Objectives. Whether extremely low birth weight (ELBW) infants are at risk of cerebral hypoperfusion is uncertain because key issues concerning their cerebral blood flow (CBF) and mean arterial pressure (MAP) are unresolved: (1) whether CBF is pressure-passive or autoregulated; (2) the normal level of MAP; and (3) whether inotropic drugs used to increase MAP might inadvertently impair CBF. We addressed these issues in ELBW infants undergoing intensive care.
Methods. CBF (measured by near-infrared spectroscopy) and MAP were measured in 17 infants aged 1.5 to 40.5 hours.
Results. Five infants remained normotensive (MAP 37 ± 2 mm Hg, [mean ± SEM]); twelve became hypotensive (MAP 25 ± 1 mm Hg) and were treated with dopamine (10–30 μg · kg−1 per min). CBF of hypotensive infants (14 ± 1 mL · 100 g−1 per min) was lower than the CBF of normotensive infants (19 ± mL · 100 g−1 per min). After commencement of dopamine in hypotensive infants, MAP increased (29 ± 1 mm Hg) and CBF also increased (18 ± 1 mL · 100g−1 per min). CBF was correlated with MAP in hypotensive infants before (R = 0.62) and during (R = 0.67) dopamine, but not in normotensive infants. A breakpoint was identified in the CBF versus MAP autoregulation curve of untreated infants at MAP = 29 mm Hg; no breakpoint was evident in dopamine-treated infants.
Conclusions. In ELBW infants (1) cerebral autoregulation is functional in normotensive but not hypotensive infants; (2) a breakpoint exists at ∼30 mm Hg in the CBF-MAP autoregulation curve; and (3) dopamine improves both MAP and CBF.
Autoregulation of cerebral blood flow (CBF) is a vital protection against disturbances in arterial pressure that is commonly presumed to be present in clinically stable, preterm infants and impaired in those with cardiorespiratory illness, but published data are conflicting. Since the first description of impaired autoregulation in a group of sick preterm infants in whom CBF depended on the level of mean arterial pressure (MAP), ie, CBF was pressure-passive,1 a number of other studies have also concluded that cerebral autoregulation is lost in preterm infants who become ill.2,3 However, autoregulation may be absent not only in sick preterm infants but also in those that are comparatively well.1,4 Other confounding evidence is found in studies that have shown CBF to be independent of MAP over a wide pressure range.5,6
Maintaining an adequate MAP is vital for ensuring sufficient CBF in infants with a pressure-passive cerebral circulation regardless of whether deficient cerebral autoregulation might be naturally lacking in preterm infants or lost as a result of illness. Yet the ideal range in which to maintain the MAP in extremely low birth weight (ELBW) infants is unresolved despite several studies7,8 and published guidelines9 that have attempted to define acceptable levels. Many neonatologists attempt to maintain the MAP above 30 mm Hg in preterm infants in light of reports suggesting that lower blood pressures lead to hypoperfusion and increased risk of white matter injury10 and cerebral hemorrhage.11 Clearly, definition of the lower MAP limit for autoregulation of CBF would aid in determining an acceptable level of arterial pressure, but as yet no CBF-MAP autoregulation curve has been established for the cerebral circulation of ELBW infants.
Inotropic agents are widely used for the management of MAP in ELBW infants judged to be hypotensive, with the naturally occurring endogenous catecholamine, dopamine, being the most commonly used.12,13 Although dopamine is effective in increasing MAP in preterm infants, its effectiveness in simultaneously elevating CBF remains uncertain, and the issue of whether the cerebral circulation is pressure-passive or autoregulated remains unresolved. Moreover, it is possible that dopamine exerts a direct vasoconstrictor effect on the cerebral vasculature of the ELBW infants, as it does in the immature cerebral circulation of lambs.14
This study of CBF and MAP in ELBW infants undergoing dopamine therapy addressed 3 major questions: (1) whether the cerebral circulation is pressure-passive or autoregulated over the range of arterial pressures commonly experienced by the very immature infant; (2) how the clinically significant value of MAP of 30 mm Hg relates to autoregulation of CBF; and (3) whether dopamine administration, although effective in elevating MAP, might not increment CBF or might even decrement it.
