Early Weight Loss and Other Factors Associated With Clinically Significant Weight Loss in Two Commercial Weight Loss Programmes

ABSTRACT

This secondary analysis assessed potential factors associated with achieving ≥ 5% or ≥ 10% weight loss among adults in a clinical trial evaluating two commercial weight loss programmes. Participants [n = 124, 45.7 (13.1) years, 20.2% male, BMI 34.4 (3.4) kg/m²] were randomised to the Medifast or OPTAVIA weight loss programme, both of which included meal replacements and behavioural support. Multivariable logistic regression analyses were used to assess early weight loss thresholds (percent weight loss at 2 and 4 weeks) and potential demographic, anthropometric, behavioural, and lifestyle factors associated with ≥ 5% or ≥ 10% weight loss at 16 weeks. Age, education, ≥ 1% weight loss by Week 2, ≥ 2% weight loss by Week 4, number of support contacts and programme adherence through Week 4 were significantly (p < 0.05) associated with achieving ≥ 5% weight loss. Participants losing ≥ 2% by Week 4 had 24.6 (95% CI: 2.49, 243) times greater odds of losing ≥ 5% by Week 16. For ≥ 10% weight loss at 16 weeks, positive factors associated included age, assignment to the OPTAVIA programme, ≥ 3% weight loss by Week 4 and programme adherence through Week 4 (p < 0.05). Participants losing ≥ 3% by Week 4 had 19.4 (95% CI: 2.02, 186) times greater odds of losing ≥ 10%. While several variables were significantly associated with achieving clinically significant weight loss, 4‐week weight loss was strongest. These results may be useful for early identification of individuals at risk for failing to achieve clinically significant weight loss who may benefit from intensification of counselling to assist in attaining weight loss goals.

Summary.

• What is already know about this subject?

  • ○
    Early weight loss during an intervention is an important factor associated with long‐term weight loss success.
  • ○
    Behavioural weight loss interventions using meal replacements and support can be more effective than self‐directed approaches.
• What this study adds

  • ○
    This analysis further expands the generalizability of the importance of early weight loss to clinical studies of commercial weight loss programmes.
  • ○
    This study emphasises the value of adherence and (programme‐specific) support in enhancing the effectiveness of structured weight loss interventions.

1. Introduction

More than two‐thirds of adults ages 20 years and older are living with overweight and obesity in the United States [1] and about 10% of US adults report being on some type of low‐energy or weight‐loss diet at a given time [2]. Both obesity rates and the reported use of weight‐loss diets continue to increase, so developing effective weight‐loss interventions is critical to meet consumer health needs. Intensive behaviour‐based weight‐loss interventions are common among commercial programmes and can be more effective long‐term (> 1 year) compared to minimally intensive interventions [3, 4, 5]. However, there is a high degree of variability in both the specific approaches utilised within these types of programmes and the weight‐loss responses among their users [4, 6]. Additionally, new generation obesity management medications (OMMs), particularly, glucagon‐like peptide‐1 receptor agonists (GLP‐1 RAs), demonstrate greater weight loss compared to lifestyle interventions but pose challenges for long‐term maintenance due to high discontinuation rates and subsequent weight regain [7, 8, 9, 10]. Given this, comprehensive lifestyle interventions remain essential for sustained weight management and overall health [8, 11].

Results from some studies suggest that early weight loss during an intervention is a critical factor in achieving longer‐term weight loss objectives [12, 13, 14, 15, 16]. The authors of a recent study found that percentage weight change during the initial 2, 4 and 8 weeks of a weight loss intervention programme, delivered within primary care, significantly predicted percent weight change at 6, 12 and 24 months, respectively [14]. Thus, investigating the relationship between early and longer‐term weight loss within different types of obesity treatment programmes (e.g., intensive, behaviour‐based commercial programmes) is warranted, as is assessing other factors associated with weight loss programme response.

A previously published 16‐week randomised, controlled trial (RCT) evaluated the efficacy of two commercial weight loss programmes (Medifast [MED] and OPTAVIA [OPT]), each compared to a self‐directed, reduced‐energy control diet. The study arms mimicked the commercial experience through the utilisation of the products (portion‐controlled meal replacements), meal plans, educational materials and behavioural support provided to actual customers of these programmes [17]. Results indicate that participants assigned to either the MED or the OPT weight loss programmes experienced significantly more weight loss (−5.0% and−5.7%, respectively, p < 0.0001), body fat reduction (p < 0.0001) and decreases in waist and hip measurements (p < 0.003) compared to the control group. Weight loss was directly correlated to the usage of meal replacements for both groups and adherence to coach support calls for the OPT group. While these findings support the use of the MED and OPT commercially available weight loss interventions for body weight reduction, individual responses varied and over half of the participants did not achieve a weight loss amount determined to be clinically significant (≥ 5%) [18]. Therefore, the purpose of this post hoc exploratory analysis was twofold: (1) to examine the predictive ability of the 2‐ and 4‐week weight loss thresholds (i.e., 1%, 2% and 3%) in identifying participants who did or did not achieve clinically significant weight loss (≥ 5% or ≥ 10%) at 16 weeks and (2) to evaluate potential factors associated with achieving at least 5% or 10% weight loss over a 16‐week period among the study participants who were randomised to either the MED or OPT weight loss programmes.

