Early
Human Development,
Elsevier Scientific
Publishers
21 (1991)
Ireland
207
207-221
Ltd.
EHD 01203
The influence of respiratory distress syndrome heart rate variability in very preterm infants
on
Conny M.A. van Ravenswaaij-Arts”, Jeroen C.W. Hopman”, Louis AA. Kollke”, Joop P.L. van Amen”, Gerard B.A. Stoelinga” and Herman P. van Geijnb “Department
of Pediatrics,
University
Hospital
Free University
(Received
6 May 1991;
Nijmegen and hDepariment
Hospital
Amsterdam
oJ Obstetrics and Gynecology,
(The Netherlands)
revision received 31 October 1991; accepted 6 November 1991)
Summary In a multi-parametric study the influence of pathological neonatal conditions on heart rate variability was investigated in 60 preterm infants born at a gestational age below 33 weeks. Measurements were performed during the first 3 days of life. Four times a day, RR-intervals, respiration curve and rate, transcutaneously measured blood gases and observed body movements were recorded while the infants were asleep. All data were stored simultaneously in a micro-computer. Severity of respiratory distress syndrome (RDS), patency of ductus arteriosus and periventricular haemorrhage were documented as well. Four sets of short- (STV) and long-term variability (LTV) indices were calculated. Severe RDS was associated with a signiticant decrease in LTV. The influence of RDS on LTV persisted after correction for conceptional age, postnatal age, behavioural state and variations in respiratory rate and in transcutaneous PO,. Infants with a symptomatic patent ductus arteriosus had lower LTV than controls with the same severity of RDS. STV was predominantly influenced by postnatal and conceptional age, and tended to be lower in infants with periventricular haemorrhage. heart rate variability; respiratory distress syndrome; patent ductus periventricular haemorrhage; neonatal monitoring; preterm infant
arteriosus;
Correspondence to: C.M.A. van Ravenswaaij-Arts, Department of Pediatrics. University Hospital Nijmegen, P.O. Box 9101, 6500 HB Nijmegen, The Netherlands. Abbreviations: AGA, appropriate for gestational age; CPAP, continuous positive airway pressure: HRV, heart rate variability; IPPV, inspiratory positive pressure ventilation; LTV, long term variability: RDS, respiratory distress syndrome; RR, RR-interval length; SGA, small for gestational age; STV. short term variability; tcPco,, transcutaneous Pco?: tcPo,, transcutaneous PO?.
0378-3782/91/$3.50 0 1991 Elsevier Scientific Published and Printed in Ireland
Publishers
Ireland
Ltd.
208
Introduction The spontaneous cardiac rhythm is modified by the activity of autonomic nerves and by neurotransmitters. The cardioregulatory mechanisms at the basis of this phenomenon of heart rate variability (HRV) are affected by respiratory activity, baroreceptor activity and the peripheral vascular resistance [9]. Fast fluctuations (short term variability; STV) are mainly caused by respiration and mediated by the parasympathetic system. Slow fluctuations (long term variability; LTV) are caused by changes in vascular resistance and mediated by both the vagal and sympathetic system [3]. After several publications in the seventies and early eighties [ 13,I4,16] on the relation between various pathological neonatal conditions and HRV, recording of HRV seemed to be attractive as a simple tool for routine monitoring of neonatal condition in the intensive care unit. HRV, though, is influenced by many factors which can hamper correct interpretation. The effect of clinical condition on neonatal HRV may be partially explained by the influence of confounding maturational and physiological factors. For instance, newborns with respiratory distress syndrome have a higher respiratory rate, often show disturbed blood gases and are less mature than newborns without respiratory insufficiency. To investigate the relative contributions of neonatal complications on HRV a multi-parametric study was performed. A study group of preterm newborns