Adverse Effects of Pertussis and Rubella Vaccines

Adverse Effects of Pertussis and Rubella Vaccines: A Report of the Committee to Review the Adverse Consequences of Pertussis and Rubella Vaccines.

Evidence Concerning Pertussis Vaccines and Central Nervous System Disorders, Including Infantile Spasms, Hypsarrhythmia, Aseptic Meningitis, and Encephalopathy


Clinical Description

Infantile spasms are a type of epileptic disorder in young children characterized by flexor (34 percent), extensor (22 percent), and mixed flexor-extensor (42 percent) seizures that tend to occur in clusters or flurries (Kellaway et al., 1979). The earliest manifestations of infantile spasms can be subtle and are easily missed, making it difficult to identify the precise age at onset.

Infantile spasms, in combination with an electroencephalogram (EEG) pattern of hypsarrhythmia and psychomotor retardation or regression, is referred to as West syndrome. Approximately 80 percent of infants with infantile spasms have, at some time, a characteristic EEG pattern of hypsarrhythmia, whereas this pattern is seen in only ~4 percent of cases with other types of epilepsy (Jeavons and Bower, 1964). The hypsarrhythmic EEG pattern usually disappears with maturation, and ~50 percent of cases may have normal EEGs by age 8 years, although ~65 percent of children with infantile spasms will go on to have other types of seizures (Glaze and Zion, 1985).

Descriptive Epidemiology

Age-specific incidence rates are not available, although the vast majority of studies report a peak onset between ages 4 and 6 months (Cowan and Hudson, in press). For 85 to 90 percent of cases, onset of spasms is within the first year of life. Incidence rates of infantile spasms range from 0.25 per 1,000 live births in Denmark and the United States to 0.4 per 1,000 live births in Finland (Leviton and Cowan, 1981).

Most investigators divide infantile spasms cases into two categories which are defined on the basis of the presence or absence of a presumed cause and the child’s developmental status prior to the onset of spasms. What are commonly referred to as “symptomatic cases” are those in whom a presumed cause can be identified. Idiopathic cases are defined as infants with no identifiable causes for their spasms. This group is further subdivided by some into cryptogenic (those for whom there is no known cause of infantile spasms and whose development was essentially normal prior to the onset of spasms; ~10 percent of all cases) and doubtful (those for whom there is no known cause of infantile spasms but whose development prior to the onset of spasms may have been delayed).

Those cases considered to be idiopathic range between 30 and 50 percent (Cowan and Hudson, in press), although this proportion may be declining because of more sensitive diagnostic methods, such as neuroimaging techniques and positron tomography (Chugani et al., 1990). However, although approximately 70 to 90 percent of infantile spasms cases are reported to have abnormal computed tomography (CT) scans (Glaze and Zion, 1985; Pinsard and Saint-Jean, 1985), the significance of some CT diagnoses, for example, cortical atrophy, has been questioned (Ludwig, 1987). Thus, it is unclear that the proportion of infantile spasms cases considered to be idiopathic is really decreasing because of improved diagnosis of cerebral anomalies.

Among symptomatic cases, presumed causes are frequently grouped according to the timing of the suspected insult as occurring pre-, peri-, or postnatally. Prenatal factors are thought to account for 20 to 30 percent of cases. This category includes cerebral anomalies, chromosomal disorders, neurocutaneous syndromes such as tuberous sclerosis, inherited metabolic disorders, intrauterine infections, family history of seizures, and microcephaly (Bobele and Bodensteiner, 1990; Kurokawa et al., 1980; Ohtahara, 1984; Riikonen and Donner, 1979). Perinatal factors are thought to account for from 25 to 50 percent of infantile spasms cases. This category includes perinatal hypoxia, birth trauma, and metabolic disorders (Kurokawa et al., 1980; Pollack et al., 1979). Approximately 8 to 14 percent of infantile spasms are attributed to postnatal factors, including central nervous system (CNS) infections, trauma, immunizations, and intracranial hemorrhage (Bobele and Bodensteiner, 1990; Gibbs et al., 1954; Kurokawa et al., 1980; Lombroso, 1983a). Few of these factors have been subjected to systematic investigation, however, and the etiology of infantile spasms remains unknown for 30 to 50 percent of cases (Cowan and Hudson, in press).

