Developmental Patterns of Heart Rate and Heart Rate Variability ...

Sleep, 5(1):28-38 © 1982 Raven Press, New York

Developmental Patterns of Heart Rate and Heart Rate Variability During Sleep and Waking in Normal Infants and Infants at Risk for the Sudden Infant Death Syndrome *Ronald M. Harper, *Barbara Leake, tJoan E. Hodgman, and tToke Hoppenbrouwers *Department of Anatomy. and the Brain Research Institute. University of California at Los Angeles; and tDepartment of Pediatrics, University of Southern California and Newhorn Division, LAC-USC Medical Center, Los Angeles, California

Summary: The developmental sequence of heart rate and heart rate variability was examined during sleep and waking states in 22 normal infants and 22 siblings of sudden infant death syndrome (SIDS) victims, using 12-h polygraphic recordings at 1 week and at 1,2,3,4, and 6 months of age. Heart rate was higher in siblings of SIDS victims than in normal infants during quiet sleep over the first 6 months of life and was higher in the waking state at 3 months of age. The sibling group also had lower variability at I week during quiet sleep. Gender contributed no significant differences to heart rate, but females at risk for SIDS had lower waking cardiac variability than males. Key Words: Heart rate-Heart rate variability-Sleep-Infants-Development-Sudden infant death syndrome-Gender.

State-related alterations in cardiac activity may yield important clues to the mechanisms of the sudden infant death syndrome (SIDS). Victims of this disorder are apparently healthy infants who die unexpectedly during sleep, apparently from asphyxia (1-3). In a previous study (4), we found evidence to suggest that there are state-related alterations in cardiac activity in infants at risk for SIDS. This earlier study indicated sleep-related cardiac disturbances in infants at risk, but the group sizes were too small to indicate the source of variability. For these reasons, this larger study was undertaken to partition the role of gender as a source of rate Accepted for publication September 1981. Address correspondence and reprint requests to Dr. Harper at Department of Anatomy, School of Medicine, UCLA, Los Angeles, California 90024.

28

DEVELOPMENTAL CARDIAC PATTERNS

29

and variability differences. Further, we wished to examine the possibility of specific heart rate differences in quiet and active sleep states for these at-risk infants. MATERIALS AND METHODS The experimental groups for this study comprised 22 normal infants and 22 subsequent siblings of SIDS victims. All infants had I-min Apgar scores between 8 and 10, and all were full-term and appropriate weight for gestational age, according to the intrauterine growth curve of Usher and McLean (5). Each infant was admitted at 5 PM to the sleep laboratory for 12-h all-night recording sessions that were held at 1 week and again at 1, 2, 3, 4, and 6 months of age. The mothers of both normal infants and subsequent siblings were screened for any history of diabetes or other illness. All parents were informed about the nature and objectives of the study, and provided written consent prior to participation. Subsequent siblings of SIDS victims (SSIDS) were chosen as an at-risk group because of a demonstrated higher risk for SIDS than the general population (6,7). The at-risk and normal infants were matched according to the educational level of their parents. Parity in educational level was considered important to avoid bias, since educational and socioeconomic levels are highly correlated and some epidemiological evidence suggests that socioeconomic level may be inversely related to incidence of SIDS (1). Both sleep and cardiopulmonary variables were monitored and recorded by procedures described in detail elsewhere (8,9). Each minute of every recording was coded into quiet sleep (QS), active sleep (AS), waking (A W), or indeterminate state by personnel trained to an 80% agreement on state classification. The stateselection criteria and decision-making rules have been described previously (9). State codes were stored on digital tape for correlation with heart rate and variability data. The R wave of the electrocardiograph (ECG) signal was identified by an electronic peak detector with an accuracy of 2 ms. Intervals between R waves were then calculated and stored on digital tape. For each minute, the heart rate value was computed by measuring the median R-R interval during that minute and converting this value to beats/min. The interquartile range of the intervals was used to measure cardiac variability; it was obtained by computing the absolute value of the difference between the first and third quartile limits of R-R interval lengths during a given minute and dividing this value by 60 to give the variability within that minute. The median and interquartile ranges were chosen as measures of cardiac rate and variability, respectively, because they are relatively insensitive to occasional abnormal or artifactual deviations in recordings. The median and interquartile range values for each infant at every age and state were averaged to give representative cardiac rate and variability values for statistical analysis and description. For convenience, these averaged median and interquartile range values for each infant will be referred to as individual heart rate and variability values. Artifacts in the ECG signal consisted primarily of excessively long or short R-R intervals. Since inclusion of such data in the calculations would have given erroneous results, artifact removal procedures were employed as described in Harper

