Abstract
Background
Hypoalbuminemia has been proven to be a powerful predictor of mortality in adult patients. However, prognostic value of serum albumin in neonates is not clear.
Objective
To assess the relationship between serum albumin level within the first day of life and outcome in preterm infants born before 32 weeks of gestation.
Methods
The study was conducted prospectively in Baskent University Hospital between October 2008 and November 2009. Patients were divided by gestational age into two groups as below or of 28 weeks and above 28 weeks. Then serum albumin percentile groups were established within each gestational age group and were defined as <25, 25–75, and >75 percentile groups by combining percentile groups between the two gestational age groups. Three serum albumin percentile groups were compared regarding neonatal outcomes.
Results
A total of 199 infants with mean birth weight of 1,272 ± 390 g and mean gestational age of 29.2 ± 2.2 weeks were admitted to the study. The mean serum albumin level was 30.6 ± 4.7 g/l for all patients. The mean serum albumin levels were 25.5 ± 3.8, 30.1 ± 2.7, and 35.3 ± 3.7 g/l for <25, 25–75, and >75 percentile groups, respectively. Prevalence of infants with respiratory distress syndrome and prevalence of infants with sepsis and mortality were significantly higher in <25 percentile group. Logistic regression analysis showed that serum albumin <25 percentile and birth weight were independent predictive variables of mortality. Albumin concentrations lower than 27.2 g/l was associated with mortality, with a sensitivity of 71% and a specificity of 86%.
Conclusion
Low serum albumin level within the first day of life is an independent predictor of mortality in preterm infants.
Keywords: hypoalbuminemia, preterm, neonate, mortality
Introduction
Diminished circulating level of albumin is common in critically ill patients and hypoalbuminemia has been proven to be a powerful predictor of mortality in adult patients 1. Serum albumin level has been added as one of the component parameters in the Acute Physiology and Chronic Health Evaluation (APACHE) III score 2. However, prognostic value of serum albumin in neonates and children has been studied in relatively few numbers of studies 3, 4, 5, 6, 7, 8. Morris et al. evaluated the relationship between serum albumin and mortality in preterm infants in a retrospective study and have shown an association between lowest serum albumin level and mortality in very low birth weight infants 4. The aim of this study is to establish the relationship between serum albumin level within the first day of life and clinical outcome in preterm infants born before 32 weeks of gestation.
Materials and Methods
The study was conducted prospectively in the Neonatal Intensive Care Unit of Baskent University Hospital between October 2008 and November 2009. Preterm infants who were born before 32 weeks of gestation were included in the study. Gestational age in completed weeks was determined by obstetric measures (last menstrual period, standard obstetric parameters, and ultrasonography). Infants with major congenital anomalies and infants transferred from other hospitals after the first 24 h of life were excluded. This study was approved by Ethical Committee of Baskent University Hospital and parental informed consent was obtained in all cases prior to study. Perinatal data (maternal preeclempsia, diabetes, premature rupture of membranes, preterm labor and use of antenatal steroids, gestational age, gender, birth weight, growth restriction at birth) and followup of neonatal and outcome data (respiratory distress syndrome, first blood gasses and lactate level, duration of mechanical ventilation and oxygen support, sepsis, patent ductus arteriosus, intracranial hemorrhage, periventricular leukomalacia, necrotizing enterocolitis, retinopathy of prematurity, bronchopulmonary dysplasia, length of stay, mortality) were assessed at discharge or at death.
Serum albumin concentration was measured by bromocresol green binding method with an automated clinical chemistry analyzer (Roche/Hitachi 912). Serum albumin level is known to be correlated with gestational age 9, 10. So, it is difficult to define the accurate normal ranges for serum albumin level in preterm infants. Therefore, we could not classify the patients into the low, normal, and high serum albumin level groups. We planned to group the patients according to serum albumin percentiles. However as the albumin levels change with the gestational ages, we tried to analyze serum albumin levels of the patients according to their gestational ages. There were not adequate patients for each week of gestational age. So neonates were first divided by gestational age into two groups as the infants with gestational age above 28 weeks and those with gestational age of 28 weeks and below. Then, serum albumin percentiles were established within each gestational age group. Finally, the results were analyzed by combining the percentile groups between the two gestational age groups. Patients with serum albumin level lower than 25 percentile were considered as lower serum albumin group and defined as <25 percentile group. Patients with serum albumin level between 25–75 percentiles were considered as normal serum albumin group and defined as 25–75 percentile group. Patients with serum albumin level higher than 75 percentile were considered as higher serum albumin group and defined as >75 percentile group. In each percentile group, there were patients with gestational age below or of 28 weeks and above 28 weeks. These three serum albumin percentile groups were compared regarding neonatal outcomes.
