Risk for Intraventricular Hemorrhage Among VLBW Infants
Risk for Intraventricular Hemorrhage Among VLBW Infants
The objective of our study was to evaluate the relationship between IVH incidence, inter-hospital transport and known IVH risk factors. Among VLBW infants, a number of clinical variables have been suggested as risk factors for IVH. Included among the potential antenatal risk factors are maternal age, prolonged rupture of membranes, mode of delivery, premature rupture of membranes and chorioamnionitis. In addition, several postnatal variables have been suggested as IVH risk factors, including postnatal resuscitation, sepsis, RDS and patent ductus arteriosus. It has also been suggested that maternal steroids, surfactant and prenatal tocolytics may offer protective effects against the development of IVH.
However, the majority of the initial studies that attempted to identify IVH risk factors incorporated only univariate analyses, and it has been suggested that many of the risk factors identified may actually be confounding variables that are simply associated with prematurity and therefore only describe a population at risk. In an attempt to control for the role of prematurity itself, Linder et al. conducted a case–control study with matched pairs of VLBW infants with and without IVH to determine independent risk factors. As a result, a number of clinical variables continued to be significantly related to IVH, including early sepsis, low hematocrit, pneumothorax, increased number of suctioning procedures and fertility treatment, whereas antenatal steroids were found to have a protective association.
In addition, previous studies have suggested a relationship between IVH and transport itself. In 1992, Yoder found increased rates of severe IVH among postnatally transported 1000–2000 g infants compared with inborn or prenatally transported infants. This prompted further research, and similarly increased IVH incidence was subsequently identified among transported premature infants between 24 and 34 weeks gestation, infants between 500 and 1200 g greater than 24 weeks gestation, and among infants less than 28 weeks gestation.
It was subsequently noted, however, that these retrospective studies incorporated only univariate analyses to compare inborn and outborn groups, again potentially ignoring the confounding influence of a number of other IVH risk factors. In an attempt to address this concern, a recent retrospective study using a large, national database compared the incidence of IVH in the first 48 h of life between inborn and transported infants <1500 g, while attempting to account for a number of possible confounding variables. After controlling for gender, race, birth asphyxia, RDS, persistent pulmonary hypertension of the newborn, pneumothorax, pulmonary hemorrhage, PDA, sepsis, necrotizing enterocolitis, maternal hypertension, chorioamnionitis, antepartum hemorrhage, cord prolapse, nuchal cord and instrument delivery, and stratifying the groups by birth weight (<1000 g, 1000 to 1499 g), the study still found significantly increased rates of both overall and severe IVH among the transported VLBW and extremely low birth weight infants. Notably, however, the study failed to account for a number of other meaningful clinical variables, including antenatal steroids, mode of delivery, gestational age and Apgar scores. In addition, stratification of infants by birth weight would not fully account for the role of prematurity as an incremental risk factor for IVH.
In one of the few prospective analyses of location status as a risk factor for IVH, Palmer et al. evaluated the association between inter-hospital transport and IVH among 894 ventilated infants born between 23 and 32 weeks gestation. As a secondary analysis within a randomized, placebo-controlled study of preemptive morphine analgesia on neonatal outcomes, they identified a significantly increased risk of both IVH and mortality among outborn infants, along with significant differences between inborn and outborn groups with respect to several obstetric and neonatal risk factors. After accounting for treatment group, gestational age and illness severity, the differences in mortality were no longer significant, but the risk of IVH remained significantly increased among outborn infants. This effect was no longer significant, however, after the model was adjusted for maternal steroid treatment before delivery. Although these findings suggest that neonatal and obstetric variables may account for the previously described differences in IVH rates between inborn and outborn premature infants, it involves only a specific subset of infants at risk for IVH. Only mechanically ventilated infants were included in the study and those with birth asphyxia, 5-min Apgar score <4, intrauterine growth restriction and/or major congenital anomalies were excluded, potentially discounting a significant population at risk.
