The effects of varying protein and energy intakes on the growth and body composition of very low birth weight infants

The researchers compared three different formula diets in very low birth weight infants:

**Group A (control):** Standard preterm formula providing 129 kcal/kg/day of energy and 3.7 g/kg/day of protein.

**Group B (high protein, high energy):** Formula providing 150 kcal/kg/day and 4.2 g/kg/day of protein.

**Group C (very high protein, high energy):** Formula providing 150 kcal/kg/day and 4.7 g/kg/day of protein.

A separate group of breastfed infants (n=6) was included as a descriptive reference only — they were not randomised and were not compared statistically to the formula groups.

The primary outcome was **fat-free mass (FFM) accretion** — essentially, how much lean tissue (muscle, organs, bone, water) the infants gained. Secondary outcomes included weight gain, length gain, head circumference gain, body mass index (BMI), fat mass, and blood markers for safety (glucose, urea, ammonia, cholesterol, triglycerides, pH, base excess).

**Sample size:** 38 preterm infants total. 32 were randomised into three formula groups (Group A: n=8, Group B: n=12, Group C: n=12). 6 breastfed infants served as a non-randomised reference group.

**Population:** Very low birth weight (VLBW) infants weighing under 1500 g at birth, born at 32 weeks gestation or fewer, with weights appropriate for their gestational age (not growth-restricted). All were admitted to the neonatology ward of Hospital Clinic in Barcelona, Spain.

**Inclusion criteria:** Infants had to have recovered their birth weight, be gaining weight, tolerate full enteral (oral) feeds, and have stopped mechanical ventilation and parenteral nutrition at least 5 days before the study began.

**Exclusion criteria:** Intrauterine growth restriction, chromosomal abnormalities, malformations, chronic diseases, or need for oxygen treatment.

**Baseline characteristics:** Groups were similar in sex distribution, rates of caesarean section, Apgar scores, and complications of prematurity (respiratory distress syndrome, sepsis, intraventricular haemorrhage, patent ductus arteriosus). However, Group C had more males (9 of 12) and more cases of respiratory distress syndrome (8 of 12) compared to other groups.

**Anthropometry:** Weight (electronic scales accurate to 1 g), length, and head circumference (non-stretch tape) were measured weekly by the same investigator for 4 weeks. Z-scores were calculated using neonatal growth curves from Catalonia, Spain (based on >200,000 newborns), adjusted for sex and post-menstrual age.

**Body composition:** Total body electrical impedance analysis (BIA) was performed by a single investigator using a Bioscan Spectrum device. An 800-μA, 50-kHz alternating current was passed through electrodes placed on the wrist and ankle in a standardised body position (prone, with specific hip, knee, and ankle angles). Total body water was calculated using the Tang et al. equation: Total body water = (0.016 + 0.674 × weight - 0.038 × weight² + 3.84 × foot length²) / resistance. Fat-free mass (FFM) was derived by dividing total body water by the water percentage of FFM (based on Fomon and Ziegler reference data for infants). Fat mass (FM) was then calculated as body weight minus FFM.

**Blood markers:** Serum glucose, total protein, ammonia, pH, base excess, urea, cholesterol, and triglycerides were measured once during the third week of the study to detect nutrition-related adverse effects.

**Duration:** The intervention lasted approximately 28 days (4 weeks), with measurements taken at baseline (Day 0) and at the end (approximately Day 28).

**Study design:** This was a longitudinal, interventional, randomised clinical trial (RCT). Infants who were not breastfed were randomly assigned to one of three formula groups (A, B, or C). The breastfed group was not randomised and served only as a descriptive reference.

**Randomisation:** Randomisation was performed by nurses who prepared the formula in the morning, using sealed envelopes, in blocks of 6. The nurses who prepared the formula were the only individuals who knew the group assignments, and they did not provide care for the infants. This is a form of allocation concealment, which helps prevent selection bias.

**Blinding:** The paper states that "during the duration of the randomised trial, the blinding remained intact." However, it is unclear whether the investigators who measured outcomes (weight, length, head circumference, BIA) were blinded to group assignment. The nurses who prepared the formula knew the assignments, but they did not care for the infants. The clinicians caring for the infants and the parents were likely blinded, but this is not explicitly stated. This is a **moderate weakness** — if outcome assessors were not blinded, there is a risk of measurement bias, especially for subjective measures like BIA electrode placement.

**Duration:** The intervention lasted approximately 28 days. This is a reasonable duration for detecting changes in growth and body composition in rapidly growing preterm infants. However, it is too short to assess long-term outcomes like neurodevelopment or later metabolic health.

