Vol. 101, No. 5
AMERICAN JOURNAL OF EPIDEMIOLOGY
Copyright © 1975 by The Johns Hopkins University
Printed mU.S A.
THE SUDDEN INFANT DEATH SYNDROME AND EPIDEMIC VffiAL DISEASE KENRAD E. NELSON,1 MARK A. GREENBERG,2 MAURICE A. MUFSON, VIJAI K. MOSES Nelson, K. E. (Dept. of Preventive Medicine and Community Health. U. of Illinois College of Medicine, Chicago, IL 60612), M. A. Greenberg, M. A. Mufson, V. K. Moses. The sudden infant death syndrome and epidemic viral disease. Am J Epidemiol 101:423-430. 1975.—A study was done to investigate the relationship between the sudden infant death syndrome (SIDS) and epidemic respiratory viral disease among hospitalized children under 18 months of age. During the 42 month-period of this study, there were 778 sudden infant deaths in Chicago and 3244 hospital admissions of children under 18 months for respiratory disease. Four outbreaks of respiratory syncytial (RS) virus infections, three outbreaks of influenza A virus infections, and several small clusters of parainfluenza virus infections occurred during the course of this study. Influenza A was the only virus infection found to have a statistically significant association with SIDS. Although environmental temperature was also significantly correlated with SIDS, the association with influenza A virus infection was independent of this temperature effect and neither association was strong. Infant mortality; respiratory tract diseases
The sudden infant death syndrome (SIDS), also known as crib death or cot death, is one of the major causes of mortality among infants one week to one year of age. In developed countries, it accounts for about one-third of all deaths in infants within this age group. Although the etiology of SIDS remains unknown, the consistent epidemiologic characteristics of SIDS Received for publication October 7, 1974. Abbreviations: RS, respiratory syncytial; SIDS, sudden infant death syndrome. From the Departments of Preventive Medicine and Community Health and of Medicine, Abraham Lincoln School of Medicine, University of Illinois, Chicago, IL 60612. 1 To whom requests for reprints should be addressed at Department of Preventive Medicine and Community Health, University of Illinois College of Medicine, 835 South Wolcott Street, Chicago, IL 60612. 2 Present address: Department of Medicine, Montefiore Hospital, Bronx, NY. The authors acknowledge the technical assistance of Helen E. Krause, Hilda E. Mocega-Gonzalez and Raymond E. Wasil. Sandra Feiveson and Joan Greenberg provided statistical and clerical assistance. Dr. Andrew Toman and his staff at the Cook County Coroner's office were very helpful in allowing us access to data and assisting in these studies.
include an increased frequency in wintertime in temperate climates and an inverse relationship to socioeconomic class (1-4). These two features, together with the common pathologic signs in the respiratory tract of intrathoracic petechiae, minimal pulmonary congestion or edema, and basement membrane thickening and edema in the larynx (2, 5), have led a number of investigators to hypothesize that a virus infection of the respiratory tract is an etiologic factor (6-12). Several investigators have isolated viruses from a substantial number of such infants (6, 8, 12-19). Indeed, two recent controlled studies have found an increased frequency of virus isolations from SIDS infants as compared to controls (8, 13). However, in none of these studies could a single virus or a group of related viruses be implicated in the etiology of SIDS; moreover in most studies, a virus could be recovered in fewer than 50 per cent of the cases. Another approach to investigating the viral hypothesis is afforded by examining
423
Downloaded from https://academic.oup.com/aje/article-abstract/101/5/423/120395 by Insead user on 18 August 2018
424
NELSON, GREENBERG, MUFSON AND MOSES
the relationships of SIDS to epidemic clusters of virus infections of the respiratory tract among infants in the community. In fact viruses that could play a role in SIDS might be difficult or impossible to recover from autopsy specimens either because of technical reasons or the characteristics of the virus. For example, if the virus were extremely labile to temperature or pH change, it would be difficult to recover from such specimens, since most often the infant is dead for 12-24 hours or more before an examination is performed (2). Further, an "incomplete" virus might be involved as has recently been shown in acute post-measles encephalitis (20) as well as in various chronic "slow virus" diseases such as subacute sclerosing panencephalitis and progressive multifocal leukoencephalopathy (21). If an incomplete virus were involved, recovery of the agent would be impossible using conventional methodology. If those viruses occurring in well-defined epidemics play an etiologic role in SIDS, their occurrence should be related temporally to clusters of SIDS providing the population contains sufficient infants in the appropriate age group. This study examines the relationship between the occurrence of SIDS and epidemic clusters of respiratory virus infections reflected in hospital admissions for serious respiratory tract disease of children between one week and 18 months of age in Chicago during a 180-week interval from January 1, 1967, to June 30, 1970. As far as the authors are aware, this is the first statistical study of the temporal association between SIDS and epidemics of respiratory virus disease among a large urban population. MATERIAL AND METHODS
Data for this study were obtained from two sources. First, the number of SIDS by week was ascertained by reviewing records of the Cook County Coroner's Office and death certificates from the office of Vital
Statistics of the Department of Public Health, State of Illinois. Secondly, the identification of specific viral infections associated with serious lower respiratory tract disease in infants hospitalized at the Cook County Hospital was determined on a weekly basis during an ongoing prospective study of these hospital admissions (22, 23). Since Cook County is a large public hospital, and one which serves that portion of the Chicago population which has the highest rate of SIDS, these virologic data represented a satisfactory sentinel of serious epidemic viral respiratory disease in the general population under study. The relationship between the weekly incidence of SIDS and infants hospitalized with specific virus infections or total hospital admissions for respiratory infections was examined using multiple linear regression analyses. The t-test was used to test for significance of the regression coefficients and to evaluate the strength of relationships between the occurrence of SIDS and specific viral infections. The mean number of viral infections occurring during weeks of outbreaks was compared with those weeks of the study without infections. The standardized regression coefficients and the unique variance were used to estimate the relative strength and the independent contributions of the two variables with a significant relationship to SIDS: influenza A virus infection and average environmental temperature. The unique variance expresses the amount of the variance in the dependent variable that could be accounted for by the influence of each of the significant independent variables operating without interaction. This allowed us to determine whether the relationship between SIDS and influenza A virus infection was a spurious one which might be accounted for by their common association with variations in temperature, or temperature dependent (but unmeasured) variables. The standardized coefficient also expresses the relative
