- Email: [email protected]
Effects of neonatal polycythemia and hemodilution on capillary perfusion
Effects of neonatal polycythemia and hemodilution on capillary perfusion M i k a e l N o r m a n , MD, B e n g t F a g r e l l , MD, PhD, a n d P e t e r Herin, MD, PhD From the Department of Pediatrics, Karolinska and St. G~ran's Hospitals, and the Department of Medicine, Karolinska Hospital, Karolinska Institute, Stockholm, Sweden
The effects of neonatal polycythemia on nutritive capillary perfusion were investigated by a television microscopy technique. The capillary blood flow velocity in skin was measured in 12 neonates with polycythemia before and after treatment with hemodilution, and in 13 healthy control infants. The capillary blood flow velocity in the patients was 0.11 (0.02 to 0.34) mm/sec and in the healthy control infants 0.30 (0.17 to 0.44) mm/sec (p <0.01, median and range values). In relation to the absolute hematocrit c h a n g e after treatment (range, - 2 0 % to 0%), the capillary blood flow velocity increased nonlinearly (range, +733% to -14%; r = -0.98; p <0.001). The postnatal a g e was found to contribute significantly to the variation in results--the neonates with polycythemia studied during the first day of life had a very slow skin capillary circulation and responded to treatment with a more pronounced increase in capillary blood flow velocity than did the older patients. This in vivo model for capillary perfusion indicates that an insufficient microcirculation may be involved in the pathophysiology responsiblefor the morbidity associated with neonatal polycythemia. (J PEDIATR1992;121:103-8) Neonatal polycythemia is a common clinical problem that is associated with functional failure of various vital organ systems. 1, 2 The hyperviscosity of the blood in these patients is believed to reduce tissue perfusion. 1 However, the hemorrheologic events related to the in vivo hemodynamies are not clear; peripheral blood flow measurements in neonates with polycythemia show no close link between hyperviscosity and perfusion. 3, 4 In previous circulatory studies of neonatal polycythemia, the blood flow was measured in large blood vessels or whole organs. These results may not reflect the effects of polycythemia on the nutritive vasculature (i.e., the capillaries) because the volume fraction of erythrocytes in the capillary differs from that in a large blood vessel. Experimental and adult human data show that different mechanisms in the Supported by grants from the Samaritan Foundation, the General Maternity Hospital Foundation, the Swedish Medical Research Council (No. 6835), the First of May Flower Annual Campaign for Children's Health, the Swedish Society of Medicine, and Karolinska Institute's research foundations in Stockholm, Sweden. Submitted for publication Aug. 29, 1991; accepted Jan. 14, 1992. Reprint requests: Mikael Norman, MD, Department of Pediatrics, Karolinska Hospital, Box 60-500, 104-01 Stockholm, Sweden. 9/23/36394
microcirculation, both passive (Fahraeus-Lindqvist effect) 5 and active (precapillary vasomotor activity),6 act to reduce the capillary hematocrit. The significance of these factors on capillary perfusion during systemic erythrocyte overload in the neonate is, however, unknown. CBV Capillary blood flow velocity Hctcap Capillary hematocrit Hctven Venoushematocrit In human beings, noninvasive and direct observation of capillary hemodynamics can be performed only in exposed organs, such as the skin. We previously studied skin capillary perfusion in healthy neonates and found an inverse correlation between hematocrit and capillary blood flow] The aim of this study was to investigate the skin capillary blood flow in neonatal polycythemia, before and after treatment with isovolemic hemodilution. METHODS We studied the skin microcirculation in 12 newborn infants (six boys) admitted to the neonatal unit for polycythemia (i.e., with a peripheral venous hematocrit >__70%) and in 13 healthy control infants with skin-prick hematocrit 103
10 4
Norman, Fagrell, and Herin
The Journal of Pediatrics July 1992
800
plethora, one or more of the following symptoms: tachypnea (3 infants), vomiting and feeding problems (2), jaundice (2), and jitteriness (1). The treatment was initiated immediately after the first microcirculatory investigation and consisted of an isovolemic exchange transfusion via an umbilical vein catheter. An estimated blood volume of 85 ml/kg 8 and an expected hematocrit of 55% were used to calculate the exchange volume (ExVol), according to the following equation:
0
700 600 500 > 400 m 0 300
ExVol (ml) = 85 (ml/kg) x BW (kg) X (HCtobs-- 55)/Hctobs
O9
200' 100
9
O 0
9
0
-IOC
................................................................................
-2o
-i5-
" I b O ~
o
.............. [ ............
o
A Hematocrit, % Fi9, t. Relative percentage of change in capillary blood flow velocity (% CBV) versus drop in skin-prickhematocrit (AHematocrit, %) after partial exchange transfusion. Significant linear (y = -18.0 • -63.9; r = -0.84; p <0.01) and polynomial regressions (y = 2.13 x2 + 16.8 x -2.81; r = -0.98; p <0.001) were found between the two variables 2 to 4 hours after hemodilution (n = 9, filled circles). A similar relationship was found between data obtained 16 to 24 hours after treatment (n = 6, clear circles).