2. Materials and Methods

These exploratory analyses from the larger study [17] (#NCT02835092) focus on examining a comprehensive list of potential variables associated with clinically significant weight loss among 124 participants randomised to either the MED or OPT programmes. Of the 67 participants randomised to MED and 65 to OPT, 64 and 60 participants completed the study with no missing weight data, respectively, and were included in the final analysis. Healthy adults (18–65 years) with a BMI of 27.0–42.0 kg/m², stable weight (≤ 5% change in the past 6 months) and an interest in losing weight were considered for the study. Exclusion criteria included clinically significant abnormal lab results; recent use (past 6 months) of medications, supplements or programmes for weight loss; history of weight loss surgery; unstable use of medications affecting body weight; medical conditions requiring intensive medical management (e.g., Type 1 diabetes or medicated Type 2 diabetes); history of eating disorders or alcohol abuse; pregnancy or lactation or allergies/intolerances to study diet components. Briefly, participants in the MED group were instructed to follow the Medifast 4&2&1 Plan (1100–1300 kcal), which consisted of four Medifast meal replacements, two lean and green meals (5–7 oz. lean protein, three servings non‐starchy vegetables and up to two healthy fat servings) and one healthy snack daily (i.e., a serving of fruit, dairy or whole grains); each participant also received written programme materials (meal plan guide, food journals, recipes, self‐directed behavioural workbook), a 10–15 min introductory call with the Medifast nutrition support team and had telephone/email access to this team throughout the study period. Exercise, as recommended by the American College of Sports Medicine, was encouraged, but not required [19]. Participants randomised to OPT were instructed to follow the Optimal Weight 5&1 Plan (800–1000 kcal), which consisted of five meal replacements and one lean and green meal daily and 45 min of light to moderate intensity exercise daily was encouraged. Each participant also received written programme materials (meal plan guide, food journals, recipes, Dr. A's Habits of Health book [20] and workbook [21]), one‐on‐one telephonic coaching at various times throughout the study period and access to online support tools and phone/email access to the nutrition support team throughout the study. Additional details of the MED and OPT programmes, accompanying behavioural support and assessments can be found in the previous publication [17].

The percent change in body weight from baseline was calculated for the 2‐, 4‐ and 16‐week visits and summarised by categories of weight loss. Participants were stratified into the following categories of weight loss at 2 and 4 weeks: (1) < 1% and ≥ 1% weight loss; (2) < 2% and ≥ 2% weight loss; (3) < 3% and ≥ 3% weight loss. Additionally, the proportion of participants achieving clinically meaningful weight loss (≥ 5% and ≥ 10%) by the 16‐week visit within the above 2‐ and 4‐week weight loss categories was determined. Results from prior studies have suggested that a threshold of ~2% weight loss over 4 weeks has shown comparatively high specificity and low false positive frequency for identification of weight loss intervention participants who do not lose a clinically relevant amount of weight during lifestyle interventions [22, 23]. Additional thresholds (1% and 3%) and a 2‐week timepoint were also evaluated for comparison to the 2% threshold and 4‐week timepoint.

Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were calculated to assess the performance of early weight loss thresholds at 2 and 4 weeks (1%, 2% and 3%) in identifying participants who did or did not achieve clinically meaningful weight loss (≥ 5% or ≥ 10%) at 16 weeks. Four groups were created with the goal of predicting a positive response: (1) true positives: achieved the weight loss threshold at Week 2 or 4 and Week 16, (2) false negatives: failed to achieve the weight loss threshold at Week 2 or 4 but achieved the weight loss threshold at Week 16, (3) false positives: achieved the weight loss threshold at Week 2 or 4 but failed to achieve the weight loss threshold at Week 16 and (4) true negatives: failed to achieve the weight loss threshold at Weeks 2 or 4 and 16. Sensitivity and specificity were calculated in the following manner: sensitivity = [true positives/(true positives + false negatives)] and specificity = [true negatives/(true negatives + false positives)]. PPV and NPV were calculated in the following manner: PPV = [true positives/(true positives + false positives)] and NPV = [true negatives/(true negatives + false negatives)].