Downloaded from https://academic.oup.com/aje/article-abstract/101/5/423/120395 by Insead user on 18 August 2018
425
SIDS AND EPIDEMIC VIRAL DISEASE
strength and direction of the statistical relationship between the dependent variables and the independent variable. The statistical analyses were done on an IBM Model 370/155 computer using programs from the Statistical Package for the Social Sciences (SPSS) and the Data-Text System (24). To determine all sudden and unexpected deaths that occurred among infants 7 days to 1 year of age in the city of Chicago for the period of the study, January 1, 1967, to June 30, 1970, records of the Cook County Coroner's Office including autopsy results, other records describing the nature of the death, and data on the death certificates were reviewed. If any of these data suggested that a death might not have been sudden or unexpected, or that the infant might have had an underlying significant illness, the infant was excluded from the study. Especially at the beginning of the study, all infants who died suddenly and unexpectedly outside a hospital and whose autopsy findings or other clinical history data obtainable from professional or other sources did not suggest another specific cause of death were included in this study despite the fact that the death certificates rarely indicated that these deaths were from SIDS. The criteria for SIDS in our study were similar to that utilized by Peterson (4) in a retrospective study of the syndrome in Seattle. The virologic data were obtained from an ongoing prospective cross-sectional study conducted during the same period (22, 23). In this study, approximately one-half of the infants less than 18 months of age with lower respiratory tract disease admitted to Cook County Hospital were tested for virus infections by isolation techniques or serologic procedures or both. Oropharyngeal swab specimens were obtained from each infant and inoculated into HEp-2, human diploid fibroblast (strain WI-38) and rhesus monkey kidney cells, and occasionally human embryonic kidney cells when avail-
able. Acute- and convalescent-phase sera were obtained from slightly more than one-half of the infants tested for virus isolation. Paired sera were tested by complement fixation using microtiter procedures for antibodies to respiratory syncytial virus, influenza A and parainfluenza 1, 2 and 3 viruses. RESULTS
To identify epidemic clusters, 1724 (53.1 per cent) of the 3244 infants admitted to Cook County Hospital for lower respiratory disease during the study period were tested for virus infections. During the same period a total of 778 deaths that met our criteria for SIDS occurred in Chicago; 260 of these were white and 518 were black infants (table 1). The weekly incidence of five myxovirus infections, the total admissions of infants with lower respiratory tract disease and the occurrence of SIDS during each of the 180 weeks of this study are depicted graphically in figure 1. Respiratory syncytial (RS) virus, influenza A virus and, to a lesser degree, parainfluenza 1 virus tended to occur in clusters. Parainfluenza 3 virus TABLE 1
Association of sudden infant deaths with mean weekly environmental temperature and hospitalizations of infants with myxovirus infections of the respiratory tract in Chicago from January 1, 1967, to June 30, 1970 Standardized regression coefficients for SIDS by race Black White (N = 518) (N - 260) Virus Influenza A Respiratory syncytial Parainfluenza 1 Parainfluenza 2 Parainfluenza 3 Average temperature * p < 0.05.
t p < 0.01.
Downloaded from https://academic.oup.com/aje/article-abstract/101/5/423/120395 by Insead user on 18 August 2018
Total (N = 778)
0.176* 0.096
0.037 0.062
0.162* 0.110
0.020 -0.043 -0.022 -0.218t
0.029 0.066 0.016 0.141
0.032 0.000 0.090 -0.247*
426
NELSON, GREENBERG, MUFSON AND MOSES Lower Respirotor> Troct Disease
. .J.I
,
i
Paramf luenzo 1 Virus 1 1. .1 Parainfluenza 2 Virus
ol
0" 50
.
4...IJ.
.
Ill
J._
-
Poramfluenzo I Virus
... Liil LJiL
fcLki 1966
I
iJL 1969
1970
WEEKS
FIGURE 1. Total infant admissions for lower respiratory tract disease and infections with five myxoviruses by week, Cook County Hospital, and sudden infant deaths by week in Chicago from January 1, 1967 to June 30, 1970.
was isolated with fair regularity throughout this period, and parainfluenza 2 virus was an uncommon isolate. The peaks in admission rates of infants with lower respiratory tract disease paralleled the clusters of RS virus infections. The limited number of influenza A infections detected were associated with increases of hospital admissions to a lesser degree. The incidence of SIDS, however, did not appear to parallel total hospital admissions for lower respiratory tract disease. The correlation between the weekly incidence of SIDS and the weekly admissions for lower respiratory tract disease was not statistically significant, (r = 0.083, p > 0.25). The relationship between the weekly incidence of SIDS cases and virus isolations by week among hospitalized infants at Cook County Hospital was studied by simple regression analysis. The only statistically significant regression was between
SIDS and influenza A virus infections, p < 0.05 (table 1). The weekly incidence of SIDS did not bear a statistically significant relationship to the other myxovirus infections studied. These findings were not changed by examining the statistical association between SIDS and virus infections detected 1 or 2 weeks before, 1 or 2 weeks after, or for a 4-week average of the number of infections. To determine that the association between SIDS and influenza A infections was not merely a reflection of the known inverse association between SIDS and temperature (that is, the increased frequency of both SIDS and influenza in the winter), the weekly incidence of SIDS was regressed on the weekly average ambient temperature in Chicago. There was a statistically significant negative regression of SIDS and weekly average temperature, p < 0.01 (table 1). Furthermore, when influenza A virus infections and temperature effects on SIDS were examined using a multiple linear regression model (according to the formula y = a + ft,x, + ^2*2, where y = weekly number of SIDS, Xi = the weekly average temperature, and x2 = weekly influenza A virus infections) a significant independent relationship with both of these two variables was shown. Both influenza A virus infections and temperature contributed about 5 per cent unique variance to the incidence of SIDS when the other variable was controlled (table 2). This suggests that the significant relationship between SIDS and influenza A virus infections was not merely a chance association of these two variables in winter. It should also be emphasized that the contribution of influenza A infections to the total SIDS incidence during this study, though significant, was not major. DISCUSSION
Perhaps the most important findings in our study are the lack of a significant temporal association between SIDS cases
Downloaded from https://academic.oup.com/aje/article-abstract/101/5/423/120395 by Insead user on 18 August 2018
SIDS AND EPIDEMIC VIRAL DISEASE
427
