Early versus late administration of amino acids in preterm infants receiving parenteral nutrition

  • Review
  • Intervention

Authors


Abstract

Background

Observational studies in preterm newborns suggest that delay in administering amino acids could result in a protein catabolic state and could impact on growth and development.

Objectives

To determine the effect of early administration of amino acids in premature newborns on growth, neurodevelopmental outcome, mortality and clinically important side effects.

Search methods

The standard search strategy of the Neonatal Review Group as outlined in The Cochrane Library was used. Relevant randomised controlled trials were identified by searching the Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane Library 2012 I ssue 9 ), MEDLINE, EMBASE and CINAHL from their earliest dates to September 2012. The trial registry portal of the World Health Organization's International Cilinical Trial Registry Platform and ClinicalTrials.gov (US National Institute of Health) was searched to identify ongoing and completed but unpublished studies.

Selection criteria

Randomised controlled trials comparing early administration of amino acids with late administration in premature newborn infants were included. Early administration of amino acid solution was defined as the administration of amino acids in isolation or with total parenteral nutrition within the first 24 hours of birth; late initiation was defined as the administration of amino acids in isolation or with total parenteral nutrition after the first 24 hours of birth. The primary outcome measures were growth, neurodevelopmental outcome and mortality at 28 days. The secondary outcomes were biochemical abnormalities, sepsis and mortality.

Data collection and analysis

Both review authors independently selected trials, assessed trial quality and extracted data from the included studies. We contacted authors for further information. Fixed-effect analyses were performed. The treatment effect was expressed as mean difference for continuous variables and as risk difference and risk ratio for dichotomous variables. All results included 95% confidence intervals (CIs).

Main results

Seven randomised controlled trials were included in this review. One randomised controlled trial reported no difference in crown-heel length and occipitofrontal head circumference by day 10. Four trials that enrolled 93 premature infants showed positive nitrogen balance (The mean difference with 95% CI was 250.42 (224.91 to 275.93 P value < 0.00001). Four trials showed a significant difference in the level of blood urea nitrogen (BUN) in the first 48 hours (P value < 0.00001). Early administration of amino acids did not result in metabolic acidosis in the first 24 hours.

Authors' conclusions

There is no available evidence of the benefits of early administration of amino acids on mortality, early and late growth and neurodevelopment. There is evidence from four randomised controlled trials included in this review that early administration of amino acids is associated with a positive nitrogen balance. The clinical relevance of this finding is not known. Acid-base status and ammonia levels were normal in the infants who received amino acids early. Given the small number of infants in the randomised controlled trials included in this review, the clinical heterogeneity among them, and the lack of data on important clinical outcomes, there is insufficient evidence to guide practice regarding the early versus late administration of amino acids to infants less than 37 weeks gestation.

Résumé scientifique

Administration précoce versus différée d'acides aminés chez les nourrissons prématurés recevant une nutrition parentérale

Contexte

Les études observationnelles menées chez des nourrissons prématurés suggèrent que le fait de retarder l'administration d'acides aminés risque de provoquer un état catabolique des protéines et d'avoir un impact sur la croissance et le développement.

Objectifs

Déterminer l'effet de l'administration précoce d'acides aminés à des nourrissons prématurés sur la croissance, le développement neurologique, la mortalité et les effets secondaires cliniquement importants.

Stratégie de recherche documentaire

Nous avons utilisé la stratégie de recherche standard du Groupe thématique Cochrane sur la néonatologie telle que décrite dans The Cochrane Library. Les essais contrôlés randomisés (ECR) pertinents ont été identifiés à l'issue de recherches effectuées dans le registre Cochrane des essais contrôlés (CENTRAL, The Cochrane Library 2012, numéro 9 ), MEDLINE, EMBASE et CINAHL, depuis leur date de création respective jusqu'à septembre 2012. Nous avons utilisé le portail de recherche du Système d'enregistrement international des essais cliniques de l'Organisation Mondiale de la Santé (ICTRP) et ClinicalTrials.gov (US National Institute of Health) afin d'identifier des études en cours et achevées mais non publiées.

Critères de sélection

Nous avons inclus les essais contrôlés randomisés (ECR) ayant comparé l'administration précoce à l'administration différée d'acides aminés à des nourrissons prématurés. L'administration précoce d'une solution d'acides aminés a été définie comme étant l'administration d'acides aminés de façon isolée ou avec une nutrition parentérale totale dans les 24 premières heures qui suivent la naissance ; l'initiation différée a été définie comme étant l'administration d'acides aminés de façon isolée ou avec une nutrition parentérale totale après les 24 premières heures qui suivent la naissance. Les principaux critères de jugement mesurés étaient la croissance, le développement neurologique et la mortalité à 28 jours. Les critères de jugement secondaires étaient les anomalies dans les paramètres biochimiques, la septicémie et la mortalité.

Recueil et analyse des données

Les deux auteurs ont sélectionné les essais, évalué leur qualité méthodologique et extrait les données des études incluses de manière indépendante. Nous avons contacté les auteurs des essais pour obtenir des informations supplémentaires. Nous avons analysé les données avec des modèles à effets fixes. L'effet du traitement a été exprimé par la différence moyenne pour les variables continues et par la différence de risque et le risque relatif pour les variables dichotomiques. Les résultats incluaient tous des intervalles de confiance (IC) à 95 %.

Résultats principaux

Sept essais contrôlés randomisés ont été inclus dans cette revue. Un essai contrôlé randomisé n'a rapporté aucune différence au niveau de la taille tête-pieds et du tour de tête occipito-frontal au jour 10. Quatre essais qui avaient recruté 93 nourrissons prématurés ont mis en évidence un bilan azoté positif (la différence moyenne avec IC à 95 % était de 250,42 (224,91 à 275,93, valeur P < 0,00001). Quatre essais ont révélé une différence significative du taux d'azote uréique sanguin (BUN) au cours des 48 premières heures (valeur P < 0,00001). L'administration précoce d'acides aminés n'a pas provoqué d'acidose métabolique au cours des 24 premières heures.

Conclusions des auteurs

Nous n'avons trouvé aucune preuve en faveur des bénéfices de l'administration précoce d'acides aminés pour la mortalité, la croissance précoce et retardée et le développement neurologique. Il existe des preuves issues de quatre essais contrôlés randomisés inclus dans cette revue que l'administration précoce d'acides aminés est associée à un bilan azoté positif. On ne connaît pas l'importance clinique de cette observation. L'équilibre acide-base et les taux d'ammoniaque étaient normaux chez les nourrissons qui ont reçu des acides aminés dans le cadre d'une administration précoce. Compte tenu du petit nombre de nourrissons recrutés dans les essais contrôlés randomisés inclus dans cette revue, de l’hétérogénéité clinique entre les essais, et du manque de données sur les critères de jugement cliniquement importants, les preuves sont insuffisantes pour orienter la pratique en ce qui concerne l'administration précoce versus différée d'acides aminés à des nourrissons nés à moins de 37 semaines de gestation.