Statistical analyses were conducted using SAS for Windows (version 9.4, Cary, NC, USA). Descriptive statistics included means and standard deviations (SDs) and frequencies were expressed as number and percent of subjects. All tests of significance were assessed at alpha = 0.05, two‐sided. Univariate and multivariable logistic regression analyses were used to determine variables associated with ≥ 5% and ≥ 10% weight loss at the 16‐week visit. Variables evaluated were baseline measures of age, gender, race, ethnicity, education level, smoking status, alcohol use, heart rate, fasting plasma glucose, high‐sensitivity C‐reactive protein, systolic and diastolic blood pressures; baseline anthropometrics (height, body weight, body mass index [BMI], body fat mass index, percent body fat, arm circumference, chest circumference, waist circumference, hip measurement, thigh measurement, waist‐to‐hip ratio android fat mass, android lean mass, gynoid fat mass, gynoid lean mass, visceral adipose tissue (VAT) mass, VAT volume and total bone mineral content); weight loss programme (MED vs. OPT); programme adherence (self‐reported using a 100 mm visual analogue scale, with higher values indicating greater adherence); physical activity level (PAL; calculated based on degree of leisure time and work/school physical activity and ranged from very light: 1.4 to heavy: 2.3 [24]); number of nutrition support team contacts; quality of life and/or behavioural/emotional measures (Impact of Weight on Quality of Life [IWQOL]: physical function, self‐esteem, sexual life, public distress, work, total score and RAND36: physical functioning, role limitations due to physical health problems, role limitations due to emotional problems, energy/fatigue, emotional well‐being, social functioning, pain, general health); and early weight loss success (based on 2‐ and 4‐week criteria). Missing data for VAT mass and volume were imputed using multiple imputation, for the remaining factors, only observed data were included in the analysis. The multiple imputation procedure used the Markov chain Monte Carol method with a single chain to create 25 imputations. Pearson's or Spearman's correlation analysis was used to explore relationships between continuous variables.

Variables were included in the multivariable analysis if the p‐value for its overall effect was less than 0.10 in the univariate analysis. Body fat mass index (total fat mass (kg)/m²) was included in the multivariable analysis instead of absolute or relative body fat mass since body fat mass index adjusts for height. Systolic blood pressure was excluded from the multivariable analysis due to a large proportion of missing values (20%). Multivariable models were reduced using backward stepwise elimination with a retention p‐value of 0.10. Odds ratios and their corresponding 95% confidence intervals and p‐values are reported for the final multivariable logistic regression models. Logistic regression analyses were conducted for all participants and in participants who completed > 12 weeks of the programme separately as sensitivity analyses.

3. Results

Baseline characteristics of the participants included in the analysis (N = 124) are presented in Table 1. On average, participants were 45.7 (13.1) years of age and had a mean BMI of 34.4 (3.4) kg/m² which falls within the obesity Class I category.

TABLE 1.

Baseline demographics and characteristics of participants included in the analysis.

Characteristic	Meal plan
	All (N = 124)	OPT (N = 60)	MED (N = 64)
Age, years	45.7 (13.1)	45.3 (12.5)	46.1 (13.7)
Males, n (%)	25 (20.2%)	12 (20.0%)	13 (20.3%)
Education (high school, trade, some college)	54 (43.5%)	28 (46.7%)	26 (40.6%)
Education (college, graduate/professional)	70 (56.5%)	32 (53.3%)	38 (59.4%)
Weight (kg)	95.8 (13.3)	95.4 (14.2)	96.2 (12.6)
Body mass index (kg/m2)	34.4 (3.4)	34.4 (3.3)	34.4 (3.5)
Waist measurement (cm)	109.9 (11.0)	110.2 (11.4)	109.7 (10.8)
Hip measurement (cm)	120.0 (8.2)	119.9 (7.5)	120.1 (8.9)
Android fat mass (kg)	4.08 (1.10)	4.07 (1.08)	4.08 (1.12)
VAT mass (kg)	1.49 (0.81)	1.40 (0.74)	1.56 (0.87)
VAT volume (cm3)	1574 (857)	1487 (781)	1656 (921)
Body fat, %	45.5 (5.3)	45.8 (5.3)	45.3 (5.4)

Note: Values are presented as mean (SD) or n (%).

Abbreviations: MED, Medifast; OPT, OPTAVIA; VAT, visceral adipose tissue.

Tables 2 and 3 display the proportion of participants who achieved ≥ 5% or ≥ 10% weight loss by Week 16, stratified by their initial weight loss at 2 and 4 weeks, as well as sensitivity and specificity data. Notably, among the participants who lost ≥ 5% of weight by Week 16, 60.7% and 88.5% lost ≥ 3% of weight by Weeks 2 and 4, respectively, indicative of sensitivity values. Participants had 27.0 (95% CI: 10.1, 72.4; p < 0.001) times greater odds of losing at least 5% of body weight by Week 16 if they lost at least 3% of their body weight by Week 4. Overall, the 3% threshold resulted in the greatest specificity (87.3% at 2 weeks and 77.8% at 4 weeks), but lowest sensitivity, as compared to the 1% and 2% thresholds. No participants achieved at least 5% weight loss at 16 weeks if they did not achieve at least 1% weight loss by Week 4.

