Defining the impact of dietary macronutrient balance on PCOS traits

Polycystic ovary syndrome (PCOS) is a common endocrine disorder affecting 8–13% of reproductive-age women and characterized by reproductive, metabolic, and endocrine abnormalities, including hyperandrogenism, anovulation, obesity, insulin resistance, and dyslipidemia. Lifestyle interventions (diet, exercise, behavioral strategies) are recommended for all women with PCOS, yet the optimal dietary composition remains unclear due to limited, heterogeneous, and short-term human studies with poor compliance. Prior research suggests weight loss improves PCOS features regardless of diet composition, but it is unknown how specific macronutrient balances influence PCOS traits. The authors hypothesized that dietary macronutrient balance (protein, carbohydrate, fat) would significantly affect the development of PCOS features and that a systematic, controlled evaluation in an animal model using the Geometric Framework for nutrition could identify optimal macronutrient combinations to ameliorate PCOS traits.

Existing evidence indicates high-fat diets aggravate obesity and PCOS traits; studies of low/high protein and low/high carbohydrate diets show variable effects. Low-carbohydrate or low glycemic load/high-protein hypocaloric diets may reduce insulin resistance, cholesterol, and weight, but overall weight loss rather than macronutrient composition appears to drive many benefits. Modest weight loss (5–15%) improves key PCOS features. However, attrition and compliance issues in human trials limit firm conclusions, and a recent systematic review found insufficient evidence to define optimal macronutrient composition for PCOS management. Animal studies have shown macronutrient balance affects reproductive function and cardiometabolic health, motivating a controlled, systematic assessment of macronutrient effects on PCOS traits.

Design: Controlled animal study combining a hyperandrogenic PCOS mouse model with the Geometric Framework (GF) to evaluate individual and interactive effects of protein (P), carbohydrate (C), and fat (F) on reproductive and metabolic PCOS traits. Mice: Female C57BL/6J mice. PCOS induction: Peripubertal androgenization by subcutaneous SILASTIC implant containing ~10 mg dihydrotestosterone (DHT) at 3 weeks of age; control mice received empty implants. DHT implants provide steady-state release for at least 6 months and were verified at collection. Housing: Standard conditions, 12-h light/dark, ad libitum food and water. Diets: At 7 weeks, mice switched from chow to one of 10 experimental diets varying in P, C, F (Specialty Feeds) for 10 weeks, ad libitum. Diet selection based on prior work to optimize response surface modeling coverage. Diet 3 was discontinued after some mice lost ≥20% body weight or failed to thrive; thus n for diet 3 was reduced (3 control, 3 PCOS completed). Sample sizes: Across diets, 187 mice were studied (~93 control, ~94 PCOS for major outcomes). Intake measurement: Weekly food intake measured with custom two-chamber inserts to capture spillage; energy intake calculated. Outcomes and assays: - Reproductive: Ovarian histology and corpora lutea (CL) counts (every third section) after 10 weeks; estrous cycles monitored via daily vaginal cytology for 11 days. - Hormones: Serum DHT and testosterone by LC-MS/MS (LOQ 0.05 ng/mL for DHT; 0.01 ng/mL for testosterone). LH and FSH by MILLIPLEX bead assays in diestrus/pseudo-diestrus. - Body composition: DEXA (PIXImus2) for fat and lean mass. - Adipose tissue: Parametrial fat pad weights; histology (H&E) and adipocyte size quantification (ImageJ). - Adiponectin: Serum total adiponectin by ELISA (R&D Systems). - Lipids: Serum total cholesterol and triglycerides enzymatically (Wako kits). - Glycemia: 6-h fasting glucose and intraperitoneal glucose tolerance test (2 g/kg) with glucose measured at 0, 15, 30, 60, 90 min; AUC computed. - Metabolic phenotyping: Indirect calorimetry in metabolic cages for 72 h (after 2-day acclimation) on chow: food intake, locomotor activity (beam breaks), energy expenditure (absolute and body weight–corrected), VO2, VCO2, RER (17 control, 16 PCOS). - Diet preference: Same mice given ad libitum access for 7 days (2-day acclimation, 5-day measurement) to 3 diets: high P (60% P, 20% C, 20% F), high C (75% C, 5% P, 20% F), high F (75% F, 5% P, 20% C); macronutrient intake calculated. Statistical analysis: GF response surfaces and modeling using generalized additive models (GAMs; thin-plate and cubic regression splines) to capture non-linear effects of P, C, F intakes on outcomes; mixture models (Scheffé polynomials) for intake vs composition. Counts modeled with negative binomial distribution; transformations applied as needed with results back-transformed. Surface subtraction used to compute pointwise differences between control and PCOS response surfaces with confidence limits to infer significance. Group comparisons by t-test or Mann–Whitney where appropriate; proportions by Fisher’s exact; significance at P<0.05.

• Validation of hyperandrogenism: Serum DHT elevated in PCOS vs control (1.34 ± 0.17 vs 0.27 ± 0.04 ng/mL; P<0.001). Testosterone comparable (0.05 ± 0.02 vs 0.07 ± 0.01 ng/mL). - Ovulatory function: PCOS mice had markedly fewer CL than controls across diets (0.5 ± 0.2 vs 4.5 ± 0.6 CL/ovary; P<0.001). Nevertheless, ovulation was restored in a subset of PCOS mice (8/35) on a specific macronutrient intake niche: C 20–30 kJ/day and F 15–25 kJ/day at median P ~8 kJ/day. Estrous cyclicity was present in a subset of PCOS mice (15/94) within similar intake ranges, confirming functional restoration. Response surfaces showed C×F intake as the main drivers of ovulation, with control and PCOS responses converging when C>20 and F>25 kJ/day. - Body weight and composition: PCOS mice were heavier regardless of diet (23.5 ± 0.3 g vs 21.7 ± 0.2 g; P<0.001) with increased body fat (P=0.0005) and lean mass (P<0.001). All macronutrients drove weight gain in both groups, but PCOS mice gained substantially more weight at lower macronutrient intakes, particularly when C and F were each <25 kJ/day (significant response difference). - Intake and metabolism: Food and energy intake patterns were similar between groups across compositions; both decreased as dietary protein increased. No difference in macronutrient preference in choice tests. Indirect calorimetry on chow showed no significant differences in day/night food intake, energy expenditure (absolute and BW-corrected), VO2, VCO2, RER, or night locomotor activity; a small reduction in day locomotor activity in PCOS (P=0.037). - Adipose tissue and adipokines: PCOS mice had larger adipocytes (1729 ± 54.3 vs 1573 ± 50.6 µm²; P=0.04) and lower serum adiponectin (11221 ± 453.8 vs 16806 ± 758.7 ng/mL; P<0.001). In controls, adiponectin increased with higher C (P=0.002) and was influenced by P (P=0.032); in PCOS, macronutrient intake did not significantly modify adiponectin, indicating dominance of PCOS pathophysiology. - Lipids: Trend to higher cholesterol in PCOS (61.5 ± 2.7 vs 53.1 ± 2.2 mg/dL; P=0.06). Control mice maintained low cholesterol across macronutrient intakes; in PCOS, fat intake drove increases in cholesterol (P<0.001), with lower F (<20 kJ/day) bringing levels closer to controls. Triglycerides were similar between groups (overall 40–65 mg/dL) and driven mainly by P (and P+C) intake; responses to diet were overall comparable. - Glycemia: Fasting glucose elevated in PCOS vs controls (9.2 ± 0.2 vs 8.8 ± 0.2 mmol/L; P<0.05). Controls maintained stable fasting glucose across diets; PCOS showed higher and more variable fasting glucose, particularly with higher P and F intakes. Comparable fasting glucose to controls occurred only when F <12 kJ/day. GTT responses (AUC) did not differ significantly between groups across most diets. - Overall pattern: Reproductive traits (ovulation/cyclicity) were more sensitive to dietary macronutrient balance and could be improved by a low-protein, medium-carbohydrate and fat intake, whereas metabolic traits (adiposity, adiponectin, cholesterol, fasting glucose) were minimally ameliorated by diet and largely driven by hyperandrogenic PCOS pathophysiology.

This study demonstrates that macronutrient balance shapes the development of PCOS traits in a hyperandrogenic mouse model. A specific intake niche—low protein with medium carbohydrate and fat—restored ovulatory function and estrous cyclicity in a subset of PCOS mice, indicating reproductive traits are particularly diet-sensitive. In contrast, metabolic traits (body weight sensitivity at lower intakes, reduced adiponectin, dyslipidemia, elevated fasting glucose) were less modifiable by macronutrient changes and appeared predominantly determined by the underlying androgen-excess pathology. The optimal macronutrient ratio identified (~14% protein, 47% carbohydrate, 39% fat) aligns with ranges of a Mediterranean-style diet, which in humans is associated with improved lipid profiles, reduced adiposity, and lower diabetes risk. These findings support dietary manipulation as an adjunct strategy to address PCOS-related infertility, while reinforcing current recommendations emphasizing weight management and exercise for broader metabolic benefit. The work underscores the need to translate macronutrient-based strategies to human studies to determine whether benefits derive from macronutrient balance per se or associated food choices and patterns.

Using a controlled PCOS mouse model and the Geometric Framework, the study establishes that dietary macronutrient balance can selectively ameliorate reproductive PCOS traits, with a low-protein, medium-carbohydrate and fat diet restoring ovulatory function in a subset of hyperandrogenic mice. Conversely, metabolic traits (adiposity, adiponectin, cholesterol, fasting glucose) showed limited improvement with macronutrient manipulation, indicating a dominant influence of PCOS pathophysiology. These results provide a mechanistic basis for developing evidence-based dietary interventions targeting reproductive outcomes in PCOS and support current clinical emphasis on weight loss and physical activity for metabolic health. Future research should include long-term, high-quality human RCTs comparing macronutrient compositions, and investigations to disentangle the effects of macronutrient balance from specific dietary patterns (e.g., Mediterranean diet) on PCOS outcomes.

Limitations were not explicitly detailed, but several are implied: (1) Translation from a hyperandrogenic mouse model to human PCOS may be limited; human compliance and long-term dietary adherence were not addressed experimentally. (2) One experimental diet (Diet 3) was discontinued due to weight loss/failure to thrive, resulting in reduced sample size for that arm. (3) While the macronutrient balance restored ovulation in a subset, most metabolic traits remained refractory to dietary manipulation, suggesting limited generalizability of dietary effects across PCOS phenotypes. (4) The study examined macronutrient balance rather than specific food matrices or micronutrients, so effects attributable to dietary patterns vs macronutrient ratios cannot be separated.