**Statistical approach:** The primary analysis used UNIANOVA (analysis of covariance) with group as the factor and the baseline value of the outcome plus the duration of intervention (in days) as covariates. This adjusts for any baseline differences and for slight variations in study duration. If the overall group effect was p < 0.1, pairwise comparisons were performed. Sample size was calculated based on preliminary data from the first 5 infants in Group A and the first 5 in Groups B/C combined, aiming for 80% power at α = 0.05. This gave a target of 12 per group (10 plus 2 for dropouts). However, Group A only reached n=8, which means the study was **underpowered** for comparisons involving Group A.

**What this design can and cannot prove:**

**Can prove:** That the different formula compositions caused differences in growth and body composition over 28 days, assuming randomisation successfully balanced confounders and blinding was maintained.

**Cannot prove:** Long-term effects on neurodevelopment, metabolic health, or later obesity risk. Cannot prove that these results generalise to infants with intrauterine growth restriction, chronic lung disease, or other complications. Cannot prove that the high-protein, high-energy diet is safe beyond 28 days. Cannot prove that the breastfed reference group is comparable (since they were not randomised).

**Major methodological weaknesses:**

1. **Small sample size:** Group A had only 8 infants (not the planned 12), reducing statistical power.

2. **Unclear blinding of outcome assessors:** The paper does not explicitly state that the investigator measuring anthropometry and performing BIA was blinded to group assignment.

3. **Breastfed group not randomised:** The breastfed group (n=6) was included as a "descriptive reference" but cannot be compared statistically. This is a missed opportunity.

4. **BIA validation in preterm infants:** BIA equations for total body water were developed in different populations, and the water percentage of FFM in preterm infants is variable. This introduces measurement error.

5. **Short duration:** 28 days is insufficient to assess long-term safety or efficacy.

6. **Industry funding not disclosed:** The paper does not state whether the formula manufacturers (Nestle, Abbott, Nutricia) provided funding or products. This is a potential conflict of interest.

**Primary outcome — Fat-free mass (FFM) accretion:**

Groups B and C (high energy, high protein) showed greater increases in fat-free mass compared to Group A (standard formula).

The paper reports that FFM accretion was "higher" in Groups B and C, but the exact numerical values for FFM gain (in g/kg/day) are not presented in the abstract or the truncated results section. The preliminary data used for sample size calculation showed FFM accretion of 15.09 ± 2.14 g/kg/day in Group A and 19.85 ± 4.15 g/kg/day in Groups B/C combined — a difference of approximately 4.8 g/kg/day (about 32% higher in the supplemented groups).

**Secondary outcomes — Weight gain:**

Groups B and C displayed greater weight gains than Group A. Again, exact numerical values are not provided in the available text, but the trend was consistent across the high-energy groups.

**Secondary outcomes — Length and head circumference:**

No significant differences were reported between groups for length gain or head circumference gain. This suggests that the extra energy and protein primarily affected weight and body composition, not linear growth.

**Secondary outcomes — Fat mass:**

The paper does not report significant differences in fat mass between groups, suggesting that the additional weight gain was primarily lean tissue, not fat.

**Safety outcomes — Blood markers:**

The only statistically significant difference in blood markers was higher urea levels in Groups B and C compared to Group A (p = 0.032). This is expected with higher protein intake and indicates that the extra protein was being metabolised.

No significant differences were found in glucose, total protein, ammonia, pH, base excess, cholesterol, or triglycerides. This suggests the high-protein, high-energy diets were well tolerated over 28 days.

**Tolerance:**

Energy intake up to 150 kcal/kg/day and protein intake up to 4.7 g/kg/day were "well tolerated by all subjects from both the clinical and analytical points of view."

Based on the preliminary data (which the authors used for sample size calculation), the high-energy, high-protein diets increased fat-free mass accretion by approximately **4.8 g/kg/day** — about a **32% increase** compared to the standard formula. To put this in perspective: for a 1.5 kg infant, this would mean an extra ~7.2 g of lean tissue gained per day, or about 200 g over 28 days. This is a substantial effect in a population where every gram of lean mass matters for neurodevelopment and long-term health.

The elevated urea levels in the high-protein groups indicate that the extra protein was being broken down and excreted, which is a normal metabolic response. The clinical significance of this is unclear — mildly elevated urea is generally not harmful, but it does suggest that not all of the extra protein was used for tissue building.

**Acknowledged by authors:**

The small sample size, particularly in Group A (n=8), which was below the calculated target of 12.

