OBJECTIVE: To describe the influence that gestational age and chronological age have on amino acid and acylcarnitine profiles in an at-risk population of premature infants.
METHODS: Metabolic profiles (15 amino acids and 35 acylcarnitines) were obtained by using standard newborn techniques on infants born between 23 and 31 completed weeks of gestation. The profiles were drawn within the first 24 hours after birth and on approximately days 7, 28, and 42 of life or at discharge. A single, central, contract laboratory analyzed and managed the samples.
RESULTS: We studied 995 patients; none was subsequently diagnosed with an inborn error of metabolism. Of the 3579 samples, there were 257 (7.2%) amino acid or acylcarnitine alerts reported in 214 infants (21.5% of infants studied). Both gestational age and postbirth chronological age significantly influenced the metabolic profile. Twenty-nine percent of infants at 23 to 26 weeks’ gestational age had an abnormal metabolic profile compared with 17% of infants at 29 to 31 weeks’ gestational age (P < .01). On the day of birth, 12% of the profiles were abnormal compared with 2% on day 28 (P < .01). The highest rate of abnormal values occurred on day 7 in the infants 23 to 26 weeks’ gestational age (21%).
CONCLUSIONS: These results demonstrate the complexity of understanding the impact of immaturity and disease on metabolic profiles used to screen for inborn errors of metabolism. Our data provide reference values for studies aimed at better understanding metabolism in preterm infants.
- ANOVA —
- analysis of variance
- EGA —
- estimated gestational age
What’s Known on This Subject:
Prematurely born infants commonly have abnormal metabolic screens.
What This Study Adds:
Both gestational and chronological age influence metabolic profiles used to screen for inborn errors of metabolism.
We previously reported that the amount of amino acids provided in parenteral nutrition influences blood amino acid and acylcarnitine values. Other investigators have reported similar results. 1 – 5 Establishing a consensus regarding what represents a normal metabolic profile in premature infants would allow investigators to identify infants with abnormal values and to evaluate the impact of those abnormal values on important health outcomes (liver injury, growth, and neurodevelopmental outcomes). Failure to establish normal values undermines our understanding of amino acid and fatty acid metabolism in premature infants. The purpose of this study is to describe the influence that gestational age and postbirth chronological age have upon amino acid and acylcarnitine profiles, to better understand what represents “normal” for this stressed and at-risk population.
We included inborn infants between the gestational ages of 23 weeks and 0/7 days and 31 weeks and 0/7 days and excluded infants if they had major congenital anomalies. Parental consent was obtained for each subject. The protocol was approved by the Western Institutional Review Board and/or each site’s local institutional review board.
Fluid management and nutritional guidelines were used as per site specific protocol. Data on precisely what was received by the study subject and growth parameters were collected on postnatal days 1, 7, 28, and 42.
An experienced health care professional used newborn screening consensus protocols for obtaining study and state screening blood samples. There was no charge for study samples, and the samples were drawn when other laboratory tests were done. Study filter papers were provided at the site initiation visit, and the precise procedure for collection of blood samples was reviewed. If scavenged blood was used, the sample had to be obtained from a line that did not contain amino acids, glucose, or lipids to prevent contamination. The date and time of the sample was recorded on the case report form and the filter paper. Cord blood was not used for the initial sample. Filter papers were not batched; they were mailed to the central laboratory within 24 hours of collection and analyzed within 48 hours of arrival.
Metabolic profiles were collected within the first 24 hours after birth (median age in hours = 16 hours), on approximately day 7 (7–8), on day 28 (27–29), and on day 42 (41–43) of life or at discharge, whichever came first. All patients were followed until the time of discharge to collect data on morbidity and mortality. Investigators were instructed to obtain the initial metabolic profile sample before the start of parenteral nutrition. This approach proved difficult to accomplish because all sites included some amino acids in the initial intravenous fluids (range, 1–3 g/100 mL).
A central, contract laboratory (PerkinElmer Genetics, Bridgeville, PA) analyzed the study samples by using tandem mass spectrometry. Alert values were set as defined by Chace et al. 6 , 7 The reference values for healthy newborn term infants were defined as per our previous study. 8 Investigators were notified if a sample value fell outside of the reference range for normal term newborns and a plan for follow-up was discussed. We collected data on all abnormal results reported by the state and study laboratories.
An electronic system was used to assign unique study codes and collect data on each subject. Clinical Research Associates monitored each site for adherence to the protocol and data accuracy.
