- Email: [email protected]
Extreme and Conventional Cardiorespiratory Events and Epidemiologic Risk Factors for SIDS
Extreme and Conventional Cardiorespiratory Events and Epidemiologic Risk Factors for SIDS TOKE HOPPENBROUWERS, PHD, JOAN E. HODGMAN, MD, ANUSHA RAMANATHAN, MD,
AND
FRED DOREY, PHD
Objective To test the hypotheses that there is a lack of correlation between extreme events and epidemiologic risk factors for sudden infant death syndrome (SIDS), and if conventional events are normal, their numbers should increase once a circadian decrease in breathing rate is established. In addition, the number of events should decrease with maternal smoking. Study design Three outcome variables were derived from the Collaborative Home Infant Monitoring Evaluation (CHIME) of 1082 infants: (1) at least 1 extreme event lasting >30 seconds, (2) at least 1 conventional event lasting >20 seconds, and (3) being part of the 50% of infants with the most events. Results Multivariate logistic regression analyses found that extreme events were not statistically associated with any known SIDS risk factors and occurred less often during the early morning. Healthy term infants had significantly fewer of these events compared with preterm infants, subsequent siblings of infants with SIDS, and infants with an apparent life-threatening event, a finding that was not evident after 43 weeks (3 weeks postterm). Conventional events increased during the night, whereas maternal smoking was associated with a decrease in conventional events. Apneic episodes persisting for >40 seconds occurred in 1.8% of the infants. Conclusions Extreme events are associated with immaturity and do not seem to be immediate precusors of or causally related to SIDS. (J Pediatr 2008;152:636-41) he historical focus on apnea as either a cause of sudden infant death syndrome (SIDS) or a mechanism of death has decreased because studies have failed to implicate these events directly.1,2 Before the 1990s, home monitoring equipment typically targeted apneic events as short as 20 seconds, which have been shown to be ubiquitous.3 These short apneic events have been designated conventional events based on traditional clinical practice. A second category of events, identified from documented monitoring, has been arbitrarily designated extreme events. These comprise episodes of bradycardia alone or apnea persisting for more than 30 seconds with or without bradycardia and oxygen desaturation.4 Their role in the etiology of SIDS also is questionable. From the Division of Neonatal Medicine and Biostatistics (F.D.), LAC ⫹ USC Medical The Collaborative Home Infant Monitoring Evaluation (CHIME) was carried out Center, Women’s and Children’s Hospital during the 1990s, when having infants sleep on their back was recommended to all and Children’s Hospital Los Angeles (T.H.), Keck School of Medicine, University of Southparticipants.5 More than 1000 infants participated in this study in 5 US clinical centers, ern California, Los Angeles, CA (T.H., J.H.) representing the largest cohort for which systematic home monitoring has been performed and New York Presbyterian Hospital-Weill Cornell Medical Center, New York, NY to date. CHIME found a high incidence of extreme cardiorespiratory events at early ages (A.R.). in preterm infants. However, at the age of highest risk for SIDS, the incidence of these Supported by the National Institute of 3 events was similar in preterm and term infants. This finding has raised further questions Child Health and Human Development (grants HD 29067, HD 29071, HD 28971, about the role of apnea in the etiology of SIDS and the possibility that either extreme or HD 29073, HD 29056, and HD 34625) conventional events can predict imminent SIDS deaths. and a generous donation from the Orange Circadian modulation of cardiac and respiratory rates, with decreases during the County Guild for Infant Survival. Submitted for publication Feb 20, 2007; last nighttime hours, is first observed after the first month of life in term infants.6 Because revision received Sep 18, 2007; accepted lower respiratory rates correlate with a higher incidence of breathing pauses, it is not Oct 1, 2007. 7 unreasonable to expect a higher incidence of conventional events during that period. Reprint requests: Toke Hoppenbrouwers,
T
ALTE AOI CHIME CI OR
636
Apparent life-threatening event Apnea of infancy Collaborative Home Infant Monitoring Evaluation Confidence interval Odds ratio
PMA ROC SIDS SSIDS
Postmenstrual age Receiver operating characteristic curve Sudden infant death syndrome Subsequent sibling of sudden infant death syndrome
PhD, Room 9L19, Women and Children’s Hospital, LAC ⫹ USC Medical Center, 1240 Mission Road, Los Angeles, CA 90033. E-mail: [email protected]. 0022-3476/$ - see front matter Copyright © 2008 Mosby Inc. All rights reserved. 10.1016/j.jpeds.2007.10.003
Table I. ORs for having at least 1 event based on SIDS and other risk factors after adjusting for preterm birth, ALTE, SSIDS1 and number of hours monitored PMA <44 weeks Extreme event SIDS risk factor
OR
P
Male predominance African-American infant Parity ⱖ2 Maternal age ⱕ18 years Maternal smoking Maternal education ⬍12 years Caucasian Hispanic Asian Placenta previa
0.77 0.54 0.79 0.62 1.0 1.5 1.01 1.0 1.5 6.63
.24 .054 .34 .28 .93 .14 .97 .99 .27 .003
95% CI 0.49 0.29 0.49 0.26 0.6 0.89 0.64 0.58 0.74 1.86
to to to to to to to to to to
PMA >44 weeks,
Conventional event
1.2 1 1.3 1.5 1.8 2.4 1.6 1.7 2.8 23.6
OR
P
1.4 0.56 1.07 1.29 0.66 0.72 1.33 1.43 0.87 0.33
.032 .009 .72 .50 .037 .11 .08 .08 .59 .17
95% CI 1.03 0.37 0.75 0.62 0.44 0.48 0.97 0.96 0.51 0.07
to to to to to to to to to to
1.91 0.87 1.53 2.65 0.97 1.08 1.82 2.12 1.46 1.5
Conventional event OR
P
0.83 0.51 1.15 1.45 0.54 1.01 1.0 1.23 1.65 1.04
.33 .045 .55 .51 .02 .79 .98 .45 .1 .96
95% CI 0.58 0.26 0.73 0.48 0.32 0.95 0.68 0.73 0.91 0.26
to to to to to to to to to to
1.2 0.99 1.8 4.35 0.91 1.1 1.46 2.08 2.99 4.2
SSIDS, subsequent siblings of SIDS.
Moreover, if extreme events were precursors to SIDS, then we would expect to see an excess of such events during the night. Our specific hypotheses were as follows: 1. Extreme events are not precursors to SIDS, as evidenced by the lack of correlation with known SIDS epidemiologic risk factors. 2. If conventional events were normal, then their numbers should increase once a circadian decrease in breathing rate during the night is established. Their numbers should decrease with maternal smoking, secondary to an increased breathing rate to compensate for hypoxia.8
METHODS Participants Of the 1082 participants in the CHIME study, the home monitoring records of 306 healthy term; 170 siblings of SIDS infants (SSIDS), 50 of whom were preterm (ⱕ37 weeks postmenstrual age [PMA]); 152 infants with an idiopathic apparent life-threatening event (ALTE), 45 of whom were preterm; and 454 preterm infants of ⱕ34 weeks gestation were included in the present study. Infants were excluded from the CHIME study if they had any of the following: pneumonia confirmed by chest radiograph; home treatment with continuous oxygen, bronchodilators, diuretics, steroids, or medications for gastroesophageal reflux or seizure; congenital heart disease, except asymptomatic ductus arteriosus, atrial septal defect, or small muscular ventricular septal defect; ventricular peritoneal shunt; congenital brain anomaly that would result in non-SIDS diagnosis in the event of sudden death; chromosomal abnormality; midfacial hypoplasia or cleft palate; inborn error of metabolism; caregiver currently using illicit drugs; or parental inability to communicate (eg, due to language barrier or no telephone). The ALTEs were screened to rule out known etiology and thus belonged to the group with apnea of infancy (AOI). An autopsy was required to confirm the status of SSIDS. If there was any doubt, or if the autopsy cause of
death was not listed as SIDS but seemed like SIDS, then the case was individually examined by CHIME investigators and sometimes by the Chair of the Pathology Subcommittee, Dr Henry Krous. The California autopsy protocol was issued in 1991 but was used in only a few sites. During the 1990s, death scene investigations were not consistently implemented, but all sites asked a standardized set of questions regarding the circumstances of death. Criteria for healthy term infants included 38 to 42 weeks gestation at birth, normal Apgar scores, discharge on or before date of maternal discharge, no medications, no family history of SIDS in siblings, and no other family history of SIDS in the last 10 years.3 The Institutional Review Boards at the 5 clinical sites approved the study. Written informed consent was obtained from all parents or caregivers.
