Self-Reported Gastrointestinal Symptoms in Type 2 Diabetes Improve With an Intensive Lifestyle Intervention: Results From the Action for Health in Diabetes (Look AHEAD) Clinical Trial

Look AHEAD was a multicenter, randomized, controlled trial evaluating the effect of an intensive weight loss program in overweight or obese individuals with type 2 diabetes on major cardiovascular events. Volunteers were aged 45–76 years at enrollment with a BMI ≥25 kg/m² (27 kg/m² if using insulin), A1C <11% (<97 mmol/mol), systolic blood pressure <160 mmHg, diastolic blood pressure <100 mmHg, and triglycerides <600 mg/dL (18). These individuals underwent a maximal graded exercise test to ensure that exercise could be safely prescribed and completed 2 weeks of self-monitoring. All informed consent procedures were approved by local institutional review boards before use, and participants signed consent forms. The trial was registered at ClinicalTrials.gov (identifier: NCT00017953).

The ILI was designed to achieve and sustain an average group weight loss of ≥7%, primarily through diet modification and increased physical activity. Caloric intake goals were 1,200–1,500 for individuals weighing <250 lb at baseline and 1,500–1,800 for individuals weighing >250 lb. Diets were developed to avoid large glycemic loads and maximize cardiovascular health. As such, they included a maximum of 30% of total calories from fat, a maximum of 10% of total calories from saturated fat, and a minimum of 15% of total calories from protein (19). The physical activity component of the ILI consisted mostly of home-based exercise, with a goal of 175 min/week of moderate-intensity physical activity.

The first 6 months of ILI included three group meetings and one personal session per month. For the remainder of the first year, individual sessions remained the same, but group meetings became biweekly. The leaders of each session were interventionists trained in nutrition and exercise counseling. In months 13–48, participants attended monthly individual meetings that were followed ∼14 days later with phone calls or e-mails by the interventionists. Optional monthly group meetings were also offered during these months.

During the initial 6 months, lifestyle strategies were the main focus of ILI. Beginning in month 7, the “toolbox” algorithm began, which included the optional use of a weight loss medication (orlistat) and/oradvanced behavioral strategies for participants who had not achieved the 10% individual weight loss goal (19).

Participants assigned to DSE were invited to three group sessions annually throughout the 4-year study period (20). These sessions utilized a standardized protocol and focused on diet, physical activity, or social support. Information on behavioral strategies was not presented, and participants were not weighed at these sessions.

At baseline and annually thereafter, participants self-reported the presence and severity of the following GI symptoms within the past 4 weeks: abdominal pain above the navel, abdominal pain below the navel, constipation, diarrhea, feeling very full after eating little, heartburn, nausea, bloating or distention, regurgitation, and vomiting. Each symptom was rated on a scale ranging from 0 to 3, where 0 indicated that a symptom did not occur, and 1, 2, and 3 indicated mild, moderate, or severe symptoms, respectively. Participants who underwent gastric bypass surgery were excluded from analyses.

Information on demography, smoking, alcohol use, and cardiovascular disease (CVD) history was based on self-report. Weight and height were measured in duplicate using a digital scale and stadiometer. A maximal graded exercise test was administered as a measure of fitness (METS) (21). One MET is approximately resting metabolism; 4 METS approximates walking on flat ground at just under 4 miles per hour. Fasting A1C was analyzed by the Central Biochemistry Laboratory (Northwest Lipid Research Laboratories, University of Washington, Seattle, Wash.) using standardized laboratory procedures. Participants brought current prescription medications to assessments to update medication records. Hypertension was based on use of antihypertensive medications or measurement >140/90 mmHg. Depression was based on a Beck Depression Inventory (BDI) (22) score ≥11, indicating elevated depression symptoms. Staff collecting assessments were masked to intervention assignment.

