Lifestyle Interventions for Cardiovascular Disease Risk Reduction: A Systematic Review of the Effects of Diet Composition, Food Provision, and Treatment Modality on Weight Loss

Abstract

The purpose of this systematic review was to evaluate, synthesize, and interpret findings from recent randomized controlled trials (RCTs) of dietary and lifestyle weight loss interventions examining the effects of 1) diet composition, 2) use of food provision, and 3) modality of treatment delivery on weight loss. Trials comparing different dietary approaches indicated that reducing carbohydrate intake promoted greater initial weight loss than other approaches but did not appear to significantly improve long-term outcomes. Food provision appears to enhance adherence to reduction in energy intake and produce greater initial weight losses. The long-term benefits of food provision are less clear. Trials comparing alternative treatment modalities suggest that phone-based treatment produce short- and long-term weight reductions equivalent to face-to-face interventions. The use of Internet and mobile technologies are associated with smaller reductions in body weight than face-to-face interventions. Based on this review, clinical implications and future research directions are provided.

Introduction

Energy reduction, physical activity promotion, and behavioral strategies are fundamental components of evidence-based lifestyle intervention for weight management. However, there are a number of specific characteristics that can vary considerably across programs. Fo r instance, weight loss interventions can differ in the type of nutritional focus or dietary recommendations provided, the frequency and duration of contacts between participants and interventionists, the method of communication between participants and treatment staff, among other factors. It is important to understand how these variations in treatment content and delivery may influence participants’ adherence and their success in achieving weight loss so that the most effective and feasible strategies are retained and/or incorporated into evidence-based treatments.

In fact, a considerable amount of research has focused on these aspects of treatment, although there has often been variability in findings and subsequent conclusions. In addition, there have been substantial differences in research designs, populations, and overall methodological strengths and limitations of studies. Therefore, it is worthwhile to review the scientific literature to interpret and integrate the most recent findings on these issues relevant to weight loss treatment delivery. The purpose of this systematic review was to critically evaluate and synthesize recent research findings on treatment characteristics that may significantly affect participants’ treatment adherence and weight management. In this review, we summarized findings from recent randomized controlled trials (RCTs) of lifestyle interventions for weight loss. In particular, we focused on the effects of three aspects of treatment, including: 1) diet composition (e.g., low-carbohydrate, low-fat recommendations), 2) structure of meal plans (e.g., provision of foods), and 3) modalities of treatment delivery (e.g., in-person, telephone). Systematic reviews were conducted for each of these topics to summarize the effects of these program characteristics on participants’ initial and sustained weight loss.

Methods

A review of the literature was conducted using PubMed to identify RCTs of dietary and lifestyle interventions for weight loss. Trials that specifically examined the effects of one of the following aspects of treatment were included: 1) diet composition, 2) use of structured or prepackaged meal plans, and 3) modality of treatment. The following criteria were used to evaluate trials to determine their eligibility for inclusion in this systematic review:

• Inclusion criteria: Published within the last ten years (January 2003 to April 2014); RCTs including weight-related outcomes; conducted with free-living adults (≥18 years); ≥15 participants per trial arm; initial treatment lasting ≥2 months; published in English in a peer-reviewed journal.

• Exclusion criteria: Non-randomized trials; interventions without an adequate comparison group; attrition >40% at any time point; treatment not focusing mostly (if not exclusively) on overweight or obese subjects; not specifically focused on weight or cardiovascular risk.

Initial follow-up was typically defined as a weight-related assessment that occurred ≤6 months after randomization. When available, extended follow-up was generally defined as a weight-related assessment that occurred ≥12 months after randomization. For studies that included more than one assessment during the initial and/or extended follow-up periods, the longer of the available time points was generally included in the summary unless a particular time point signified a distinct transition in the type or intensity of care. Studies that included components of more than one topic area (e.g., trials comparing different modalities of delivery that also included food provision) were included in only one of the reviewed domains based on the primary focus of the original trial.

Results

Trials Comparing Different Diet Compositions

Overview of Included Trials

Twenty-two unique weight loss trials that compared different dietary approaches for reducing energy intake and/or modifying macronutrient compositions were included for review [1–22]. Dietary targets examined in one or more trials included: 1) “conventional” reduced-calorie, reduced-fat; 2) reduced-carbohydrate, 3) high-protein, 4) high or low glycemic-index foods, 5) Mediterranean diet, and 6) vegan diet, although the precise definition and recommendations within each of these general categories differed across trials. Eighteen of the 22 trials included a comparison of one diet to an alternative active dietary intervention [2–8, 11–21]. Of these, 14 included a conventional, reduced-fat intervention arm in comparisons [2, 4, 7, 11–21], while four included differing versions of alternative diets (e.g., high-protein vs. high-carbohydrate; [3, 5, 6, 8]). The remaining four trials compared an active intervention to an education-only condition or similar minimal treatment comparator [1, 9, 10, 22]. Of the 22 trials, eight reported initial outcomes only (≤6 months; [1–8]), three reported extended outcomes only (≥12 months; [9–11]), and 11 reported both initial and extended outcomes [12–22].