MATERIALS AND METHODS
Infants were studied over a 10-month period while receiving care in the neonatal intensive care unit at Monash Medical Centre, Melbourne, Australia. The study had ethical approval from the Monash Medical Centre Human Research and Ethics Committee and written, informed parental consent was obtained before each study. Seventeen infants (9 male and 8 female) of mean gestation 26 ± 0.4 weeks and mean body weight 772 ± 53 g were studied during the 48 hours after delivery. All infants were intubated from birth and were receiving mechanical ventilation during the study period. Clinical management of the infants including the decision to commence dopamine was at the sole discretion of the attending pediatrician.
Infants were categorized into 3 groups according to the level and stability of MAP and whether they were receiving dopamine therapy. The first group (control, n = 5) consisted of infants who were normotensive (MAP ≥30 mm Hg) and were not treated with dopamine; the second group (predopamine, n = 12) were those who exhibited MAP consistently <30 mm Hg within the first 24 hours of life. The third group comprised all infants from the predopamine group after they commenced dopamine therapy. Once commencing dopamine, this third group was recategorized as the dopamine group (n = 12). Dopamine was initially infused at a rate of 10 μg · kg−1 per min, then titrated up to a maximum of 30 μg · kg−1 per min.
Arterial blood pressure was continually measured from an indwelling peripheral or umbilical arterial catheter connected to a transducer monitoring system zeroed at the midaxillary line (Abbott Critical Care Systems Transpac IV, Abbott Australasia, Kurnell, New South Wales, Australia). The wave form and numerical values of MAP were recorded (Agilent CMS 2000, Component Monitoring System, Agilent Technologies Inc, Andover, MA) and continually displayed on a bedside monitor (Viridia CMS 2000 series, Hewlett-Packard, Waltham, MA). The same bedside monitor was also used to measure and display cardiorespiratory data including arterial oxygen saturation (Spo2).
CBF was determined by using a commercially available near-infrared spectroscopy (NIRS) spectrophotometer (NIRO-500 Hamamatsu Photonics KK, Hamamatsu City, Japan). The NIRS technique and its application to studying the cerebral circulation of newborn infants has been verified in a number of studies.15–18 The NIRS spectrophotometer produces light at 4 frequencies in the near-infrared spectrum between 700 and 1000 nm, a frequency range at which light can penetrate biological tissues including the skull and the brain. Light was transmitted via a fiber optic cable and delivered to the infant by an optode secured to the head by an elasticized bandage over the temporoparietal region; a second receiving optode was positioned exactly 4 cm away. Because CBF measurements using NIRS require a pulse oximeter configured to provide beat-to-beat Spo2 measurements, a dedicated oximeter (Nellcor N200, Nellcor Inc, Hayward, CA) was attached to the right hand of each infant. As soon as the physiologic recordings and the clinical condition of the infant were stable we induced a change in Spo2 of 2% to 5% over 2 to 5 seconds by a sudden increase in fraction of inspired oxygen. Changes in oxygenated hemoglobin (Hb) ([Hbo2]), deoxygenated Hb ([HbH]), and total Hb [Hbtot]) were recorded continuously using the NIRS and, along with Spo2, exported and stored on a portable computer. CBF was calculated from NIRS and Spo2 data offline, using equations detailed previously16 and standard software (Excel, Microsoft Corporation, Seattle, WA). We used a time integral of 7 seconds and a differential path length factor of 4.55.19 Data were accepted for analysis according to previously published strict criteria.19
During the first 48 hours of life multiple measurements of CBF were performed on each infant (range: 2–12; median: 4), for a total of 160 measurements overall. To avoid the possibility of bias arising from repeated measurements in individual infants, replicate CBF and MAP measurements for each infant were grouped into 30-minute bins and averaged. Using binned data, linear regression of CBF versus postnatal age, MAP versus postnatal age, and CBF versus MAP was performed for each of the 3 study groups (Excel, Microsoft Corporation).
