Donald A. Williamson’s research while affiliated with Pennington Biomedical Research Center and other places

Importance Age-related macular degeneration (AMD) is a leading cause of blindness with no treatment available for early stages. Retrospective studies have shown an association between metformin and reduced risk of AMD. Objective To investigate the association between metformin use and age-related macular degeneration (AMD). Design, Setting, and Participants The Diabetes Prevention Program Outcomes Study is a cross-sectional follow-up phase of a large multicenter randomized clinical trial, Diabetes Prevention Program (1996-2001), to investigate the association of treatment with metformin or an intensive lifestyle modification vs placebo with preventing the onset of type 2 diabetes in a population at high risk for developing diabetes. Participants with retinal imaging at a follow-up visit 16 years posttrial (2017-2019) were included. Analysis took place between October 2019 and May 2022. Interventions Participants were randomly distributed between 3 interventional arms: lifestyle, metformin, and placebo. Main Outcomes and Measures Prevalence of AMD in the treatment arms. Results Of 1592 participants, 514 (32.3%) were in the lifestyle arm, 549 (34.5%) were in the metformin arm, and 529 (33.2%) were in the placebo arm. All 3 arms were balanced for baseline characteristics including age (mean [SD] age at randomization, 49 [9] years), sex (1128 [71%] male), race and ethnicity (784 [49%] White), smoking habits, body mass index, and education level. AMD was identified in 479 participants (30.1%); 229 (14.4%) had early AMD, 218 (13.7%) had intermediate AMD, and 32 (2.0%) had advanced AMD. There was no significant difference in the presence of AMD between the 3 groups: 152 (29.6%) in the lifestyle arm, 165 (30.2%) in the metformin arm, and 162 (30.7%) in the placebo arm. There was also no difference in the distribution of early, intermediate, and advanced AMD between the intervention groups. Mean duration of metformin use was similar for those with and without AMD (mean [SD], 8.0 [9.3] vs 8.5 [9.3] years; P = .69). In the multivariate models, history of smoking was associated with increased risks of AMD (odds ratio, 1.30; 95% CI, 1.05-1.61; P = .02). Conclusions and Relevance These data suggest neither metformin nor lifestyle changes initiated for diabetes prevention were associated with the risk of any AMD, with similar results for AMD severity. Duration of metformin use was also not associated with AMD. This analysis does not address the association of metformin with incidence or progression of AMD.

Background The 2018 World Cancer Research Fund/American Institute for Cancer Research (WCRF/AICR) 3rd expert report highlights up-to-date Cancer Prevention Recommendations that may reduce burdens of many chronic diseases, including diabetes. This study examined if following a lifestyle that aligns with the recommendations – assessed via the 2018 WCRF/AICR Score – was associated with lower risk of type 2 diabetes in high-risk adults participating in the Diabetes Prevention Program Outcomes Study (DPPOS). Methods The Diabetes Prevention Program (DPP) randomized adults at high risk for diabetes to receive a lifestyle intervention (ILS), metformin (MET) or a placebo (PLB) (mean: 3.2 years), with additional follow-up in DPPOS for 11 years (mean: 15 years total). 2018 WCRF/AICR Scores included seven components: body weight, physical activity, plant-based foods, fast foods, red and processed meat, sugar-sweetened beverages, and alcohol; the optional breastfeeding component was excluded. Scores ranged 0-7 points (with greater scores indicating greater alignment with the recommendations) and were estimated at years 0, 1, 5, 6, 9, and 15 ( N =3,147). Fasting glucose and HbA1c were measured every six months and oral glucose tolerance tests were performed annually. Adjusted Cox proportional hazard ratios (HRs) and 95% confidence intervals (CIs) were used to examine the association of both Score changes from years 0-1 and time-dependent Score changes on diabetes risk through DPP and year 15. Results Scores improved within all groups over 15 years ( p <0.001); ILS Scores improved more than MET or PLB Scores after 1 year ( p <0.001). For every 1-unit improvement from years 0-1, there was a 31% and 15% lower diabetes risk in ILS (95% CI: 0.56-0.84) and PLB (95% CI: 0.72-0.97) through DPP, and no significant association in MET. Associations were greatest among American Indian participants, followed by non-Hispanic White and Hispanic participants. Score changes from years 0-1 and time-dependent Score changes in ILS and PLB remained associated with lower risk through year 15. Conclusions Score improvements were associated with long-term, lower diabetes risk among high-risk adults randomized to ILS and PLB, but not MET. Future research should explore impact of the Score on cancer risk. Trial registration Diabetes Prevention Program: NCT00004992 ; Diabetes Prevention Program Outcomes Study: NCT00038727