values --<65%. All infants were vaginally delivered and had their cords clamped within the first minute. The Apgar scores at 1 minute of age were >7. Consent for the investigations was obtained from the parents before each study, and the protocol was approved by the local ethics committee. The average birth weight was 3150 (2240 to 4880) gm in the neonates with polycythemia and 3380 (3030 to 4340) gm in the control infants. The corresponding gestational ages were 38 (35 to 42) and 40 (38 to 40) weeks. In seven patients and five control subjects the postnatal age varied between 6 and 14 hours (referred to as <1 day of age), whereas the remaining subjects were 44 to 96 hours of age (>1 day). The polycythemia group included two slightly premature infants and three neonates with signs of inappropriate intrauterine growth (two were small for gestational age, and one was large for gestational age). Maternal gestational diabetes was present in two cases. One patient was a twin; the sibling had a normal hematocrit. On admission to the hospital, 6 neonates had, in addition to
where BW is the birth weight and HCtobs is the observed hematocrit. The blood withdrawn was replaced by either fresh frozen plasma (n = 5) or a 5% albumin solution (n = 7). The procedures of umbilical vein catheterization and exchange transfusion caused no complications in any subject. The healthy control infants were studied once, whereas the neonates with polycythemia were investigated before and at 2 to 4 hours and 16 to 24 hours after hemodilution. The capillary perfusion was investigated at room temperature (23 ~ + 0.5 ~ C) with the infant clothed and sleeping in the prone position. Details of the methods used for determinations of capillary blood flow velocity have been described elsewhere.7, 9 The nail-fold capillary loops of the thumb were visualized on a television screen with a light microscope connected to a television camera. The image was stored on videotape for subsequent analysis with a videophotometric technique. A pair of photometric windows were generated on the television screen and placed over a clearly visible capillary. Different optical densities within the capillary, produced by the passage of blood cells and plasma gaps, resulted in two time-separated but otherwise similar lines of optical signals from the dual windows. The CBV was quantified by computing the interwindow transit time between similar optical signals. This type of image analysis has been found to give reliable (coefficient of variation for repeated measuremehts ~ 3%) and valid results when used in neonates.9 It was possible to measure the CBV in one to four capillaries in each subject because of variations in recording time and in capillary visualization. Twenty-five capillaries in the neonates with polycythemia and 37 in the control infants were studied. The mean individual CBV values during an effective measuring period of 1 to 5 min/capillary are presented. In each individual, comparisons of the CBV before and after treatment were always made in the same capillary and at the same intracapillary location. Successful CBV recordings were obtained in nine of the neonates with polycythemia at the first follow-up (2 to 4 hours) and in six at the second follow-up (16 to 24 hours) after hemodilution.
Volume 121 Number 1
Effects o f polycythemia and hemodilution on capillary perfusion
1 05
T a b l e . Skin capillary perfusion in healthy neonates (n = 13) and in neonates with polycythemia before (n = 12), 2 to 4 hours (n = 9) and 16 to 24 hours (n = 6) after treatment with hemodilution Infants with p o l y c y t h e m i a
Heart rate (min-1) MAP (mm Hg) Skin temperature (~ Body temperature (~ Hctven (%) Hct~av (%) CBV (mm/sec) % CBV
Control infants
Before HD
2-4 hr after liD
16-24 hr after HD
125 (98-143) 54 (39-68) 30.1 (28.6-32.2) 36.6 (36.5-37.0) -61 (52-65) 0.30 (0.17-0.44)
122 (109-151) 56 (42-63) 29~2 (27.4-32.2)* 36.5 (36.2-36.9)* 74 (70-79) 78 (70-85)* 0.11 (0.02-0.34)*
125 (93-149) 51 (36-72) 29.0 (28.2-32.2) 36.4 (36.2-37.3) 61 (57-69)]68 (59-76)]0.21 (0.07-0.38)t +50 (-14-+350)]"
124 (100-146) 53 (46-64) 29.1 (28.3-31.4) 36.5 (36.2-36.6) 62 (55-69) 67 (60-72) 0.23 (0.07-0.50) +61 (-12-+733)
The data are presentedas median(range)values. HD, Hemodilution;MAP, meanarterial bloodpressure. *The differencescomparedwith controlinfantsstatisticallysignificant(p <0.05). tThe differencescomparedwith beforehemodilutionstatisticallysignificant(p <0.05).
The hematocrits were sampled in duplicate 75 mm heparinized microtubules and blood was obtained from a peripheral vein (Hctven) and from a skin prick of the infant's unwarmed heel (Hcte~p). In the control group, venous samples were not obtained, because the Hctoap in healthy neonates has been found to correlate with the consistently lower Hctven.10, 11 Whereas the Hctven was primarily used to define polycythemia, the capillary perfusion measurements should preferably be related to the hematocrit in the capillary. With regard to the cutaneous microcirculation, the Hctcap (i.e., the hematocrit in a group of cutaneous microvessels) was considered as a more accurate estimate of the capillary hematocrit than the corresponding venous value. After centrifugation at 10,000 rpm for 10 minutes, the hematocrit values were read manually to avoid the diagnostic error reported with automated blood cell counters. 12 The methodologic error for this type of determination, incurred during sampling and instrumentation, is 3.4% of the venous13 and 1% of the observed skin-prick values. 1~ The leukocyte and platelet counts were also determined in each neonate with polycythemia. Other variables studied at each CBV observation included the body and local skin temperatures of the thumb and the oscillometric determinations of the heart rate and mean arterial blood pressure. Because the data showed a skewed distribution, the results are presented as median and range values. For comparisons between groups of data, the Mann-Whitney U test and the Wilcoxon signed rank test were used. To elucidate any associations between the CBV and the other variables studied, we used a stepwise regression model (F value to enter >--_4). RESULTS The CBV in the neonates with polycythemia was 0.11 (0.02 to 0.34) mm/sec and in the healthy control infants
0.30 (0.17 to 0.44) mm/sec (p <0.01). According to regression analysis, this finding was associated with the difference in Hctcap between the two groups (p <0.01), whereas the differences in temperatures and postnatal age did not contribute to the CBV discrepancy (Table). Within groups, a weak correlation was found between postnatal age and CBV in the infants with polycythemia (r = 0.62; p = 0.05; regression slope = 0.002; intercept = 0.06), but not in the control infants. After treatment, the hematocrit decreased and the CBV increased, whereas the other variables remained unchanged (Table). The degree of hemodilution achieved varied considerably (Fig. 1). The CBV, 2 to 4 hours after hemodilution (n = 9), increased as a function of the decrease in hematocrit. The absolute change in CBV (measured in millimeters per second) correlated inversely with the AHctcap (r = -0.70; p <0.05) and with the AHctven (r = -0.80; p <0.05). The corresponding correlation coefficients for the associations between the relative percentage of change in CBV and the two types of hematocrit determinations were -0.84 (p <0.01) and -0.75 (p <0.05), respectively. When the percentage of CBV and the AHctcap values were entered into a polynomial regression model (second degree), the correlation increased (r = -0.98; p <0.001), indicating a strong nonlinear relationship (Fig. 1). The data obtained 16 to 24 hours after treatment (n = 6) showed the same relationship between the CBV and the AHctcap (Fig. 1). In addition to the AHctven, the postnatal age was also found to contribute to the percentage of CBV after treatment (r = - 0 . 8 8 ; p <0.05) (slopeage = -2.7; slopeaHct = --1 1.6, intercept 61). To clarify the associations found between the CBV and the postnatal age, we subgrouped the subjects with polycythemia. As shown in Fig. 2, the less-than-l-day-old neonates with polycythemia (n = 7) had a very slow skin capillary perfusion before treatment,
106
Norman, Fagrell, and Herin
The Journal of Pediatrics July 1992
.5 A
.4
CD
o~
0
.3
J 9
A
0 0
0
A
8
.2 A
.1
0
50
55
60
65
70
75
80
85
90
Hematocrit, % Fig. 2. Capillary blood flow velocity (CBV) versus skin-prick hematocrit in neonates with polycythemia (filled symbols) and healthy control infants (dear symbols). CBV was significantlyslower in neonates with polycythemia than in healthy control infants (p <0.01 ). This differencewas most pronouncedin neonates < 1 day of age (triangles), whereas several older neonates with polycythemia (circles) had CBV similar to that in control infants.