CBF-MAP autoregulation curves were determined for (1) combined data of infants not receiving dopamine therapy (normotensive control infants and hypotensive infants before receiving dopamine therapy) and (2) infants receiving dopamine therapy. To identify the breakpoint in the CBF-MAP curve, we used an analytical process based on bilinear regression analysis and analysis of variance.20 In brief, CBF values are fitted with 2 regression lines intersecting at a test point (breakpoint) that is iteratively moved across the range of MAP, summing the residual sums of squares at each iteration. The breakpoint is identified as the MAP at which the residual sums of squares reaches a minimum, that is, where the intersecting lines best fit the data (Fig 1). Once the breakpoint has been identified, a first regression line for the data ≤MAPBR is fitted to the total data set; however, data points which are >MAPBR are modified to equal this breakpoint. A second line of zero slope is fitted just to the data ≤MAPBR, passing through the mean CBF for data ≤MAPBR. A significantly improved fit to the data were achieved by using bilinear regression through the breakpoint compared with a single regression line (F1,32 = 8.8; P <.01). Group means for clinical and physiologic data were contrasted using Student’s t test (normal distribution) or Mann-Whitney U test (skewed distribution, SigmaStat Version 2,SPSS Inc., Chicago, IL). In all statistical tests P <.05 was considered significant.
No gestational age, birth weight, Apgar score, Pao2, Paco2, or acid-base status differences existed between the infants of the control group and those of the hypotensive group before receiving dopamine (Table 1). All infants were clinically stable throughout the study, and none developed cranial ultrasound abnormalities within the study period.
Table 2 summarizes MAP and CBF of the ELBW infants before and after initiation of dopamine therapy. Control infants with MAP ≥ 30 mm Hg generally remained normotensive throughout the study. Infants who were deemed hypotensive by the attending physician because of MAP being consistently < 30 mm Hg recorded significantly lower MAP (25 ± 1 mm Hg) than controls (37 ± 2 mm Hg). After the initiation of dopamine treatment in these infants, MAP increased significantly in the first hour, with no additional increase in subsequent hours. Before commencing dopamine therapy, the CBF of hypotensive infants (14 ± 1 mL · 100 g−1 per min) was significantly lower than the CBF of normotensive infants (19 ± 1 mL · 100 g−1 per min). During the hour after commencement of dopamine, CBF also increased (18 ± 1 mL · 100 g−1 per min) from levels in the preceding hour.
The results of linear regression analysis are shown in Table 3. When the relationship between circulatory measurements and the time after birth was examined, MAP was not related to time in control infants nor in hypotensive infants. CBF increased with time in control infants but was unrelated to time in hypotensive infants. By contrast, both MAP and CBF were positively correlated with time in hypotensive infants after commencement of dopamine therapy. Regression of CBF with MAP revealed no correlation of CBF with MAP in the control group of infants. In hypotensive infants before dopamine therapy, CBF was positively correlated with MAP (R = 0.62). There was also a positive, linear correlation of CBF with MAP during dopamine therapy (R = 0.67). Positive correlation of CBF with MAP before and after receiving dopamine is illustrated in Fig 2.
The results of the CBF versus MAP autoregulation curve analysis are illustrated in Fig 1. In infants not receiving dopamine therapy (normotensive control infants and hypotensive infants before receiving dopamine therapy), a breakpoint in the autoregulation curve was identified at MAP equal to 29 mm Hg (Fig 1 A and B). No breakpoint was evident for infants when receiving dopamine therapy (Fig 1 C and D).
This study provides the first description of the CBF versus MAP autoregulation curve of ELBW infants. Our findings suggest that CBF is autoregulated above (and pressure-passive below) a breakpoint that averages ∼30 mm Hg in ELBW infants. Reassuringly, elevating MAP >30 mm Hg by dopamine infusion can restore CBF of hypotensive ELBW infants. On a cautionary note, CBF remains pressure-passive in the higher-pressure range so that careful monitoring of MAP levels is mandatory during dopamine infusion.
