First-Week Protein and Energy Intakes Are Associated With 18-Month Developmental Outcomes in Extremely Low Birth Weight Infants

METHODS

During the study period (January 1, 2000, to December 31, 2001), 156 ELBW (≤1000 g), infants were born at ≤32 weeks' gestation at Women and Infants Hospital and survived to discharge. Because our outcome of interest was 18-month neurodevelopment and growth, nonsurvivors were not included in the study cohort. Eight infant survivors were excluded from the study cohort, 7 because of missing data and 1 because of congenital anomaly (duodenal atresia). Of the remaining 148 eligible infants, 124 infants (84%) returned for follow-up at 18 months' corrected age and were included in the analysis.

Daily total enteral and parenteral intakes for the first 4 weeks of life were collected, and mean daily protein and energy intakes were calculated for weeks 1, 2, 3, and 4 based on what the infants actually received according to the bedside nursing flowsheets. Infants received intravenous fluid with dextrose starting at birth. During the study period, a set of nutritional guidelines was used for the initiation and advancement of parenteral nutrition. Parenteral nutrition on day 1 of life was composed of carbohydrate in a monohydrate form that provides 14.3 kJ (3.4 kcal)/g of carbohydrate infused; protein as Trophamine, which provides 16.8 kJ (4.0 kcal)/g of protein; and lipid as Intralipid 20% emulsion, providing 8.4 kJ (2.0 kcal)/mL of Intralipid. Protein was started in the first bag of parenteral nutrition at 1.0 g/kg per day and advanced by 0.5 g/kg per day up to 2.5 to 3.5 g/kg per day. We followed the same guidelines throughout the study period. There was no change in unit policy. Enteral feeds were composed of breast milk (nutritional content presumed that of mature breast milk with 84.0 kJ [20.0 kcal]/oz and 1.54 g of protein per 420 kJ (100 kcal)) with or without Similac Human Milk Fortifier (additional 58.8 joules [14.0 calories] and 1.0 g of protein per 100 mL) or preterm formula initiated at 84.0 kJ (20.0 kcal)/oz and advanced to 100.8 to 113.4 kJ (24–27 kcal)/oz once full enteral feeds were tolerated. The formulations of the 2 preterm formulas used provided 2.7 and 3.0 g of protein per 420 kJ (100 kcal).

Neonatal data were collected including birth weight, gestational age, gender, number of days on supplemental oxygen, number of days on a ventilator, and length of hospitalization. The presence of morbidities, including CLD, NEC, IVH, and PVL, were tabulated. Morbidities were defined using standard published definitions. CLD was defined as oxygen requirement at 36 weeks' postconceptional age. NEC was Bell's stage 2 or higher,²⁵ defined as clinical and radiographic evidence of NEC resulting in treatment with bowel rest, total parenteral nutrition, and intravenous antibiotics for ≥7 days or need for surgical intervention. IVH was defined as grade 3 or 4 on a cranial ultrasound using the grading system from Papile et al.²⁶ Maternal socioeconomic data, including race, maternal age, education, and marital status, were recorded.

Infants were seen for follow-up evaluation at 18 months' corrected age by developmental specialists. At these evaluations, the children's primary caretakers were interviewed to review medical, social, and demographic histories. The 18-month physical assessment included standard measurement of weight, length, and head circumference,⁸ as well as a developmental evaluation. Body weight was obtained using digital electronic scales reading to the nearest gram. Recumbent length was measured using a length board. Occipitofrontal head circumference was measured by placing a paper measurement tape firmly around the head at the most prominent part of the frontal bulge and the part of the occiput that gave the largest circumference. The developmental evaluation was performed using the Bayley Scales of Infant Development II²⁴ administered by testers trained to reliability. An MDI and PDI were derived for each child. MDI and PDI scores have a mean of 100, with an SD of 15 points in a normal population.

Correlations were performed to evaluate the relationship between protein and energy intake during weeks 1, 2, 3, and 4, and MDI, PDI, weight, length, or head circumference at 18 months' corrected age. Multiple regression models were used to evaluate the independent contributions of energy intake and protein intake to growth and development while controlling for factors known to impact on growth or development. Separate logistic regression models were performed for weekly protein intake in grams per kilogram per day and weekly energy intake in kilojoules (kilocalories) per kilogram per day to predict weight, length, or head circumference <10th percentile at 18 months' corrected age. Independent variables known to impact growth, including birth weight, small for gestational age (SGA) status, CLD, and NEC, were entered into the models. Linear regression models were also developed to predict Bayley MDI or PDI scores at 18 months' corrected age. Separate models were performed for weekly protein intake in grams per kilogram per day and weekly energy intake in kilojoules (kilocalories) per kilogram per day. Independent variables known to impact on outcome, including birth weight, gender, maternal education >12 years, and presence of CLD, IVH/PVL, and NEC, were entered into each model. The final variable, protein (mean grams per kilogram per day) or energy (mean kilojoules [kilocalories] per kilogram per day) in week 1, was entered into each model as a continuous variable.