TABLE 2.

Classification performance of 2‐ and 4‐week WL thresholds to predict ≥ 5% WL at 16 weeks.

Threshold	n	≥ 5% WL	< 5% WL	OR (95% CI)	p	Sensitivity (%)	Specificity (%)	PPV (%)	NPV (%)
2‐week WL category
< 1%	30	2	28
≥ 1%	94	59	35	23.59 (5.30, 105.1)	< 0.001	96.7	44.4	62.8	93.3
< 2%	52	12	40
≥ 2%	72	49	23	7.10 (3.15, 16.0)	< 0.001	80.3	63.5	68.1	76.9
< 3%	79	24	55
≥ 3%	45	37	8	10.60 (4.30, 26.1)	< 0.001	60.7	87.3	82.2	69.6
4‐week WL category
< 1%	21	0	21
≥ 1%	103	61	42	61.00 (3.59, 1035.81)	< 0.001	100.0	33.3	59.2	100.0
< 2%	36	1	35
≥ 2%	88	60	28	74.99 (9.77, 575)	< 0.001	98.4	55.6	68.2	97.2
< 3%	56	7	49
≥ 3%	68	54	14	27.00 (10.1, 72.4)	< 0.001	88.5	77.8	79.4	87.5

Note: Sensitivity, specificity, PPV and NPV were calculated using early weight loss thresholds (1%, 2%, and 3% at Weeks 2 and 4) to classify participants who did or did not achieve ≥ 5% weight loss at 16 weeks. Sensitivity = percentage of clients achieving ≥ 5% weight loss at Week 16 who are correctly identified using the initial weight loss threshold [true positives/(true positives + false negatives)]. Specificity = the percentage of clients who did not lose ≥ 5% weight at Week 16 who are correctly identified using the initial weight loss threshold [true negatives/(true negatives + false positives)]. PPV: The percentage of participants who met the early weight loss threshold and went on to achieve ≥ 5% at 16 weeks. PPV = true positives/(true positives + false positives). NPV: The percentage of participants who did not meet the early weight loss threshold and did not achieve ≥ 5% at 16 weeks. NPV = true negatives/(true negatives + false negatives). In cases where one cell contained zero events (e.g., 0 participants achieving ≥ 5% weight loss in the < 1% 4‐week weight loss group), the odds ratio and confidence intervals were calculated by substituting 0.5 for the zero cell count. Fisher's exact test was used to assess the statistical significance of the observed association.

Abbreviations: CI, confidence interval; NPV, negative predictive value; OR, odds ratio; PPV, positive predictive value; WL, weight loss.

TABLE 3.

Classification performance of 2‐ and 4‐week WL thresholds to predict ≥ 10% WL at 16 weeks.

Threshold	n	≥ 10% WL	< 10% WL	OR (95% CI)	p	Sensitivity (%)	Specificity (%)	PPV (%)	NPV (%)
2‐week WL category
< 1%	30	1	29
≥ 1%	94	26	68	11.09 (1.44, 85.6)	0.021	96.3	29.9	27.7	96.7
< 2%	52	4	48
≥ 2%	72	23	49	5.63 (1.81, 17.5)	0.003	85.2	49.5	31.9	92.3
< 3%	79	6	73
≥ 3%	45	21	24	10.65 (3.85, 29.5)	< 0.001	77.8	75.3	46.7	92.4
4‐week WL category
< 1%	21	0	21
≥ 1%	103	27	76	14.92 (0.87, 255.09)	0.007	100.0	21.6	26.2	100.0
< 2%	36	0	36
≥ 2%	88	27	61	31.87 (1.89, 538.72)	< 0.001	100.0	37.1	30.7	100.0
< 3%	15	1	14
≥ 3%	68	26	42	34.04 (4.44, 261)	< 0.001	96.3	56.7	38.2	98.2

Note: Sensitivity, specificity, PPV and NPV were calculated using early weight loss thresholds (1%, 2%, and 3% at Weeks 2 and 4) to classify participants who did or did not achieve ≥ 10% weight loss at 16 weeks. Sensitivity = percentage of clients achieving ≥ 10% weight loss at Week 16 who are correctly identified using the initial weight loss threshold [true positives/(true positives + false negatives)]. Specificity = the percentage of clients who did not lose ≥ 10% weight at Week 16 who are correctly identified using the initial weight loss threshold [true negatives/(true negatives + false positives)]. PPV: The percentage of participants who met the early weight loss threshold and went on to achieve ≥ 10% at 16 weeks. PPV = true positives/(true positives + false positives). NPV: The percentage of participants who did not meet the early weight loss threshold and did not achieve ≥ 10% at 16 weeks. NPV = true negatives/(true negatives + false negatives). In cases where one cell contained zero events (e.g., 0 participants achieving ≥ 10% weight loss in the < 1% or < 2% 4‐week weight loss groups), the odds ratios and confidence intervals were calculated by substituting 0.5 for zero cell counts. Fisher's exact test was used to assess the statistical significance of the observed association.