Categorical variables were evaluated by 2-tailed χ2 and Fisher’s exact test. Continuous variables were compared by using a 2-tailed analysis of variance (ANOVA) for parametrically distributed data and Kruskal-Wallis ANOVA for nonparametrically distributed data. ANOVA for repeated measures was used to evaluate the changes in laboratory and nutritional values over time, and within each estimated gestational age (EGA) group. A comparison for all possible pairs was done by using Tukey–Kramer tests. To summarize and compare the values obtained for each analyte, we calculated a z score for each patient by using the mean and SD values from a normal term infant population sample reported in our previous work. 8 The z score for each value was calculated: observed value minus the mean value of a normal term infant/divided by the SD from the normal term infant sample. A median z score of >1 or less than −1 means that 50% of the patients had a value that was >1 SD away from the mean value seen in term infants (Figs 1 and 2).
There were 1002 unique subjects enrolled from 23 sites in 17 states. Six patients had anomalies that were found after they were enrolled; these patients were excluded (2 with Trisomy 21, 1 bowel atresia, 1 chromosomal abnormality not defined, 1 Nager Syndrome/type II, and 1 chromosome 22Q duplication). One subject had a duplicate study code generated; leaving 995 subjects for analyses. On follow-up, no study subject was diagnosed with an inborn error of metabolism.
Characteristics are described as median (10–90th percentile): birth weight 1057 (644–1514) grams, EGA was 28 (24–30) weeks; Apgar at 1 minute 6 (1–8); Apgar at 5 minutes 8 (5–9; Table 1).
Before analysis, the 995 infants were subdivided into 3 EGA groups; 23 to 26 weeks (n = 293); 27 to 28 weeks (n = 277); and 29 to 31 weeks (n = 425). These subgroups were selected based on sample size (goal >200) and degree of prematurity. Each EGA group is distinctive; the more immature infants are smaller, had lower Apgar scores, less often were white, and were more often treated with steroids, surfactant, and insulin. The most immature infants (23 to 26 weeks) had more morbidity and it took them longer to reach full feeds (Table1).
Parenteral nutrition was started at a median of 2 (1 to 25) hours; the median amount of glucose given on day 1 was 7 (5–9) g/kg per day. Lipids were started at a median of 27 hours (9–48) at 2 (1–3) g/kg per day. In 390 (39%) infants, carnitine was added to the parenteral nutrition and the proportion of infants given carnitine increased with decreasing gestational age. The majority of early nutrition (≤7 days) was provided by intravenous glucose, amino acids, and fatty acids. Most (78%) infants did not receive enteral feedings during the 24 hours after birth. Infants who were fed received <30 cc/kg per day (Table 2). Infants in the 23 to 26 EGA group were exposed to higher volumes of intravenous fluids, were on parenteral nutrition longer, and took more time to get to full feedings (Tables 1 and 2). The peak use of intravenous lipids and amino acids was on day 7. Over time there was a shift from parenteral to enteral nutrition. By day 42, all 3 groups approached 120 calories/kg per day and most infants (84%) were receiving strictly enteral feedings. Growth velocities for each EGA group were similar, but the highest growth velocity was in the more mature infants (29 to 31 weeks) and during the enteral phase of nutrition.
Abnormal Newborn Screens
Differences Across State Newborn Screen
We requested that each site provide their newborn state screen panel and alert values. We received 15 of the 17 state references. Although the panels and the alert limits were similar, none were exactly the same. What is measured, units of measurement, alert levels, and analytes used as primary and secondary markers for identification of disease were different for each state. (The National Newborn Screening and Genetics Resource Center Web site http://genes-r-us.uthscsa.edu/sites/genes-r-us/files/nbsdisorders.pdf).
State Newborn Screen Results
There is an important limitation to our observations. Each state has its own panel and recommended procedures for identifying infants with inborn errors of metabolism and there are no accepted standards that are universally applied. 9 – 11
The initial state screening test was abnormal and a repeat study was requested in 461 (46%) of the 995 infants studied. The more immature infants most often had an abnormal state screen (Table 1; 64% vs 36%, P < .01). The most common abnormalities reported by the state laboratory were as follows: low thyroxine or high thyroid-stimulating hormone (n = 263, 26.4%); elevated 17-hydroxyprogestrone (n = 78, 7.8%); abnormal amino-acid profile (n = 89, 8.9%), and abnormal acylcarnitine profile (n = 26, 2.6%). The most common abnormal amino acids were elevations of methionine (n = 43, 4.3%), tyrosine (n = 26, 2.6%), and/or leucine (n = 16, 1.6%). The most common acylcarnitine elevation was isovalerylcarnitine (C5; n = 14, 1.4%).