Statistical analyses were performed with the SPSS 15.0 for Windows (SPSS, Chicago, IL). A chi‐square test was used to analyze categorical variables. Student's t‐test and Mann–Whitney U test were used to analyze continuous parametric and nonparametric variables, respectively. A logistic regression model was constructed for assessing the independent effects on mortality. The results are presented as prevalence odd ratios (OR) and their 95% confidence intervals.
Results
A total of 271 neonates born before 32 weeks of gestation were screened during the study. Thirteen infants with major congenital anomalies and 59 infants transferred from other hospitals after the first 24 h of life were excluded from the study. Finally, a total of 199 infants, with 98 females (49%) and 101 males (51%) were admitted to this study over the 1‐year period. The mean birth weight was 1,272 ± 390 g and the mean gestational age was 29.2 ± 2.2 weeks. The mean serum albumin level was 30.6 ± 4.7 g/l and serum albumin level was correlated with gestational age (r = 0.529, P = 0.01; Fig. 1). Serum albumin percentiles established according to gestational age groups are shown in Table 1. Infants were classified into three serum albumin groups according to their albumin percentiles as <25, 25–75, and >75 percentile groups. The mean serum albumin levels were 25.5 ± 3.8 g/l, 30.1 ± 2.7 g/l, and 35.3 ± 3.7 g/l for <25, 25–75, and >75 percentile groups, respectively. Perinatal data of the infants according to the serum albumin groups are presented in Tables 2 and 3. No significant difference was found in birth weight, gestational age, and gender of infants and in proportion of infants with growth restriction at birth between serum albumin groups. Proportion of infants of preeclamtic mothers was significantly lower, but proportion of infants without antenatal steroid exposure and proportion of infants of mothers with preterm labor was significantly higher in <25 percentile serum albumin group compared with 25–75 and >75 percentile groups. Neonatal and outcome data of the infants according to the serum albumin groups are presented in Tables 4 and 5. Prevalence of infants with respiratory distress syndrome and prevalence of infants with early‐onset sepsis and mortality were significantly higher in <25 percentile serum albumin group compared with 25–75 and >75 percentile groups. Logistic regression analysis showed that serum albumin lower than 25 percentile and birth weight were independent predictive variables of mortality (Table 6). Serum albumin concentrations lower than 27.2 g/l was associated with mortality, with a sensitivity of 71% and a specificity of 86% and serum lactate levels greater than 3.5 mmol/l was associated with mortality, with a sensitivity of 72% and a specificity of 68%. Figure 2 shows ROC curves for serum albumin concentrations (AUC = 0.83, P < 0.001), serum lactate levels (AUC = 0.27, P < 0.001), and mortality.
Figure 1.
Mean serum albumin levels according to gestational age.
Table 1.
Serum Albumin Percentile Cutoff Points in Two Gestational Age Groups
| Percentile cutoff points | ||||||||
|---|---|---|---|---|---|---|---|---|
| 5 | 10 | 25 | 50 | 75 | 90 | 95 | ||
| Serum albumin levels (g/l) | Gestational age ≤28 weeks | 19.3 | 22.4 | 25.1 | 27.4 | 30.1 | 31.4 | 32.9 |
| Gestational age >28 weeks | 25.8 | 27.4 | 29.3 | 32.7 | 34.7 | 37.0 | 38.4 | |
Table 2.