OBJECTIVE To determine glycemic and nonglycemic risk factors that contribute to the presence of diabetic retinopathy (DR) before and after the onset of type 2 diabetes (T2D). RESEARCH DESIGN AND METHODS During the Diabetes Prevention Program (DPP) and DPP Outcome Study (DPPOS), we performed fundus photography over time in adults at high risk for developing T2D, including after they developed diabetes. Fundus photographs were graded using the Early Treatment Diabetic Retinopathy Study (ETDRS) grading system, with DR defined as typical lesions of DR (microaneurysms, exudates, hemorrhage, or worse) in either eye. RESULTS By DPPOS year 16 (∼20 years after random assignment into DPP), 24% of 1,614 participants who had developed T2D and 14% of 885 who remained without diabetes had DR. In univariate analyses, using results from across the entire duration of follow-up, American Indian race was associated with less frequent DR compared with non-Hispanic White (NHW) race, and higher HbA1c, fasting and 2-h plasma glucose levels during an oral glucose tolerance test, weight, and history of hypertension, dyslipidemia, and smoking, but not treatment group assignment, were associated with more frequent DR. On multivariate analysis, American Indian race was associated with less DR compared with NHW (odds ratio [OR] 0.36, 95% CI 0.20–0.66), and average HbA1c was associated with more DR (OR 1.92, 95% CI 1.46–1.74 per SD [0.7%] increase in HbA1c). CONCLUSIONS DR may occur in adults with prediabetes and early in the course of T2D. HbA1c was an important risk factor for the development of DR across the entire glycemic range from prediabetes to T2D.

To examine associations between sleep disturbance and changes in weight and body composition, and the mediating role of changes of appetite and food cravings in the POUNDS Lost 2-year weight loss diet intervention trial, this study included 810 overweight or obese individuals with baseline sleep disturbance assessment, who were randomly assigned one of four diets varying in macronutrient composition. Changes in body weight and fat distribution were assessed by dual x-ray absorptiometry (DXA) and computed tomography (CT) during the 2-year intervention. Participants were asked to provide sleep disturbance levels (no, slight, moderate, or great) at baseline and to recall their sleep disturbances since last visit at 6, 12, 18 and 24 months. Weight loss during the first 6 months was followed by 1.5 years of steady weight regain. Participants with greater sleep disturbance from baseline to 6 months showed significant losses of body weight (P trend < 0.001) and waist circumference (P trend = 0.002) at 6 months, after multivariate adjustment. Compared with individuals without sleep disturbance at all from baseline to 6 months, those with slight, moderate, or great sleep disturbance showed an elevated risk of failure to lose weight (−5% or more loss) at 6 months, when the maximum weight loss was achieved, with OR of 1.24 (95% CI 0.87, 1.78), 1.27 (95% CI 0.75, 2.13) or 3.12 (95% CI 1.61, 6.03), respectively. In addition, we observed that the repeatedly measured levels of sleep disturbance over 2 years were inversely associated with the overall weight loss rate (weight changes per 6 months) (P trend < 0.001). Further, sleep disturbances during weight loss from baseline to 6 months and weight regain from 6 months to 24 months were significantly predictive of total fat, total fat mass percent and trunk fat percent changes during the 2 years. Our results also indicated that food cravings for carbohydrates/starches, fast food fats, sweets; cravings, prospective consumption, hunger of appetite measurements; dietary restraint, disinhibition and hunger subscales measured at 6 months significantly mediated the effects of sleep disturbance on weight loss. Our results indicate that more severe sleep disturbance during weight loss is associated with an elevated risk of failure to lose weight during the dietary intervention. Food cravings and eating behaviors may partly mediate these associations.