χ² and t tests were used to assess the balance between intervention groups at baseline. The prevalence of GI symptoms across follow-up between groups was assessed using generalized estimation equations with baseline prevalence as reference. The odds of transitioning to more severe GI symptoms from baseline between intervention groups was assessed using multivariate multinomial mixed models with covariate adjustment for potential confounding baseline factors (23). The impact of including 1-year weight change as a covariate was assessed in the full model. We also examined whether symptoms varied among ILI participants grouped according to patterns of weight loss: those who maintained year-1 weight loss at year 4, lost at year 1 then gained at year 4, gained at year 1 then lost at year 4, and had no loss at either years 1 or 4 (24). Use of insulin and metformin were included as time-varying covariates in supporting analyses. The effect of orlistat use on GI symptom severity over time was evaluated by excluding orlistat users. All analyses were conducted using SAS version 9.3 (SAS Institute, Cary, N.C.).

Our analyses included the 4,986 (96.9%) of 5,145 Look AHEAD participants who provided at least one follow-up assessment of GI symptoms during the first 4 years and had not had bariatric surgery at the time of their follow-up visit. Visits were excluded if they occurred after bariatric surgery, resulting in exclusion of 11 visits at year 1, 32 visits at year 2, 51 visits at year 3, and 71 visits at year 4. Baseline characteristics were similar between the intervention groups (Table 1), except for BDI score, for which slightly more participants in ILI had elevated depressive symptoms (14.7% in ILI vs. 12.4% for DSE, P = 0.02).

The ILI participants included in this report lost an average of 8.6% (SD 6.8%) of their BMI at year 1, compared to 0.7% (SD 4.6%) for DSE participants. At year 4, mean losses were 4.5% (SD 7.6%) for ILI participants and 0.7% (SD 7.2%) for DSE participants. Differences in weight losses between groups were highly significant (P <0.0001) throughout all 4 years of follow-up.

At baseline, bloating was the most common GI symptom, being reported by 40% of participants, with >12% rating it as either moderate or severe. The next most commonly reported symptom was heartburn, experienced by 35.3% of participants, with 8.4% considering it either moderate or severe. More than 32% of participants reported having diarrhea, with almost 8% considering it moderate or severe. Constipation was reported by 29% of participants, with 7.1% rating it moderate or severe. The remaining symptoms were reported by <20% of the cohort. The symptom that occurred least often among participants at baseline was vomiting, reported by 4.3% of participants, with only 1.3% reporting it as moderate or severe.

At years 1–4, 96.5, 94.0, 93.4, and 91.9% of the participants provided data on symptoms, respectively; follow-up was balanced between intervention groups. Table 2 presents the prevalence of symptoms at annual assessments by group. In general, the prevalence of symptoms tended to increase with time, with significant (P <0.05) time trends for all except diarrhea. The average post-randomization prevalence across follow-up was significantly lower among ILI than among DSE participants for diarrhea, feeling full after eating little, heartburn, and regurgitation, but significantly higher for constipation. The time course for the prevalence of symptoms also varied for several symptoms. For feeling full after eating little, heartburn, nausea, bloating or distention, and regurgitation, there was a pattern for relative decreases in the prevalence of symptoms among ILI participants at year 1 that waned or disappeared over time. For diarrhea, the relative increase in prevalence among ILI participants that was evident at year 1 also disappeared by year 4.

Figure 1 summarizes the relative intervention effects on the odds of progressing to more severe GI symptoms (e.g., from none to mild, from mild to moderate, and so on) across 4 years, with adjustment for baseline levels of all factors in Table 1. As can be seen, 95% confidence intervals (CIs) favoring overall relative benefit for ILI participants across the 4 years excluded 1.00 for abdominal pain below the navel, diarrhea, fullness, heartburn, regurgitation, and vomiting. The associated odds ratios (ORs) for these symptoms ranged from 0.74 (heartburn) to 0.87 (abdominal pain below the navel), translating to overall 13–26% reductions in the odds of increasing severity. For only one symptom (constipation) was there a (nonsignificant) trend toward relative worsening in symptom severity among ILI participants.