whereas several of the older patients (n = 5), although equally polycythemie, had CBV values similar to those found in the control infants. Moreover, 2 to 4 hours after hemodilution, the younger neonates had a significantly greater decrease in hematocrit (AHCtcap = - 1 6 % ; AHctven =--16% [median values]) than did the older ones (AHctcap = -7%; AHetven = --9%; p <0.05). This difference in absolute hematocrit reduction was accompanied by a much larger CBV increase (+234%) in the younger infants than in the infants who were > 1 day of age when receiving treatment (median CBV increase, +3%; p <0.05). The postnatal age in the infants treated with fresh frozen plasma was 12 (9 to 92) hours (n = 5), and it was 14 (6 to 52) hours in the albumin-treated subgroup (n = 7). Before treatment, there was no significant difference in hematocrit or CBV values between these two subgroups. After treatment, the use of two different exchange solutions did not contribute to the variation in treatment effects. However, the microcirculatory and hemodiluting effects of the two exchange solutions could not be conclusively compared because of the small number of infants in each subgroup from which CBV values could be obtained 2 to 4 hours after hemodilution (n = 3 and 6, respectively), and because of the age-related variation in therapy outcome. There was no correlation between the .I~c~ap and the Hctvenin the neonates with polycythemia before hemodilution. However, 2 to 4 hours after treatment the two types of hematocrit determinations showed a significant correlation (r = 0.72; p <0.01), which increased 16 to 24 hours after treatment (r = 0.87; p <0.05).
The leukocyte count was 13.7 (6.2 to 31.1 ) • 109/L, and the platelet count was 211 (121 to 321) x 109/L before treatment. These values did not change significantly after exchange transfusion. Although no association was found between the leukocyte count and the CBV, the number of leukocytes correlated inversely with the postnatal age (r = • p <0.01). With the exception of jaundice, the symptoms seen in some of the neonates with polycythemia on admission to the hospital rapidly disappeared after hemodilution. DISCUSSION The relationship between neonatal polycythemia and a decreased CBV can be interpreted in one of two ways: it reflects either a phys!~logic hemodynamic adaptation to increased oxygen delivery (i.e., hemoglobin mass) or insufficient tissue perfusion in polycythemia (indicating a pathophysiologic phenomenon). Our finding of a manifold increase in capillary perfusion after a moderate reduction in hematocrit supports the view that the hematocrit plays a major role as a flow-resistance factor in neonatal skin capillaries, and that perfusion before treatment was impeded. The current explanation of the reason that an elevated hematocrit can cause reduced perfusion is that the blood becomes hyperviscous and increases the resistance to flow. The skin CBV in several of the neonates with polycythemia who were > 1 day of age was the same as in the control infants, so factors other than the hematoerit may be involved in the development of microcirculatory deceleration. In vitro theologic data give no indication of an unfavorable
Volume 121 Number 1
Effects of polycythemia and hemodilution on capillary perfusion
blood composition at birth with respect to other determinants of viscous resistance, such as plasma viscosity and erythrocyte deformabilityJ 4, z5 On the contrary, these factors seem to be adjusted to enhance neonatal blood flow and reduce viscosity in artificial microvessels in the presence of a high hematocrit, s, 16 In vivo, the microvascular hematocrit is determined not only by the composition of the blood but also by the precapillary vasomotor activity. Thus the fraction of blood cells flowing through the skin capillaries i n the adult has been found to decrease intermittently, a finding that has been attributed to periodic vasocontractions in the supplying precapillary resistance vessels. 6 In healthy neonates, the cutaneous vasomotor activity appears to be less pronounced 7 or even absent during the first day of life. 17 Such an inactive vasoregulatory mechanism in combination with polycythemia may explain the very slow skin capillary circulation seen in the patients < 1 day of age. Beneficial circulatory effects and functional improvement after hemodilution have been described in several studies of neonatal polycythemia), 18, 19 Our results demonstrate that the hemodynamic effects of systemic hematocrit changes may be reflected particularly at the level of the microcirculation in the neonate. A pronounced increase in capillary perfusion was noted in response to hemodilution, despite an unchanged heart rate, blood pressure, and skin temperature. Whether the increased microcirculation is a result of the altered blood composition alone or also involves a change in microvascular performance remains to be studied. A high hematocrit has been associated with decreased skin microvascular reactivity in healthy neonates, 2~ indicating that polycythemia per se may interact with the microvascular regulation. Although the treatment was aimed at achieving a fixed hematocrit value, the hemodiluting effect varied considerably. An intraindividual, age-related variation in hemodilution might be expected because the natural course of the hematocrit change after birth is characterized by an initial hemoconcentration during the first 2 to 4 hours, after which the hematocrit falls to a more stable level at 18 hours of age.lO, 11 Interindividual variations in blood volume may also contribute to the differences in effect. Hypervolemia caused by delayed clamping of the umbilical cord with an excessively large placental transfusion l~ is a less probable cause of polycythemia in our patients, because the cord is routinely clamped within the first minute after birth in our clinic. However, neonates with intrauterine growth retardation often have blood volumes exceeding the value of 85 m l / k g used in our calculations, z1 which may have invalidated some of the exchange volume estimates. The postnatal age in our study varied from a few hours to several days. During this postnatal period, the cutaneous
1 07
vasomotor activity in healthy neonates has been found to increase. 17 The relationship between hematocrit reduction and CBV increase may therefore exhibit age-specific characteristics. Because of the varying degree of hematocrit reduction, the microcirculatory effects of a more pronounced hemodilution could be studied only in the younger ( < 1 day old) infants. By inference, the validity of the regression data may be limited to this specific postnatal age. Whether these findings in the cutaneous capillary circulation also apply to vital organs is an unanswered question. Further studies are required to elucidate whether the skin microcirculation in the human neonate reflects that of the internal organs.