Plain language summary

Early versus late administration of amino acids in preterm infants receiving parenteral nutrition

Early administration of amino acids results in positive nitrogen balance. Metabolic acidosis, elevated serum ammonia and hypoglycaemia are not a complication of early administration of amino acids. Elevated blood urea nitrogen (BUN) is consistently associated with early administration of amino acids. We included seven trials in this review. The number of study participants is small and the protocols of the studies included are not identical.

Résumé simplifié

Administration précoce versus différée d'acides aminés chez les nourrissons prématurés recevant une nutrition parentérale

L'administration précoce d'acides aminés a pour résultat un bilan azoté positif. L'acidose métabolique, les taux sériques élevés d'ammoniaque et l'hypoglycémie ne sont pas des complications de l'administration précoce d'acides aminés. L'azote uréique sanguin élevé (BUN) est systématiquement associé à l'administration précoce d'acides aminés. Nous avons inclus sept essais dans cette revue. Le nombre de participants aux études est faible et les protocoles des études incluses ne sont pas identiques.

Notes de traduction

Traduit par: French Cochrane Centre 4th September, 2013
Traduction financée par: Pour la France : Minist�re de la Sant�. Pour le Canada : Instituts de recherche en sant� du Canada, minist�re de la Sant� du Qu�bec, Fonds de recherche de Qu�bec-Sant� et Institut national d'excellence en sant� et en services sociaux.

Background

Premature infants frequently receive nutritional support via parenteral nutrition. Optimal postnatal nutritional intake should ideally follow in-utero fetal growth rates (AAP 1998). However, this goal is difficult to achieve and postnatal growth failure is common (Ehrenkranz 1999; Lemons 2001).

The amino acid (AA) uptake by the fetus is far in excess of that needed to meet accretion requirement and is utilised for fetal energy production. Maximal weight-specific protein gain occurs before 32 weeks of gestation (Micheli 1993).

Description of the condition

A protein intake of approximately 3 gm/kg/day [3.85 gm/kg/day in extremely low birth weight (ELBW) newborns] and an energy intake of 90 kcal/kg/day support intrauterine rates of growth and nitrogen accretion in newborns (Zlotkin 1981; Kashyap 1988; Kashyap 1994; Ziegler 1994). After birth, most infants lose 10% to 20% of weight (Thureen 2005). Half of this loss is due to extracellular water loss, but the remainder represents either loss of, or failure to accrete lean mass and/or fat at a reasonable rate. ELBW infants receiving only carbohydrate as a substrate can lose 1% to 2% of their endogenous protein stores or 0.6 to 1.1 grams of protein per kg per day (van Lingen 1992; Rivera 1993; Van Goudoever 1995; Heird 1999; Denne 2007). Protein losses in neonates receiving intravenous glucose increase with decreasing gestational age (Denne 2007).

Loss of lean mass may have adverse consequences (Thureen 2005). Delaying AA intake in a preterm infant could potentially result in a catabolic state of nutrition, leading to early postnatal growth failure and a widening gap in lean body mass between neonates and fetuses of the same gestational age. The resultant protein deficit may be difficult if not impossible to recoup and may impact on both short-term and long-term morbidities (Denne 2007). Time to regain birth weight could be longer and catch-up requirements are higher. Undernutrition during early periods can have permanent effects on various aspects of development, cognition and behaviour, as well as somatic growth (Ziegler 2002). Higher protein intake can improve growth and possibly reduce neurodevelopmental deficits (Lucas 1990; Lucas 1994; Lucas 1998). Improved anthropometric growth outcomes (weight, length and head circumference) were reported in a secondary analysis of ELBW infants at a postmenstrual age of 36 weeks who received > 3 gm/kg/day of AAs at five days or less of postnatal age compared with those who did not (Poindexter 2006). However, in the same analysis, no difference was seen in neurodevelopmental outcomes at 18 months of age. A cohort of infants who had received early AAs exhibited improved growth with significantly fewer infants in the early AA cohort having weight less than 10th percentile at a postmenstrual age of 32 weeks (Kotsopoulos 2006).

Concerns about metabolic acidosis, elevated ammonia levels and blood urea metabolic derangements precluded the early administration of AAs. These short-term toxicities were not reported to be higher in newborns receiving AA solutions (Ziegler 2002; Denne 2007). High protein intake could lead to cholestasis and the phosphate content of AA may increase the tendency of neonates to develop hypocalcaemia (Andronikou 1983).

The benefits of higher protein intake include greater growth of lean tissue and bones, higher levels of blood constituents, greater synthesis of hormones and enzymes, and the improved maintenance of oncotic pressure (Fomon 1993). Renal hypertrophy and increased circulating insulin-like growth factor-1 have been reported secondary to high protein intake (Murray 1993). In an animal study, higher protein intake was shown to accelerate maturation of the renal tubules (Jakobsson 1990). Deficiency of protein in infants leads to growth failure, causing oedema and decreased resistance to infection (Nayak 1989). High protein intake in early life may increase the risk of long-term obesity and development of diabetes (Rolland-Cachera 1995; Scaglioni 2000; Raiha 2001).

Description of the intervention

Early administration of AA solution was defined as the administration of AAs with or without any other parenteral nutritional intake within the first 24 hours of birth. Late initiation was defined as the administration of AAs with or without any other parenteral nutrition intake.

How the intervention might work

Administering high rates of AA with glucose decreases protein breakdown (van Lingen 1992; Rivera 1993).

Protein intake in the first two weeks of life in ELBW infants is an independent prognostic determinant of growth (Berry 1997). Replenishing protein intravenously can avoid catabolism, achieve protein accretion and improve growth. Reversal of catabolism has been a consistent finding in numerous studies, despite differences in the composition of the AA used (Denne 2007).

Why it is important to do this review

Nutritional management of very low birth weight infants varies from nursery to nursery and diverse nutritional practices can exist within institutions (Ziegler 2002). Analysis of a neonatal nutritional survey reported the practice of administration of protein on first postnatal day in many units in the USA (Hans 2009). With observational studies suggesting that protein catabolism with lack of early AA supplementation and the negative impact of this catabolism on growth and development, it is important to critically review the evidence and evaluate safety, efficacy and balance of benefits versus risks of early supplementation of AAs in order to optimise nutrition and improve long-term growth.

Objectives

To determine the effect of the initiation of early AAs on growth, development and biochemical outcomes in preterm infants receiving parenteral nutrition.

Primary comparison

  1. Early versus late initiation of AAs with or without any other parenteral nutrition intake.

Sub group analysis

  1. Premature infants with a gestational age < 30 weeks versus premature infants with a gestational age ≥ 30 weeks.

  2. Birth weight

    1. Premature low birth weight infants (less than 1500 grams) versus premature infants with birth weight of 1500 grams or more;

    2. Premature very low birth weight infants (less than 1000 grams) versus premature infants with birth weight of 1000 grams or more.

  3. High (≧ 2 gm/kg/day at commencement) or low (< 2 gm/kg/day at commencement) intake

  4. AA solution in isolation versus AA solution with glucose and/or lipids.

  5. Premature infants born with intrauterine growth restriction versus premature infants born with gestational age- appropriate birth weight.

Methods

Criteria for considering studies for this review

Types of studies

Randomised, quasi-randomised and cluster-randomised trials were eligible for inclusion. The control and experimental group differed only in the timing of the initiation of AAs, with the other components of parenteral nutrition remaining the same.