Abbreviations: CI, confidence interval; NPV, negative predictive value; OR, odds ratio; PPV, positive predictive value; WL, weight loss.

Among participants who lost at least 10% of their body weight by Week 16, 77.8% had lost at least 3% weight by Week 2 and 96.3% had at least 3% weight loss by Week 4, indicating high sensitivity. Participants had 34.0 (95% CI: 4.44, 261; p ≤ 0.001) times greater odds of losing at least 10% of body weight by Week 16 if they lost at least 3% of their body weight by Week 4. Overall, the 3% threshold resulted in the greatest specificity (75.3% at 2 weeks and 56.7% at 4 weeks) as compared to the 1% and 2% thresholds. None of the participants achieved at least 10% weight loss by 16 weeks if they did not achieve at least 2% weight loss by Week 4. Sensitivity and NPV were generally higher for the 4‐week thresholds compared to the 2‐week thresholds, whereas specificity and PPV varied more substantially by cutoff.

Select physical and anthropometric univariate logistic regression of factors associated with at least 5% and 10% weight loss at Week 16 is provided in Tables S1 and S2. Age was positively associated with achieving both ≥ 5% and ≥ 10% weight loss. Mean (SD) age was 48.3 (12.5) years for those achieving at least 5% weight loss versus 43.2 (13.2) years for those losing less than 5% body weight (p = 0.0313). Among participants who lost at least 10% of their body weight by Week 16, mean (SD) age was 53.6 (9.9) years versus 43.6 (13.0) years for those who did not (p = 0.0010). While no other baseline physical or anthropometric parameters were significantly associated with ≥ 5% or ≥ 10% weight loss achievers at 16 weeks, hip measurement, android fat mass, VAT mass and VAT volume all showed trends (p < 0.1) towards being associated with at least 10% weight loss at Week 16 and thus were included in the multivariable logistic regression models (Table S2).

Tables S3 and S4 show select behavioural or lifestyle univariate logistic regression of factors associated with at least 5% and 10% weight loss at Week 16, respectively. Adherence to either weight loss programme (MED or OPT) through Week 4 was significantly associated with achieving at least 5% or 10% weight loss by 16 weeks (p < 0.001). However, when assessing each weight loss programme individually, only adherence to the OPT weight loss programme through Week 4 was significantly associated with achieving at least 5% or 10% weight loss by 16 weeks (p < 0.001). A greater number of nutrition support contacts through Week 4 was also significantly associated with at least 5% weight loss by Week 16 (p = 0.012) and a marginally significant factor associated with at least 10% body weight loss (p = 0.054). Each 0.1‐unit increase in physical activity level (ranging from 1.4 to 2.3) at baseline and at Week 4 resulted in 1.45 times greater odds of achieving ≥ 10% weight loss by Week 16 (95% CI: 1.04, 2.04 at baseline; 95% CI: 1.08, 1.95 at Week 4).

Table 4 presents the factors associated with ≥ 5% or ≥ 10% weight loss achievers at Week 16 after adjusting for weight loss programme, early weight loss categorizations, age, select anthropometrics, adherence, support contact, select behavioural/lifestyle characteristics and education level (≥ 5% weight loss only). After adjustments, the odds of losing ≥ 5% of weight at 16 weeks were 13.1 (95% CI: 1.25, 137; p = 0.032) times greater for participants who lost ≥ 1% compared to < 1% of weight at Weeks 2 and 24.6 (95% CI: 2.49, 243; p = 0.006) times greater for participants who lost ≥ 2% compared to < 2% of weight at Week 4. Furthermore, the number of nutrition support contacts through Week 4, programme adherence through Week 4 and age were positively associated with achieving at least 5% weight loss (p < 0.05). Lower educational attainment was negatively associated with achieving at least 5% weight loss by Week 16 (p = 0.047).

TABLE 4.

Multivariable logistic regression of factors associated with ≥ 5% and ≥ 10% weight loss at Week 16.