Autoregulation of the cerebral circulation is a key protection for the brain against varying arterial blood pressure, but its presence in immature infants has been left uncertain despite intensive study. Among the many potential explanations for the varied findings, the postnatal age range of the infants studied represents a major confounding element, because cerebral perfusion changes rapidly after birth. After an initial fall from fetal levels after the onset of air-breathing, CBF begins a progressive increase from 10 hours of postnatal age.21,22 Similarly, hypotension is a clinical problem that is common for the ELBW newborn in the first hours after birth.1,9 We are confident that age-related changes of CBF and MAP did not contribute to our findings, because the age range that we studied was narrow (1.5–40.5 hours). Confirming previous findings,9,21 we identified an increase of CBF, but not in MAP, with time after birth in control infants. However, no correlations existed between time after birth and CBF or MAP in hypotensive infants who had not received dopamine. As expected, significant correlation of MAP with time emerged with the institution of dopamine therapy to the hypotensive group of infants. In dopamine-treated infants, CBF too was significantly related to time, most likely reflecting the increase of MAP and persistence of the “pressure-passive” correlation of CBF with MAP that was evident before them receiving dopamine.
Recently, it was stressed that the physiologic mechanism of autoregulation is not a simple static process but rather a dynamic, rate-sensitive one.3,4 Accordingly, it has been argued that “static” or single estimates of CBF made using 133Xenon clearance23 or NIRS5 make them less sensitive to rapid changes in CBF occurring in response to transient MAP variations and thus less likely to detect that autoregulation is impaired. Yet, more rapid-responding methods of assessing autoregulation are also in disagreement. Intact autoregulation has been found in 1 assessment of preterm infants based on the coherence in an ultralow frequency range of 0 to 0.01 Hz of variations of intravascular oxygenation, measured by NIRS with changes of MAP.3 Another, based on rapid CBF velocity changes occurring in association with spontaneous, transient changes of arterial pressure, found autoregulation to be absent in neurologically normal preterm infants, but present in their term counterparts.4
As a unifying suggestion, we propose that in addition to the well-known postnatal age-related differences in the level of CBF, there may be significant age-related differences in the effectiveness of the intrinsic vasodilatory mechanism. We propose that the speed of autoregulation is least in the preterm brain, which seems able to respond to persisting or static MAP changes, as we and others5 have found, and also to slow dynamic MAP changes,3 but not to rapid dynamic changes.4 With increasing age, the ability to respond to rapid MAP changes seems also to be expressed in the full-term brain.4 A parallel hierarchy of sensitivity to hypoxic-ischemic insult may also contribute to inconsistency of published findings. Our proposal is that the ability for cerebral vessels to respond to rapid, dynamic MAP changes is most vulnerable to failure, and the first lost in asphyxiated infants, whereas the ability to respond to persistent, static MAP changes is more robust. This hierarchy of sensitivity is in keeping with the findings that dynamic autoregulation is commonly absent in the sick preterm infant, whereas static autoregulation is present, as we and others5 have found.
Absence of autoregulation, hypotension and low CBF is closely related to the subsequent appearance of cerebral injury, but the exact pathogenesis of injury is poorly understood.3,24–26 Likewise, the possibilities for therapeutic interventions aimed at restoring autoregulation and preventing injury remain to be determined. Complicating the issue is the possibility that therapeutic agents directed at maintaining blood pressure above an arbitrary level blood pressure may themselves be harmful.5 As recently discussed,27 there may be 2 circulatory mechanisms of hemorrhagic injury. We suggest that the first, a low CBF-reperfusion injury that impairs autoregulation, may be the result of a persisting insult arising from failure of the very slow or static autoregulation. The second, the transmission of pressure to a pressure passive circulation, may be due to impairment of dynamic autoregulation. The distinction between dynamic and static autoregulation merits additional evaluation, because it may aid resolution of issues relating to pathogenesis and prevention of injury.
