- Email: [email protected]
Low superior vena cava flow and neurodevelopment at 3 years in very preterm infants
ORIGINAL ARTICLES
LOW SUPERIOR VENA CAVA FLOW AND NEURODEVELOPMENT AT 3 YEARS IN VERY PRETERM INFANTS ROD W. HUNT, FRACP, NICK EVANS, DM, MRCPCH, INGRID RIEGER, FRACP, AND MARTIN KLUCKOW, FRACP, PHD
Objectives Low superior vena cava (SVC) flow is common in the first hours after very preterm birth and has a strong association with subsequent periventricular/intraventricular hemorrhage. We report the neurodevelopmental outcome at 3 years of age of very preterm babies who had serial echocardiographic studies, including measures of SVC flow, during the first 48 hours after birth. Study design A prospective observational study was performed on a cohort of 126 babies (<30 weeks), 103 of whom survived to discharge. Neurodevelopmental follow-up data, which included abnormal developmental quotient, abnormal motor score, and cerebral palsy, were available for 93% of this cohort at 3 years of age. Relations between 3-year outcome and early hemodynamic measures and clinical parameters were explored. Results After controlling for confounding variables, average SVC flow over the first 24 hours of life was significantly associated with the primary outcome of death or survival with any disability (P = .004) and with the secondary outcome of abnormal developmental quotient (P = .006). A greater number of low SVC flow readings during the first 24 hours was significantly related to death and adverse developmental outcome, but the individual lowest SVC flow was not, suggesting the importance of duration of low SVC flow. After adjustment, there was no significant association between average mean blood pressure over the first 24 hours and abnormal developmental outcome, whereas the proportion of mean blood pressure readings less than the gestational age showed a trend toward an association with death and any disability. Conclusions Low early postnatal blood flow to the upper body and brain may be one factor in the causal pathway of impaired preterm neurodevelopmental outcome. (J Pediatr 2004;145:588-92)
dvances in newborn intensive care have led to improved survival of the preterm infant. However, this has increased the burden of neurodevelopmental disability. Some of this disability is mediated through periventricular and intraventricular hemorrhage (PIVH), although survivors who do not have these lesions may still have impaired developmental progress.1 The underlying cause of developmental disability is almost certainly multifactorial and remains to be fully elucidated. An understanding of the pathophysiologic processes is essential to development of preventive strategies. Ischemia appears to have a major role in the evolution of perinatal brain injury,2 and the hypothesis that low cerebral blood flow in preterm infants plays a role in the See editorial, p 573. pathogenesis of subsequent neurodisability is not new.3 Our group has described the Doppler assessment of superior vena cava (SVC) flow as a measure of blood flow returning From RPA Newborn Care, Royal from the upper body and brain.4 We have shown in babies born before 30 weeks that low Prince Alfred Hospital and University SVC flow occurs in approximately 35%, with the nadir in flow usually occurring in the first of Sydney, Camperdown, NSW, and 12 hours after birth. There was a strong association between low SVC flow and PIVH, Department of Neonatal Medicine and University of Sydney, Royal North which occurred after the flow had improved, implicating hypoperfusion-reperfusion in the Shore Hospital, St Leonards, NSW, pathogenesis of IVH. However, not all babies with low SVC flow had ultrasound evidence Australia. 5 6 of cerebral injury. It is also clear from both magnetic resonance imaging studies and Submitted for publication Jun 26, 2003; last revision received Jun 17, follow-up that much preterm cerebral injury is not identified by ultrasound. We 2004; accepted Jun 28, 2004. hypothesized that infants who had low SVC flow would be at increased risk of cerebral Reprint requests: Dr Nick Evans, RPA injury and abnormal neurodevelopmental outcome. Newborn Care, Royal Prince Alfred
A
DQ PIVH
588
Developmental quotient Periventricular/intraventricular hemorrhage
SVC
Superior vena cava
Hospital, Missenden Road, Camperdown, NSW 2050, Australia. E-mail: [email protected]. 0022-3476/$ - see front matter Copyright ª 2004 Elsevier Inc. All rights reserved. 10.1016/j.jpeds.2004.06.056
Our goal was to describe the neurodevelopmental outcome at 3 years in preterm infants with low SVC flow and to explore the relation between that outcome and early SVC flow together with other hemodynamic and perinatal variables.
The entry criteria for this study were birth before 30 weeks’ gestation and informed parental consent. The 126 babies enrolled between March 1995 and December 1996 represented 85% of eligible babies. Refused consent (n = 5) and investigator not available (n = 19) were the reasons for nonenrolment.4,5
development was created, encompassing abnormal motor score and/or those children who had cerebral palsy on physical examination. Spastic cerebral palsy was diagnosed when there was increased tone with exaggerated stretch reflexes (claspknife effect) plus or minus brisk reflexes. Ataxic cerebral palsy was diagnosed when there was incoordination resulting in difficulties with voluntary movement with or without secondary signs of intention tremor and poor balance. When infants were unable to attend the hospital follow-up service, information was obtained by writing to the currently treating pediatricians. When this information did not include formal follow-up data, information regarding mortality and the presence of cerebral palsy was collected.
Echocardiography
Statistics
Echocardiography was performed at 5, 12, and 24 hours of age, with SVC flow measured as previously described.4,5 The measures of SVC flow included the lowest SVC flow, as a measure of the severity of low flow, and the average of the 3 SVC flow readings, as a marker of the duration of low flow. The number of echocardiograms at which SVC flow was measured below 30 mL/kg per minute was also recorded because below this level, the infants in this study were previously shown to be most likely to have development or extension of an IVH.5
The primary outcome was death or survival with any disability (combining abnormal motor development and/or abnormal DQ). Secondary outcomes were death, abnormal motor development, and abnormal DQ, analyzed as separate variables. The majority of the analyses were performed for the entire cohort to determine if SVC flow was associated with outcome. Further analyses of the variables of SVC flow were performed in which the 13 infants who died in the neonatal period from primarily respiratory or infectious causes were excluded from the cohort, leaving survivors and infants who had intensive care withdrawn because of grade 3 or 4 IVH. The goal was to evaluate the impact of low SVC flow on adverse neurodevelopmental outcome combined with death from a severe neurologic insult. Statistical analysis was performed in SAS and included logistic regression with SVC flow as a continuous variable and the number of times SVC flow was less than 30 mL/kg per minute as a categoric variable. These results are reported as odds ratios with 95% confidence intervals. With the large number of analyses being performed, a conservative value of P < .01 was used to determine statistical significance. Central Sydney Area Health Ethics Review Committee (RPA Zone) approved the study, with informed written consent.