Sleep. Vol. 5, No.1, 1982

30

R. M. HARPER ET AL.

et al. (9). Fewer than 10% of the minutes in any recording had to be removed because of artifactual contamination; most of these contaminated minutes occurred during wakefulness. Except for these data deletions, R-R intervals over the entire 12-h recordings were used to derive results. Statistical assessment of the influence of age, state, group (i.e., at-risk or normal), and gender on cardiac activity was achieved by submitting individual heart rate and heart rate variability values to a mixed-model, repeated-measures analysis of variance program (10). Subanalyses were performed for each state separately when significant interactions involving this factor were found. Specific changes in cardiac activity with age and state were described by Duncan's multiple range tests. RESULTS Heart Rate Averaged median heart rates for normal and SSIDS infants are plotted in Fig. 1 as a function of age and state. This figure suggests that the SSIDS had higher heart rates than the normal infants in QS during the first 6 months of life. Analysis of variance performed separately for each state, with age as a trial factor and atrisklnonrisk status as a grouping variable, did indeed reveal a significantly higher overall mean heart rate in the SSIDS group during QS (p < 0.025). As suggested by Fig. 1, the overall differences in the two groups' heart rates during AS and AW did not reach a 5% level of significance. Since group differences at certain ages are of interest, further tests were performed within each age and state. These tests indicated that the 3-month heart rate of the at-risk group was significantly higher than that of the normal group in both QS and A W (p < 0.05). When the gender of the individual infants was included in the analysis of variance program as a second grouping variable, the preceding results were essentially repeated, although the overall group difference in QS dropped slightly to a 3% level of significance and the group difference at 3 months in A W rose to a 2% level. The higher mean heart rates observed in the sibling as opposed to the normal group were thus not derived from differing numbers of male and female infants. The effect of gender on heart rate also was examined independently in both normal and at-risk infants. Although the mean heart rates of the females were generally greater than those of the males, none of the differences was significant at the 5% level (Fig. 2). The developmental course of mean heart rate for both normal and SSIDS groups decreased significantly between birth and 6 months of age in all states (p < 0.001). Nevertheless, some group and state differences were observed within these developmental trends. While the heart rate of both groups increased sharply between 1 week and 1 month of age in QS and AS (p < 0.005), the normal group also showed a significant heart rate increase during this period in AW (p < 0.01), whereas the at-risk group's increase in waking heart rate was not significant. Then, during the 1- to 2-month age period, the normal group's heart rate dropped significantly (p < 0.05) but relatively slowly in both sleep states, whereas the SSIDS

Sleep, Vol. 5, No.1, 1982

31

DEVELOPMENTAL CARDIAC PATTERNS AS

AW

170

OS

ss CT

120 110 lwk 1

2

3

4

6

AGE IN MONTHS *P 0.005) and subsequently increased between 1 and 3 months (p < 0.01) and again between 4 and 6 months of age (p < 0.05). The siblings showed a similar pattern, with a decline between 1 week and 1 month (p < 0.05), followed by a pronounced increase between 1 and 3 months (p < 0.001). The normal infants' AS heart rate variability declined between 1 and 3 months (p < 0.05) and again between 3 and 6 months of age. In contrast, the at-risk infants had a steep drop in their AS cardiac variability at 1 month (p < 0.005), similar to the pattern in QS. This initial drop in their AS variability values was followed by a significant increase between 1 and 2 months (p < 0.05), and then decreases in the 2- to 4-month and 4- to 6-month age periods (p < 0.01). In A W, both the normal and at-risk infants' cardiac variability increased between 1 week and 2 months of age (p < 0.01 and 0.05, respectively). Subsequently, however, the cardiac variability of the normal group declined between 2 and 4 months (p < 0.005), whereas the SSIDS group's variability values did not change significantly. As can be inferred from Fig. 3, both the SSIDS and normal groups had lower heart rate variability during QS than in either of the two other states between birth and 6 months (p < 0.001). Occasionally, there were also occurrences of significantly lower cardiac variability in AS than in AW. By 4 months of age, heart rate variability by state separated further, with highest values in A W, followed by intermediate values in AS, and lowest values in QS (p < 0.05). DISCUSSION This study confirms the preliminary findings of increased heart rate in both sleep and waking states during the first postnatal month, followed by decrements in heart rate over the next 5 months (4). Further, it supports a previous suggestion that infants at risk for SIDS have a lag in the 1- to 3-month heart rate decline seen

, Sleep. Vol. 5. No.1. 1982

R. M. HARPER ET AL.

34

03

1

CT

ss

AS

~, \~~M

M

F

FIG. 4. Heart rate variability partitioned by gender for CT and SS infants over the first 6 months of life during sleep and waking states. Note the change in order of variability for CT and SS infants during A W.