REFERENCES
1. Van der Elst CW, Molteno CD, Malan AF, Van Hesse H. The management of polycythemia in the newborn infant. Early Hum Dev 1980;4:393-403. 2. Wiswell TE, Cornish JD, Northam RS. Neonatal polycythemia: frequency of clinical manifestations and other associated findings. Pediatrics 1986;78:26-30. 3. Bergqvist G, Zetterstr6m R. Blood viscosity: and peripheral circulation in newborn infants. Acta Paediatr Scand 1974; 63:865-8. 4. Linderkamp O, Strohhacker I, Versmold HT, Klose H, Riegel KP, Betke K. Peripheral circulation in the newborn: interaction of peripheral blood flow, blood pressure, blood volume and blood viscosity. Eur J Pediatr 1978;129:73-8i. 5. Zilow EP, Linderkamp O. Viscosity reduction of red blood cells~ from preterm and full-term neonates and adults in narrow tubes (Fahraeus-Lindqvist effect). Pediatr Res 1989;25:595-9. 6. Fagrell B, Intaglietta M, Ostergren J. Relative hematocrit in human skin capillaries and its relation to capillary blood flow velocity. Microvasc Res 1980;20:327-35. 71 Norman M, Herin P, Fagrell B, Zetterstr6m R. Capillary blood cell velocity in full-term infants as determined in skin by videophotometric microscopy. Pediatr Res 1988;23:585-8. 8. Rawlings JS, Pettett G, Wiswell TE, Clapper J. Estimated blood volumes in polycythemic neonates as a function of birth weight. J PEDIATR 1982;101:594-9. 9. Norman M, Herin P, Fagrell B. An evaluation of skin capillary blood flow determinations in neonates using a computerized videophotometric method. Microvasc Res (in press). 10. Oh W, Lind J. Venous and capillary hematocrit in newborn infants and placental transfusion. Acta Paediatr Scand 1966; 55:38-48. 11. Ramamurthy RS, Berlanga M. Postnatal alterations in hematocrit and viscosity in normal and polycythemic infants. J PEDIATR 1987;110:929-34. 12. Villalta IA, Pramanik AK, Diaz-Blanco J, Herbst JJ. Diagnostic errors in neonatal polycythemia based on method of hematocrit determination. J PEDIATR 1989;115:460-2. 13. Shohat M, Merlob P, Reisner SH. Neonatal p01ycythemia. I. Early diagnosis and incidence relating to time of sampling. Pediatrics 1984;73:7-10. 14. Linderkamp O. Blood rheology in the newborn infant. Baillieres Clin Haematol 1987;1(3):801-25. 15. Rampling MW, Whittingstall P, Martin G, et al. A compar-
108
Norman, Fagrell, and Herin
The Journal of Pediatrics July 1992
ison of the rheologic properties of neonatal and adult blood. Pediatr Res 1989;25:457-60. 16. Stadler A, Linderkamp O. Flow behaviour of neonatal and adult erythrocytes in narrow capillaries. Microvasc Res 1989; 37:267-79. 17. P6schl J, Weiss T, Diem C, Linderkamp O. Periodicvariations in skin perfusion in full-term and preterm neonates using laser Doppler technique. Acta Paediatr Scand 1991;80:999-1007. 18. Swetnam SM, Yabek SM, Alverson DC. Hemodynamic consequences of neonatal polycythemia. J PEDIATR 1987;110: 443-7.
19. Maertzdorf WJ, Tangelder GJ, Slaaf DW, Blanco CE. Effects of partial plasma exchange transfusion on cerebral blood flow velocity in polycythaemic preterm, term and small for date newborn infants. Eur J Pediatr 1989;148:774-8. 20. Norman M, Gu L, Herin P, Fagrell B. Reactive hyperemia in term neonates and adults: a laser Doppler fluxmetry study of skin microcirculation. Microvasc Res 1991;41:229-38. 21. Maertzdorf WJ, Aldenhuyzen-Dorland W, Slaaf DW, Tangelder GJ, Blanco CE. Circulating blood volume in appropriate and small for gestational age full term and preterm polycythaemic infants. Acta Paediatr Scand 1991;80:620-7.
FELLOWSHIPS Fellowships available in pediatric subspecialties and those for general academic pediatric training are listed once a year, in January, in THE JOURNAL OF PEDIATRICS. Each June, forms for listing fellowships available for the academic year beginning 18 months after publication are sent to the Chairman of the Department of Pediatrics at major hospitals in the United States and Canada. In addition, a copy of the application form appears in the July, August, and September issues of THE JOURNAL (please use the current form). Should you desire to list fellowships, a separate application must be made each year for each position. All applications must be returned to Mosby-Year Book, Inc., by October 15 preceding the listing year to ensure publication. Additional forms will be supplied on request from the Journal Editing Department, Mosby-Year Book, Inc., 11830 Westline Industrial Drive, St. Louis, MO 63146-3318/800-325-4l 77, ext. 4317, or 314453-4317.
