Higher protein intake during caloric restriction improves diet quality and attenuates loss of lean body mass

Abstract

Objective:

Higher protein intake during weight loss is associated with better health outcomes, but whether this is due to improved diet quality is not known. The purpose of this study was to examine how the change in self-selected protein intake during calorie restriction (CR) alters diet quality and lean body mass (LBM).

Methods:

In this analysis of pooled data from multiple weight loss trials, 207 adults with overweight or obesity were examined before and during six months of CR (approximately 10 food records/person). Body composition was measured by dual-energy x-ray absorptiometry. Diet quality was assessed using the Healthy Eating Index (HEI-2015) in two groups: lower (LP) and higher (HP) protein intake.

Results:

Participants (54±11 years; 29±4 kg/m²) lost 5.0±5.4% weight. Protein intake was 78±9 g/d (1.0±0.2 g/kg/d) and 58±6 g/d (0.8±0.1 g/kg/d) in the HP and LP groups, respectively (p<0.05) and there was an attenuated LBM (kg) loss in the HP (−0.6±1.5%) compared to LP (−1.2±1.4%) group (p<0.01). The increased HEI score in the HP compared to LP group is attributed to greater total protein and green vegetable, and reduced refined grain and added-sugar intake (p<0.05).

Conclusions:

Increasing dietary protein during CR improves diet quality and may be another reason for reduced LBM but requires further study.

INTRODUCTION

Individuals with overweight or obesity often have poor quality diets that lack fruits and vegetables, whole grains, legumes, and contain excessive amounts of added-sugar, and saturated, compared to unsaturated fatty acids (1) that can lead to higher risk of greater weight gain and chronic disease (2, 3). Weight loss (WL) of 5-10% may prevent chronic disease (4); however, WL diets that restrict energy also reduce healthy food and micronutrient intake (5). The effect of higher protein intake during WL on health outcomes has been reported extensively and there is evidence that it can promote a healthy body weight, attenuate loss of muscle mass (6, 7) and reduce chronic disease (8–10). In addition, it has been shown that dietary protein contributes to nutrient adequacy in the general population (11). However, the impact of self-selected dietary protein on diet quality has not been examined in a longitudinal study (11), such as during caloric restriction (CR). The link between protein intake and diet quality is important since diet quality is suboptimal in the USA, and higher protein WL diets are popular. In addition, the nutrient adequacy in populations consuming lower energy intake, such as in young children and older individuals, or during calorie restricted diets requires more research. It is possible that if dietary protein affects intake of other foods and diet quality, that this can provide further insight into outcomes associated with low calorie higher protein diets.

Epidemiological studies indicate that use of diet quality indices, rather than single nutrients or food intake in isolation provides a comprehensive analysis of dietary intake (12). The Healthy Eating Index (HEI) aligns with key recommendations of the Dietary Guidelines of Americans (DGA) with a goal to achieve recommended nutrient intakes within the recommended energy intakes (13). The proportional scoring structure of the HEI is an appropriate metric to examine longitudinal change such as in a WL trial (14). Evidence indicates that a better diet quality has health benefits and is associated with reduced weight gain (2) or greater WL (15, 16), yet there is little known about how HEI is influenced by protein content of the diet.

To address the knowledge gaps, a pooled analysis of completed trials was performed to maximize the number of participants undergoing a similar protocol for moderate WL. Lifestyle modification (i.e. diet and behavior therapy) using group counseling was delivered in 16 sessions over six months (17, 18) similar to that recommended by the Guidelines for Management of Overweight and Obesity in Adults (18). The primary goal in this study was to determine how changes in self-selected protein intake during CR affect intake of other foods (with low or zero protein), diet quality using HEI-2015, and nutrient adequacy during six months of WL in adults with overweight and obesity. Whether changes in protein intake and other foods in the diet are associated with lean body mass (LBM) after WL in this population was also examined. It was hypothesized that higher protein intake during WL would improve diet quality and attenuate LBM loss compared to lower protein intake.

METHODS

Trial Designs

This analysis included pooled data of multiple trials from the same laboratory at Rutgers University (19–23) in which participants followed a 6-12-month WL intervention, with weekly counseling sessions during the first 8 weeks and at least twice monthly sessions thereafter with a registered dietitian nutritionist (RD/RDN). Because all participants completed 6-months of WL, this time point was used for the current analysis.

