Abstract
Background
Case reports have suggested that vaccines may trigger transverse myelitis (TM) or acute disseminated encephalomyelitis (ADEM), but the evidence for a causal association is inconclusive. We analyzed the association of immunization and subsequent development of TM or ADEM.
Methods
We identified all cases of TM and ADEM in the Vaccine Safety Datalink population. Using a case-centered method, we compared vaccination of each case to vaccination of all matched persons in the study population, who received the same type of vaccine, with respect to whether or not their vaccination occurred during a predetermined exposure interval. We calculated a risk difference (excess risk) of TM and ADEM for each vaccine.
Results
Following nearly 64 million vaccine doses, only 7 cases of TM and 8 cases of ADEM were vaccinated during the primary exposure window 5–28 days prior to onset. For TM, there was no statistically significant increased risk of immunization. For ADEM, there was no statistically significant increased risk following any vaccine except for Tdap (adolescent and adult tetanus, reduced diphtheria, acellular pertussis) vaccine. Based on 2 exposed cases, the odds ratio for Tdap exposure 5–28 days prior to ADEM onset was 15.8 (95% confidence interval [CI], 1.2–471.6; P = .04), and the estimated excess risk was 0.385 (95% CI, −.04 to 1.16) cases per million doses.
Conclusions
We found no association between TM and prior immunization. There was a possible association of ADEM with Tdap vaccine, but the excess risk is not likely to be more than 1.16 cases of ADEM per million vaccines administered.
Keywords: vaccine, immunization, transverse myelitis, acute disseminated encephalomyelitis
Concerns that vaccines might trigger autoimmune demyelination have existed for many years [1–3]. Most publications on the subject are anecdotal case reports [4–7], with limited scientific evidence to support these concerns, and multiple studies have found no association [8–15]. The Institute of Medicine has judged the evidence to be inconclusive with respect to a causal association between specific acute demyelinating events (ADEs) and any vaccine [16]. However, many parents, patients, and providers consider the possibility of an association between vaccines and ADEs to be both plausible and worrisome [3, 17].
Studies of possible associations of ADEs with vaccinations have been problematic for multiple reasons [1, 2]; primarily, ADEs are rare and very large populations are needed to study them with adequate statistical power. Furthermore, for many ADEs, incidence rates peak during early and middle adulthood, times of life when fewer people are receiving vaccines of any kind, compared with young children and older adults. Evaluation for associations between previous immunizations and disease onset is therefore difficult because long periods of time have often elapsed since vaccination. In addition, it may not be possible to identify an unvaccinated comparison group similar to the vaccinees; for observational studies, many differences between vaccinated and unvaccinated groups may not be measured and thus cannot be controlled in an analysis [18, 19]. A significant challenge for earlier studies of ADEs in general is that there is limited understanding about the underlying pathophysiology or causes of these diseases. For example, seasonality may be an important confounder in studies of vaccine adverse events, particularly if the adverse event might also be precipitated by an infection, such as influenza.
Case-series, risk-interval, or case-centered analyses may be useful for addressing some of the challenges in evaluating possible associations between vaccinations and ADEs. Adverse events that are most amenable to study using these methodologies are events that are relatively abrupt in onset, clearly defined, occur relatively soon after vaccination, have a limited period of risk (termed the “risk interval” [20]), and are serious enough that people seek medical care for them. If an illness advances slowly over a long period of time prior to diagnosis, or has a long latency period, determining the time of actual symptom onset may be very challenging. This clinical complexity may make any determination of a possible relationship between disease development and timing of vaccination very difficult. In this study, we analyzed the association of immunization and subsequent development of TM and ADEM, 2 diseases which have a relatively acute onset and for which patients seek medical care in a timely fashion.
Acute transverse myelitis (TM) is a rare acquired demyelinating disorder that presents with sudden onset of neurologic deficits due to spinal cord lesions. Annual incidence in the general population is thought to be between 1 and 8 per million [7, 21, 22], with most studies showing a significant proportion with onset of symptoms preceded within a few weeks by an infectious disease.