METHODS
Blood Pressure Routine intensive care included the use of intra-arterial catheters where possible. In 106 babies (84%), physiologic data including blood pressure were downloaded every 5 minutes from the Hewlett-Packard Merlin monitors to a central database (Docvue; Hewlett-Packard, Andover, Mass) for the first 60 hours of admission. This was not possible for technical reasons in the other 20 babies, in whom hourly recordings of blood pressure were taken from the nursing charts. The average of the mean blood pressure readings over 12 or 24 hours was calculated. To assess percentage of time that an infant was hypotensive, the percentage of mean blood pressure readings that were less than the infant’s gestational age at birth in weeks was calculated.
Cranial Ultrasound Serial cranial ultrasound examinations were performed at the same time as the echocardiograms and between days 4 and 7 and on day 28. The presence of IVH was noted and graded according to the Papile classification.7
Neurodevelopmental Follow-Up Surviving infants were assessed with the Griffiths Mental Development Scales and physical examination in the Newborn Follow-up Clinic at 3 years of age. These assessments were performed by one investigator, a developmental pediatrician who was blinded to the early hemodynamic observations. Developmental quotient (DQ) and motor scores were defined as abnormal if they were greater than 2 SD below the mean. A dichotomous variable of abnormal motor Low Superior Vena Cava Flow and Neurodevelopment at 3 Years in Very Preterm Infants
RESULTS The babies had a mean gestation of 27 weeks (range, 23 to 29 weeks) and a mean birth weight of 991 grams (range, 420 to 1630 grams). Fifty-two percent were boys and 91% were inborn. Eighteen percent (n = 23) of the babies died in the neonatal period. Seven babies died from respiratory failure on day 1 or 2. Of the remaining 16 babies who died, 10 had intensive care withdrawn as a result of severe IVH, 5 of late sepsis, and 1 died late of respiratory failure.
Neurodevelopmental Outcome Of the 103 surviving infants, the results of Griffiths Developmental assessment were available for 86 infants. Information was collected from currently treating pediatricians for a further 10 infants, providing data for 96 of 103 (93%) of the surviving cohort. At 3 years of age, the median DQ was 94, 589
Table I. Results of univariate analysis of demographic and clinical variables with adverse outcome Death P value
Variable Level of maternal education Mode of delivery Sex Gestation Birth weight percentile 1-min Apgar 5-min Apgar Antenatal steroids Hypertensive disease of pregnancy Use of postnatal steroids Need for mechanical ventilation
Normal outcome
9.87 (2.15–45.4) .003 0.98 (0.83–1.17) .80 1.93 (0.77–4.85) .16 0.56 (0.41–0.76) .0003 1.0 (0.98–1.03) .073 0.73 (0.56–0.95) .02 0.74 (0.54–1.02) .07 0.69 (0.20–2.34) .55 1.86 (0.70–4.92) .21 1.16 (0.43–3.11) .77 — —
P value 0.67 1.05 0.95 1.55 1.00 1.13 1.13 1.20 0.90 0.33 0.57
(0.24–1.87) (0.91–1.21) (0.45–2.02) (1.20-2.02) (0.99–1.02) (0.94–1.40) (0.86–1.49) (0.39–3.72) (0.38–2.11) (0.13–0.85) (0.09–3.56)
Table II. Results of multivariate analysis, adjusted for gestational age, need for postnatal steroids, and level of maternal education, of average SVC flow (for each increase of 10 mL/kg per minute) over the first 24 hours of life with outcome
Death and any disability (n = 51/101) Death (n = 23/126) Abnormal DQ (n = 34/85) Abnormal motor (n = 18/86)
OR
95% CI
P value
0.72 0.66 0.64 0.74
0.58–0.90 0.50–0.88 0.47–0.88 0.53–1.03
.004 .005 .006 .07
with a range of 28 to 137. Median scores (and range) for individual subscales of the Griffiths assessment are as follows: motor, 93 (30 to 123); personal-social, 98 (29 to 142); language, 94 (26 to 188); eye-hand coordination, 81 (26 to 123); performance, 94 (25 to 136); and practical reasoning, 98 (50 to 133).
