Vaccine
Safety > Research
Risk of Hospitalization Due to Aseptic Meningitis
Following Measles-Mumps-Rubella Vaccination in One to Two Year Old Children:
An Analysis of the Vaccine Safety Datalink (VSD) Project
As published in
Pediatric Infectious Disease Journal 1997; 16: 500-3
Authors: |
Steven Black,
MD; Henry Shinefield, MD; Paula Ray, MPH; Edwin Lewis, MPH; Robert
Chen, MD; John Glasser, PhD; Steven Hadler, MD; Janet Hardy, MPH, MS;
Phil Rhodes, PhD; Emmet Swint, MA; Robert Davis, MD; Robert Thompson,
MD; John Mullooly, PhD; Michael Marcy, MD; Connie Vadheim, PhD; Joel
Ward, MD; Suresh Rastogi, PhD; Robert Wise, MD |
Abstract
Objective. To assess the level of increased risk, if any, of hospitalizations for
aseptic meningitis after Jeryl-Lynn mumps strain measles-mumps-rubella (MMR) vaccine in
the Vaccine Safety Datalink population.
Study Design. A possible increased risk of aseptic meningitis 8 to 14 days after
receipt of MMR was observed in a preliminary screening analysis of automated data from the
Vaccine Safety Datalink (VSD) project Year 2 analysis. To further evaluate this
association a retrospective ten year matched case-control study was undertaken in the four
health maintenance organizations (HMOs) in the VSD project. Cases ascertained from a broad
scan of the automated data were validated against a standard case definition. Two controls
matched on age, sex, HMO and HMO membership were assigned per case.
Results. The VSD project involves the cooperative collection of automated
vaccination and medical outcome data from four large HMOs that currently have 500,000
children younger than seven years of age under surveillance. Review of automated screening
results from the first two years of data revealed a possible increased risk of aseptic
meningitis 0 to 14 days after MMR with a relative risk of 3.61 (95% confidence interval,
1.0 to 13.1) although the total number of cases was small. Although the automated data had
suggested a possible association of aseptic meningitis with MMR containing the
Jeryl-Lynn
strain of mumps, review of validated hospitalized cases during the observation
period did not reveal evidence of an increased risk of aseptic meningitis after MMR
containing the Jeryl-Lynn strain of mumps (odds ratio < 1.0 for all analyses).
Conclusion. Although it is recognized that hospitalized cases represent a minority
of the total cases of aseptic meningitis, it is reassuring that in this evaluation no
increased risk of aseptic meningitis after MMR vaccine was found.
Introduction
Before introduction of mumps vaccine the mumps virus was the leading cause of viral
meningoencephalitis in the United States.1 Live attenuated
Jeryl-Lynn strain mumps virus vaccine was introduced into the United States in 1967 and
has been recommended for universal use in children 12 to 18 months of age since 1976.
Since its introduction there has been a steady decrease in the incidence of mumps and
mumps virus infection.2 However, MMR and mumps vaccines in other
countries containing Urabe strain attenuated mumps virus have been associated with cases
of aseptic meningitis, which has resulted in the cessation of mumps vaccination in Japan
and other countries.3-7 Aseptic meningitis after Urabe mumps vaccine has been
reported with an incidence as high as one per 3,800 doses of vaccine.3
Until 1991 the evaluation of potential vaccine safety issues depended on passive
reporting of events in the Vaccine Adverse Event Reporting System or on ad hoc studies.
Recognizing the potential of large clinical databases within several HMOs to provide a
resource for evaluation of vaccine safety issues, in 1991 the Centers for Disease Control
and Prevention (CDC) funded the Vaccine Safety Datalink (VSD).8 This
project links outcome and vaccine exposure information, as well as demographic and other
covariate information, from the automated clinical databases for 500,000 children younger
than seven years of age enrolled at Group Health of Puget Sound and the Kaiser Permanente
Medical Care Programs in Oregon and Northern California. Southern California Kaiser
Permanente was added in 1992. Within this project annual data tapes are utilized to screen
for possible associations of adverse events after vaccination and to evaluate existing
hypotheses from the Institute of Medicine2 and other sources.