Types of participants

All neonates born at less than 37 weeks of gestation were eligible.

Neonates with metabolic diseases affecting protein metabolism were excluded.

Types of interventions

Timing of the initiation of AA solutions: early administration of AA solution was defined as the administration of AAs with or without any other parenteral nutrition intake within the first 24 hours of birth and late initiation was defined as the administration of AAs with or without any other parenteral nutrition intake after the first 24 hours of birth.

Other parenteral nutrition included intravenous dextrose, solution containing a any combination of dextrose, electrolytes, trace elements administered with or without lipids.

Types of outcome measures

Primary outcomes

1. During the first month of life:

  • weight gain in gm/kg/week;

  • linear growth in cm/week;

  • head circumference in cm/week.

2. Discharge weight in grams.

3. Neurodevelopmental outcome at two years of age (range 18 months to 30 months) assessed using the Bayely Scales of Infant and Toddler Development or Griffiths Mental Development Scales.

4. All-cause mortality at 28 days and before discharge from hospital.

Secondary outcomes

1. Biochemical abnormalities:

  • nitrogen balance, assessed by measuring total urinary nitrogen excretion in the first seven days of life and subtracting it from the nitrogen intake over the same period, and then identifying it as positive or negative;

  • incidence of metabolic acidosis, defined as pH < 7.35 and bicarbonate levels of < 12 mmol/L in the first seven days of postnatal life;

  • incidence of elevated ammonia (> 100 μmol/L or other equivalent unit and blood urea nitrogen (BUN) levels (> 5.0 mmol/L or other equivalent unit) in the first seven days of postnatal life;

  • incidence of cholestasis (elevated alkaline phosphatase > 450 IU/L, serum level of direct bilirubin > 20% of total serum bilirubin or serum level of direct bilirubin > 34 mmol/L);

  • incidence of low serum albumin (< 20 gm/L);

  • incidence of hypocalcaemia (< 2.2 mmol/L) and hypophosphataemia (< 1.35 mmol/L); and

  • incidence of blood sugar level < 2.6 mmol/L or > 6.0 mmol/L or equivalent in mg/dL in the first seven days of postnatal life.

2. Other secondary outcomes:

  • time to regain birth weight in days;

  • incidence of sepsis (positive bacterial culture in cerebral spinal fluid, urine or blood);

  • chronic lung disease (oxygen requirement at or beyond a postmenstrual age of 36 weeks);

  • duration of admission in days;

  • time to full feeds in days.

Search methods for identification of studies

See: Cochrane Neonatal Group methods used in reviews.

We used the criteria and standard methods of The Cochrane Collaboration and the Neonatal Review Group. We included studies identified in the search if they met the inclusion criteria. We did not identify any unpublished studies.

Electronic searches

Two review authors conducted computerised searches. We searched the Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane Library 2012 Issue 9), MEDLINE, EMBASE and CINAHL from their earliest dates to September 2012. We used the MeSH terms infant, newborn OR infant, premature OR infant, low birth weight OR infant, very low birth weight OR infant, extremely low birth weight OR infant, premature diseases OR infant, newborn diseases AND amino acids. In a second strategy we replaced amino acids with parenteral nutrition. The EMTREE vocabulary for EMBASE differs slightly and the terms used in this search strategy were prematurity OR newborn OR low birth weight OR very low birth weight Or extremely low birth weight OR newborn diseases AND amino acids; following this we replaced amino acids with total parenteral nutrition.

We limited the searches to controlled clinical trials and then to randomised controlled trials. Also, with changes in the CINAHL interface after publication of the protocol, studies were restricted to randomised controlled trials. We did not enforce language restrictions. We restricted PubMed clinical queries to therapy with the scope of a broad sensitive search.

Searching other resources

We searched the trial registry portal of the World Health Organization's International Cilinical Trial Registry Platform and ClinicalTrials.gov (US National Institute of Health) to identify ongoing and completed but unpublished studies. We also searched abstracts and conferences and symposia proceedings from Society of Pediatric Research and the American Academy of Pediatrics. We reviewed cross references independently for additional relevant titles and abstracts.

Data collection and analysis

Each review author performed data collection independently and resolved differences by discussion.

Selection of studies

We included the studies that met the prespecified criteria in the review. We merged the search results and duplicate records were removed. We examined the full text of potentially relevant studies.

Data extraction and management

Each review author performed data extraction independently and resolved differences by discussion.

Assessment of risk of bias in included studies

We assessed bias as low risk/unclear risk/high risk and entered assessments into the table 'Characteristics of included studies'. We entered the following information into the 'Risk of bias' table.

  1. Sequence generation: was the allocation sequence adequately generated?

  2. Allocation concealment: was the allocation adequately concealed?

  3. Blinding of participants, personnel and outcome assessors: was knowledge of the allocated intervention adequately blinded during the study? at the study entry? at the outcome assessments?

  4. Incomplete outcome data: were incomplete outcome data adequately addressed?

  5. Selective outcome reporting: is the study report free of the suggestion of selective outcome reporting?

  6. Other sources of bias: was the study free of other problems that could put it at high risk of bias?

Measures of treatment effect

We carried out statistical analysis using the standard methods of the Cochrane Neonatal Review Group. We expressed the treatment effect as mean difference for continuous variables, and as risk difference and risk ratio for dichotomous variables. All the results included 95% confidence intervals.

Unit of analysis issues

As none of the studies included in this review were randomised at cluster level, 'unit-of-analysis' was not an issue.

Assessment of heterogeneity

We used Chi2 and the I2 statistics to detect potential heterogeneity.

Assessment of reporting biases

We used funnel plots to identify publication bias.

Data synthesis

We considered studies that conform to the definition of early and late administration of AAs and included the above-mentioned study populations in the review. We used a fixed-effect model for meta-analysis with 95% confidence intervals (CIs).

Subgroup analysis and investigation of heterogeneity

We did not perform subgroup analysis as none of the criteria mentioned in the protocol for such an analysis were met.

Sensitivity analysis

We did not perform sensitivity analysis as we did not identify any issues suitable for such analysis in the review process.

Results

Description of studies

See:Characteristics of included studies;Characteristics of excluded studies

Seven randomised controlled trials met the inclusion criteria for this review. All seven trials included premature infants who in the 'early-arm' had amino acids administered within the first 24 hours after birth.

Results of the search

After performing the search 23 relevant titles were identified. After reviewing full text of these studies, seven were included in the review and 16 were excluded. No ongoing or completed but unpublished studies were identified.

Included studies

Seven randomised controlled trials were included in this review (Blanco 2008; Heimler 2010; Murdock 1995; Rivera 1993; Saini 1989; Tang 2009; te Braake 2005; van den Akker 2006); van den Akker 2006 reported a subgroup of patients from the study by te Braake 2005. All studies were single centre studies. Three were performed in the USA, two in the UK and one in the Netherlands and one in China.