Neurodevelopmental Outcome and SVC Flow Univariate analyses were performed to examine the association between each of death, abnormal DQ and abnormal motor development, and potential confounding perinatal variables. These variables included highest level of maternal education (primary, secondary, or tertiary), mode of delivery, sex, gestation, birth weight percentile, Apgar scores at 1 and 5 minutes, antenatal steroids, hypertensive disease of pregnancy, and use of postnatal steroids. Gestation, use of postnatal steroids, and level of maternal education were significant confounders, at P < .05 on the outcome variables, and were therefore adjusted for in all subsequent multivariate regression modelling (Table I). The SVC flow averaged from 3 scans over the first 24 hours was significantly associated with adverse outcome after adjustment for confounding variables (Table II). For every 590
Hunt et al
Abnormal DQ
.44 .50 .90 .0008 .77 .21 .39 .75 .81 .02 .55
Abnormal motor
P value 0.34 0.92 0.75 0.80 0.99 1.02 1.05 0.98 0.92 3.27 0.98
(0.08–1.54) (0.78–1.08) (0.32–1.81) (0.60–1.06) (0.97–1.01) (0.83–1.25) (0.76–1.45) (0.26–3.76) (0.34–2.54) (1.21–8.81) (0.16–6.19)
.16 .29 .58 .11 .47 .89 .77 .97 .88 .02 .98
P value 4.24 (0.44-41.2) 0.97 (0.80-1.17) 0.73 (0.26-2.12) 0.83 (0.60-1.15) 1.0 (0.98-1.02) 0.97 (0.78-1.24) 0.94 (0.64-1.38) 0.57 (0.13-2.44) 1.27 (0.39-4.11) 5.29 (1.75-16.03) —
.21 .74 .57 .26 .87 .81 .74 .44 .69 .003
increase of 10 mL/kg per minute in average SVC flow, the odds of death or survival with any disability were decreased by 28% (P = .004) and the odds of having an abnormal DQ were decreased by 36% (P = .006). The odds of surviving with abnormal motor development were reduced but did not reach statistical significance (P = .07). Analysis of the cohort that included survivors and those who died of major IVH revealed a stronger association between average SVC flow and disability combined with death caused by severe IVH (P = .002). The lowest SVC flow within the first 24 hours of life was not significantly associated with any outcome measured for the entire cohort. For a 10 mL/kg per minute increase in lowest SVC flow, the odds of death or survival with any disability (with 95% CI) were 0.87 (0.68 to 1.12, P = .28). The odds of death were 1.12 (0.75 to 1.65, P = .59), the odds of abnormal DQ were 0.79 (0.58 to 1.08, P = .14), and the odds of abnormal motor score were 0.88 (0.59 to 1.32, P = .54). Nor was there a significant association with disability and death from severe IVH after exclusion of those infants who died of respiratory failure or sepsis in the neonatal period (P = .04). The risk of death or survival with any disability (Cochran-Mantel-Haenszel [CMH] x2 = 11.09, P = .0009, 1 df) increased significantly, as the number of times in the first 24 hours that the SVC flow was less than 30 mL/kg per minute increased from zero to 3 (Table III). The individual outcomes also showed significant trends. All babies with 2 or more very low SVC recordings had abnormal DQ, and all babies with 3 low recordings had abnormal DQ and abnormal motor scores. Analysis of the cohort that included survivors and those who died of major IVH showed an even stronger trend for the increasing risk of incurring disability or death caused by severe IVH (CMH x2 = 13.82, P < .001, 1 df).
Neurodevelopmental Outcome and Mean Blood Pressure All of the same potential confounders were considered, and again, only gestation, use of postnatal steroids, and level of The Journal of Pediatrics November 2004
Table III. Outcome for increasing number of SVC measures less than 30 mL/kg per minute for entire cohort Measures of SVC flow ,30 ml/kg per min
0 (n = 90)
1 (n = 19)
2 (n = 5)
3 (n = 5)
CMH x2
P value
Death and any disability (%) Death (%) Abnormal DQ (%) Abnormal motor (%)
45 11 37 16
63 30 38 31
100 60 100 50
100 60 100 100
11.09 15.77 3.05 9.60
.0009 ,.0001 .08 .002
Table IV. Results of multivariate analysis of mean blood pressure (considered as units of 10 mm Hg) and percentage of mean blood pressure readings less than gestational age (considered as 10% change of percentage) against outcome adjusted for gestation, use of postnatal steroids, and level of maternal education Average MBP over first 12 h Outcome Death and any disability (n = 51/101) Death (n = 23/126) Abnormal DQ (n = 34/85) Abnormal motor (n = 18/86)
Average MBP over first 24 h
% Readings MBP less than GA: First 12 h
% Readings MBP less than GA: First 24 h
P P OR (95% CI) value OR (95% CI) value
OR (95% CI)
P value
OR (95% CI)
P value
0.84 (0.52–1.34)
.45
0.68 (0.42–1.09)
.11
1.25 (1.03–1.51)
.02
1.48 (1.07–2.04)
.02
0.98 (0.43–2.26) 0.86 (0.52–1.44) 1.18 (0.51–2.73)
.97 .57 .69
0.49 (0.26–0.93) 0.79 (0.47–1.31) 1.01 (0.48–2.10)
.03 .36 .98
1.46 (1.15–1.84) 1.15 (0.93–1.43) 0.98 (0.74–1.30)
.002 .20 .87
1.47 (1.06–2.04) 1.48 (1.01–2.16) 1.17 (0.77–1.77)
.02 .04 .46
maternal education had significant effects in the model. The multivariate analysis was adjusted for these three potential confounders. When the average of the mean blood pressure measurements was considered for the first 12 hours of life or for the first 24 hours of life, there was no significant association between average mean blood pressure and any of the neurodevelopmental outcomes (Table IV). The number of mean blood pressure readings that were below gestational age were considered as a percentage of the total number of readings for the first 12 hours and 24 hours, reflecting percentage of time that the infant was hypotensive (Table IV). For the first 12 hours, there was a significant association with death (OR, 1.46; 1.15-1.84; P = .002) and a trend that did not reach significance for death or survival with any disability (OR, 1.25; 1.03-1.51; P = .02). When the time period of the first 24 hours was adjusted for the same confounders, there was a trend toward significance with death or survival with any disability (OR, 1.48; 1.07-2.04; P = .02) and with death alone (OR, 1.47; 1.06-2.04; P = .02). Other individual outcome variables showed no significant associations.