The effects of neonatal polycythemia on nutritive capillary perfusion were investigated by a television microscopy technique. The capillary blood flow velocity in skin was measured in 12 neonates with polycythemia before and after treatment with hemodilution, and in 13 healthy control infants. The capillary blood flow velocity in the patients was 0.11 (0.02 to 0.34) mm/sec and in the healthy control infants 0.30 (0.17 to 0.44) mm/sec (p <0.01, median and range values). In relation to the absolute hematocrit c h a n g e after treatment (range, - 2 0 % to 0%), the capillary blood flow velocity increased nonlinearly (range, +733% to -14%; r = -0.98; p <0.001). The postnatal a g e was found to contribute significantly to the variation in results--the neonates with polycythemia studied during the first day of life had a very slow skin capillary circulation and responded to treatment with a more pronounced increase in capillary blood flow velocity than did the older patients. This in vivo model for capillary perfusion indicates that an insufficient microcirculation may be involved in the pathophysiology responsiblefor the morbidity associated with neonatal polycythemia. (J PEDIATR1992;121:103-8) Neonatal polycythemia is a common clinical problem that is associated with functional failure of various vital organ systems. 1, 2 The hyperviscosity of the blood in these patients is believed to reduce tissue perfusion. 1 However, the hemorrheologic events related to the in vivo hemodynamies are not clear; peripheral blood flow measurements in neonates with polycythemia show no close link between hyperviscosity and perfusion. 3, 4 In previous circulatory studies of neonatal polycythemia, the blood flow was measured in large blood vessels or whole organs. These results may not reflect the effects of polycythemia on the nutritive vasculature (i.e., the capillaries) because the volume fraction of erythrocytes in the capillary differs from that in a large blood vessel. Experimental and adult human data show that different mechanisms in the Supported by grants from the Samaritan Foundation, the General Maternity Hospital Foundation, the Swedish Medical Research Council (No. 6835), the First of May Flower Annual Campaign for Children's Health, the Swedish Society of Medicine, and Karolinska Institute's research foundations in Stockholm, Sweden. Submitted for publication Aug. 29, 1991; accepted Jan. 14, 1992. Reprint requests: Mikael Norman, MD, Department of Pediatrics, Karolinska Hospital, Box 60-500, 104-01 Stockholm, Sweden. 9/23/36394
microcirculation, both passive (Fahraeus-Lindqvist effect) 5 and active (precapillary vasomotor activity),6 act to reduce the capillary hematocrit. The significance of these factors on capillary perfusion during systemic erythrocyte overload in the neonate is, however, unknown. CBV Capillary blood flow velocity Hctcap Capillary hematocrit Hctven Venoushematocrit In human beings, noninvasive and direct observation of capillary hemodynamics can be performed only in exposed organs, such as the skin. We previously studied skin capillary perfusion in healthy neonates and found an inverse correlation between hematocrit and capillary blood flow] The aim of this study was to investigate the skin capillary blood flow in neonatal polycythemia, before and after treatment with isovolemic hemodilution. METHODS We studied the skin microcirculation in 12 newborn infants (six boys) admitted to the neonatal unit for polycythemia (i.e., with a peripheral venous hematocrit >__70%) and in 13 healthy control infants with skin-prick hematocrit 103
10 4
Norman, Fagrell, and Herin
The Journal of Pediatrics July 1992
800
plethora, one or more of the following symptoms: tachypnea (3 infants), vomiting and feeding problems (2), jaundice (2), and jitteriness (1). The treatment was initiated immediately after the first microcirculatory investigation and consisted of an isovolemic exchange transfusion via an umbilical vein catheter. An estimated blood volume of 85 ml/kg 8 and an expected hematocrit of 55% were used to calculate the exchange volume (ExVol), according to the following equation:
0
700 600 500 > 400 m 0 300
ExVol (ml) = 85 (ml/kg) x BW (kg) X (HCtobs-- 55)/Hctobs
O9
200' 100
9
O 0
9
0
-IOC
................................................................................
-2o
-i5-
" I b O ~
o
.............. [ ............
o
A Hematocrit, % Fi9, t. Relative percentage of change in capillary blood flow velocity (% CBV) versus drop in skin-prickhematocrit (AHematocrit, %) after partial exchange transfusion. Significant linear (y = -18.0 • -63.9; r = -0.84; p <0.01) and polynomial regressions (y = 2.13 x2 + 16.8 x -2.81; r = -0.98; p <0.001) were found between the two variables 2 to 4 hours after hemodilution (n = 9, filled circles). A similar relationship was found between data obtained 16 to 24 hours after treatment (n = 6, clear circles).