Monitoring Procedure CHIME staff members evaluated maternal and infant characteristics at study entry to establish eligibility, and at various times thereafter.3 Cardiorespiratory waveforms were recorded in the home from each enrolled infant by the CHIME monitor (Non-Invasive Monitoring Systems, Miami, FL). This instrument monitors rib cage and abdominal excursions, using respiratory inductance plethysmography. Using standard disposable infant electrodes, an R-wave detection algorithm determines heart rate. In addition, a pulse oximeter (Ohmeda Corp, Liberty Center, NJ) was used to measure hemoglobin oxygen saturation. Risk factors for SIDS are listed in Table I. Additional variables examined include events during early morning hours, ALTE and SSIDS status; Asian, Caucasian, or Hispanic ethnicity; placenta previa; and abruptio placenta. Data Analysis The memory monitor cartridges with events were downloaded every 2 to 4 weeks. On study completion, waveforms were analyzed by trained technicians unaware of study
Extreme and Conventional Cardiorespiratory Events and Epidemiologic Risk Factors for SIDS
637
group or clinical status, who used a software tool to score the data.9
Identification of Events Conventional events were defined as (1) apnea persisting for at least 20 seconds, (2) heart rate ⬍60 beats per minute (bpm) for at least 5 seconds or ⬍80 bpm for 15 seconds in infants ⬍44 weeks PMA, or (3) heart rate ⬍50 bpm for at least 5 seconds or ⬍60 bpm for at least 15 seconds in infants ⱖ44 weeks PMA. Extreme events were defined as (1) apnea persisting for at least 30 seconds, (2) heart rate ⬍60 bpm for at least 10 seconds in infants ⬍44 weeks PMA, or (3) heart rate ⬍50 bpm for at least 10 seconds in infants ⱖ44 weeks PMA. Either of these events could be accompanied by oxygen desaturation. Statistical Analyses Ramanathan et al3 focused on the dependent variable of having at least 1 extreme event. In this work, we evaluated 3 definitions of outcome events: having (1) at least 1 extreme event, (2) at least 1 conventional event without having an extreme event, or (3) numerous extreme or conventional events, defined as being in the upper 50th percentile of infants with at least 1 event. Thus, the infants belonged to the half with the most events and had a tendency to have multiple extreme or conventional events. In addition, we examined the presence of extra long apneic periods (ⱖ40 seconds). Multivariate logistic regression was the primary analysis strategy. In the original study, prematurity was found to be highly predictive of having at least 1 extreme event; thus, this variable was always used as a covariate.3 The hours of monitoring differed widely among participants.10 The monitoring hours were stratified into 5 categories (⬍100, 100 to 500, 501 to 1000, 1001 to 2000, and ⬎2000), with approximately 200 infants in each category. This variable was treated as continuous and included as a covariate. Multivariate analyses were performed first for all infants and all events during the entire monitoring period and then separately for the events occurring before and at 44 weeks PMA and those occurring after 44 weeks PMA. This 2-age strategy was predicated on the age of highest risk for SIDS in term infants between 50 and 52 weeks PMA. The highest risk for SIDS in preterm infants is earlier and has a much wider dispersion.11 The paucity of extreme events after 44 weeks PMA, however, precluded a multivariate analysis that permitted unequivocal conclusions (Figure 1). In the logistic regression analysis, the odds ratio (OR) was used as a measure of the effect and the area under the receiver operating characteristic curve (ROC) was used as a measure of model strength. Goodness of fit was tested using the Hosmer-Lemeshow test with 10 groupings.12 Randomeffects logistic regression was used to establish the effect of time of night on the incidence of events, using repeated observations for each subject. The relationship between extreme and conventional events (ie, their rate of decline as a function of maturity) was examined as well. 638
Hoppenbrouwers et al
Figure 1. Percentage of infants (ordinate) with conventional events (top line) and extreme events (bottom line) as a function of postmenstrual age (PMA, abscissa). The graphs were adjusted for number of hours monitored.
RESULTS Subject Characteristics Smoking during pregnancy and in the neonatal period was observed in 12% of Asian, 16% of African-American, 32% of Caucasian, 4% of Hispanic, and 24% of other mothers (Table II). Events in ALTE, SSIDS, and Preterm Infants In the cohort ⱕ44 weeks PMA, the ALTE, SSIDS, and preterm infants were at significantly increased risk of at least 1 extreme event compared with healthy term infants. A comparison of SSIDS and ALTE infants with preterm infants revealed that the preterm infants had significantly more events (P ⫽ .003). The ROC result for this model was 80.5%, and the goodness of fit was P ⫽ .42. According to the original study,3 the effect of study group disappeared after 43 weeks PMA, a finding confirmed in the present study. Because approximately 1/3 of the ALTE and SSIDS infants had a PMA of 37 weeks or less, we investigated whether prematurity alone can explain the increased risk for exhibiting at least 1 extreme event. In both cases, there was a significant excess chance of having at least 1 extreme event over and beyond the variables of prematurity and number of hours monitored (ALTE: OR, 5.77; P ⫽ .005; 95% CI ⫽ 1.75 19.39; SSIDS: OR, 5.8; P ⫽ .03; 95% CI ⫽ 1.79 to 18.76). SSIDS and preterm infants were at increased risk of having at least 1 conventional event without having an extreme event (SSIDS: OR ⫽ 2.18; P ⫽ 0; 95% CI ⫽ 1.43 to 3.32; preterm: OR ⫽ 2.71; P ⫽ 0; 95% CI ⫽ 1.92 to 3.83). The ROC was 74.7%, and the goodness of fit was P ⫽ .19. Because of these findings, these study groups were used as covariates in subsequent analyses of extreme and conventional events. Of the infants who had extreme events, 87% experienced these extreme events both before and after 44 weeks PMA. Analysis of the cohort with events before and after 44 weeks PMA combined revealed no significant relationship The Journal of Pediatrics • May 2008
Table II. Maternal and infant characteristics
Maternal Age, years, mean (SD) ⱕ18 years, n (%) Education ⱕ11 years, n (%) Parity ⱖ2, n (%) Race, number (%) Asian African American Caucasian Hispanic Other Married, n (%) Cigarette use, n (%) Infant Male, n (%) Gestational age at birth, weeks, mean (SD) Birth weight, g, mean (SD) Placenta previa, n (%) Median hours monitored
Healthy term (n ⴝ 306)
SSIDS (n ⴝ 170)
ALTE (n ⴝ 152)
Preterm <34 weeks (n ⴝ 454)
30 (6) 11 (4) 25 (8) 201 (66)
29 (6) 7 (4) 39 (23) 170 (100)
26 (6) 11 (7) 38 (25) 95 (63)
28 (6) 25 (6) 97 (21) 293 (65)
33 (11) 47 (15) 172 (56) 18 (6) 35 (11) 227 (74) 40 (13)
12 (7) 25 (15) 102 (60) 22 (13) 9 (5) 112 (66) 53 (31)
18 (12) 18 (12) 79 (52) 16 (11) 21 (14) 78 (51) 53 (35)
44 (10) 113 (25) 164 (36) 107 (24) 24 (5) 264 (74) 93 (20)
162 (53) 39.4 (1) 3312 (306) 4 (1.3) 204
87 (51) 38.8 (2.3) 3330 (556) 1 (0.6) 1438
79 (52) 38.4 (2.3) 3137 (589) 1 (0.7) 780
232 (51) 29.7 (2.5) 1255 (315) 9 (1.2) 800
between extreme events and any of the SIDS risk factors examined.