This group of WL trials in our laboratory that were funded by the National Institutes of Health (AG-12161) are included in the Osteoporosis, Weight Loss and Endocrine database (OWLE). Trial registration of original studies is at ClinicalTrials.gov NCT01631292, NCT00473031, NCT00472745, and NCT00472680 (Table S1). These clinical trials were selected because participants had body mass index (BMI) >25 kg/m² and participated in at least 6 months of nutrition education and behavior modification using similar protocols. Additionally, there was consistency among these trials with study staff adhering to evidence-based practices (i.e., registered dietitian nutritionist), and standard operating procedures to enhance validity.

In the original trials, participants received calcium and vitamin D supplementation, or assignment to high or normal protein groups. All participants were encouraged to lose weight following a 500-calorie deficit diet and to consume healthier foods using the Academy of Nutrition and Dietetics/American Diabetes Association’s Food Lists for Weight Management. Participants included in these studies were advised to consume protein intake (18% of the calories) from less processed sources (i.e., poultry, red meat, fish, legumes, dairy). The subset (n=24) included in this study who were encouraged to consume higher protein intake (23), reported intake ranging from 13-29% which was similar to the other 183 participants who were asked to consume 18% protein (10-28%).

Participants were informed on procedures for reporting accurate food records during initial phases of screening and intervention. Baseline 24-hour recall and monthly food records were documented. Physical activity level (PAL) was based on the reported time spent in walking for exercise or in other walking and scored from 0-3 (24). All participants were advised to maintain their usual physical activity during the intervention. Height at baseline, and monthly weight were measured on a stadiometer and digital balance scale, respectively. Body composition (fat mass and lean body mass, LBM) was assessed using dual energy x-ray absorptiometry (Lunar Prodigy-Advanced, GE-Lunar. Madison, WI, USA) at baseline and 6 months.

Participants

In the original WL trials, adult women and men with overweight and obesity were screened between 2000 and 2012 for eligibility in the primary WL studies (19–23). Participants were recruited from the NY-NJ-PA metropolitan area in the United States. Participants between the ages of 24-75 years with a BMI 25-40 kg/m² (≥23 kg/m² if Asian) were recruited for these studies. A brief physical exam, for screening purposes, confirmed all participants met inclusion criteria, as described previously (19–23). In these WL trials, participants randomized to weight maintenance were excluded from this analysis. Participants were also excluded from this analysis if they did not report a 24-hour recall at baseline or no food records during the intervention. Participants completed an informed consent before enrollment in the original clinical trials, approved by the Rutgers University Institutional Review Board.

Power Analysis-

In a previous study (25), with a 16% difference in HEI between groups (WL with a 2.3% greater protein intake than weight maintenance controls), a sample size of at least 21 per group would provide 90% power to detect a difference in HEI due to WL. In another study, examining the HEI among individuals who eat animal protein products, an estimated group of at least 63 persons is needed to detect a 4.8 score difference between groups (β of 0.90 and α of .05) (26). Based on these two studies, and allowing for some missing data, we estimated that with one covariate, 100 persons per group would be needed.

Dietary Assessment and Calculations

Food records that were collected from the original trials were never previously analyzed or reported. Baseline 24-hour recall and food records were collected monthly using forms provided by the laboratory and validated by an RD. Over the WL duration, 3-day food records (2 weekdays, 1 weekend per week) were entered at 1 month, 3 months, and 6 months. These food records were entered by the RD or staff (verified by the RD) into Automated Self-Administered 24-Hour Dietary Assessment Tool (ASA24, USA version; NCI-NIH). Diet quality, nutrient intake, and sources of dietary protein (red meat, processed meat, poultry, organ meat, seafood, eggs, soy, nuts, seeds, and legumes) were examined.