Perinatal (Continuous) Data of Infants According to the Serum Albumin Percentile Groups
| Serum albumin percentile groups | ||||
|---|---|---|---|---|
| <25 percentile | 25–75 percentile | >75 percentile | P | |
| Birth weight (g)a | 1,208 ± 391 | 1,290 ± 366 | 1,300 ± 434 | 0.404 |
| Gestational age (week)a | 28.86 ± 2.36 | 29.29 ± 1.98 | 29.31 ± 2.36 | 0.473 |
| pHa | 7.19 ± 0.14 | 7.26 ± 0.10 | 7.25 ± 0.11 | 0.004 |
| Base Excessa (mmol/l) | –9.64 ± 5.69 | –6.81 ± 4.02 | –5.80 ± 5.50 | 0.001 |
| Lactatea (mmol/l) | 4.74 ± 3.84 | 3.20 ± 2.21 | 3.34 ± 1.89 | 0.003 |
| Total number of patients | 51 (25.6%) | 97 (48.8%) | 51(25.6%) | 199 |
Mean ± SD.
Table 3.
Perinatal (Categorical) Data of Infants According to the Serum Albumin Percentile Groups
| Serum albumin percentile groups | ||||
|---|---|---|---|---|
| <25 percentile | 25–75 percentile | >75 percentile | P | |
| Gender | ||||
| Female | 21/51 (41.2%) | 55/97 (56.7%) | 22/51 (43.1%) | 0.120 |
| Male | 30/51 (58.8%) | 42/97 (43.3%) | 29/51 (56.9%) | |
| Small for gestational agea | 7/51 (13.7%) | 14/97 (14.4%) | 9/51 (17.6%) | 0.836 |
| Preeclampsia | 8/51 (15.7%) | 24/97 (24.7%) | 21/51 (41.1%) | 0.006 |
| Gestational diabetes | 6/51 (11.8%) | 9/97 (9.3%) | 3/51 (5.9%) | 0.569 |
| Premature rupture of membranes | 13/51 (25.5%) | 26/97 (26.8%) | 8/51 (15.7%) | 0.276 |
| Preterm labor | 37/51 (72.5%) | 71/97 (73.2%) | 27/51 (52.9%) | 0.013 |
| Antenatal steroid usageb | 20/51 (39.2%) | 53/97 (54.6%) | 38/51 (74.5%) | 0.001 |
| Total number of patients | 51 (25.6%) | 97 (48.8%) | 51 (25.6%) | 199 |
Small for gestational age: infants with a birth weight below the tenth percentile for gestational age.
Two doses of 12 mg betamethasone given intramusculary, 24 h apart.
Table 4.
Neonatal Outcome (Continuous) Data of Infants According to Albumin Percentile Groups
| Serum albumin percentile groups | ||||
|---|---|---|---|---|
| <25 percentile | 25–75 percentile | >75 percentile | P | |
| Duration of mechanical ventilation (day)a | 5.2 ± 8.2 | 3.8 ± 8.0 | 3.3 ± 7.7 | 0.489 |
| Duration of oxygen therapy (day)a | 11.6 ± 13.9 | 13.1 ± 20.0 | 13.4 ± 27.3 | 0.886 |
| Duration of hospital stay (day)a | 24.6 ± 18.0 | 30.7 ± 22.7 | 26.1 ± 26.7 | 0.224 |
Mean ± SD.
Table 5.
Neonatal Outcome (Categorical) Data of Infants According to Albumin Percentile Groups
| Serum albumin percentile groups | |||||
|---|---|---|---|---|---|
| <25 percentile | 25–75 percentile | >75 percentile | Total | P | |
| Respiratory distress syndrome | 34/51 (66.6%) | 60/97 (61.8%) | 22/51 (43.1%) | 116 (58.3%) | 0.033 |
| Early‐onset sepsisa | 28/51 (54.9%) | 49/97 (50.5%) | 15/51 (29.4%) | 92 (%46.2) | 0.016 |
| Patent ductus arteriosus | 11/51 (21.6%) | 15/97 (15.5%) | 7/51 (13.7%) | 33 (16.6%) | 0.521 |
| Intracranial hemorrhage | 6/51 (11.7%) | 3/97 (3.3%) | 2/51 (3.9%) | 11 (5.5%) | 0.058 |
| Broncopulmonary dysplasia | 8/51 (15.6%) | 13/97 (13.4%) | 8/51 (15.6%) | 29 (14.6%) | 0.602 |
| Necrotizing enterocolitis | 6/51 (11.8%) | 12/97 (12.4%) | 3/51 (5.8%) | 21 (10.5%) | 0.450 |
| Retinopathy of premature | 3/51 (5.8%) | 4/97 (4.1%) | 2/51 (3.9%) | 9 (4.5%) | 0.663 |
| Periventricular leukomalacia | 1/51 (1.9%) | 5/97 (5.1%) | 3/51 (5.9%) | 9 (4.5%) | 0.405 |
| Mortality | 17/51 (33.3%) | 10/97 (10.3%) | 8/51 (15.6%) | 35 (17.6%) | 0.003 |
| Total | 51 (25.6%) | 97 (48.8%) | 51 (25.6%) | 199 | |
Early‐onset sepsis: clinical or culture proven sepsis that present in the first 3 days of life.