In reviewing the combined automated data from the first two years of the project, a
possible association of aseptic meningitis after receipt of MMR was noted. During the 8
to14 day time window after receipt of MMR, there was an increased risk of aseptic
meningitis with a relative risk of 3.61 (95% confidence interval, 1.0-13.1) based on three
cases of aseptic meningitis. As of the date of review of these screening results,
outpatient data within the VSD were available from only two sites and then only since the
beginning of the study in 1991, whereas hospitalization data were available from the three
original sites for ten years. Therefore it was decided to make an initial evaluation of
the possible association through a retrospective review of ten years of hospitalization
data to identify all hospitalized cases of aseptic meningitis in children between 12 and
23 months of age, the most common age for administration of MMR.
Methods
To evaluate the risk of hospitalization for aseptic meningitis following all
immunizations, a matched case-control study was performed. Hospitalization data tapes have
been routinely maintained at three of the four HMOs for the ten year period 1984 to1993
although automated immunization information was not available for this entire time period.
At the Southern California site records were available for review for the two-year period
1992-1993. Potential cases were identified through these computerized hospitalization
records at the four centers. Children between 12 and 23 months of age were chosen because
the overwhelming majority of MMR doses over the study period were administered to this age
group. All medical records with discharge diagnoses with ICD-9 codes 045.2, 047.*, 048,
049.*, 072.1, 321.2, or 322.* among children one to two years of age were identified as
potential cases to be further investigated by medical records review.
Cases of aseptic meningitis were reviewed against a predefined case definition of no
evidence of prior underlying meningitis or underlying disease caused by
toxoplasmosis,
syphilis, cytomegalovirus, neonatal herpes simplex or human immunodeficiency virus. (The
same exclusion criteria were also used for controls.) In addition bacterial, mycobacterial
and fungal cultures of the cerebrospinal fluid must have been negative, and the patient
must have had a cerebrospinal fluid white blood cell count of > 10 cells/cubic
millimeter. Patients in whom the diagnosis was unclear or for whom there was a potential
diagnosis of encephalitis were referred to the principal investigator at each site or to a
pediatric infectious disease specialist for an evaluation that was blinded as to vaccine
status. Cases meeting the case validation criteria were accepted based on this medical
record review. Two controls were then selected per validated case matched by HMO, age (+
1 month), sex, and HMO membership status. Vaccination status for cases and controls was
ascertained from medical record review.
In a matched case-control analysis, the risk of developing aseptic meningitis within 8
to 14 days after vaccination or within 14 days or 30 days after vaccination was compared
with the risk of developing aseptic meningitis within the same time interval after the
index date in the matched controls. Relative risks with 95% confidence intervals were
estimated by the odds ratio obtained by fitting conditional logistic regression models to
the matched case-control data.9,10
Historical pharmacy records and membership files were used to estimate the total number
of children vaccinated and the number of children in the overall study population.
Confidence intervals for incidence were estimated assuming that the incidence may be
approximated by a Poisson distribution.
Results
Of the total initial number of 93 possible cases identified from the automated data, 59
cases of aseptic meningitis were identified as well as the additional diagnoses for the
possible cases not meeting the case definition are also shown in Table 1 .
Although this study was not intended to evaluate the risk of encephalopathy after
vaccination, it is of interest that none of the five cases of encephalopathy/encephalitis
identified had their onset within 30 days after the receipt of any vaccine.
The vaccines and vaccine combinations given within 30 days before any case, or the
corresponding reference date in controls, are shown in Table 2. As can be seen MMR was
given most commonly with other vaccines.
The results of the case-control analysis are shown in Table 3. There was no increased
risk of aseptic meningitis demonstrated within 8 to14 days, 14 days or 30 days after MMR
or any other vaccination. Also there were no cases of aseptic meningitis within seven days
of any vaccine. Approximately 300,000 doses of MMR were given to children 12 through 23
months of age during the approximately 350,000 person-years available for observation in
this age group during the study period at the four HMOs. One child was hospitalized with
aseptic meningitis within 14 days of MMR and three children were hospitalized within 30
days. The crude rate of disease was one case per 300,000 doses within 14 days and one case
per 100,000 doses within 30 days. This is equivalent to an incidence per 100,000
person-years of 8.7 within 14 days and 12.2 within 30 days, with 95% confidence internals
of 0.2 to 48.5 (14 days) and 2.5 to 35.6 (30 days). The upper bounds of the these
intervals are equivalent to one case occurring within the 14-day window per 54,000 doses
and one case within the 30-day window per 34,000 doses, permitting us to rule out
hypotheses that Jeryl-Lynn strain MMR is followed by hospitalized aseptic meningitis any
more often than these upper bounds. These are statistical estimates of the possible
attributable risk given the power of this study and the overall findings of the study do
not support an association of aseptic meningitis with receipt of MMR vaccine.