:~~F ~1O

~M

~9

~8 ~7 w ~6

5 4 "1M< 1 2 3 4 6 AGE N tvlONTHS

in normal infants, along with higher overall heart rates at 3 months of age. With the addition of 12 infants to both normal and SSIDS groups, however, striking group differences have emerged in QS: the at-risk infants display significantly higher heart rates in this state during the first 6 months of life and lower heart rate variability in the postnatal period.

Sleep. Vul. 5, Nu. I, 1982


35

The increased number of infants also has permitted an examination of gender effects on cardiac activity. These results include a lack of gender influence on heart rate (although both normal and SSIDS analyses indicate that a larger study might reveal higher heart rates in females around 1 month of age in QS), higher heart rate variability for females in the normal group during wakefulness, and higher waking heart rate variability for males in the at-risk group. The significance of the gender difference in waking heart rate variability, with at-risk females showing lower variation than at-risk males, is unknown. As most waking heart rate variability results from sources that are not related to respiration, these variability results may reflect differences in motility (11). The state-related differences in cardiac rate and variability observed here are also in accord with our previous studies (4,9). In particular, both the normal and sibling groups had highest heart rates in wakefulness, with intermediate rates in AS, and lowest rates in QS. Heart rate variability also was lowest in QS, with higher values in AS and AW. Although the reasons for the higher heart rates in the at-risk infants are unknown, there is suggestive anatomical evidence that SIDS victims have suffered from chronic hypoxia. This evidence includes increased muscle tissue in pUlmonary arteries, increased volume of chromaffin cells in the adrenal medulla, and abnormal retention of brown fat cells in the adrenals (12-14). The existence of a chronic hypoxemic condition might help to explain the abnormally high heart rates in the sibling group; the infant would attempt to compensate for the oxygen lack with an increase in cardiac output. A developing infant who possesses a heart with very little residual volume may be able to meet this demand only by an increase in heart rate (15,16). Further, chronic hypoxia would be expected to produce hyperpnea, and such a respiratory pattern has been observed in some of the SSIDS infants (17). Within the context of these results, it is necessary to consider why the disturbance of cardiac activity should be so predominant in quiet sleep. Phillipson (18) has suggested that control of respiration during QS is particularly dependent on chemoreceptor and vagal afferent activity, whereas respiration during phasic AS additionally appears to be controlled by a variety of other mechanisms (19). During wakefulness, of course, respiration is regulated by voluntary central mechanisms, as well as by vagal and chemoreceptor action. If the infant is particularly dependent on the integrity of chemoreceptor response during QS, then a reduction in central or peripheral chemoreceptor gain might lead to a relative hypoxia during that state and to compensatory increased heart rate. Chronic increased upper airway resistance during sleep might also lead to oxygen de saturation and increased heart rate. However, one would expect some external sign of such resistance such as the very loud snoring observed in adults, and such signs were not observed in these infants. Any causative mechanism proposed for these cardiac changes must accommodate results that indicate that the mechanisms are operative shortly after birth, since some of the signs of dysfunction appear in the first week oflife. As 3 months is a period of maximum risk for SIDS, and as the group differences in heart rate