Table 6.
Logistic Regression Analysis of Gestational Age, Birth Weight, Albumin Level <25 Percentile, BE, and Lactate Level for the Mortality Estimation
| 95% CI | ||||
|---|---|---|---|---|
| P | OR | Lower limit | Upper limit | |
| Gestational age | 0.066 | 0.743 | 0.541 | 1.020 |
| Birth weight | 0.006 | 0.997 | 0.994 | 0.999 |
| Albumin <25 percentile | 0.024 | 3.400 | 1.178 | 9.813 |
| Base Excess | 0.173 | 0.915 | 0.805 | 1.040 |
| Lactate | 0.637 | 1.055 | 0.845 | 1.318 |
Figure 2.
ROC curve of serum albumin level and lactate level for the estimation of mortality. AUC for serum albumin level = 0.83 (P < 0.001) and AUC for lactate level = 0.27 (P < 0.001).
Discussion
Serum albumin is considered as a marker of disease severity in critically ill patients and low albumin levels have been shown to be associated with poor outcome and mortality in adult patients. In a meta‐analysis reviewing 90 cohort studies with 291,433 total patients, hypoalbuminemia was shown to be a powerful, concentration‐dependent, independent risk factor for mortality 1. However, significance of hypoalbuminemia in critically ill children is evaluated in a relatively few numbers of studies. In the study by Durward et al., hypoalbuminemia was common on admission to pediatric intensive care unit but was not an independent predictor of mortality; mean serum albumin concentrations were similar between survivors and nonsurvivors 6. Horowitz et al. found that hypoalbuminemia was a significant marker of mortality in children admitted to pediatric intensive care unit 8. Leite et al., while investigating the role of serum albumin as a predictor of clinical outcome in children who undergo operative correction of congenital heart defects, reported that hypoalbuminemia may be associated with higher risks of mortality in pediatric patients undergoing cardiac surgery 7. Wong et al. evaluated the association between serum albumin and death among pediatric patients with end‐stage renal disease and concluded that low serum albumin at dialysis initiation was an important marker of mortality risk in pediatric patients 5. Prognostic value of serum albumin in neonates was studied in two studies. Kenny et al. investigated the prevalence of hypoalbuminemia in 37 surgical neonates receiving parenteral nutrition and its prognostic effect on outcome 3. Fifteen infants were preterm and 22 infants were term. Study patients were divided into hypoalbuminemia and normal albumin groups according to their lowest measured serum albumin value and hypoalbuminemia was defined as serum albumin level <27.9 g/l. Mortality was higher among hypoalbuminemic infants. Morris et al. evaluated the relationship between lowest recorded serum albumin level and mortality in 107 preterm very low birth weight infants 4. There was a significant association between lowest albumin levels and mortality that persisted when adjusted for birth weight and gestational age. However, this study was a retrospective study and albumin measurements were not recorded at standardized times. In the present study, we have searched the relationship between serum albumin level within the first day of life and the mortality in preterm infants born before 32 weeks of gestation. We have compared the mortality in preterm infants who were classified into three serum albumin groups according to their albumin percentiles because serum albumin level is known to be correlated with gestational age in preterm infants. Mortality was significantly higher in <25 percentile serum albumin group compared with 25–75 and >75 percentile groups. We have also demonstrated that serum albumin lower than 25 percentile and birth weight were independent predictive variables of mortality. High serum lactate levels were previously ascertained as a predictor of mortality in preterm infants 11, 12. We have compared the area under the curve (AUC) for albumin and lactate as a predictor of mortality. We have found that the sensitivities of hypoalbuminemia and hyperlactatemia were similar but specificity of hypoalbuminemia was better than hyperlactatemia for predicting mortality.