The origins of theories specifying dietary restraint as a cause of eating disorders can be traced to the 1970s. This paper will present an overview of the origins of dietary restraint theories and a brief historical review of evidence will be summarized. Recent research will be presented, including the results from the CALERIE Phase 1 study, as well as CALERIE Phase 2, which were NIH-sponsored randomized controlled trials. CALERIE 2 provided a test of the effect of two years of caloric restriction (CR) on the development of eating disorder syndromes and symptoms in comparison to a control group that did not alter eating behavior or body weight. The intervention was effective for inducing a chronic (two-year) reduction in total energy expenditure and increased dietary restraint but did not increase symptoms of eating disorders. The results of this recent investigation and other studies have not provided experimental support for conventional dietary restraint theories of eating disorders. These findings are discussed in terms of potential revisions of dietary restraint theory, as well as the implications for a paradigm shift in public health messaging related to dieting.

Objective This study aimed to determine the medical cost impact and return on investment (ROI) of a large, commercial, digital, weight‐management intensive lifestyle intervention (ILI) program (Real Appeal). Methods Participants in this program were compared with a control group matched by age, sex, geographic region, health risk, baseline medical costs, and chronic conditions. Medical costs were defined as the total amount paid for all medical expenses, inclusive of both the insurers’ and the study participants’ responsibility. Results In the 3 years following program registration, the intent‐to‐treat (ITT) cohort had significantly lower medical expenditures than the matched controls, with an average of −$771 or 12$3,693,090 compared with total program costs of $1,639,961 translated into a 2.3:1 ROI. Program completers (n = 3,990), who attended more sessions than the overall ITT group, had greater mean weight loss (−4.4$956 or 14%), and an ROI of 2.0:1 over the 3‐year time frame compared with matched controls. Conclusions The findings demonstrated that the digital weight‐management ILI was associated with a significantly positive ROI. Employers and payers willing to cover the cost of an ILI that produces both weight loss and demonstrated cost benefits can improve health and save money for their population with overweight or obesity.

Aim: To investigate whether the genetic risk score (GRS) for lean body mass (LBM) modified the effects of weight-loss diets on changes in appetite and adiposity among overweight and obese individuals. Participants and methods: In the 2-year Preventing Overweight Using Novel Dietary Strategies (POUNDS Lost) trial, we included 692 adults who were randomly assigned to one of four diets varying in macronutrient intake. A GRS was calculated using five single nucleotide polymorphisms associated with LBM. Results: The LBM-GRS was not associated with the baseline LBM measured by dual-energy x-ray absorptiometry in a subgroup (50%) of the study population. We found that the LBM-GRS had significantly different associations with changes in appetite from baseline to 6 months according to low- or high-fat diet group (P-interaction < 0.001, 0.021, 0.005 and 0.024 for total appetite score, hunger, fullness and prospective consumption, respectively). Lower LBM-GRS (indicating a greater genetic predisposition to LBM) was associated with greater decreases in the total appetite score (P < 0.001), hunger (P = 0.01), fullness (P = 0.001) and prospective consumption (P = 0.019) in participants in the low-fat diet group, whereas no significant associations with these appetite measures were observed in the high-fat diet group. In addition, lower LBM-GRS was associated with greater reduction in body weight (P = 0.003) and waist circumference (P = 0.011) among participants in the low-fat diet group, while no associations were observed in the high-fat diet group. The interactions attenuated, along with weight regain, from 6 months to 2 years. Conclusions: Our findings suggest that genetic variation in LBM may be differentially associated with appetite changes, and may subsequently be related to changes in body weight and waist circumference, according to dietary fat intake.