In the current study, we evaluated the differences in IVH incidence between VLBW infants born at a tertiary facility or transported within 48 h of delivery, while attempting to account for potentially confounding known IVH risk factors. Similar to previous research, our initial univariate analysis found that infants transported from an outside facility had an increased incidence of IVH and IVH grade within 48 h of life compared with inborn infants, as well as significant differences in a number of important clinical variables. After accounting for known IVH risk factors in the multivariable analyses, however, the association between transport and IVH was no longer statistically significant. Recognizing that large P-values do not provide evidence for 'no association,' the CIs for the estimated ORs do not rule out the possibility of clinically relevant associations of transport with IVH or mortality, even after covariate adjustment. Larger studies would be needed to provide more precision to rule out such associations.
Using multivariable models, gestational age, mode of delivery, PDA and 5-min Apgar scores were all found to be significantly associated with IVH. All of these are documented IVH risk factors with potential contributions to either vessel fragility or fluctuations in systemic blood pressure that cannot be attenuated by the cerebral vasculature. Gestational age is an obvious surrogate for immature vasculature, with accompanying decreases in vessel stability and autoregulation with decreased gestational age. Vaginal delivery has also been previously found to be associated with higher rates of IVH among VLBW infants and in the current study was found to explain at least a portion of the difference in IVH and IVH grade between the groups. The increased rates of premature labor in the outborn group suggest that these infants may have been more likely to result from precipitous deliveries and that the mothers in question were not eligible for transport before delivery. Conversely, the higher incidence of PIH among the mothers of inborn infants may result in an increased cesarean delivery rate. This independent relationship between mode of delivery and IVH is a novel finding, as this variable has previously only been evaluated in univariate analyses and was not included as a potential confounder in prior studies that utilized multivariate analyses to investigate the relationship between IVH and transport.
Patent ductus arteriosus has been previously suggested as a risk factor for IVH and has been shown to alter cerebral blood flow as a result of significant ductal steal. This lesion is particularly common among premature infants and it can be expected that the resulting alterations in cerebral pressures may be poorly managed by the immature germinal matrix vasculature, resulting in an increased incidence in IVH. It is also possible that the coincidence of PDA and prematurity make it a potential surrogate marker for a population at risk for IVH rather than a risk factor itself.
In the present study, we also found that outborn infants had significantly worse initial clinical status than inborn. As a marker of clinical status during initial resuscitative efforts in the delivery room, Apgar scores are likely simply a descriptor of the VLBW population that is at increased risk for IVH. Premature infants with poor respiratory effort, poor oxygenation, depressed neurologic function and/or hemodynamic instability will surely constitute a population at risk for wide fluctuations in systemic blood pressure that will be transmitted onto their immature, fragile cerebral vasculature. In addition, infants with lower Apgar scores will likely be subjected to an increased number of interventions (for example, positive-pressure ventilation, endotracheal intubation, chest compressions) that may result in wide variations of cerebral pressure that will increase their risk of developing IVH. This independent association between clinical status, IVH and location status represents a unique finding. Although Palmer et al. found that illness severity did not fully account for the increased IVH incidence among outborn ventilated VLBW infants, additional research which found a persistent association between IVH incidence and transport did not include ratings of clinical illness severity as a potential confounding variable in the analysis.
In this study, we also attempted to identify factors associated with mortality among VLBW infants. After including those clinical variables with the strongest relationship with mortality determined via univariate analysis in a multivariable model, only gestational age and 5-min Apgar remained significant predictors of mortality. Because so many clinical variables among VLBW infants may be redundant descriptors of an overall at-risk population, the identification of independent risk factors is an important step toward identifying those variables that bear an actual relationship with mortality in this group of vulnerable patients.
It is unclear from our current study why infants born at an outside facility demonstrate worse initial clinical status than those born at a tertiary care facility. It could be suggested that the outborn infants were transported because they represent a sicker population of patients born at other hospitals and therefore are more predisposed to IVH as a group than inborn patients. In this study, however, all but one of the outborn infants were transported either specifically because of their low birth weight or for surgical evaluation. Only a single infant was transported due to poor clinical status and no patients had known IVH before transfer. In other words, although the two groups may have different initial clinic status, this difference cannot be accounted for by a selection bias.