Blanco 2008 recruited 62 ELBW infants admitted between 2002 and 2005 at the University Hospital, SanAntonio, Texas. Inclusion criteria were birth weight < 1000 grams, age < 12 hours and gestational age ≥ 24 weeks. Infants were randomised to either a standard AAs protocol or an early and high AA protocol. The standard AA group received 0.5 gm/kg/day of intravenous AAs (Aminosyn PF; Abbott Laboratories, Chicago, Illinois with 40 mg/kg/day of cysteine hydrochloride) starting between the first 24 to 36 hours of life with increases of 0.5 gm/kg/day every 24 hours to a maximum of 3.0 kg/day. The early and high AA group received 2 gm/kg/day of AAs within the first 24 hours after birth with increases of 1 gm/kg/day every day to reach 4 gm/kg/day. All infants were prescribed intravenous lipids.

Heimler 2010 recruited 17 appropriate for gestational age preterm infants of < 34 weeks gestation. The mean gestational age of the group that received AAs early was 29.6 weeks and 30.2 weeks of the late group. The early group received 1.5 gm/kg/day of AAs within the 24 hours after birth along with intravenous glucose, calcium gluconate and vitamins. The AA infusion rate was advanced by 0.5 gm/kg/day, up to 2.5 gm/kg/day by day three and continued at this rate during the trial. The other group received intravenous glucose, calcium gluconate and vitamins for three days. AA infusion was added in this group after three days age at 1 gm/kg/day, advanced by 0.5 gm/kg/day up to a maximum of 2.5 kg/day by day seven.

Murdock 1995 recruited 29 infants weighing < 2000 grams who, for clinical reasons, could not receive enteral feeds immediately after birth. The mean gestational age of the infants in the glucose-only group was 31 weeks and 32.8 weeks in the glucose and AAs group.These infants were allocated to three intravenous fluid regimens. The first group received glucose 10% for two days, the second group received glucose 10% and intravenous AAs 1 gm/kg/day on day one and 1.4 gm/kg/day on day two (Vamin 9, Pharmacia, MIlton Keynes, UK) and the third group received glucose 10%, intravenous AAs and lipids. As per the protocol of this review, the third group was not included in this review.

Rivera 1993 enrolled 23 preterm infants with a mean birth weight of 1.07 kg and compared parenteral glucose versus glucose plus AAs during the first three days of life. The mean gestational age was 28.5 weeks in both the groups. The AAs were infused at 1.5 gm/kg/day. The AA solution use in this study was Aminosyn PF; Abbott Laboratories, Chicago, Illinois with cysteine hydrochloride additive which was omitted inadvertently from the nutrient solutions administered to the first four infants. None of the infants in the study received enteral feeds or parenteral lipid infusion.

Saini 1989 enrolled preterm infants of less than 30 weeks gestation into one of the two intravenous feeding regimen groups. The early group received intravenous glucose and AAs (Vamin 9, KabiVitrum) within 24 hours after birth while the late group received intravenous glucose alone and nitrogen only after 72 hours. The nitrogen delivery in the groups was increased progressively over three days from 1 gm/kg/day to 3 gm/kg/day. Intravenous fat was introduced in both the regimens after seven days. The criteria for inclusion were that the infants competed serial 24-hour balance studies of nitrogen and energy over the first 10 days of life and that at least 75% of their nitrogen and energy intake was delivered intravenously. Thirty-two infants were initially recruited. Eleven infants in the early group and 10 in the late group met the criteria for analysis. Six infants in the early group had periods up to 48 hours when they could not receive AAs due to a variety of reasons.

Tang 2009 randomised 106 preterm infants with birth weight between 1000 grams to 2000 grams in three arms. The high AA group received 2.4 gm/kg/day of AA within 24 hours after birth, increasing by increments of 1.2 gm/kg/day to a maximum of 3.6 gm/kg/day. The medium AA group received 1.0 gm/kg/day of AA, 24 hours after birth, increasing by 0.5 gm/kg/day until a maximum of 3.0 gm. The low AA group received 0.5 gm/kg/day of AA at day three and increased by increments of 0.5 gm/kg/day until maximum of 3.0 gm/kg/day as the final dose. The third group was not included in this review.

te Braake 2005 recruited 135 preterm infants with a birth weight less than 1500 grams. Mean gestational age in both the groups was 28.4 weeks. On day one, the infants in the intervention arm received glucose and 2.4 gm/kg/day of AAs and the control arm received glucose only. AA infusion continued at the same rate in the intervention arm and in the control arm infusion rate was increased to 2.4 gm/kg/day, over 48 hours. The same amount of intravenous lipids were administered in both groups from day two onward. The AA infusion in this study was Primene 10%, Baxter, Clintec Benelux NV, Brussels, Belgium.

van den Akker 2006 reported a subgroup of patients from the study by te Braake 2005. Thirty-two premature infants were included in this trial.

Excluded studies

The protocol for this review defined early administration of AA solution as the administration of AAs in isolation or with total parenteral nutrition within the first 24 hours of birth. Anderson 1979; Black 1981; Weinstein 1987; van Lingen 1992 and Bassoiuny 2009 were not included in the review as the age at randomisation differed from that defined in the protocol for this review. In three studies (Chessex 1985; Thureen 2003; Kadrofske 2006), the difference was in the dose of amino acids and not the timing. The study groups differed in the intake of lipids in Gunn 1978; Ibrahim 2004 and Makay 2007. Jadhav 2007 was not a randomised controlled trial. The study groups differed in the intake of carbohydrates in Wilson 1997. Van Goudoever 1995 was excluded as the control arm did not receive AA.

Risk of bias in included studies

Allocation

Blanco 2008 reported that assignment to a treatment group was carried out by the clinical pharmacist with cards in sealed sequential envelopes. Murdock 1995 and Tang 2009 reported random allocation but did not mention the method of randomisation. Heimler 2010 and Rivera 1993 mentioned an envelope method of randomisation. Saini 1989 mentioned sequential allocation but did not describe it any further. te Braake 2005 mentioned the trial as randomised but did not provide any more details. van den Akker 2006 reported a subset of participants from te Braake 2005 and mentioned their trial as randomised and open.

Blinding

Saini 1989; Rivera 1993; Murdock 1995; Heimler 2010 ; Blanco 2008 and Tang 2009 did not report blinding of intervention and outcome measurements. te Braake 2005 and van den Akker 2006 reported that their trials were open.

Incomplete outcome data

The outcomes studied by the trials included in this review were short-term. No obvious bias from incomplete outcome data could be identified.

Selective reporting

The protocol for the trial by Blanco 2008 was registered with ClinicalTrials.gov. There was no mention of registration of protocols by any of the other included studies.

Other potential sources of bias

The trial by Rivera 1993 was supported by a grant from Abbott Laboratories and Gerber Companies Foundation. Heimler 2010 was supported in part by a grant from the Children's Hospital at Wisconsin. Details of financial support were not mentioned by any of the other studies included in this review. Publication bias was not obvious for the nitrogen balance or BUN estimates within the first two days of age. No other potential sources of bias could be identified.

Figure 1 and Figure 2 summarise risk of bias.

Figure 1.

'Risk of bias' graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Figure 2.

'Risk of bias' summary: review authors' judgements about each risk of bias item for each included study.