Objective This study evaluated weight changes after cessation of the 10‐year intensive lifestyle intervention (ILI) in the Look AHEAD (Action for Health in Diabetes) study. It was hypothesized that ILI participants would be more likely to gain weight during the 2‐year observational period following termination of weight‐loss–maintenance counseling than would participants in the diabetes support and education (DSE) control group. Methods Look AHEAD was a randomized controlled trial that compared the effects of ILI and DSE on cardiovascular morbidity and mortality in participants with overweight/obesity and type 2 diabetes. Look AHEAD was converted to an observational study in September 2012. Results Two years after the end of the intervention (EOI), ILI and DSE participants lost a mean (SE) of 1.2 (0.2) kg and 1.8 (0.2) kg, respectively (P = 0.003). In addition, 31% of ILI and 23.9% of DSE participants gained ≥ 2% (P < 0.001) of EOI weight, whereas 36.3% and 45.9% of the respective groups lost ≥ 2% of EOI weight (P = 0.001). Two years after the EOI, ILI participants reported greater use of weight‐control behaviors than DSE participants. Conclusions Both groups lost weight during the 2‐year follow‐up period, but more ILI than DSE participants gained ≥ 2% of EOI weight. Further understanding is needed of factors that affected long‐term weight change in both groups.

Context Eating habits and food craving are strongly correlated with weight status. It is currently not well understood how psychological and behavioral factors influence both weight loss and weight regain. Objective To examine the associations between psychological and behavioral predictors with weight changes and energy intake in a randomized controlled trial on weight loss. Design and setting The Prevention of Obesity Using Novel Dietary Strategies (POUNDSLOST) is a dietary intervention trial, which examined the efficacy of four diets on weight loss over 2 years. Participants were 811 overweight (BMI: 25-40.9 kg/m2, age: 30-70 years) otherwise healthy adults. Results Every 1-point increase in craving score for high-fat foods at baseline was associated with greater weight loss (-1.62kg, P=0.0004) and a decrease in energy intake (r = -0.10, P =0.01) and fat intake (r = -0.16, P <0.0001) during the weight loss period. In contrast, craving for carbohydrates/starches was associated with both less weight loss (P <0.0001) and more weight regain (P =0.04). Greater cognitive restraint of eating at baseline was associated with both less weight loss (0.23kg, P <0.0001) and more weight regain (0.14kg, P =0.0027) whereas greater disinhibition of eating was only associated with more weight regain (0.12kg, P =0.01). Conclusions Craving for high-fat foods is predictive of greater weight loss while craving for carbohydrates is predictive of less weight loss. Cognitive restraint is predictive of less weight loss and more weight regain. Interventions targeting different psychological and behavioral factors can lead to greater success in weight loss.

Efforts to identify a preferable diet for weight management based on macronutrient composition have largely failed, but recent evidence suggests that satiety effects of carbohydrates may depend on the individual’s insulin-mediated cellular glucose uptake. Therefore, using data from the POUNDS LOST trial, pre-treatment fasting plasma glucose (FPG), fasting insulin (FI), and homeostatic model assessment of insulin resistance (HOMA-IR) were studied as prognostic markers of long-term weight loss in four diets differing in carbohydrate, fat, and protein content, while assessing the role of dietary fiber intake. Subjects with FPG <100 mg/dL lost 2.6 (95% CI 0.9;4.4, p = 0.003) kg more on the low-fat/high-protein (n = 132) compared to the low-fat/average-protein diet (n = 136). Subjects with HOMA-IR ≥4 lost 3.6 (95% CI 0.2;7.1, p = 0.038) kg more body weight on the high-fat/high-protein (n = 35) compared to high-fat/average-protein diet (n = 33). Regardless of the randomized diet, subjects with prediabetes and FI below the median lost 5.6 kg (95% CI 0.6;10.6, p = 0.030) more when consuming ≥35 g (n = 15) compared to <35 g dietary fiber/10 MJ (n = 16). Overall, subjects with normal glycemia lost most on the low-fat/high-protein diet, subjects with high HOMA-IR lost most on the high-fat/high protein diet, and subjects with prediabetes and low FI had particular benefit from dietary fiber in the diet.