Prevention of weight re-gain is difficult for many individuals [12–14]. The main finding of this investigation was that subsequent to substantial weight loss on a VLED, a low carbohydrate diet and a low fat diet, combined with a clinical weight management program, were similar and effective to prevent weight re-gain over 6 months. For the low carbohydrate group, body weight remained approximately 19% below baseline body weight and the low fat group remained approximately 18% below baseline body weight.
Although both diets were similar to prevent weight re-gain, not all participants responded uniformly to either intervention. It was not surprising that some participants from both dietary groups regained weight after VLED as this has been reported elsewhere . Further, it is well-known that not all individuals that lose weight are successful for weight maintenance [1, 2, 16]. However, both dietary groups showed similar variation in weight change during the weight maintenance period. Fifty-five percent of participants in the low carbohydrate and 50% of participants in the low fat group continued to decrease their body weight during weight maintenance while the remainder re-gained a portion of their body weight (Figures 2 &3).
To attempt to explain the variability in weight change within each group, we examined energy and macronutrient intake differences between weight gainers and losers and found they were generally not significant; however, a couple of trends are interesting. For the low carbohydrate group, there was no difference in energy intake between the weight gainers and losers but the weight losers averaged 13 g of carbohydrate/day less than the weight gainers (P = 0.16). For the low fat group, weight losers consumed 178 kJ/day less (P = 0.09) and 7 g of fat/day less (P = 0.12) than the weight gainers. These trends imply that the level of carbohydrate or fat restriction for each dietary group may be important for subsequent weight change. Accordingly, the carbohydrate or fat level consumed by the weight gainers may not have been sufficient to maintain energy balance or produce an energy deficit. This is especially likely for the low carbohydrate gainers as the average consumption of carbohydrates was higher than reported in other studies [3, 4, 6, 7].
One statistically significant difference between the low carbohydrate weight gainers and losers is noteworthy. The low carbohydrate weight losers consumed an average of 15 grams of protein/day more than the low carbohydrate weight gainers (P = 0.02). This is consistent with human and rodent studies that report that increasing protein intake may be beneficial for weight loss and prevention of weight re-gain [17, 18]. For instance, Westerterp-Plantenga et al reported that additional protein intake (18% vs. 15%) resulted in less weight re-gain after 4 weeks of weight loss on a VLED. In addition, increasing the ratio of protein to carbohydrate, as reported by Layman et al, may also be important for continued weight loss and maintenance . Thus, we cannot rule out the possibility that the increased protein intake for the low carbohydrate group weight losers contributed to their continued decrease in body weight. Regardless of the reason for the variability in weight change within each dietary group, it is likely that both diets, if appropriately applied and adhered to, will yield a measure of success for weight maintenance in some individuals. Perhaps one of the most interesting questions arising from this study for future investigations is "how to determine or predict which individuals are most likely to succeed consuming a specific diet".
For the present study, attrition was similar for both groups. The primary reason for attrition during weight maintenance for both groups was lack of attendance at group meetings. Participants would not always provide reasons for their unwillingness to continue attendance or they intentionally discontinued correspondence and were removed from the study after dropping below the required attendance level (75%). As a result, we are left to conjecture as to why some participants discontinued attendance. We recognize that the high attrition in both groups is unfortunate and represents a weakness of the study. However, other similar studies have reported high rates of attrition up to 38% for low carbohydrate groups and 46% for low fat groups [4, 6, 7]. In a recent meta-analysis of 5 low carbohydrate vs. low fat trials reported by Nordmann et al, attrition rates were 30% and 43% for a low carbohydrate and low fat diet, respectively, after 6 months and 38% and 46%, respectively, after 12 months . As mentioned earlier, in attempt to limit attrition bias, we included an intent to treat analysis which did not change the statistical significance for any variable.
Physical activity is an important component of successful weight maintenance [21, 22]. Interventions that promote lifestyle changes along with PA have shown better weight maintenance than interventions that do not have these components . For the present study, PA was an important component and was similar for both groups. Both groups were prescribed an identical amount of PA and there was little variation. As a result, any difference in body weight change between groups during weight maintenance is not likely due to differences in PA.
Some participants were on prescription medications during the study. However, medication use was generally stable and statistical adjustment for medication use did not significantly influence body weight outcomes. As a result, we do not believe that medication use was a confounding factor. Further, the results are likely to generalize fairly well to an overweight or obese adult population, who are typically taking 1 or more medications such as blood pressure, depression, and lipids .
There were more total adverse events in the low carbohydrate group than in the low fat group. However, it should be noted that when adverse events were considered excluding the re-feeding period (month 4) the total number of adverse events reported were essentially the same for both groups. It is possible that the transition from a liquid VLED to solid food is more difficult when consuming a low carbohydrate diet or that our method of re-feeding can be improved to smooth this transition. Commonly reported adverse events for the low carbohydrate group were consistent with other studies, specifically, constipation, and diarrhea [7, 24]. One participant in the low carbohydrate group had an unexplainable increase in total cholesterol after 3 months on the low carbohydrate diet (total cholesterol increased from 136 to 306 mg/dL). This participant was advised to seek medical attention immediately.
We recognize that there are several limitations with this study. 1) The diet and PA data were self-reported. There are known biases and limitations with self-reported data, such as under-reporting energy intake . Nevertheless, diet records are commonly used and acceptable research instruments. 2) There were considerably more women than men and so results were not reported by gender. 3) We chose to use a quazi-experimental design with the site being assigned as either low carbohydrate of low fat rather than to randomize individual participants to a particular group. This was done because the popularity of the Atkins diet was at its height during our data collection. We felt that if participants assigned to different dietary protocols were in the same group or location there would be increased likelihood of data contamination by participants choosing to follow the dietary protocol of their choice rather than their assignment. Further, we did not use a cluster design and analyzed the data by individual participants. Had we analyzed the data by clinic assignment, the sample size would have been n = 2, insufficient for a cluster, and would likely have biased the results. Nevertheless, we recognize that the study design can be improved for future studies by randomization of participants or randomization by clinic using a cluster design of sufficient sample size. 5) The data collected for adverse events may be biased due to the assessment method. Adverse event data was collected from participants by administering a single sided page that listed specific adverse events seen in other low carbohydrate and low fat studies. We may have inadvertently prompted the participants to consider a specific adverse event they would not necessarily have reported had it not been listed.