County Hospital, Further, the highest rates of SIDS occurred among the population of infants living in the same census tracts and from the same socioeconomic class as those having a high rate of admission to Cook County Hospital (26). The Standardized Unique criteria for diagnosis of SIDS in this study regression Variable variancet coefficient was similar to that used in a study from Seattle reported by Peterson (4). 0.2201: 0.047 Influenza A virus infection Recent studies have shown an increased Average temperature -0.212}: 0.044 rate of isolation of viruses from infants * Analysis of variance: F = 9.26, p < 0.001: the dying suddenly and unexpectedly as comhigh level of significance supports the appropriateness pared to explained deaths (8, 13). The of the use of the multiple linear regression model in predominant viruses isolated include this analysis (see Materials and Methods). mainly enteroviruses, adenoviruses and tThe "'unique variance" is the variance associated rhinoviruses. Most of these studies have with each independent variable when it is entered as the last variable after all other independent variables not implicated a single agent or a group of have been added (24). In this case, the nearly equal closely related agents (6, 8, 13, 18). In one values of unique variance indicate that influenza A study, the respiratory virus infections devirus infection and average temperature explain equal tected in cases of SIDS were not found to amounts of the variation in the weekly SIDS deaths. be unusually extensively invasive or associ%p < 0.01. ated with viremia (27). Furthermore, imand hospital admission for lower respira- portant lower respiratory pathogens for tory tract disease or RS virus infections. this age group such as influenza, parainfluThese data are in agreement with those enza and RS viruses have been recovered reported by Foy and Ray (25) in which no uncommonly from SIDS cases. Ray and association was apparent between the rates Beckwith (8) failed to isolate influenza or of the clinical syndromes of pneumonia, RS viruses from 80 SIDS cases; they recroup or bronchiolitis among infants and covered only 1 RS virus from families of children whose families were enrolled in a SIDS cases, while 3 RS viruses were relarge group health plan and the incidence covered from their 30 control cases. Brandt of SIDS in Seattle. Some biases are inher- (13) isolated RS virus from only 1 of 70 ent in studies of this type since it is SIDS cases, whereas 20 per cent of 451 necessary to sample large populations of patients with acute respiratory disease young infants for considerable periods of were shedding RS virus. Urquhart and time in order to obtain sufficient data for Grist (19) also failed to isolate RS virus or analysis. The virologic data in our study influenza virus from 72 sudden infant were obtained from an ongoing study of deaths. In contrast to these studies was a respiratory virus infections in a hospital recent report by Ferris et al. in which population of infants under 18 months of small quantities of several pathogenic age (22, 23). While these data relate specif- myxoviruses including RS virus were isoically to this hospital population, data lated or identified by indirect immunofluofrom other clinical and public health labo- rescent staining of post-mortem lung tissue ratories in Chicago indicate that similar in 13 of 51 sudden infant deaths (12). epidemics of RS virus and influenza virus Technical factors may be responsible for infections were in progress among infants the frequent failure to isolate RS virus, throughout the Chicago area that were influenza virus and parainfluenza viruses. coincident with those discovered at Cook These agents tend to be quite fastidious TABLE 2
Multiple linear regression analysis of weekly sudden infant deaths and hospitalizations of infants for influenza A virus infections with weekly ambient temperature in Chicago from January 1. 1967, to June 30, 1970*
Downloaded from https://academic.oup.com/aje/article-abstract/101/5/423/120395 by Insead user on 18 August 2018
428
NELSON, GREENBERG, MUFSON AND MOSES
and are more sensitive to deleterious effects of temperature and pH variation than enteroviruses. Since SIDS infants are examined only after death and the length of time prior to testing is variable and often unknown, the conditions are rarely optimal for demonstrating the most fastidious viruses. Ray et al. (8) increased their recovery of viral agents from infants with SIDS from 16 to 53 per cent by examining all specimens unfrozen. There are reasons to suspect that respiratory viruses, especially RS, parainfluenza and influenza viruses, might be more relevant to SIDS than the enteroviruses or adenoviruses which have been isolated more commonly. Myxovirus infections tend to be more common in the winter season, are known to be commonly pathogenic among young infants of the age group usually affected by SIDS, and primarily involve the respiratory tract where histopathologic changes have been found in infants with SIDS. Furthermore, the syndrome of hyperacute reactions to natural RS virus infections in children previously exposed to an inactive RS virus vaccine suggests a mechanism of heightened immunologic responsiveness to this agent that might be so acute as to. culminate in a sudden death in some infants (28-30). If the immunologic hyperreactive state were to be present naturally without exposure to the killed vaccine, sudden death might occur in an infant (12, 29). One can postulate that maternal antibodies might sensitize the infant. These antibodies would disappear at about the time in life that the frequency of SIDS decreases markedly. In contrast, since the viruses that have been isolated most frequently from SIDS cases in most studies are often associated with low-grade pathogenicity or the carrier state, it is hard to be certain that a virus isolated from such an infant after death is related to the death. While comparison with a suitably selected control population is useful in establishing this relationship, it
would also be essential to demonstrate an increase of SIDS cases during an epidemic of viral infections to be sure the association was real. For all of these reasons, it is of considerable importance to explore the temporal association between viral infections that are known to be epidemic and the occurrence of SIDS in the same population. Our study, which took place over a 180-week period in Chicago, showed a significant association between the weekly incidence of SIDS and infants hospitalized with respiratory tract disease associated with influenza A virus infection. No such association was detected for RS virus despite the fact that four epidemics of RS virus infection occurred in this population during the course of this study. Peaks of virus infections occurred in each year of the study period as did increases of SIDS; both occurred during the winter. Nevertheless, the maximum prevalence of RS virus activity either preceded or followed increases in SIDS incidence usually by 4-6 weeks and the relationship between the two variables was inconstant. These data in conjunction with other recent studies (8, 13, 19) are taken as evidence against an important role of RS virus infections in the sudden infant death syndrome. The widely held belief that most SIDS is a hyperacute reaction to RS virus infection may not be correct (11, 12). The association of SIDS with influenza A virus infection in our study was of interest. Although influenza A virus was not frequently isolated from SIDS cases in other studies, influenza A virus was isolated from two SIDS in the study of Ferris et al. (12). Although morbidity and mortality from influenza A virus has been thought to involve principally adults, especially the elderly, a recent report by Brocklebank et al. (31) indicates that some influenza epidemics may cause frequent severe disease in very young children and infants. The excess mortality observed in the 1957 and
Downloaded from https://academic.oup.com/aje/article-abstract/101/5/423/120395 by Insead user on 18 August 2018
429
SIDS AND EPIDEMIC VIRAL DISEASE