Sleep. Vol. 5. No. I, 1982

36

R. M. HARPER ET AL.

are particularly pronounced at this time, one might speculate that initial disturbances in chemoreceptor response, continued over a long period of time, may greatly alter chemosensitivity by 3 months of age. There is a normal reduction in CO 2 response during QS (20) that might combine with impaired chemoreceptor sensitivity to predispose an infant to a prolonged and possibly fatal apneic episode. Reduced response to chemoreceptor challenge has been suggested from pulmonary tests in near-miss SIDS infants (21), and an impaired hypoxic response has been found in a group of siblings of SIDS victims (similar to the group under study here) (22). Such tests were not performed on the infants in this study, but could clearly help to elucidate the mechanisms underlying sudden death in infants. We have, to this point, considered chemoreceptor response in terms of a ventilatory response to chemical stimulation. It is important to note that part of this response might include arousal from a sleep state, so that a whole new set of neural respiratory-control mechanisms might then operate. If there is a deficit in this arousal mechanism (as distinct from the ventilatory response to chemical stimulation within a state), then the consequences could be fatal to an infant who needs to arouse from a sleep state in order to resume breathing. It might not be necessary to postulate a deficit in chemoreceptor mechanisms under this paradigm; instead, defects in the brain mechanisms of arousal would be the focus of interest. There is evidence that infants at risk for SIDS do indeed have an impaired ability to arouse from sleep (23). The principal difference in variability of heart rate was observed early in life, with at-risk infants having lower variability in QS. However, there is very little heart rate variability during QS that is not attributable to respiration (24). Thus, reduced variability in SSIDS probably results from a diminution in respiratoryrelated variation (sinus arrhythmia), as opposed to variation contributed from all other sources (such as movement). Mazza et al. (25) have demonstrated a linear relationship of heart rate to instantaneous R-R interval-change values in sleeping infants. Part of the decreased variability which we observed is undoubtedly related to the higher heart rates in QS, although only the findings at 1 week were significantly different in variability, and the principal rate differences were present at 3 months. Increased heart rate and decreased heart rate variability in both QS and AS have been noted in short sleep recordings from another group of infants at risk for SIDS, that of near-miss or aborted SIDS infants (26). Because these infants also demonstrate a shorter Q-T interval, increased sympathetic activity or increased levels of circulating catecholamines, perhaps secondary to hypoxia, have been suggested as a mechanism for these findings. Although this is certainly a possibility, reduced vagal tone might also be considered, since decreased parasympathetic activity will also directly affect ventricular repolarization (27). The decrease in cardiac rate variation does not appear to be specific to infants at risk for SIDS. Reduced variability also appears to be present in infants with respiratory distress syndrome (28- 30) and in at-risk infants with perinatal anoxia (31). In fact, the presence of minimal heart rate variation, or fixed heart rate, is a sign for poor prognosis in infants with respiratory distress syndrome (32). The

Sleep. Vol. 5, No. I, 1982


37

SSIDS group does not manifest fixed heart rate, but the group does tend to show reduced cardiac variation early in life, particularly in QS. Given the prominence of respiratory-related heart rate variation during QS, a study of the etiology of SIDS might well consider the state-related mechanisms for reduced cardiac variation from respiratory sources, as well as mechanisms for state-specific patterns of heart rate differences between normal and at-risk infants. ACKNOWLEDGMENT We thank S. Geidel, B. Havens, J. R. Mason, L. Miyahara, and D. Taube for their valuable contributions to this study. This research was supported by grants HD-1-4608, HD-4-2810, and HD-2-2777 from the National Institute of Child Health and Human Development, NIH. REFERENCES 1. Beckwith JB. The sudden infant death syndrome. Curr Prohl Pediatr 1973; 3:1-36. 2. Guntheroth WB, Breazeale D, McGough GA. The significance of pulmonary petechiae in crib death. Pediatrics 1973; 52:601-3. 3. Patrick JR. Cardiac or respiratory death. In: Bergman AB, Beckwith JB, Ray CG, eds, Sudden infant death syndrome. Seattle: University of Washington Press, 1970: 130. 4. Harper RM, Leake B, Hoppenbrouwers T, Sterman MB, McGinty DJ, Hodgman J. Polygraphic studies of normal infants and infants at risk for the sudden infant death syndrome: heart rate and variability as a function of state. Pediatr Res 1978; 12:778-85. 5. Usher R, McLean F. Intrauterine growth of live-born caucasian infants at sea level: standards obtained from measurements in seven dimensions of infants born between 25 and 44 weeks of gestation. J Pediatr 1969; 74:901-10. 6. Froggatt P, Lynas MA, MacKenzie G. Epidemiology of sudden unexpected deaths in infants (cot death) in Northern Ireland. Br J Prev Soc Med 1971; 25:119-34. 7. Peterson DR, Chinn NM, Fisher LD. The sudden infant death syndrome: repetitions in families. J Pediatr 1980; 97:265-7. 8. Hoffman E, Havens B, Geidel S, Hoppenbrouwers T, Hodgman JE. Longterm continuous monitoring of mUltiple physiological parameters in newborn and young infants. A procedural manual. Acta Pediatr Scand Suppl 1977; 266: 1-24. 9. Harper RM, Hoppenbrouwers T, Sterman MB, McGinty DJ, Hodgman J. Polygraphic studies of normal infants during the first six months oflife. I. Heart rate and variability as a function of state. Pediatr Res 1976; 10:945-51. 10. Dixon WJ. BMDP: biomedical computer programs. Los Angeles: UCLA, 1977. 11. Coons S, Guilleminault C, Challamel MJ. Sleep and wake patterns in the first six months. Sleep Res 1981; in press. 12. Naeye RL. Hypoxemia and the sudden infant death syndrome. Science 1974; 186:837-8. 13. Naeye RL. Brain stem and adrenal anomalies in the sudden infant death syndrome. Am J Clin Pathol 1976; 66:526-30. 14. Naeye RL. The sudden infant death syndrome. Arch Pathol Lab Med 1977; 101:165-7. 15. Gribbe P, Hirvonen L, Peltonen T. Cineangiocardiographic studies of puppies and full-grown dogs. Acta Physiol Scand 1961; 51:169. 16. Dawes OS. Sudden death in babies: physiology of the fetus and newborn. Am J Cardiol 1968; 22:469-78. 17. Hoppenbrouwers T, Hodgman JE, McGinty DJ, Harper RM, Sterman MB. Sudden infant death syndrome: sleep apnea and respiration in subsequent siblings. Pediatrics 1980; 66:205-14. 18. Phillipson EA. Respiratory adaptations in sleep. Annu Rev Physiol 1978; 40: 133-56. 19. Netick A, Foutz AS, Dement WC. Sleep state effects upon respiration following vagotomy and cord transection in the cat. Soc Neurosci 1977; 3:47. 20. Bulow K. Respiration and wakefulness in man. Acta Physiol Scand 1963; 59:209.

