A Nutrient Ratio-Based, Web-Enabled Food Quality Score Is Associated With Weight and Blood Pressure Compared With Leading Nutrient Profiling Systems Authors: Christopher J Damman, Cara L Frankenfeld Published: 2026 ABSTRACT AbstractBackgroundRising prevalence of chronic cardiometabolic conditions may be partly driven by shifts in dietary patterns. Nutrient profiling systems (NPSs) aim to guide healthier food choices through labeling and consumer-facing technologies but vary in accessibility and how well they distinguish food healthfulness.ObjectivesThe primary objective was to compare a web-enabled, ratio-based NPS, Nutrient Consume Score (NCS), and its underlying nutrient ratios with 4 other leading NPSs, Nutri-Score, Health Star Rating, NOVA Classification, and Food Compass 2.0, by examining associations with obesity and blood pressure. Secondary objectives included assessing associations with cardiometabolic biomarkers and identifying food categories contributing most to each score.MethodsCross-sectional study of NHANES 2015–2016 data from 9971 adults aged ≥20 y were analyzed. Dietary intake was assessed using a single 24-h dietary recall (day 1), and NPS scores were calculated. Multivariable regression models adjusted for sociodemographic and health factors examined associations with obesity, blood pressure, and cardiometabolic biomarkers. Compositional analysis evaluated food categories driving scores.ResultsHigher NCS was associated with lower BMI [β = −0.64 kg/m2 per 10-unit increase; 95% confidence interval (CI): −0.86, −0.41; P < 0.0001], waist circumference (β = −1.63 cm; 95% CI: −2.23, −1.02; P < 0.0001), systolic blood pressure (β = −1.01 mm Hg; 95% CI: −1.67, −0.35; P < 0.0051), and diastolic blood pressure (β = −0.56 mm Hg; 95% CI: −0.90, −0.23; P < 0.0026), with effect sizes comparable with other NPSs. Calorie-to-weight, sodium-to-potassium, and saturated-to-unsaturated fat ratios were associated with weight outcomes, whereas sodium-to-potassium and carbohydrate-to-fiber ratios were associated with blood pressure outcomes.ConclusionsThe NCS was associated with weight and blood pressure in this cross-sectional analysis, with effect estimates comparable in magnitude with other commonly used NPSs. These findings support the potential utility of a ratio-based, web-enabled NPS for assessing diet quality in relation to cardiometabolic risk factors.Keywords: nutrient ratios, obesity, hypertension, nutrient profiling systems, diet quality Introduction The rising incidence and prevalence of obesity [1], metabolic syndrome [2], and other noncommunicable disease [3] in the United States and globally [4] has been linked to shifts in the types of foods available to consumers including increased ultraprocessed [5] and hyperpalatable foods [6] with resulting changes in dietary habits [7]. Complex, confusing [8], and conflicting [9] nutrition guidance makes it difficult for individuals to make informed, healthy choices on food purchases, preparation, and consumption [10]. Nutrient profiling systems (NPSs) aim to simplify nutrition complexity into easier-to-follow ratings that assess the health quality of foods and beverages. These systems can offer real-time guidance through package labeling [11], online websites [12], and smartphone technologies [13,14], helping individuals make healthier food choices. Additionally, they can support public health recommendations [15] and guide food companies in creating healthier products [16,17]. Although there are over 100 reported NPSs [15,18], a small subset has been rigorously evaluated for their association with health outcomes. The most studied NPSs with strongest links to improved dietary choices and better health are Nutri-Score (NS) [19], Health Star (HS) [20], NOVA Classification (NC) [21], and Food Compass (FC) [22,23]. NS [24], popular in several European countries, and HS [25], widely used in Australia and New Zealand, rank foods based on nutrient profiles and food categories. The NC categorizes foods by their degree of processing, popularizing the term ultraprocessed food [26]. FC evaluates foods across 9 health domains, including categories like nutrient ratios, fiber, and phytochemicals [23,27]. These systems are variably effective in assessing the relative quality of different ultraprocessed foods [28] and not available on front of product packaging in many countries including the United States [23]. Despite the availability of multiple NPSs, a gap remains for a rigorously studied, web-enabled algorithm that is independent of country-specific labeling practices and that can scale, based on relative rather than absolute nutrient measures. The Nutrient Consume Score (NCS) [12] is a recently developed web-based algorithm that focuses on nutrient ratios—carbohydrate-to-fiber [29], saturated fat-to-unsaturated fat [30], sodium-to-potassium [31], and calorie-to-weight [32]—as simple, measurable proxies for the degree of food processing and previously identified as strong predictors of food quality [33]. It also incorporates alcohol and adjusts for food categories high in bioactive components that support a healthy gut microbiome, including polyphenols, bioactive fats, fermentable fibers, and fermentation products [34]. The primary aim of this study was to compare NCS, a nutrient ratio-based, web-enabled