History of Suspected Association with Pertussis Vaccines

Among the earliest case reports suggesting a possible link between infantile spasms and pertussis immunization are those of Baird and Borofsky (1957). They described 24 children who had hypsarrhythmia and infantile myoclonic seizures and whose development prior to the onset of spasms was apparently normal. Nine cases of infantile spasms were reported to have occurred between 1 and 5 days after DPT vaccination. Three of these nine children also had a history of perinatal complications that the authors thought might have been related to a risk of infantile spasms. The authors also stated, on the basis of a review of published EEG tracings, that hypsarrhythmia was present in two of the affected children described by Byers and Moll (1948). Since these early case reports, additional cases of infantile spasms in association with pertussis immunization have been described in the literature (Fukuyama et al., 1977; Millichap, 1987; Portoian-Shuhaiber and Al Rashied, 1986). The time intervals reported between vaccination and the onset of infantile spasms have been from minutes to weeks (Melchior, 1971).

Evidence from Studies in Humans

Case Reports and Case Series

One of the largest case series of infantile spasms following pertussis immunization was published by Millichap (1987). Six children ranging in age from 2 to 9 months were included. The time interval from immunization to the onset of spasms was from 6.5 hours to 5 days, and first seizures were reported to have occurred in conjunction with the first, second, or third doses of pertussis vaccine. Except for one case who had experienced myoclonic seizures since birth, no mention was made of the children having seizures prior to immunization. In reviewing the etiology and treatment of infantile spasms, Millichap (1987) listed the postulated mechanisms for pertussis-related seizures as (1) a direct neurotoxic effect, (2) an immediate immune reaction, (3) delayed cellular hypersensitivity reaction, and (4) vaccine-induced activation of a latent neurotropic virus infection.

In addition to the variability in age at the time of onset of spasms, associated vaccine dose, and time from immunization to the onset of spasms, there was no consistent pattern in the types of neurologic abnormalities reported in conjunction with infantile spasms. These included spastic diplegia, psychomotor retardation, hypotonic diplegia, and progressive neurologic deterioration. Not all children with infantile spasms have other neurologic or developmental problems, and when they do, diversity of expression of these associated neurologic conditions is typically reported (Lacy and Penry, 1976). This case series provides some of the better clinical descriptions available in the published literature of seizures occurring after immunization with DPT. Although typical of many cases of infantile spasms, information from this series also suggests that there is no consistent syndrome of neurologic manifestations among children whose spasms follow DPT immunization.

Fukuyama and colleagues (1977) studied 185 cases of infantile spasms seen in the Department of Pediatrics of the Tokyo Women’s Medical College from 1968 to 1972. Table 2 of their paper lists “DPT or DT” as one of the types of vaccines to which cases were exposed, whereas the text and all other tables and figures refer to “DPT or DP.” Thus, although there is some uncertainty about the precise vaccines to which these children were exposed, the committee considered DP to be the exposure the authors intended to describe. Complete information on immunization histories and health status prior to vaccination was available for 110 of the 185 infantile spasms cases. Of these 110 children, 22 (20 percent) had been immunized within 1 month of the onset of spasms, 10 with DPT or DP vaccine alone, 5 with DPT vaccine in combination with one or more other vaccines, 4 with smallpox vaccine alone, 2 with Japanese encephalitis vaccine alone, and 1 with polio vaccine alone. Of the 15 cases of infantile spasms with onset after immunization with either DPT or DP vaccine alone or DPT vaccine in combination with another vaccine, onset occurred after the first immunization in 3 cases, after the second in 10 cases, and after the third in 2 cases. The interval from immunization to the reported onset of spasms ranged from less than 48 hours to more than 7 days. The remaining cases had been vaccinated either more than 1 month before or more than 1 month after the onset of spasms (n = 44, 40 percent) or had never been immunized (n = 44, 40 percent). The authors gave no indication that any of the cases had had whooping cough, either before or after the onset of infantile spasms.

