TOWARD, a metabolic health intervention, demonstrates robust 1-year weight loss and cost-savings through deprescription

Behavioral and lifestyle changes are considered a viable first-line treatment for diet-related noncommunicable diseases, including Type 2 Diabetes (T2D), prediabetes, hypertension, metabolic syndrome and obesity (1). The prevalence of these diseases is increasing, with 38% of adults in the United States having prediabetes or diabetes and 42% having obesity (2, 3). The economic burden of metabolic disease is also high, with $412 billion spent on T2D and $173 billion on obesity annually in the form of direct care cost (4).

Novel medications, including GLP-1 receptor agonists (GLP-1 RA), are becoming a cornerstone of obesity management of metabolic disease given demonstrated improvements in glycemia, adiposity, and cardiovascular outcomes (5). However, these medications are not without limitations. Trials have consistently found weight regain upon discontinuation (6), necessitating a lifetime commitment to the drug to retain weight loss and other health benefits. Some patients are nonresponders and do not meet the clinically significant weight loss goal of 5%. Fully 25% of subjects in the semaglutide 2.4 mg "STEP 5" trial experienced a < 5% reduction in body weight over 104 weeks (6). This trial also reported that cravings and hunger started to return toward baseline at 20 weeks to nearly or entirely equal to the placebo group at week 104 (6). An analysis of the SUSTAIN 8 clinical trial found that 30 to 40% of weight loss came from lean body mass (muscle) with the use of semaglutide over 52 weeks (7). Lean mass loss with GLP-1RA is of potential long-term concern because of the implications of sarcopenia on metabolism and its association with increased all-cause mortality (8).

Serious and widespread side effects of GLP-1 RAs are additional cause for concern. Some 82.2% of subjects in the STEP 5 trial were reported to experience gastrointestinal side effects on semaglutide 2.4 mg at 2 years (5). A small but notable increase in risk for all thyroid cancer and medullary thyroid cancer was also observed after 1–3 years treatment on GLP-1RAs in an analysis from the French national health care insurance system database (9).

For people with diabetes, outcomes may be less robust than reported. For instance, a follow-up study to the SCOPE trial (semaglutide 2.4 mg) reported significantly reduced HbA1c, yet this observation was derived from data on only 30 out of 343 subjects (10). Further, in a real-world retrospective study of 82,624 adults (mostly men), GLP-1RAs were found to be the least well tolerated medication, with a deprescription rate of more than 50%, among 5 medications prescribed for people with T2D. This deprescription rate was 10% higher than the next-most deprescribed drug (11). Another real-world study published in July 2023 investigated claims of 16 million commercially insured members and found that almost 70% of patients stopped GLP-1RA treatment within less than a year after starting (12). A prominent concern with these weight loss therapeutics is the price tag, with more than 142 million Americans qualifying for the medication and a price ranging from $800 to $1700/month, annual costs on the healthcare system could reach over a trillion dollars (13). The need for cost-effective alternatives as either a replacement or adjunct to GLP-1 RAs is therefore paramount.

Two studies have documented sustained weight loss maintenance with GLP-1RA discontinuations. A supervised exercise intervention was found to maintain body weight and body composition for 1 year following the discontinuation of GLP-1RA (14). Likewise, a retrospective analysis on 154 individuals demonstrated that a ketogenic diet prevented weight regain and protected against any increase in HbA1c at both six and 12 months following deprescription from GLP-1 RAs (15). The mean HbA1c of the deprescribed cohort rose 0.2% at 12 months yet remained in the non-diabetic range (15). These data suggest that programs that incorporate intensive health coaching, exercise prescriptions and TCR may represent important interventions to mitigate cost as well as limit weight regain and potential other harms from GLP-1 RA use.

Historically, few employee wellness programs have documented strong efficacy in managing diabetes or obesity. Our previously published 6-month data using the TOWARD approach demonstrated significant results on weight, Atherosclerotic Cardiovascular Disease (ASCVD) risk and cost-savings (16). In this paper, we present a focused review of the one-year weight loss results and cost savings from medication deprescription from 50 patients who self-selected to participate.

Data from 50 subjects were included in the intention-to-treat analysis. All subjects completed the first visit and nine subjects dropped out. Three of the subjects are no longer working for the company, one subject cited family reasons for dropping out, and the other dropouts did not give a reason for leaving the program (Figure 3). Of the 50 subjects, 23 (46%) were male and the median age at first visit was 52 (10). The mean starting weight was 123.8 kg (Table 1).

The mean weight loss was 19.5 ± 11.8 kg in the ITT analysis with a mean weight loss of 20.5 kg ± 11.8 in the PP analysis. This corresponded with an average weight loss of 15.5% (7.9%) of total body weight in the ITT analysis. 92% of individuals lost at least 5% of their body weight, 70% lost at least 10, 52% lost at least 15, and 30% lost >20%. All 50 out of 50 participants lost weight compared to their baseline. Of the 12 participants that did not continue to lose weight, mean weight loss was 12.3 kg. with a range of 1.6 to 23.2 kgs. Of the 50 participants, 38 continued to lose weight through 12 months and two stopped losing weight after month three and ten stopped losing weight after month six (Figure 4).