Effects of interventions

Blanco 2008: The primary aim of this study was to evaluate the effect of early and higher AA supplementation on the incidence of hyperkalaemia. Elevated serum BUN was noted in infants who were in the early and high AA group. AA infusion was discontinued in six infants in the early (mean 3.5 days) and high AA group either due to elevated ammonia or elevated BUN. No difference in sodium bicarbonate levels was seen at days one and two. At day three, the bicarbonate level was higher in the early and high AA group. Sodium bicarbonate therapy did not differ in the groups.

Heimler 2010: The relevant outcomes were nitrogen balance, return to birth weight, head circumference, serum glucose, serum ammonia and nitrogen balance. There was no difference in the head circumference in the two groups in the first two weeks. Serum glucose and ammonia levels were normal and there was no difference in the two groups. The group that receive AAs early had positive nitrogen balance (P value <0.0001) and elevated serum BUN (P value 0.01).

Murdock 1995: Incidence of intervention due to blood sugar level below 2.6 mmol/L, AA profile and peak mean serum bilirubin levels were reported as outcomes. The odds of hypoglycaemia were high in the group that received glucose and AAs, however, the confidence interval was very wide and the result was not statistically significant (RR 0.13, 95% CI 0.01 to 1.44)

Rivera 1993: The relevant outcomes for this study were nitrogen balance and plasma ammonia levels. The plasma ammonia concentrations at day three were similar in the two groups and ranged from 45 to 90 μmol/L and from 46 to 82 μmol/L for the infants infused with glucose alone and glucose and AAs, respectively. The nitrogen balance was calculated as intake minus urinary nitrogen excretion. The infants who received glucose and AAs were in positive nitrogen balance and retained 88 mg/kg/day of nitrogen, whereas the infants who received glucose only had a mean loss of 135 gm/kg/day of nitrogen.

Saini 1989: The outcome of interest was nitrogen balance. The nitrogen balance was calculated as intake minus nitrogen excretion. Nitrogen retention differed between the two study groups in the first 72 hours (P value 0.001). The early-group retained a mean of 120 mg/kg/day of nitrogen and the late-group was in a negative nitrogen balance of 133 mg/kg/day. This study also looked at weight change in gm/kg/day, occipitofrontal head circumference in cm/week and crown heel length in cm/week, in the first 10 days. The anthropometric measurements did not differ in the two groups.

Tang 2009: Postnatal weight loss, length of stay in neonatal intensive care unit (NICU), days to tolerate enteral nutrition, days to reach 2000 grams, time to regain birth weight, blood transfusion, antibiotic treatment, duration of continuous positive airway pressure (CPAP) and a number of biochemical parameters were reported. Intensive and early administration of amino acids was reported to improve growth and tolerance to enteral feeding. It also reduced the cost of hospitalisation and the incidence of sepsis. Levels of BUN, ammonia, pH and bicarbonate were not different in the groups.

te Braake 2005: The outcome measures were BUN, glucose, acid-base status and nitrogen balance. The nitrogen balance was studied at days two and four. The nitrogen balance was calculated as intake minus urinary nitrogen excretion. Nitrogen balance was positive in the intervention arm, which received AAs early compared with the control arm, which had a negative nitrogen balance. Nitrogen balance was reported in the form of a bar chart and actual values were not reported. Base excess, pH and bicarbonate were not different in the two groups at 12 hours and two days. Mean glucose levels were not different at 12 hours. At day two, the mean glucose levels were different in the two groups in a statistically significant way but were normal.

van den Akker 2006: Leucine and glucose kinetics were the outcomes studied. The other reported biochemical parameters were blood glucose levels, pH, base excess, BUN and nitrogen balance. Nitrogen balance from this report was included in this review. The infants in the intervention arm who received AAs early were in a mean positive nitrogen balance of 151mg/kg/day. The mean loss of nitrogen in the infants in the control arm was 99 mg/kg/day.

Primary outcomes

Length and occipitofrontal circumference did not differ between the two groups (Saini 1989; Tang 2009). Heimler 2010 reported no difference in the occipitofrontal circumference in the two weeks. The occipitofrontal circumference data from Heimler 2010 was not included in the meta-analysis as it was not presented as per the protocol of this review.

None of the other primary outcomes (weight gain, discharge weight, neurodevelopmental outcome or mortality) were reported.

Secondary outcomes

1. Biochemical abnormalities:

  • Nitrogen balance

Four studies (Rivera 1993; Saini 1989; van den Akker 2006; Heimler 2010) reported nitrogen balance as an outcome. Premature infants who received AA early had positive nitrogen balance. Negative nitrogen balance was recorded in those who received AA late.

The mean difference with 95% CI was 250.42 (224.91 to 275.93). The P value was < 0.00001. The P value for the Chi2 statistic for assessment of heterogeneity was not significant (0.06) and the I2 statistic showed a moderate heterogeneity (59%)

  • Metabolic acidosis

Mild metabolic acidosis was seen in both the groups with a pH just below 7.35 at days one and two (te Braake 2005; Tang 2009). The difference in pH in the two groups was not statistically significant. The bicarbonate level within the first 24 hours (te Braake 2005; Blanco 2008; Tang 2009), at two days (te Braake 2005; Blanco 2008) and after two days (Blanco 2008; Tang 2009) were normal in two groups.

  • Ammonia

Ammonia levels at day one (Rivera 1993) and days three/four (Rivera 1993; Tang 2009; Heimler 2010) were normal in both early and late AA groups.

  • BUN

An elevated level of BUN in the early AA group in the first two days was reported by Rivera 1993;te Braake 2005; Blanco 2008; and Tang 2009. The meta-analysis showed statistically significant elevation in BUN level (P value < 0.00001, mean difference and 95% CI 2.06 [1.5 to 2.62]). Substantial heterogeneity was noted on the I2 statistic (90%).

On days three and four BUN levels were elevated (Rivera 1993; te Braake 2005;Tang 2009 and Heimler 2010) in the early AA group with a P value of < 0.00001 and a mean difference and 95% CI of 2.12 [1.40 to 2.84]). Substantial heterogeneity was noted on the I2 statistic.

Meta-analysis of BUN level on days six and seven (te Braake 2005; Blanco 2008; and Tang 2009) showed a significantly elevated level in the early AA group. (P value < 0.00001 with mean difference and 95% CI of 2.44 [1.59 to 3.29]). Considerable heterogeneity was noted on the I2 statistic (90%).

  • Cholestatsis

Neither of the groups had cholestasis (direct serum bilirubin > 34 mmol/L) at days one and four (Tang 2009).

  • Blood glucose

te Braake 2005 found a statistically significant difference in the blood glucose level at day two in the experimental and control group. The mean difference and 95% CI were -0.90 and [-1.58 to -0.22] respectively. However, The mean blood glucose was normal in both the study groups at days one and two (te Braake 2005). Murdock 1995 showed no significant difference if the incidence of hypoglycaemia in the experimental and control group [RR 0.61, 95% CI (0.34 to 1.08) and P value 0.1]. Heimler 2010 also reported no difference in the mean blood sugar level in the two groups. These data were not included in the meta-analysis as the time of the test was not reported.

Other secondary outcomes

  • Days to regain birth weight

Tang 2009 reported that the number of mean days to regain birth weight was less in infants who received early AA (11.7 days) compared with those who received it late (14.1 days). No difference was seen by Heimler 2010. The meta-analysis showed a P value of 0.02 favouring early administration of amino acids.