The authors considered vaccination as the etiology of infantile spasms if cases met the following three criteria: (1) no other identifiable cause, (2) normal development prior to the onset of spasms, and (3) the interval from immunization to the onset of spasms was within 48 hours for pertussis-containing vaccines and within 18 days for smallpox, polio, and Japanese encephalitis vaccines. Given these criteria, 5 of the 110 cases were considered by the authors to have infantile spasms caused by vaccination. It was not possible to determine from the data given in the paper how many of these five cases followed administration of DPT vaccine, since detailed information was given only for three of the five cases. At least one of the five cases occurred following smallpox vaccination alone, and at least two occurred following administration of DP vaccine.

It could not be determined from the information provided whether cases were representative of all those with infantile spasms from a defined geographic area or whether they were a selected group who were referred to these experts in pediatric neurology. The investigators acknowledged that because there is no biologic marker for vaccine-associated infantile spasms, the assignment of cause was made “solely from the clinical standpoint.” They stated that because of the diversity of the etiology of infantile spasms, “there is still free space for any agent to be suspected as an injurious factor causative of infantile spasms” (Fukuyama et al., 1977, p. 229).

Jeavons and colleagues (1970) reported on a follow-up of 98 cases of infantile spasms, 13 of which were attributed to immunization (type not specified). The follow-up ranged from 4 to 12 years. Outcomes were similar in the cryptogenic and immunization groups, among whom the survivorship, percent without neurologic abnormality at follow-up, and percent in regular school were higher than for those cases of infantile spasms attributed to perinatal or other causes (e.g., tuberous sclerosis).

Factors that should be considered in evaluating the study findings are that the patient groups were highly selected, the different lengths of follow-up were not considered in comparing outcomes among the groups, criteria for defining mental outcome were not given, and developmental status at follow-up was not ascertained uniformly for all cases. The first weakness affects the generality of the findings, and the last three problems given above make it difficult to compare outcomes between the groups studied.

Fifty-eight cases of infantile spasms (International Classification of Disease [ICD] 9 code 345.6 includes hypsarrhythmia and drop seizures) occurring within 28 days of DPT immunization were reported through the Centers for Disease Control’s (CDC’s) Monitoring System for Adverse Events Following Immunization(MSAEFI) system from 1978 to 1990, a period in which approximately 80.1 million doses of DPT vaccine were administered through public mechanisms in the United States (J. Mullen, Centers for Disease Control, personal communication, 1990). Of these 58 cases, 41 (71 percent) also received at least one other vaccine at the time of DPT immunization. No follow-up of the cases was made, and a physicians’s diagnosis was not required.

Controlled Epidemiologic Studies

If pertussis immunization were an important cause of infantile spasms, then one could expect a change in the ages at which immunizations were given to be followed by a change in the ages at the time of onset of infantile spasms. This issue was specifically addressed in a study by Melchior (1977) that examined changes in the distributions of ages of onset of infantile spasms and changes in the ages of immunization in Denmark. Prior to April 1, 1970, DPT vaccine was given to Danish children at ages 5, 6, 7, and 15 months. After that date, monovalent pertussis vaccine was given at ages 5 and 9 weeks and 10 months.

Melchior (1977) compared the distributions of ages at the time of onset of infantile spasms for two time periods, 1957 to 1967 and 1970 to 1975, which encompassed the different immunization schedules. Although there was some increase from the first to the second time period in the percentage of cases with onset under age 3 months (12 versus 23 percent), there was no significant difference in the overall distributions of age at onset for the two time periods. In both time intervals, the peak ages at onset for infantile spasms were in the 4- to 6-month range.