Differences between intervention groups on symptom severity tended to be largest at year 1 and to diminish over time, with significant (P <0.05) attenuation based on tests of interaction for bloating or distention, constipation, fullness, heartburn, nausea, and regurgitation. Figure 2 portrays the characteristic longitudinal pattern, as seen for heartburn. Including year-1 weight change in the models of symptom severity attenuated the effects of treatment group such that there were no longer significant differences for the main effect of treatment. Weight-change patterns over time were related to increase in bloating or distention, as well as feeling full after eating little for participants who did not lose weight at either year 1 or year 4 (bloating OR 2.00 [95% CI 1.27–3.13]; fullness OR 1.84 [1.17–2.90]) and in diarrhea for participants who lost at year 1 and then gained at year 4 (OR 1.29 [1.04–1.60]) compared to participants who maintained their weight loss at years 1 and 4.

Use of metformin increased over time among DSE participants (P <0.001); the OR (95% CI) of current compared to baseline use rose from OR 1.16 (1.09–1.24) at year 1 to OR 1.49 (1.36–1.64) at year 4. Use of metformin initially decreased among ILI participants (OR 0.90 [0.85–0.96]) at year 1, but was unchanged at year 4 (OR 1.09 [0.99–1.18]). Overall, metformin use was associated with a greater prevalence of diarrhea (OR 1.88 [1.67–2.11]), feeling full after eating little (OR 1.18 [1.03–1.36]), heartburn (OR 1.28 [1.13–1.45]), nausea (OR 1.25 [1.10–1.42]), and regurgitation (OR 1.19 [1.03–1.37]). Inclusion of metformin use as a time-varying covariate comparing the prevalence of symptoms between intervention groups, however, did not materially affect results.

Use of insulin was associated with an increased prevalence of bloating or distention (OR 1.20 [1.05–1.39]), feeling full after eating little (OR 1.33 [1.14–1.57]), nausea (OR 1.29 [1.12–1.50]), and vomiting (OR 1.37 [1.12–1.69]). Similar to metformin, including insulin use as a time-varying covariate did not materially alter the results from comparisons of the prevalence of symptoms between intervention groups. Low prevalence of use of exenatide and acarbose prohibited analysis of their effects on GI symptom expression.

At some point during the 4 years of follow-up, 693 participants reported orlistat use (ILI: baseline n = 1, year 1 n = 409, year 2 n = 324, year 3 n = 148, year 4 n = 64; DSE: baseline n = 0, year 1 n = 10, year 2 n = 3, year 3 n = 2, year 4 n = 1). Excluding these individuals from analyses did not alter findings.

We examined a cluster of GI symptoms that are prevalent in the general population, but more so in individuals with diabetes (6). We found that the Look AHEAD ILI yielded a modest, statistically significant overall reduction in the prevalence and severity of GI symptoms across 4 years. The reductions were largest for bloating, heartburn, and regurgitation. Benefits tended to be greatest during the first year of the intervention, when weight losses were greatest, and to wane over time. Inclusion of 1-year weight losses in models accounted for intervention effects, suggesting that these effects could be at least partially attributable to weight loss. The GI benefits of the intervention were evident after covariate adjustment for metformin and insulin use. Excluding orlistat users from analyses did not materially alter findings.

Participants in the DSE group had significantly higher odds of reporting a more severe symptom at follow-up than they had reported at baseline for several GI symptoms. In contrast, those in the ILI group reported symptoms that were significantly less severe symptoms than at baseline for these same symptoms or no difference in likelihood of a more severe symptom than at baseline. The observed improvements in GI symptoms for the ILI group are consistent with our hypothesis.

It is interesting to note that weight regain at years 1 and 4 appeared to attenuate the beneficial effects to some extent, suggesting that weight maintenance is an important factor in sustaining the apparent benefit of weight loss on GI symptoms. At year 1, those in the ILI group had a loss of 8.6% of initial weight versus 0.7% in DSE, which was significant (P <0.001) (25). This effect of magnitude of weight loss is also likely reflected in the finding that intervention effects tended to be greatest at the point of maximal weight loss at the end of year 1, with attenuation found in subsequent years of follow-up as some weight regain occurred.

The findings of this study mirror the results found in a previous investigation, in which a combination of a healthy diet and higher levels of physical activity, both of which were core features of the ILI in Look AHEAD, resulted in a reduction in GI symptoms in a weight loss intervention (26). Our analyses covered a longer follow-up (48 vs. 24 months) and involved a more controlled intervention. Other studies have also reported a direct relationship between GI symptoms and either BMI or obesity but did not examine the effect of changes in weight on symptoms in a population of people with type 2 diabetes (14,27,28).