It has been previously suggested that mothers of outborn infants represent a different population than those of inborn infants, and it is possible that the reason for delivery at one facility versus another is responsible for the clinical status that ultimately predisposes a VLBW infant to IVH. Although preterm labor as a reason for preterm delivery was significantly higher among outborn infants in our study, PPROM was more common among inborn infants, a finding consistent with previous research. This may suggest that outborn infants experience a higher rate of precipitous deliveries than inborn, precluding the possibility of adequate antibiotic or maternal steroid administration before delivery. In addition, outborn infants experienced a significantly higher rate of placental abruption than inborn, putting them at greater risk of hemodynamic stability that could predispose to the development of IVH. Nonetheless, although inborn and outborn groups differed in a number of clinical variables, the multivariable regression analysis employed in this study demonstrated that once these variables were accounted for, the relationship between transport and IVH was no longer significant.
This retrospective, observational study has limitations. Obvious ethical and logistical considerations preclude the possibility of a randomized, controlled trial of the relationship between inter-facility transport and IVH incidence among VLBW infants. We have therefore attempted to use a multivariate statistical analysis to fully account for clinical differences between the inborn and outborn groups that cannot be accounted for through randomization. Despite our attempts to include the clinical variables which have been previously suggested as significantly associated with IVH, it is possible that other relevant clinical variables were not included. For example, some research has indicated that the use of high-frequency oscillatory ventilation may have a protective effect on the development of IVH, whereas other efforts have failed to confirm this relationship. Although this represents an important target for future research efforts, the exclusion of this variable represents a limitation of the current study. In addition, we experienced significant missingness of data regarding the indication for premature delivery. Although we feel that the multiple imputation strategy and the subsequent sensitivity analyses employed represent an effective means to account for these missing data, this nonetheless remains a limitation. Finally, this study evaluates data from a single center with a regional referral base and the results may not be generalizable to other populations.
Discussion
The objective of our study was to evaluate the relationship between IVH incidence, inter-hospital transport and known IVH risk factors. Among VLBW infants, a number of clinical variables have been suggested as risk factors for IVH. Included among the potential antenatal risk factors are maternal age, prolonged rupture of membranes, mode of delivery, premature rupture of membranes and chorioamnionitis. In addition, several postnatal variables have been suggested as IVH risk factors, including postnatal resuscitation, sepsis, RDS and patent ductus arteriosus. It has also been suggested that maternal steroids, surfactant and prenatal tocolytics may offer protective effects against the development of IVH.
However, the majority of the initial studies that attempted to identify IVH risk factors incorporated only univariate analyses, and it has been suggested that many of the risk factors identified may actually be confounding variables that are simply associated with prematurity and therefore only describe a population at risk. In an attempt to control for the role of prematurity itself, Linder et al. conducted a case–control study with matched pairs of VLBW infants with and without IVH to determine independent risk factors. As a result, a number of clinical variables continued to be significantly related to IVH, including early sepsis, low hematocrit, pneumothorax, increased number of suctioning procedures and fertility treatment, whereas antenatal steroids were found to have a protective association.
In addition, previous studies have suggested a relationship between IVH and transport itself. In 1992, Yoder found increased rates of severe IVH among postnatally transported 1000–2000 g infants compared with inborn or prenatally transported infants. This prompted further research, and similarly increased IVH incidence was subsequently identified among transported premature infants between 24 and 34 weeks gestation, infants between 500 and 1200 g greater than 24 weeks gestation, and among infants less than 28 weeks gestation.