with a gestational age below 33 weeks was chosen. Newborns below this age are at increased risk for neonatal complications: i.e. respiratory distress syndrome (RDS), patent ductus arteriosus and periventricular haemorrhage. In a previous paper [27] we focused on the influence of the most important maturational (birthweight, conceptional age and age after birth) and physiological (respiration, blood gases and behavioural state) factors on HRV in 29 spontaneously breathing preterm infants. The present paper reports on the influence of RDS, patent ductus arteriosus and periventricular haemorrhage on HRV in 60 very preterm infants. The main question to be answered in the study was whether the negative influence of RDS on HRV found by others [1,2,5,13,14,16] could be confirmed or should be attributed to changes in confounding variables (e.g. higher respiratory rate or increased transcutaneous PC02 in infants with RDS). Since periventricular haemorrhage and patent ductus arteriosus are well known complications in very preterm infants, the influence of these conditions on neonatal HRV was also investigated. Subjects and Methods Subjects Infants were allowed to enter the study if they were born at a gestational age below 33 weeks and if they were admitted to the neonatal intensive care unit within 12 h after birth. Exclusion criteria were presence of congenital malformations, need for resuscitation at birth and use of cardiovascular or sedative agents by mother or infant. The entire protocol, lasting until the age of 72 h after birth, could be completed for 50 of the 69 infants that were originally included into the study group. Ten in-
209
fants had to be excluded after the first day because of anti-epileptic treatment (n = 3) use of dopamine (n = 6) and early discharge (n = 1). The data of these 10 infants which were collected on day 1 only, were used in the data-analyses performed per day (n = 60, 50 and 50 on days 1, 2 and 3, respectively). Nine infants had to be excluded before the end of the first day. The data of these 9 infants were not included in the analyses. The most important patient characteristics are tabulated in Table I. Apgar scores after 1 and 5 min did not significantly differ between the RDS groups. Infants were treated with indomethacin if a symptomatic patent ductus arteriosus occurred (loading dose 0.2 mg/kg i.v., maintenance dose 0.1 mg/kg i.v. 12 and 36 h thereafter). Data acquisition Heart rate, respiration, transcutaneous POD and PCO~ and motility were assessed at regular intervals (09:30, 13:30, 17:OO and 2 I:30 h) until 72 h after birth. This period was chosen since most of the adaptive processes develop during the first 3 days after birth. Recordings were obtained while the infants were asleep and at least 1 h after the last intervention (e.g. feeding, endotracheal suctioning). TABLE
I
Patient characteristics Number Number of infants (n) conceptional age (weeks) birthweight (g) Small-for-gestational ageb (n) conceptional age (weeks) birthweight (g) Patent ductus arteriosus spontaneous closure in (n) at (h) indomethacin
started
in (n)
haemorrhagee
Range
30 1415
650-2280
60 (50)” 26-32
6 30.5 890
27-31 650- 1000
II
O-63
38
12-42
Number
Overall RDS’
Number of infants
32 (23)’ 25d (2O)C
at (h) Periventricular
Median
Grade Grade Grade
I 2 3
I (I)” 5 (2) 3 (3)
No Mild Moderate
12 (II)” 15 (15) 20 (15)
Grade
4
1 (0)
Severe
13 (9)
“Number of completely measured infants (until the age of 72 h) in parentheses. bBirthweight below tenth percentile [15]. CBetween brackets: within study period (9 spontaneous closure before first measurement. 5 indomethacin treatment started after the third day of life). dOf the 25 infants treated with indomethacin. 13 had moderate and I2 severe RDS. Wassification according to Trounce and Levene [23]. ‘Overall RDS = worst RDS score according to Kero 1141 that was given at least three times to an infant (see Table IV for distribution of RDS severity of each postnatal day).