SIDS Risk Factors/Additional Variables and at Least 1 Extreme Event in Infants <44 Weeks PMA (Table I) None of the SIDS risk factors were significantly related to having at least 1 extreme event. Note that the variable of African-American race was almost significant (P ⫽ .054); however, the OR was ⬍1, suggesting that African-American infants have fewer, not more extreme events, a finding in line with their increased risk for SIDS. Only placenta previa had a significant OR (6.63; P ⫽ .003). Being in the Upper Half of Infants with Extreme Events The results for this third outcome variable differed little from those for the other outcomes. Placenta previa, the only variable that significantly increased the incidence of having at least 1 extreme event, was again significant. The variable African-American race yielded an OR ⬍1, but this was not significant; however, the variable Asian race yielded a significant OR (2.36; P ⬍.02; 95% CI ⫽ 1.18 to 4.72). SIDS Risk Factors/Additional Variables and Conventional Events in Infants <44 Weeks PMA African-American infants had OR ⫽ 0.56, P ⫽ .009, and 95% CI ⫽ 0.37 to 0.87, and those infants whose mother smoked during pregnancy had OR ⫽ 0.66, P ⫽ .037, and 95% CI ⫽ 0.44 to 0.97, indicating fewer conventional events. Males were more likely to exhibit conventional events (OR ⫽ 1.4; P ⫽ .032). Being a SSIDS doubled the likelihood of exhibiting conventional events (OR 2.09; P ⫽ .015), whereas
ALTE status and being a healthy term infant reduced the likelihood of exhibiting these events (OR ⫽ 0.369; P ⫽ .05 and OR ⫽ 0.52; P ⫽ .006, respectively).
Being in the Upper Half of Infants with Conventional Events Infants whose mothers smoked during pregnancy had significantly fewer conventional events. Males were again more likely to have conventional events (OR ⫽ 1.43; P ⫽ .03; 95% CI ⫽ 1.04 to 1.98). With this dependent variable, being African American appeared to be protective, but the value was not significant. Conventional Events in Infants >44 Weeks PMA In older infants, being African-American and maternal smoking during pregnancy were associated with fewer conventional events. None of the additional variables examined exhibited a significant OR. Time of Night and Extreme and Conventional Events Having at least 1 extreme event was less likely in the early morning hours, between 4:00 AM and 7:00 AM (OR ⫽ 0.61; P ⫽ .013; 95% CI ⫽ 1.08 to 1.57) (Figure 2). The incidence of conventional events was higher during 2 nightly intervals: 10:00 PM to 1:00 AM and 1:00 AM to 4:00 AM (OR ⫽ 1.28; P ⫽ .001; 95% CI ⫽ 1.1 to 1.5). Reexamination of infants only after they had reached ⬎44 weeks PMA revealed a similar distribution for conventional events as found in the entire cohort, with more events occurring between 10:00 PMand 4:00 AM than later at night. There were too few extreme events in infants ⬎44 weeks PMA to allow statistical establishment of a nightly distribution of events. The finding
Extreme and Conventional Cardiorespiratory Events and Epidemiologic Risk Factors for SIDS
639
Figure 2. Mean number of events, adjusted for hours monitored, as a function of interval during the night. Stars indicate significant values compared with other intervals.
of fewer extreme events between 4:00 AM and 7:00 AM in the entire cohort was not significant in this older group (P ⫽ .069), although the direction was the same.
Apnea >40 Seconds Nineteen infants exhibited a total of 118 apneic episodes of this duration. Of these infants, 13 were preterm, including 1 SSIDS, 4 term SSIDS, 2 ALTE, and no healthy terms. With 4 exceptions, the extra-long apneic periods occurred in infants ⱕ44 weeks PMA. Rate of Decline of Extreme and Conventional Events as a Function of Maturity The rate of decline of extreme events with increased maturity paralleled that of conventional events (Figure 1).
DISCUSSION This study supports our first hypothesis, that extreme events are not statistically associated with known SIDS risk factors. An excess of extreme events was not found at the time of night when infants are most likely to die. In contrast, extreme events occurred less often during the early morning, when infants tend to be at higher risk for SIDS. Moreover, Asians, the ethnic group at lowest epidemiologic risk for SIDS, exhibited a tendency toward a higher rate of extreme events.13 Healthy terms had significantly fewer of these events compared with the other study groups, but these differences disappeared after 43 weeks PMA, well before the peak incidence of SIDS.3 The results also support our second hypothesis, becauses conventional events increased when breathing rates begin to decline during the night, a manifestation of a circadian modulation.14 Maternal smoking during pregnancy, which would be expected to increase breathing rates in the infant secondary to mild hypoxia, was associated with a decrease in conventional events in both young and older infants. Based on the similar rate of decline in these events with increasing age, 640
Hoppenbrouwers et al
the conceptual separation of conventional and extreme events in terms of mechanism or etiology seems unwarranted. Epidemiologic risk factors for SIDS have been changing over the past few decades, partly as a result of the Back to Sleep campaign and partly because the incidences of other modifiable risk factors have changed. This includes, with 1 exception,15 a decline in smoking,16 a decrease in teenage pregnancies, and increased breast-feeding. Hoffman et al17 reported a 70% rate of smoking in SIDS mothers during the 1980s, compared with 42% in the 1990s.15 The revisions that have occurred historically do not invalidate the risk factors examined here, however.18 Events are strongly associated with immaturity, the conventional events less so than the extreme events. Because of the large number of extreme events at ⱕ44 weeks PMA and the even larger number of conventional events throughout the study period, it is very likely that the absence of significant associations between events and risk factors is reliable. A highly significant relationship of extreme events and placenta previa was based on only 15 cases. These infants, who perhaps suffered from fairly severe hypoxia during pregnancy,19 had a 6-fold greater risk of exhibiting extreme events ⱕ44 weeks PMA. With 1 exception,20 epidemiologic studies have not implicated placenta previa or placenta abruptio as SIDS risk factors.17,21 Conventional events occur more often in subsequent siblings and during the time of night at which infants tend to die. Could this indicate that these events make infants vulnerable? We believe that these conventional events are normal, even a sign of health, not just because they are ubiquitous (an obvious argument against pathology), but also because they occur more often in healthy term infants. There are several other reasons to assume that conventional events do not confer risk: 1. We would expect extreme events to be more risky, but they turned out to be not associated with the time of highest risk in the morning. 2. A lack of some degree of variability in autonomic nervous system function has been associated with danger.22 3. Respiratory rates that correlate with short apneic periods, similar to conventional events, typically decline during that same time interval.23 4. Like extreme events, conventional events decline dramatically by the age of highest SIDS risk. 5. These conventional events occur less often in infants stressed by maternal smoking. Maternal smoking during pregnancy increases the breathing rate in the neonatal period.8 The data favor the alternative conclusion that these events are normal and perhaps even a sign of health. The finding of significantly fewer conventional events in African-American infants cannot be explained by an increased incidence of maternal smoking during pregnancy. Half as many of the African-American mothers as Caucasian mothers smoked in this cohort (16.4% vs 31.5%). The Journal of Pediatrics • May 2008
The absence of any correlation between extreme events and SIDS risk factors strongly suggests that extreme events are not causally related to SIDS. The so-called “apnea hypothesis,” which has been questioned,1,2 apparently still has some adherents.24 A few infants died while being monitored, and although gasping was often observed, establishing the primary mechanism of failure has been difficult.25 In the final analysis, the mechanism of death inevitably will be either respiratory or cardiac arrest. The findings from the previous CHIME study3 demonstrated that the extreme events were most prominent before the peak SIDS incidence in both preterm and term infants. This result is particularly relevant in light of the recent report of abnormalities in the serotonin transporter system in the brainstem of SIDS infants.26 Serotonin has been implicated in the regulation of breathing and arousal. Because the present results show that extreme events are not associated with SIDS risk factors and time of death in SIDS, the onus of these serotonin abnormalities should probably be placed not so much on their effect on the respiratory system as on their effect on a putative failure to arouse in SIDS. In conclusion, neither extreme nor conventional events were associated with primary epidemiologic risk factors for SIDS, supporting our hypothesis that they are not immediate precursors of or causally related to SIDS. Conventional events seem to be a hallmark of health; they occur less frequently in infants of mothers who smoked during pregnancy. Both types of events decline with maturation in a remarkably parallel fashion. Therefore, conventional and extreme events are probably normal, part of a continuum and not fundamentally different in mechanism. The duration of any event that marks entry into a danger zone that will inevitably be followed by an infant’s death remains to be determined. We thank Xiaowei Liu for his contributions to data management and organization and Denis Rybin for his assistance with the CHIME data base. Dr. Hoppenbrouwers was a member of the CHIME steering committee. We thank the steering committee members and staff of CHIME who designed and executed the original study.