values --<65%. All infants were vaginally delivered and had their cords clamped within the first minute. The Apgar scores at 1 minute of age were >7. Consent for the investigations was obtained from the parents before each study, and the protocol was approved by the local ethics committee. The average birth weight was 3150 (2240 to 4880) gm in the neonates with polycythemia and 3380 (3030 to 4340) gm in the control infants. The corresponding gestational ages were 38 (35 to 42) and 40 (38 to 40) weeks. In seven patients and five control subjects the postnatal age varied between 6 and 14 hours (referred to as <1 day of age), whereas the remaining subjects were 44 to 96 hours of age (>1 day). The polycythemia group included two slightly premature infants and three neonates with signs of inappropriate intrauterine growth (two were small for gestational age, and one was large for gestational age). Maternal gestational diabetes was present in two cases. One patient was a twin; the sibling had a normal hematocrit. On admission to the hospital, 6 neonates had, in addition to
where BW is the birth weight and HCtobs is the observed hematocrit. The blood withdrawn was replaced by either fresh frozen plasma (n = 5) or a 5% albumin solution (n = 7). The procedures of umbilical vein catheterization and exchange transfusion caused no complications in any subject. The healthy control infants were studied once, whereas the neonates with polycythemia were investigated before and at 2 to 4 hours and 16 to 24 hours after hemodilution. The capillary perfusion was investigated at room temperature (23 ~ + 0.5 ~ C) with the infant clothed and sleeping in the prone position. Details of the methods used for determinations of capillary blood flow velocity have been described elsewhere.7, 9 The nail-fold capillary loops of the thumb were visualized on a television screen with a light microscope connected to a television camera. The image was stored on videotape for subsequent analysis with a videophotometric technique. A pair of photometric windows were generated on the television screen and placed over a clearly visible capillary. Different optical densities within the capillary, produced by the passage of blood cells and plasma gaps, resulted in two time-separated but otherwise similar lines of optical signals from the dual windows. The CBV was quantified by computing the interwindow transit time between similar optical signals. This type of image analysis has been found to give reliable (coefficient of variation for repeated measuremehts ~ 3%) and valid results when used in neonates.9 It was possible to measure the CBV in one to four capillaries in each subject because of variations in recording time and in capillary visualization. Twenty-five capillaries in the neonates with polycythemia and 37 in the control infants were studied. The mean individual CBV values during an effective measuring period of 1 to 5 min/capillary are presented. In each individual, comparisons of the CBV before and after treatment were always made in the same capillary and at the same intracapillary location. Successful CBV recordings were obtained in nine of the neonates with polycythemia at the first follow-up (2 to 4 hours) and in six at the second follow-up (16 to 24 hours) after hemodilution.
Volume 121 Number 1
Effects o f polycythemia and hemodilution on capillary perfusion
1 05
T a b l e . Skin capillary perfusion in healthy neonates (n = 13) and in neonates with polycythemia before (n = 12), 2 to 4 hours (n = 9) and 16 to 24 hours (n = 6) after treatment with hemodilution Infants with p o l y c y t h e m i a
Heart rate (min-1) MAP (mm Hg) Skin temperature (~ Body temperature (~ Hctven (%) Hct~av (%) CBV (mm/sec) % CBV
Control infants
Before HD
2-4 hr after liD
16-24 hr after HD
125 (98-143) 54 (39-68) 30.1 (28.6-32.2) 36.6 (36.5-37.0) -61 (52-65) 0.30 (0.17-0.44)
122 (109-151) 56 (42-63) 29~2 (27.4-32.2)* 36.5 (36.2-36.9)* 74 (70-79) 78 (70-85)* 0.11 (0.02-0.34)*
125 (93-149) 51 (36-72) 29.0 (28.2-32.2) 36.4 (36.2-37.3) 61 (57-69)]68 (59-76)]0.21 (0.07-0.38)t +50 (-14-+350)]"
124 (100-146) 53 (46-64) 29.1 (28.3-31.4) 36.5 (36.2-36.6) 62 (55-69) 67 (60-72) 0.23 (0.07-0.50) +61 (-12-+733)
The data are presentedas median(range)values. HD, Hemodilution;MAP, meanarterial bloodpressure. *The differencescomparedwith controlinfantsstatisticallysignificant(p <0.05). tThe differencescomparedwith beforehemodilutionstatisticallysignificant(p <0.05).
The hematocrits were sampled in duplicate 75 mm heparinized microtubules and blood was obtained from a peripheral vein (Hctven) and from a skin prick of the infant's unwarmed heel (Hcte~p). In the control group, venous samples were not obtained, because the Hctoap in healthy neonates has been found to correlate with the consistently lower Hctven.10, 11 Whereas the Hctven was primarily used to define polycythemia, the capillary perfusion measurements should preferably be related to the hematocrit in the capillary. With regard to the cutaneous microcirculation, the Hctcap (i.e., the hematocrit in a group of cutaneous microvessels) was considered as a more accurate estimate of the capillary hematocrit than the corresponding venous value. After centrifugation at 10,000 rpm for 10 minutes, the hematocrit values were read manually to avoid the diagnostic error reported with automated blood cell counters. 12 The methodologic error for this type of determination, incurred during sampling and instrumentation, is 3.4% of the venous13 and 1% of the observed skin-prick values. 1~ The leukocyte and platelet counts were also determined in each neonate with polycythemia. Other variables studied at each CBV observation included the body and local skin temperatures of the thumb and the oscillometric determinations of the heart rate and mean arterial blood pressure. Because the data showed a skewed distribution, the results are presented as median and range values. For comparisons between groups of data, the Mann-Whitney U test and the Wilcoxon signed rank test were used. To elucidate any associations between the CBV and the other variables studied, we used a stepwise regression model (F value to enter >--_4). RESULTS The CBV in the neonates with polycythemia was 0.11 (0.02 to 0.34) mm/sec and in the healthy control infants
0.30 (0.17 to 0.44) mm/sec (p <0.01). According to regression analysis, this finding was associated with the difference in Hctcap between the two groups (p <0.01), whereas the differences in temperatures and postnatal age did not contribute to the CBV discrepancy (Table). Within groups, a weak correlation was found between postnatal age and CBV in the infants with polycythemia (r = 0.62; p = 0.05; regression slope = 0.002; intercept = 0.06), but not in the control infants. After treatment, the hematocrit decreased and the CBV increased, whereas the other variables remained unchanged (Table). The degree of hemodilution achieved varied considerably (Fig. 1). The CBV, 2 to 4 hours after hemodilution (n = 9), increased as a function of the decrease in hematocrit. The absolute change in CBV (measured in millimeters per second) correlated inversely with the AHctcap (r = -0.70; p <0.05) and with the AHctven (r = -0.80; p <0.05). The corresponding correlation coefficients for the associations between the relative percentage of change in CBV and the two types of hematocrit determinations were -0.84 (p <0.01) and -0.75 (p <0.05), respectively. When the percentage of CBV and the AHctcap values were entered into a polynomial regression model (second degree), the correlation increased (r = -0.98; p <0.001), indicating a strong nonlinear relationship (Fig. 1). The data obtained 16 to 24 hours after treatment (n = 6) showed the same relationship between the CBV and the AHctcap (Fig. 1). In addition to the AHctven, the postnatal age was also found to contribute to the percentage of CBV after treatment (r = - 0 . 8 8 ; p <0.05) (slopeage = -2.7; slopeaHct = --1 1.6, intercept 61). To clarify the associations found between the CBV and the postnatal age, we subgrouped the subjects with polycythemia. As shown in Fig. 2, the less-than-l-day-old neonates with polycythemia (n = 7) had a very slow skin capillary perfusion before treatment,
106
Norman, Fagrell, and Herin
The Journal of Pediatrics July 1992
.5 A
.4
CD
o~
0
.3
J 9
A
0 0
0
A
8
.2 A
.1
0
50
55
60
65
70
75
80
85
90
Hematocrit, % Fig. 2. Capillary blood flow velocity (CBV) versus skin-prick hematocrit in neonates with polycythemia (filled symbols) and healthy control infants (dear symbols). CBV was significantlyslower in neonates with polycythemia than in healthy control infants (p <0.01 ). This differencewas most pronouncedin neonates < 1 day of age (triangles), whereas several older neonates with polycythemia (circles) had CBV similar to that in control infants.