1968 outbreaks of influenza A involved primarily the elderly; however, SIDS was not specifically studied in these outbreaks. Certainly further studies of this relationship seem warranted. It should be emphasized, however, that the association between influenza A and SIDS was weak, accounting for only about 5 per cent of the variance in SIDS incidence. Its effect was independent of but of similar magnitude to the seasonal effect, which was studied by examining the relationship between SIDS and ambient temperature. The latter association has been noted consistently in nearly every study of the epidemiology of SIDS (1-4). It is quite possible that viral infections may be only one of a variety of stimuli that act through a final common pathway leading to SIDS. The sensitivity of the end organ, perhaps the respiratory tract or the central nervous system, in the predisposed child might be as important a determinant in an infant's risk of SIDS as the nature of the stimulus, which might vary from one infant to the next (32). Several stimuli have been postulated or studied by various research workers in connection with SIDS including changes in serum electrolytes, environmental pollutants, nasal obstruction, aberrant or poorly developed neurologic reflexes such as the diving reflex, apneic spells and others. The numerous etiologic theories have been reviewed by Valdes-Dapena (1, 33, 34). Steinschneider has recently reported the possible association of SIDS with prolonged apneic episodes (35). In addition Naeye has noted hypertrophy and hyperplasia of smooth muscle fibers in the small pulmonary arteries in cases of SIDS (36). He interprets this finding to represent a response to alveolar hypoxia, possibly as a result of previous apneic episodes. It is possible that in an abnormal infant with chronic recurrent apneic episodes, a virus infection that in a normal host might be mild or only moderately severe might be
associated with sudden death. It seems clear also that if an association exists between virus infections and SIDS it is not a simple or direct one. REFERENCES
1. Valdes-Dapena MA: Sudden and unexpected death in infancy: A review of the world's literature. Pediatrics 39:123-138, 1967 2. Adelson L, Kinney ER: Sudden and unexpected death in infancy. Pediatrics 17:663-699, 1956 3. Bergman AB, Ray, CG, Pomeroy, MA, et al: Studies of the sudden infant death syndrome in King County, Washington. III. Epidemiology. Pediatrics 49:860-870, 1972 4. Peterson DR: Sudden unexpected deaths in infants. Am J Epidemiol 84:478-482, 1966 5. Adelson L: Specific studies of infant victims of sudden death. Proceedings of the Conference of Causes of Sudden Death in Infants. Edited by RJ Wedgewood and EP Benditt. Bethesda, NICHHD, USPHS Pub No 1412, 1963, pp 11-40 6. Gold E, Carver DH, Heinberg H, et al: Viral infection, a possible cause of sudden unexpected death in infants. N Engl J Med 264:53-60, 1961 7. Valdes-Dapena MA, Hummeler K: Sudden and unexpected deaths in infants. II. Viral infections as causative factors. J Pediatr 63:398-401, 1963 8. Ray CG, Beckwith JB: Studies of the S.I.D.S. in King County, Washington. I. The role of viruses. JAMA 211:610-623, 1970 9. Blattner RJ: Possible role of virus infections in SIDS. J Pediatr 65:631-633, 1964 10. Gunther M: Cot deaths: Anaphylactic reaction after intrauterine infection as another possible cause. Lancet 1:912-914, 1966 11. Jackson GG: Respiratory syncytial viral infections. In Cecil-Loeb, Textbook of Medicine, 13th ed. Edited by PB Beeson and W McDermott. London, WB Saunders, 1971, pp 367-368 12. Ferris JAJ, Aherne WA, Locke WS, et al: Sudden and unexpected deaths in infants: histology and virology. Br Med J 2:439-441, 1973 13. Brandt CD: Infectious agents from SIDS cases and members of their community. In Sudden Infant Death Syndrome. Proceedings of the Second International Conference on Causes of Sudden Death in Infants. Edited by AB Bergman, JB Beckwith and CG Ray. Seattle, University of Washington Press, 1970, pp 161-174 14. Horstmann DM, Hsiung GD: Myxovirus infections and respiratory illnesses in children. Clin Pediatr 2:378-386, 1963 15. Moore ML, Hooser LE, Davis EV, et al: Sudden unexpected death in infancy and isolations of ECHO Type 7 virus. Proc Soc Exp Biol Med
Downloaded from https://academic.oup.com/aje/article-abstract/101/5/423/120395 by Insead user on 18 August 2018
430
NELSON, GREENBERG, MUFSON AND MOSES
116:231-234, 1964 16. Balduzzi PC, Greendyke RM: Sudden unexpected death in infancy and viral infection. Pediatrics 38:201-206, 1966 17. Houstek J, Holy J: Respiratory viruses and the syndrome of sudden and unexpected death in babies and small infants. Cesk Pediatr 23:243-252, 1968 18. O'Reilly MJJ, Whiley MK: Cot deaths in Brisbane. Med J Austral 2:1084-1087, 1967 19. Urquhart GED, Grist NR: Virological studies of sudden unexplained infant deaths in Glasgow, 1967-1970. J Clin Pathol 25:443-446, 1972 20. ter Meulen V, Muller D, Kackell Y, et al: Isolation of infectious measles virus in measles encephalitis. Lancet 2:1172-1175, 1972 21. Weiner LP, Johnson RT, Herndon RM: Viral infections and demyelinating diseases. N Engl J Med 288:1103-1110, 1973 22. Mufson MA, Krause HE, Mocega HE, et al: Viruses, Mycoplasma pneumoniae and bacteria associated with lower respiratory tract disease among infants. Am J Epidemiol 91:192-202, 1970 23. Mufson MA, Levine HD, Wasil RE, et al: Epidemiology of respiratory syncytial virus infection among infants and children. Am J Epidemiol 98: 88-95, 1973 24. Armor DJ, Couch AS: Data-Text Primer—An Introduction to Computerized Social Data Analysis. New York, The Free Press, 1972, p 227 25. Foy HM, Ray CG: Epidemiology of sudden infant death syndrome and lower respiratory tract disease in young children: a comparison. Am J Epidemiol 98:69-71, 1973 26. Greenberg MA, Nelson KE, Carnow BW: A study of the relationship between sudden infant death syndrome and environmental factors. Am J Epidem 98:412-422, 1973 27. Ray CG, Hebestreit NM: Studies of the sudden
28.
29. 30.
31. 32.
33.
34. 35. 36.
infant death syndrome in King County, Washington. II. Attempts to demonstrate evidence of viremia. Pediatrics 48:79-82, 1971 Chanock RM, Parrott RH, Kapikian AZ, et al: Possible role of immunological factors in pathogenesis of RS virus lower respiratory tract disease. Perspect Virol 6:125, 1968 Urquhart GED, Logan RW, Izatt MM: Sudden unexplained death in infancy and hyperimmunization. J Clin Pathol 24:736-739, 1971 Gardner PS, McQuillin J, Court SDM: Speculation on pathogenesis in death from respiratory syncytial virus infection. Br Med J 1:327-330, 1970 Brocklebank JT, Court SDM, McQuillin J, et al: Influenza A infection in children. Lancet 2:497-500, 1972 Bergman AB: Summary statement. Sudden Infant Death Syndrome. Proceedings of the Second International Conference on Causes of Sudden Death in Infants. Edited by AB Bergman, JB Beckwith, CG Ray, Seattle, University of Washington Press, 1970, pp 209-212 Valdes-Dapena MA: Theories of etiology, review of the literature. In Sudden infant Death Syndrome. Proceedings of the Second International Conference on Causes of Sudden Death in Infants. Edited by AB Bergman, JB Beckwith and CG Ray. Seattle, University of Washington Press, 1970 Valdes-Dapena MA: Sudden, unexpected and unexplained ueath in infancy—A status report—1973. N Engl J Med 289:1195-1197,1973 Steinschne' ier A: Prolonged apnea and the sudden infant death syndrome: Clinical and laboratory observations. Pediatrics 50:646-654, 1972 Naeye RL: Pulmonary arterial abnormalities in the sudden-infant-death syndrome. N Engl J Med 289:1167-1170, 1973
Downloaded from https://academic.oup.com/aje/article-abstract/101/5/423/120395 by Insead user on 18 August 2018
AMERICAN JOURNAL OF EPIDEMIOLOGY
Copyright © 1975 by The Johns Hopkins University
Printed mU.S A.