Although this study was not designed to identify the mechanism(s) responsible for the beneficial effects on GI symptoms of ILI compared to DSE, it is well known that weight loss interventions have favorable effects on various facets of both physical and psychological functioning in people with diabetes (29–31). More specifically, previous studies have shown that mood disorders can have a significant impact on GI symptoms (3,12,13). Of note, previous research with the Look AHEAD cohort found that ILI reduces depression after 1 year (31).

There are several strengths and some limitations of this study. Its strengths include its large size, the fact that participants were randomized, and that there was significant weight loss in the ILI group. Individuals who volunteer for clinical trials may not represent clinical populations; this may limit the generalizability of our findings. In addition, although this study examined GI symptoms, it is unclear whether the reported reductions in symptoms are clinically meaningful.

In summary, ILI yielded beneficial effects on GI symptoms, with some variability in the strength of these effects depending on the specific symptom and a general increase in the magnitude of the beneficial effect with greater weight loss. Potential modifiers to the effect were analyzed for several variables, but, despite this, ILI retained an association with improvement in GI symptoms. The intervention may have affected other aspects of the study participants’ lives that may, in turn, affect GI symptomology, such as medication use and depression, but we view these as potential mediators of change rather than confounding factors to the interpretation of our findings. Our findings suggest that weight loss through an intensive lifestyle change intervention may be beneficial to many obese individuals with type 2 diabetes who suffer from most common GI symptoms.

RHN performed the analyses and wrote the manuscript. JJR participated in the development, coauthored several sections, and provided feedback on drafts. WBA participated in the development and writing and provided critical feedback. JMC, WCK, GAB, and LJC reviewed the work and provided critical feedback. MAE participated in the development of the manuscript, coauthored several sections, reviewed the work, and provided critical feedback. RHN is the guarantor of this work and, as such, had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. A complete listing if the Look AHEAD clinical sites and staff members can be found online in Supplementary Table 1.

This study is supported by the U.S. Department of Health and Human Services through the following cooperative agreements from the National Institutes of Health (NIH): DK57136, DK57149, DK56990, DK57177, DK57171, DK57151, DK57182, DK57131, DK57002, DK57078, DK57154, DK57178, DK57219, DK57008, DK57135, and DK56992. The following federal agencies have contributed support: NIDDK, NHLBI, National Institute of Nursing Research, National Center on Minority Health and Health Disparities, Office of Research on Women’s Health, CDC, and the Department of Veterans Affairs. This research was supported in part by the NIDDK Intramural Research Program. The Indian Health Service (IHS) provided personnel, medical oversight, and use of facilities. The opinions expressed in this article are those of the authors and do not necessarily reflect the views of the IHS or other funding sources.

Additional support was received from The Johns Hopkins Medical Institutions Bayview General Clinical Research Center (M01RR02719); the Massachusetts General Hospital Mallinckrodt General Clinical Research Center and the Massachusetts Institute of Technology General Clinical Research Center (M01RR01066); the University of Colorado Health Sciences Center General Clinical Research Center (M01RR00051) and Clinical Nutrition Research Unit (P30 DK48520); the University of Tennessee at Memphis General Clinical Research Center (M01RR0021140); the University of Pittsburgh General Clinical Research Center (M01RR000056); the Clinical Translational Research Center funded by the Clinical & Translational Science Award (UL1 RR 024153) and NIH grant (DK 046204); and the Frederic C. Bartter General Clinical Research Center (M01RR01346).

The following organizations committed to making major contributions to Look AHEAD: FedEx Corporation; Health Management Resources; LifeScan, Inc., a Johnson & Johnson Company; OPTIFAST® of Nestle HealthCare Nutrition, Inc.; Hoffmann-La Roche Inc.; Abbott Nutrition; and the Slim-Fast Brand of Unilever North America.

Some of the information contained herein was derived from data provided by the Bureau of Vital Statistics, New York City Department of Health and Mental Hygiene.

No potential conflicts of interest relevant to this article were reported.