Sleep. Vol. 5, No. I, 1982

38

R. M. HARPER ET AL.

21. Shannon DC, Kelly D. Impaired regulation of alveolar ventilation and the sudden infant death syndrome. Science 1977; 197:367-8. 22. Brady JP, Partridge C, Durand M. Response to mild hypoxia differs in siblings of SIDS. Pediatr Res 1981; 15:652. 23. Harper RM, Leake B, Hoffman H, et al. Periodicity of sleep states is altered in infants at risk for the sUdden infant death syndrome. Science 1981; 213:1030-2. 24. Harper RM, Walter DO, Leake B, et al. Development of sinus arrhythmia during sleeping and waking states in normal infants. Sleep 1978; 1:33-48. 25. Mazza NM, Epstein MAF, Haddad GG, Law HS, Mellins RB, Epstein RA. Relation of beat-tobeat variability to heart rate in normal sleeping infants. Pediatr Res 1980; 14:232-5. 26. Leistner HL, Haddad GG, Epstein RA, Lai TL, Epstein MAF, Mellins RB. Heart rate and heart rate variability during sleep in aborted sudden infant death syndrome. J Pedialr 1980; 97:51-5. 27. Prystowsky EN, Jackman WM, Rinkenberger RL, Heger 11, Zipes DP. Effect of autonomic blockade on ventricular refractoriness and atrioventricular nodal conduction in humans: evidence supporting a direct cholinergic action on ventricular muscle refractoriness. eirc Res 1981; 49:511-18. 28. Kero P. Heart rate variation in infants with the respiratory distress syndrome. Acta Pediatr Scand Suppl 1974; 250: 1-70. 29. Vallbona C, Desmond MM, Rudolph AJ, Papp LF, Franklin RR, Rush JB. Cardiodynamic studies in the newborn. II. Regulation of the heart rate. BioI Neonate 1963; 5: 159-99. 30. Rudolph AJ, Vall bona C, Desmond MM. Cardiodynamic studies in the newborn. III. Heart rate patterns in infants with idiopathic respiratory distress syndrome. Pediatrics 1965; 36:551-9. 31. Miyazaki S, Watanabe K, Hara K. Heart rate variability in full-term normal and abnormal newborn infants during sleep. Brain Dev 1979; 1:57-60. 32. Urbach JR, Phuvichit B, Zweizig H, et al. Instantaneous heart-rate patterns in newborn infants. Am J Obstet Gynecol 1965; 93:965-74.

Sleep, Vol. 5, Nu. 1,1982