210
The R-waves in the ECG, respiration curve and rate (Hewlett Packard-monitor 78801B), transcutaneous (tc) Paz and Pcoz (Novametrix tcozmette 809A or Sensormedics Transend) were stored on-line in a PDP-1 l/23 micro-computer (see Ref. 27). Movements were observed by one examiner and the incidence and duration were stored in the micro-computer simultaneously with the other parameters. Distinction was made between small movements (hand or foot), isolated limb movements, startles and more complex movements (general, rotating movements and stretches). In 36 of the 60 infants both tcPc@ and tcPo2, while in the others only tcPoz was measured. Registrations were continued until three periods of 3 min in stable condition were marked by the observer. The infant was then screened for patent ductus arteriosus and periventricular haemorrhage by echo-Doppler investigation. Periventricular haemorrhage was scored according to Trounce and Levene [23]. Patent ductus arteriosus was considered to be present if both a continuous forward and a diastolic backward flow were found in the main pulmonary artery [6]. The severity of RDS was scored according to Kero [ 161 as class I (mild), II (moderate) or III (severe), based on ventilator therapy, blood gases, fraction of inspired oxygen, chest X-ray and clinical symptoms. The total number of measurement series was 600. Data processing After an artefact-rejection procedure [27], for each 3-min period four long term variability (LTV) and four short term variability (STV) indices were calculated according to Corometrics [13] (LTV-l (bpm) and STV-1 (ms)), Yeh [28] (LTV-2 and STV-2), Huey [12] (LTV-3 (ms) and STV-3 (ms)) and Van Geijn [24] (LTV-4 (ms) and STV4). Also the mean RR-interval (mean RR (ms)) and mean and standard deviation (S.D.) of respiratory frequency (per min) and transcutaneous blood gases (mean and S.D. tcPo* and tcPc@ (mmHg)) were calculated for each 3-min period. Sleep state was scored as state coincidence 1 (Cl, ‘quiet sleep’) if none or only a few small movements occurred and if respiration was regular during the entire 3-min period. A period was classified as state coincidence 2 (C2, ‘active sleep’) if respiration was irregular during the 3 min. Non-classifiable periods or periods with a transition from Cl to C2 or reverse were classified as no-coincidence (NC). This simplified classification is based on the vectors of behavioural state described by Prechtl and O’Brien [ 181, while the name ‘state coincidence’ was adopted from Nijhuis et al. [ 171. In spontaneously breathing infants, periods with evident periodic breathing (i.e. short periods of apnea lasting 5-10 s, alternated by regular breathing periods [4]) were marked. Statistics The registrations of the 50 infants who were measured during the entire 3-daystudy period were used in most analyses. The data of the 10 infants who were measured only during the first day of life were used when analyses were performed for each postnatal day separately. Each measurement series consisted of three periods. The median values of HRV, mean RR and mean and S.D. of respiratory frequency, tcPo2 and tcPcoz were used to investigate the influence of RDS, patent ductus arteriosus, periventricular haemorrhage and indomethacin administration.
211
As in our previous study regarding spontaneously breathing preterm infants, the influence of conceptional age, birthweight, age after birth, respiration and blood gases on HRV has been analyzed by calculating Spearman correlation coefficients (see Ref. 27). These analyses have been repeated in the current study to investigate whether the relationships found in preterm infants with an uncomplicated course were also valid in ill infants. The effects of sleep state, indomethacin administration, closure of the ductus arteriosus and artificial ventilation were investigated intra-individually with the paired t-test. The effect of periventricular haemorrhage and symptomatic patent ductus on HRV was studied in matched patient-control pairs (paired t-test). Conconceptional age, postnatal age and trols were matched for RDS severity, behavioural state. Kruskal-Wallis test was used to analyze the differences between RDS classes regarding HRV, respiration and blood gases. Finally, a stepwise regression procedure [21] was used to explore which of the independent variables conceptional age, birthweight, age, mean RR, mean and SD. of respiratory rate and tcPO,, extent of movements, RDS score, existence of patent ductus arteriosus and periventricular haemorrhage and indomethacin-administration contributed at the most to STV and LTV. The incompletely measured and the small for gestational age @GA) infants were not included in this analysis. The latter were excluded since in SGA infants HRV appeared to be less related to postnatal age than in appropriate for gestational age (AGA) infants [27]. All analyses were repeated for state coincidence 1 and 2 and for each postnatal day separately. The results of these extra analyses are only mentioned when additional information was obtained. Results were considered significant at the 0.05 (ttest, Kruskal-Wallis test) or 0.01 level (Spearman and stepwise regression). Ethical approval Informed parental consent was obtained for each infant and the study was approved by the Ethical Committee of the University Hospital Nijmegen. Results Maturational and physiological influences on HR V The correlation coefficients between HRV and conceptional age, birthweight, age after birth, heart rate, respiration and blood gases were comparable to those found in our previous study [27]. A positive correlation between HRV and conceptional age, birthweight, age, standard deviation of respiratory frequency and standard deviation of tcPo* was found. LTV was negatively correlated with respiratory rate (Table II). Therefore all these parameters were included in the final stepwise regression analysis. Sleep states were distributed as follows: C 1, C2 and NC during, respectively, 47X, 44% and 9% of all 3-min periods. During C2 periods a significant increase in all LTV-indices and standard deviation of respiratory frequency was found compared to Cl periods (Table III). To exclude the effect of movements, the analysis was limited to periods with small (hand or foot) movements.