The overall rate of aseptic meningitis was 59 cases in an estimated 350,000 person
years of follow-up time or 16.9 cases per 100,000 person years (95% confidence interval,
12.5 to 21.2).
Discussion
In this retrospective analysis of hospitalization caused by aseptic meningitis, there
was no increased risk of aseptic meningitis after MMR vaccine containing the
Jeryl-Lynn
strain of mumps. This contrasts with several reports of an increased risk of aseptic
meningitis after Urabe strain MMR3-7 and is consistent with the observations of
Colville and Pugh3 and others that limited an increased risk of aseptic
meningitis to the Urabe strain of mumps vaccine.
Although this study analyzed only hospitalized cases, it is believed that the results
of this study are reflective of the risk of severe aseptic meningitis in that it was
standard practice to hospitalize almost all children with recognized aseptic meningitis in
this age group during the study period. However, milder cases or unrecognized cases may
have occurred without seeking medical treatment or without being recognized by physicians.
Because there was no case selection in vaccine exposures, analysis of the data based on
hospitalized cases should be valid. Also because the most severe cases are most likely to
be recognized and because these cases are potentially the most worrisome in terms of
sequelae, the results reported herein are reassuring.
The overall rate of confirmed cases of aseptic meningitis in the study population was
17.2 cases per 100,000 person years (95% confidence interval 12.5 to 21.2). This rate is
very close to the background rate of 16.2 cases per 100,000 person years in children one
to four years of age observed in Olmstead County, Minnesota.11 This similarity
between these rates also supports the generalizability of the negative results from this
study. In addition the overall rate of aseptic meningitis within 30 days after receipt of
vaccine of one case per 100,000 doses contrasts sharply with the reported risk of aseptic
meningitis after Urabe strain of one case per 3,800 doses.3
This study was performed utilizing the cooperative framework of the Vaccine Safety
Datalink Project of the CDC and the automated data resources of four large HMOs. The
initial review of automated data tapes from the project with codes specific for aseptic
meningitis had revealed a possible association of aseptic meningitis with MMR. This
initial review was based primarily on hospital cases although outpatient cases were
included from one site. To assure complete case ascertainment for the current analysis,
the set of ICD-9 codes reviewed was expanded to include all the codes likely to be used
for coding of an aseptic meningitis case. Chart review was then utilized to identify
validated cases of aseptic meningitis from the pool of possible cases. This methodology
allowed maximal sensitivity while maintaining specificity through the chart review
process. Within the framework of the VSD project, analyses such as the one reported here
are planned when the automated data for specific codes indicate a possible association or
when such an association is suggested by other reports. One strength of the VSD automated
databases is that they allow rapid identification of potential cases using specific codes
for screening analysis and identification of a broader pool using a wider range of codes
when analysis of chart review validated cases is warranted. However, as is demonstrated
here all possible associations must be confirmed, and any definitive conclusions based on
systemic medical record review of the potential cases identified.
In this review of hospitalized cases of aseptic meningitis, no increased risk of
aseptic meningitis after Jeryl-Lynn strain MMR vaccine was found among an estimated total
cohort of 300,000 children 12 to 23 months of age who received MMR vaccine. Although the
results of this study are reassuring, further analyses incorporating cases identified
through prospective review of automated laboratory data, automated outpatient records and
hospitalized records are planned.
