OBJECTIVE: With male gender as a strong predictor of cardiovascular instability, we hypothesized that gender-specific differences in circulating carbon monoxide levels contributed to dysregulated microvascular function in preterm male infants.
METHODS: Infants born at 24 to 34 weeks of gestation (N = 84) were studied in a regional tertiary neonatal unit. Carboxyhemoglobin levels were measured through spectrophotometry in umbilical arterial blood and at 24, 72, and 120 hours after birth. Microvascular blood flow was determined through laser Doppler flowmetry.
RESULTS: Carboxyhemoglobin levels demonstrated a strong inverse relationship with gestational age (r = −0.636; P < .001) and were higher in boys (P = .032). Repeated-measures analysis of variance showed a significant decrease in arterial carboxyhemoglobin levels over time (P < .001), with significant between-subjects effects for gestational age (P = .011) and gender (P = .025). Positive correlations with microvascular blood flow at 24 hours of age (r = 0.495; P < .001) and 120 hours of age (r = 0.548; P < .001) were observed. With controlling for gestational age, carboxyhemoglobin levels at 72 hours were greater for infants who died in the first week of life (P = .035).
CONCLUSIONS: The gestational age- and gender-specific differences in carboxyhemoglobin levels and the relationship with dysregulated microvascular blood flow, a state related to greater illness severity and hypotension, are novel findings not confined solely to sick preterm infants. Both inducible heme oxygenase-dependent and non–heme oxygenase-dependent pathways may initially play a central role in carbon monoxide production, inducing pathophysiologic processes in a gender-specific manner.
Low blood pressure is a common hemodynamic sequela of preterm birth, particularly in the first days of life, with gender-specific differences in cardiorespiratory parameters being well recognized. Significant gender-related differences in arterial blood pressure at 12 to 24 hours, with increased need for inotropic support and incidence of “resistant hypotension” in male infants, also having been reported.1 Death related to respiratory distress syndrome (RDS) exhibits the greatest male excess.2 The physiologic basis for these gender-related differences in the newborn period has not been investigated extensively.
Nitric oxide (NO) and its second messenger guanosine 3′,5′-cyclic monophosphate (cGMP) mediate vascular smooth muscle relaxation and play an active role in the regulation of blood pressure in preterm infants.3 In preterm infants with RDS, however, increased cGMP levels occur in the absence of upregulation of the NO pathway,3,4 which highlights the likely roles of other cGMP-generating pathways for this high-risk population. Carbon monoxide (CO)-mediated increases in cGMP have been demonstrated to play significant roles in the regulation of blood pressure and cardiac function in both physiologic and pathologic conditions5 and therefore have been proposed to contribute to low blood pressure in infants with RDS who are undergoing ventilation.6
Approximately 85% of endogenous CO production can be ascribed to the metabolism of heme by inducible heme oxygenase, which is also called heat shock protein 32. This has been demonstrated to be increased in response to oxidative stress and inflammation.7 CO binds competitively to hemoglobin, in preference to oxygen, to form carboxyhemoglobin, which represents an in vivo sink for CO. Carboxyhemoglobin levels have been demonstrated to be elevated significantly during stress, sepsis, and shock in both adult and pediatric populations.8,9
Decreased peripheral vascular resistance has been proposed to contribute to low blood pressure in preterm infants.10 We demonstrated significant relationships between dysregulated peripheral microvascular tone, illness severity, and arterial blood pressure in preterm infants.11 In addition, we showed that dysregulation of peripheral blood flow demonstrated gender-specific differences conferring a male disadvantage.12 We propose that this dysregulation in peripheral blood flow in male preterm infants may be attributable to changes in microvascular tone through a CO-dependent rather than a NO-dependent pathway. The aim of this study was to investigate the relationships between arterial carboxyhemoglobin levels, endogenous NO production (determined through serial measurements of 24-hour NO excretion in urine), and peripheral microvascular blood flow, with respect to gender and neonatal outcomes.
The study was approved by the local ethics committee. After informed parental consent was obtained, premature infants with gestational ages of ≤34 weeks who were admitted to the NICU between April 2005 and May 2006 were studied prospectively. Exclusion criteria included birth at another hospital (with subsequent transfer for ongoing neonatal intensive care) and chromosomal or major congenital abnormalities.
Umbilical cord blood and arterial blood samples were obtained at 24, 72, and 120 hours of age, for determination of carboxyhemoglobin levels. Peripheral microvascular blood flow in the skin was determined at 24, 72, and 120 hours of life, as described previously, by using a Periflux 5001 laser Doppler system (Perimed, Järfälla, Sweden).