212
?? ? ? ? ?????????
??
????
\D\D\Dp‘m*Dmm 00000000
d
0. 0’ 0. 0. 0. d
VVIV-VVV
-Po\bP\o*ooO
$Lp?~~v?x~x~
0.
213
Patent ductus arteriosus and indomethacin A paired t-test between HRV, derived from the last measurement before and the first measurement after cessation of the ductal left-to-right shunt, revealed no signilicant differences in mean RR and HRV (n = 37). Also no difference in HRV before and after the first administration of indomethacin was found (n = 20). To explore the influence of symptomatic ductus patency the last measurement before indomethacin administration was compared with a measurement in a control infant matched for RDS class, age, conceptional age and behavioural state (n = 7). Significant differences were found for mean RR, LTV-2, LTV-4, STV-3 and STV-4 (Table III). Periventricular haemorrhage Although the 6 infants with periventricular haemorrhage (Table I) tended to have lower HRV (especially STV) than their matched controls, a t-test on the differences in HRV of the matched pairs did not show significance (Table III). Respiratory distress syndrome Since RDS severity may change from day to day it was scored at each measurement series. The most obvious results of the Kruskal-Wallis test regarding the influence of RDS class on conceptional age- and birthweight-distribution, HRV, heart
TABLE Influence
III of behavioural
state, symptomatic
patent
ductus
arteriosus
and periventricular
haemorrhage.
Mean values (S.E.M.) of LTV and STV according to Corometrics (-1. (bpm) and (ms)), Yeh (-2. (x lo-‘) and (dimensionless)), Huey (-3. (ms) and (ms)), Van Geijn (-4. (ms) and (dimensionless)), SD. of respiratory frequency (SD. RF, (per min)) and RR-interval length (RR. (ms)) during Cl and C2 periods. in infants with symptomatic patent ductus arteriosus (PDA) and their matched controls and in infants with periventricular haemorrhage (PVH) and their matched controls (during Cl).
Cl c2
n
LTV- I
LTV-2
LTV-3
LTV-4
SD.
39 39
8.0 (0.6) 10.3 (0.7)
76 (9) 93 (9)
Human Development,
Elsevier Scientific
Publishers
21 (1991)
Ireland
207
207-221
Ltd.
EHD 01203
The influence of respiratory distress syndrome heart rate variability in very preterm infants
on
Conny M.A. van Ravenswaaij-Arts”, Jeroen C.W. Hopman”, Louis AA. Kollke”, Joop P.L. van Amen”, Gerard B.A. Stoelinga” and Herman P. van Geijnb “Department
of Pediatrics,
University
Hospital
Free University
(Received
6 May 1991;
Nijmegen and hDepariment
Hospital
Amsterdam
oJ Obstetrics and Gynecology,
(The Netherlands)
revision received 31 October 1991; accepted 6 November 1991)
Summary In a multi-parametric study the influence of pathological neonatal conditions on heart rate variability was investigated in 60 preterm infants born at a gestational age below 33 weeks. Measurements were performed during the first 3 days of life. Four times a day, RR-intervals, respiration curve and rate, transcutaneously measured blood gases and observed body movements were recorded while the infants were asleep. All data were stored simultaneously in a micro-computer. Severity of respiratory distress syndrome (RDS), patency of ductus arteriosus and periventricular haemorrhage were documented as well. Four sets of short- (STV) and long-term variability (LTV) indices were calculated. Severe RDS was associated with a signiticant decrease in LTV. The influence of RDS on LTV persisted after correction for conceptional age, postnatal age, behavioural state and variations in respiratory rate and in transcutaneous PO,. Infants with a symptomatic patent ductus arteriosus had lower LTV than controls with the same severity of RDS. STV was predominantly influenced by postnatal and conceptional age, and tended to be lower in infants with periventricular haemorrhage. heart rate variability; respiratory distress syndrome; patent ductus periventricular haemorrhage; neonatal monitoring; preterm infant
arteriosus;
Correspondence to: C.M.A. van Ravenswaaij-Arts, Department of Pediatrics. University Hospital Nijmegen, P.O. Box 9101, 6500 HB Nijmegen, The Netherlands. Abbreviations: AGA, appropriate for gestational age; CPAP, continuous positive airway pressure: HRV, heart rate variability; IPPV, inspiratory positive pressure ventilation; LTV, long term variability: RDS, respiratory distress syndrome; RR, RR-interval length; SGA, small for gestational age; STV. short term variability; tcPco,, transcutaneous Pco?: tcPo,, transcutaneous PO?.