Summary of Findings

Half of the trials (11 of 22) found no weight loss differences between dietary conditions at any time point [3–8, 11–15], while the other 11 trials reported significant differences in weight loss at one or more assessment ([1, 2, 9, 10, 15–22]; see Table 1). Two of the eight trials that included only initial outcomes (≤6 months) found significant differences between conditions [1, 2]. One of these short-term studies reported that a Mediterranean diet achieved greater reductions in BMI than usual care [1], while the other study found that a low-carbohydrate diet achieved greater weight loss than a low-fat diet [2]. However, four short-term trials comparing various levels of dietary fat, carbohydrates, and protein composition yielded no differences in weight loss [3–6], and two trials comparing Mediterranean diets to either a low-fat diet [7] or an educational control [8] showed no differences. Two of the three trials that included only extended outcomes found that a Mediterranean-style diet achieved greater weight loss than education-only or limited-treatment conditions [9, 10]. The other long-term study, which compared a Mediterranean diet to a more intensive comparison treatment, failed to show differences in weight loss [11].

Table 1.

						Pre treat ment	Treatment		Follow-up
Reference	n	Attrition	Contact with Clinician	Characteristics of Intervention	Inclusion criteria/ characteristics	Original weight (kg)	Duration (mo.)	Wt. change (kg)	Duration (mo.)	Wt. change (kg)
Trials Comparing Different Diet Compositions
Dansinger et al. (2005) [12]  Carbohydrate restriction (Atkins)  Macronutrient balance (Zone)  Calorie restriction (Weight Watchers)  Fat restriction (Ornish)	160 40 40 40 40	21% (M2) 38% (M6) 42% (M12)	4 group sessions	Diet, physical activity, limited behavioral strategies	Men and women BMI=27–42 kg/m2 ≥1 cardiac risk factor Aged 22–72 years	100.0 99.0 97.0 103.0	2	−3.6 −3.8 −3.5 −3.6	12	−2.1 −3.2 −3.0 −3.3
Ebbeling et al. (2007)[13]  Low-glycemic load  Low-fat	73 36 37	10% (M6) 25% (M12) 30% (M18)	23 group workshops; 1 individual session; 5 telephone calls	Diet, behavioral strategies	Men and women BMI ≥30 kg/m2 Aged 18–35 years	103.5 103.3	6	NSa	18	NSa
Elhayany et al. (2010)[11]  Low-carb Mediterranean  Traditional Mediterranean  ADA diet	259 85 89 85	17% (M3) 22% (M6) 25% (M9) 31% (M12)	24 sessions	Diet, physical activity	Men and women with type 2 diabetes BMI 27–34 kg/m2 Aged 30–65 years	86.7 85.5 87.9	---	---	12	−8.9 −7.4 −7.6
Esposito et al. (2003)[9]  Low-energy Mediterranean (LEM)  Educational control	120 60 60	7% (M12)	18 group sessions (LEM only)	Diet, physical activity, behavioral strategies (LEM only)	Premenopausal women BMI≥30 kg/m2 Aged 20–46 years	95.0 94.0	---	---	24	14.0*< −3.0
Esposito et al. (2004)[10]  Mediterranean diet  Control diet (low-fat recommendations)	180 90 90	9% (M24)	18 group sessions (LEM); 12 group sessions (control)	Diet, physical activity, behavioral strategies	Men and women with metabolic syndrome	78.0 77.0	---	---	24	−4.0* −1.2
Esposito et al. (2009; 2014)b, [16, 49]  Low-carb Mediterranean  Low-fat	215 108 107	9% (M48)	30 group sessions	Diet, physical activity, limited behavioral strategies	Men and women with newly-diagnosed type 2 diabetes Aged 30–75 years BMI>25 kg/m2	86.0 85.7	12	−6.2* −4.2	48	−3.8 −3.2
Estruch et al. (2006)[8]  Mediterranean + olive oil  Mediterranean + nuts  Education control (low-fat recommendations)	772 257 258 257	<1% (M3)	1 dietary advice session (all diets); 1 educational group session (Med diets only)	Diet	Men (aged 55–80 years) and women (aged 60–80 years) with type 2 diabetes or 3+ other CHD risk factors	Not reported	3	−0.19 −0.26 −0.24	---	---
Foster et al. (2003)[17]  Low-carbohydrate (Atkins)  Low-fat	63 33 30	22% (M3) 33% (M6) 41% (M12)	4 brief sessions (focused on diet)	Diet	Obese men and women Mean age=44 years Mean BMI=34 kg/m2	98.7 98.3	3	−6.8* −2.7 (%change)	12	−4.4 −2.5 (%change)
Foster et al. (2010)[18]  Low-carbohydrate  Low-fat	307 153 154	7% (M3) 14% (M6) 26% (M12) 37% (M24)	38 group sessions	Diet, physical activity, behavioral strategies	Men and women BMI 30 – 40 kg/m2 Aged 18–65 years	103.3 103.5	3	−9.5* −8.4	24	−6.3 −7.4
Gardner et al. (2007)[20]  Low-carbohydrate (Atkins)  Macronutrient balance (Zone)  Low-fat (LEARN)  Very low-fat (Ornish)	311 77 79 79 76	20% (M12)	8 group sessions	Diet, physical activity, behavioral strategies (varied based on content of diets)	Premenopausal women BMI 27–40 kg/m2 Aged 25–50 years	86.0 84.0 85.0 86.0	2	Atkins > weight loss than other 3 *c	12	−4.7* −1.6 −2.2 −2.6
Iqbal et al. (2010)[14]  Low-carbohydrate  Low-fat	144 70 74	30% / 12%‡ (M6) 47% / 19%‡ (M12) 53% / 13%‡ (M24)	27 group sessions	Diet, physical activity	Women and men with type 2 diabetes BMI≥30 kg/m2 Aged ≥18 years	118.3 115.5	6	−2.8 −2.0	24	−1.5 −0.2
Luscombe-Marsh et al. (2005)[3]  Low-fat, high-protein  High-fat, standard-protein	73 36 37	22% (M4)	8 dietary counseling visits coinciding with visits for food provision	Diet (including provision of key foods to achieve macronutrient composition)	Men and women BMI 27–40 kg/m2 Aged 20–65 years	M/W 100.3/ 90.5 111.6/ 90.0	4	−9.7 −>10.2	---	---