REFERENCES 1. Hodgman JE, Hoppenbrouwers T. Home monitoring for SIDS: the case against. Ann NY Acad Sci 1988;533:164-75. 2. Harper RM, Bandler R. Finding the failure mechanism in sudden infant death syndrome. Nat Med 1998;4:157-8. 3. Ramanathan R, Corwin MJ, Hunt CE, Lister G, Tinsley L, Baird L, et al, and the Collaborative Home Infant Monitoring Evaluation (CHIME) Study Group. Cardiorespiratory events recorded in the home: comparison of healthy infant with those at increased risk for SIDS. JAMA 2001;285:2199-207.
4. Hodgman JE, Gonzalez F, Hoppenbrouwers T, Cabal L. Apnea, transient episodes of bradycardia and periodic breathing in preterm infants. Am J Dis Child 1990;144:54-7. 5. Neuman MR, Watson HL, Mendenhall RS, Zoldak JT, DiFiore JM, Peucker M, et al, and the CHIME Study Group. Cardiopulmonary monitoring at home: the CHIME monitor. Physiol Meas 2001;22:267-86. 6. Hoppenbrouwers T, Hodgman JE, Sterman MB, Harper RM. Temporal distribution of sleep states, somatic and autonomic activity during the first half year of life. Sleep 1982;5:131-44. 7. Hoppenbrouwers T, Hodgman JE, Harper RM, Sterman MB. Respiration during the first six months of life in normal infants. IV. Gender differences. Early Human Dev 1980;4:167-77. 8. Stick SM, Burton PR, Gurrin L, Sly PD, Le Souëf PN. Effects of smoking during pregnancy and a family history of asthma on respiratory function in newborn babies. Lancet 1996;348:1060-4. 9. Corwin MJ, Lister G, Silvestri JM, Peucker M, Brooks LJ, Davidson Ward SL, et al, and the CHIME Study Group. Agreement among raters in assessment of physiologic waveforms recorded by a cardiorespiratory monitor for home use. Pediatr Res 1998;44:682-90. 10. Silvestri JM, Lister G, Corwin MJ, Smok-Pearsall SM, Hunt CE, Baird TM, et al, and the CHIME Study Group. Factors that influence use of a home cardiorespirator for infants: the Collaborative Home Infant Monitoring Evaluation. Arch Pediatr Adolesc Med 2005;159:18-24. 11. Malloy MH, Freeman DH Jr. Birth weight and gestational age–specific sudden infant death syndrome mortality: United States, 1991 versus 1995. Pediatrics 2000;105: 1227-31. 12. Metz CE. Basic principles of ROC analysis. Semin Nuclear Med 1978;VIII; 283-98. 13. Nelson T, To K-F, Wong Y-Y, Dickinson J, Choi K-C, Yu, L-M, et al. Hong Kong case-control study of sudden unexpected infant death. NZ Med J 2005;118: U1788. 14. Hoppenbrouwers T, Jensen D, Hodgman JE, Harper RM, Sterman MB. The emergence of a circadian pattern in respiratory rates: comparison between infants at low and increased risk for SIDS. Pediatr Res 1980;14:345-51. 15. Blair PS, Sidebotham P, Berry PJ, Evans M, Fleming PJ. Major epidemiological changes in sudden infant death syndrome: a 20-year population-based study in the UK. Lancet 2006;367:314-9. 16. Li DK, Petitti DB, Willinger M, McMahon R, Odouli R, Vu H, et al. Infant sleeping position and the risk of sudden infant death syndrome in California, 19972000. Am J Epidemiol 2003;157:446-55. 17. Hoffman HJ, Damus K, Hillman L, Krongrad E. Risk factors for SIDS: results of the National Institutes of Child Health and Human Development SIDS Cooperative Epidemiological Study. Ann NY Acad Sci 1988;533:1. 18. American Academy of Pediatrics, Task Force on Sudden Infant Death Syndrome. Diagnostic coding shifts, controversies regarding the sleeping environment, and new variables to consider in reducing risk. Pediatrics 2005;116:1245-55. 19. Crane JM, van den Hof MC, Dodds L, Armson BA, Liston R. Neonatal outcomes with placenta previa. Obstet Gynecol 1999;93:541-4. 20. Li DK, Wi S. Maternal placental abnormality and the risk of sudden infant death syndrome. Am J Epidemiol 1999;149:608-11. 21. Leach CE, Blair PS, Fleming PJ, Smith IJ, Platt MW, Berry PJ, et al. Epidemiology of SIDS and explained sudden infant deaths. CESDI SUDI Research Group. Pediatrics 1999;104:43. 22. Hodgman J, Pavlova Z, Hoppenbrouwers T, Moore K, Peckham J, Sleep S, et al. Clinical predictors of imminent brain death. Clin Res 1987;35:234A. 23. Hoppenbrouwers T. Ontogenesis of circadian respiratory rhythms. J Interdisc Cycle Res 1986;17:140-1. 24. Thach B. The role of respiratory control disorders in SIDS. Res Physiol Neurobiol 2005;149:343-53. 25. Poets CE, Meny RG, Chobanian MR, Bonofiglo RE. Gasping and other cardiorespiratory patterns during sudden infant deaths. Pediatr Res 1999;45:350-4. 26. Patterson DS, Trachtenberg FL, Thompson EG, Belliveau RA, Beggs AH, Darnall R, et al. Multiple serotonergic brainstem abnormalities in sudden infant death syndrome. JAMA 2006;296:2124-32.