THE SUDDEN INFANT DEATH SYNDROME AND EPIDEMIC VffiAL DISEASE KENRAD E. NELSON,1 MARK A. GREENBERG,2 MAURICE A. MUFSON, VIJAI K. MOSES Nelson, K. E. (Dept. of Preventive Medicine and Community Health. U. of Illinois College of Medicine, Chicago, IL 60612), M. A. Greenberg, M. A. Mufson, V. K. Moses. The sudden infant death syndrome and epidemic viral disease. Am J Epidemiol 101:423-430. 1975.—A study was done to investigate the relationship between the sudden infant death syndrome (SIDS) and epidemic respiratory viral disease among hospitalized children under 18 months of age. During the 42 month-period of this study, there were 778 sudden infant deaths in Chicago and 3244 hospital admissions of children under 18 months for respiratory disease. Four outbreaks of respiratory syncytial (RS) virus infections, three outbreaks of influenza A virus infections, and several small clusters of parainfluenza virus infections occurred during the course of this study. Influenza A was the only virus infection found to have a statistically significant association with SIDS. Although environmental temperature was also significantly correlated with SIDS, the association with influenza A virus infection was independent of this temperature effect and neither association was strong. Infant mortality; respiratory tract diseases
The sudden infant death syndrome (SIDS), also known as crib death or cot death, is one of the major causes of mortality among infants one week to one year of age. In developed countries, it accounts for about one-third of all deaths in infants within this age group. Although the etiology of SIDS remains unknown, the consistent epidemiologic characteristics of SIDS Received for publication October 7, 1974. Abbreviations: RS, respiratory syncytial; SIDS, sudden infant death syndrome. From the Departments of Preventive Medicine and Community Health and of Medicine, Abraham Lincoln School of Medicine, University of Illinois, Chicago, IL 60612. 1 To whom requests for reprints should be addressed at Department of Preventive Medicine and Community Health, University of Illinois College of Medicine, 835 South Wolcott Street, Chicago, IL 60612. 2 Present address: Department of Medicine, Montefiore Hospital, Bronx, NY. The authors acknowledge the technical assistance of Helen E. Krause, Hilda E. Mocega-Gonzalez and Raymond E. Wasil. Sandra Feiveson and Joan Greenberg provided statistical and clerical assistance. Dr. Andrew Toman and his staff at the Cook County Coroner's office were very helpful in allowing us access to data and assisting in these studies.
include an increased frequency in wintertime in temperate climates and an inverse relationship to socioeconomic class (1-4). These two features, together with the common pathologic signs in the respiratory tract of intrathoracic petechiae, minimal pulmonary congestion or edema, and basement membrane thickening and edema in the larynx (2, 5), have led a number of investigators to hypothesize that a virus infection of the respiratory tract is an etiologic factor (6-12). Several investigators have isolated viruses from a substantial number of such infants (6, 8, 12-19). Indeed, two recent controlled studies have found an increased frequency of virus isolations from SIDS infants as compared to controls (8, 13). However, in none of these studies could a single virus or a group of related viruses be implicated in the etiology of SIDS; moreover in most studies, a virus could be recovered in fewer than 50 per cent of the cases. Another approach to investigating the viral hypothesis is afforded by examining
423
Downloaded from https://academic.oup.com/aje/article-abstract/101/5/423/120395 by Insead user on 18 August 2018
424
NELSON, GREENBERG, MUFSON AND MOSES
the relationships of SIDS to epidemic clusters of virus infections of the respiratory tract among infants in the community. In fact viruses that could play a role in SIDS might be difficult or impossible to recover from autopsy specimens either because of technical reasons or the characteristics of the virus. For example, if the virus were extremely labile to temperature or pH change, it would be difficult to recover from such specimens, since most often the infant is dead for 12-24 hours or more before an examination is performed (2). Further, an "incomplete" virus might be involved as has recently been shown in acute post-measles encephalitis (20) as well as in various chronic "slow virus" diseases such as subacute sclerosing panencephalitis and progressive multifocal leukoencephalopathy (21). If an incomplete virus were involved, recovery of the agent would be impossible using conventional methodology. If those viruses occurring in well-defined epidemics play an etiologic role in SIDS, their occurrence should be related temporally to clusters of SIDS providing the population contains sufficient infants in the appropriate age group. This study examines the relationship between the occurrence of SIDS and epidemic clusters of respiratory virus infections reflected in hospital admissions for serious respiratory tract disease of children between one week and 18 months of age in Chicago during a 180-week interval from January 1, 1967, to June 30, 1970. As far as the authors are aware, this is the first statistical study of the temporal association between SIDS and epidemics of respiratory virus disease among a large urban population. MATERIAL AND METHODS
Data for this study were obtained from two sources. First, the number of SIDS by week was ascertained by reviewing records of the Cook County Coroner's Office and death certificates from the office of Vital
Statistics of the Department of Public Health, State of Illinois. Secondly, the identification of specific viral infections associated with serious lower respiratory tract disease in infants hospitalized at the Cook County Hospital was determined on a weekly basis during an ongoing prospective study of these hospital admissions (22, 23). Since Cook County is a large public hospital, and one which serves that portion of the Chicago population which has the highest rate of SIDS, these virologic data represented a satisfactory sentinel of serious epidemic viral respiratory disease in the general population under study. The relationship between the weekly incidence of SIDS and infants hospitalized with specific virus infections or total hospital admissions for respiratory infections was examined using multiple linear regression analyses. The t-test was used to test for significance of the regression coefficients and to evaluate the strength of relationships between the occurrence of SIDS and specific viral infections. The mean number of viral infections occurring during weeks of outbreaks was compared with those weeks of the study without infections. The standardized regression coefficients and the unique variance were used to estimate the relative strength and the independent contributions of the two variables with a significant relationship to SIDS: influenza A virus infection and average environmental temperature. The unique variance expresses the amount of the variance in the dependent variable that could be accounted for by the influence of each of the significant independent variables operating without interaction. This allowed us to determine whether the relationship between SIDS and influenza A virus infection was a spurious one which might be accounted for by their common association with variations in temperature, or temperature dependent (but unmeasured) variables. The standardized coefficient also expresses the relative
Downloaded from https://academic.oup.com/aje/article-abstract/101/5/423/120395 by Insead user on 18 August 2018