Carboxyhemoglobin levels were determined through spectrophotometry by using an ABL700 blood gas analyzer (Radiometer, Copenhagen, Denmark); the carboxyhemoglobin concentration was expressed as a proportion of the total hemoglobin concentration. A zeroing calibration was performed at least every 4 hours. Serum bilirubin levels were measured, as part of clinical care, by quality-ensured hospital laboratories. Clinical illness severity on day 1 of life was assessed with the Clinical Risk Index for Babies II scoring system.13
Twenty-four hour urine samples were collected by using disposable diapers (Kimberley Clark, Sydney, Australia)14 containing pure cellulose pads (Kimberley Clark). All preterm infants were placed on appropriately sized diapers, which were closed appropriately to limit evaporation. The common interval for changing diapers was 4 hours. Weighing of the diapers before and after collection allowed exact calculation of the 24-hour urine output. Urine was extracted by using a manual press and was pooled over 24 hours. Because humidity can contribute to diaper weight, the degree and length of time in humidity were recorded for weight adjustments for NO metabolite and creatinine concentrations according to volume. The recalculated 24-hour urine measures were used for all analyses.15 The total nitrate/nitrite concentration in pooled neonatal urine samples, an accepted reliable measure of endogenous whole-body NO production,16 was determined with a commercial nitrate/nitrite colorimetric assay (Cayman Chemical Co, Ann Arbor, MI). The intraassay variation was 7.2%. Creatinine levels were determined with a commercial creatinine colorimetric assay (Cayman Chemical Co). Results were expressed as the total nitrate/nitrite concentration/urinary creatinine concentration ratio.
Data are summarized as means and SEMs or median and interquartile ranges where appropriate. Statistical analyses of differences between groups were performed with the Mann-Whitney test or Fisher's exact test. The relationship between carboxyhemoglobin levels and microvascular blood flow was assessed through partial correlation controlling for the influence of gestational age. Correlations between non–parametrically distributed variables were assessed by using Spearman's r correlation. Two-way analysis of variance was used to investigate the influence of gestational age and gender on umbilical arterial carboxyhemoglobin levels, with posthoc analysis with independent-sample t tests and a Bonferroni adjustment to the degree of freedom. Temporal changes in arterial carboxyhemoglobin levels and 24-hour urinary NO metabolite levels were evaluated through analysis of variance for repeated measures, with posthoc analysis with paired t tests and a Bonferroni adjustment to the degree of freedom.
Eighty-four neonates were included in the study (43 girls and 41 boys). Patient information and clinical characteristics are shown in Table 1. There were no differences between male and female infants with respect to gestational age, birth weight percentile, Apgar score, or Clinical Risk Index for Babies II score. Patent ductus arteriosus requiring treatment with indomethacin occurred more frequently for boys than for girls (P = .01).
Requirements for ventilation and measures of respiratory function are shown in Table 2. All infants of gestational age of <29 weeks received surfactant (Curosurf, 120 mg/kg; Chiesi Farmaceutica, Parma, Italy) at the time of admission from the labor ward. Normal clinical practice entailed immediate extubation and continuous positive airway pressure support after surfactant administration if the infant was in clinically stable condition. No difference between the genders with respect to any respiratory parameter at 24, 72, or 120 hours of age was observed. Seven boys and 4 girls did not survive to discharge.
Umbilical arterial carboxyhemoglobin levels were related inversely to gestational age (r = −0.636; P < .0001) (Fig 1A). In addition, carboxyhemoglobin levels were lower for girls than for boys (F1,76 = 2.89; P = .032) (Fig 1B). Repeated-measures analysis of variance with gestational age and gender as independent factors and time (24, 72, and 120 hours) as the repeated measure was conducted with the arterial carboxyhemoglobin level data. A main effect for time was observed (F2,106 = 13.88; P < .001), with posthoc analysis (paired t tests with a Bonferroni correction to the α level) demonstrating lower carboxyhemoglobin levels at 120 hours, compared with both 24 hours (P < .001) and 72 hours (P < .001).