Extreme and Conventional Cardiorespiratory Events and Epidemiologic Risk Factors for SIDS
641
AND
FRED DOREY, PHD
Objective To test the hypotheses that there is a lack of correlation between extreme events and epidemiologic risk factors for sudden infant death syndrome (SIDS), and if conventional events are normal, their numbers should increase once a circadian decrease in breathing rate is established. In addition, the number of events should decrease with maternal smoking. Study design Three outcome variables were derived from the Collaborative Home Infant Monitoring Evaluation (CHIME) of 1082 infants: (1) at least 1 extreme event lasting >30 seconds, (2) at least 1 conventional event lasting >20 seconds, and (3) being part of the 50% of infants with the most events. Results Multivariate logistic regression analyses found that extreme events were not statistically associated with any known SIDS risk factors and occurred less often during the early morning. Healthy term infants had significantly fewer of these events compared with preterm infants, subsequent siblings of infants with SIDS, and infants with an apparent life-threatening event, a finding that was not evident after 43 weeks (3 weeks postterm). Conventional events increased during the night, whereas maternal smoking was associated with a decrease in conventional events. Apneic episodes persisting for >40 seconds occurred in 1.8% of the infants. Conclusions Extreme events are associated with immaturity and do not seem to be immediate precusors of or causally related to SIDS. (J Pediatr 2008;152:636-41) he historical focus on apnea as either a cause of sudden infant death syndrome (SIDS) or a mechanism of death has decreased because studies have failed to implicate these events directly.1,2 Before the 1990s, home monitoring equipment typically targeted apneic events as short as 20 seconds, which have been shown to be ubiquitous.3 These short apneic events have been designated conventional events based on traditional clinical practice. A second category of events, identified from documented monitoring, has been arbitrarily designated extreme events. These comprise episodes of bradycardia alone or apnea persisting for more than 30 seconds with or without bradycardia and oxygen desaturation.4 Their role in the etiology of SIDS also is questionable. From the Division of Neonatal Medicine and Biostatistics (F.D.), LAC ⫹ USC Medical The Collaborative Home Infant Monitoring Evaluation (CHIME) was carried out Center, Women’s and Children’s Hospital during the 1990s, when having infants sleep on their back was recommended to all and Children’s Hospital Los Angeles (T.H.), Keck School of Medicine, University of Southparticipants.5 More than 1000 infants participated in this study in 5 US clinical centers, ern California, Los Angeles, CA (T.H., J.H.) representing the largest cohort for which systematic home monitoring has been performed and New York Presbyterian Hospital-Weill Cornell Medical Center, New York, NY to date. CHIME found a high incidence of extreme cardiorespiratory events at early ages (A.R.). in preterm infants. However, at the age of highest risk for SIDS, the incidence of these Supported by the National Institute of 3 events was similar in preterm and term infants. This finding has raised further questions Child Health and Human Development (grants HD 29067, HD 29071, HD 28971, about the role of apnea in the etiology of SIDS and the possibility that either extreme or HD 29073, HD 29056, and HD 34625) conventional events can predict imminent SIDS deaths. and a generous donation from the Orange Circadian modulation of cardiac and respiratory rates, with decreases during the County Guild for Infant Survival. Submitted for publication Feb 20, 2007; last nighttime hours, is first observed after the first month of life in term infants.6 Because revision received Sep 18, 2007; accepted lower respiratory rates correlate with a higher incidence of breathing pauses, it is not Oct 1, 2007. 7 unreasonable to expect a higher incidence of conventional events during that period. Reprint requests: Toke Hoppenbrouwers,
T
ALTE AOI CHIME CI OR
636
Apparent life-threatening event Apnea of infancy Collaborative Home Infant Monitoring Evaluation Confidence interval Odds ratio
PMA ROC SIDS SSIDS
Postmenstrual age Receiver operating characteristic curve Sudden infant death syndrome Subsequent sibling of sudden infant death syndrome
PhD, Room 9L19, Women and Children’s Hospital, LAC ⫹ USC Medical Center, 1240 Mission Road, Los Angeles, CA 90033. E-mail: [email protected]. 0022-3476/$ - see front matter Copyright © 2008 Mosby Inc. All rights reserved. 10.1016/j.jpeds.2007.10.003
Table I. ORs for having at least 1 event based on SIDS and other risk factors after adjusting for preterm birth, ALTE, SSIDS1 and number of hours monitored PMA <44 weeks Extreme event SIDS risk factor
OR
P
Male predominance African-American infant Parity ⱖ2 Maternal age ⱕ18 years Maternal smoking Maternal education ⬍12 years Caucasian Hispanic Asian Placenta previa
0.77 0.54 0.79 0.62 1.0 1.5 1.01 1.0 1.5 6.63
.24 .054 .34 .28 .93 .14 .97 .99 .27 .003
95% CI 0.49 0.29 0.49 0.26 0.6 0.89 0.64 0.58 0.74 1.86
to to to to to to to to to to
PMA >44 weeks,
Conventional event
1.2 1 1.3 1.5 1.8 2.4 1.6 1.7 2.8 23.6
OR
P
1.4 0.56 1.07 1.29 0.66 0.72 1.33 1.43 0.87 0.33
.032 .009 .72 .50 .037 .11 .08 .08 .59 .17
95% CI 1.03 0.37 0.75 0.62 0.44 0.48 0.97 0.96 0.51 0.07
to to to to to to to to to to
1.91 0.87 1.53 2.65 0.97 1.08 1.82 2.12 1.46 1.5
Conventional event OR
P
0.83 0.51 1.15 1.45 0.54 1.01 1.0 1.23 1.65 1.04
.33 .045 .55 .51 .02 .79 .98 .45 .1 .96
95% CI 0.58 0.26 0.73 0.48 0.32 0.95 0.68 0.73 0.91 0.26
to to to to to to to to to to
1.2 0.99 1.8 4.35 0.91 1.1 1.46 2.08 2.99 4.2
SSIDS, subsequent siblings of SIDS.
Moreover, if extreme events were precursors to SIDS, then we would expect to see an excess of such events during the night. Our specific hypotheses were as follows: 1. Extreme events are not precursors to SIDS, as evidenced by the lack of correlation with known SIDS epidemiologic risk factors. 2. If conventional events were normal, then their numbers should increase once a circadian decrease in breathing rate during the night is established. Their numbers should decrease with maternal smoking, secondary to an increased breathing rate to compensate for hypoxia.8
METHODS Participants Of the 1082 participants in the CHIME study, the home monitoring records of 306 healthy term; 170 siblings of SIDS infants (SSIDS), 50 of whom were preterm (ⱕ37 weeks postmenstrual age [PMA]); 152 infants with an idiopathic apparent life-threatening event (ALTE), 45 of whom were preterm; and 454 preterm infants of ⱕ34 weeks gestation were included in the present study. Infants were excluded from the CHIME study if they had any of the following: pneumonia confirmed by chest radiograph; home treatment with continuous oxygen, bronchodilators, diuretics, steroids, or medications for gastroesophageal reflux or seizure; congenital heart disease, except asymptomatic ductus arteriosus, atrial septal defect, or small muscular ventricular septal defect; ventricular peritoneal shunt; congenital brain anomaly that would result in non-SIDS diagnosis in the event of sudden death; chromosomal abnormality; midfacial hypoplasia or cleft palate; inborn error of metabolism; caregiver currently using illicit drugs; or parental inability to communicate (eg, due to language barrier or no telephone). The ALTEs were screened to rule out known etiology and thus belonged to the group with apnea of infancy (AOI). An autopsy was required to confirm the status of SSIDS. If there was any doubt, or if the autopsy cause of
death was not listed as SIDS but seemed like SIDS, then the case was individually examined by CHIME investigators and sometimes by the Chair of the Pathology Subcommittee, Dr Henry Krous. The California autopsy protocol was issued in 1991 but was used in only a few sites. During the 1990s, death scene investigations were not consistently implemented, but all sites asked a standardized set of questions regarding the circumstances of death. Criteria for healthy term infants included 38 to 42 weeks gestation at birth, normal Apgar scores, discharge on or before date of maternal discharge, no medications, no family history of SIDS in siblings, and no other family history of SIDS in the last 10 years.3 The Institutional Review Boards at the 5 clinical sites approved the study. Written informed consent was obtained from all parents or caregivers.