TABLE ONE*: Classification of 93 possible cases of
aseptic meningitis identified from automated data
Diagnostic Group |
Number of Cases |
Aseptic meningitis |
59 |
Febrile seizure |
6 |
Bacterial meningitis |
5 |
Encephalopathy / Encephalitis |
5 |
Partially treated bacterial meningitis |
4 |
Viral syndrome with total CSF white cell count < 10 |
4 |
Cerebellar ataxia |
2 |
Septic shock and anoxic encephalopathy |
1 |
Brain abscess |
1 |
Possible aseptic meningitis ( no lumbar puncture done) |
1 |
Pneumococcal bacteremia |
1 |
Other1 |
4 |
1 Includes one case each of dermoid cyst, congenital
myopathy, congenital brain stem cyst, and tuberous sclerosis.
TABLE TWO*: Vaccine combinations within 14 and 30 days
before onset of illness for the 59 cases and the corresponding index date in the 118
matched controls.
Vaccine
Combination |
Within 14 days--
# Cases |
Within 14 days--
# Controls |
Within
30 days--
# Cases |
Within
30 days--
# Controls |
MMR Alone |
0 |
1 |
2 |
3 |
MMR with Hib |
0 |
1 |
0 |
2 |
MMR with DPT, Hib, & OPV |
0 |
2 |
0 |
2 |
MMR with HDPT, OPV, HepB |
1 |
0 |
1 |
0 |
TOTAL MMR |
1 |
4 |
3 |
7 |
Other combinations without MMR |
1 |
4 |
4 |
11 |
TOTAL all vaccines combinations |
2 |
8 |
7 |
18 |
TABLE THREE*: Risk of aseptic meningitis in relation to recent exposure to
MMR or any vaccine: matched case-control analysis in 59 cases and 118 controls.
Vaccine & Time Window
|
Cases
|
Controls
|
Odds Ratio |
95% C.I.
|
P-Value |
MMR within 14 days |
1 |
4 |
0.50 |
0.1 - 4.5 |
0.54 |
MMR within 30 days |
3 |
7 |
0.84 |
0.2 - 3.5 |
0.81 |
MMR within 8-14 days |
1 |
2 |
1.00 |
0.1 - 9.2 |
0.99 |
Any vaccine within 14 days |
2 |
8 |
0.44 |
0.1 - 2.1 |
0.30 |
Any vaccine within 30 days |
7 |
18 |
0.75 |
0.3 - 1.9 |
0.55 |
Any vaccine within 8-14 days |
2 |
4 |
1.00 |
0.2 - 5.6 |
0.99 |
References:
- Cherry JD, Shielfs WD. Encephalitis and
meningoencephalitis.
In Feigin RD, Cherry JD eds. Textbook of pediatric infectious diseases. Philadelphia:
Saunders, 1987:484-96.
- Centers for Disease Control. Mumps: United States 1980-1993. MMWR
1995, 44: 55.
-
Colville A, Pugh S. Mumps meningitis and measles,
mumps, rubella vaccine. Lancet 1992, 340:786.
-
Farrington, P. Pugh S, Colville A, et al. A new
method for active surveillance of adverse events from Diphtheria/tetanus/pertussis and
measles/mumps/rubella vaccines. Lancet 1995, 345:567.
-
Sugiura A, Lyamada A. Aseptic meningitis as a
complication of mumps vaccination. Pediatr Inf Dis J 1991, 10:209-13.
-
Fujinaga T, Motegi Y, Tamura H, Kuroume T. A
prefecture-wide survey of mumps meningitis associated with measles, mumps, rubella
vaccine. Pediatr Infect Dis J 1991, 10:204-9.
-
Miller E, Goldacre M, Pugh S, et al. Risk of
aseptic meningitis after measles, mumps, rubella vaccine in UK Children. Lancet 1993,
341:979.
-
Wassilak SG, Glasser JW, Chen
RT, Hadler SC et al.
Utility of Large Linked Databases in Vaccine Safety, particularly in distinguishing
independent and synergistic effects. Ann NY Acad Sci 1995; 754:377-82.
-
Schlesselman J. Case Control Studies: Design,
Conduct, and Analyses. Oxford University Press. New York, NY, 1982, 227-290.
-
SAS Release 6.09; SAS Institute, Inc.; Cary, N.C.
-
Institute of Medicine, Adverse Events Associated with
Childhood Vaccines. (Washington, D.C.:National Academy Press, 1994),130.
|