Factor	Estimate (SEM)	Odds ratio (95% CI)	p
≥ 5% weight loss at Week 16
Intercept	−11.019 (2.559)		< 0.001
Age (years)	0.046 (0.022)	1.05 (1.00, 1.09)	0.036
Education a	−1.170 (0.589)	0.31 (0.10, 0.99)	0.047
Either weight loss programme adherence through Week 4 (%) b	0.049 (0.018)	1.05 (1.01, 1.09)	0.007
Number of nutrition support team contacts through Week 4	0.918 (0.425)	2.50 (1.09, 5.76)	0.031
Week 2 weight loss category (≥ 1% vs. < 1%)	2.572 (1.198)	13.10 (1.25, 137)	0.032
Week 4 weight loss category (≥ 2% vs. < 2%)	3.203 (1.170)	24.60 (2.49, 243)	0.006
≥ 10% weight loss at week 16
Intercept	−18.427 (4.578)		< 0.001
Age (years)	0.079 (0.029)	1.08 (1.02, 1.15)	0.007
Weight loss programme (OPT vs. MED)	1.688 (0.690)	5.41 (1.40, 20.9)	0.015
Either weight loss programme adherence through Week 4 b	0.114 (0.042)	1.12 (1.03, 1.22)	0.006
Week 4 weight loss category (≥ 3% vs. < 3%)	2.965 (1.153)	19.39 (2.02, 186)	0.010

Note: Variables that were considered for the multivariable analysis included weight loss programme, 1% weight loss categories at Week 2, 2% or 3% weight loss categories at Week 4, age (years), hip measurement (cm), android fat mass (g), VAT mass (g), VAT volume (cm³), weight (kg), adherence to either weight loss programme (%), number of nutrition support team contacts, physical activity level at baseline and Week 4, Impact of Weight on Quality of Life (IWQOL): physical function, IWQOL: sexual life, IWQOL: total score, RAND36: role limitations due to emotional problems, RAND36: energy/fatigue, RAND36: pain. The referent for categorical factors is the second level listed within the parentheses in the factor column. Estimates, SEM, odds ratios, 95% CI and p values are derived from a logistic regression models predictive of 5% and 10% weight loss or more at Week 16. Stepwise method model significance level = 0.10.

Abbreviations: CI, confidence interval; MED, Medifast Weight Loss Programme; OPT, OPTAVIA Weight Loss Programme; SEM, standard error of the mean.

^a Education = High school; trade; Some College vs. college degree; grad/professional.

^b Programme adherence (%) was self‐reported using a 100 mm visual analogue scale, with higher values indicating greater adherence.

When evaluating ≥ 10% weight loss, the odds were 19.4 (95% CI: 2.02, 186; p = 0.0102) times greater for participants who lost ≥ 3% of weight compared to < 3% weight at Week 4 (after adjusting for other factors) (Table 4). In addition, the odds were 5.41 (95% CI: 1.40, 20.9; p = 0.015) times greater to achieve at least 10% weight loss for participants following the OPT versus MED weight loss programme. Age and adherence to either weight loss programme were both positively associated with (p < 0.05) achieving at least 10% body weight loss by Week 16.

4. Discussion

The present investigation used data collected from the two active treatment arms (MED and OPT) of a clinical trial for which results have been previously published [17] and explored various potential factors associated with achieving at least 5% or 10% weight loss over a 16‐week period among study participants. In multivariable analyses, variables associated with a greater likelihood of achieving at least 5% weight loss included losing ≥ 1% by Week 2 or ≥ 2% weight loss at Week 4, older age, higher education attainment, greater dietary adherence to the weight loss programme through Week 4 and a greater number of nutrition support contacts through Week 4. In addition, classification analysis showed that a ≥ 2% weight loss at Week 4 provided a strong overall balance of sensitivity (98%), specificity (56%), PPV (68%) and NPV (97%) for identifying participants who achieved ≥ 5% weight loss. Factors associated with the achievement of at least 10% weight loss included early weight loss (losing ≥ 3% body weight by Week 4), older age, greater adherence through Week 4 and randomisation to the OPT (vs. MED) weight loss programme. Across all 4‐week thresholds, NPV values were high (≥ 98%), reinforcing their ability to identify those unlikely to achieve ≥ 10% weight loss if early progress is not made. Together, these data suggest that initial weight loss, as well as older age and adherence to the weight loss programme, appear to be important determinants of whether clinically significant weight loss will be achieved.