It was subsequently noted, however, that these retrospective studies incorporated only univariate analyses to compare inborn and outborn groups, again potentially ignoring the confounding influence of a number of other IVH risk factors. In an attempt to address this concern, a recent retrospective study using a large, national database compared the incidence of IVH in the first 48 h of life between inborn and transported infants <1500 g, while attempting to account for a number of possible confounding variables. After controlling for gender, race, birth asphyxia, RDS, persistent pulmonary hypertension of the newborn, pneumothorax, pulmonary hemorrhage, PDA, sepsis, necrotizing enterocolitis, maternal hypertension, chorioamnionitis, antepartum hemorrhage, cord prolapse, nuchal cord and instrument delivery, and stratifying the groups by birth weight (<1000 g, 1000 to 1499 g), the study still found significantly increased rates of both overall and severe IVH among the transported VLBW and extremely low birth weight infants. Notably, however, the study failed to account for a number of other meaningful clinical variables, including antenatal steroids, mode of delivery, gestational age and Apgar scores. In addition, stratification of infants by birth weight would not fully account for the role of prematurity as an incremental risk factor for IVH.
In one of the few prospective analyses of location status as a risk factor for IVH, Palmer et al. evaluated the association between inter-hospital transport and IVH among 894 ventilated infants born between 23 and 32 weeks gestation. As a secondary analysis within a randomized, placebo-controlled study of preemptive morphine analgesia on neonatal outcomes, they identified a significantly increased risk of both IVH and mortality among outborn infants, along with significant differences between inborn and outborn groups with respect to several obstetric and neonatal risk factors. After accounting for treatment group, gestational age and illness severity, the differences in mortality were no longer significant, but the risk of IVH remained significantly increased among outborn infants. This effect was no longer significant, however, after the model was adjusted for maternal steroid treatment before delivery. Although these findings suggest that neonatal and obstetric variables may account for the previously described differences in IVH rates between inborn and outborn premature infants, it involves only a specific subset of infants at risk for IVH. Only mechanically ventilated infants were included in the study and those with birth asphyxia, 5-min Apgar score <4, intrauterine growth restriction and/or major congenital anomalies were excluded, potentially discounting a significant population at risk.
In the current study, we evaluated the differences in IVH incidence between VLBW infants born at a tertiary facility or transported within 48 h of delivery, while attempting to account for potentially confounding known IVH risk factors. Similar to previous research, our initial univariate analysis found that infants transported from an outside facility had an increased incidence of IVH and IVH grade within 48 h of life compared with inborn infants, as well as significant differences in a number of important clinical variables. After accounting for known IVH risk factors in the multivariable analyses, however, the association between transport and IVH was no longer statistically significant. Recognizing that large P-values do not provide evidence for 'no association,' the CIs for the estimated ORs do not rule out the possibility of clinically relevant associations of transport with IVH or mortality, even after covariate adjustment. Larger studies would be needed to provide more precision to rule out such associations.
Using multivariable models, gestational age, mode of delivery, PDA and 5-min Apgar scores were all found to be significantly associated with IVH. All of these are documented IVH risk factors with potential contributions to either vessel fragility or fluctuations in systemic blood pressure that cannot be attenuated by the cerebral vasculature. Gestational age is an obvious surrogate for immature vasculature, with accompanying decreases in vessel stability and autoregulation with decreased gestational age. Vaginal delivery has also been previously found to be associated with higher rates of IVH among VLBW infants and in the current study was found to explain at least a portion of the difference in IVH and IVH grade between the groups. The increased rates of premature labor in the outborn group suggest that these infants may have been more likely to result from precipitous deliveries and that the mothers in question were not eligible for transport before delivery. Conversely, the higher incidence of PIH among the mothers of inborn infants may result in an increased cesarean delivery rate. This independent relationship between mode of delivery and IVH is a novel finding, as this variable has previously only been evaluated in univariate analyses and was not included as a potential confounder in prior studies that utilized multivariate analyses to investigate the relationship between IVH and transport.
Patent ductus arteriosus has been previously suggested as a risk factor for IVH and has been shown to alter cerebral blood flow as a result of significant ductal steal. This lesion is particularly common among premature infants and it can be expected that the resulting alterations in cerebral pressures may be poorly managed by the immature germinal matrix vasculature, resulting in an increased incidence in IVH. It is also possible that the coincidence of PDA and prematurity make it a potential surrogate marker for a population at risk for IVH rather than a risk factor itself.
In the present study, we also found that outborn infants had significantly worse initial clinical status than inborn. As a marker of clinical status during initial resuscitative efforts in the delivery room, Apgar scores are likely simply a descriptor of the VLBW population that is at increased risk for IVH. Premature infants with poor respiratory effort, poor oxygenation, depressed neurologic function and/or hemodynamic instability will surely constitute a population at risk for wide fluctuations in systemic blood pressure that will be transmitted onto their immature, fragile cerebral vasculature. In addition, infants with lower Apgar scores will likely be subjected to an increased number of interventions (for example, positive-pressure ventilation, endotracheal intubation, chest compressions) that may result in wide variations of cerebral pressure that will increase their risk of developing IVH. This independent association between clinical status, IVH and location status represents a unique finding. Although Palmer et al. found that illness severity did not fully account for the increased IVH incidence among outborn ventilated VLBW infants, additional research which found a persistent association between IVH incidence and transport did not include ratings of clinical illness severity as a potential confounding variable in the analysis.
In this study, we also attempted to identify factors associated with mortality among VLBW infants. After including those clinical variables with the strongest relationship with mortality determined via univariate analysis in a multivariable model, only gestational age and 5-min Apgar remained significant predictors of mortality. Because so many clinical variables among VLBW infants may be redundant descriptors of an overall at-risk population, the identification of independent risk factors is an important step toward identifying those variables that bear an actual relationship with mortality in this group of vulnerable patients.
It is unclear from our current study why infants born at an outside facility demonstrate worse initial clinical status than those born at a tertiary care facility. It could be suggested that the outborn infants were transported because they represent a sicker population of patients born at other hospitals and therefore are more predisposed to IVH as a group than inborn patients. In this study, however, all but one of the outborn infants were transported either specifically because of their low birth weight or for surgical evaluation. Only a single infant was transported due to poor clinical status and no patients had known IVH before transfer. In other words, although the two groups may have different initial clinic status, this difference cannot be accounted for by a selection bias.
It has been previously suggested that mothers of outborn infants represent a different population than those of inborn infants, and it is possible that the reason for delivery at one facility versus another is responsible for the clinical status that ultimately predisposes a VLBW infant to IVH. Although preterm labor as a reason for preterm delivery was significantly higher among outborn infants in our study, PPROM was more common among inborn infants, a finding consistent with previous research. This may suggest that outborn infants experience a higher rate of precipitous deliveries than inborn, precluding the possibility of adequate antibiotic or maternal steroid administration before delivery. In addition, outborn infants experienced a significantly higher rate of placental abruption than inborn, putting them at greater risk of hemodynamic stability that could predispose to the development of IVH. Nonetheless, although inborn and outborn groups differed in a number of clinical variables, the multivariable regression analysis employed in this study demonstrated that once these variables were accounted for, the relationship between transport and IVH was no longer significant.
Limitations
This retrospective, observational study has limitations. Obvious ethical and logistical considerations preclude the possibility of a randomized, controlled trial of the relationship between inter-facility transport and IVH incidence among VLBW infants. We have therefore attempted to use a multivariate statistical analysis to fully account for clinical differences between the inborn and outborn groups that cannot be accounted for through randomization. Despite our attempts to include the clinical variables which have been previously suggested as significantly associated with IVH, it is possible that other relevant clinical variables were not included. For example, some research has indicated that the use of high-frequency oscillatory ventilation may have a protective effect on the development of IVH, whereas other efforts have failed to confirm this relationship. Although this represents an important target for future research efforts, the exclusion of this variable represents a limitation of the current study. In addition, we experienced significant missingness of data regarding the indication for premature delivery. Although we feel that the multiple imputation strategy and the subsequent sensitivity analyses employed represent an effective means to account for these missing data, this nonetheless remains a limitation. Finally, this study evaluates data from a single center with a regional referral base and the results may not be generalizable to other populations.
Source...