There was no difference between carboxyhemoglobin levels at 24 and 72 hours of age. Significant between-subjects effects were also observed for gestational age (F1,53 = 7.02; P = .011) and gender (F1,53 = 5.31; P = .025). Carboxyhemoglobin levels were higher for the most-preterm infants. In addition, girls were found to have lower carboxyhemoglobin levels, compared with boys (Fig 2A). To investigate the role of hemolysis in CO production, repeated-measures analysis of variance with gestational age and gender as independent factors and time (24, 72, and 120 hours) as the repeated measure revealed a significant increase in serum bilirubin levels with increasing postnatal age (F2,73 = 45.67; P < .001). However, serum bilirubin levels were not influenced by either gestational age or infant gender.
Repeated-measures analysis of variance with gestational age and gender as independent factors and time (24, 72, and 120 hours) as the repeated measure was also conducted with the 24-hour urinary NO metabolite data. A main effect of time was demonstrated (F2,62 = 2.65; P = .048), with posthoc testing revealing increases in urinary NO metabolite levels between 24 and 120 hours (P = .004) and between 72 and 120 hours (P = .012) but no significant difference between 24 and 72 hours (P = .332) (Fig 2B). Urinary NO metabolite levels were not observed to be influenced by either gestational age or infant gender.
Partial correlation analysis was used to explore the relationship between arterial carboxyhemoglobin levels and basal microvascular blood flow. There were strong positive correlations between carboxyhemoglobin levels and baseline microvascular blood flow at 24 hours (r = 0.495; P < .001) and 120 hours (r = 0.548; P < .001) (Fig 3), with elevated levels of carboxyhemoglobin being associated with greater basal microvascular blood flow. Inspection of the 0-order correlation (24 hours: r = 0.52; 120 hours: r = 0.56) suggested that controlling for gestational age had very little effect on the strength of the relationship between these 2 variables. At 72 hours, the correlation between carboxyhemoglobin levels and microvascular blood flow did not reach significance (r = 0.202; P = .119).
Eight infants, all with gestational ages of <28 weeks, died in the first week of life. With controlling for gestational age, carboxyhemoglobin levels were found to be higher at 72 hours (P = .035) and 120 hours (P = .063) of age for infants who died in the first week of life (Fig 4). No significant difference was observed at 24 hours of age.
Both at birth and during the first days of life, arterial carboxyhemoglobin levels were found to demonstrate gestational age- and gender-specific differences, being higher in boys and in the most-premature infants. In addition, significant relationships between carboxyhemoglobin levels and microvascular perfusion were observed. CO-mediated upregulation of plasma cGMP levels has been proposed to result in increased vasodilation of the systemic vascular bed in preterm infants.6 The findings of the current study suggest that gender-specific differences in CO production by preterm infants may contribute to the greater incidence of hemodynamic instability observed for male infants as a result of dysregulated microvascular tone in the immediate newborn period.
The temporal changes in arterial carboxyhemoglobin levels are in keeping with the clinical time course of acute neonatal RDS, a pathophysiologic condition known to be associated with elevated carboxyhemoglobin and cGMP levels,6 and with changes in vascular tone and mean arterial blood pressure observed after preterm birth.17 The opposite was observed for 24-hour urinary NO metabolite levels, a finding similar to that of Krediet et al,3 who examined NO production in preterm infants with and without RDS in the first week of life. CO-mediated upregulation of plasma cGMP levels has been proposed as an alternate mechanism that results in increased vasodilation of the systemic vascular bed in preterm infants. Taken together, these findings suggest a greater role for CO in the mechanisms maintaining vascular homeostasis in the transitional circulation of preterm neonates, a process previously thought to be principally dependent on NO synthesis after upregulation of calcium-dependent, constitutive NO synthase.4
Total body stores of CO are influenced by 2 principle processes, that is, the production and excretion of endogenous CO and the respiratory exchange of exogenous CO.18 Under physiologic conditions, CO is produced primarily as a result of hemolysis, with oxidative degradation of heme producing equimolar amounts of biliverdin, CO, and iron.19 Biliverdin is then reduced immediately by the cytoplasmic enzyme biliverdin reductase to form bilirubin.20 Unlike carboxyhemoglobin levels, however, serum bilirubin levels were found to increase significantly from 24 to 120 hours of age, an observation that was not affected by gestational age or infant gender.