DISCUSSION This study follows a cohort of babies in whom SVC flows were serially measured in the first 48 hours of life and in whom a strong association between SVC flow and subsequent PIVH was previously reported.5 The majority of babies (8 of 11) with severe late PIVH died. Despite this, we now report Low Superior Vena Cava Flow and Neurodevelopment at 3 Years in Very Preterm Infants
that low SVC flow in the perinatal period is also associated with abnormal neurodevelopment. Specifically, the lower the average SVC flow, or the greater the number of times the SVC flow was <30 mL/kg per minute in the perinatal period, the more likely infants are to have a developmental score >2 SD below the mean on the Griffiths Mental Development Scales at 3 years of age. Lowest SVC flow had no significant association with outcome, suggesting that prolonged low SVC flow is a more significant risk factor than is a transient decrease in SVC flow. Previous studies reported an association between hypotension and an increased risk of adverse neurodevelopmental outcome.8,9 Once confounders were controlled for, we found no significant association between average mean blood pressure and neurodevelopmental outcome. However, there was a significant association between duration of mean blood pressure below gestational age in the first 12 hours and death and a trend toward a significant association between duration of mean blood pressure below gestational age in the first 24 hours and death or survival with any disability. These results support the concept that normal lower limits of blood pressure should relate to the maturity and or size of the baby. Cerebral blood flow can be independent of mean arterial blood pressure in preterm infants,10 and adequate cerebral perfusion pressure can be maintained across a broad range of mean arterial blood pressures. This suggests that SVC flow is a more sensitive marker of cerebral hemodynamic pathology than the more commonly used measure of blood pressure. An important 591
limitation is that therapy in this study was directed at low blood pressure, so the association between low blood pressure and adverse outcome may have been masked. The demonstration of an association between low cerebral blood flow and impaired neurodevelopment is not new. The association has been implied from studies that used ultrasound measurements of cerebral blood flow velocity.11 An increased pulsatility index of the anterior cerebral artery has been related to adverse neurodevelopmental outcome at 2 years of age. ‘‘Stolen’’ cerebral flow during extracorporeal membrane oxygenation has been linked to neurodevelopmental deficits among its survivors.12 In the only other study in which cerebral blood flow was measured directly (with 133Xe clearance), a similar association was found,3 although numbers were small (n = 19) and duration of longest follow-up was approximately 12 months. We report the largest cohort to date and used a cotside technique, which assesses the portion of the systemic blood flow that incorporates cerebral blood flow.4,5,13 In the first report of this cohort,5 ultrasound markers of brain injury were not universal for infants with low SVC flow. This study has demonstrated that ischemia, as measured by low SVC flow, is associated with poor outcome. The small number of infants in our study who had SVC flow <30 mL/kg per minute in 2 or 3 of 3 measurements either died or had adverse neurodevelopmental outcomes. This association was present despite the fact that most babies who had major IVH, after a period of low SVC flow, had intensive care withdrawn. However, postnatal ischemia is not the only mechanism by which cerebral injury is mediated. Thirty-seven percent of the babies in whom all SVC flow measures were greater than 30 mL/kg per minute had an adverse neurodevelopmental outcome. White matter injury is the consequence of a number of possible insults, and the role of free radical attack on immature oligodendroglia is well described.1 It is increasingly recognized that white matter injury may be the consequence of an antenatal insult. De Vries et al14 reported ultrasound evidence of antenatal white matter injury in 7% of surviving preterm babies and 27% of those who died. The potential for intrapartum hypoxia ischemia is also suggested by our previous observation that early and late PIVH have different risk factors, with vaginal delivery being the dominant risk factor for early IVH.5,15 It is important to emphasize that although the associations between low SVC flow and adverse outcome are statistically significant, they remain associations, and causation cannot be proven. Establishment of causation would require prevention of the low flow state leading to improvement in neurodevelopmental outcome. There have been a number of pharmacologic measures used in an attempt to reduce the
592
Hunt et al
incidence of IVH,16 although none has proven to be of consistent benefit. Inotropes such as dopamine and dobutamine do not consistently increase upper body and cerebral perfusion as measured by SVC flow.17 True benefit may only come from the prevention of the low flow state rather than from the treatment of low flow once it has been diagnosed. Prevention of the ischemic phase of the ischemia-reperfusion cycle may significantly reduce the vulnerability of the white matter to injury.
REFERENCES 1. Volpe JJ. Neurologic outcome of prematurity. Arch Neurol 1998;55: 297-300. 2. Volpe JJ. Neurology of the Newborn. 4th edition. Philadelphia, Pa: WB Saunders Co., 2001. 3. Lou HC, Skov J, Pedersen H, Koffler H. Low cerebral blood flow: a risk factor in the neonate. J Pediatr 1979;95:606-9. 4. Kluckow M, Evans N, Cummins SK. Superior vena cava flow in newborn infants: a novel marker of systemic blood flow. Arch Dis Child 2000;82:F182-7. 5. Kluckow M, Evans N. Low superior vena cava flow and intraventricular haemorrhage. Arch Dis Child 2000;82:F188-94. 6. Inder TE, Huppi PS, Warfield S, Kikinis R, Zientara G, Barnes PD, et al. Periventricular white matter injury in the premature infant is associated with a reduction in cerebral cortical gray matter volume at term. Ann Neurol 1999;46:755-60. 7. Papile LA, Burnstein J, Burnstein R, et al. Incidence and evolution of subependymal and intraventricular haemorrhage: a study of infants with birth weight less than 1500 grams. J Pediatr 1978;92:529-34. 8. Grether JK, Nelson KB, Emery ES III, et al. Prenatal and perinatal factors and cerebral palsy in very low birth weight infants. J Pediatr 1996;128: 407-14. 9. Murphy DJ, Hope PL, Johnson A. Neonatal risk factors for cerebral palsy in very preterm babies: case-control study. BMJ 1997;314:404-8. 10. Tyszczuk L, Meek J, Elwell C, Wyatt JS. Cerebral blood flow is independent of mean arterial blood pressure in preterm infants undergoing intensive care. Pediatrics 1998;102:337-41. 11. Van Bel F, den Ouden L, van de Bor M, Stijnen T, Baan J, Ruys JH. Cerebral blood-flow velocity during the first week of life of preterm infants and neurodevelopment at two years. Dev Med Child Neurol 1989;31:320-8. 12. Graziani LJ, Streletz LJ, Mitchell DG, Merton DA, Kubichek M, Desai HJ, et al. Electroencephalographic, neuroradiologic, and neurodevelopmental studies in infants with subclavian steal during ECMO. Pediatr Neurol 1994; 10:97-103. 13. Evans N. Echocardiography on neonatal intensive care units in Australia and New Zealand. J Paediatr Child Health 2000;36:169-71. 14. De Vries LS, Eken P, Groenendaal F, Rademaker KJ, Hoogervorst B, Bruinse HW. Antenatal onset of haemorrhagic and/or ischaemic lesions in preterm infants: prevalence and associated obstetric variables. Arch Dis Child 1998;78:F51-6. 15. Osborn DA, Evans N, Kluckow M. Hemodynamic and antecedent risk factors of early and late peri/intraventricular hemorrhage in premature infants. Pediatrics 2003;112:33-9. 16. Ment LR, Ehrenkranz RA, Duncan CC. Intraventricular hemorrhage of the preterm neonate: prevention studies. Semin Perinatol 1988;12:359-72. 17. Osborn D, Evans N, Kluckow M. Randomized trial of dobutamine versus dopamine in preterm infants with low systemic blood flow. J Pediatr 2002;140:183-91.