Diet Quality Index -

The Healthy Eating Index encourages intake of fruits, vegetables, whole grains, and protein foods high in fiber and unsaturated fats. Additionally, the HEI discourages intake of saturated fatty acids (SFA), refined grains, added sugar, and sodium. The HEI (calculated with SAS v9.4 code; National Cancer Institute and US Department of Agriculture) has 13 food components (scored as 0-5 or 0-10) with a maximum score of 100 (14). The food components include: Whole Grains; Total Fruit (includes fruit juice); Whole Fruit (excludes fruit juice); Total Vegetables; Greens and Beans (dark green vegetables and all types of legumes); Total Protein Foods (meat; poultry; eggs; legumes); Seafood and Plant Proteins (fish, shellfish, nuts, seeds, legumes, soy foods [not beverages]); Dairy (milk, yogurt, cheese, and fortified soy beverages); Fatty Acids ratio (polyunsaturated and monounsaturated fatty acids to SFA). Foods that are unfavorable to the HEI include: Refined grains, Sodium, Added-Sugars, and SFA (13). The HEI incorporates energy intake into the score making it applicable for CR diets that vary in calories. Unlike other indices that use population means or dichotomous scoring structure for comparing populations, the HEI uses proportional scoring making it better suited to track changes over time. It also includes a greater range of protein sources to address questions related to dietary protein in this study.

Statistical Analysis

Descriptive statistics were used to define baseline characteristics. Data were assessed for skewness and normality, and Levene’s test for equal variances analyzed homoscedasticity. To examine the change in dietary protein from baseline to CR, the sample population was divided by the median (lower protein, LP, and higher protein, HP). Mixed models ANCOVA was used to assess how the longitudinal changes in dietary protein affected change in other foods in the lower or higher protein groups (controlled for baseline protein intake). One-way ANCOVA assessed the change in body composition between protein groups, controlling for age and sex. A sensitivity analysis was also performed using mixed model ANCOVA with dietary protein examined as a continuous variable rather than categorical. Multiple imputation was used to handle partial missing body composition data (n=8). A sensitivity analysis was completed to compare findings with and without multiple imputation for body composition. Multiple linear regression was used to determine the extent that protein sources contributed to change in protein intake and LBM in the total sample. In addition, multiple linear regression was used to determine how the change in protein intake contributed to the change in micronutrient intake (adjusted for age, sex, and BMI) and to determine which food components predicted improvement in HEI during CR. The IBM-SPSS v27 statistical software (Armonk, NY) was utilized for the analyses.

RESULTS

Participants Characteristics

There were 314 individuals who participated in the trials, and 272 were assigned to WL and measured for body composition at baseline and 6 months (Figure S1). Participants (n=209) who recorded intake at baseline and during the intervention were eligible for this study. Two participants were removed for lack of feasibility of reported dietary intake: one due to very low protein and energy intake during WL and one as an outlier for high protein intake (>3 SD). Analysis included 24-hour recalls at baseline (n=207) and 1,870 food records during CR (~10 days of intake/person). This sample is predominantly female (88%), Caucasian (85%) and overweight (29.1±4.1 kg/m²; Table 1).Baseline body weight did not differ significantly between LP (77.3±14.1 kg) and HP groups (81.1±14.7 kg). Fat mass and LBM also did not differ between groups [LP; (LBM: 42.6±8.3 kg, fat mass: 32.3±7.7 kg) and HP; (LBM: 44.9±8.6 kg, fat mass: 34.8±9.0 kg)].

Table 1.

Baseline Characteristics

Age (years)	54.4 ± 10.7
Female	183 (88%)
Race/Ethnicity
Caucasian	176 (85%)
African American	24 (12%)
Other	7 (3%)
Weight	79.7 ± 14.6
Lean Mass	33.5 ± 8.4
Fat Mass	43.8 ± 8.5
BMI (kg/m2)	29.1 ± 4.1
Overweight (25-29.9)	138 (67%)
Obese (≥30)	69 (33%)
Dietary Intake
Energy (kcal/d)	1829 ± 626
Protein (%kcal)	17.5 ± 5.3
Carbohydrate (% kcal)	48.9 ± 9.9
Fat (% kcal)	33.7 ± 8.5

Categorical variables presented as n (%). Age represents Mean ± SD. n=207

Calorie restriction

Participants lost 5.0±5.4% (p<0.001) of their baseline body weight (79.7±14.6 kg). During CR, the LP group consumed less protein at 58.3±6.6 g/d compared to the HP group,78.6±9.4 g/d (Table 2). Also, the percent protein intake during CR was lower in the LP (17.8±2.9%) compared to HP group (19.9±2.8%) (p<0.001). During CR, carbohydrates (LP, 51±5.3% vs HP, 47.4±5.8%) differed between groups (p<0.001). Dietary fat intake averaged 32.8±4.8% and did not differ between the groups. The PAL score was 1.0±0.7 and physical activity did not differ significantly between LP and HP groups or change over time. Both protein groups lost a similar amount of weight and fat mass during CR, but there was a greater decrease in LBM in the LP compared to HP group (p<0.01), whether calculated as percent (not shown) or kilogram loss (Figure 1).