whereas several of the older patients (n = 5), although equally polycythemie, had CBV values similar to those found in the control infants. Moreover, 2 to 4 hours after hemodilution, the younger neonates had a significantly greater decrease in hematocrit (AHCtcap = - 1 6 % ; AHctven =--16% [median values]) than did the older ones (AHctcap = -7%; AHetven = --9%; p <0.05). This difference in absolute hematocrit reduction was accompanied by a much larger CBV increase (+234%) in the younger infants than in the infants who were > 1 day of age when receiving treatment (median CBV increase, +3%; p <0.05). The postnatal age in the infants treated with fresh frozen plasma was 12 (9 to 92) hours (n = 5), and it was 14 (6 to 52) hours in the albumin-treated subgroup (n = 7). Before treatment, there was no significant difference in hematocrit or CBV values between these two subgroups. After treatment, the use of two different exchange solutions did not contribute to the variation in treatment effects. However, the microcirculatory and hemodiluting effects of the two exchange solutions could not be conclusively compared because of the small number of infants in each subgroup from which CBV values could be obtained 2 to 4 hours after hemodilution (n = 3 and 6, respectively), and because of the age-related variation in therapy outcome. There was no correlation between the .I~c~ap and the Hctvenin the neonates with polycythemia before hemodilution. However, 2 to 4 hours after treatment the two types of hematocrit determinations showed a significant correlation (r = 0.72; p <0.01), which increased 16 to 24 hours after treatment (r = 0.87; p <0.05).
The leukocyte count was 13.7 (6.2 to 31.1 ) • 109/L, and the platelet count was 211 (121 to 321) x 109/L before treatment. These values did not change significantly after exchange transfusion. Although no association was found between the leukocyte count and the CBV, the number of leukocytes correlated inversely with the postnatal age (r = • p <0.01). With the exception of jaundice, the symptoms seen in some of the neonates with polycythemia on admission to the hospital rapidly disappeared after hemodilution. DISCUSSION The relationship between neonatal polycythemia and a decreased CBV can be interpreted in one of two ways: it reflects either a phys!~logic hemodynamic adaptation to increased oxygen delivery (i.e., hemoglobin mass) or insufficient tissue perfusion in polycythemia (indicating a pathophysiologic phenomenon). Our finding of a manifold increase in capillary perfusion after a moderate reduction in hematocrit supports the view that the hematocrit plays a major role as a flow-resistance factor in neonatal skin capillaries, and that perfusion before treatment was impeded. The current explanation of the reason that an elevated hematocrit can cause reduced perfusion is that the blood becomes hyperviscous and increases the resistance to flow. The skin CBV in several of the neonates with polycythemia who were > 1 day of age was the same as in the control infants, so factors other than the hematoerit may be involved in the development of microcirculatory deceleration. In vitro theologic data give no indication of an unfavorable
Volume 121 Number 1
Effects of polycythemia and hemodilution on capillary perfusion
blood composition at birth with respect to other determinants of viscous resistance, such as plasma viscosity and erythrocyte deformabilityJ 4, z5 On the contrary, these factors seem to be adjusted to enhance neonatal blood flow and reduce viscosity in artificial microvessels in the presence of a high hematocrit, s, 16 In vivo, the microvascular hematocrit is determined not only by the composition of the blood but also by the precapillary vasomotor activity. Thus the fraction of blood cells flowing through the skin capillaries i n the adult has been found to decrease intermittently, a finding that has been attributed to periodic vasocontractions in the supplying precapillary resistance vessels. 6 In healthy neonates, the cutaneous vasomotor activity appears to be less pronounced 7 or even absent during the first day of life. 17 Such an inactive vasoregulatory mechanism in combination with polycythemia may explain the very slow skin capillary circulation seen in the patients < 1 day of age. Beneficial circulatory effects and functional improvement after hemodilution have been described in several studies of neonatal polycythemia), 18, 19 Our results demonstrate that the hemodynamic effects of systemic hematocrit changes may be reflected particularly at the level of the microcirculation in the neonate. A pronounced increase in capillary perfusion was noted in response to hemodilution, despite an unchanged heart rate, blood pressure, and skin temperature. Whether the increased microcirculation is a result of the altered blood composition alone or also involves a change in microvascular performance remains to be studied. A high hematocrit has been associated with decreased skin microvascular reactivity in healthy neonates, 2~ indicating that polycythemia per se may interact with the microvascular regulation. Although the treatment was aimed at achieving a fixed hematocrit value, the hemodiluting effect varied considerably. An intraindividual, age-related variation in hemodilution might be expected because the natural course of the hematocrit change after birth is characterized by an initial hemoconcentration during the first 2 to 4 hours, after which the hematocrit falls to a more stable level at 18 hours of age.lO, 11 Interindividual variations in blood volume may also contribute to the differences in effect. Hypervolemia caused by delayed clamping of the umbilical cord with an excessively large placental transfusion l~ is a less probable cause of polycythemia in our patients, because the cord is routinely clamped within the first minute after birth in our clinic. However, neonates with intrauterine growth retardation often have blood volumes exceeding the value of 85 m l / k g used in our calculations, z1 which may have invalidated some of the exchange volume estimates. The postnatal age in our study varied from a few hours to several days. During this postnatal period, the cutaneous
1 07
vasomotor activity in healthy neonates has been found to increase. 17 The relationship between hematocrit reduction and CBV increase may therefore exhibit age-specific characteristics. Because of the varying degree of hematocrit reduction, the microcirculatory effects of a more pronounced hemodilution could be studied only in the younger ( < 1 day old) infants. By inference, the validity of the regression data may be limited to this specific postnatal age. Whether these findings in the cutaneous capillary circulation also apply to vital organs is an unanswered question. Further studies are required to elucidate whether the skin microcirculation in the human neonate reflects that of the internal organs.