Monitoring Procedure CHIME staff members evaluated maternal and infant characteristics at study entry to establish eligibility, and at various times thereafter.3 Cardiorespiratory waveforms were recorded in the home from each enrolled infant by the CHIME monitor (Non-Invasive Monitoring Systems, Miami, FL). This instrument monitors rib cage and abdominal excursions, using respiratory inductance plethysmography. Using standard disposable infant electrodes, an R-wave detection algorithm determines heart rate. In addition, a pulse oximeter (Ohmeda Corp, Liberty Center, NJ) was used to measure hemoglobin oxygen saturation. Risk factors for SIDS are listed in Table I. Additional variables examined include events during early morning hours, ALTE and SSIDS status; Asian, Caucasian, or Hispanic ethnicity; placenta previa; and abruptio placenta. Data Analysis The memory monitor cartridges with events were downloaded every 2 to 4 weeks. On study completion, waveforms were analyzed by trained technicians unaware of study
Extreme and Conventional Cardiorespiratory Events and Epidemiologic Risk Factors for SIDS
637
group or clinical status, who used a software tool to score the data.9
Identification of Events Conventional events were defined as (1) apnea persisting for at least 20 seconds, (2) heart rate ⬍60 beats per minute (bpm) for at least 5 seconds or ⬍80 bpm for 15 seconds in infants ⬍44 weeks PMA, or (3) heart rate ⬍50 bpm for at least 5 seconds or ⬍60 bpm for at least 15 seconds in infants ⱖ44 weeks PMA. Extreme events were defined as (1) apnea persisting for at least 30 seconds, (2) heart rate ⬍60 bpm for at least 10 seconds in infants ⬍44 weeks PMA, or (3) heart rate ⬍50 bpm for at least 10 seconds in infants ⱖ44 weeks PMA. Either of these events could be accompanied by oxygen desaturation. Statistical Analyses Ramanathan et al3 focused on the dependent variable of having at least 1 extreme event. In this work, we evaluated 3 definitions of outcome events: having (1) at least 1 extreme event, (2) at least 1 conventional event without having an extreme event, or (3) numerous extreme or conventional events, defined as being in the upper 50th percentile of infants with at least 1 event. Thus, the infants belonged to the half with the most events and had a tendency to have multiple extreme or conventional events. In addition, we examined the presence of extra long apneic periods (ⱖ40 seconds). Multivariate logistic regression was the primary analysis strategy. In the original study, prematurity was found to be highly predictive of having at least 1 extreme event; thus, this variable was always used as a covariate.3 The hours of monitoring differed widely among participants.10 The monitoring hours were stratified into 5 categories (⬍100, 100 to 500, 501 to 1000, 1001 to 2000, and ⬎2000), with approximately 200 infants in each category. This variable was treated as continuous and included as a covariate. Multivariate analyses were performed first for all infants and all events during the entire monitoring period and then separately for the events occurring before and at 44 weeks PMA and those occurring after 44 weeks PMA. This 2-age strategy was predicated on the age of highest risk for SIDS in term infants between 50 and 52 weeks PMA. The highest risk for SIDS in preterm infants is earlier and has a much wider dispersion.11 The paucity of extreme events after 44 weeks PMA, however, precluded a multivariate analysis that permitted unequivocal conclusions (Figure 1). In the logistic regression analysis, the odds ratio (OR) was used as a measure of the effect and the area under the receiver operating characteristic curve (ROC) was used as a measure of model strength. Goodness of fit was tested using the Hosmer-Lemeshow test with 10 groupings.12 Randomeffects logistic regression was used to establish the effect of time of night on the incidence of events, using repeated observations for each subject. The relationship between extreme and conventional events (ie, their rate of decline as a function of maturity) was examined as well. 638
Hoppenbrouwers et al
Figure 1. Percentage of infants (ordinate) with conventional events (top line) and extreme events (bottom line) as a function of postmenstrual age (PMA, abscissa). The graphs were adjusted for number of hours monitored.
RESULTS Subject Characteristics Smoking during pregnancy and in the neonatal period was observed in 12% of Asian, 16% of African-American, 32% of Caucasian, 4% of Hispanic, and 24% of other mothers (Table II). Events in ALTE, SSIDS, and Preterm Infants In the cohort ⱕ44 weeks PMA, the ALTE, SSIDS, and preterm infants were at significantly increased risk of at least 1 extreme event compared with healthy term infants. A comparison of SSIDS and ALTE infants with preterm infants revealed that the preterm infants had significantly more events (P ⫽ .003). The ROC result for this model was 80.5%, and the goodness of fit was P ⫽ .42. According to the original study,3 the effect of study group disappeared after 43 weeks PMA, a finding confirmed in the present study. Because approximately 1/3 of the ALTE and SSIDS infants had a PMA of 37 weeks or less, we investigated whether prematurity alone can explain the increased risk for exhibiting at least 1 extreme event. In both cases, there was a significant excess chance of having at least 1 extreme event over and beyond the variables of prematurity and number of hours monitored (ALTE: OR, 5.77; P ⫽ .005; 95% CI ⫽ 1.75 19.39; SSIDS: OR, 5.8; P ⫽ .03; 95% CI ⫽ 1.79 to 18.76). SSIDS and preterm infants were at increased risk of having at least 1 conventional event without having an extreme event (SSIDS: OR ⫽ 2.18; P ⫽ 0; 95% CI ⫽ 1.43 to 3.32; preterm: OR ⫽ 2.71; P ⫽ 0; 95% CI ⫽ 1.92 to 3.83). The ROC was 74.7%, and the goodness of fit was P ⫽ .19. Because of these findings, these study groups were used as covariates in subsequent analyses of extreme and conventional events. Of the infants who had extreme events, 87% experienced these extreme events both before and after 44 weeks PMA. Analysis of the cohort with events before and after 44 weeks PMA combined revealed no significant relationship The Journal of Pediatrics • May 2008
Table II. Maternal and infant characteristics
Maternal Age, years, mean (SD) ⱕ18 years, n (%) Education ⱕ11 years, n (%) Parity ⱖ2, n (%) Race, number (%) Asian African American Caucasian Hispanic Other Married, n (%) Cigarette use, n (%) Infant Male, n (%) Gestational age at birth, weeks, mean (SD) Birth weight, g, mean (SD) Placenta previa, n (%) Median hours monitored
Healthy term (n ⴝ 306)
SSIDS (n ⴝ 170)
ALTE (n ⴝ 152)
Preterm <34 weeks (n ⴝ 454)
30 (6) 11 (4) 25 (8) 201 (66)
29 (6) 7 (4) 39 (23) 170 (100)
26 (6) 11 (7) 38 (25) 95 (63)
28 (6) 25 (6) 97 (21) 293 (65)
33 (11) 47 (15) 172 (56) 18 (6) 35 (11) 227 (74) 40 (13)
12 (7) 25 (15) 102 (60) 22 (13) 9 (5) 112 (66) 53 (31)
18 (12) 18 (12) 79 (52) 16 (11) 21 (14) 78 (51) 53 (35)
44 (10) 113 (25) 164 (36) 107 (24) 24 (5) 264 (74) 93 (20)
162 (53) 39.4 (1) 3312 (306) 4 (1.3) 204
87 (51) 38.8 (2.3) 3330 (556) 1 (0.6) 1438
79 (52) 38.4 (2.3) 3137 (589) 1 (0.7) 780
232 (51) 29.7 (2.5) 1255 (315) 9 (1.2) 800
between extreme events and any of the SIDS risk factors examined.