This study was approved by the Women and Infants' Hospital Institutional Review Board. Because of the retrospective design, informed consent was waived.

RESULTS

Mean birth weight of our cohort was 787 g (±133 g; range: 460–1000g). Mean gestational age was 25.9 weeks (±1.6 weeks; range: 23–31 weeks). The population was composed of 43% boys and 13% SGA infants. The infants required supplemental oxygen for an average of 57 days (±33 days; range: 0–131 days), ventilator support for an average of 19 days (±17 days; range: 0–116 days), and hospitalization in the NICU for an average of 89 days (±27 days; range: 13–235 days). Thirty-two percent had CLD, 16% were diagnosed with NEC, 4% had grade 3 to 4 IVH, and 1% had PVL. A total of 124 (84%) of 148 eligible infants were seen in follow-up at 18 months' corrected age.

Seventy-two percent of the population were white, 13% were black, 11% were Hispanic, and 4% were Asian. Fifty-seven percent of the mothers were married at time of delivery, and 66% were married at the time of the 18-month visit; 93% of the mothers were high school graduates, and 41% were college graduates. The mean maternal age at delivery was 29 years (±6 years; range: 15–44 years).

Daily mean protein and energy intakes for the first week of life and weekly mean protein and energy intakes for the first month of life are shown in Table 1. Energy intake increased from 252 kJ (60 kcal)/kg per day in week 1 to 441 kJ (105 kcal)/kg per day by weeks 3 and 4. Protein intake increased from 1.8 g/kg per day in week 1 to between 3.3 and 3.5 in weeks 2 to 4. The majority of energy and protein intakes during the first week of life were parenteral (98% of protein and 98% of energy). Protein and energy intakes were highly correlated (r = 0.7; P < .001).

Neurodevelopmental and growth outcomes at 18-month follow-up are shown in Table 2. Mean MDI and PDI, as well as percentage of infants with scores >2 SDs below the mean, are shown. Mean weight, length, and head circumference at 18 months, as well as percentage of infants below the 10th percentile for each growth parameter, are also shown. Of the total cohort, 29% had MDI <70, 22% had PDI <70, and one third were growth restricted below the 10th percentile at 18 months. Eight (50%) of 16 infants born SGA remained below the 10th percentile at 18 months. Thirty-two of 108 infants born appropriate for gestation age (30%) had fallen to below the 10th percentile at 18 months.

Attrition analysis revealed some differences between those infants who were seen in follow-up and those who were not. Infants seen in follow-up had a lower mean birth weight (787 vs 848 g; P = .0383) and were more likely to be girls (92% vs 75%; P = .004). No differences were seen in week 1, 2, 3, or 4 energy or protein intake between those infants seen and those not seen at follow-up.

First-week energy intake correlated with higher MDI scores (r = 0.3; P = .0004) and PDI scores (r = 0.3; P = .0004) at 18 months. First-week protein intake correlated with higher MDI scores alone (r = 0.3; P = .004). Significant associations between energy or protein intakes during weeks 2, 3, or 4 and neurodevelopmental outcomes were not seen. Energy and protein intakes in the first week did not correlate with weight, length, or head circumference at 18 months.

Multiple linear regression analyses were performed to assess the independent contribution of energy and protein intakes to MDI and PDI scores at 18 months while controlling for birth weight, gender, maternal college education, and morbidities associated with neurodevelopmental outcomes (CLD, IVH, and NEC). As shown in Table 3, mean energy intake in the first week of life made a significant independent contribution to MDI (b = 0.46; P = .0134) at 18 months. Energy intake was entered into the model as kilojoules (kilocalories) per kilogram per day, so for each 4.2 kilojoules (1 kilocalorie) per kilogram per day increase in energy intake during week 1, there was an associated 0.46-point increase in MDI at 18 months, and each 42-kJ (10 kcal/kg per day increase in energy intake was associated with a 4.6-point increase in MDI at 18 months. This model accounted for 18% of the variance in MDI at 18 months.

As shown in Table 4, mean protein intake during week 1 also had an independent effect on MDI at 18 months (b = 8.21; P = .0274). Protein was entered into the model as grams per kilogram per day, so for every 1-g/kg per day increase in protein intake during week 1, there was an associated 8.2-point increase in Bayley MDI at 18 months. This model contributed 17% to the variance in MDI at 18 months.

In each of our models, birth weight and male gender had significant independent associations with cognitive outcomes. For every gram increase in birth weight, there was an associated 0.03-point higher MDI score at 18 months (b = 0.03, P = .0244 and b = 0.03, P = .0277). Male gender was associated with a lower MDI score at 18 months by >8 points (b = 8.23, P = .0055 and b = 8.72, P = .0033). The presence of common neonatal morbidities was not significantly associated with MDI at 18 months in our models.

The relationships between energy intake and MDI and protein and MDI were examined and were linear. An effect of energy or protein intake on PDI was not seen (models not shown).

Multiple logistic regression analyses were performed to assess the independent contribution of energy and protein intakes to growth restriction (weight, length, or head circumference <10th percentile), while controlling for birth weight, SGA status at birth, and morbidities associated with growth restriction (CLD and NEC). Higher protein intake was associated with lower likelihood of length <10th percentile at 18 months (odds ratio: 0.260 [CI: 0.076 to 0.907]; P = .0345). Energy and protein intake were unrelated to weight and head circumference at 18 months (models not shown).

DISCUSSION

This is the first study to quantify the independent contribution of first-week energy and protein intakes to neurodevelopmental outcome in the ELBW infant. An increase in first-week energy intake of 42 kJ (10 kcal)/kg per day was independently associated with an ∼5-point increase in MDI, and an increase in first-week protein intake of 1 g/kg per day was independently associated with a >8-point increase in MDI. This has important implications for neonatal clinical practice and allocation of post-NICU resources for early intervention services.

Our models predict 17% to 18% of MDI at 18 months, thus accounting for a generous amount of the variance in such a complicated outcome as long-term cognitive score. We did not perform this study as an attempt to predict Bayley scores at 18 months nor to identify all of the factors that contribute to cognitive outcomes. Rather, this was a study to determine the contribution of nutrition to 18-month cognitive outcomes while controlling for neonatal covariates that may impact on both neonatal nutrition and neurodevelopmental outcomes.

There was a wide range of protein and energy intakes among infants in this study. During the entire study period, the same set of nutritional guidelines was used for the initiation and advancement of parenteral nutrition. The source of variation in protein intakes was the varied comfort levels among individual neonatologists with regard to starting and/or advancing parenteral nutrition in the first week of life. There was no change in unit policy. Enteral nutrition was started after the first few days of life, also at the discretion of the neonatologist, and advanced gradually, again based on clinical judgment. The majority of energy and protein intakes given during the first week of life were parenteral. Although severity of illness and neonatal morbidities may affect these decisions, with sicker infants having nutrition started and advanced more cautiously, our models controlled for neonatal morbidities; thus, associations seen are thought to be real.

There is accumulating evidence that malnutrition during periods of vulnerability alters the growth of the developing brain and may have permanent negative developmental consequences.^27–29 In humans, the most critical developmental period of brain growth and function occurs during the third trimester of pregnancy and the first 2 years of postnatal life.^27,30 Thus, the provision of adequate nutrition to infants during this period is critical. Yet, nutritional practices vary greatly among NICUs, and energy-containing macronutrients (protein, carbohydrate, and lipid) are often introduced slowly and increased cautiously because of concerns of intolerance. This results in a period of nutritional deficiency that is common and accepted as inevitable but may lead to early malnutrition.^18,31

In much larger preterm infants (<1750 g), energy deprivation (defined as intakes of <357 kJ [<85 kcal]/kg per day) has been directly related to poor head growth, and prolonged periods of energy deprivation (>4 weeks in duration) has been associated with lower PDI scores at 12 months.³² Poor in utero and extrauterine head growth in VLBW infants has been shown to predict lower MDI scores at 15 to 20 months^33,34 and lower verbal and performance IQ, receptive and expressive language, and academic abilities at school age.³⁵ In SGA infants, head circumference “catch-up” growth has been correlated with higher energy intake during the first 10 days of life and higher IQs later in life.³⁶ However, the direct effects of energy intake on cognitive outcomes have not been shown. Although a recent Cochrane review³⁷ suggests that higher protein intake accelerates weight gain in low birth weight neonates, the effects of protein intake on cognitive outcomes have not been demonstrated.