McLaughlin et al. (2006)[4]  Low-fat, high-carbohydrate  High-fat, moderate-carbohydrate	65 34 31	12% (M4)	1 initial nutrition education session + 16 brief sessions with dietician (to review food records)	Diet (including provision of meal plans), limited behavioral strategies (self-monitoring)	Men and women with insulin resistance BMI 29–36 kg/m2	94.3 95.0	4	−5.7 −6.9	---	---
McMillan-Price et al. (2006)[5]  High-carb, high-GI  High-carb, low-GI  High-protein, high-GI  High-protein, low-GI	129 32 32 32 33	10% (M3)	12 visits with dietician coinciding with visits for food provision	Diet (including provision of some key foods and provision of meal plans)	Men and women BMI ≥25 kg/m2 Aged 18–40 years	86.0 87.1 87.7 88.4	3	−3.7 −4.8 −5.3 −4.4	---	---
Noakes et al. (2005)[6]  High-protein, low-fat  High-carbohydrate, low-fat	119 58 61	16% (M3)	3 consultation visits with dietician + 3 food preparation sessions	Diet (including provision of some key foods), physical activity, limited behavioral strategies (self-monitoring)	Women BMI 27–40 kg/m2 Aged 20–65 years	87.0 86.0	3	−7.6 −6.9	---	---
Sacks et al. (2009)[15]  Low-fat, average-protein  Low-fat, high-protein  High-fat, average-protein  High-fat, high-protein	811 204 202 204 201	20% (M24)	54 group sessions + 12 individual sessions	Diet, physical activity, behavioral strategies	Men and women BMI 25–40 kg/m2 Aged 30–70 years	94.0 92.0 92.0 94.0	6	NS d	24	NS d
Shai et al. (2008)[21]  Low-fat  Mediterranean  Low-carbohydrate	322 104 109 109	2% (M6) 5% (M12) 13% (M18) 16% (M24)	18 group sessions; 6 motivational telephone calls (as needed)	Diet, physical activity, behavioral strategies	Predominantly men (86%) BMI ≥27 kg/m2 Aged 40–65 years, Or, presence of type 2 diabetes or CHD (regardless of BMI or age)	91.3 91.1 91.8	6	Low-carb > low-fat, Med * e	24	-2.9* -4.4 -4.7
Stern et al. (2004)[19]  Low-carbohydrate  Low-fat	132 64 68	40% / 11%‡ (M6) 34% / 5%‡ (M12)	15 group sessions	Not specified	Women and men BMI ≥35 kg/m2 Aged ≥18 years	130.0 132.0	6	Low-carb > low-fat*f	12	-5.1 -3.1
Toobert et al. (2003)[1]  Mediterranean  Usual care	279	12% (M6)	Initial 3-day retreat + 24 sessions (Med only)	Diet, physical activity, behavioral strategies (Med only)	Postmenopausal women with type 2 diabetes Aged <75 years	BMI 35.3 34.9	6	BMI -0.37 * +0.20	---	---
Turner-McGrievy et al. (2007)[22]  Vegan diet  National Cholesterol Education Program (NCEP) guidelines	62 31 31	15% (M12) 23% (M24)	14 group sessions (all participants)+ 26 additional support groups (34 participants)	Diet, physical activity	Postmenopausal women BMI 26–44 kg/m2 Aged 44–73 years	87.4 86.4	12	−4.9* −1.8	24	−3.1* −0.8
Vincent-Baudry et al. (2005)[7]  Mediterranean  Low-fat	212 102 110	20% (M3)	Not specified	Diet (nutritional recommendations and booklet)	Men and women with moderate CVD risk Aged 18–70 years	BMI 28.7 28.7	3	BMI −1.5 −1.2	---	---
Yancy et al. (2004)[2]  Low-carb  Low-fat	119 59 60	34% (M6)	9 group sessions	Diet, physical activity	Women and men with hyperlipidemia BMI 30–60 kg/m2 Aged 18–65 years	97.8 96.8	6	−12.0 −6.5*	---	---
Trials Comparing Methods of Meal Provision
Hannum et al. (2004)[27]  Self-selected diet  Portion-controlled entrees	53 27 26	g12% (M2)	8 dietary education sessions	Diet	Women BMI 26–42 kg/m2 Aged 24–60 years	85.3 86.7	2	-3.6 -5.6*	---	---
Hannum et al. (2006)[28]  Self-selected diet  Portion-controlled entrees	51 26 25	15% (M2)	8 dietary education sessions	Diet	Men BMI 26–42 kg/m2 Aged 24–60 years	98.0 96.8	2	−5.1 −7.4*	---	---
Li et al. (2005)[23]  Slimfast meal replacement plan  Prescribed 500 calorie/day deficit	82 46 36	21% (M6) 26% (M12)	9 dietary education sessions	Diet	Men and women with type 2 diabetes BMI 27–40 kg/m2 Aged ≥30 years	Not reported	3	−5.6* −2.9	12	−4.4* −2.4
Rock et al. (2010)[24]  In-person treatment + prepackaged meals  Phone-based treatment + prepackaged meals  Usual care	442 167 164 111	8% (M24)	104 individual sessions	Diet, physical activity, behavioral strategies (treatment groups only)	Women BMI 25–40 kg/m2 Aged ≥18 years ≥15 kg over ideal weight	92.2 92.9 91.0	6	−9.2i −8.3 −2.9*	24	−7.4i −6.2 −2.0*
Tsai et al. (2012)[25]  2 prepackaged meals/day  1 prepackaged meal/day	50 25 25	26% (M4)	8 individual sessions	Diet, physical activity, behavioral strategies	Men and women BMI 30–49.9 kg/m2 Aged ≥18 years	108.9 106.5	4	−5.9 −5.3	---	---
Webber et al. (2013)[26]  Internet program  Internet program + portion-controlled diet	50 25 25	6% (M3)	1 in-person session at baseline	Diet, physical activity, behavioral strategies	Men and women BMI 30 – 45 kg/m2 Aged 25–65 years	BMI 35.2 35.0	3	Weight −4.1 −5.7	---	---
Trials Comparing Different Treatment Modalities
Appel et al. (2011)[31]  Remote  Remote + in-person  Self-directed	415 139 138 138	12% (M6) 4% (M24)	33 telephone contacts (remote); 39 in-person or phone contacts (remote+in-person)	Diet, physical activity, behavioral strategies	Men and women One or more cardiovascular risk factor Aged ≥21 years	102.1 105.0 104.4	6	−6.1h −5.8 −1.4	24	−4.6 −5.1 −0.8