0378-3782/91/$3.50 0 1991 Elsevier Scientific Published and Printed in Ireland
Publishers
Ireland
Ltd.
208
Introduction The spontaneous cardiac rhythm is modified by the activity of autonomic nerves and by neurotransmitters. The cardioregulatory mechanisms at the basis of this phenomenon of heart rate variability (HRV) are affected by respiratory activity, baroreceptor activity and the peripheral vascular resistance [9]. Fast fluctuations (short term variability; STV) are mainly caused by respiration and mediated by the parasympathetic system. Slow fluctuations (long term variability; LTV) are caused by changes in vascular resistance and mediated by both the vagal and sympathetic system [3]. After several publications in the seventies and early eighties [ 13,I4,16] on the relation between various pathological neonatal conditions and HRV, recording of HRV seemed to be attractive as a simple tool for routine monitoring of neonatal condition in the intensive care unit. HRV, though, is influenced by many factors which can hamper correct interpretation. The effect of clinical condition on neonatal HRV may be partially explained by the influence of confounding maturational and physiological factors. For instance, newborns with respiratory distress syndrome have a higher respiratory rate, often show disturbed blood gases and are less mature than newborns without respiratory insufficiency. To investigate the relative contributions of neonatal complications on HRV a multi-parametric study was performed. A study group of preterm newborns with a gestational age below 33 weeks was chosen. Newborns below this age are at increased risk for neonatal complications: i.e. respiratory distress syndrome (RDS), patent ductus arteriosus and periventricular haemorrhage. In a previous paper [27] we focused on the influence of the most important maturational (birthweight, conceptional age and age after birth) and physiological (respiration, blood gases and behavioural state) factors on HRV in 29 spontaneously breathing preterm infants. The present paper reports on the influence of RDS, patent ductus arteriosus and periventricular haemorrhage on HRV in 60 very preterm infants. The main question to be answered in the study was whether the negative influence of RDS on HRV found by others [1,2,5,13,14,16] could be confirmed or should be attributed to changes in confounding variables (e.g. higher respiratory rate or increased transcutaneous PC02 in infants with RDS). Since periventricular haemorrhage and patent ductus arteriosus are well known complications in very preterm infants, the influence of these conditions on neonatal HRV was also investigated. Subjects and Methods Subjects Infants were allowed to enter the study if they were born at a gestational age below 33 weeks and if they were admitted to the neonatal intensive care unit within 12 h after birth. Exclusion criteria were presence of congenital malformations, need for resuscitation at birth and use of cardiovascular or sedative agents by mother or infant. The entire protocol, lasting until the age of 72 h after birth, could be completed for 50 of the 69 infants that were originally included into the study group. Ten in-
209
fants had to be excluded after the first day because of anti-epileptic treatment (n = 3) use of dopamine (n = 6) and early discharge (n = 1). The data of these 10 infants which were collected on day 1 only, were used in the data-analyses performed per day (n = 60, 50 and 50 on days 1, 2 and 3, respectively). Nine infants had to be excluded before the end of the first day. The data of these 9 infants were not included in the analyses. The most important patient characteristics are tabulated in Table I. Apgar scores after 1 and 5 min did not significantly differ between the RDS groups. Infants were treated with indomethacin if a symptomatic patent ductus arteriosus occurred (loading dose 0.2 mg/kg i.v., maintenance dose 0.1 mg/kg i.v. 12 and 36 h thereafter). Data acquisition Heart rate, respiration, transcutaneous POD and PCO~ and motility were assessed at regular intervals (09:30, 13:30, 17:OO and 2 I:30 h) until 72 h after birth. This period was chosen since most of the adaptive processes develop during the first 3 days after birth. Recordings were obtained while the infants were asleep and at least 1 h after the last intervention (e.g. feeding, endotracheal suctioning). TABLE
I
Patient characteristics Number Number of infants (n) conceptional age (weeks) birthweight (g) Small-for-gestational ageb (n) conceptional age (weeks) birthweight (g) Patent ductus arteriosus spontaneous closure in (n) at (h) indomethacin
started
in (n)
haemorrhagee
Range
30 1415
650-2280
60 (50)” 26-32
6 30.5 890
27-31 650- 1000
II
O-63
38
12-42
Number
Overall RDS’
Number of infants
32 (23)’ 25d (2O)C
at (h) Periventricular
Median
Grade Grade Grade
I 2 3
I (I)” 5 (2) 3 (3)
No Mild Moderate
12 (II)” 15 (15) 20 (15)
Grade
4
1 (0)
Severe
13 (9)
“Number of completely measured infants (until the age of 72 h) in parentheses. bBirthweight below tenth percentile [15]. CBetween brackets: within study period (9 spontaneous closure before first measurement. 5 indomethacin treatment started after the third day of life). dOf the 25 infants treated with indomethacin. 13 had moderate and I2 severe RDS. Wassification according to Trounce and Levene [23]. ‘Overall RDS = worst RDS score according to Kero 1141 that was given at least three times to an infant (see Table IV for distribution of RDS severity of each postnatal day).