food quality score with other leading NPSs in relation to weight and blood pressure, key prognostic factors in cardiometabolic disease and mortality. We hypothesized that NCS would demonstrate associations with these outcomes that are at least comparable in magnitude with those observed for other commonly used NPSs. Secondary aims included evaluating associations with cardiometabolic biomarkers and identifying food categories contributing most to each score. Methods Data sources and analytic population The NHANES is a repeated, cross-sectional analysis that has been conducted continuously since 1999 and data are released in 2-y cycles, with exceptions for recent cycles because of COVID-19 disruptions in survey data collection [35]. The survey includes interview and physical measurement components and is conducted in a nationally representative sample of ∼5000 persons each year from persons located in 15 counties that are randomly selected each year. For the primary analysis to compare NCS and other NPSs, data from the continuous survey data releases 2015–2016 were used for this analysis because complete previously published data were available for all evaluated NPS scores for this NHANES data cycle [36]. In the 2015–2016 data cycle, 9971 individuals participated. The following exclusions were applied in this order to obtain the analytic subsample used (n excluded): age <20 y (n = 4252), missing education (n = 5), missing poverty-to-income ratio (PIR) (n = 0), missing day 1 dietary intake data (n = 699), physiologically unsustainable low dietary intake of <500 kcal (n = 42), physiologically unsustainable high dietary intake >5000 kcal (n = 57), missing smoking information (n = 20), missing waist circumference (n = 203), missing BMI (n = 9), missing blood pressure (n = 51), or missing physical activity (n = 35). NHANES does not have missing data for age, gender, or race/ethnicity. The analytic sample for primary analyses used included 4598 adult individuals. A secondary analysis was undertaken to evaluate associations of the NCS score with other cardiometabolic biomarkers in NHANES cycles 2005–2018. Additional exclusions to the 4598 individuals in the analytic sample were missing LDL, HDL, or total cholesterol, plasma glucose, or blood triglycerides (n = 2647). The analytic sample for secondary analyses of cardiometabolic biomarkers included 1951 adult individuals. Dietary intake assessment Dietary intake is assessed for NHANES using 2 multiple pass 24-h recalls. The first 24-h recall is conducted in-person during the visit to the Mobile Examination Center (MEC). The second 24-h recall is conducted via phone 3 to 10 d after the MEC visit. All NHANES participants are eligible for the 24-h recall interviews. The interviews are conducted using a midnight-to-midnight time frame for the 24-h period before the interview. The USDA Food Survey Research Group conducts the dietary data collection methodology, maintenance of databases used to code and process data, and data review and processing. The dietary data are released in 2 files: individual foods and total nutrients. The individual foods file lists each food reported by the participant, along with details of the consumption such as a USDA Food and Nutrient Database for Dietary Studies (FNDDS) code (food code), eating occasion, and amount of food/beverage consumed in grams. The total nutrient intake file is a daily aggregate of nutrients from the reported foods consumed for that 24-h recall. NCS and nutrient ratio scores NCS is an online NPS based on nutrient ratios as proxies for the degree of food processing—carbohydrate-to-fiber, saturated fat-to-unsaturated fat, sodium-to-potassium, and calorie-to-weight—along with additional nutrients (protein, alcohol, vitamin D, iron, calcium) available on United States Nutrition Facts labels. All scores used in this paper are available through the online portal (https://gutbites.org/carb-fiber-ratio-calculator). Positive adjustments are made for food categories like fruits, vegetables, nuts, seeds, and whole grains to account for microbiome-supportive and bioactive factors not listed on food labels: polyphenols, prebiotic fibers, bioactive fats, and fermentation products. In addition, negative adjustments are made for alcohol, soft drinks, processed meats, and processed potatoes to reflect additives and food components linked to negative health outcomes. Three versions of NCS were evaluated in the analysis: NCS proper, NCS without alcohol included in calculations to compare with other algorithms that do not include alcohol, and NCS without bioactive adjustments. Nutrient ratios: carbohydrate-to-fiber (fiber ratio), saturated fat-to-unsaturated fat (fat ratio), sodium-to-potassium (salt ratio), and calorie-to-weight (energy ratio) were also analyzed separately to evaluate their independent contributions to observed associations. Nutrient ratio scores were mapped to a quantized scale of 1 to 3 in 0.125 increments. Other NPSs To contextualize the performance of the NCS, we compared it with several leading and widely used NPSs that represent distinct approaches to assessing food quality. NS is a scoring algorithm that balances positively weighted nutrients (e.g., fruits, vegetables, and fiber) against negatively weighted ones (e.g., saturated fat, sugars, and