Albumin has many important physiological functions, including maintenance of colloid osmotic pressure, buffering, protein binding and transport, free radical scavenging, platelet function inhibition, antitrombosis, and alterations of vascular permeability 4. A variety of pathophysiological processes cause hypoalbuminemia in preterm infants, which contributes to poor outcome. During critical illness, increased capillary permeability causes escape of albumin into the extravascular compartment 13. Albumin synthesis capacity is limited in ill preterm neonates and albumin synthesis also decreases in response to reprioritization of protein synthesis in favor of acute reactant proteins. Inflammatory mediators can also directly inhibit the gene transcription responsible for albumin synthesis 14.
Although inadequate protein and caloric intake decreases the rate of albumin synthesis, serum albumin is not a reliable marker of nutritional status in preterm infants 15, 16. Likewise, no significant difference was found in proportion of infants with growth restriction at birth between serum albumin groups, in our study. We have postulated that albumin metabolism is not influenced from fetal malnutrition.
In this study, we did not search the pathophysiologic mechanisms contributing to high mortality in hypoalbuminemic infants. Nevertheless, prevalence of infants with respiratory distress syndrome and prevalence of infants with early‐onset sepsis were significantly higher in <25 percentile serum albumin group. Low albumin concentrations in infants with respiratory distress syndrome may have been related to increased alveolar capillary permeability and increased protein leakage into alveolar space. Bland et al. first reported the value of a low cord‐blood protein determination as a screening tool of selection of high risk infants for respiratory distress syndrome 17. Moison et al. analyzed the postnatal changes in plasma levels of albumin in preterm infants with respiratory distress syndrome and chronic lung disease 18. Hypoalbuminemia was present at birth in all patients with respiratory disease. The albumin levels began to rise after day 4 in babies who recovered from respiratory distress syndrome but hypoalbuminemia persisted in babies with chronic lung disease. In our study, proportion of infants without antenatal steroid exposure was also significantly higher in <25 percentile serum albumin group. This finding supports the data that antenatal steroid therapy improves the cytostructural maturation of lungs and decreases pulmonary protein leakage. On the other hand, albumin is assumed as a negative acute phase protein. Inflammation mediated by cytokines leads to decreased synthesis of albumin and causes albumin redistribution associated with increased capillary permeability 13, 14, 19, 20. Proportion of infants of mothers with preterm labor was significantly higher in <25 percentile serum albumin group in our study. Preterm labor may have been related with fetal systemic inflammatory response syndrome, which leads to early neonatal sepsis and low serum albumin levels.
In contrast to pediatric and adult studies, we were unable to document higher long‐term neonatal morbidites in hypoalbuminemic preterm infants born before 32 weeks of gestation. Duration of mechanical ventilation and oxygen support, prevalence of patent ductus arteriosus, intracranial hemorrhage, periventricular leukomalacia, necrotizing enterocolitis, retinopathy of prematurity, bronchopulmonary dysplasia, and length of hospital stay were similar between serum albumin groups. However, we did not follow the subsequent changes in serum albumin levels of patients, so we could not establish an association between low serum albumin levels and long‐term neonatal morbidities.
In conclusion, this study showed that low serum albumin level within the first day of life is an independent predictor of mortality in preterm infants born before 32 weeks of gestation. Serum albumin level may be considered for incorporation in illness severity scores based on physiologic indexes such as CRIB 2 and SNAPPE 2. Further studies are needed to determine whether including serum albumin values in the predictive models could improve accurate risk estimates.