210
The R-waves in the ECG, respiration curve and rate (Hewlett Packard-monitor 78801B), transcutaneous (tc) Paz and Pcoz (Novametrix tcozmette 809A or Sensormedics Transend) were stored on-line in a PDP-1 l/23 micro-computer (see Ref. 27). Movements were observed by one examiner and the incidence and duration were stored in the micro-computer simultaneously with the other parameters. Distinction was made between small movements (hand or foot), isolated limb movements, startles and more complex movements (general, rotating movements and stretches). In 36 of the 60 infants both tcPc@ and tcPo2, while in the others only tcPoz was measured. Registrations were continued until three periods of 3 min in stable condition were marked by the observer. The infant was then screened for patent ductus arteriosus and periventricular haemorrhage by echo-Doppler investigation. Periventricular haemorrhage was scored according to Trounce and Levene [23]. Patent ductus arteriosus was considered to be present if both a continuous forward and a diastolic backward flow were found in the main pulmonary artery [6]. The severity of RDS was scored according to Kero [ 161 as class I (mild), II (moderate) or III (severe), based on ventilator therapy, blood gases, fraction of inspired oxygen, chest X-ray and clinical symptoms. The total number of measurement series was 600. Data processing After an artefact-rejection procedure [27], for each 3-min period four long term variability (LTV) and four short term variability (STV) indices were calculated according to Corometrics [13] (LTV-l (bpm) and STV-1 (ms)), Yeh [28] (LTV-2 and STV-2), Huey [12] (LTV-3 (ms) and STV-3 (ms)) and Van Geijn [24] (LTV-4 (ms) and STV4). Also the mean RR-interval (mean RR (ms)) and mean and standard deviation (S.D.) of respiratory frequency (per min) and transcutaneous blood gases (mean and S.D. tcPo* and tcPc@ (mmHg)) were calculated for each 3-min period. Sleep state was scored as state coincidence 1 (Cl, ‘quiet sleep’) if none or only a few small movements occurred and if respiration was regular during the entire 3-min period. A period was classified as state coincidence 2 (C2, ‘active sleep’) if respiration was irregular during the 3 min. Non-classifiable periods or periods with a transition from Cl to C2 or reverse were classified as no-coincidence (NC). This simplified classification is based on the vectors of behavioural state described by Prechtl and O’Brien [ 181, while the name ‘state coincidence’ was adopted from Nijhuis et al. [ 171. In spontaneously breathing infants, periods with evident periodic breathing (i.e. short periods of apnea lasting 5-10 s, alternated by regular breathing periods [4]) were marked. Statistics The registrations of the 50 infants who were measured during the entire 3-daystudy period were used in most analyses. The data of the 10 infants who were measured only during the first day of life were used when analyses were performed for each postnatal day separately. Each measurement series consisted of three periods. The median values of HRV, mean RR and mean and S.D. of respiratory frequency, tcPo2 and tcPcoz were used to investigate the influence of RDS, patent ductus arteriosus, periventricular haemorrhage and indomethacin administration.