425
SIDS AND EPIDEMIC VIRAL DISEASE
strength and direction of the statistical relationship between the dependent variables and the independent variable. The statistical analyses were done on an IBM Model 370/155 computer using programs from the Statistical Package for the Social Sciences (SPSS) and the Data-Text System (24). To determine all sudden and unexpected deaths that occurred among infants 7 days to 1 year of age in the city of Chicago for the period of the study, January 1, 1967, to June 30, 1970, records of the Cook County Coroner's Office including autopsy results, other records describing the nature of the death, and data on the death certificates were reviewed. If any of these data suggested that a death might not have been sudden or unexpected, or that the infant might have had an underlying significant illness, the infant was excluded from the study. Especially at the beginning of the study, all infants who died suddenly and unexpectedly outside a hospital and whose autopsy findings or other clinical history data obtainable from professional or other sources did not suggest another specific cause of death were included in this study despite the fact that the death certificates rarely indicated that these deaths were from SIDS. The criteria for SIDS in our study were similar to that utilized by Peterson (4) in a retrospective study of the syndrome in Seattle. The virologic data were obtained from an ongoing prospective cross-sectional study conducted during the same period (22, 23). In this study, approximately one-half of the infants less than 18 months of age with lower respiratory tract disease admitted to Cook County Hospital were tested for virus infections by isolation techniques or serologic procedures or both. Oropharyngeal swab specimens were obtained from each infant and inoculated into HEp-2, human diploid fibroblast (strain WI-38) and rhesus monkey kidney cells, and occasionally human embryonic kidney cells when avail-
able. Acute- and convalescent-phase sera were obtained from slightly more than one-half of the infants tested for virus isolation. Paired sera were tested by complement fixation using microtiter procedures for antibodies to respiratory syncytial virus, influenza A and parainfluenza 1, 2 and 3 viruses. RESULTS
To identify epidemic clusters, 1724 (53.1 per cent) of the 3244 infants admitted to Cook County Hospital for lower respiratory disease during the study period were tested for virus infections. During the same period a total of 778 deaths that met our criteria for SIDS occurred in Chicago; 260 of these were white and 518 were black infants (table 1). The weekly incidence of five myxovirus infections, the total admissions of infants with lower respiratory tract disease and the occurrence of SIDS during each of the 180 weeks of this study are depicted graphically in figure 1. Respiratory syncytial (RS) virus, influenza A virus and, to a lesser degree, parainfluenza 1 virus tended to occur in clusters. Parainfluenza 3 virus TABLE 1
Association of sudden infant deaths with mean weekly environmental temperature and hospitalizations of infants with myxovirus infections of the respiratory tract in Chicago from January 1, 1967, to June 30, 1970 Standardized regression coefficients for SIDS by race Black White (N = 518) (N - 260) Virus Influenza A Respiratory syncytial Parainfluenza 1 Parainfluenza 2 Parainfluenza 3 Average temperature * p < 0.05.
t p < 0.01.
Downloaded from https://academic.oup.com/aje/article-abstract/101/5/423/120395 by Insead user on 18 August 2018
Total (N = 778)
0.176* 0.096
0.037 0.062
0.162* 0.110
0.020 -0.043 -0.022 -0.218t
0.029 0.066 0.016 0.141
0.032 0.000 0.090 -0.247*
426
NELSON, GREENBERG, MUFSON AND MOSES Lower Respirotor> Troct Disease
. .J.I
,
i
Paramf luenzo 1 Virus 1 1. .1 Parainfluenza 2 Virus
ol
0" 50
.
4...IJ.
.
Ill
J._
-
Poramfluenzo I Virus
... Liil LJiL
fcLki 1966
I
iJL 1969
1970
WEEKS
FIGURE 1. Total infant admissions for lower respiratory tract disease and infections with five myxoviruses by week, Cook County Hospital, and sudden infant deaths by week in Chicago from January 1, 1967 to June 30, 1970.
was isolated with fair regularity throughout this period, and parainfluenza 2 virus was an uncommon isolate. The peaks in admission rates of infants with lower respiratory tract disease paralleled the clusters of RS virus infections. The limited number of influenza A infections detected were associated with increases of hospital admissions to a lesser degree. The incidence of SIDS, however, did not appear to parallel total hospital admissions for lower respiratory tract disease. The correlation between the weekly incidence of SIDS and the weekly admissions for lower respiratory tract disease was not statistically significant, (r = 0.083, p > 0.25). The relationship between the weekly incidence of SIDS cases and virus isolations by week among hospitalized infants at Cook County Hospital was studied by simple regression analysis. The only statistically significant regression was between
SIDS and influenza A virus infections, p < 0.05 (table 1). The weekly incidence of SIDS did not bear a statistically significant relationship to the other myxovirus infections studied. These findings were not changed by examining the statistical association between SIDS and virus infections detected 1 or 2 weeks before, 1 or 2 weeks after, or for a 4-week average of the number of infections. To determine that the association between SIDS and influenza A infections was not merely a reflection of the known inverse association between SIDS and temperature (that is, the increased frequency of both SIDS and influenza in the winter), the weekly incidence of SIDS was regressed on the weekly average ambient temperature in Chicago. There was a statistically significant negative regression of SIDS and weekly average temperature, p < 0.01 (table 1). Furthermore, when influenza A virus infections and temperature effects on SIDS were examined using a multiple linear regression model (according to the formula y = a + ft,x, + ^2*2, where y = weekly number of SIDS, Xi = the weekly average temperature, and x2 = weekly influenza A virus infections) a significant independent relationship with both of these two variables was shown. Both influenza A virus infections and temperature contributed about 5 per cent unique variance to the incidence of SIDS when the other variable was controlled (table 2). This suggests that the significant relationship between SIDS and influenza A virus infections was not merely a chance association of these two variables in winter. It should also be emphasized that the contribution of influenza A infections to the total SIDS incidence during this study, though significant, was not major. DISCUSSION
Perhaps the most important findings in our study are the lack of a significant temporal association between SIDS cases
Downloaded from https://academic.oup.com/aje/article-abstract/101/5/423/120395 by Insead user on 18 August 2018
SIDS AND EPIDEMIC VIRAL DISEASE
427