References
- 1. Vincent JL, Dubois MJ, Navickis RJ, Wilkes MM. Hypoalbuminemia in acute illness: Is there a rationale for intervention? A meta‐analysis of cohort studies and controlled trials. Ann Surg 2003;237(3):319–334. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Knaus WA, Wagner DP, Draper EA, et al. The APACHE III prognostic system. Risk prediction of hospital mortality for critically ill hospitalized adults. Chest 1991;100(6):1619–1636. [DOI] [PubMed] [Google Scholar]
- 3. Kenny SE, Pierro A, Isherwood D, Donnell SC, van Saene HK, Lloyd DA. Hypoalbuminaemia in surgical neonates receiving parenteral nutrition. J Pediatr Surg 1995;30(3):454–457. [DOI] [PubMed] [Google Scholar]
- 4. Morris I, McCallion N, El‐Khuffash A, Molloy EJ. Serum albumin and mortality in very low birth weight infants. Arch Dis Child Fetal Neonatal Ed 2008;93(4):F310–F312. [DOI] [PubMed] [Google Scholar]
- 5. Wong CS, Hingorani S, Gillen DL, et al. Hypoalbuminemia and risk of death in pediatric patients with end‐stage renal disease. Kidney Int 2002;61(2):630–637. [DOI] [PubMed] [Google Scholar]
- 6. Durward A, Mayer A, Skellett S, et al. Hypoalbuminaemia in critically ill children: Incidence, prognosis, and influence on the anion gap. Arch Dis Child 2003;88(5):419–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Leite HP, Fisberg M, de Carvalho WB, de Camargo Carvalho AC. Serum albumin and clinical outcome in pediatric cardiac surgery. Nutrition 2005;21(5):553–558. [DOI] [PubMed] [Google Scholar]
- 8. Horowitz IN, Tai K. Hypoalbuminemia in critically ill children. Arch Pediatr Adolesc Med 2007;161(11):1048–1052. [DOI] [PubMed] [Google Scholar]
- 9. Cartlidge PH, Rutter N. Serum albumin concentrations and oedema in the newborn. Arch Dis Child 1986;61(7):657–660. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Zlotkin SH, Casselman CW. Percentile estimates of reference values for total protein and albumin in sera of premature infants (less than 37 weeks of gestation). Clin Chem 1987;33(3):411–413. [PubMed] [Google Scholar]
- 11. Hussain F, Gilshenan K, Gray PH. Does lactate level in the first 12 hours of life predict mortality in extremely premature infants? J Paediatr Child Health 2009;45(5):263–267. [DOI] [PubMed] [Google Scholar]
- 12. Nadeem M, Clarke A, Dempsey EM. Day 1 serum lactate values in preterm infants less than 32 weeks gestation. Eur J Pediatr 2010;169(6):667–670. [DOI] [PubMed] [Google Scholar]
- 13. Fleck A, Raines G, Hawker F, et al. Increased vascular permeability: A major cause of hypoalbuminaemia in disease and injury. Lancet 1985;1(8432):781–784. [DOI] [PubMed] [Google Scholar]
- 14. Brenner DA, Buck M, Feitelberg SP, Chojkier M. Tumor necrosis factor‐alpha inhibits albumin gene expression in a murine model of cachexia. J Clin Invest 1990;85(1):248–255. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Nadeau L, Forest JC, Masson M, Morrissette I, Larivière F, Caron M. Biochemical markers in the assessment of protein‐calorie malnutrition in premature neonates. Clin Chem 1986;32(7):1269–1273. [PubMed] [Google Scholar]
- 16. Polberger SK, Fex GA, Axelsson IE, Räihä NC. Eleven plasma proteins as indicators of protein nutritional status in very low birth weight infants. Pediatrics 1990;86(6):916–921. [PubMed] [Google Scholar]
- 17. Bland RD. Cord‐blood total protein level as a screening aid for the idiopathic respiratory‐distress syndrome. N Engl J Med 1972;287(1):9–13. [DOI] [PubMed] [Google Scholar]
- 18. Moison RM, Haasnoot AA, van Zoeren‐Grobben D, Berger HM. Plasma proteins in acute and chronic lung disease of the newborn. Free Radic Biol Med 1998;25(3):321–328. [DOI] [PubMed] [Google Scholar]
- 19. Castell JV, Gómez‐Lechón MJ, David M, Fabra R, Trullenque R, Heinrich PC. Acute‐phase response of human hepatocytes: Regulation of acute‐phase protein synthesis by interleukin‐6. Hepatology 1990;12(5):1179–1186. [DOI] [PubMed] [Google Scholar]
- 20. Ruot B, Papet I, Bechereau F, et al. Increased albumin plasma efflux contributes to hypoalbuminemia only during early phase of sepsis in rats. Am J Physiol Regul Integr Comp Physiol 2003;284(3):R707–R713 [DOI] [PubMed] [Google Scholar]