The Journal of Pediatrics November 2004
LOW SUPERIOR VENA CAVA FLOW AND NEURODEVELOPMENT AT 3 YEARS IN VERY PRETERM INFANTS ROD W. HUNT, FRACP, NICK EVANS, DM, MRCPCH, INGRID RIEGER, FRACP, AND MARTIN KLUCKOW, FRACP, PHD
Objectives Low superior vena cava (SVC) flow is common in the first hours after very preterm birth and has a strong association with subsequent periventricular/intraventricular hemorrhage. We report the neurodevelopmental outcome at 3 years of age of very preterm babies who had serial echocardiographic studies, including measures of SVC flow, during the first 48 hours after birth. Study design A prospective observational study was performed on a cohort of 126 babies (<30 weeks), 103 of whom survived to discharge. Neurodevelopmental follow-up data, which included abnormal developmental quotient, abnormal motor score, and cerebral palsy, were available for 93% of this cohort at 3 years of age. Relations between 3-year outcome and early hemodynamic measures and clinical parameters were explored. Results After controlling for confounding variables, average SVC flow over the first 24 hours of life was significantly associated with the primary outcome of death or survival with any disability (P = .004) and with the secondary outcome of abnormal developmental quotient (P = .006). A greater number of low SVC flow readings during the first 24 hours was significantly related to death and adverse developmental outcome, but the individual lowest SVC flow was not, suggesting the importance of duration of low SVC flow. After adjustment, there was no significant association between average mean blood pressure over the first 24 hours and abnormal developmental outcome, whereas the proportion of mean blood pressure readings less than the gestational age showed a trend toward an association with death and any disability. Conclusions Low early postnatal blood flow to the upper body and brain may be one factor in the causal pathway of impaired preterm neurodevelopmental outcome. (J Pediatr 2004;145:588-92)
dvances in newborn intensive care have led to improved survival of the preterm infant. However, this has increased the burden of neurodevelopmental disability. Some of this disability is mediated through periventricular and intraventricular hemorrhage (PIVH), although survivors who do not have these lesions may still have impaired developmental progress.1 The underlying cause of developmental disability is almost certainly multifactorial and remains to be fully elucidated. An understanding of the pathophysiologic processes is essential to development of preventive strategies. Ischemia appears to have a major role in the evolution of perinatal brain injury,2 and the hypothesis that low cerebral blood flow in preterm infants plays a role in the See editorial, p 573. pathogenesis of subsequent neurodisability is not new.3 Our group has described the Doppler assessment of superior vena cava (SVC) flow as a measure of blood flow returning From RPA Newborn Care, Royal from the upper body and brain.4 We have shown in babies born before 30 weeks that low Prince Alfred Hospital and University SVC flow occurs in approximately 35%, with the nadir in flow usually occurring in the first of Sydney, Camperdown, NSW, and 12 hours after birth. There was a strong association between low SVC flow and PIVH, Department of Neonatal Medicine and University of Sydney, Royal North which occurred after the flow had improved, implicating hypoperfusion-reperfusion in the Shore Hospital, St Leonards, NSW, pathogenesis of IVH. However, not all babies with low SVC flow had ultrasound evidence Australia. 5 6 of cerebral injury. It is also clear from both magnetic resonance imaging studies and Submitted for publication Jun 26, 2003; last revision received Jun 17, follow-up that much preterm cerebral injury is not identified by ultrasound. We 2004; accepted Jun 28, 2004. hypothesized that infants who had low SVC flow would be at increased risk of cerebral Reprint requests: Dr Nick Evans, RPA injury and abnormal neurodevelopmental outcome. Newborn Care, Royal Prince Alfred
A
DQ PIVH
588
Developmental quotient Periventricular/intraventricular hemorrhage
SVC
Superior vena cava
Hospital, Missenden Road, Camperdown, NSW 2050, Australia. E-mail: [email protected]. 0022-3476/$ - see front matter Copyright ª 2004 Elsevier Inc. All rights reserved. 10.1016/j.jpeds.2004.06.056
Our goal was to describe the neurodevelopmental outcome at 3 years in preterm infants with low SVC flow and to explore the relation between that outcome and early SVC flow together with other hemodynamic and perinatal variables.
The entry criteria for this study were birth before 30 weeks’ gestation and informed parental consent. The 126 babies enrolled between March 1995 and December 1996 represented 85% of eligible babies. Refused consent (n = 5) and investigator not available (n = 19) were the reasons for nonenrolment.4,5
development was created, encompassing abnormal motor score and/or those children who had cerebral palsy on physical examination. Spastic cerebral palsy was diagnosed when there was increased tone with exaggerated stretch reflexes (claspknife effect) plus or minus brisk reflexes. Ataxic cerebral palsy was diagnosed when there was incoordination resulting in difficulties with voluntary movement with or without secondary signs of intention tremor and poor balance. When infants were unable to attend the hospital follow-up service, information was obtained by writing to the currently treating pediatricians. When this information did not include formal follow-up data, information regarding mortality and the presence of cerebral palsy was collected.
Echocardiography
Statistics
Echocardiography was performed at 5, 12, and 24 hours of age, with SVC flow measured as previously described.4,5 The measures of SVC flow included the lowest SVC flow, as a measure of the severity of low flow, and the average of the 3 SVC flow readings, as a marker of the duration of low flow. The number of echocardiograms at which SVC flow was measured below 30 mL/kg per minute was also recorded because below this level, the infants in this study were previously shown to be most likely to have development or extension of an IVH.5
The primary outcome was death or survival with any disability (combining abnormal motor development and/or abnormal DQ). Secondary outcomes were death, abnormal motor development, and abnormal DQ, analyzed as separate variables. The majority of the analyses were performed for the entire cohort to determine if SVC flow was associated with outcome. Further analyses of the variables of SVC flow were performed in which the 13 infants who died in the neonatal period from primarily respiratory or infectious causes were excluded from the cohort, leaving survivors and infants who had intensive care withdrawn because of grade 3 or 4 IVH. The goal was to evaluate the impact of low SVC flow on adverse neurodevelopmental outcome combined with death from a severe neurologic insult. Statistical analysis was performed in SAS and included logistic regression with SVC flow as a continuous variable and the number of times SVC flow was less than 30 mL/kg per minute as a categoric variable. These results are reported as odds ratios with 95% confidence intervals. With the large number of analyses being performed, a conservative value of P < .01 was used to determine statistical significance. Central Sydney Area Health Ethics Review Committee (RPA Zone) approved the study, with informed written consent.
METHODS
Blood Pressure Routine intensive care included the use of intra-arterial catheters where possible. In 106 babies (84%), physiologic data including blood pressure were downloaded every 5 minutes from the Hewlett-Packard Merlin monitors to a central database (Docvue; Hewlett-Packard, Andover, Mass) for the first 60 hours of admission. This was not possible for technical reasons in the other 20 babies, in whom hourly recordings of blood pressure were taken from the nursing charts. The average of the mean blood pressure readings over 12 or 24 hours was calculated. To assess percentage of time that an infant was hypotensive, the percentage of mean blood pressure readings that were less than the infant’s gestational age at birth in weeks was calculated.
Cranial Ultrasound Serial cranial ultrasound examinations were performed at the same time as the echocardiograms and between days 4 and 7 and on day 28. The presence of IVH was noted and graded according to the Papile classification.7
Neurodevelopmental Follow-Up Surviving infants were assessed with the Griffiths Mental Development Scales and physical examination in the Newborn Follow-up Clinic at 3 years of age. These assessments were performed by one investigator, a developmental pediatrician who was blinded to the early hemodynamic observations. Developmental quotient (DQ) and motor scores were defined as abnormal if they were greater than 2 SD below the mean. A dichotomous variable of abnormal motor Low Superior Vena Cava Flow and Neurodevelopment at 3 Years in Very Preterm Infants
RESULTS The babies had a mean gestation of 27 weeks (range, 23 to 29 weeks) and a mean birth weight of 991 grams (range, 420 to 1630 grams). Fifty-two percent were boys and 91% were inborn. Eighteen percent (n = 23) of the babies died in the neonatal period. Seven babies died from respiratory failure on day 1 or 2. Of the remaining 16 babies who died, 10 had intensive care withdrawn as a result of severe IVH, 5 of late sepsis, and 1 died late of respiratory failure.
Neurodevelopmental Outcome Of the 103 surviving infants, the results of Griffiths Developmental assessment were available for 86 infants. Information was collected from currently treating pediatricians for a further 10 infants, providing data for 96 of 103 (93%) of the surviving cohort. At 3 years of age, the median DQ was 94, 589
Table I. Results of univariate analysis of demographic and clinical variables with adverse outcome Death P value
Variable Level of maternal education Mode of delivery Sex Gestation Birth weight percentile 1-min Apgar 5-min Apgar Antenatal steroids Hypertensive disease of pregnancy Use of postnatal steroids Need for mechanical ventilation
Normal outcome
9.87 (2.15–45.4) .003 0.98 (0.83–1.17) .80 1.93 (0.77–4.85) .16 0.56 (0.41–0.76) .0003 1.0 (0.98–1.03) .073 0.73 (0.56–0.95) .02 0.74 (0.54–1.02) .07 0.69 (0.20–2.34) .55 1.86 (0.70–4.92) .21 1.16 (0.43–3.11) .77 — —
P value 0.67 1.05 0.95 1.55 1.00 1.13 1.13 1.20 0.90 0.33 0.57
(0.24–1.87) (0.91–1.21) (0.45–2.02) (1.20-2.02) (0.99–1.02) (0.94–1.40) (0.86–1.49) (0.39–3.72) (0.38–2.11) (0.13–0.85) (0.09–3.56)
Table II. Results of multivariate analysis, adjusted for gestational age, need for postnatal steroids, and level of maternal education, of average SVC flow (for each increase of 10 mL/kg per minute) over the first 24 hours of life with outcome
Death and any disability (n = 51/101) Death (n = 23/126) Abnormal DQ (n = 34/85) Abnormal motor (n = 18/86)
OR
95% CI
P value
0.72 0.66 0.64 0.74
0.58–0.90 0.50–0.88 0.47–0.88 0.53–1.03
.004 .005 .006 .07
with a range of 28 to 137. Median scores (and range) for individual subscales of the Griffiths assessment are as follows: motor, 93 (30 to 123); personal-social, 98 (29 to 142); language, 94 (26 to 188); eye-hand coordination, 81 (26 to 123); performance, 94 (25 to 136); and practical reasoning, 98 (50 to 133).
Neurodevelopmental Outcome and SVC Flow Univariate analyses were performed to examine the association between each of death, abnormal DQ and abnormal motor development, and potential confounding perinatal variables. These variables included highest level of maternal education (primary, secondary, or tertiary), mode of delivery, sex, gestation, birth weight percentile, Apgar scores at 1 and 5 minutes, antenatal steroids, hypertensive disease of pregnancy, and use of postnatal steroids. Gestation, use of postnatal steroids, and level of maternal education were significant confounders, at P < .05 on the outcome variables, and were therefore adjusted for in all subsequent multivariate regression modelling (Table I). The SVC flow averaged from 3 scans over the first 24 hours was significantly associated with adverse outcome after adjustment for confounding variables (Table II). For every 590
Hunt et al
Abnormal DQ
.44 .50 .90 .0008 .77 .21 .39 .75 .81 .02 .55
Abnormal motor
P value 0.34 0.92 0.75 0.80 0.99 1.02 1.05 0.98 0.92 3.27 0.98
(0.08–1.54) (0.78–1.08) (0.32–1.81) (0.60–1.06) (0.97–1.01) (0.83–1.25) (0.76–1.45) (0.26–3.76) (0.34–2.54) (1.21–8.81) (0.16–6.19)
.16 .29 .58 .11 .47 .89 .77 .97 .88 .02 .98
P value 4.24 (0.44-41.2) 0.97 (0.80-1.17) 0.73 (0.26-2.12) 0.83 (0.60-1.15) 1.0 (0.98-1.02) 0.97 (0.78-1.24) 0.94 (0.64-1.38) 0.57 (0.13-2.44) 1.27 (0.39-4.11) 5.29 (1.75-16.03) —
.21 .74 .57 .26 .87 .81 .74 .44 .69 .003
increase of 10 mL/kg per minute in average SVC flow, the odds of death or survival with any disability were decreased by 28% (P = .004) and the odds of having an abnormal DQ were decreased by 36% (P = .006). The odds of surviving with abnormal motor development were reduced but did not reach statistical significance (P = .07). Analysis of the cohort that included survivors and those who died of major IVH revealed a stronger association between average SVC flow and disability combined with death caused by severe IVH (P = .002). The lowest SVC flow within the first 24 hours of life was not significantly associated with any outcome measured for the entire cohort. For a 10 mL/kg per minute increase in lowest SVC flow, the odds of death or survival with any disability (with 95% CI) were 0.87 (0.68 to 1.12, P = .28). The odds of death were 1.12 (0.75 to 1.65, P = .59), the odds of abnormal DQ were 0.79 (0.58 to 1.08, P = .14), and the odds of abnormal motor score were 0.88 (0.59 to 1.32, P = .54). Nor was there a significant association with disability and death from severe IVH after exclusion of those infants who died of respiratory failure or sepsis in the neonatal period (P = .04). The risk of death or survival with any disability (Cochran-Mantel-Haenszel [CMH] x2 = 11.09, P = .0009, 1 df) increased significantly, as the number of times in the first 24 hours that the SVC flow was less than 30 mL/kg per minute increased from zero to 3 (Table III). The individual outcomes also showed significant trends. All babies with 2 or more very low SVC recordings had abnormal DQ, and all babies with 3 low recordings had abnormal DQ and abnormal motor scores. Analysis of the cohort that included survivors and those who died of major IVH showed an even stronger trend for the increasing risk of incurring disability or death caused by severe IVH (CMH x2 = 13.82, P < .001, 1 df).
Neurodevelopmental Outcome and Mean Blood Pressure All of the same potential confounders were considered, and again, only gestation, use of postnatal steroids, and level of The Journal of Pediatrics November 2004
Table III. Outcome for increasing number of SVC measures less than 30 mL/kg per minute for entire cohort Measures of SVC flow ,30 ml/kg per min
0 (n = 90)
1 (n = 19)
2 (n = 5)
3 (n = 5)
CMH x2
P value
Death and any disability (%) Death (%) Abnormal DQ (%) Abnormal motor (%)
45 11 37 16
63 30 38 31
100 60 100 50
100 60 100 100
11.09 15.77 3.05 9.60
.0009 ,.0001 .08 .002
Table IV. Results of multivariate analysis of mean blood pressure (considered as units of 10 mm Hg) and percentage of mean blood pressure readings less than gestational age (considered as 10% change of percentage) against outcome adjusted for gestation, use of postnatal steroids, and level of maternal education Average MBP over first 12 h Outcome Death and any disability (n = 51/101) Death (n = 23/126) Abnormal DQ (n = 34/85) Abnormal motor (n = 18/86)
Average MBP over first 24 h
% Readings MBP less than GA: First 12 h
% Readings MBP less than GA: First 24 h
P P OR (95% CI) value OR (95% CI) value
OR (95% CI)
P value
OR (95% CI)
P value
0.84 (0.52–1.34)
.45
0.68 (0.42–1.09)
.11
1.25 (1.03–1.51)
.02
1.48 (1.07–2.04)
.02
0.98 (0.43–2.26) 0.86 (0.52–1.44) 1.18 (0.51–2.73)
.97 .57 .69
0.49 (0.26–0.93) 0.79 (0.47–1.31) 1.01 (0.48–2.10)
.03 .36 .98
1.46 (1.15–1.84) 1.15 (0.93–1.43) 0.98 (0.74–1.30)
.002 .20 .87
1.47 (1.06–2.04) 1.48 (1.01–2.16) 1.17 (0.77–1.77)
.02 .04 .46
maternal education had significant effects in the model. The multivariate analysis was adjusted for these three potential confounders. When the average of the mean blood pressure measurements was considered for the first 12 hours of life or for the first 24 hours of life, there was no significant association between average mean blood pressure and any of the neurodevelopmental outcomes (Table IV). The number of mean blood pressure readings that were below gestational age were considered as a percentage of the total number of readings for the first 12 hours and 24 hours, reflecting percentage of time that the infant was hypotensive (Table IV). For the first 12 hours, there was a significant association with death (OR, 1.46; 1.15-1.84; P = .002) and a trend that did not reach significance for death or survival with any disability (OR, 1.25; 1.03-1.51; P = .02). When the time period of the first 24 hours was adjusted for the same confounders, there was a trend toward significance with death or survival with any disability (OR, 1.48; 1.07-2.04; P = .02) and with death alone (OR, 1.47; 1.06-2.04; P = .02). Other individual outcome variables showed no significant associations.