  • Duration of admission

Duration of admission in number of days was less in those who received AAs early (Tang 2009). The P value was 0.05.

  • Time to full feeds

Infants who received early AA had fewer days to reach enteral nutrition compared with those who were commenced on AA late (Tang 2009). The mean number of days in infants the early AA was 13.9 compared with 16.1 in the late AA group.

Chronic lung disease and infection as outcomes were also reported by Tang 2009. However, their definitions were not described.

Discussion

Summary of main results

This review considered trials that examined early administration versus late administration of amino acids (AA). The review identified no short-term difference in length and occipitofrontal circumference between the experimental and control groups. The impact of early administration of AA on the other primary outcomes of interest, weight gain, discharge weight, neurodevelopmental outcome at two years of age, all-cause mortality at 28 days and before discharge, were not reported in the included studies.

Nitrogen balance improved with early administration of AAs. Nitrogen accretion in early life compared to nitrogen loss makes physiological sense and emulates foetal growth but its potential clinical significance is not known. Metabolic acidosis and hypoglycaemia were not a clinically significant problem with early administration of AAs. Early administration of AAs was not associated with clinically significant low pH and low bicarbonate levels in the first 48 hours. Statitistically significant differences in the values of serum bicarbonate were reported by two studies but the mean values were normal in both groups and this finding is not likely to be of clinical significance. An elevated level of BUN was associated with early administration of AAs. An elevated BUN does not necessarily mean intolerance to AA. It could be a reflection of oxidation of AA, just like in utero, where AAs are partly oxidized and partly used for protein synthesis. Ammonia and direct bilirubin levels did not differ between the two groups.

One study showed that length of stay, days to enteral nutrition, duration of admission and days to regain birth weight were all better in the group that received AAs in the first 24 hours.

Overall completeness and applicability of evidence

All participants were premature infants with low birth weight. The mean gestation age was less than 33 weeks and the birth weight was less than 2000 grams. Positive nitrogen balance seen in association with early administration of amino acids in preterm infants had statistical significance with a P value of < 0.00001. Positive nitrogen balance was consistent across all the trials that reported nitrogen balance as an outcome. The clinical significance of the statistically significant observation mentioned above is not clear. Normoglycaemia and normal acid base status in premature infants who have received early amino acid is clinically relevant.  

The data for this review come from short-term studies with small number of study participants. Given the quality of the evidence and small number of participants in the study population, the applicability of the current evidence is not strong.

Quality of the evidence

Several limitations in the quality of evidence should be noted. Only three studies described their method of randomisation. Generation of sequence was not described in any of the studies. It seems that the AA infusion was added separately to glucose and these studies were not blinded to the intervention. It is also not clear if those who measured outcomes were blinded to the allocation. Only one trial reported their protocol to be registered.

The number of study participants is small and significant clinical heterogeneity was observed in the included studies. The inclusion criteria for participants, intervention and outcomes studied were not exactly the same. The mean gestational age of the infants included in this review ranged from 32.8 weeks to 26.3 weeks. The range of mean birth weight was from 1635 grams to 768 grams. Given the small number of infants in this review, we did not perform subgroup analysis. Early AA administration was carried out within the first 24 hours after birth, however, the protocol for late administration was unique for each included study.

The dosage of AA administered was different in all studies. Early AA ranged from 1 gm/kg/day to 2.4 gm/kg/day. Some studies administered the same amount of lipids to both groups. As a result, the amount of energy provided to the treatment and control arms is quite variable between the studies. Also, this meta-analysis has trials that used different preparations of AAs.

Potential biases in the review process

None identified.

Agreements and disagreements with other studies or reviews

There are no other systematic reviews comparing early administration of AA with late administration.

Authors' conclusions

Implications for practice

The evidence gathered from this review suggests no clinically significant biochemical derangements from starting AA early along with benefits in nitrogen accretion. The impact of early AA administration on long-term neurodevelopmental outcome has not been reported. Clinical and statistical heterogeneity of the studies suggests insufficient evidence from randomised controlled trials to guide practice.

Implications for research

Well designed and executed randomised controlled trials with appropriate power involving premature infants looking at impact of early administration of AAs on early growth and mortality as outcomes would help in consolidating evidence for such practice. Further trials that compare the impact of different amino acid solutions on nitrogen accretion will provide evidence in regards to the choice of amino acid solutions.

Data and analyses

Download statistical data

Comparison 1. Growth
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Crown heel length in first 10 days of postnatal life (cm/week)121Mean Difference (IV, Fixed, 95% CI)0.0 [-0.41, 0.41]
2 Occipitofrontal head circumference287Mean Difference (IV, Fixed, 95% CI)0.05 [-0.03, 0.14]
Analysis 1.1.

Comparison 1 Growth, Outcome 1 Crown heel length in first 10 days of postnatal life (cm/week).

Analysis 1.2.

Comparison 1 Growth, Outcome 2 Occipitofrontal head circumference.

Comparison 2. Nitrogen balance
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Nitrogen balance493Mean Difference (IV, Fixed, 95% CI)250.42 [224.91, 275.93]
Analysis 2.1.

Comparison 2 Nitrogen balance, Outcome 1 Nitrogen balance.

Comparison 3. Metabolic acidosis
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 pH within first 24 hours2201Mean Difference (IV, Fixed, 95% CI)-0.01 [-0.03, 0.01]
2 pH at 2 days1135Mean Difference (IV, Fixed, 95% CI)-0.01 [-0.03, 0.01]
3 Bicarbonate first 24 hours3262Mean Difference (IV, Fixed, 95% CI)-0.17 [-0.80, 0.47]
4 Bicarbonate at day 22196Mean Difference (IV, Fixed, 95% CI)-0.66 [-1.30, -0.03]
5 Bicarbonate at day 3-42127Mean Difference (IV, Fixed, 95% CI)1.60 [0.52, 2.69]
Analysis 3.1.

Comparison 3 Metabolic acidosis, Outcome 1 pH within first 24 hours.

Analysis 3.2.

Comparison 3 Metabolic acidosis, Outcome 2 pH at 2 days.

Analysis 3.3.

Comparison 3 Metabolic acidosis, Outcome 3 Bicarbonate first 24 hours.

Analysis 3.4.

Comparison 3 Metabolic acidosis, Outcome 4 Bicarbonate at day 2.

Analysis 3.5.

Comparison 3 Metabolic acidosis, Outcome 5 Bicarbonate at day 3-4.

Comparison 4. Ammonia
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Plasma ammonia at day 1114Mean Difference (IV, Fixed, 95% CI)18.0 [-7.69, 43.69]
2 Plasma ammonia at day 3-4396Mean Difference (IV, Fixed, 95% CI)-1.86 [-7.90, 4.19]
Analysis 4.1.

Comparison 4 Ammonia, Outcome 1 Plasma ammonia at day 1.

Analysis 4.2.

Comparison 4 Ammonia, Outcome 2 Plasma ammonia at day 3-4.