In addition to the comparison of the age distributions, medical records of the 113 cases of infantile spasms from 1970 to 1975 were examined to determine possible etiologies. Sixty cases were considered by the authors to be symptomatic, 40 were considered to be cryptogenic, and 13 were due to immunization. Of the 13 cases attributed to vaccination, 6 occurred after receipt of the monovalent pertussis vaccine and 7 occurred after receipt of diphtheria-tetanus-polio triple vaccine. Thus, infantile spasms occurring after immunization were reported in approximately equal numbers following administration of pertussis-and non-pertussis-containing vaccines.

After mid-1970, the “potency of the pertussis vaccine was reduced by 20 percent and the aluminum adjuvant was removed” (Shields et al., 1988, p. 802). Thus, immunization schedule was not the only factor that was different in the two time periods. In addition, the total number of immunizations given in the population for pertussis and for diphtheria-tetanus-polio was not reported, and therefore, the rate of infantile spasms associated with each type of immunization cannot be determined and, therefore, it is not possible to determine whether the risks are equivalent.

Another potential limitation of Melchior’s (1977) study is that cases identified for the first time interval (i.e., 1957 to 1967) were taken from a previous study and did not represent a nationwide survey or a national sample of all cases. Thus, it is possible that they had an unusual distribution of onset ages and were not appropriate for comparison with the 1970 to 1975 cases, which included all children with infantile spasms in Denmark. However, the range of peak age at the time of onset for the cases from the earlier interval corresponds to that usually reported, and thus, they are probably not a biased group with respect to age.

A similar analysis, also based on data from Denmark, was done by Shields and colleagues (1988). The study considered the frequencies of epilepsy, febrile seizures, infantile spasms (as a subgroup of all cases of epilepsy), and CNS infections (bacterial meningitis and aseptic meningitis) in children aged 1 month to 2 years identified from hospital or outpatient clinic records from 12 of 22 pediatric departments in Denmark. Two time periods, 1967 to 1968 and 1972 to 1973, were selected for comparison to reflect changes in the immunization schedule and in vaccine composition.

The exact dates of pertussis immunization were known for 372 children in the first time period and for 432 children in the second time period. Comparison of the distributions of the ages at the time of immunization for the two time intervals showed a marked difference in the frequency of immunization at different ages, corresponding with the ages at which immunizations were recommended. That is, in the 1967 to 1968 interval the peak ages at immunization were 5, 6, 7, and 15 months, while for the 1972 to 1973 interval immunizations peaked at ages 5 and 9 weeks and 10 months. Despite this difference, however, there was no significant difference in the age distributions of incident cases of infantile spasms in the two time periods. The results of this study are thus not consistent with the hypothesis that pertussis immunization is associated with the risk of infantile spasms, since there was no change in the distribution of ages at the time of onset when the ages at immunization were changed. However, only 80 cases were included in the study, and given this relatively small sample size, the study had a low statistical power to detect a difference in the distributions unless the association of infantile spasms and pertussis immunization was relatively large (see Appendix D). For instance, even if 29 percent of all cases of infantile spasms were caused by DPT immunization, the data of Shields and colleagues would have only about a 50 percent chance of finding a significant difference. To have an 80 percent power, about 40 percent of all infantile spasms cases would have to be caused by DPT. The data abstracters were not masked to the hypothesis of the study, but all events in a defined population were included, and no attempt was made during data collection to relate the events to the time of immunization.

The North West Thames Study (Pollock and Morris, 1983) describes voluntary reports of suspected vaccine reactions from 1975 through 1981 and a separate review of hospitalized cases of neurologic disorders in children for 1979. During the 7 years of the study, approximately equal numbers of children in the population completed courses of DPT and DT immunizations (134,700 and 133,500, respectively). Most of these children were also given oral polio vaccine. During this 7-year interval, 1,172 reports of ”vaccine-associated” events were received. Of these, 926 (79 percent) were considered to be “simple” reactions. Of the remaining 246 reports, 114 (10 percent) children experienced anaphylaxis or collapse, convulsions, neurologic disorders, or death. Forty-five (39 percent) of these more serious events were observed following receipt of DPT or monovalent pertussis vaccines, 20 (18 percent) occurred following DT immunization, 37 (32 percent) followed administration of the measles vaccine, and the remaining 12 (11 percent) followed immunization for rubella or other infectious diseases.