Total weight loss was comparable in patients with and without GLP-1RA (Figure 5). Of the 96 medications that were deprescribed, four were GLP-1RA, including two semaglutide 2 mg, semaglutide 1 mg, and dulaglutide 3 mg. Mean time to last follow up after beginning taper and complete deprescription was 487 days and 270 days, respectively. One subject on semaglutide 2.0 mg stopped their medication without a taper. After complete cessation, subjects experienced a 4%, −8%, 2%, and − 7% change in weight (Figure 6).

A total of ninety six medications were deprescribed and eight medications were prescribed. Eighteen of these medications were for T2D or prediabetes, nine were for Gastroesophageal Reflux Disease (GERD), and 26 were for hypertension (). The eight medications added included amlodipine-olmesartan, two prescriptions of tirzepatide, levothyroxine, pravastatin and three prescriptions of rosuvastatin. One prescription of tirzepatide was added to help an individual who continued to score highly on the Yale Food Addiction Scale (YFAS) screening despite maintaining TCR for several months and another prescription of tirzepatide was started for persistently elevated HbA1c and history of amputations. Overall, direct cost savings from deprescription over the course of 1 year were approximately $111,740.00 from deprescription, whereas net cost savings over the year from medication changes was approximately $83,285.00, which annualizes to a savings of about $1,700.00 per subject (). Supplementary material Supplementary material

Our 1-year data of 50 subjects who underwent a TOWARD comprehensive metabolic clinic intervention demonstrated continued weight loss of −15.5% in an ITT analysis, with −16.6% weight loss in the PP analysis, and in the setting of medication deprescription, including GLP-1 RA drugs, with resultant cost savings. Our average weight loss was comparable to Semaglutide 2.4 mg in the STEP 1 trial which had an average − 14.9% weight loss from baseline at 68 weeks (5). However, as our sensitivity analysis showed, it is unlikely that this similarity is due to some of our patients using GLP-1RAs as these were only a small proportion of our cohort (8 out of 50 patients) and weight loss was comparable in patients with and without GLP1-RA. A main difference between our intervention and semaglutide 2.4 mg is cost; the annualized costs of semaglutide is roughly $13,000 while the TOWARD approach provided annualized cost savings of $1,700 through medication deprescription. We also reported four subjects who discontinued their GLP-1 RA with ongoing weight loss or < 5% weight regain.

These data add to a growing body of evidence that a multi-modal approach to lifestyle that includes TCR is a cost-effective strategy for improving metabolic health. In the UK, interventions that employ TCR have been shown to reduce costs associated with diabetes while demonstrating impressive clinical results with T2D remission (27). In companies with a large market capitalization, a telemedicine approach to TCR has demonstrated durable T2D remission along with cost-savings at 2 years (21). Given that only 6.8% of the adults in the United States have optimal cardiometabolic health (28), there is an urgent need to find sustainable, cost-effective methods like TOWARD to reverse the metabolic health epidemics in more diverse settings.

Critics of lifestyle interventions for weight loss and diabetes remission have suggested the lack of long-term durability secondary to adherence as a major limiting factor for persistent efficacy and weight loss maintenance (29). In the UK, a study demonstrated long term diabetes remission in 91 patients (27). In our present data, 76% continued to lose weight after 12 months in the ITT analysis. This suggests that intensive lifestyle interventions are viable and durable long-term options for chronic disease (27).

Limitations of this pilot study included absence of a formal control group, although we can and did compare our data to GLP-1RA published results at a similar time point. Further, the patients self-selected to apply for the program, which introduces bias to the results. The TOWARD intervention is also multimodal; thus, it is not possible to identify the relative contribution of each TOWARD component, heterogeneity of response to individual components throughout the cohort, and whether certain components synergized to improve metabolic health outcomes or adherence. The TOWARD intervention, being a lifestyle intervention, is also necessarily limited by present, widespread socioeconomic obstacles, including access to food, education, community, social norms, and metabolically informed healthcare supports. Thus, these pilot data, while promising, need to be taken in context of broader socioeconomic issues limiting efficacy of lifestyle interventions in the United States and elsewhere. Still, our results have many strengths, including potential for scalability given the nature of the intervention, relatively high rates of adherence despite participants not being educated on metabolic health at baseline, and a large effect size in terms of weight loss that is comparable to current presumed best practices (GLP-1RA), while deprescribing 96 medications, which resulted in medication cost savings for the employer.

Additional limitations include heterogenous data collection (which is true for any real-world clinical intervention). Given the nature of home, remote-monitoring and patient preference, many data points were heterogeneously collected. For example, due to technical limitations of the remotely-monitored home scale, inconsistencies arose caused by participants' home-use, including issues such as wearing socks and shoes during measurements, making it impossible to ensure reliable and consistent body composition data across the cohort. As a result, we focused on body weight in this analysis. Lastly, although patients endorsed following a low-carbohydrate diet during clinic visits, the exact amount of carbohydrates eaten per day is unknown.

In conclusion, the TOWARD approach demonstrated robust improvements in weight and metabolic parameters with net deprescription of 88 medications across 50 patients. Additionally, there was significant cost savings through stopping medications that were no longer necessary due to lifestyle changes that were maintained for 1 year duration.