Table 2.

Mixed models assessment of diet quality food scores before and during calorie restriction (CR) by protein intake.

	Baseline (n=207)	Caloric Restriction		P value
Dietary Protein		Lower Protein	Higher Protein		Interaction
grams/day	77.8 ± 29.2	58.29 ± 6.61	78.56 ± 9.36		<0.001
grams/kg/day	0.99 ± 0.38	0.84 ± 0.14	1.01 ± 0.21		<0.001
HEI components (total score)
Total Protein (5)	4.4 ± 1.2	4.1 ± 0.7	4.5 ± 0.5		<0.001
Seafood/Plant Protein (5)	2.9 ± 2.3	2.5 ± 1.2	3.2 ± 1.1		<0.001
Total Dairy (10)	5.3 ± 3.4	6.1 ± 1.8	6.1 ± 1.8		0.028
Total Vegetables (5)	3.5 ± 1.6	3.9 ± 0.8	3.8 ± 0.7		0.821
Greens & Beans (5)	2.4 ± 2.3	2.4 ± 1.2	2.6 ± 1.1		0.004
Total Fruit (5)	2.9 ± 2.0	3.5 ± 1.1	3.1 ± 1.2		0.431
Whole Fruit (5)	3.1 ± 2.2	3.6 ± 1.2	3.5 ± 1.1		0.822
Whole Grains (10)	3.4 ± 3.6	3.7 ± 2.1	4.1 ± 2.2		0.533
Refined Grains a (10)	5.8 ± 3.8	6.4 ± 1.9	6.7 ± 2.1		<0.001
MUFA+PUFA/SFA (10)	5.0 ± 3.9	5.7 ± 1.8	5.5 ± 1.6		0.642
SFA a (10)	6.0 ± 3.6	7.0 ± 1.5	6.4 ± 1.9		0.205
Added Sugar a (10)	8.0 ± 2.3	7.8 ± 1.5	8.3 ± 1.2		0.039
Sodium a (10)	3.6 ± 3.3	3.7 ± 1.9	3.2 ± 1.6		0.873
HEI (100)	56.5 ± 18.5	60.4 ± 8.0	60.9 ± 7.6		0.010

Values are reported as Mean ± SD.

The lower protein (n=104) and higher protein (n=103) are shown using linear mixed models ANCOVA (controlling for baseline protein intake). Baseline diet quality and HEI component scores did not differ significantly between groups, and dietary protein (% of energy intake) was 17.8 ± 2.9% and 19.9 ± 2.8% in the LP and HP groups, respectively (p<0.001).

^aHEI moderation component (higher score indicates lower intake and healthier diet).

Abbreviations: Healthy eating index-2015 (HEI), monounsaturated fatty acid (MUFA), polyunsaturated fatty acids (PUFA, saturated fatty acids (SFA)

Figure 1.

Body composition change due to six months of weight loss. Values are mean ± SD for change in fat, lean mass, and total weight (kilograms, kg) and compare the lower protein (n=104) and higher protein (n=103) groups using ANCOVA (adjusted for age and sex).

Diet quality

The HEI score improved in the entire sample from baseline to CR over 6 months (p<0.05). Multiple linear regression models were constructed to determine foods that contributed to the improved HEI with CR indicating that total fruit, whole grains, refined grains, dairy, and seafood/plant protein sources all contributed to the rise in HEI in the unadjusted and adjusted models (p<0.05; Table S3). In the HP compared to LP group, HEI component scores that improved during CR include dairy, added sugar (p<0.05), total protein, seafood/plant protein, refined grains, and vegetable greens and beans (p<0.01; Table 2). Diet quality (HEI) improved over time in both groups with a greater increase in the HP compared to LP group (p<0.05; Table 2). Intake of SFA (23.3±14.5 g) added sugar (46.8 ±38.1 g), refined grains (171±83 g), and sodium (31.8±11.8 g) decreased during CR (p<0.05) but did not differ between groups. In the sensitivity analysis using protein as a continuous variable, protein intake also contributed significantly to the increase in HEI and the same dietary components. The change in protein intake by group was also analyzed (mixed models ANCOVA) and indicated that HP improves scores for HEI and certain food components (Figure S2).

Total protein intake was 68.4±3.0 g/day in the entire sample during CR and was largely from animal protein (46.8±16.3 g/d). Protein sources that accounted for 70% of variance for the change in protein intake from baseline to CR were poultry, unprocessed red meat, seafood, cured meat, cheese, milk, eggs, and nuts/seeds (F=50.6; p<0.001; Table S2). Protein sources that were not consumed in significant quantity and did not contribute to the change in protein intake were organ meats, legumes, yogurt, and soy products. In addition, of all protein sources, only poultry accounted for variance in LBM change with β=0.433 (95%CI, 0.042, 0.824; p<0.05) (not shown). Furthermore, when examining whether food categories (fruit, vegetables, grains, protein, dairy, oils/fat), only total protein could explain the variance for LBM change (β=1.919; 95%CI 1.384, 2.455; p<0.001).

Micronutrients

Baseline nutrient intake was below the RDA for calcium, magnesium, potassium, choline, fiber, vitamins D and E. Micronutrient intake further decreased (p<0.05) for all micronutrients during CR, except vitamin K, D, C, B₁₂ and vitamin A. In addition, a greater intake of protein during CR was associated with most vitamins and minerals (excluding vitamin B₁, folate, vitamin E, copper, and iron; Table 3).

Table 3.

Regression coefficients and 95% CI examining the contribution of protein intake on the change in micronutrient intake from baseline to calorie restriction.

Vitamins	B (95% CI)	P value	Mineral	B (95% CI)	P value
Vitamin B3	0.267 (0.224, 0.310)	<0.001	Selenium	1.087 (0.894, 1.281)	<0.001
Choline	3.706 (3.142, 4.271)	<0.001	Phosphorus	9.670 (8.277, 11.062)	<0.001
Vitamin B6	0.023 (0.017, 0.029)	<0.001	Potassium	15.320 (11.086, 19.553)	<0.001
Vitamin D	0.090 (0.059, 0.121)	<0.001	Sodium	12.884 (7.374, 18.395)	<0.001
Vitamin B2	0.010 (0.006, 0.013)	<0.001	Zinc	0.053 (0.03, 0.077)	<0.001
Vitamin K	1.935 (0.819, 3.051)	<0.001	Magnesium	1.075 (0.502, 1.649)	<0.001
Vitamin B12	0.077 (0.027, 0.127)	0.003	Calcium	2.519 (0.266, 4.772)	0.029
Vitamin A	8.393 (2.948, 13.838)	0.003	Copper	0.007 (−0.001, 0.016)	0.088
Vitamin C	0.518 (0.032, 1.004)	0.037	Iron	−0.022 (−0.057, 0.012)	0.205
Vitamin B1	0.004 (−0.002, 0.009)	0.165
Folate	0.138 (−1.029, 1.305)	0.816
Vitamin E	0.005 (−0.028, 0.037)	0.767

n=207; Controlled for energy, age, sex, and body mass index.