SIDS Risk Factors/Additional Variables and at Least 1 Extreme Event in Infants <44 Weeks PMA (Table I) None of the SIDS risk factors were significantly related to having at least 1 extreme event. Note that the variable of African-American race was almost significant (P ⫽ .054); however, the OR was ⬍1, suggesting that African-American infants have fewer, not more extreme events, a finding in line with their increased risk for SIDS. Only placenta previa had a significant OR (6.63; P ⫽ .003). Being in the Upper Half of Infants with Extreme Events The results for this third outcome variable differed little from those for the other outcomes. Placenta previa, the only variable that significantly increased the incidence of having at least 1 extreme event, was again significant. The variable African-American race yielded an OR ⬍1, but this was not significant; however, the variable Asian race yielded a significant OR (2.36; P ⬍.02; 95% CI ⫽ 1.18 to 4.72). SIDS Risk Factors/Additional Variables and Conventional Events in Infants <44 Weeks PMA African-American infants had OR ⫽ 0.56, P ⫽ .009, and 95% CI ⫽ 0.37 to 0.87, and those infants whose mother smoked during pregnancy had OR ⫽ 0.66, P ⫽ .037, and 95% CI ⫽ 0.44 to 0.97, indicating fewer conventional events. Males were more likely to exhibit conventional events (OR ⫽ 1.4; P ⫽ .032). Being a SSIDS doubled the likelihood of exhibiting conventional events (OR 2.09; P ⫽ .015), whereas
ALTE status and being a healthy term infant reduced the likelihood of exhibiting these events (OR ⫽ 0.369; P ⫽ .05 and OR ⫽ 0.52; P ⫽ .006, respectively).
Being in the Upper Half of Infants with Conventional Events Infants whose mothers smoked during pregnancy had significantly fewer conventional events. Males were again more likely to have conventional events (OR ⫽ 1.43; P ⫽ .03; 95% CI ⫽ 1.04 to 1.98). With this dependent variable, being African American appeared to be protective, but the value was not significant. Conventional Events in Infants >44 Weeks PMA In older infants, being African-American and maternal smoking during pregnancy were associated with fewer conventional events. None of the additional variables examined exhibited a significant OR. Time of Night and Extreme and Conventional Events Having at least 1 extreme event was less likely in the early morning hours, between 4:00 AM and 7:00 AM (OR ⫽ 0.61; P ⫽ .013; 95% CI ⫽ 1.08 to 1.57) (Figure 2). The incidence of conventional events was higher during 2 nightly intervals: 10:00 PM to 1:00 AM and 1:00 AM to 4:00 AM (OR ⫽ 1.28; P ⫽ .001; 95% CI ⫽ 1.1 to 1.5). Reexamination of infants only after they had reached ⬎44 weeks PMA revealed a similar distribution for conventional events as found in the entire cohort, with more events occurring between 10:00 PMand 4:00 AM than later at night. There were too few extreme events in infants ⬎44 weeks PMA to allow statistical establishment of a nightly distribution of events. The finding
Extreme and Conventional Cardiorespiratory Events and Epidemiologic Risk Factors for SIDS
639
Figure 2. Mean number of events, adjusted for hours monitored, as a function of interval during the night. Stars indicate significant values compared with other intervals.
of fewer extreme events between 4:00 AM and 7:00 AM in the entire cohort was not significant in this older group (P ⫽ .069), although the direction was the same.
Apnea >40 Seconds Nineteen infants exhibited a total of 118 apneic episodes of this duration. Of these infants, 13 were preterm, including 1 SSIDS, 4 term SSIDS, 2 ALTE, and no healthy terms. With 4 exceptions, the extra-long apneic periods occurred in infants ⱕ44 weeks PMA. Rate of Decline of Extreme and Conventional Events as a Function of Maturity The rate of decline of extreme events with increased maturity paralleled that of conventional events (Figure 1).
DISCUSSION This study supports our first hypothesis, that extreme events are not statistically associated with known SIDS risk factors. An excess of extreme events was not found at the time of night when infants are most likely to die. In contrast, extreme events occurred less often during the early morning, when infants tend to be at higher risk for SIDS. Moreover, Asians, the ethnic group at lowest epidemiologic risk for SIDS, exhibited a tendency toward a higher rate of extreme events.13 Healthy terms had significantly fewer of these events compared with the other study groups, but these differences disappeared after 43 weeks PMA, well before the peak incidence of SIDS.3 The results also support our second hypothesis, becauses conventional events increased when breathing rates begin to decline during the night, a manifestation of a circadian modulation.14 Maternal smoking during pregnancy, which would be expected to increase breathing rates in the infant secondary to mild hypoxia, was associated with a decrease in conventional events in both young and older infants. Based on the similar rate of decline in these events with increasing age, 640
Hoppenbrouwers et al
the conceptual separation of conventional and extreme events in terms of mechanism or etiology seems unwarranted. Epidemiologic risk factors for SIDS have been changing over the past few decades, partly as a result of the Back to Sleep campaign and partly because the incidences of other modifiable risk factors have changed. This includes, with 1 exception,15 a decline in smoking,16 a decrease in teenage pregnancies, and increased breast-feeding. Hoffman et al17 reported a 70% rate of smoking in SIDS mothers during the 1980s, compared with 42% in the 1990s.15 The revisions that have occurred historically do not invalidate the risk factors examined here, however.18 Events are strongly associated with immaturity, the conventional events less so than the extreme events. Because of the large number of extreme events at ⱕ44 weeks PMA and the even larger number of conventional events throughout the study period, it is very likely that the absence of significant associations between events and risk factors is reliable. A highly significant relationship of extreme events and placenta previa was based on only 15 cases. These infants, who perhaps suffered from fairly severe hypoxia during pregnancy,19 had a 6-fold greater risk of exhibiting extreme events ⱕ44 weeks PMA. With 1 exception,20 epidemiologic studies have not implicated placenta previa or placenta abruptio as SIDS risk factors.17,21 Conventional events occur more often in subsequent siblings and during the time of night at which infants tend to die. Could this indicate that these events make infants vulnerable? We believe that these conventional events are normal, even a sign of health, not just because they are ubiquitous (an obvious argument against pathology), but also because they occur more often in healthy term infants. There are several other reasons to assume that conventional events do not confer risk: 1. We would expect extreme events to be more risky, but they turned out to be not associated with the time of highest risk in the morning. 2. A lack of some degree of variability in autonomic nervous system function has been associated with danger.22 3. Respiratory rates that correlate with short apneic periods, similar to conventional events, typically decline during that same time interval.23 4. Like extreme events, conventional events decline dramatically by the age of highest SIDS risk. 5. These conventional events occur less often in infants stressed by maternal smoking. Maternal smoking during pregnancy increases the breathing rate in the neonatal period.8 The data favor the alternative conclusion that these events are normal and perhaps even a sign of health. The finding of significantly fewer conventional events in African-American infants cannot be explained by an increased incidence of maternal smoking during pregnancy. Half as many of the African-American mothers as Caucasian mothers smoked in this cohort (16.4% vs 31.5%). The Journal of Pediatrics • May 2008
The absence of any correlation between extreme events and SIDS risk factors strongly suggests that extreme events are not causally related to SIDS. The so-called “apnea hypothesis,” which has been questioned,1,2 apparently still has some adherents.24 A few infants died while being monitored, and although gasping was often observed, establishing the primary mechanism of failure has been difficult.25 In the final analysis, the mechanism of death inevitably will be either respiratory or cardiac arrest. The findings from the previous CHIME study3 demonstrated that the extreme events were most prominent before the peak SIDS incidence in both preterm and term infants. This result is particularly relevant in light of the recent report of abnormalities in the serotonin transporter system in the brainstem of SIDS infants.26 Serotonin has been implicated in the regulation of breathing and arousal. Because the present results show that extreme events are not associated with SIDS risk factors and time of death in SIDS, the onus of these serotonin abnormalities should probably be placed not so much on their effect on the respiratory system as on their effect on a putative failure to arouse in SIDS. In conclusion, neither extreme nor conventional events were associated with primary epidemiologic risk factors for SIDS, supporting our hypothesis that they are not immediate precursors of or causally related to SIDS. Conventional events seem to be a hallmark of health; they occur less frequently in infants of mothers who smoked during pregnancy. Both types of events decline with maturation in a remarkably parallel fashion. Therefore, conventional and extreme events are probably normal, part of a continuum and not fundamentally different in mechanism. The duration of any event that marks entry into a danger zone that will inevitably be followed by an infant’s death remains to be determined. We thank Xiaowei Liu for his contributions to data management and organization and Denis Rybin for his assistance with the CHIME data base. Dr. Hoppenbrouwers was a member of the CHIME steering committee. We thank the steering committee members and staff of CHIME who designed and executed the original study.