Like TM, acute disseminated encephalomyelitis (ADEM) is a rare disease, often preceded by a respiratory or gastrointestinal illness. It has also been reported after a number of immunizations, mostly as case reports unable to support causality [23, 24]. Patients with ADEM present with fever and headache, and then progress rapidly to manifest symptoms due to brain or spinal cord inflammation, such as obtundation or motor deficits.
METHODS
The study used a case-centered design, which compares vaccination patterns during an exposure interval prior to the outcome in cases vs the entire study population, matched by age, sex, and site.
Study Population
The Vaccine Safety Datalink (VSD) is a collaboration between the Centers for Disease Control and Prevention’s (CDC) Immunization Safety Office and several integrated healthcare systems across the United States. These health systems provide essentially all healthcare services and capture data from outpatient department visits, emergency department visits, and hospitalizations. The VSD has data on >9 million subjects annually, including 2.1 million children and 7.2 million adults. Participating VSD sites maintain automated databases of healthcare encounters, including immunization registries. All sites have the capability to access medical records and other data sources to provide detailed information on specific healthcare encounters. Our study population included all children and adults of any age enrolled in the health plans at 6 VSD sites who received 1 or more vaccines during the period from 1 January 2007 through 31 December 2012.
Case Selection
We utilized each VSD site’s internal electronic medical record (EMR) diagnostic text codes to identify first-ever TM and ADEM diagnoses in any setting (hospital, emergency, or outpatient clinic), and required at least 1 diagnosis by a neurologist within 3 months of the initial diagnosis. These text codes are more specific than International Classification of Diseases, Ninth Revision codes, and are what the medical provider sees and chooses when making a diagnosis. We used case selection criteria for acute idiopathic TM and ADEM and excluded cases due to other causes, such as compression, multiple sclerosis, or neuromyelitis optica. We excluded individuals with either a prior or subsequent history of multiple sclerosis. Trained medical records analysts (MRAs) reviewed all identified cases to verify that a neurologist made the diagnosis within 3 months of the first electronic diagnosis, to validate the date of onset, to ensure the diagnosis did not change over time, and to abstract pertinent information. Following this, a neurologist with expertise in demyelinating illnesses reviewed each case. For ADEM, cases were confirmed using the Brighton Collaboration Criteria [25],and we accepted diagnostic levels of certainty 1, 2, and 3 as cases. For TM we used the Transverse Myelitis Consortium Working Group criteria [26], because Brighton Criteria were not available. The date of onset used for analysis was that determined by medical record review.
Vaccines
We evaluated all administered vaccines, both individually and combined. To increase power for the study, we combined all inactivated influenza vaccines together, including trivalent, quadrivalent, and high dose. We analyzed meningococcal conjugate vaccines (both Menactra and Menveo), and pneumococcal conjugate vaccines (7- and 13-valent) both separately and combined. We also looked at any vaccine (all vaccines combined).
Exposure Intervals
On the basis of prior studies and expert opinion, we used 2 exposure intervals: (1) 5–28 days as the most likely interval following immunization to result in a demyelination illness and (2) 2–42 days as reassurance that we are not missing an increased risk with the same type of vaccine, beyond the shorter 5- to 28- day exposure interval. All persons (cases and comparisons) in the study were immunized with at least 1 vaccine within 9 months prior to the cases’ onsets. We used the period after the longer exposure interval, 43 days through 9 months, as the comparison interval. Nine months was chosen as the comparison interval to avoid duplicate influenza vaccines over 2 seasons.