County Hospital, Further, the highest rates of SIDS occurred among the population of infants living in the same census tracts and from the same socioeconomic class as those having a high rate of admission to Cook County Hospital (26). The Standardized Unique criteria for diagnosis of SIDS in this study regression Variable variancet coefficient was similar to that used in a study from Seattle reported by Peterson (4). 0.2201: 0.047 Influenza A virus infection Recent studies have shown an increased Average temperature -0.212}: 0.044 rate of isolation of viruses from infants * Analysis of variance: F = 9.26, p < 0.001: the dying suddenly and unexpectedly as comhigh level of significance supports the appropriateness pared to explained deaths (8, 13). The of the use of the multiple linear regression model in predominant viruses isolated include this analysis (see Materials and Methods). mainly enteroviruses, adenoviruses and tThe "'unique variance" is the variance associated rhinoviruses. Most of these studies have with each independent variable when it is entered as the last variable after all other independent variables not implicated a single agent or a group of have been added (24). In this case, the nearly equal closely related agents (6, 8, 13, 18). In one values of unique variance indicate that influenza A study, the respiratory virus infections devirus infection and average temperature explain equal tected in cases of SIDS were not found to amounts of the variation in the weekly SIDS deaths. be unusually extensively invasive or associ%p < 0.01. ated with viremia (27). Furthermore, imand hospital admission for lower respira- portant lower respiratory pathogens for tory tract disease or RS virus infections. this age group such as influenza, parainfluThese data are in agreement with those enza and RS viruses have been recovered reported by Foy and Ray (25) in which no uncommonly from SIDS cases. Ray and association was apparent between the rates Beckwith (8) failed to isolate influenza or of the clinical syndromes of pneumonia, RS viruses from 80 SIDS cases; they recroup or bronchiolitis among infants and covered only 1 RS virus from families of children whose families were enrolled in a SIDS cases, while 3 RS viruses were relarge group health plan and the incidence covered from their 30 control cases. Brandt of SIDS in Seattle. Some biases are inher- (13) isolated RS virus from only 1 of 70 ent in studies of this type since it is SIDS cases, whereas 20 per cent of 451 necessary to sample large populations of patients with acute respiratory disease young infants for considerable periods of were shedding RS virus. Urquhart and time in order to obtain sufficient data for Grist (19) also failed to isolate RS virus or analysis. The virologic data in our study influenza virus from 72 sudden infant were obtained from an ongoing study of deaths. In contrast to these studies was a respiratory virus infections in a hospital recent report by Ferris et al. in which population of infants under 18 months of small quantities of several pathogenic age (22, 23). While these data relate specif- myxoviruses including RS virus were isoically to this hospital population, data lated or identified by indirect immunofluofrom other clinical and public health labo- rescent staining of post-mortem lung tissue ratories in Chicago indicate that similar in 13 of 51 sudden infant deaths (12). epidemics of RS virus and influenza virus Technical factors may be responsible for infections were in progress among infants the frequent failure to isolate RS virus, throughout the Chicago area that were influenza virus and parainfluenza viruses. coincident with those discovered at Cook These agents tend to be quite fastidious TABLE 2
Multiple linear regression analysis of weekly sudden infant deaths and hospitalizations of infants for influenza A virus infections with weekly ambient temperature in Chicago from January 1. 1967, to June 30, 1970*
Downloaded from https://academic.oup.com/aje/article-abstract/101/5/423/120395 by Insead user on 18 August 2018
428
NELSON, GREENBERG, MUFSON AND MOSES
and are more sensitive to deleterious effects of temperature and pH variation than enteroviruses. Since SIDS infants are examined only after death and the length of time prior to testing is variable and often unknown, the conditions are rarely optimal for demonstrating the most fastidious viruses. Ray et al. (8) increased their recovery of viral agents from infants with SIDS from 16 to 53 per cent by examining all specimens unfrozen. There are reasons to suspect that respiratory viruses, especially RS, parainfluenza and influenza viruses, might be more relevant to SIDS than the enteroviruses or adenoviruses which have been isolated more commonly. Myxovirus infections tend to be more common in the winter season, are known to be commonly pathogenic among young infants of the age group usually affected by SIDS, and primarily involve the respiratory tract where histopathologic changes have been found in infants with SIDS. Furthermore, the syndrome of hyperacute reactions to natural RS virus infections in children previously exposed to an inactive RS virus vaccine suggests a mechanism of heightened immunologic responsiveness to this agent that might be so acute as to. culminate in a sudden death in some infants (28-30). If the immunologic hyperreactive state were to be present naturally without exposure to the killed vaccine, sudden death might occur in an infant (12, 29). One can postulate that maternal antibodies might sensitize the infant. These antibodies would disappear at about the time in life that the frequency of SIDS decreases markedly. In contrast, since the viruses that have been isolated most frequently from SIDS cases in most studies are often associated with low-grade pathogenicity or the carrier state, it is hard to be certain that a virus isolated from such an infant after death is related to the death. While comparison with a suitably selected control population is useful in establishing this relationship, it
would also be essential to demonstrate an increase of SIDS cases during an epidemic of viral infections to be sure the association was real. For all of these reasons, it is of considerable importance to explore the temporal association between viral infections that are known to be epidemic and the occurrence of SIDS in the same population. Our study, which took place over a 180-week period in Chicago, showed a significant association between the weekly incidence of SIDS and infants hospitalized with respiratory tract disease associated with influenza A virus infection. No such association was detected for RS virus despite the fact that four epidemics of RS virus infection occurred in this population during the course of this study. Peaks of virus infections occurred in each year of the study period as did increases of SIDS; both occurred during the winter. Nevertheless, the maximum prevalence of RS virus activity either preceded or followed increases in SIDS incidence usually by 4-6 weeks and the relationship between the two variables was inconstant. These data in conjunction with other recent studies (8, 13, 19) are taken as evidence against an important role of RS virus infections in the sudden infant death syndrome. The widely held belief that most SIDS is a hyperacute reaction to RS virus infection may not be correct (11, 12). The association of SIDS with influenza A virus infection in our study was of interest. Although influenza A virus was not frequently isolated from SIDS cases in other studies, influenza A virus was isolated from two SIDS in the study of Ferris et al. (12). Although morbidity and mortality from influenza A virus has been thought to involve principally adults, especially the elderly, a recent report by Brocklebank et al. (31) indicates that some influenza epidemics may cause frequent severe disease in very young children and infants. The excess mortality observed in the 1957 and
Downloaded from https://academic.oup.com/aje/article-abstract/101/5/423/120395 by Insead user on 18 August 2018
429
SIDS AND EPIDEMIC VIRAL DISEASE