211
As in our previous study regarding spontaneously breathing preterm infants, the influence of conceptional age, birthweight, age after birth, respiration and blood gases on HRV has been analyzed by calculating Spearman correlation coefficients (see Ref. 27). These analyses have been repeated in the current study to investigate whether the relationships found in preterm infants with an uncomplicated course were also valid in ill infants. The effects of sleep state, indomethacin administration, closure of the ductus arteriosus and artificial ventilation were investigated intra-individually with the paired t-test. The effect of periventricular haemorrhage and symptomatic patent ductus on HRV was studied in matched patient-control pairs (paired t-test). Conconceptional age, postnatal age and trols were matched for RDS severity, behavioural state. Kruskal-Wallis test was used to analyze the differences between RDS classes regarding HRV, respiration and blood gases. Finally, a stepwise regression procedure [21] was used to explore which of the independent variables conceptional age, birthweight, age, mean RR, mean and SD. of respiratory rate and tcPO,, extent of movements, RDS score, existence of patent ductus arteriosus and periventricular haemorrhage and indomethacin-administration contributed at the most to STV and LTV. The incompletely measured and the small for gestational age @GA) infants were not included in this analysis. The latter were excluded since in SGA infants HRV appeared to be less related to postnatal age than in appropriate for gestational age (AGA) infants [27]. All analyses were repeated for state coincidence 1 and 2 and for each postnatal day separately. The results of these extra analyses are only mentioned when additional information was obtained. Results were considered significant at the 0.05 (ttest, Kruskal-Wallis test) or 0.01 level (Spearman and stepwise regression). Ethical approval Informed parental consent was obtained for each infant and the study was approved by the Ethical Committee of the University Hospital Nijmegen. Results Maturational and physiological influences on HR V The correlation coefficients between HRV and conceptional age, birthweight, age after birth, heart rate, respiration and blood gases were comparable to those found in our previous study [27]. A positive correlation between HRV and conceptional age, birthweight, age, standard deviation of respiratory frequency and standard deviation of tcPo* was found. LTV was negatively correlated with respiratory rate (Table II). Therefore all these parameters were included in the final stepwise regression analysis. Sleep states were distributed as follows: C 1, C2 and NC during, respectively, 47X, 44% and 9% of all 3-min periods. During C2 periods a significant increase in all LTV-indices and standard deviation of respiratory frequency was found compared to Cl periods (Table III). To exclude the effect of movements, the analysis was limited to periods with small (hand or foot) movements.
212
?? ? ? ? ?????????
??
????
\D\D\Dp‘m*Dmm 00000000
d
0. 0’ 0. 0. 0. d
VVIV-VVV
-Po\bP\o*ooO
$Lp?~~v?x~x~
0.
213
Patent ductus arteriosus and indomethacin A paired t-test between HRV, derived from the last measurement before and the first measurement after cessation of the ductal left-to-right shunt, revealed no signilicant differences in mean RR and HRV (n = 37). Also no difference in HRV before and after the first administration of indomethacin was found (n = 20). To explore the influence of symptomatic ductus patency the last measurement before indomethacin administration was compared with a measurement in a control infant matched for RDS class, age, conceptional age and behavioural state (n = 7). Significant differences were found for mean RR, LTV-2, LTV-4, STV-3 and STV-4 (Table III). Periventricular haemorrhage Although the 6 infants with periventricular haemorrhage (Table I) tended to have lower HRV (especially STV) than their matched controls, a t-test on the differences in HRV of the matched pairs did not show significance (Table III). Respiratory distress syndrome Since RDS severity may change from day to day it was scored at each measurement series. The most obvious results of the Kruskal-Wallis test regarding the influence of RDS class on conceptional age- and birthweight-distribution, HRV, heart
TABLE Influence
III of behavioural
state, symptomatic
patent
ductus
arteriosus
and periventricular
haemorrhage.
Mean values (S.E.M.) of LTV and STV according to Corometrics (-1. (bpm) and (ms)), Yeh (-2. (x lo-‘) and (dimensionless)), Huey (-3. (ms) and (ms)), Van Geijn (-4. (ms) and (dimensionless)), SD. of respiratory frequency (SD. RF, (per min)) and RR-interval length (RR. (ms)) during Cl and C2 periods. in infants with symptomatic patent ductus arteriosus (PDA) and their matched controls and in infants with periventricular haemorrhage (PVH) and their matched controls (during Cl).
Cl c2
n
LTV- I
LTV-2
LTV-3
LTV-4
SD.
39 39
8.0 (0.6) 10.3 (0.7)
76 (9) 93 (9)
