talley et al.

AJG – Vol. 96, No. 1, 2001

Figure 2. Effect of diabetes type on quality of life compared with age-relevant Australian norms.

0.9), vitality (p ⫽ 0.2), social functioning (p ⫽ 0.7), and mental health (p ⫽ 0.1). In a further multivariate analysis, the effect of diabetes type on quality of life was found to be confounded by patient characteristics on a number of quality-of-life domains. After controlling for age, gender, and alcohol and tobacco use, the effect of diabetes was significant for pain (p ⫽ 0.03) but not for physical functioning (p ⬎ 0.9), role—physical (p ⫽ 0.1), general health (p ⫽ 0.5), vitality (p ⫽ 0.8), social functioning (p ⫽ 0.1), role— emotional (p ⬎ 0.9), and mental health (p ⬎ 0.9). The effect of GI symptoms, however, remained statistically significant on all

domains of quality of life after controlling for diabetes type and patient characteristics.

DISCUSSION Information about the relationship between quality of life and GI symptoms among people with diabetes mellitus is limited. We have shown that GI symptom complexes are associated with reduced HRQOL in both type 1 and 2 diabetes in outpatients and in a community sample of people with diabetes. This is the first study to specifically address this issue, although others have observed, using the SF-36,

Figure 3. Effect of GI symptoms (Sx) on quality of life scores across SF-36 subscales.

AJG – January, 2001

Diabetes Mellitus: Chronic Gastrointestinal Symptoms and HRQOL

that differences in HRQOL in diabetes were driven primarily by diabetic complications (17). Diabetes mellitus is a chronic illness known to impair quality of life (6 –10). Our results confirm these observations and show that diabetes particularly impairs physical functioning and general health perceptions. GI symptom complexes in diabetes had a notable impact not only on these two domains but on all HRQOL dimensions measured by the SF-36. These results are important because GI symptoms in diabetes are common. In this study, we observed that constipation, dyspepsia, abdominal pain, and reflux symptoms were the most frequent set of complaints reported. Although controversial (4), several studies have suggested that GI symptoms are more prevalent in those with diabetes than in healthy controls (2, 3). It was recently reported that in 483 US subjects with diabetes, 50% reported one or more upper-GI symptoms, compared with 38% in matched controls (18). Similar results were observed in Sweden (19). We adjusted for a number of factors that may influence quality of life, including type of diabetes. We observed that type 2 diabetics, who were older, had more impaired quality of life; however, adjusting for diabetes type as well as age, there remained a significant impact of GI symptoms on quality of life. The association of impaired quality of life and GI symptoms also remained when gender, smoking, and alcohol use were considered. Another potential confounder is glycemic control, which we did not consider in our analyses. A few studies have examined the relationship between glycemic control and quality of life, with varying outcomes. Some investigators have found no association (20), suggesting that the demands of achieving very tight control might obviate the less tangible benefits, such as a reduction in the number or severity of complications. By contrast, Nerenz et al. demonstrated an “inverted U-shape” association between glucose control and quality of life in type 1 diabetes so that both tight and poor control resulted in poorer quality of life than a middle-road option (21). Among type 2 diabetics, a linear relationship existed such that better control was associated with improved quality of life. A number of other studies, however, have found no evidence of a relationship between glycemic control and quality of life in diabetic subjects (22, 23). The relationship between quality of life, glycemic control, and GI symptoms now needs to be carefully examined prospectively. The present study was relatively large and was strengthened by the use of both a validated GI symptom questionnaire and a well-established generic quality-of-life measure. We could not obtain data on GI symptom status or on quality of life in the nonresponders from the community diabetic population, but because similar results were obtained in outpatients and community diabetics, any selection bias is likely to have been small. In conclusion, quality of life seems to be significantly impaired by GI symptoms in type 1 and type 2 diabetes. Our results imply that successful management of GI symptoms

75

in diabetics is likely to improve quality of life, although this hypothesis will require prospective evaluation. Reprint requests and correspondence: Nicholas J. Talley, M.D., Ph.D., Professor of Medicine, Department of Medicine, Nepean Hospital, Clinical Sciences Building, Corner of Somerset & Derby Streets, Kingswood, NSW 2747 Australia. Received May 15, 2000; accepted Aug. 15, 2000.