Clearly, knowledge of the level of an infant’s MAP in relation to the breakpoint in the CBF-MAP curve will aid in establishing whether autoregulation has failed or whether the infant’s MAP places it in the pressure-passive range of the autoregulation curve. Decision is complicated by the fact that the MAP of preterm infants is normally located close to the lower breakpoint of the CBF-MAP autoregulatory curve, as it is in the cerebral circulation of the immature lamb.28 We identified a breakpoint at 29 mm Hg in the autoregulatory curve of ELBW infants not receiving dopamine. In this situation, infants whose MAP fell below the clinically significant level of 30 mm Hg10,11 exhibited pressure-passive CBF and low absolute CBF levels. When MAP fell in the range > 30 mm Hg, higher plateau levels of CBF were found. Absolute levels of CBF in these infants averaged 19 mL · 100 g−1 per min, with a range of 12 to 27 mL · 100 g−1 per min, values similar to those made with NIRS5 and other techniques such as 133Xenon clearance17 and positron emission tomography.29
The clinical implication of the CBF-MAP autoregulation curve we have described is that CBF may be considered normal in ELBW infants whose MAP exceeds 30 mm Hg and abnormal in those whose MAP is persistently < 30 mm Hg. These findings are consistent with previous clinical observations that blood pressures < 30 mm Hg lead to hypoperfusion and increased risk of white matter injury10 and cerebral hemorrhage11 and lend support to the clinical practice of maintaining MAP >30 mm Hg. There is also a recommendation, in guidelines published by the British Association of Perinatal Medicine and the Royal College of Physicians,9 for MAP in mm Hg to be maintained greater than or equal to the infant’s gestational age in weeks. This might seem to infer that infants of gestational age <29 weeks lie below the CBF-MAP curve breakpoint. However, our data encompass a gestational age range of 23 to 30 weeks and our estimate of the breakpoint has insufficient precision with respect to age to draw this inference. Additional research is needed to discriminate between these recommendations.
We found no evidence in our study group of hypoxic ischemic encephalopathy, no seizures, no intraventricular hemorrhage, and no serious brain injury on serial ultrasound examination. Nevertheless, all infants were sufficiently at risk to require intubation, mechanical ventilation, and arterial catheterization. Twelve of the 17 infants studied were considered by their attending neonatologists to be persistently hypotensive and in need of therapy to raise MAP; in these infants, dopamine therapy was managed independent of the investigators.
Inotropic agents are widely used for the management of hypotension, with the naturally occurring, endogenous catecholamine dopamine the most commonly used.12,13 Dopamine is effective in increasing MAP in preterm infants by β-receptor inotropic stimulation of the myocardium and by α-receptor mediated vasoconstriction in the peripheral circulation.30,31 However whether dopamine concurrently raises CBF has been uncertain.13,32 In our study group, dopamine used clinically in a dose range of 10 to 30 μg · kg−1 per min was effective in simultaneously raising both MAP and CBF. Although concern has been expressed that pressor agents such as dopamine may inappropriately constrict the cerebral circulation of preterm infants,5 just as in the immature brain of lambs,14 we found no evidence of cerebral vasoconstriction in these ELBW infants.