REFERENCES 1. Hodgman JE, Hoppenbrouwers T. Home monitoring for SIDS: the case against. Ann NY Acad Sci 1988;533:164-75. 2. Harper RM, Bandler R. Finding the failure mechanism in sudden infant death syndrome. Nat Med 1998;4:157-8. 3. Ramanathan R, Corwin MJ, Hunt CE, Lister G, Tinsley L, Baird L, et al, and the Collaborative Home Infant Monitoring Evaluation (CHIME) Study Group. Cardiorespiratory events recorded in the home: comparison of healthy infant with those at increased risk for SIDS. JAMA 2001;285:2199-207.
4. Hodgman JE, Gonzalez F, Hoppenbrouwers T, Cabal L. Apnea, transient episodes of bradycardia and periodic breathing in preterm infants. Am J Dis Child 1990;144:54-7. 5. Neuman MR, Watson HL, Mendenhall RS, Zoldak JT, DiFiore JM, Peucker M, et al, and the CHIME Study Group. Cardiopulmonary monitoring at home: the CHIME monitor. Physiol Meas 2001;22:267-86. 6. Hoppenbrouwers T, Hodgman JE, Sterman MB, Harper RM. Temporal distribution of sleep states, somatic and autonomic activity during the first half year of life. Sleep 1982;5:131-44. 7. Hoppenbrouwers T, Hodgman JE, Harper RM, Sterman MB. Respiration during the first six months of life in normal infants. IV. Gender differences. Early Human Dev 1980;4:167-77. 8. Stick SM, Burton PR, Gurrin L, Sly PD, Le Souëf PN. Effects of smoking during pregnancy and a family history of asthma on respiratory function in newborn babies. Lancet 1996;348:1060-4. 9. Corwin MJ, Lister G, Silvestri JM, Peucker M, Brooks LJ, Davidson Ward SL, et al, and the CHIME Study Group. Agreement among raters in assessment of physiologic waveforms recorded by a cardiorespiratory monitor for home use. Pediatr Res 1998;44:682-90. 10. Silvestri JM, Lister G, Corwin MJ, Smok-Pearsall SM, Hunt CE, Baird TM, et al, and the CHIME Study Group. Factors that influence use of a home cardiorespirator for infants: the Collaborative Home Infant Monitoring Evaluation. Arch Pediatr Adolesc Med 2005;159:18-24. 11. Malloy MH, Freeman DH Jr. Birth weight and gestational age–specific sudden infant death syndrome mortality: United States, 1991 versus 1995. Pediatrics 2000;105: 1227-31. 12. Metz CE. Basic principles of ROC analysis. Semin Nuclear Med 1978;VIII; 283-98. 13. Nelson T, To K-F, Wong Y-Y, Dickinson J, Choi K-C, Yu, L-M, et al. Hong Kong case-control study of sudden unexpected infant death. NZ Med J 2005;118: U1788. 14. Hoppenbrouwers T, Jensen D, Hodgman JE, Harper RM, Sterman MB. The emergence of a circadian pattern in respiratory rates: comparison between infants at low and increased risk for SIDS. Pediatr Res 1980;14:345-51. 15. Blair PS, Sidebotham P, Berry PJ, Evans M, Fleming PJ. Major epidemiological changes in sudden infant death syndrome: a 20-year population-based study in the UK. Lancet 2006;367:314-9. 16. Li DK, Petitti DB, Willinger M, McMahon R, Odouli R, Vu H, et al. Infant sleeping position and the risk of sudden infant death syndrome in California, 19972000. Am J Epidemiol 2003;157:446-55. 17. Hoffman HJ, Damus K, Hillman L, Krongrad E. Risk factors for SIDS: results of the National Institutes of Child Health and Human Development SIDS Cooperative Epidemiological Study. Ann NY Acad Sci 1988;533:1. 18. American Academy of Pediatrics, Task Force on Sudden Infant Death Syndrome. Diagnostic coding shifts, controversies regarding the sleeping environment, and new variables to consider in reducing risk. Pediatrics 2005;116:1245-55. 19. Crane JM, van den Hof MC, Dodds L, Armson BA, Liston R. Neonatal outcomes with placenta previa. Obstet Gynecol 1999;93:541-4. 20. Li DK, Wi S. Maternal placental abnormality and the risk of sudden infant death syndrome. Am J Epidemiol 1999;149:608-11. 21. Leach CE, Blair PS, Fleming PJ, Smith IJ, Platt MW, Berry PJ, et al. Epidemiology of SIDS and explained sudden infant deaths. CESDI SUDI Research Group. Pediatrics 1999;104:43. 22. Hodgman J, Pavlova Z, Hoppenbrouwers T, Moore K, Peckham J, Sleep S, et al. Clinical predictors of imminent brain death. Clin Res 1987;35:234A. 23. Hoppenbrouwers T. Ontogenesis of circadian respiratory rhythms. J Interdisc Cycle Res 1986;17:140-1. 24. Thach B. The role of respiratory control disorders in SIDS. Res Physiol Neurobiol 2005;149:343-53. 25. Poets CE, Meny RG, Chobanian MR, Bonofiglo RE. Gasping and other cardiorespiratory patterns during sudden infant deaths. Pediatr Res 1999;45:350-4. 26. Patterson DS, Trachtenberg FL, Thompson EG, Belliveau RA, Beggs AH, Darnall R, et al. Multiple serotonergic brainstem abnormalities in sudden infant death syndrome. JAMA 2006;296:2124-32.
Extreme and Conventional Cardiorespiratory Events and Epidemiologic Risk Factors for SIDS
641