REFERENCES 1. Feldman M, Schiller LR. Disorders of gastrointestinal motility associated with diabetes mellitus. Ann Intern Med 1983;98: 378 – 84. 2. Schvarcz E, Palmer M, Ingberg CM, et al. Increased prevalence of upper gastrointestinal symptoms in long-term type 1 diabetes mellitus. Diabet Med 1996;13:478 – 81. 3. Bytzer PM, Talley NJ, Leemon M, et al. Diabetes mellitus is associated with an increased prevalence of gastrointestinal symptoms: Results from a population-based survey of 15,000 adults. Gastroenterology 2000;118:A716 (abstract). 4. Enck P, Rathmann W, Spiekermann M, et al. Prevalence of gastrointestinal symptoms in diabetic patients and non-diabetic subjects. Z Gastroenterol 1994;32:637– 41. 5. Last JM, ed. Dictionary of epidemiology. Oxford, UK: Oxford University Press, 1995. 6. Larsson D, Lager I, Nilsson PM. Socio-economic characteristics and quality of life in diabetes mellitus—Relation to metabolic control. Scand J Public Health 1999;2:101–105. 7. Silvers D, Kipners M, Broadstone V, et al. Domperidone in the management of symptoms of diabetic gastroparesis: Efficacy, tolerability, and quality-of-life outcomes in a multicenter controlled trial. DOM-USA-5 Study Group. Clin Ther 1998;20: 438 –53. 8. Mayou R, Bryant B, Turner R. Quality of life in non-insulin dependent diabetes and a comparison with insulin-dependent diabetes. J Psych Res 1990;34:1–11. 9. Stewart, AL, Greenfield S, Hayes RD, et al. Functional status and well-being of patients with chronic conditions. JAMA 1989;262:907–13. 10. Ward J, Lin M, Heron G, et al. Comprehensive audit of quality-of-care and quality-of-life for patients with diabetes. J Qual Clin Pract 1997;17:91–100. 11. World Health Organization. Prevention of diabetes mellitus. WHO Technical Report Series. Geneva: World Health Organization, 1994:844. 12. Lin M, Ward JE. Reliability of the MOS SF-36 health status measure in Australian general practice. Aust Fam Physician 1998;27(suppl 2):S94 – 8. 13. Brazier JE, Harper R, Jones NMB, et al. Validating the SF-36 health survey questionnaire: New outcome measure for primary care. BMJ 1992;305:160 – 4. 14. McCallum, J. The SF-35 in an Australian sample: Validating a new, generic health status measure. Aust J Public Health 1995;19:160 – 66. 15. Talley NJ, Phillips SF, Melton J III, Wiltgen C, Zinsmeister AR. A patient questionnaire to identify bowel disease. Ann Intern Med 1989;111:671– 4. 16. Talley NJ, Boyce P, Owen BK, et al. Initial validation of a bowel symptom questionnaire and measurement of chronic gastrointestinal symptoms in Australians. Aust NZJ Med 1995;25:302– 8. 17. Gulliford MC, Mahabir D. Relationship of health-related quality of life to symptom severity in diabetes mellitus: A

76

Talley et al.

study in Trinidad and Tobago. J Clin Epidemiol 1999;52: 773– 80. 18. Ricci JA, Siddique R, Stewart WF, Sandler RS, Sloan S, Farup CE. Upper gastrointestinal symptoms in a U.S. national sample of adults with diabetes. Scand J Gastroenterol 2000;35: 152–9. 19. Spangeus A, El-Salhy M, Suhr O, Eriksson J, Lithner F. Prevalence of gastrointestinal symptoms in young and middleaged diabetic patients. Scand J Gastroenterol 1999;34:1192– 202. 20. Hanestad BR, Ho¨rnquist JO, Albrektsen G. Self-assessed quality of life and metabolic control in persons with insu-

AJG – Vol. 96, No. 1, 2001

lin-dependent diabetes mellitus. Scand J Soc Med 1991;19: 57– 65. 21. Nerenz DR, Repasky DP, Whitehouse FW, et al. Ongoing assessment of health status in patients with diabetes mellitus. Med Care 1992;30:MS112–24. 22. Weinberger M, Kirkman S, Samsa GP, et al. The relationship between glycaemic control and health-related quality of life in patients with non-insulin dependent diabetes mellitus. Med Care 1994;32:1173– 81. 23. Wredling R, Stålhammar J, Adamson U, et al. Well-being and treatment satisfaction in adults with diabetes: A Swedish population-based study. Qual Life Res 1995;4:515–22.