Befort et al. (2010) [39]  Group phone  Individual phone	34 16 18	15% (M4) 21% (M6)	20 sessions; either group or individual	Diet, physical activity, behavioral strategies	Women living in rural areas BMI 25 – 44.9 kg/m2 Aged 22–65 years	95.7 90.5	4	−13.5 * −9.2	6	−14.9* −9.5
Carter et al. (2013)[43]  Self-monitoring via app  Self-monitoring via website  Self-monitoring via paper diary	128 43 43 41	38% (M6)	None, except for weight assessment visits	Self-monitoring intervention only	Men and women BMI ≥27 kg/m2 Aged 18–65 years	96.84 97.9 96.4	6	−4.6i −2.9 −1.3	---	---
Chambliss et al. (2011)[44]  Computerized self-monitoring  Computerized self-monitoring + enhanced behavioral treatment  Control	120 45 45 30	21% (M3)	Enhanced program only: 1 introductory seminar, 3 telephone consultations	Diet, physical activity, behavioral strategies (limited behavioral strategies only in self-monitoring and control groups)	Women and men BMI 25–35 kg/m2 Mean age = 45.0 years	97.1 98.2 100.1	3	−2.7j −2.5 0.3	---	---
Collins et al. (2013)[42]  Basic online  Basic online + personalized feedback and reminders	301 143 158	21% (M6)	No clinician contact – personalized feedback was automatically generated based on participant-reported data	Diet, physical activity, behavioral strategies	Men and women BMI 25–40 kg/m2 Aged 18–60 years	94.4 93.4	3	−2.7 −3.3	6	−3.6 −4.3
Donnelly et al. (2007)[32]  Phone  Face-to-face  Control	74 25 27 22	24% (M6.5)	26 group sessions; either by phone or in person	Diet (including meal provision), physical activity, behavioral strategies	Men and women Mean BMI = 33.2 Aged 25–68 years	102.5 95.6 88.2	4	−10.6k −12.7 −0.25	6.5	−12.8j −12.5 n/a
Donnelly et al. (2013)[33]  Phone  Face-to-face	395 201 194	15% (M6) 27% (M18)	42 group sessions	Diet (including meal provision), physical activity, behavioral strategies	Men and women BMI 25–44.9 kg/m2 Aged 18–65 years	100.0 101.4	6	−12.6 −13.5	18	−6.2 −7.1
Gold et al. (2007)[38]  Therapist-led online  Basic online	88 40 48	29% (M12)	24 weekly meetings (therapist-led program only)	Diet, physical activity, behavioral strategies	Men and women BMI 25–39.9 kg/m2 Aged ≥18 years	92.0 90.2	6	−6.8* −3.3	12	−5.1* −2.6
Harvey-Berino et al. (2010)[34]  Internet  In-person  Hybrid	481 161 158 162	4% (M6)	24 group sessions	Diet, physical activity, behavioral strategies	Men and women BMI 25–50 kg/m2 Mean age = 46.6 years	97.2 97.2 96.5	6	−5.5l −8.0 −6.0	---	---
Hunter et al. (2008)[29]  Internet  Usual care	446 224 222	17% (M6)	Internet group: 2 phone sessions, weekly written feedback	Diet, physical activity, behavioral strategies only in internet group	Military personnel Men and women BMI ≥25 for women BMI ≥27.5 for men Aged 18–65 years	87.4 86.6	6	−1.3* 0.6	---	---
Micco et al. (2007)[37]  Internet-only  Internet + in-person support (IPS)	123 62 61	37% (M12)	52 weekly online meetings; IPS group included 12 monthly in-person sessions	Diet, physical activity, behavioral strategies	Men and women BMI 25–39.9 kg/m2 Aged ≥18 years	92.0 86.1	6	−6.8 −5.1	12	−5.1 −3.5
Patrick et al. (2009)[30]  Standard weight advice  Tailored text messages	65 32 33	66% (M4)m	3–5 text messages per day in treatment group, plus 4 monthly phone calls	Diet, physical activity, limited behavioral strategies	Men and women BMI 25–39.9 kg/m2 Aged 25–55 years	88.0 88.8	4	−0.17 −4.6*	---	---
Perri et al. (2008)[35]  Face-to-face  Phone  Educational control	234 83 72 79	6% (M18)	26 biweekly sessions (or mailings for control group)	Diet, physical activity, behavioral strategies	Women BMI >30 kg/m2 Aged 50–75 years	97.8 96.4 95.0	6	−10.0n	18	−8.9 −8.2 −6.8*
Sherwood et al. (2010)[40]  Self-directed  10 phone sessions  20 phone sessions	63 21 21 21	19% (M6)	10 or 20 phone sessions over 6 months in intervention groups	Diet, physical activity, behavioral strategies	Men and women BMI 30–39 kg/m2 Aged ≥18 years	95.5 93.6 96.7	6	−2.3 −3.2 −4.9	---	---
Tate et al. (2003)[36]  Basic internet  Internet behavioral counseling	92 46 46	16% (M12)	Enhanced: weekly emails from counselor (5/week during month 1, weekly during months 2–12)	Diet, physical activity, behavioral strategies	Men and women BMI 27–40 kg/m2 Mean age = 48.5 years ≥1 risk factor for type 2 diabetes	89.4 86.2	3	−2.7 −4.1*	12	−2.0 −4.5*
Turner-McGrievy (2009)[45]  Weight loss podcast  Theory-based weight loss podcast	78 37 41	17% (M3)	24 podcasts, no clinician contact	Diet, physical activity, behavioral strategies	Men and women BMI 25–40 kg/m2 Mean age = 38.6 years	89.0 91.9	3	−0.3 −2.9*	---	---
Turner-McGrievy et al. (2011)[46]  Podcast only  Podcast + internet support	96 49 47	10% (M6)	24 podcasts + 32 mini-podcasts, no clinician support	Diet, physical activity, behavioral strategies	Men and women BMI 25–45 kg/m2 Aged 18–60 years	BMIo 32.2 32.9	6	−2.7 −2.7	---	---
Webber et al. (2008)[41]  Internet  Enhanced internet	66 33 33	0% (M4)	No clinician meetings in internet group; Enhanced group received 16 chat group sessions	Diet, physical activity, behavioral strategies	Women BMI 25–50 kg/m2 Aged 22–65 years	82.5 82.1	4	−5.2 −3.7	---	---