Five of the 114 children with more serious vaccine-associated reactions identified through the voluntary reporting system were diagnosed with infantile spasms. Among these five children, four had received DPT vaccine from 8 days to 6 weeks prior to the onset of spasms, and 1 had received the DT vaccine. The onset of infantile spasms reportedly occurred 1 month prior to immunization in the latter case. On the basis of these data, the relative risk (RR) is 4.0, but the 95 percent confidence interval (CI) is wide: 0.6 to 25.2. Despite the large denominators for these rates, the power of this test is low: 50 percent for an RR of 6.3 and 80 percent for an RR of 14.0.

In the review of discharge diagnoses for 1979, there were 682 children less than age 2 years who had relevant neurologic illnesses, and hospital records were obtained for 642 of them (94 percent). Five hundred twenty six (82 percent) of these children had febrile convulsions, but only three children with infantile spasms in association with immunization were reported from the review of discharge diagnoses. One child with infantile spasms attributed to Haemophilus influenzae meningitis had received DPT vaccine 19 days prior to the onset of spasms. A second child developed infantile spasms 6 weeks after DPT immunization, and the third child had onset of infantile spasms 12 weeks after immunization with the DT vaccine. Neither the expected number of cases of infantile spasms in a population of the size studied nor the number of cases identified in children who had not been immunized was reported. Thus, it is not possible to determine whether the observed cases were in excess of the expected number.

Results based on data from voluntary reporting of events thought to be associated with immunization and those based on data from review of discharge diagnoses are somewhat different. Although the number of cases of infantile spasms is small in both instances, voluntary reporting might suggest that infantile spasms occurred more often after DPT than after DT immunization, whereas review of discharge diagnoses found one case occurring after DPT immunization and one after DT immunization. The opportunity for bias is greater in the voluntary reporting data, since if a particular exposure is under suspicion as a cause of infantile spasms (in this case, the exposure being DPT), it is more likely that events occurring in temporal association with that exposure will be reported.

Walker and colleagues (1988) identified from medical and pharmacy records all cases of neurologic illnesses without an apparent predisposing cause in approximately 26,600 children born in Group Health Cooperative hospitals from 1972 to 1983. Medical records for cases and a control group born at the same hospitals during the same calendar period were reviewed for information on immunization status. Fifty-five cases of first afebrile seizures were identified; two of these children had infantile spasms, but the onset of spasms did not occur within 30 days of DPT immunization in either of them. The authors pointed out that since adrenocorticotropic hormone and steroids were not among the drugs for which pharmacy records were screened, some cases of infantile spasms may have been missed. However, only if these children had also not been hospitalized would they have been completely excluded from the study. In addition, children recently immunized with DPT vaccine would have to be more likely to be missed than children immunized more than 30 days prior to the onset of spasms.

The largest controlled study of the association between immunization and risk of infantile spasms was done among cases identified as part of the British National Childhood Encephalopathy Study (NCES) (Bellman et al., 1983a). This study is described in more detail later in this chapter. Briefly, the study included 269 children aged 2 to 35 months admitted to hospitals in England, Scotland, and Wales with a diagnosis of infantile spasms. Of these cases, 64 percent had EEGs with typical or atypical hypsarrhythmia, 30 percent had other EEG abnormalities, and 6 percent were reported to have normal EEGs (Bellman, 1983). Two controls were chosen for each case and were matched for age, sex, and area of residence. Immunization histories of cases and controls were obtained from the records of the children’s general practitioners. Risk of infantile spasms associated with immunization was assessed within four time intervals, defined by the following days postimmunization: 0 to 6 days, 7 to 13 days, 14 to 20 days, and 21 to 28 days. For the first period, the RR was 1.2 with a 95 percent CI of 0.5 to 3.0 (Miller et al., 1988). With a sample of the size used, there was 50 percent power to detect an RR of 2.5 and 80 percent power to detect an RR of 3.7.