Comparison 5. BUN
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 BUN levels in first two days4285Mean Difference (IV, Fixed, 95% CI)2.06 [1.50, 2.62]
2 BUN Level on day 3 and 44242Mean Difference (IV, Fixed, 95% CI)2.12 [1.40, 2.84]
3 BUN level at day 6-73262Mean Difference (IV, Fixed, 95% CI)2.44 [1.59, 3.29]
Analysis 5.1.

Comparison 5 BUN, Outcome 1 BUN levels in first two days.

Analysis 5.2.

Comparison 5 BUN, Outcome 2 BUN Level on day 3 and 4.

Analysis 5.3.

Comparison 5 BUN, Outcome 3 BUN level at day 6-7.

Comparison 6. Cholestasis
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Direct Bilirubin at day 1166Mean Difference (IV, Fixed, 95% CI)-0.90 [-3.85, 2.05]
2 Direct Bilirubin at day 4166Mean Difference (IV, Fixed, 95% CI)-0.5 [-3.15, 2.15]
Analysis 6.1.

Comparison 6 Cholestasis, Outcome 1 Direct Bilirubin at day 1.

Analysis 6.2.

Comparison 6 Cholestasis, Outcome 2 Direct Bilirubin at day 4.

Comparison 7. Blood glucose
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Blood glucose at day 11135Mean Difference (IV, Fixed, 95% CI)-0.40 [-1.36, 0.56]
2 Blood glucose at day 21135Mean Difference (IV, Fixed, 95% CI)-0.90 [-1.58, -0.22]
3 Hypoglycaemia121Risk Ratio (M-H, Fixed, 95% CI)0.61 [0.34, 1.08]
Analysis 7.1.

Comparison 7 Blood glucose, Outcome 1 Blood glucose at day 1.

Analysis 7.2.

Comparison 7 Blood glucose, Outcome 2 Blood glucose at day 2.

Analysis 7.3.

Comparison 7 Blood glucose, Outcome 3 Hypoglycaemia.

Comparison 8. Days to regain birth weight
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Days to regain birthweight283Mean Difference (IV, Fixed, 95% CI)-2.39 [-4.48, -0.31]
Analysis 8.1.

Comparison 8 Days to regain birth weight, Outcome 1 Days to regain birthweight.

Comparison 9. Duration of admission
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Duration of admission in days166Mean Difference (IV, Fixed, 95% CI)-7.20 [-14.37, -0.03]
Analysis 9.1.

Comparison 9 Duration of admission, Outcome 1 Duration of admission in days.

Comparison 10. Time to full feeds
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Days to enteral nutrition166Mean Difference (IV, Fixed, 95% CI)-2.20 [-4.38, -0.02]
Analysis 10.1.

Comparison 10 Time to full feeds, Outcome 1 Days to enteral nutrition.

Contributions of authors

Both review authors independently performed the literature search and discussed the results. The manuscript was written by Amit Trivedi and reviewed by John Sinn.

Declarations of interest

No conflict of interest.

Sources of support

Internal sources

  • No sources of support supplied

External sources

  • Eunice Kennedy Shriver National Institute of Child Health and Human Development National Institutes of Health, Department of Health and Human Services, USA.

    Editorial support of the Cochrane Neonatal Review Group has been funded with federal funds from the Eunice Kennedy Shriver National Institute of Child Health and Human Development National Institutes of Health, Department of Health and Human Services, USA, under Contract No. HHSN275201100016C

Differences between protocol and review

  1. Additional subgroup analysis based on the presence or absence of intra-uterine growth restriction was added as per suggestion of the peer reviewer.

  2. The interface of CINHAL electronic database changed after publication of the protocol and limits to randomised controlled trial were applied to the search.

  3. The risk of bias changed from yes/no can't tell to low risk, high risk and unclear risk of bias

  4. The data analysis in the review shows biochemical parameters as continuous variables rather than the dichotomous outcomes mentioned in the protocol.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Blanco 2008

MethodsSingle centre randomised controlled trial in the USA
Participants62 ELBW infants with birth weight < 1000 grams were included in this study at age < 12 hours. Exclusion criteria were gestational age < 24 weeks (because of high mortality of infants with gestational age < 24 weeks), major congenital anomaly, imminent death before enrolment and previable infants
InterventionsStandard amino acid group was commenced on amino acid infusion 24 hours after birth and early and high amino acid group received intravenous amino acid within 24 hours of postnatal age
OutcomesSerum potasium, glucose, BUN and insulin levels, urine urea level
NotesAll infants were prescribed lipids
Risk of bias
BiasAuthors' judgementSupport for judgement
Allocation concealment (selection bias)Low riskSequential opaque envelopes were used by a clinical pharmacist
Adequate sequence generationUnclear riskDetails not described
Incomplete outcome dataLow risk 
BlindingHigh risk 
Free of selective reportingLow riskThe protocol was registered with ClinicalTrial.gov
Free of other bias?Low riskDetails of financial support, if any, were not mentioned

Heimler 2010

MethodsSingle centre randomised controlled trial in the USA
Participants17 appropriate for gestational age preterm infants of < 34 weeks gestation were studied during the first week of life. Infants with major congenital anomalies or sepsis were excluded from the study
InterventionsOne group of infants received amino acids within first 24 hours along with intravenous glucose, calcium gluconate and vitamins. The other group received glucose solution with vitamins and calcium gluconate for the first three days. Amino acids were introduced after 72 hours of age in this group
OutcomesBody fluid compartment measurement, plasma amino acids, nitrogen loss and balance, serum ammonia, serum glucose, return to birth weight in days, head circumference increment in two weeks and base excess
Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Allocation concealment (selection bias)Low riskEnvelope randomisation was used
Adequate sequence generationUnclear riskDetails not described
Incomplete outcome dataLow risk 
BlindingHigh risk 
Free of selective reportingUnclear riskRegistration of the protocol was not reported
Free of other bias?Low riskThe study was supported in part by a grant from the Children's Hospital of Wisconsin

Murdock 1995

MethodsSingle centre randomised controlled trial in the United Kingdom. Blinding of randomisation, intervention and outcome measurement was not described. Loss to follow- up was not mentioned
Participants29 premature infants with birth weight < 2000 grams were included in the study. Exclusion criteria were not mentioned. The infants could not receive enteral feeds immediately after birth
InterventionsThese infants were allocated to three intravenous fluid regimens (glucose only, glucose 10% and intravenous AAs, and glucose 10%, intravenous AAs and lipids)
OutcomesHypoglycaemia, plasma triglyceride and cholesterol, peak serum bilirubin
NotesThe third group received intravenous lipids and data from this group were not included in this review
Risk of bias
BiasAuthors' judgementSupport for judgement
Allocation concealment (selection bias)Unclear riskDetails of randomisation were not described
Adequate sequence generationUnclear riskDetails not described
Incomplete outcome dataLow risk 
BlindingHigh risk 
Free of selective reportingUnclear riskRegistration of the protocol was not reported
Free of other bias?Low riskDetails of financial support, if any, were not mentioned

Rivera 1993

MethodsSingle centre randomised controlled trial in the USA. Blinding of randomisation was done by envelopes. Blinding of intervention and outcome measurements was not mentioned. Loss to follow- up was not described
Participants23 preterm infants with respiratory distress syndrome were included in this study. Mean gestational age was 28.5 ∓ 1.8 weeks in the intervention group and 28.5 ∓ 1.8 weeks in the control group
InterventionsOn postnatal day one, the intervention group received glucose and amino acids and the control group received glucose only. Intravenous lipid was not included in the regimen
Outcomes