Central Laboratory Alerts Based on Chace et al Values 6
Of 3579 samples, there were 257 (7.2%) amino acid or acylcarnitine alerts reported in 214 infants (21.5% of infants studied). We did not include thyroid screening as a part of our study alerts. Similar to the state screening results, the most common alerts were in the most immature infants and most often occurred on days 1 and 7 (Table 1). Twenty-nine percent of infants who were 23 to 26 weeks’ EGA had an abnormal metabolic profile compared with 17% of infants 29 to 31 weeks’ EGA (P < .01). On the day of birth, 12% of the profiles were abnormal compared with 2% on day 28 (P < .01); the highest rate of abnormal values occurred on day 7 in the infants 23 to 26 weeks’ EGA (21%).
Of the 257 alerts, the most common reports were as follows: elevated methionine (n = 89); generalized elevation of short-chain and medium chain acylcarnitine possibly consistent with L-carnitine supplementation (n = 69); mild elevation of multiple amino acids including phenylalanine, leucine, and methionine (n = 22); elevation in concentrations of propionylcarnitine (C3) and other indices, such as the relative ratios of propionylcarnitine (C3) to acetylcarnitine (C2) or propionylcarnitine (C3) to palmitoylcarnitine (C16; n = 15); elevation in tyrosine (n = 15); elevation in methionine and tyrosine (n = 7); and moderate elevation of multiple amino acids including phenylalanine, leucine, and methionine (n = 5).
Results of Metabolic Profiles
Comparison of Preterm to Term Controls
With the exception of arginine, serine, free carnitine, malonylcarnitine (C3-DC)+3-hydroxyoctanoylcarnitine (C8-OH), hydroxybutyrylcarnitine (C4OH), methylmalonylcarnitine (C4-DC), 3-hydroxyisovalerylcarnitine (C5-OH), glutarylcarnitine (C5-DC)+3-hydroxydecanoylcarnitine (C10-OH), hexanoylcarnitine (C6), methylglutarylcarnitine (C6DC), hydroxyhexanoylcarnitine (C6OH), dodecenoylcarnitine (C12:1), 3-hydroxydodecanoylcarnitine (C12-OH), hydroxytetradecadienylcarnitine (C14OH), hydroxypalmitoylcarnitine (C16OH), 3-hydroxystearoylcarnitine (C18-OH), and 3-hydroxyoleoylcarnitine (C18:1-OH), all analytes were significantly different from term controls (Supplemental Table 3) and changed over time (Figs 1 and 2). Every possible pattern of change was observed (up over time, down over time, up then down, down then up). The time points most consistently found to be different from the other time points were days 1 and 7 (peak of intravenous nutrition was day 7; the peak of enteral nutrition was day 42).
Comparison of EGA Groups Over Time
Amino Acid Profile
Effect of EGA
Across EGA groups, there were significant differences in alanine, aspartate, citrulline, glutamine, glutamate, glycine, phenylalanine, serine, and valine (Supplemental Table 3). Alanine, glutamate and serine were lower in the 23- to 26-week EGA group compared with the other 2 groups. Aspartate, phenylalanine, and valine were higher in the 23 to 26 EGA group compared with the other 2 groups. Citrulline and glycine were significantly higher in the 29 to 31 EGA group compared with the other 2 groups.
Effect of Time
With the exception of serine, all the amino acids changed over time, and every possible pattern of change was observed. The time points most consistently found to be different from the other time points were days 1 and 7. With the exception of citrulline, glutamine, and tyrosine, there were no differences between the amino acid profiles obtained on days 28 and 42; citrulline, glutamine, and tyrosine were slightly higher on day 42 compared with day 28 (Supplemental Table 3).
Interaction Between EGA and Time
The pattern of change over time for each EGA group was similar, but the degree of change over time was significantly different for histidine, phenylalanine, aspartate, citrulline, glutamate, glutamine, glycine, and tyrosine. The most distinctive changes were in citrulline (increase over time especially in the 29 to 31 EGA group), leucine-isoleucine (increased on day 7 and then decrease with time for all EGA groups), and tyrosine (decreased on day 7 and then increase with time especially in the 23 to 26 EGA group; Supplemental Table 3). There was a weak correlation between leucine-isoleucine values and the dose of amino acids (g/kg per day; R 2 = 0.1, P < .01) that was independent of EGA group and age at when the blood sample was done. On day 7, the type of amino acids also influenced the amino acid profile (data not shown).