^*indicates significant between-group difference(s); NS = not significant.

^‡includes weights retrieved from medical record.

^aEbbeling et al. (2007) did not report specific weight change values for dietary conditions. Weight changes were not significantly different at months 6 or 18.

^bEsposito et al. (2014) extended observational follow-up to six years post-randomization from the original trial (Esposito et al., 2009). At Year 6, the mean between-group weight difference was 0.4 kg and no longer significant (95% CI −0.1 to 0.7 kg). Because the original trial reported 4-year outcomes, which includes a much longer follow-up than most other trials, we have focused on these extended outcomes in the table.

^cGardner et al. (2007) did not report specific weight change values for dietary conditions at month 2. At month 2, the Atkins diet was significantly different than the other three diets. At month 12, the Atkins diet was significantly different from the Zone diet.

^dSacks et al. (2009) did not report specific weight change values for each dietary condition. Weight changes differed by <0.5 kg between all conditions and was not significant at any time points.

^eShai et al. (2008) did not report specific values for weight change at month 6. However, at month 6, weight loss was significantly greater in the low-carb condition vs. the low-fat or Mediterranean diet conditions, which did not differ from each other. At month 24, weight loss was significantly greater in the low-carb and Mediterranean diet conditions vs. the low-fat condition. The low-carb and Mediterranean diet conditions did not differ significantly from each other at month 24.