The use of BIA, which is an indirect method for measuring body composition and relies on assumptions about hydration status that may not hold in preterm infants.

The short duration of the study (28 days).

**Not acknowledged but critical:**

**Unclear blinding:** The paper does not explicitly state that outcome assessors were blinded. If the investigator measuring weight, length, and BIA knew which group an infant was in, this could bias results.

**No intention-to-treat analysis:** The paper does not mention whether any infants dropped out or were excluded after randomisation, and if so, how this was handled.

**Industry funding:** The paper does not disclose whether formula manufacturers provided financial support or products. Given that Nestle, Abbott, and Nutricia products were used, this is a potential conflict of interest.

**Generalisability:** The study excluded infants with intrauterine growth restriction, chronic diseases, or oxygen dependence. These are common in VLBW populations, so results may not apply to sicker infants.

**No long-term follow-up:** 28 days is too short to assess whether the increased lean mass translates to better neurodevelopmental outcomes or whether it increases later metabolic risk.

**Breastfed group not comparable:** The breastfed group was not randomised and had different baseline characteristics (e.g., more cases of necrotising enterocolitis), making any comparison meaningless.

For someone running their own n=1 experiment (e.g., an adult interested in optimising protein and energy intake for lean mass gain):

### What to test

**Intervention:** Increase daily protein intake from ~1.6 g/kg body weight (typical for maintenance) to ~2.2 g/kg body weight, while simultaneously increasing total energy intake by ~15% (e.g., from 30 kcal/kg to 35 kcal/kg). This mirrors the proportional increase used in the study (protein increased by ~14-27%, energy by ~16%).

**Dose:** Based on the study, a protein-to-energy ratio of approximately 2.8 g protein per 100 kcal (as in Group B) was effective. For a 70 kg adult, this would mean ~154 g protein and ~5,500 kcal per day — which is extremely high and likely unnecessary. A more practical translation: increase protein by 0.5-0.6 g/kg/day above your current intake, with a modest calorie surplus.

### Minimum meaningful duration

**28 days minimum.** The study showed measurable differences in body composition within 4 weeks. For adults, changes in lean mass are slower, so 8-12 weeks would be more reliable.

### What to measure

**Primary metric:** Fat-free mass (lean body mass). Use bioelectrical impedance analysis (BIA), DEXA scan, or skinfold measurements with validated equations. Weigh yourself daily under standardised conditions (morning, after voiding, before eating).

**Secondary metrics:** Body weight, waist circumference, strength (e.g., grip strength or 1-rep max on a compound lift), and subjective energy levels.

**Safety markers:** Blood urea nitrogen (BUN) — expect it to rise with higher protein intake. Also monitor kidney function (creatinine, eGFR) and liver enzymes if you have any pre-existing conditions.

### Key confounds to control for

**Total energy intake:** If you increase protein without increasing total calories, the extra protein may be used for energy rather than tissue building. Keep total energy constant or slightly elevated.

**Training status:** If you start a new resistance training programme at the same time, you won't know whether the diet or the training caused the changes. Stabilise your training for 2-4 weeks before changing diet.

**Hydration status:** BIA measurements are sensitive to hydration. Measure at the same time of day, after the same pre-measurement routine (e.g., no exercise for 12 hours, no caffeine for 4 hours).

**Sleep and stress:** Both affect cortisol and growth hormone, which influence protein synthesis. Keep sleep consistent (7-9 hours/night) and log perceived stress.

**Protein source:** The study used a mix of whey and casein (from formula and supplements). Different protein sources have different absorption rates and amino acid profiles. Stick to one source (e.g., whey protein isolate) throughout the experiment.

### What a positive result would look like

**Lean mass gain:** An increase of 0.5-1.0 kg of fat-free mass over 4 weeks (in adults), measured by BIA or DEXA, with stable or slightly reduced body fat percentage.

**Weight gain:** 0.5-1.5 kg total weight gain over 4 weeks, with the majority being lean mass (not fat).

**Strength gain:** A 5-10% increase in 1-rep max on compound lifts (e.g., squat, deadlift, bench press) over 4-8 weeks.

**Subjective:** Improved recovery from workouts, stable energy levels, no digestive distress.

**Safety:** BUN may rise by 10-20% but should remain within normal range. No increase in serum creatinine or liver enzymes. No symptoms of dehydration or kidney discomfort.

**Caveat:** This study was done in preterm infants, not adults. The principles of protein-energy balance likely apply across ages, but the optimal doses and ratios will