DISCUSSION

A higher quality diet has been associated with WL, and is largely attributed to higher fiber, fruit and vegetable intake, and controlled portion sizes (15, 16). Consistent with this, individuals in the current study show an improvement in diet quality during CR compared to baseline. In addition, higher protein intake during CR has been shown to contribute to certain health outcomes (10). These benefits may be attributed to a higher protein intake alone and/or alterations in dietary patterns caused by a change in protein intake. To our knowledge, no previous study has examined how the amount of protein intake during CR affects diet quality. In this study, we use multiple food records during 6 months of WL to examine whether a change in self-selected dietary protein affects diet quality and food component scores. We found that that individuals with higher protein (78 g/d) compared to lower protein (58 g/d) intake during CR had a greater improvement in diet quality. In addition, the change in protein intake in the HP compared to LP groups resulted in not only greater total protein and dairy scores (as expected), but also greater intake of dark green vegetables, and reduced intake of refined grains and added sugar. Overall, WL in individuals consuming higher compared to lower protein intake attenuated loss of LBM, which was somewhat expected, but they also altered low or zero-protein foods which improved diet quality. In this study, individuals who self-selected higher protein intake during CR, lost less LBM than those who consumed lower protein intake. An increased protein intake of 1.0 g/kg/d is not especially high, and adherence to this should be achievable with nutrition counseling, even in individuals who typically consume low protein intakes. The benefits of more LBM and greater insulin sensitivity with increased protein intake have been studied extensively (7, 27–30), yet there is still lack of clarity whether the health benefits are entirely due to protein intake alone when individuals are self-selecting sources of protein during WL (31). The Pounds Lost trial examined 2 levels of protein intake (18% and 20% of calories) indicating a trend for greater LBM loss in the lower protein group (32). This protein intake is comparable to the LP and HP groups in our current study, which indicated a significantly greater loss of LBM due to LP intake. The Pounds Lost study (57% women) showed there is a greater LBM loss in women than (33), and this may one reason for greater LBM differences between groups in the current study that also had a greater proportion of women (88%).

In the Health ABC study, self-selected protein intake was examined over 3 years in 2,066 older individuals who were not undergoing WL (30). In this study (30), the lower and upper two quintiles of protein consumed were 55 g/d compared to 79 g/d and change in LBM was −0.9 kg and −0.5 kg, in the two groups, respectively. In the current WL study, the lower and higher protein groups consumed 58 and 78g protein/d with LBM change of −1.0 and −0.6 kg, respectively. Together, these findings suggest that a higher dietary protein of about 80 g/d (or 1 g/kg/d) preserves LBM compared to a normal protein intake (of ~60 g/d or 0.8 g/kg/d) and may be especially important to consider in females and the elderly who are more susceptible to consuming inadequate dietary protein. The current study indicated that during reduced energy intake, lean protein sources (largely poultry) accounted for a significant proportion of variance in LBM change during WL. Other factors, such as the reported physical activity, was low in this population and did not differ between groups or over time, and therefore does not explain the attenuated LBM loss in the HP group. In individuals who are not losing weight, diet quality is associated with higher LBM, with or without physical activity (34). Understanding how to preserve LBM during CR continues to be a concern in the field and future studies might explore further whether protein sources in combination with other foods affect LBM.

Identifying high quality proteins and the amount required for sufficient intake is influenced by several variables and remains controversial. Most measures of protein quality by amino acid composition and digestibility (protein digestibility-corrected amino acid score, and digestible indispensable amino acid score) suggest proteins from animal sources are more complete to varying degrees than plant sources (35). However, other nutrients often consumed in excess in the Western diet that tend to accompany animal-protein intake may reduce the benefit of including ‘complete’ proteins in the diet. For example, observational studies show that in individuals who are not dieting, protein consumption from animal sources is associated with a dietary pattern that has a greater intake of saturated fat, cholesterol, sodium, and added sugar, and less fiber, all of which lower diet quality (36, 37). However, unlike reported high intakes of added sugar and sodium, SFA intake in the USA (while still high) is closer to recommended levels (≤10% of total energy intake; or ≤ 22 g/d) (38). This is consistent with findings in this study, showing that SFA was 23 g/d at baseline and decreased in both protein groups during CR. In this study, a greater intake of total protein during CR was largely from lean meats like poultry and unprocessed red meat, as well as seafood, dairy, and nuts and seeds. Consequently, HEI components such as sodium and SFA that may have worsened with increased total protein intake, were at healthier levels during CR and did not differ between protein groups. Multiple other studies have shown that individuals consuming plant-based diets meet protein needs and have a higher quality diet than omnivores (36, 39), but this tends to be driven by fruit and vegetable intake, not necessarily due to higher quality protein intake (35). One study concludes that diets with protein sources that are predominantly from lean meat or plant sources have similar diet quality (40). In this study we show that higher protein intake (from lean animal and plant sources), with dietary counseling to support WL, can contribute to a healthier dietary pattern, as indicated by consumption of less refined grains and added sugar, and more green vegetables.