Among the cases, 9 percent had been immunized with DPT vaccine within the preceding 28 days and 8 percent had been immunized with DT vaccine during the same time interval. Comparable percentages for the matched controls were 13 percent for DPT vaccine and 9 percent for DT vaccine. Immunization with neither DPT nor DT vaccine was statistically significantly associated with an increased risk of infantile spasms in any 7-day interval examined. However, risks of infantile spasms were higher within the first 7 days following administration of both DPT and DT vaccines than they were for the other three time periods, when there appeared to be a deficit of infantile spasms cases (RRs for the four time periods 0 to 6, 7 to 13, 14 to 20, and 21 to 28 days were 1.2, 0.6, 0.4, and 0.6, respectively, following DPT immunization and 1.3, 0.7, 0.8, and 0.5, respectively, following DT immunization). These differences in risk across time periods, however, were not statistically significant. Similar results were observed when analyses were confined to the 152 cases who were apparently neurologically normal prior to the onset of infantile spasms (RRs for the four time periods 0 to 6, 7 to 13, 14 to 20, and 21 to 28 days were 2.5, 0.3, 0.5, and 1.5, respectively, following DPT immunization and 2.0, 0.4, 1.0, and 0.3, respectively, following DT immunization). Whether the apparent clustering of cases that was observed within the first 6 days after immunization for both DPT and DT represents a triggering phenomenon, bias in assigning the date of onset of spasms, or simply a chance observation cannot be determined from these data. Looking at cases immunized within 28 days of diagnosis (a period similar to that used in the other controlled studies on infantile spasms), the RR was 0.6 (95 percent CI, 0.4 to 1.0) for all children in the NCES study and 0.7 (95 percent CI, 0.5 to 1.6) for previously normal children (Bellman et al., 1983a). The power of a test based on these data is somewhat higher than one based on data from the early period only (i.e., 0 to 6 days). For all children in the study, there was 50 percent power to detect an RR of 1.6 and 80 percent power to detect an RR of 2.0. For the previously normal children, the respective RRs were 1.9 and 2.4.

The NCES is the largest population-based, controlled study of the association of immunization and risk of infantile spasms. A limitation of the NCES data with respect to infantile spasms was the lack of a uniform case definition, in that children were considered infantile spasms cases if they were so designated by the admitting physician (Bellman et al., 1983b). Those conducting the NCES were notified of cases by physicians from all of England, Scotland, and Wales, and no set of standardized clinical criteria were used. In addition, 41 percent (48 of 116) of previously normal infantile spasms cases were in the “normal-normal” group (Alderslade et al., 1981). That is, they were considered to be neurologically normal both before their initial admission for infantile spasms and at 15 days postadmission or discharge. Although the prognosis for children with infantile spasms without a known cause and who are developmentally normal prior to the onset of spasms is reported to be better than that for symptomatic cases (Lacy and Penry, 1976), 41 percent is a rather high proportion of cases to “recover” from infantile spasms within 2 weeks. This raises the question as to whether these children really had infantile spasms, because the diagnosis was not confirmed and no uniform rules for diagnosis were applied to the group of potential cases. What effect the inclusion of children without infantile spasms would have had on the analysis depends on the true nature of the associations of their conditions with pertussis vaccination.

Comparisons of the estimates of risk of infantile spasms done separately for DPT and DT vaccinees can be used to examine the influence of the pertussis component of the vaccine. The fact that nearly identical results were observed for children who received the DPT and DT vaccines suggests that exposure to the pertussis component of the DPT vaccine does not increase the risk of infantile spasms.

The Study of Neurological Illness in Children (SONIC) was a large case-control investigation of the association between the risk of serious acute neurologic illness and DPT immunization in young children. A detailed description of SONIC is given later in this chapter. Briefly, the study was conducted in the states of Washington and Oregon from August 1, 1987, through July 31, 1988, and included children aged 1 to 24 months. Cases were identified primarily through systematic review of emergency room, outpatient clinic, and inpatient discharge listings. A panel of international experts on neurologic illnesses in children confirmed diagnoses by review of medical records and the use of uniform, prespecified criteria. The panel was unaware of the immunization history of cases. Two controls per case were selected from birth certificate registries of the states of Washington and Oregon. Controls were matched to cases by age (within 5 days), sex, and county of birth. Immunization histories for both cases and controls were obtained from interviews with parents, and attempts were made to validate these data by using medical records.