The current data support potential contributions from other CO-producing processes in the perinatal and immediate neonatal period, pathways that demonstrate gender-related differences. Oxidative stress upregulates both the principle source of CO, that is, enzyme-inducible heme oxygenase, and non–heme oxygenase-related pathways that result in CO production (in particular, lipid peroxidation).21 Although the associations between carboxyhemoglobin and serum bilirubin levels and carboxyhemoglobin levels and mortality rates have been investigated in the adult literature,22 the data are commonly reported with adjustment for gender, because of the male excess of morbidity and death observed for critically ill adults.23
The evidence for greater susceptibility to oxidative stress in male subjects is limited. In preterm infants, however, periventricular/intraventricular hemorrhage, necrotizing enterocolitis, chronic lung disease, and retinopathy of prematurity all share a pathogenesis related to excess oxidative stress24 relative to antioxidant defenses,25 with clear gender-specific differences in the incidence and severity of these common complications of preterm birth.26,27 Elevated umbilical levels of 8-isoprostane, a product of lipid peroxidation that has been used as a measure of free radical exposure and injury and is known to be associated with increased neonatal morbidity and mortality rates, do not demonstrate any relationship with gender.28 Similarly, no fetal gender-specific differences in levels of lipid and protein oxidation products in pregnancies complicated by maternal asthma are observed.29
The current data may be explained by differences in CO excretion. CO excretion has been shown to be dependent on alveolar ventilation, the diffusing capacity of the lung, and the mean oxygen tension in the pulmonary capillaries.18 Although a greater incidence of acute neonatal RDS in the most-preterm infants may explain some of the gestational age-related differences in arterial carboxyhemoglobin levels, differences in CO excretion would not explain the significantly higher umbilical arterial carboxyhemoglobin levels observed for male infants. No gender-related differences in ventilatory requirements or respiratory parameters up to day 5 of life were observed, but carboxyhemoglobin levels remained significantly higher in male patients during this period.
Male infants had a significantly higher incidence of patent ductus arteriosus requiring pharmacologic treatment. Ductus arteriosus patency is associated with greater ventilatory and oxygen requirements30 and a transitional circulation characterized by low systemic blood flow.31 By 72 hours of age, however, only 10% of male patients continued to require mechanical ventilation. Although CO has a role in the prenatal patency of the ductus arteriosus,32 there are no literature data on the role of CO in postnatal patency or on pulmonary CO excretion. The respective contributions of CO production and CO excretion in the immediate newborn period have yet to be fully elucidated; however, the relationships between carboxyhemoglobin levels and microvascular blood flow add to the limited data on the physiologic actions of inducible heme oxygenase and CO in preterm infants and provide insights into possible pathophysiologic mechanisms underlying dysregulated vascular tone in the immediate newborn period.
CO-induced vasorelaxation is endothelium independent, is not a consequence of tissue hypoxia, and is independent of any adrenergic effect. In preterm infants, CO has been reported to be related indirectly to the increased incidence of hypotension associated with RDS6 and has been implicated in the pathophysiologic processes of periventricular/intraventricular hemorrhage,33 through augmented cGMP production. In addition, CO has been demonstrated to inhibit vascular tone by blocking the cytochrome P450-mediated production of the vasoconstrictor endothelin-134 and to mediate vascular relaxation through the activation of potassium channels in a tissue-dependent manner.35 These findings indicate that different signaling pathways mediate vasodilation induced by CO and that the relative contributions of these pathways depend on the vascular source.
Dysregulation of microvascular tone has been demonstrated to contribute to hemodynamic instability and clinical illness severity after preterm birth.11 In addition, gender-specific differences in microvascular function in premature neonates were described recently.12 The relationships between arterial carboxyhemoglobin levels and microvascular blood flow demonstrated in the current study are novel and provide one potential explanation for the increased incidence of hypotension and its clinical sequelae in male patients. This may explain a mechanism for some of our previously reported microvascular findings in premature infants. It is of note that these observations were not confined to the most-preterm, sick infants but involved infants up to 34 weeks of gestational age with varying degrees of respiratory illness, including none at all.
Significant differences in carboxyhemoglobin levels were observed with respect to death, with elevated carboxyhemoglobin levels at 72 hours, but not earlier, being predictive of death in the first week of life. Such a finding may be explained by the presence of multiple redundant pathways in vascular control, using many systemically and locally produced vasoactive mediators in the systemic and pulmonary circulations, of which CO is merely one. Only recently have the potential roles of these vascular mediators been investigated in the perinatal setting. The underlying pathophysiologic processes that lead to ongoing elevated CO production in preterm neonates are yet to be fully elucidated.