In addition to a high quality diet, micronutrient intake contributes to nutrient adequacy during CR and high protein foods are good sources of minerals (11). Our data indicate that many of the same nutrients that are below recommended intakes in the general population (41) were also low in participants in this study. Not surprisingly, these micronutrients became further compromised during modest WL. However, greater protein density was positively associated with multiple B vitamins, choline, and vitamins A, C, D, K, as well as multiple minerals. Others have found that protein from animal sources increase consumption of zinc, potassium, vitamin B₁₂, riboflavin, and folate (11, 40). Protein from plant sources contributes to higher intakes of calcium, copper, folate, potassium, magnesium, and thiamin (11, 42). Our data during CR indicate that protein intake as well as the changes in dietary pattern associated with altered protein intake are important contributors to increased micronutrient intake.

A strength of this work is that no previous study has examined how protein intake during an energy restricted diet alters diet quality and patterns of food intake. This sample included pre- and post-menopausal women as well as men, increasing generalizability to the American population, but it is largely limited to a Caucasian population. Additionally, all participants were counseled by dietitians using the same nutrition education behavior modification intervention, increasing the consistency of nutrition counseling. This diet prescription and the counseling would be expected to and did improve diet quality. However, because the studies were conducted within one laboratory, this could limit generalizability of the findings. While the lower protein intake may reflect general lower adherence to the provided advice, both groups lost a similar amount of weight suggesting that adherence to reduced calorie intake was similar between groups. Since methods used in this study to measure LBM do not discern between organ and muscle mass, the greater loss in the LP group during WL cannot be attributed only to muscle mass and its associated health benefits. However, muscle mass (but not organ mass) correlates with WL (43). Another limitation is that individuals with overweight and obesity typically underestimate intake (44). However, because we analyzed 10 food records to estimate dietary intake, and a dietitian educated them to accurately report food intake and reviewed this with participants to reinforce validity of intake, it is expected that the quality of the nutrient analysis was enhanced (45).

Conclusions

These findings indicate that a moderately higher protein intake during CR improves diet quality and attenuates loss of LBM. The self-selected higher protein intake during WL improves diet quality largely due to consumption of low-fat protein sources, greater intake of green vegetables, and reduced intake of refined grains and added sugar to better align with the DGAs. While the findings in this pooled analysis indicate that only dietary protein explained the variability in LBM changes during CR, the link to other food choices should be explored as a possibility in future studies. Accordingly, it would be interesting to determine if higher self-selected protein intake that improves diet quality, as compared to a protein supplement alone differentially affects LBM or other health outcomes. Also, future WL studies are needed to determine if the dietary shifts in relation to self-selected protein intake are consistent in individuals consuming different ethnic food patterns, or in vulnerable populations at risk for meeting nutrient adequacy such as in children or the elderly.

Study Importance Questions.

What is already known?

• Obesity is a heterogeneous disease often associated with a poor-quality diet.

• Intensive nutrition counseling for weight loss can increase diet quality, yet the role of higher protein intake on diet quality during caloric restriction is not known.

• Understanding the link between diet quality and protein intake during weight loss is important since higher dietary protein is associated with attenuated loss of lean body mass and other reported health benefits.

What are the new findings?

• Individuals with overweight and obesity improved the quality of their diet more with a higher (79 g/d) compared lower (58 g/d) protein intake.

• Individuals who self-select a diet higher protein during caloric restriction (CR) compared to lower intake, also reduced intake of low or zero-protein foods including refined grains and added sugar, and increased intake of green vegetables.

• Greater protein intake and diet quality during CR attenuates loss of lean body mass.

How might these results change the focus of clinical practice?

• Moderately higher protein intake (1.0 g/kg/d) at 20% of energy intake with weight loss counseling can be encouraged for successful weight loss, improved diet quality, and attenuated loss of lean body mass.

• Counseling for weight management that recognizes there is a wide potential range of protein intake which is interconnected to other food choices, may improve the quality of advice to patients.

Data Availability:

Not applicable due to IRB regulations on open data access for these studies. The data that support the findings of this study are available from the corresponding author upon reasonable request