DISCUSSION This study follows a cohort of babies in whom SVC flows were serially measured in the first 48 hours of life and in whom a strong association between SVC flow and subsequent PIVH was previously reported.5 The majority of babies (8 of 11) with severe late PIVH died. Despite this, we now report Low Superior Vena Cava Flow and Neurodevelopment at 3 Years in Very Preterm Infants
that low SVC flow in the perinatal period is also associated with abnormal neurodevelopment. Specifically, the lower the average SVC flow, or the greater the number of times the SVC flow was <30 mL/kg per minute in the perinatal period, the more likely infants are to have a developmental score >2 SD below the mean on the Griffiths Mental Development Scales at 3 years of age. Lowest SVC flow had no significant association with outcome, suggesting that prolonged low SVC flow is a more significant risk factor than is a transient decrease in SVC flow. Previous studies reported an association between hypotension and an increased risk of adverse neurodevelopmental outcome.8,9 Once confounders were controlled for, we found no significant association between average mean blood pressure and neurodevelopmental outcome. However, there was a significant association between duration of mean blood pressure below gestational age in the first 12 hours and death and a trend toward a significant association between duration of mean blood pressure below gestational age in the first 24 hours and death or survival with any disability. These results support the concept that normal lower limits of blood pressure should relate to the maturity and or size of the baby. Cerebral blood flow can be independent of mean arterial blood pressure in preterm infants,10 and adequate cerebral perfusion pressure can be maintained across a broad range of mean arterial blood pressures. This suggests that SVC flow is a more sensitive marker of cerebral hemodynamic pathology than the more commonly used measure of blood pressure. An important 591
limitation is that therapy in this study was directed at low blood pressure, so the association between low blood pressure and adverse outcome may have been masked. The demonstration of an association between low cerebral blood flow and impaired neurodevelopment is not new. The association has been implied from studies that used ultrasound measurements of cerebral blood flow velocity.11 An increased pulsatility index of the anterior cerebral artery has been related to adverse neurodevelopmental outcome at 2 years of age. ‘‘Stolen’’ cerebral flow during extracorporeal membrane oxygenation has been linked to neurodevelopmental deficits among its survivors.12 In the only other study in which cerebral blood flow was measured directly (with 133Xe clearance), a similar association was found,3 although numbers were small (n = 19) and duration of longest follow-up was approximately 12 months. We report the largest cohort to date and used a cotside technique, which assesses the portion of the systemic blood flow that incorporates cerebral blood flow.4,5,13 In the first report of this cohort,5 ultrasound markers of brain injury were not universal for infants with low SVC flow. This study has demonstrated that ischemia, as measured by low SVC flow, is associated with poor outcome. The small number of infants in our study who had SVC flow <30 mL/kg per minute in 2 or 3 of 3 measurements either died or had adverse neurodevelopmental outcomes. This association was present despite the fact that most babies who had major IVH, after a period of low SVC flow, had intensive care withdrawn. However, postnatal ischemia is not the only mechanism by which cerebral injury is mediated. Thirty-seven percent of the babies in whom all SVC flow measures were greater than 30 mL/kg per minute had an adverse neurodevelopmental outcome. White matter injury is the consequence of a number of possible insults, and the role of free radical attack on immature oligodendroglia is well described.1 It is increasingly recognized that white matter injury may be the consequence of an antenatal insult. De Vries et al14 reported ultrasound evidence of antenatal white matter injury in 7% of surviving preterm babies and 27% of those who died. The potential for intrapartum hypoxia ischemia is also suggested by our previous observation that early and late PIVH have different risk factors, with vaginal delivery being the dominant risk factor for early IVH.5,15 It is important to emphasize that although the associations between low SVC flow and adverse outcome are statistically significant, they remain associations, and causation cannot be proven. Establishment of causation would require prevention of the low flow state leading to improvement in neurodevelopmental outcome. There have been a number of pharmacologic measures used in an attempt to reduce the
592
Hunt et al
incidence of IVH,16 although none has proven to be of consistent benefit. Inotropes such as dopamine and dobutamine do not consistently increase upper body and cerebral perfusion as measured by SVC flow.17 True benefit may only come from the prevention of the low flow state rather than from the treatment of low flow once it has been diagnosed. Prevention of the ischemic phase of the ischemia-reperfusion cycle may significantly reduce the vulnerability of the white matter to injury.
REFERENCES 1. Volpe JJ. Neurologic outcome of prematurity. Arch Neurol 1998;55: 297-300. 2. Volpe JJ. Neurology of the Newborn. 4th edition. Philadelphia, Pa: WB Saunders Co., 2001. 3. Lou HC, Skov J, Pedersen H, Koffler H. Low cerebral blood flow: a risk factor in the neonate. J Pediatr 1979;95:606-9. 4. Kluckow M, Evans N, Cummins SK. Superior vena cava flow in newborn infants: a novel marker of systemic blood flow. Arch Dis Child 2000;82:F182-7. 5. Kluckow M, Evans N. Low superior vena cava flow and intraventricular haemorrhage. Arch Dis Child 2000;82:F188-94. 6. Inder TE, Huppi PS, Warfield S, Kikinis R, Zientara G, Barnes PD, et al. Periventricular white matter injury in the premature infant is associated with a reduction in cerebral cortical gray matter volume at term. Ann Neurol 1999;46:755-60. 7. Papile LA, Burnstein J, Burnstein R, et al. Incidence and evolution of subependymal and intraventricular haemorrhage: a study of infants with birth weight less than 1500 grams. J Pediatr 1978;92:529-34. 8. Grether JK, Nelson KB, Emery ES III, et al. Prenatal and perinatal factors and cerebral palsy in very low birth weight infants. J Pediatr 1996;128: 407-14. 9. Murphy DJ, Hope PL, Johnson A. Neonatal risk factors for cerebral palsy in very preterm babies: case-control study. BMJ 1997;314:404-8. 10. Tyszczuk L, Meek J, Elwell C, Wyatt JS. Cerebral blood flow is independent of mean arterial blood pressure in preterm infants undergoing intensive care. Pediatrics 1998;102:337-41. 11. Van Bel F, den Ouden L, van de Bor M, Stijnen T, Baan J, Ruys JH. Cerebral blood-flow velocity during the first week of life of preterm infants and neurodevelopment at two years. Dev Med Child Neurol 1989;31:320-8. 12. Graziani LJ, Streletz LJ, Mitchell DG, Merton DA, Kubichek M, Desai HJ, et al. Electroencephalographic, neuroradiologic, and neurodevelopmental studies in infants with subclavian steal during ECMO. Pediatr Neurol 1994; 10:97-103. 13. Evans N. Echocardiography on neonatal intensive care units in Australia and New Zealand. J Paediatr Child Health 2000;36:169-71. 14. De Vries LS, Eken P, Groenendaal F, Rademaker KJ, Hoogervorst B, Bruinse HW. Antenatal onset of haemorrhagic and/or ischaemic lesions in preterm infants: prevalence and associated obstetric variables. Arch Dis Child 1998;78:F51-6. 15. Osborn DA, Evans N, Kluckow M. Hemodynamic and antecedent risk factors of early and late peri/intraventricular hemorrhage in premature infants. Pediatrics 2003;112:33-9. 16. Ment LR, Ehrenkranz RA, Duncan CC. Intraventricular hemorrhage of the preterm neonate: prevention studies. Semin Perinatol 1988;12:359-72. 17. Osborn D, Evans N, Kluckow M. Randomized trial of dobutamine versus dopamine in preterm infants with low systemic blood flow. J Pediatr 2002;140:183-91.
The Journal of Pediatrics November 2004