Preliminary findings from SONIC have been reported recently (Gale et al., 1990). In the population studied, 10 incident cases of infantile spasms were identified. Of these, three had onset of spasms within 28 days following immunization with DPT. A six fold increased risk of infantile spasms among children exposed to DPT within 28 days was observed. These results suggest the possibility that recent exposure to DPT is related to an increased risk of infantile spasms. However, the number of cases on which this estimate is based is small, and thus, the confidence interval is wide (95 percent CI = 0.6-57.7), indicating that the estimate of risk of infantile spasms observed in SONIC was very imprecise. The power of the statistical test was correspondingly low: 50 percent for an RR of 9.6 and 80 percent for an RR of 25.4. Because of the small number of cases of infantile spasms, estimates could not be calculated for exposure intervals shorter than 28 days.

Hunt (1983) reported on the association between the time of vaccination and the onset of seizures among individuals with tuberous sclerosis who responded to a survey questionnaire. Of 150 families contacted through the Tuberous Sclerosis Association of Great Britain, 97 (65 percent) responded. Of the responders, 82 (84 percent) had had seizures, 66 (80 percent) of whom had infantile spasms. The age range of cases in the survey was less than 1 to 51 years. Outcome was compared among subgroups of responders, defined on the basis of their immunization status at the time of their first seizure. Of the 82 people with tuberous sclerosis who had seizures, 20 had never been immunized, 27 had been immunized after their first seizure, 17 had been immunized within 1 month prior to their first seizure, and 18 had been immunized more than 1 month prior to their first seizure. Profoundly handicapped children, defined as those older than age 5 who could neither walk nor talk, were more often observed among the tuberous sclerosis cases with seizures who were immunized after their first seizure (8 of 27). Of those immunized after their first seizure and for whom the type of immunization was known, the frequency of profound handicap was 6 of 13 who received DT vaccine and 2 of 14 who received DPT vaccine. All of the profoundly handicapped children had their first seizure before the age of 7 months.

Although this study suggests that DPT vaccine does not add to the seizure burden among children with tuberous sclerosis or increase the risk of neurologic handicap, the study design has weaknesses that reduce the utility of its results in addressing the question of an association between pertussis and the risk of infantile spasms. For example, the sampling frame from which subjects were chosen is not representative of all people with tuberous sclerosis, and the response rate was low. The age range of cases was also very wide, with some individuals being as old as 51 years, thus introducing the possibility of recall bias regarding immunization histories.


Case reports of children who develop infantile spasms after receipt of pertussis vaccine prompt concern regarding a possible relation between immunization and seizures. However, the reported time intervals between immunization and onset vary widely, from hours to months, and no consistent pattern of timing or associated neurologic disorders is reported. Given the insidious onset of infantile spasms, the temporal relation of immunization and onset is difficult to establish with certainty.

The body of evidence concerning the possible relation between vaccination with DPT or its pertussis component and infantile spasms includes a number of case reports, case series, and four controlled observational epidemiologic studies, which are summarized in Table 4-1. Risk estimates were not consistent and varied widely across studies, ranging from 0.3 to 6.0, depending on the time interval examined. None of the risk estimates was statistically significant, and the NCES had sufficient statistical power (80 percent) to detect an RR of 2.0 to 2.4, depending on which data are used to make the comparison. Direct comparisons between studies is hampered by differences in the definitions of infantile spasms cases and the time intervals used. Although the results tended to be inconsistent, most controlled studies did not observe an increased risk. Only two studies reported risks greater than 2.5: the analysis of voluntary reporting data from the North West Thames study and SONIC. The risk estimate from SONIC is highly uncertain because of the small number of cases on which it is based.

Continue to source for remainder of study: