Comparison of the acute metabolic effect of different infant formulas and human milk in healthy adults: a randomized trial

The study addresses whether the form (intact vs partially hydrolyzed whey) and quantity of protein in infant formula acutely influence postprandial insulin and glucose responses, relative to human milk (HM). Although breastfeeding is optimal, many infants receive formula; thus aligning formula’s metabolic responses with HM is important. Milk elicits disproportionally high insulinemia relative to its low glycemic index, influenced by protein type and amino acid profile. Prior reports suggest whey may stimulate insulin more than casein, and hydrolysates may increase insulin secretion vs intact proteins, though results are mixed and often not in milk-like matrices. The authors hypothesized that insulin response would be similar between intact protein formula (INTACT; whey/casein 70/30) and partially hydrolyzed whey formula (PHw; 100% whey), and compared both to a higher-protein partially hydrolyzed whey formula (HPPHw) and an HM subgroup in healthy adults to inform formula design for infants.

Background literature indicates: (1) Cow’s milk and both infant formula and HM can produce higher insulinemic responses than expected from glycemic index alone; insulin responses depend on lactose, proteins, and lipids. (2) Protein type matters: whey can yield greater postprandial insulin than casein; certain amino acids (e.g., branched-chain AAs) are insulinotropic. (3) Protein form may matter: some studies show greater insulin secretion with hydrolysates vs intact proteins, though others do not, often with protein doses and matrices not reflective of infant milk (typically low protein, high fat, high lactose). (4) Prior infant studies suggesting higher insulin after formula vs HM have methodological limitations. (5) Evidence gaps persist regarding intact vs partially hydrolyzed formula in milk matrices, especially relative to HM. The early protein hypothesis posits higher early protein intakes increase insulinogenic amino acids and anabolic hormones, potentially elevating later adiposity risk; trials have linked higher infant protein intakes to greater weight/adiposity.

Design: Randomized, double-blinded, monocenter, cross-over trial with four study products; primary objective to demonstrate similar insulin response (Cmax) between PHw and INTACT. Secondary objectives included comparing postprandial insulin and glucose among all products and HM, and exploring other metabolic markers. Participants: Healthy adults (20–50 y), BMI 19–25 kg/m², recruited at Nestlé Research Center, Lausanne (June–October 2012). Key exclusions: metabolic diseases, dyslipidemia, anemia, recent surgery, weight loss, high-intensity exercise, allergies (e.g., lactose intolerance), chronic meds (except OCP), high alcohol, drugs, smoking >5/day, pregnancy/lactation, special diets, recent blood donation, noncompliance risk, and recent trial participation. Ethics: Approved by Commission cantonale d'éthique de la recherche sur l'être humain (protocol 398/11); written informed consent obtained; registered NCT04332510. Interventions: Each participant consumed 600 ml within 10 min of: (1) INTACT formula (1.87 g protein/100 kcal; whey/casein 70/30; 63 kcal/100 ml); (2) PHw formula (1.96 g/100 kcal; 100% whey; 63 kcal/100 ml); (3) HPPHw formula (2.79 g/100 kcal; 100% whey; 73 kcal/100 ml); and a subset also received HM from a milk bank (reference). Macronutrients were similar for fat and carbohydrate across products; HPPHw had 65% more protein than the other formulas. Nine participants received all four products (including HM); 24 received the three formulas only. Procedures: Cross-over with washouts: 2 weeks ±2 days (men) and aligned follicular phase with 4 weeks ±2 days (women). Standardized diet guidance 3 days pre-test; fixed meals provided the day prior; overnight fast after 21:00. On test days, baseline measures included two arterialized venous blood samples at -30 and 0 min via retrograde catheter with heated-hand method (55 °C). After product ingestion, arterialized blood was sampled at 15, 30, 45, 60, 90, 120, 150, 180 min. Participants remained seated; adverse events and compliance recorded. Biomarkers and assays: Plasma glucose (Siemens Dimension), insulin (ELISA IBL RE53171), C-peptide (ELISA EZHCP-20K), active GLP-1 (ELISA EGLP-35K), glucagon (ELISA EZGLU-30K), triglycerides (Siemens Dimension), free fatty acids (Wako NEFA-HR2), and plasma amino acids. Outcomes: Primary—insulin Cmax. Secondary—postprandial insulin and glucose responses among all products and HM; free fatty acids, triglycerides, C-peptide, glucagon, active GLP-1, and essential amino acids (EAA). Curve characteristics: AUC, iAUC (incremental above baseline), Cmax; negative AUC (nAUC) for FFAs. Statistics: Mixed-effects models with fixed effects for baseline, time (categorical), sex, treatment; subject as random effect; Box-Cox transformations. Time-trend differences tested via likelihood-ratio tests with permutation-based null distributions. Curve characteristics estimated from back-transformed predicted means; statistics via permutation tests. EMA bioequivalence guidance used to interpret similarity (test/reference ratios within 90% CI of 80–125%); HM used as reference for percentage differences. Analyses on per-protocol completers. R 3.5.2 with lme4, MASS, stats, logspline.

Participants: 35 randomized; 29 per-protocol analyzed for formulas; 8 for HM comparisons. Mean age 31 y; BMI ~22.3 kg/m² (Table 2). Two nausea-related AEs (one with PHw leading to withdrawal; one with HM partial consumption). Insulin: Mean iAUC (µIU/ml): HPPHw 18.66; HM 13.10; PHw 11.92; INTACT 11.79. Mean Cmax (µIU/ml): HPPHw 57.24; HM 43.36; PHw 40.29; INTACT 40.88. PHw vs INTACT showed bioequivalence: iAUC difference 1.5% (90% CI -9.7 to 12.67; p=0.95); Cmax difference -2.1% (90% CI -12.2 to 7.9; p=0.81). HPPHw higher than INTACT and PHw: iAUC differences +51% vs INTACT (90% CI 32–70; p<0.001) and +49% vs PHw (90% CI 30–68; p<0.001). Time trend for HPPHw differed vs INTACT, PHw, HM (LR p<0.001, <0.001, 0.050). Glucose: Mean iAUC (mmol/l): HPPHw 0.45; HM 0.40; PHw 0.24; INTACT 0.24. Mean Cmax (mmol/l): HPPHw 6.25; HM 6.14; PHw 5.86; INTACT 5.83. PHw vs INTACT bioequivalent: iAUC difference -7.5% (90% CI -22.9 to 7.9; p=0.55); Cmax difference 0.16% (90% CI -1.7 to 2.05; p=0.84). HPPHw higher than INTACT and PHw: iAUC +50% vs INTACT (90% CI 23–76; p=0.003) and +57% vs PHw (90% CI 32–83; p<0.001). HM vs INTACT/PHw differences were similar in magnitude but not statistically significant. C-peptide: HPPHw higher than INTACT and PHw: iAUC +55% vs INTACT (90% CI 33–78; p<0.001) and +54% vs PHw (90% CI 33–76; p<0.001). No significant time-trend differences for INTACT or PHw vs HM. Triglycerides: Time trend for HPPHw differed vs others (LR p=0.009–0.012). iAUC higher with HPPHw: +56% vs INTACT (90% CI 27–85; p=0.019); +59% vs PHw (90% CI 37–82; p=0.005). INTACT and PHw time trends did not differ vs HM. Free fatty acids: No significant differences in time trends among products. Glucagon: PHw vs INTACT differed by Cmax (+15%; 90% CI 6–23; p=0.009). No significant iAUC differences between PHw and HPPHw. Essential amino acids: HPPHw higher iAUC vs PHw (+74%; 90% CI 39–108; p<0.003) and vs HM (+65%; 90% CI 29–102; p=0.012). INTACT and PHw did not differ vs HM in time trends. GLP-1: No statistically significant differences in time trends among products. Overall: INTACT and PHw produced similar postprandial insulin and glucose responses that were close to HM; higher protein content (HPPHw) increased insulin, glucose, C-peptide, triglycerides, and EAA responses.

The primary hypothesis was supported: intact protein formula (INTACT) and partially hydrolyzed whey formula (PHw) elicited bioequivalent insulin and glucose responses in healthy adults, indicating that protein form (intact vs partially hydrolyzed whey) at typical infant-formula protein levels does not materially alter acute postprandial insulinemia or glycemia when delivered in a milk matrix. These responses were also close to those observed with human milk (HM), aligning with prior work showing similar glycemic/insulinemic responses to HM and standard formula. Conversely, increasing protein quantity (HPPHw) substantially elevated insulin and glucose responses and associated markers (C-peptide, triglycerides, essential amino acids), consistent with the early protein hypothesis that higher protein loads increase insulinogenic amino acids and anabolic hormone responses. Although tested in adults, these findings suggest that higher-protein infant formulas may drive stronger postprandial anabolic signaling than standard-protein formulas, which could have implications for growth trajectories and long-term metabolic outcomes. The largely unchanged GLP-1 and free fatty acid responses indicate protein quantity primarily influenced insulinotropic pathways rather than incretin or lipolysis suppression within the tested window. Small differences in glucagon (PHw vs INTACT by Cmax) were observed but without consistent iAUC effects. Together, the results emphasize that at conventional protein levels, the choice between intact and partially hydrolyzed whey may have minimal impact on acute insulin/glucose responses, whereas higher protein content meaningfully augments these responses. These data inform infant formula design aiming to approximate HM’s metabolic profile and support careful consideration of protein quantity. Confirmation in infants is needed due to physiological and dosing differences.

In healthy adults, formulas with intact protein (whey/casein 70/30) and partially hydrolyzed 100% whey at standard protein levels produced similar, HM-like postprandial insulin and glucose responses. A higher-protein partially hydrolyzed whey formula elicited significantly greater insulin, glucose, C-peptide, triglyceride, and essential amino acid responses. While adult data may not directly generalize to infants, the direction of effects suggests protein quantity, rather than protein form, drives larger metabolic responses. Future research should confirm these findings in infants, include measures of gastric emptying, and explore long-term implications for growth and metabolic health, including contexts where higher anabolic responses may be desirable (e.g., catch-up growth).

- Conducted in adults; infant responses may differ, limiting generalizability. - Human milk subgroup was small (n=8 analyzed), reducing power for HM comparisons. - Large test volume (600 ml, ~350–438 kcal) exceeds typical single-feed volumes for young infants; chosen to represent a small adult meal. - No direct measurement of gastric emptying, which could elucidate kinetics. - Some missing data and protocol deviations led to exclusions (29 per-protocol analyzed). - Adverse events (nausea) occurred in two participants (one with PHw, one with HM) affecting consumption/completion. - Study compared acute responses only; no long-term outcomes.