^fStern et al. (2004) did not report specific values for weight change at month 6. At month 6, weight loss was significantly greater in the low-carb condition vs. the low-fat condition.

^gRock et al. (2010) reported that at months 6 and 12, both the in-person treatment + prepackaged meals and the phone-based treatment + prepackaged meals were significantly different than usual care. The two active interventions did not differ from each other at either time point.

^hAppel et al. (2011) reported no significant differences between the remote and remote+ in-person intervention groups at any time point during the study, although both groups differed from the self-directed group at each time point.

ⁱCarter et al. (2013) reported that this pilot study was not adequately powered to detect change in anthropomorphic measures, but noted that all three groups did significantly differ in weight change.

^jChambliss et al. (2011) reported that both groups who self-monitored lost significantly more weight than the control group

^kDonnelly et al. (2007) reported median weight changes. At 4 months, all three groups differed significantly from each other. At follow-up, weights were only reported for the two treatment conditions and were not significantly different between phone and face-to-face groups.

^lHarvey-Berino et al. (2010) reported that all conditions differed significantly from each other at 6 months.

^mPatrick et al. (2009) calculated adherence as the number of text messages responded to at the stated time point.

ⁿPerri et al. (2008) randomized groups to treatment conditions following initial treatment. During the first 6 months of treatment, all groups met face to face on a weekly basis and weight change was not significantly different between Randomization to different treatment arms started following the first 6 months of the study.

^oTurner-McGrievy et al. (2011) reported only BMI, not weight, at baseline. At months 3 and 6, the authors reported kilograms lost (and not BMI change).

Of the 11 trials that included both initial and long-term outcomes, four observed no differences between diets at any time point [12–15]. These trials included 2–4 dietary arms typically comparing diets focused on fat and carbohydrate restriction and/or the quality of fat and carbohydrate intake. Four of the 11 trials found significant initial effects that were not maintained over time [16–19]. In all four trials, a low-carbohydrate approach produced significantly greater initial weight loss than low-fat diets, although these benefits attenuated over time and were no longer present ≥12 months after treatment initiation.

Finally, three of the 11 studies found significant effects at both initial and extended follow-ups [20–22]. Gardner et al. compared four dietary interventions and found that a low-carbohydrate diet achieved significantly greater initial weight loss than the other three programs [20]. Also, the low-carbohydrate diet demonstrated continued superiority to one of the other conditions (Zone diet) at month 12 [20]. Shai et al. found that a low-carbohydrate diet produced greater weight loss than low-fat or Mediterranean diets initially, and the low-fat diet continued to achieve the least amount of weight loss and differed significantly from the low-carbohydrate and Mediterranean diets at extended follow-up [21]. The final trial indicated that a vegan diet achieved greater weight loss at 12- and 24-month follow-ups compared to an educational condition [22].

Conclusions and Limitations

While a significant amount of research has focused on the effects of different dietary approaches to weight loss, findings have been inconsistent. Across studies reviewed, there is some indication that a low-carbohydrate approach may achieve relatively greater weight loss as compared to a conventional, low-fat diet, but this is not a uniform finding. Also, these benefits may be short-lived, as most studies fail to find differences at extended follow-ups. There are also some suggestions for the benefit of a Mediterranean-style diet for weight loss, although many of these significant effects have been observed with comparisons to less-intensive, educational controls. Across trials, the magnitude of weight loss observed in intervention and control conditions varied widely, which may be attributable to the intensity and duration of treatment as well as the inconsistent inclusion of behavioral strategies in treatment.

These trials evaluated whether the interventions achieved differential weight loss assuming participants were adherent to the prescribed diets. The lack of differences between diets observed in many of these trials could be partly due to dietary non-adherence and suboptimal differentiation between conditions in terms of what participants actually consumed. Despite this, these trials address an important question, as an efficacious treatment will have little impact if individuals are unable or unwilling to adhere to it.

There are several factors that make it challenging to generalize conclusions across studies. First, interventions included varied dietary recommendations in how and to what extent macronutrients were restricted. Some treatments encouraged energy reduction in conjunction with macronutrient rebalancing, while others did not. Second, there were varying levels of structure imposed across interventions, with some protocols providing some foods, detailed meal plans, and/or shopping/food lists, while others afforded more participant autonomy in food selections and meal preparation. Third, as mentioned previously, interventions varied tremendously in intensity (i.e., amount of contact with interventionists) and in the inclusion of behavioral strategies (e.g., self-monitoring, problem-solving) to promote adherence to dietary recommendations. In general, trials that incorporated more intensive behavioral strategies achieved greater weight loss regardless of nutritional focus.

Trials Comparing Methods of Meal Provision

Overview of Included Trials

Six trials examining the use of meal provision for weight loss were reviewed [23–28]. Meal provision most often included the delivery of full meal replacements (i.e., individual prepackaged meals), although commercial meal replacement shakes (i.e., Slimfast) were used in one trial [23]. Meal provision typically included at least two meal replacements per day, with some groups receiving up to two additional prepackaged snacks per day. Three of the studies compared the use of prepackaged meals to an alternative active intervention [24–26]. Of these three, one examined the use of prepackaged meals in both phone and in-person treatment conditions (as well as a third usual care group; [24]), one examined a full meal subsidy (two meals per day) versus a partial subsidy (one meal per day, with encouragement to buy an additional meal at the participant’s expense; [25]), and one compared a meal provision plan with treatment materials delivered over the internet to an internet treatment-only group [26]. The remaining three studies compared the active intervention to dietary education sessions [23, 27, 28]. Four of the trials reported initial outcomes only [25–28], while two studies reported initial and extended outcomes [23, 24].