REFERENCES
1. Van der Elst CW, Molteno CD, Malan AF, Van Hesse H. The management of polycythemia in the newborn infant. Early Hum Dev 1980;4:393-403. 2. Wiswell TE, Cornish JD, Northam RS. Neonatal polycythemia: frequency of clinical manifestations and other associated findings. Pediatrics 1986;78:26-30. 3. Bergqvist G, Zetterstr6m R. Blood viscosity: and peripheral circulation in newborn infants. Acta Paediatr Scand 1974; 63:865-8. 4. Linderkamp O, Strohhacker I, Versmold HT, Klose H, Riegel KP, Betke K. Peripheral circulation in the newborn: interaction of peripheral blood flow, blood pressure, blood volume and blood viscosity. Eur J Pediatr 1978;129:73-8i. 5. Zilow EP, Linderkamp O. Viscosity reduction of red blood cells~ from preterm and full-term neonates and adults in narrow tubes (Fahraeus-Lindqvist effect). Pediatr Res 1989;25:595-9. 6. Fagrell B, Intaglietta M, Ostergren J. Relative hematocrit in human skin capillaries and its relation to capillary blood flow velocity. Microvasc Res 1980;20:327-35. 71 Norman M, Herin P, Fagrell B, Zetterstr6m R. Capillary blood cell velocity in full-term infants as determined in skin by videophotometric microscopy. Pediatr Res 1988;23:585-8. 8. Rawlings JS, Pettett G, Wiswell TE, Clapper J. Estimated blood volumes in polycythemic neonates as a function of birth weight. J PEDIATR 1982;101:594-9. 9. Norman M, Herin P, Fagrell B. An evaluation of skin capillary blood flow determinations in neonates using a computerized videophotometric method. Microvasc Res (in press). 10. Oh W, Lind J. Venous and capillary hematocrit in newborn infants and placental transfusion. Acta Paediatr Scand 1966; 55:38-48. 11. Ramamurthy RS, Berlanga M. Postnatal alterations in hematocrit and viscosity in normal and polycythemic infants. J PEDIATR 1987;110:929-34. 12. Villalta IA, Pramanik AK, Diaz-Blanco J, Herbst JJ. Diagnostic errors in neonatal polycythemia based on method of hematocrit determination. J PEDIATR 1989;115:460-2. 13. Shohat M, Merlob P, Reisner SH. Neonatal p01ycythemia. I. Early diagnosis and incidence relating to time of sampling. Pediatrics 1984;73:7-10. 14. Linderkamp O. Blood rheology in the newborn infant. Baillieres Clin Haematol 1987;1(3):801-25. 15. Rampling MW, Whittingstall P, Martin G, et al. A compar-
108
Norman, Fagrell, and Herin
The Journal of Pediatrics July 1992
ison of the rheologic properties of neonatal and adult blood. Pediatr Res 1989;25:457-60. 16. Stadler A, Linderkamp O. Flow behaviour of neonatal and adult erythrocytes in narrow capillaries. Microvasc Res 1989; 37:267-79. 17. P6schl J, Weiss T, Diem C, Linderkamp O. Periodicvariations in skin perfusion in full-term and preterm neonates using laser Doppler technique. Acta Paediatr Scand 1991;80:999-1007. 18. Swetnam SM, Yabek SM, Alverson DC. Hemodynamic consequences of neonatal polycythemia. J PEDIATR 1987;110: 443-7.
19. Maertzdorf WJ, Tangelder GJ, Slaaf DW, Blanco CE. Effects of partial plasma exchange transfusion on cerebral blood flow velocity in polycythaemic preterm, term and small for date newborn infants. Eur J Pediatr 1989;148:774-8. 20. Norman M, Gu L, Herin P, Fagrell B. Reactive hyperemia in term neonates and adults: a laser Doppler fluxmetry study of skin microcirculation. Microvasc Res 1991;41:229-38. 21. Maertzdorf WJ, Aldenhuyzen-Dorland W, Slaaf DW, Tangelder GJ, Blanco CE. Circulating blood volume in appropriate and small for gestational age full term and preterm polycythaemic infants. Acta Paediatr Scand 1991;80:620-7.
FELLOWSHIPS Fellowships available in pediatric subspecialties and those for general academic pediatric training are listed once a year, in January, in THE JOURNAL OF PEDIATRICS. Each June, forms for listing fellowships available for the academic year beginning 18 months after publication are sent to the Chairman of the Department of Pediatrics at major hospitals in the United States and Canada. In addition, a copy of the application form appears in the July, August, and September issues of THE JOURNAL (please use the current form). Should you desire to list fellowships, a separate application must be made each year for each position. All applications must be returned to Mosby-Year Book, Inc., by October 15 preceding the listing year to ensure publication. Additional forms will be supplied on request from the Journal Editing Department, Mosby-Year Book, Inc., 11830 Westline Industrial Drive, St. Louis, MO 63146-3318/800-325-4l 77, ext. 4317, or 314453-4317.