Biochemical indices: Ammonia nitrogen, ionised calcium, sodium, potassium, chloride, carbon dioxide and urea levels

C-Peptide

NItrogen balance

Leucine metabolism

Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Allocation concealment (selection bias)Low riskEnvelope randomisation was performed
Adequate sequence generationUnclear riskDetails not described
Incomplete outcome dataLow risk 
BlindingHigh risk 
Free of selective reportingUnclear riskRegistration of the protocol was not reported
Free of other bias?Low riskGrants from National Institute of Health, Abbott laboratories and Gerber Companies foundation were reported

Saini 1989

MethodsSingle centre randomised controlled trial in United Kingdom. The enrolment was sequential. Blinding of intervention and outcome measurement were not described
Participants32 preterm infants of less than 30 weeks gestation were enrolled in the study. 11 infants were later excluded from the study as the did not meet the criteria for inclusion (completed serial 24 hour balance studies of nitrogen and energy over the first 10 days of life and at least 75% of nitrogen or energy was delivered intravenously and death
InterventionsOn postnatal day one, the early group received glucose and amino acids and the control group received glucose only
Outcomes

Nitrogen balance

Weight, Length and head circumference at 72 hours

NotesSix infants 'early' group could not receive protein solution up to 48 hours due to a variety of reasons
Risk of bias
BiasAuthors' judgementSupport for judgement
Allocation concealment (selection bias)Unclear riskThe study participants were entered sequentially on admission. No further details were provided
Adequate sequence generationUnclear riskDetails not described
Incomplete outcome dataLow risk 
BlindingHigh risk 
Free of selective reportingUnclear riskRegistration of the protocol was not reported
Free of other bias?Low riskDetails of financial support, if any, were not mentioned

Tang 2009

MethodsSingle centre randomised controlled trial in China. This trial was described as randomised, however method of randomisation and blinding were not described
Participants104 preterm infants were initially randomised in three arms. However, only 96 were included in the study. Mean gestational age was less than 32 weeks and mean birth weight was less than 1600 grams
InterventionsInfants were randomised in either high and early AA arm, medium AA arm which received AA 24 hours after birth and low AA which received AAs on day three. The amino acids were increased in a different way in the high and early AA arm and the medium and low AA arms
OutcomesThe outcomes on interest were postnatal weight loss, length of stay in NICU, days to reach 2000 grams, days to tolerate enteral nutrition, cost of hospitalisation and various biochemical markers
Notes 
Risk of bias
BiasAuthors' judgementSupport for judgement
Allocation concealment (selection bias)Unclear riskThe infants were randomly assigned to the two arm of this study. Details of randomisation were not described
Adequate sequence generationUnclear riskDetails not described
Incomplete outcome dataLow risk 
BlindingHigh risk 
Free of selective reportingUnclear riskRegistration of the protocol was not reported
Free of other bias?Low riskDetails of financial support, if any, were not mentioned

te Braake 2005

MethodsSingle centre randomised controlled trial in Netherlands. Blinding of randomisation was mentioned but not described any further. Blinding of the intervention was reported to as 'not possible'. Blinding of outcome measurement was not mentioned. Loss to follow- up was not described
Participants135 preterm infants with birth weight <1500 grams were included in the study. Mean gestational age was 28.4 ∓ 2 weeks in the intervention group and 28.4 ∓ 1.9 weeks in the control group. Exclusion criteria were congenital abnormalities, chromosomal defects, metabolic diseases and endocrine, renal, or hepatic disorder
InterventionsOn postnatal day one, the intervention group received glucose and amino acids and the control group received glucose only. Intravenous lipid was added on day two and its dose did not differ in either groups
Outcomes

Safety outcomes: Blood gas, glucose concentration, BUN and plasma amino acids concentration

Efficacy Outcomes: Nitrogen balance and age to regain birth weight

NotesNitrogen balance was mentioned only in the graph and no values were mentioned. There was no reply from the corresponding author and this data could not be used in this meta-analysis. Age to regain birth weight was mentioned in median and was not used in this review
Risk of bias
BiasAuthors' judgementSupport for judgement
Allocation concealment (selection bias)Unclear riskThe trial was described as randomised and blinded and it was mentioned that the for logistic reasons, it was not possible to perform a study using a double- blinded design.
Adequate sequence generationUnclear riskDetails not described
Incomplete outcome dataLow risk 
BlindingHigh risk 
Free of selective reportingUnclear riskRegistration of the protocol was not reported
Free of other bias?Low riskDetails of financial support, if any, were not mentioned

van den Akker 2006

  1. a

    AAs: amino acids
    BUN: blood urea nitrogen
    ELBW: extremely low birth weight
    NICU: neonatal intensive care unit

MethodsSingle centre randomised controlled trial in Netherlands
ParticipantsThe included infants were a subset of the patients included by te Braake 2005. A total of 32 premature newborns were included in this article
InterventionsAs above
OutcomesAs Above
NotesThis article reported statistics of nitrogen balance in 32 infants and only that was included in the review
Risk of bias
BiasAuthors' judgementSupport for judgement
Allocation concealment (selection bias)Unclear riskAs mentioned above this study was a subset of the patients included in the te Braake 2005 study. No further risks of bias were identified in this study. Allocation concealment risk same as that of te Braake 2005.
Adequate sequence generationUnclear riskDetails not described
Incomplete outcome dataLow risk 
BlindingHigh risk 
Free of selective reportingUnclear riskRegistration of the protocol of the original study (te Braake 2005) was not reported
Free of other bias?Low riskThere was no funding from the industry and this study was reported to be 'investigator initiated'.

Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion
Anderson 1979The age of infants at randomisation was different to that specified in the protocol
Bassoiuny 2009The age of infants at randomisation was different to that specified in the protocol
Black 1981The age of infants at randomisation was different to that specified in the protocol
Chessex 1985The difference between the experimental and control group was composition of amino acids and not the time of initiation
Gunn 1978The late group did not receive amino acids and the two groups differed in the intake of lipid as well
Ibrahim 2004The early and late amino acid groups differed in the intake of lipid
Jadhav 2007Not a randomised controlled trial
Kadrofske 2006Both experimental and control group were commenced on amino acids in different doses
Makay 2007The two groups differed in the in the intake of lipid
te Braake 2008The primary outcome was glutathione synthesis. Although difference in dextrose and dextrose plus amino acid was mentioned, the values were not published
Thureen 2003The two groups differed in the doses of amino acid intake rather than timing of amino acid intake
van den Akker 2007It was a subgroup analysis of infants in te Braake 2005 study. Albumin levels were depicted in a graph but actual values were not given. The details could not be obtained from the author
Van Goudoever 1995The 'late' group did not receive amino acids
van Lingen 1992Amino acid infusion was commenced in the early group at 2 days of postnatal age
Weinstein 1987The age of infants at randomisation was different to the that specified in the protocol
Wilson 1997The control group and the intervention group differed in glucose intake

Ancillary