These factors have been identified previously [25]; however, a key distinction between this study and the majority of prior publications is that the strength of the association observed between early weight loss and achievement of clinically significant weight loss in this study was much larger. For example, in this analysis, those who achieved 1% weight loss at 2 weeks or 2% at 4 weeks were 13 and 25 times more likely to achieve 5% weight loss at 16 weeks, respectively. Similarly, those who achieved 3% weight loss at 4 weeks were 19 times more likely to achieve 10% weight loss at 16 weeks. Results from past research indicate that those who reach early weight loss thresholds were between 3 and 11 times more likely to lose a clinically significant amount of weight compared to individuals with lower initial weight loss [13, 15, 22, 26, 27]. Although different from many prior reports, the higher likelihood of clinically significant weight loss if early thresholds are met aligns well with findings from a previous analysis of real‐world data of Medifast meal plans, where individuals who met early weight loss thresholds were 49 and 20 times more likely to achieve 5% and 10% weight loss, respectively [28]. The larger odds ratios observed in the previous analysis of real‐world data of Medifast meal plans could be related to the lower proportion of clients who were early non‐responders: only 5% did not meet the 2% threshold at 4 weeks and only 11% did not meet the 3% threshold. In the current analysis, 45% did not meet the 2% threshold at 4 weeks and 39% did not meet the 3% threshold. While this analysis shows a slightly higher proportion of early non‐responders than previous research (where approximately one‐quarter to one‐third of participants did not meet the early weight loss threshold), both the current and past analysis of Medifast meal plans show higher odds ratios of achieving clinically significant weight loss than other studies in the published literature [13, 22, 29]. Although there were differences in the proportions of early non‐responders between the two Medifast analyses, the reason(s) why the odds ratios were so much higher for both analyses of Medifast meal plans is unclear. It is possible the focus on meal replacements as a key dietary tool and the structured nature of the meal plans contributed to the higher odds. However, regardless of the likelihood of clinically significant weight loss being greater or smaller, early weight loss has been identified as the strongest determinant of weight loss success [25].

Increasing age has also been shown to be a factor associated with weight loss success [25, 30]. Age‐related factors that align with these structured interventions include a preference for consistency in behaviours [31] and higher intrinsic motivation for weight loss driven by health concerns [25, 32]. Consequently, it is perhaps not surprising that age has also been identified as a positive factor associated with higher adherence to weight loss interventions [32].

Adherence is a key factor associated with weight loss and has been previously shown to be closely linked with early weight loss outcomes [32]. The reciprocal relationship of these two factors aligns with the concept that adherence drives early results and early results reinforce adherence. When individuals adhere closely to a weight loss programme, they are more likely to see early success, and early success can serve as feedback that adherence pays off, creating a positive feedback loop [32]. Although this analysis and previous research have linked early adherence with weight loss success, this association has not been consistent [28]. This inconsistency has been attributed to minimal variability in early adherence among individuals and differences in adherence assessment methods [15, 16]. Thus, while early adherence may reinforce initial weight loss, long‐term adherence is likely more critical for overall weight loss success.

Other variables that increased the likelihood of achieving clinically significant weight loss included higher education and nutrition support team contacts (for 5% weight loss) and being on the OPT weight loss programme (for 10% weight loss). Higher education has been previously identified to have a positive association with weight loss success [25]. The nutrition support team was available as a resource for participants in both the MED and OPT groups. Contacting nutrition support may be a proxy for other factors related to weight loss (e.g., motivation) that were not measured during the study [25]. This result might also highlight the benefit of having a support resource that is available to consult with on intervention‐related questions, which helps foster confidence both in the intervention and in participants' own abilities. Finally, it was not surprising that being on the OPT programme increased the likelihood of achieving 10% weight loss relative to the MED programme given the lower energy level (800–1000 vs. 1100–1300 kcal) and additional coach support available to OPT participants [33, 34]. This may also explain why the OPT programme was associated with 10% but not 5% weight loss, as less energy combined with the additional support is likely more impactful in achieving greater weight loss rather than more modest reductions.

Even in this era of new and efficacious OMMs, particularly, GLP‐1 RAs, a recent survey of US adults further validated the relevance of this analysis [35]. A recent survey by the International Food Information Council found lifestyle changes remain the preferred weight loss method for US adults, with 54% of respondents having tried a specific eating style in the past year and, of those, 43% were doing so to lose weight [35]. Additionally, while nearly half of Americans believe prescription OMMs are effective, only three in 10 believe they are safe and only one in four would rather take a medication than change their lifestyle [35]. Another national survey of US adults found nearly three‐quarters of respondents preferred a diet change over injectable OMMs and seven in 10 are interested in trying a plant‐based diet if it could lead to significant weight loss [36]. Beyond preference, accessibility also plays a key role in highlighting the importance of lifestyle interventions, as shortages of new OMMs and barriers such as cost and limited prescribing by healthcare providers have resulted in only ~1.4% of eligible patients being prescribed these medications [37]. Furthermore, more than half of all patients discontinue GLP‐1 RAs before a year, reinforcing the ongoing importance of lifestyle interventions [7]. Despite this reported preference for lifestyle change over OMMs, for those opting to use OMMs, combining these medications with lifestyle modification also remains relevant and has the potential to enhance health and wellbeing, promote optimal nutrition as well as improve weight maintenance [8, 11].

There are several strengths of this study, including the use of two commercial weight loss programmes, a robust list of potential factors associated with clinically significant weight loss and hypothesis‐generating, clinically relevant research questions that could guide the design of future trials. Additionally, the high retention rate of the original study resulted in 94% of the originally randomised participants being included in these analyses. This minimises the potential for bias that can arise when analyses are limited to completers, particularly when those who drop out differ meaningfully from those who remain. The study also has limitations: it is a secondary analysis, and evaluable variables were limited to the data collected during the original trial. Other aspects of behavioural therapy, such as motivation, self‐efficacy and self‐regulation, could not be assessed. However, a previous systematic review has indicated that these factors do not consistently predict weight loss success [25]. The original study was a 16‐week weight loss intervention that did not include weight maintenance; however, based on previous research, early weight loss and age are both likely associated with long‐term weight maintenance [14, 30, 38]. Further, while the results of this analysis could be viewed as only being relevant to the MED and OPT programmes, a recent meta‐analysis highlights the broader relevance of these study findings [6]. Specifically, the meta‐analysis of 169 trials indicates there may be several key determinants of whether meaningful weight loss is achieved within behavioural weight management programmes [6]. These are (1) modification of diet; (2) offering partial or total meal replacement; (3) intervention provided by a psychologist, counsellor or dietitian and (4) delivery in a home setting. The MED and OPT programmes both provide each of these elements, with support being provided in a unique manner as outlined in the Methods.

5. Conclusions

Given the ongoing preference for non‐pharmacologic interventions by consumers, existing and previous shortages of new OMMs, and the low level of use of OMMs by prescribing health care providers, lifestyle behaviour change will continue to be a mainstay of obesity treatment. Identifying factors associated with success, as well as the optimal timing and criteria for early weight loss thresholds, remains crucial for effective nutrition and lifestyle interventions. Overall, these findings suggest initial weight loss, along with age and early adherence to the weight loss programme, appears to be important determinants of whether clinically significant weight loss will be achieved. Notably, participants who achieved early weight loss thresholds were decidedly more likely to experience clinically significant weight loss, highlighting the importance of monitoring early weight loss progress.

Classification analyses (sensitivity, specificity, PPV and NPV) further support the utility of early weight loss thresholds, especially at 4 weeks, for guiding targeted intervention strategies and identifying participants who may benefit from early support or programme modification. Participants assigned to the more intensive programme with a lower energy intake and greater support were more likely to achieve greater weight loss. Thus, these results highlight the benefits of more intensive lifestyle intervention and further support the view that low weight loss early in an intervention programme may help identify individuals who can be targeted for additional counselling and support, including programme alterations to improve weight loss outcomes. Interventions specifically targeting individuals with minimal early weight loss should be evaluated in RCTs to determine the optimal type and timing of such strategies for improving long‐term success.

Ethics Statement

This was a post hoc exploratory analysis of existing study data. The original study was approved by IntegReview Institutional Review Board, Austin, TX (BIO‐1607) and conducted according to the Declaration of Helsinki and Good Clinical Practice Guidelines (US 21CFR).

Consent

Written informed consent was obtained from all participants involved in the original study.

Conflicts of Interest

C.D.C., J.R.K. and S.S.J. are employees of Medifast Inc. J.L.U. is a member of the Medifast Scientific Advisory Board and receives consulting fees associated with that appointment. Within the prior 24 months, K.C.M. has received consulting fees from entities that provide products or services related to weight management, including Medifast, Eli Lilly, Pharmavite, General Mills and Novo Nordisk; he also has current research grant funding from National Dairy Council, Hass Avocado Board, Greenyn, Indiana University Foundation, Naturmega, Cargill, National Cattlemen's Beef Association and Ro. M.L.W. is an employee of and O.M.P. and M.B. are contractors for Midwest Biomedical Research which has received consulting fees from entities that provide products or services related to weight management, including Medifast, Eli Lilly, Pharmavite, General Mills and Novo Nordisk.

Supporting information

Table S1: Selected baseline physical and anthropometric univariate logistic regression factors associated with 5% weight loss or more at Week 16.

Table S2: Selected baseline physical and anthropometric univariate regression factors associated with 10% weight loss or more at Week 16.

Table S3: Selected behavioural or lifestyle univariate logistic regression factors associated with 5% weight loss or more at Week 16.

Table S4: Selected behavioural or lifestyle univariate logistic regression factors associated with 10% weight loss or more at Week 16.

Coleman C. D., Kiel J. R., Palacios O. M., et al., “Early Weight Loss and Other Factors Associated With Clinically Significant Weight Loss in Two Commercial Weight Loss Programmes,” Clinical Obesity 16, no. 1 (2026): e70046, 10.1111/cob.70046.

Data Availability Statement

Research data are not shared.