Finally, if CO-mediated increases in microvascular blood flow are in the causal pathway for increased neonatal morbidity and death, then a number of potential novel therapeutic approaches may be available for these infants. A competitive metalloporphyrin inhibitor of heme oxygenase, tin mesoporphyrin, has been tested extensively in both term and preterm populations for its effects on the prevention of bilirubin metabolism, but any vascular effect, protective or otherwise, has not been considered in detail.36
Both umbilical and subsequent arterial carboxyhemoglobin levels were found to exhibit both gestational age- and gender-specific differences, being higher in infants of greatest prematurity and male gender. Although carboxyhemoglobin levels decreased during the first 5 postnatal days, they remained elevated in infants who did not survive the first week of life. This observation may be explained by the strong relationship between arterial carboxyhemoglobin levels and microvascular dysregulation, a state known to be related to greater neonatal illness severity and incidence of hypotension. Future research should be directed toward characterization of the CO-producing pathways in premature infants and the mechanisms underlying the observed gender-related differences.
Financial support was provided by the Hunter Children's Research Foundation.
- Accepted March 30, 2009.
- Address correspondence to Ian M.R. Wright, FRACP, Mother and Babies Research Centre, Hunter Medical Research Institute, John Hunter Hospital, Lookout Road, New Lambton Heights, Newcastle, New South Wales 2310, Australia. E-mail:
Financial Disclosure: The authors have indicated they have no financial relationships relevant to this article to disclose.
What's Known on This Subject:
The vasotransmitter CO is a microvascular dilator through cGMP. Carboxyhemoglobin levels are increased in sick neonates. Microvascular dilation is associated with illness severity in premature infants and is associated independently with male gender; boys fare worse as sick premature infants.
What This Study Adds:
Carboxyhemoglobin levels are related to gestational age and gender in sick and well preterm infants and to microvascular blood flow. Inducible heme oxygenase- and non–heme oxygenase-dependent pathways may be involved in CO production, with the potential for novel therapeutic interventions.
- ↵Chen YH, Yet SF, Perrella MA. Role of heme oxygenase-1 in the regulation of blood pressure and cardiac function. Exp Biol Med (Maywood).2003;228 (5):447– 453
- ↵van Bel F, Latour V, Vreman HJ, et al. Is carbon monoxide-mediated cyclic guanosine monophosphate production responsible for low blood pressure in neonatal respiratory distress syndrome? J Appl Physiol.2005;98 (3):1044– 1049
- ↵Shi Y, Pan F, Li H, et al. Carbon monoxide concentration in paediatric sepsis syndrome. Arch Dis Child.2003;88 (10):889– 890
- ↵Stark MJ, Clifton VL, Wright IMR. Microvascular flow, clinical illness severity and cardiovascular function in the preterm infant. Arch Dis Child Fetal Neonatal Ed.2008;93 (4):F271– F274
- ↵Hegyi T, Anwar M, Carbonne MT, et al. Blood pressure ranges in premature infants, part II: the first week of life. Pediatrics.1996;97 (3):336– 342
- ↵Tenhunen R, Marver HS, Schmid R. The enzymatic conversion of heme to bilirubin by microsomal heme oxygenase. Proc Natl Acad Sci USA.1968;61 (2):748– 755
- ↵Yoshida T, Noguchi M, Kikichi G. The step of carbon monoxide liberation in the sequence of heme degradation catalyzed by the reconstituted microsomal heme oxygenase system. J Biol Chem.1982;257 (16):9345– 9348
- ↵Sullivan JL, Newton RB. Serum antoixidant activity in neonates. Arch Dis Child.1988;63 (7 spec no):748– 750
- ↵Darlow BA, Hutchinson JL, Simpson JM, Henderson-Smart DJ, Evans N. Risk factors for significant retinopathy of prematurity in babies of the Australian and New Zealand Neonatal Network. Presented at the meeting of the Perinatal Society of Australia and New Zealand; March 15th–18th, 2004; Sydney, Australia
- ↵Henderson-Smart DJ, Hutchinson JL, Donoghue DA, Evans NJ, Simpson JM, Wright I. Prenatal predictors of chronic lung disease in very preterm infants. Arch Dis Child Fetal Neonatal Ed.2006;91 (1):F40– F45
- ↵Yanowitz TD, Yao AC, Pettigrew KD, Werner JC, Oh W, Stonestreet BS. Postnatal hemodynamic changes in very-low-birthweight infants. J Appl Physiol.1999;87 (1):370– 380
- ↵Kluckow M, Evans N. Low superior vena cava flow and intraventricular haemorrhage in preterm infants. Arch Dis Child Fetal Neonatal Ed.2000;82 (3):F188– F194
- Copyright © 2009 by the American Academy of Pediatrics