sodium), resulting in a color-coded rating from A to E [24]. HS Rating separates foods into 3 groups—dairy products, nondairy beverages, and all other foods—and assigns points based on levels of protein, fats, saturated fats, energy, carbohydrates, sugar, and sodium, yielding in a 5-star scale with half-star increments [25]. The NC ranks foods by processing level, distinguishing between minimally or unprocessed foods, processed ingredients, processed foods, and ultraprocessed foods, resulting in a 4-level scale [26]. FC assesses foods across 9 domains—nutrient ratios, vitamins, minerals, food-based ingredients, additives, processing, specific lipids, fiber and protein, phytochemicals—producing a 100-point score [23,27]. Score calculation and weighting NCS values were retrieved using the NCS algorithm available online at Gutbites.org [12]. Each of the 4 other NPSs was rescaled to enable direct comparison with NCS. NS (A–E = 100–20), HS Rating (0.5–5 = 10–100), NC (1–4 = 80–20), and FC 2 (1–100 = 1–100) scores were derived from previously reported values for foods within the 2015–2016 What We Eat in America (WWEIA)/FNDDS food categories [23,27,37]. Alcoholic beverages were assigned null values in systems that did not evaluate them: NS, HS, NC, and FC. All scores and nutrient ratios were evaluated in 3 ways: weighted by calorie (kcal), weighted by weight (grams), and unweighted. Unweighted scores were calculated for each individual as the mean score of each food consumed in the 24-h period divided by the total number of food items. Scores weighted by energy and weight were calculated as the sum of the individual food items’ scores multiplied by the kcal or grams provided by the food item and then divided by the total kcal or grams for the 24-h period. cores were analyzed as continuous variables on a 10-unit scale. Food categories and contributions to scores To evaluate which foods were contributing to most of the variation in the NPS and ratio scores, a compositional approach was applied. Individual foods reported were assigned to food group classifications based on the WWEIA food categories [38]. This analysis was performed for all NPSs. Total intake of food categories was calculated for each person based on percentage energy (kcals) provided by all the foods in that category for the 24-h dietary intake recall period. Center-log transformed ratios of percentage energy contributed by food categories to overall energy were calculated [39] using the R compositions package [40] to include in regression models. Outcomes Anthropometry and blood pressure were evaluated as continuous outcomes and as binary outcomes based on established cutoffs. Anthropometric characteristics were measured during the in-person visit by trained technicians. Blood pressure was measured 1 to 4 times by trained technicians, and the mean of available measurements for each participant was used in analyses. Details about measurement and analysis are available from NHANES established procedures. The mean of the available measurements for each person was used for analysis. Obesity was defined as BMI ≥30 kg/m2 [41] and abdominal obesity was defined as waist circumference >88 cm for females and >102 cm for males [42]. High blood pressure was defined as the presence of systolic blood pressure 130 mm Hg or diastolic blood pressure 85 mm Hg [42,43]. Blood-based cardiometabolic risk factors (analysis conducted for NCS only to maintain consistency across NHANES cycles where comparable data for other NPSs were not available) evaluated included LDL, HDL, total cholesterol, and triglycerides, and plasma glucose. These risk factors were evaluated as continuous measures and in relation to established cutoffs for high or low values [42,43]. Additionally, metabolic syndrome was assessed and was classified as the presence of ≥3 of: abdominal obesity, high blood pressure, high LDL, low HDL, or elevated serum glucose [42]. Covariates Personal demographic characteristics were included to describe the analytic group and as adjustment variables in multivariable analyses. Characteristics included were age (in categories for description and in years for multivariable analyses), gender (female and male), race or ethnicity (non-Hispanic White, non-Hispanic Black, non-Hispanic other, and Hispanic), education (65, female gender, Hispanic, non-Hispanic other, 4-y college graduates, PIR of ≥3, and never smokers. Categories more represented in the lowest score tertiles included younger age, non-Hispanic White, non-Hispanic Black, lower education level, current smoker, and highest physical activity level (165+ METS). TABLE 1Study population characteristics by tertiles of energy-weighted NCS score in NHANES adults, 2015–2016TABLE 1Personal characteristicsTertile 1 (<37.08), n = 1533Tertile 2 (37.08–48.31), n = 1533Tertile 3 (>48.31), n = 1532Age (y) 20–34458 (33.3%)370 (25.4%)360 (25.4%) 35–49415 (26.8%)392 (24.8%)339 (22.6%) 50–64380 (25.2%)434 (28.6%)403 (27.4%) 65+280 (14.8%)337 (21.2%)430 (24.7%)Gender Male764 (53.3%)752 (47.1%)693 (43.0%) Female769 (46.7%)781 (52.9%)839 (57.0%)Race and non-Hispanic ethnicity Non-Hispanic White589 (67.7%)559 (68.0%)420 (58.2%) Non-Hispanic Black415 (13.8%)316 (10.4%)220 (7.6%) Non-Hispanic Other116 (4.9%)199 (8.2%)333 (16.2%) Hispanic413 (13.6%)459 (13.4%)559 (18.0%)Education