Impact of EGA
Across all EGA groups, there were significant differences between the acylcarnitine levels except for hydroxybutyrylcarnitine (C4OH), hydroxyhexanoylcarnitine (C6OH), dicarboxydodecanoylcarnitine (C12DC), and tetradecenoylcarnitine (C14:1; Supplemental Table 3). The most distinctive group was the 23 to 26 EGA group. With the exception of the 4 acylcarnitines listed above, the 23 to 26 EGA group was significantly different from the other 2 EGA groups. The 2 more mature EGA groups were similar to each other for all acylcarnitine analytes.
Impact of Time
Every acylcarnitine changed over time except hydroxytetradecadienylcarnitine (C14OH). Every possible pattern of change was observed (up over time, down over time, up then down, down then up). The time points most consistently found to be different from the other time points were day 1 and day 7. Although the acylcarnitine values obtained on days 28 and 42 were similar, free carnitine, acetylcarnitine (C2), propionylcarnitine (C3), methylmalonylcarnitine (C4-DC), dodecanoylcarnitine (C12), and tetradecenoylcarnitine (C14) were statistically higher on day 42 than on day 28. The acylcarnitine with the most striking increase from day 1 to 7 followed by a decrease was linoleoylcarnitine (C18:2; Fig 2). Linolenic acid is one of the primary fatty acids in intralipids. Linoleoylcarnitine returned to lower levels on days 28 and 42 when the primary nutrition was enteral feedings and use of intravenous intralipids was less common. There was a direct correlation between the dose of intralipids and the linoleoylcarnitine (C18:2) values that was independent of EGA and age at sample (R 2 = 0.3, P < .001).
Interaction Between EGA and Time
With the exception of dodecanoylcarnitine (C12), hydroxytetradecadienylcarnitine (C14OH), tetradecenoylcarnitine (C14:1), hydroxypalmitoylcarnitine (C16OH), 3-hydroxystearoylcarnitine (C18-OH), and 3-hydroxyoleoylcarnitine (C18:1-OH), there was an interaction between EGA and time for all the acylcarnitines. The pattern of change for all the acylcarnitines over time for each EGA group was similar, but the degree of change was significantly different for all acylcarnitines except those listed above. The most significant changes over time and within EGA group were for isovalerylcarnitine + methylbutyrylcarnitine (C5), octanoylcarnitine (C8), and linoleoylcarnitine (C18:2; Fig 2). All 3 of these acylcarnitines increased from day 1 to 7 and then progressively decreased for days 28 and 42. The most immature infants (23–26 weeks) were the most distinctive and were significantly different from the other 2 groups.
Summary Observations for All Analytes
Age at sampling was a more important factor associated with changes in metabolic profiles than EGA group. The difference in the median value for any specific analyte was smaller between EGA groups than between time groups.
Our study reveals a high rate of newborn screening alerts in premature infants. Twenty-one percent of the infants had values above the prespecified cutoff values used to identify infants with inborn errors of metabolism. None of the abnormal values could be explained by contamination of the blood sample, or by inappropriate collection methods. These results confirm what is generally observed in practice, that the use of cutoff values derived from healthy term infants results in a high rate of “abnormal” values (false-positives) in premature infants. 4 , 8 , 12 Figures 1, 2, and 3 graphically demonstrate the problem and reveal that the values for commonly measured analytes are often >1 SD above or below the mean value for healthy term infants.
Both EGA and time confound the ability to define normal. The most premature infants appear to be metabolically different from more mature premature infants and the metabolic profile changes over time for most (96%) of the measured analytes. This variability makes it impossible to define normative values without correcting for time and EGA. Based on the literature and the patterns we describe above, we must also correct for degree of illness and the type of nutritional support at the time of sampling. 4 , 6 , 8 , 12
Nutrition is a dynamic interaction between what is provided (glucose/carbohydrates, protein/amino acids, lipids/fatty-acids/carnitine), how it is provided (intravenous or enteral), how it is used for energy and growth, and how each nutritional component is metabolized. Just as there are adverse events associated with drug–drug interactions that relate to metabolism, the potential for adverse side effects related to nutritional–nutritional interactions is real.
We believe our data reveal precisely this type of problem in the most immature infants (23–26 EGA group; Figure 3). Both leucine-isoleucine (primary component of intravenous amino-acid solutions) and linoleoylcarnitine (linoleic and linolenic acids are key components of intralipids) are elevated on day 7 when the doses of amino acids and intralipids are highest. At the same time, isovalerylcarnitine + methylbutyrylcarnitine and octanoylcarnitine are elevated. The accumulation of these carnitines may lead to the formation of toxic organic acid metabolites and/or liver injury.
