Background: Protein intake above the RDA attenuates cardiometabolic risk in overweight and obese adults during weight loss. However, the cardiometabolic consequences of consuming higher-protein diets in free-living adults have not been determined.
Objective: This study examined usual protein intake [g/kg body weight (BW)] patterns stratified by weight status and their associations with cardiometabolic risk using data from the NHANES, 2001–2010 (n = 23,876 adults ≥19 y of age).
Methods: Linear and decile trends for association of usual protein intake with cardiometabolic risk factors including blood pressure, glucose, insulin, cholesterol, and triglycerides were determined with use of models that controlled for age, sex, ethnicity, physical activity, poverty-income ratio, energy intake (kcal/d), carbohydrate (g/kg BW) and total fat (g/kg BW) intake, body mass index (BMI), and waist circumference.
Results: Usual protein intake varied across deciles from 0.69 ± 0.004 to 1.51 ± 0.009 g/kg BW (means ± SEs). Usual protein intake was inversely associated with BMI (−0.47 kg/m2 per decile and −4.54 kg/m2 per g/kg BW) and waist circumference (−0.53 cm per decile and −2.45 cm per g/kg BW), whereas a positive association was observed between protein intake and HDL cholesterol (0.01 mmol/L per decile and 0.14 mmol/L per g/kg BW, P < 0.00125).
Conclusions: Americans of all body weights typically consume protein in excess of the RDA. Higher-protein diets are associated with lower BMI and waist circumference and higher HDL cholesterol compared to protein intakes at RDA levels. Our data suggest that Americans who consume dietary protein between 1.0 and 1.5 g/kg BW potentially have a lower risk of developing cardiometabolic disease.
Alex’s Notes: Last week we saw that higher protein intake in European adolescents was associated with a lower body fat percentage, a known risk factor for many diseases. We have also previously seen that higher-protein diets benefit infants. And of course, in adults, higher-protein intakes have been shown to be critical for maintenance of energy expenditure, preservation of lean body mass during dieting, and growth of lean body mass during overfeeding.
Evidence also suggests that in type-2 diabetics, consuming 25-30% of energy intake as protein results in greater weight loss, lower HbA1c, and lower blood pressure compared to a standard 15-20% protein diet. It is thus worth asking if these beneficial effects discovered in laboratory settings translate to the general population of free-living adults habitually consuming higher-protein diets. The current study was therefore an analysis of nearly 24,000 adults (>19 years) with complete and reliable 24-hour dietary recall data from the NHANES 2001-2010 cohort. Because underweight (BMI < 18.5) individuals are more likely to have an eating disorder or chronic disease, they were not included in the study population, and pregnant or lactating women were excluded as well.
Gathering various cardiometabolic risk factors from examination and medical records of the participants, several statistical models were made, of which we are only interested in the final model that controlled for sex, age, race/ethnicity, physical activity, poverty-income ratio, energy intake, carbohydrate intake, fat intake, and BMI (for the cardiometabolic risk factors only). This is because it does the best job of controlling for other dietary confounders that may influence the results, and for the statistical nerds, significance was set at p < 0.00125, meaning that the significant outcomes have a 0.125% chance of being a random act of nature.
So what is America eating?
Habitual protein intake was divided into deciles to help with analysis, and it is interesting to note that protein intake (g/kg bodyweight) more than doubled from decile one (0.69 g/kg) to ten (1.51 g/kg). Females were more prevalence in deciles 1-5, while males dominated deciles 6-10, and average age showed a consistent decline as protein intake increased. These trends held true when the data was stratified by BMI as well.
As protein intake increased, so did total energy intake, leading to greater absolute consumption of fat and carbohydrates. Fat as a percentage of energy intake also showed a small and steady increase from 30 to 35% of calories, but carbohydrate intake declined from 57 to 44% of calories, suggesting that even small increases in protein intake is mainly offset by carbohydrates within the diet, even though absolute food intake still increases. Interestingly, even in the highest protein consumers, intake never exceeds about 19% of caloric intake, which according to many experimental studies is not “high” protein. This speaks to protein’s known satiety effects that make overconsumption difficult unless one actively tries.
The above macronutrient trends were similar within the different BMI categories as well, but it is worth pointing out that although protein intake was similar when expressed as grams per day or percentage of energy intake, there was a clear decline in protein grams per kilogram bodyweight as BMI increased from normal to obese (1.64 à 1.49 à 1.33).
Keeping in mind the model we are looking at is the fully adjusted version, protein intake was significantly negatively associated with BMI and waist circumference, and every decile increase in protein intake led to a 0.47 and -1.195cm reduction in BMI and waist circumference, respectively. When stratified by weight, the significant negative association held for waist circumference in all BMI categories as well, but the significant association with BMI only remained in the overweight and obese groups. Nonetheless, as BMI increased, the magnitude of increasing protein’s effect increased.
For blood-borne markers of cardiometabolic health, protein intake was not significantly associated with blood pressure, glucose, insulin resistance, LDL-c, or triglycerides. The only significant association was with HDL-c, which showed an overall positive association with protein intake. All these associations were similar in men and women with no differences observed when stratified by sex.
So what does this mean for protein?
The two bottom lines here are that Americans habitually consume protein in excess of the RDA (0.8 g/kg), and that higher-protein diets are independently associated with lower BMI and waist circumference, and greater HDL-c. Specifically, HDL concentrations were about 15% greater in those consuming 1.5 g/kg of protein compared to those consuming the RDA. It is also important to emphasize that the lower waist circumference suggests that central and visceral adiposity was lower in those consuming higher-protein diets, and this association was present in all BMI categories, albeit stronger in those who are overweight or obese. More importantly, this analysis showed no detrimental associations between protein intake and cardiometabolic risk, and because these associations were independent of energy, carbohydrate, and fat intake, the intrinsic properties of protein appear to be partially responsible for these effects.
So what if we are interested in how protein intake influences other dietary components?
It is a good question, as we know that protein intake in and of itself has benefit limited to BMI, waist circumference, and HDL-c, but what if we step back and remove the adjustments for carbohydrate and fat intake? That is, what if we look at the researcher’s second model that controlled for sex, age, race/ethnicity, physical activity, poverty-income ratio, and energy intake only? Since consuming protein displaces other macronutrients in the diet, this would allow us to see how the small increases in fat intake and reductions in percentage of calories from carbohydrates also play a role while still eliminating total energy intake from the equation.
In this case, there was still a significant overall negative association between protein intake and BMI and waist circumference, but the BMI-stratified outcomes were no longer significant with the exception of overweight persons and waist circumference. There is also a significant positive association with fasting blood glucose, but an even more significant negative association with insulin resistance, suggesting that displacing carbohydrates in the diet with protein and fat leads to improved insulin sensitivity. Finally, the positive association with HDL-c remained significant, but the magnitude of the effect was attenuated slightly.
The bottom line
Increasing protein intake to levels that never exceed 19% of energy intake but represent about 1.5 g/kg bodyweight, on average, results in reduced BMI, waist circumference, and increased HDL-c independent of total energy intake, carbohydrate intake, or fat intake, suggesting that the outcomes are at least partially owed to an intrinsic aspect of protein. If increasing protein intake displaces carbohydrates in the diet, then improvements in insulin sensitivity can be seen as well, although it must be kept in mind that the carbohydrate reduction was from over 50% to just above 40% of caloric intake.