1968 outbreaks of influenza A involved primarily the elderly; however, SIDS was not specifically studied in these outbreaks. Certainly further studies of this relationship seem warranted. It should be emphasized, however, that the association between influenza A and SIDS was weak, accounting for only about 5 per cent of the variance in SIDS incidence. Its effect was independent of but of similar magnitude to the seasonal effect, which was studied by examining the relationship between SIDS and ambient temperature. The latter association has been noted consistently in nearly every study of the epidemiology of SIDS (1-4). It is quite possible that viral infections may be only one of a variety of stimuli that act through a final common pathway leading to SIDS. The sensitivity of the end organ, perhaps the respiratory tract or the central nervous system, in the predisposed child might be as important a determinant in an infant's risk of SIDS as the nature of the stimulus, which might vary from one infant to the next (32). Several stimuli have been postulated or studied by various research workers in connection with SIDS including changes in serum electrolytes, environmental pollutants, nasal obstruction, aberrant or poorly developed neurologic reflexes such as the diving reflex, apneic spells and others. The numerous etiologic theories have been reviewed by Valdes-Dapena (1, 33, 34). Steinschneider has recently reported the possible association of SIDS with prolonged apneic episodes (35). In addition Naeye has noted hypertrophy and hyperplasia of smooth muscle fibers in the small pulmonary arteries in cases of SIDS (36). He interprets this finding to represent a response to alveolar hypoxia, possibly as a result of previous apneic episodes. It is possible that in an abnormal infant with chronic recurrent apneic episodes, a virus infection that in a normal host might be mild or only moderately severe might be
associated with sudden death. It seems clear also that if an association exists between virus infections and SIDS it is not a simple or direct one. REFERENCES
1. Valdes-Dapena MA: Sudden and unexpected death in infancy: A review of the world's literature. Pediatrics 39:123-138, 1967 2. Adelson L, Kinney ER: Sudden and unexpected death in infancy. Pediatrics 17:663-699, 1956 3. Bergman AB, Ray, CG, Pomeroy, MA, et al: Studies of the sudden infant death syndrome in King County, Washington. III. Epidemiology. Pediatrics 49:860-870, 1972 4. Peterson DR: Sudden unexpected deaths in infants. Am J Epidemiol 84:478-482, 1966 5. Adelson L: Specific studies of infant victims of sudden death. Proceedings of the Conference of Causes of Sudden Death in Infants. Edited by RJ Wedgewood and EP Benditt. Bethesda, NICHHD, USPHS Pub No 1412, 1963, pp 11-40 6. Gold E, Carver DH, Heinberg H, et al: Viral infection, a possible cause of sudden unexpected death in infants. N Engl J Med 264:53-60, 1961 7. Valdes-Dapena MA, Hummeler K: Sudden and unexpected deaths in infants. II. Viral infections as causative factors. J Pediatr 63:398-401, 1963 8. Ray CG, Beckwith JB: Studies of the S.I.D.S. in King County, Washington. I. The role of viruses. JAMA 211:610-623, 1970 9. Blattner RJ: Possible role of virus infections in SIDS. J Pediatr 65:631-633, 1964 10. Gunther M: Cot deaths: Anaphylactic reaction after intrauterine infection as another possible cause. Lancet 1:912-914, 1966 11. Jackson GG: Respiratory syncytial viral infections. In Cecil-Loeb, Textbook of Medicine, 13th ed. Edited by PB Beeson and W McDermott. London, WB Saunders, 1971, pp 367-368 12. Ferris JAJ, Aherne WA, Locke WS, et al: Sudden and unexpected deaths in infants: histology and virology. Br Med J 2:439-441, 1973 13. Brandt CD: Infectious agents from SIDS cases and members of their community. In Sudden Infant Death Syndrome. Proceedings of the Second International Conference on Causes of Sudden Death in Infants. Edited by AB Bergman, JB Beckwith and CG Ray. Seattle, University of Washington Press, 1970, pp 161-174 14. Horstmann DM, Hsiung GD: Myxovirus infections and respiratory illnesses in children. Clin Pediatr 2:378-386, 1963 15. Moore ML, Hooser LE, Davis EV, et al: Sudden unexpected death in infancy and isolations of ECHO Type 7 virus. Proc Soc Exp Biol Med
Downloaded from https://academic.oup.com/aje/article-abstract/101/5/423/120395 by Insead user on 18 August 2018
430
NELSON, GREENBERG, MUFSON AND MOSES
116:231-234, 1964 16. Balduzzi PC, Greendyke RM: Sudden unexpected death in infancy and viral infection. Pediatrics 38:201-206, 1966 17. Houstek J, Holy J: Respiratory viruses and the syndrome of sudden and unexpected death in babies and small infants. Cesk Pediatr 23:243-252, 1968 18. O'Reilly MJJ, Whiley MK: Cot deaths in Brisbane. Med J Austral 2:1084-1087, 1967 19. Urquhart GED, Grist NR: Virological studies of sudden unexplained infant deaths in Glasgow, 1967-1970. J Clin Pathol 25:443-446, 1972 20. ter Meulen V, Muller D, Kackell Y, et al: Isolation of infectious measles virus in measles encephalitis. Lancet 2:1172-1175, 1972 21. Weiner LP, Johnson RT, Herndon RM: Viral infections and demyelinating diseases. N Engl J Med 288:1103-1110, 1973 22. Mufson MA, Krause HE, Mocega HE, et al: Viruses, Mycoplasma pneumoniae and bacteria associated with lower respiratory tract disease among infants. Am J Epidemiol 91:192-202, 1970 23. Mufson MA, Levine HD, Wasil RE, et al: Epidemiology of respiratory syncytial virus infection among infants and children. Am J Epidemiol 98: 88-95, 1973 24. Armor DJ, Couch AS: Data-Text Primer—An Introduction to Computerized Social Data Analysis. New York, The Free Press, 1972, p 227 25. Foy HM, Ray CG: Epidemiology of sudden infant death syndrome and lower respiratory tract disease in young children: a comparison. Am J Epidemiol 98:69-71, 1973 26. Greenberg MA, Nelson KE, Carnow BW: A study of the relationship between sudden infant death syndrome and environmental factors. Am J Epidem 98:412-422, 1973 27. Ray CG, Hebestreit NM: Studies of the sudden
28.
29. 30.
31. 32.
33.
34. 35. 36.
infant death syndrome in King County, Washington. II. Attempts to demonstrate evidence of viremia. Pediatrics 48:79-82, 1971 Chanock RM, Parrott RH, Kapikian AZ, et al: Possible role of immunological factors in pathogenesis of RS virus lower respiratory tract disease. Perspect Virol 6:125, 1968 Urquhart GED, Logan RW, Izatt MM: Sudden unexplained death in infancy and hyperimmunization. J Clin Pathol 24:736-739, 1971 Gardner PS, McQuillin J, Court SDM: Speculation on pathogenesis in death from respiratory syncytial virus infection. Br Med J 1:327-330, 1970 Brocklebank JT, Court SDM, McQuillin J, et al: Influenza A infection in children. Lancet 2:497-500, 1972 Bergman AB: Summary statement. Sudden Infant Death Syndrome. Proceedings of the Second International Conference on Causes of Sudden Death in Infants. Edited by AB Bergman, JB Beckwith, CG Ray, Seattle, University of Washington Press, 1970, pp 209-212 Valdes-Dapena MA: Theories of etiology, review of the literature. In Sudden infant Death Syndrome. Proceedings of the Second International Conference on Causes of Sudden Death in Infants. Edited by AB Bergman, JB Beckwith and CG Ray. Seattle, University of Washington Press, 1970 Valdes-Dapena MA: Sudden, unexpected and unexplained ueath in infancy—A status report—1973. N Engl J Med 289:1195-1197,1973 Steinschne' ier A: Prolonged apnea and the sudden infant death syndrome: Clinical and laboratory observations. Pediatrics 50:646-654, 1972 Naeye RL: Pulmonary arterial abnormalities in the sudden-infant-death syndrome. N Engl J Med 289:1167-1170, 1973
Downloaded from https://academic.oup.com/aje/article-abstract/101/5/423/120395 by Insead user on 18 August 2018