Our results reveal the complexity of understanding the impact of immaturity and postnatal age on metabolic profiles. The inherent metabolic instability manifested by the premature infant makes an abnormal result likely, even though few of these infants will ever demonstrate an inherited metabolic abnormality. Just as both high and low values of glucose can cause injury, high and low values of amino acids and acylcarnitines may cause long-term morbidity. Investigations are needed to define what values are safe in premature infants. New reference standards that include adjustments for gestational age, age at sampling, and nutritional support may better identify infants who have inborn errors of metabolism and allow us to better understand the metabolic consequences of premature birth.
Pediatrix-Obstetrix Center for Research, Education and Quality
Principal Investigators: Wesley Medical Center (KS): Barry Bloom, MD; Pinnacle Health Hospital (PA): Kimberly Costello, DO; McLeod Regional Medical Center (SC): J. Tommy Cox, MD, Joseph Harlan, MD; Tucson Medical Center (AZ): Carlos Flores, MD; Presbyterian Hospital of Dallas (TX): David W. Green, MD; INOVA Alexandria Hospital (VA): Michael Holliday, MD; Mount Carmel St. Ann’s Hospital (OH): Angela Isbel, DO; Providence Alaska Medical Center (AK): Jack Jacob, MD; Akron Children’s Hospital (OH): Anand Kantak, MD; McAllen Medical Center (TX): Joe Mendiola, MD; Covenant Medical Center (MI): Khawar Mohsini, MD; Stormont-Vail Regional Health Center (KS): Maria Bautista-Navarro, MD; Swedish Medical Center Healthone & Skyridge Medical Center (CO): Michael Paranka, MD; St. Joseph’s Hospital (TX): Hector Pierentoni, MD; Memorial Health University Medical Center (GA): Daniel Sandler, MD; Good Samaritan Medical Center (CA): Meera Sankar, MD, Kenneth Tan, MD; Charleston Area Medical Center (WV): Jayesh Shah, MD; Memorial Hospital of South Bend (IN): Vitaliy Soloveychik, MD; Carillion Clinic Children’s Hospital (VA): Ellen Szego, MD; Sacred Heart Hospital (FL): Jason Tanner, MD; Central Baptist Hospital (KY): John Walker, DO; Memorial Hospital of South Bend (IN): Robert White; East Tennessee Children’s Hospital (TN): Henry H. Wooldridge, MD.
Contributors: Tucson Medical Center (AZ): Colleen Shehan-Bakewell, NNP; Pinnacle Health Hospital (PA): Penny Barcavage, NNP; McLeod Regional Medical Center (SC): Shelley Coker, NNP; Wesley Medical Center (KS): Paula Delmore, MSN; East Tennessee Children’s Hospital (TN): Kathy Fulton RNC, NNP; St. Joseph’s Hospital (TX): Rosemary Manguba-Gibbs, MD; Carillion Clinic Children’s Hospital (VA): Dena Goldberg, PhD; Memorial Health University Medical Center (GA): Lydia Harris, MSN, NNP; Stormont-Vail Regional Health Center (KS): Renee Hunt, RNC, NNP; Swedish Medical Center Healthone (CO): Terri Molther, NNP; McAllen Medical Center (TX) and Presbyterian Hospital of Dallas (TX): Yvonne Perez, NNP.
- Accepted April 22, 2014.
- Address correspondence to Reese H. Clark, MD, Pediatrix Medical Group, Inc, 1301 Concord Terrace, Sunrise, FL 33323-2825. E-mail:
Dr Clark has contributed to the conception and design of the study. The article from first to final draft was written and edited by Dr Clark with the help of his co-authors who offered suggestions for improvement; Mrs Kelleher has contributed to the drafting and revising of the article for important intellectual content; Dr Chace has contributed to the conception and design of the study analysis and interpretation of data; Dr Spitzer has contributed to the conception and design of the study, analysis and interpretation of data, and critical revisions for important intellectual content; and all authors approved the final manuscript as submitted.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: This work was supported by Pediatrix Medical Group. No outside honorarium, grant, or other form of payment was given to anyone to produce the article.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
- Yang S ,
- Lee BS ,
- Park HW ,
- et al
- Chace DH ,
- De Jesús VR ,
- Lim TH ,
- Hannon WH ,
- Clark RH ,
- Spitzer AR
- Chace DH ,
- Kalas TA ,
- Naylor EW
- Clark RH ,
- Chace DH ,
- Spitzer AR ,
- Pediatrix Amino Acid Study Group
- Copyright © 2014 by the American Academy of Pediatrics