Summary of Findings

Four of the six trials found significant differences between conditions at one or more follow-up, while two trials found no significant weight changes between groups (see Table 1). Two of the four studies reporting only initial outcomes found significant differences in weight loss immediately following treatment, with observed benefits for the provision of portion-controlled entrees along with a variety of ‘a la carte’ food items [27, 28]. Two other studies described no initial differences in weight change between groups. One of these studies compared a full meal subsidy (2 meals per day) versus a partial subsidy (1 meal per day; [25]), while the other study examined the use of meal provision and internet-based treatment versus internet-based treatment alone [26].

Of the two studies that included initial and extended outcomes, both reported significant between-group differences in weight at both follow-ups [23, 24]. One trial examined the use of meal replacement shakes and found that participants in the meal provision group lost more weight than participants receiving basic dietary instructions on caloric reduction [23]. The second study compared two treatment groups to usual care [24]. Both treatment groups received commercially-available prepackaged meals (i.e., Jenny Craig) during the initial intervention with the addition of weekly sessions conducted either in-person or by phone. Usual care participants lost significantly less weight than both of the meal provision groups, which did not differ from each other.

Conclusions and Limitations

The evidence for meal replacement as a method for weight loss is the most limited of the three areas reviewed. While there is a robust literature on the effects of meal replacements as part of a very-low calorie diet, there is less recent research examining food provision in the context of reduced-calorie diets. Overall, the studies included in this review are fairly consistent in support for the use of meal provision for promoting weight loss in the short-term and through extended care, and the magnitude of weight loss approaches the amounts seen in conventional behavioral treatment programs without food provision. However, with a small number of studies (only two of which followed participants beyond initial treatment) it is difficult to draw firm conclusions about the overall effectiveness of meal provision for weight loss. In addition, half of the studies reviewed did not include behavioral strategies or physical activity in treatment. For the studies that included a more comprehensive behavioral approach in conjunction with meal provision, the potential benefits of meal provision was less pronounced, and there was little support for adding meal provision to existing behavioral treatment [26]. Therefore, it is unclear what relative effectiveness meal provision provides beyond more comprehensive lifestyle interventions without meal provision. Other limitations include the relatively modest sample sizes of most of the reviewed trials.

Trials Comparing Different Treatment Modalities

Overview of Included Trials

Eighteen weight loss trials comparing alternative methods of treatment delivery were included [29–46]. Alternative modalities (defined as any approach other than face-to-face sessions) examined in these trials included treatment delivered by: 1) phone, 2) online, 3) mobile technology, and/or 4) podcasts. Most of the studies (16 of 18) compared one type of treatment to at least one other active intervention, whereas two studies [29, 30] compared treatment to usual care or advice-only. Of the 16 involving at least two active intervention groups, only five compared an alternative treatment modality to a face-to-face approach [31–35]. The other 11 studies compared different treatment intensities within the same modality (e.g., “basic” versus “enhanced” internet-based treatment; [36–46]). Enhanced conditions typically included increased clinician feedback, social support, and/or a greater number of sessions. Twelve of the studies provided only initial outcomes (≤6 months; [29, 30, 32, 34, 39–46]), while six reported long-term results (≥12 months; [31, 33, 35–38]).

Summary of Findings

Six of the trials found no weight loss differences at any follow-up [33, 37, 40–42, 46], while 12 trials found significant differences at one or more assessment ([29–32, 34–36, 38, 39, 43–45]; see Table 1). As for specific modalities, six of the 18 trials assessed phone-based treatment [31–33, 35, 39, 40]. Four of these trials showed significant results at one or more follow-up. Of these four trials, three found that phone treatment (or a combined phone + internet approach) was more effective than a control condition and was equally effective to in-person treatment [31, 32, 35]. One study found that group phone treatment was more effective than individual phone treatment [39]. Similar to other comparisons of phone-based and in-person treatments, Donnelly et al. found no significant differences in weight loss between these modalities [33]. Finally, one trial found no differences in weight lost between two phone-based interventions (10 vs. 20 phone sessions) and a self-directed control group [40].

Seven of the 18 trials examined the effectiveness of internet-delivered treatment [29, 34, 36–38, 41, 42]. Four trials compared standard internet care (e.g., basic educational information with self-monitoring capabilities) to an enhanced internet program with tailored support and feedback [36, 38, 41, 42]. In two trials, enhanced internet programs led to significantly improved weight loss [36, 38], while the two other trials found no significant differences [41, 42]. One trial compared internet-based treatment to a usual care condition and found that the internet condition led to greater weight loss [29]. Finally, two of the seven internet-based conditions compared internet to face-to-face interventions [34, 37]. One trial found that an internet-only condition did not differ significantly in terms of weight loss when compared to an internet condition with supplemental in-person support [37]. The remaining trial found that in-person treatment produced the greatest weight losses, and an internet/in-person hybrid condition was more effective than an internet-only condition [34].