Several studies have demonstrated the safety of the early provision of higher levels of protein and energy intake without adverse clinical sequelae.^19–23 Recently, “early, aggressive” parenteral and enteral nutrition has been advocated to prevent nutritional deficits.^19,29,31,38 However, the data demonstrating the effects of this “aggressive” nutrition on neurodevelopmental outcome are sparse.

Three studies have demonstrated that higher protein and energy intakes in the first days of life result in better growth at NICU discharge in VLBW (<1500 g) infants.^39–41 Each demonstrated lower rates of growth restriction (parameters <10%) at discharge in those infants fed more protein and/or energy in the first weeks of life. However, none reported neurodevelopmental outcomes.

Poindexter et al¹⁰ demonstrated better head growth at 18 months' corrected age associated with a higher protein intake in ELBW infants. However, better head growth did not translate to better neurodevelopmental outcomes; no differences in rates of neurodevelopmental impairment (MDI < 70, PDI < 70, cerebral palsy, blind, or deaf) at 18 months' corrected age were identified between infants receiving higher or lower protein intakes. In our cohort, higher protein intake was associated with lower likelihood of length growth restriction at 18 months. However, restrictions of head growth and weight were not associated with first-week intakes. This is not surprising, because nutrition during the remaining weeks and months of hospitalization, as well as after NICU discharge, would be expected to contribute substantially to growth at 18 months.

An association between poor growth and neurodevelopmental outcome has been demonstrated previously. Connors et al⁴² demonstrated an association between lower developmental scores and weight <10th percentile at 2 years in a cohort of 70 high-risk ELBW infants. Ehrenkranz et al¹² demonstrated a significant relationship between in-hospital growth velocity and neurodevelopmental outcome in 490 ELBW infants from August 1994 to August 1995. Lower rates of weight gain and head growth were both significantly associated with cerebral palsy, MDI < 70, and neurodevelopmental impairment at 18 months' corrected age when controlling for confounding variables. However, neither of these studies examined nutritional intake.

A clear association between early enteral intake and neurodevelopmental outcome has been demonstrated in larger preterm infants. Lucas et al^43,44 demonstrated higher cognitive and motor scores at 18 months' corrected age and higher verbal IQs and lower rates of cerebral palsy at 7.5 to 8.0 years of age in preterm infants (<1850 g) fed preterm formula containing 2 g of protein and 336 kJ (80 kcal)/dL versus term formula containing 1.45 g of protein and 286 kJ (68 kcal)/dL for the first 4 weeks of life.

The association between breast milk intake and neurodevelopmental outcome in ELBW infants is less clear. Furman et al⁴⁵ demonstrated no effect of maternal milk intake in the first 4 weeks of life on 20-month cognitive or motor outcomes in a cohort of 98 VLBW infants. Conversely, in a cohort of 1035 ELBW infants, Vohr et al⁴⁶ found that those infants fed more breast milk versus formula during hospitalization in the NICU had higher 18-month MDI scores.

It is important to note that studies of enteral nutrition and its effects on developmental outcome fail to account for nutrition in the first days and weeks of life, because most ELBW infants do not begin to feed enterally for several days and do not reach full enteral feedings for weeks. No previous studies have clearly demonstrated neurodevelopmental effects of early, parenteral nutrition in the first few days of life.

In addition to the benefits, there may be adverse long-term effects of increased early protein and energy intake. Recent studies^47,48 suggest that rapid “catch-up growth” may shorten life span by increasing cardiovascular risk, risk of hypertension, obesity, and noninsulin dependent diabetes. We do not know whether there is a threshold effect.

The strengths of this study include analysis of total enteral and parenteral energy and protein intake, enrollment inclusive of all ELBW survivors during a 2-year period, analysis of intake related to 18-month outcome, and good follow-up rate of 84%. A limitation of this study is its retrospective design. A prospective longitudinal study is needed to fully examine the effects of early protein and energy intake in conjunction with known confounders on the outcomes of ELBW infants.

CONCLUSIONS

We have shown that first-week energy and protein intakes are strongly associated with 18-month developmental outcomes in ELBW infants. If the degree of association with cognitive outcome at 18 months can be replicated in a prospective trial, the results could have enormous implications for clinical practice. Societal implications of an 8-point increase in cognitive score are also substantial, resulting in a potentially significant decrease in the number of children and adults in need of educational services and supports. Increased emphasis should be given to providing adequate nutrition in the first week of life for all ELBW infants while awaiting results of a prospective trial.