It has been debated whether inotropic therapy is effective in raising cerebral perfusion in hypotensive infants.5 Reassuringly, dopamine elevated CBF of the hypotensive infants in this study into the range of normotensive infants within 1 hour of commencing infusion, provided MAP was concurrently increased. On a cautionary note, we found no evidence of autoregulation in ELBW infants receiving dopamine. In these infants, MAP rose into the autoregulatory range, but our analysis revealed no evidence of a breakpoint in the CBF-MAP relation (Fig 1). Rather, CBF seemed to continue along the pressure-passive curve that was evident in the infants before commencing dopamine therapy. Because constant delivery is a critical aspect of dopamine’s effectiveness in raising MAP, these data underline the need to strictly monitor MAP to ensure smooth infusion of inotropic agents by using appropriate delivery systems.33
Although NIRS enables quantitation of CBF5 and has allowed us to assemble the CBF-MAP autoregulation curve for ELBW infants, this application of NIRS is not easily applied and is unsuitable for routine bedside investigations. CBF and its autoregulation seem to be very labile in preterm infants, making simpler, continuous assessments of cerebrovascular autoregulation desirable, such as the coherence of NIRS-derived estimates of intravascular oxygenation with changes of MAP.3,4 Assessments of regional cerebral intravascular oxygenation are also possible using NIRS.34 Additionally, cerebral venous saturation (CVso2), a measure of cerebral tissue oxygenation, has been estimated using NIRS in both term35 and preterm infants.36 Measurement of CVso2 is potentially more useful than CBF measurement and intravascular measurements, because it offers direct information on cerebral tissue oxygenation and the balance of cerebral oxygen transport against utilization. Although several aspects of NIRS have been validated,5,15,17,19 NIRS-derived measurements of CVso2 await additional laboratory evaluation.
We gratefully acknowledge the assistance of Dr Daniel Grant, Dr Malcolm Wilkinson, Dr Philip Berger, and Ms Jennene Wild for their valuable comments and contributions to the data analysis.
- Accepted May 26, 2004.
- Address correspondence to Charles P. Barfield, MBBS, FRACP, Ritchie Centre for Baby Health Research, Monash Institute of Reproduction and Development, Monash University and Newborn Services, Monash Medical Centre, Victoria 3168, Australia. E-mail:
No conflict of interest declared.
- ↵Tsuji M, Saul JP, du Plessis A, et al. Cerebral intravascular oxygenation correlates with mean arterial pressure in critically ill premature infants. Pediatrics.2000;106 :625– 632
- ↵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(2 pt 1) :337– 341
- ↵Joint Working Group of the British Association of Perinatal Medicine and the Research Unit of the Royal College of Physicians. Development of audit measures and guidelines for good practice in the management of neonatal respiratory distress syndrome. Arch Dis Child.1992;67(10 spec no) :1221– 1227
- ↵Miall-Allen VM, de Vries LS, Whitelaw AG. Mean arterial blood pressure and neonatal cerebral lesions. Arch Dis Child.1987;62 :1068– 1069
- ↵Wagerle LC, Kurth CD, Roth RA. Sympathetic reactivity of cerebral arteries in developing fetal lamb and adult sheep. Am J Physiol.1990;258(5 pt 2) :H1432– H1438
- ↵Wickramasinghe YA, Livera LN, Spencer SA, Rolfe P, Thorniley MS. Plethysmographic validation of near infrared spectroscopic monitoring of cerebral blood volume. Arch Dis Child.1992;67(4 spec no) :407– 411
- ↵Elwell CE. Measurement of cerebral blood flow using NIRS. In: A Practical Users Guide to Near Infra Red Spectroscopy. Hamamatsu City, Japan: Hamamatsu Photonics KK, 2003:69–85
- ↵Volpe JJ. Hypoxic-ischemic encephalopathy. In: Volpe JJ, ed. Neurology of the Newborn. Philadelphia, PA: W.B. Saunders Company; 2001:217–276
- ↵Wyatt J, Meek J. Commentary on cerebral intravascular oxygenation correlates with mean arterial pressure in critically ill premature infants. Pediatrics.2000;106 :828
- ↵Roze JC, Tohier C, Maingueneau C, Lefevre M, Mouzard A. Response to dobutamine and dopamine in the hypotensive very preterm infant. Arch Dis Child.1993;69(1 spec no) :59– 63
- ↵Capes DF, Dunster KR, Sunderland VB, McMillan D, Colditz PB, McDonald C. Fluctuations in syringe-pump infusions: association with blood pressure variations in infants. Am J Health Syst Pharm.1995;52 :1646– 1653
- Copyright © 2004 by the American Academy of Pediatrics