Two trials focused on the use of cell phones to deliver weight loss treatment. One trial examined the use of tailored text messaging to provide educational tips, feedbacks, and brief self-monitoring capabilities and found that the tailored messaging program was more effective than standard weight advice [30]. Another trial focused solely on self-monitoring strategies and compared the effectiveness of self-monitoring using a cell phone app, a website, or paper diaries. Participants using the app to self-monitor lost similar amounts of weight as participants who used the website, but both groups lost more weight than participants using paper diaries [43]. A related trial that focused on self-monitoring compared computerized self-monitoring alone to computerized self-monitoring plus enhanced behavioral strategies. Participants lost comparable amounts of weight, both of which differed from a control group [44].

Finally, two trials examined the use of podcasts to deliver treatment. One found that 24 podcast episodes based on social cognitive theory yielded greater weight loss than basic, educational weight loss podcasts [45]. The second trial compared treatment delivered by podcast to treatment including podcasts plus self-monitoring (using an online app) and a social media component. Both groups lost comparable amounts of weight [46].

Conclusions and Limitations

Due to substantial differences in treatment intensity and content across studies, it is difficult to draw conclusions about the consistency and magnitude of effects for alternative approaches to the delivery of weight loss treatment. The most consistent evidence relates to phone-based treatments, which generally demonstrate effect sizes similar to what is observed with traditional in-person interventions. Moreover, these effects are the most consistent of those found among alternative treatment modalities. Online and mobile-based treatment strategies show the most variation between trials, smaller effect sizes, and overall inconsistent findings across studies. Results of these trials are promising, yet future research is needed to refine and strengthen these treatments to create the type of individual engagement that is a central component of successful weight loss programs.

The magnitude of weight loss observed with telephone, computer, and mobile technology interventions may be largely based on the behavioral components and intensity of the programs. A recent review of technology-based weight loss programs indicated that programs need to include self-monitoring, clinician feedback/support, social support, and use of a structured and individually-tailored program to be successful [47]. In the current review, some interventions included very limited behavioral components, whereas others incorporated a full array of behavioral strategies. Also, only five of the eighteen trials compared an alternative method of treatment to a face-to-face approach [31–35], which limits the ability to assess whether alternative modalities can be as effective as conventional, in-person interventions. In addition, a challenge for examining technology-based interventions is the time it takes to develop and evaluate a treatment program before the technology itself becomes outdated or obsolete.

Clinical Implications and Future Directions

Based on findings from this systematic review, several clinical recommendations may be appropriate. First, there is currently limited long-term support for targeting different macronutrient compositions to improve weight loss. While low carbohydrate diets appear to enhance initial weight loss, there does not appear to be any significant advantage associated with diet composition with respect to long-term weight management. As recommended by current clinical guidelines, it may be more productive to instead find a diet that achieves moderate energy reductions that participants are willing and able to follow [48].

Second, providing prepackaged foods to participants appears to enhance adherence to energy intake reductions and thereby produce greater initial weight losses. The provision of foods may help some individuals “jump start” their weight loss efforts and could be particularly useful for participants who have limited time, motivation, and/or knowledge to plan, measure, and prepare their own reduced-calorie, healthier meals. However, the long-term benefits of food provision are less clear, particularly with respect to the transition to participants’ control of food purchases.

Third, delivering weight loss treatment via telephone and/or other modalities (i.e., internet, mobile technology) offers a potential alternative to in-person treatment and may be particularly useful for populations who would otherwise have limited access to conventional, in-person programs. Phone-based interventions appear to achieve short- and long-term weight reductions comparable to in-person approaches, while the use of internet and mobile technologies are associated with smaller reductions in body weight than face-to-face interventions. Regardless of modality, it is important that alternative treatments provide a sufficient “dose” of treatment and incorporate evidence-based behavioral strategies.

Several areas for future research should be considered. Regarding interventions targeting different dietary compositions, it would be worthwhile to investigate individual characteristics (i.e., behavioral, psychological, metabolic) potentially associated with preferential responses to certain dietary recommendations. For example, some intriguing work suggests that individuals’ hormonal response (i.e., insulin secretion) may influence the rate and magnitude of weight loss achieved with a low-carbohydrate diet [13]. If future research can identify individual characteristics that allow clinicians to successfully match participants to specific dietary recommendations, this could hold great potential for improving treatment outcomes.

Future work with meal provision should examine the relative benefits of this strategy when provided with other evidence-based behavioral techniques. Additional information is also needed on the number of meals that should be provided for optimal results as well as the most effective strategies and timeline for transitioning participants from provided foods to more independent meal planning and preparation.

Future studies examining alternative modalities for treatment delivery could focus on matching individuals to treatment modality based on individual preferences or characteristics. Also, including a combination of modalities or changing modalities (e.g., from phone to in-person treatment) over time may also help promote weight loss while maintaining participant interest and engagement. This notion of utilizing or changing modalities in a stepped-care fashion based on participants’ initial response to treatment may also have significant merit for treatment tailoring and improving outcomes.