Statistical Methods
We used a case-centered method, as has been previously described [18, 27–29]. We identified all cases in the vaccinated study population. We compared each case to all similar (by age, sex, and VSD site) vaccinees who were in the study population on the diagnosis date of the case. We compared the cases to the general population, with respect to whether or not their vaccination occurred during the exposure interval. Thus, the method is equivalent to a matched case-control study, except that it utilizes all available matched controls who received the same vaccine type, rather than just a sample of them. Risk sets consisting of cases and all matched controls were constructed at each site from presummarized immunization follow-up data on the entire VSD population, and matching was done by age and sex. All cases and comparison matches were vaccinated with the same vaccine in the 9 months prior to the onset of the case. We determined the proportion of each risk set that was vaccinated within the exposure interval prior to the date of illness onset for each matched TM or ADEM case. We examined whether or not each case was vaccinated in the exposure interval or the comparison interval prior to the date of illness onset in relation to the proportion of the entire risk set that was vaccinated in the exposure interval prior to the case onset date. We used a stratified exact binomial method to calculate the odds ratios (ORs) with confidence intervals (CIs) and P values. The OR directly estimates the relative risk of being diagnosed with ADEM or TM in the exposure interval compared to the rest of the 9 months after vaccination.
Risk Difference Calculation
To take advantage of the large number of immunizations given over the study period, when only relatively small numbers of TM and ADEM cases had their onset within even a year of a vaccine, we used a Mantel–Haenszel–type weighted average [30] to estimate the excess risk (also known as the risk difference) and 95% CIs [31] associated with immunization and TM and ADEM. Wald-type P values were calculated. The risk difference amounts to the risk per day of the outcome diagnosis during the exposure interval minus the risk per day during the rest of the 9 months after vaccination. The results are scaled to the risk per dose of each vaccine. Unlike the OR, which is a relative measure where 1.0 indicates no association, the attributable risk is a difference measure in which zero indicates lack of association.
The study protocol was approved by institutional review boards at the CDC and each VSD site, and it was determined that informed consent was not required.
RESULTS
Over the study period, nearly 64 million vaccine doses were recorded and used in the analyses (Table 1).
Table 1.
Vaccines Given at Vaccine Safety Datalink Sites During the Study Period, 2007–2012
| Vaccine Type | Vaccine Count |
|---|---|
| Trivalent and quadrivalent inactivated influenza (IIV) | 18 926 060 |
| Tetanus, reduced diphtheria, acellular pertussis (Tdap) | 5 940 485 |
| Any pneumococcal conjugate (7- or 13-valent) (PCV) | 3 611 754 |
| Hepatitis A (HAV) | 3 459 116 |
| Haemophilus influenzae type b (Hib) | 2 911 711 |
| Varicella | 2 427 803 |
| 7-valent pneumococcal conjugate (PCV7) | 2 272 002 |
| Hepatitis B (HBV) | 2 147 046 |
| Quadrivalent human papillomavirus (HPV4) (Gardisil) | 2 131 486 |
| Diphtheria, tetanus, acellular pertussis (DTaP) | 1 955 611 |
| Monovalent H1N1 influenza | 1 773 442 |
| DTaP + inactivated polio (IPV) + HBV (Pediarix) | 1 712 276 |
| Any quadrivalent meningococcal conjugate (MCV4) (Menactra or Menveo) | 1 581 856 |
| Measles, mumps, rubella (MMR) | 1 576 861 |
| Quadrivalent meningococcal conjugate (MCV4) (Sanofi-Menactra) | 1 525 239 |
| 23-valent pneumococcal polysaccharide (PPSV23) | 1 509 189 |
| Rotavirus 5 | 1 453 465 |
| IPV | 1 321 888 |
| PCV13 | 1 318 310 |
| Live attenuated influenza vaccine (LAIV) | 905 363 |
| Tetanus, diphtheria (Td) | 806 054 |
| Live zoster vaccine | 670 492 |
| MMR + Varicella | 431 991 |
| Injectable typhoid | 317 895 |
| DTaP + IPV + Hib (Pentacel) | 302 904 |
| DTaP + IPV (Kinrix) | 246 986 |
| Rotavirus 1 | 186 275 |
| HAV + HBV (Twinrix) | 122 231 |
| Yellow fever | 81 314 |
| MCV4 Novartis (Menveo) | 44 317 |
| Quadrivalent meningococcal polysaccharide (MPSV4) | 5228 |
| Rabies | 1846 |
| Total number of vaccines given | 63 678 496 |
Over all VSD sites, we identified 545 potential cases of TM in the EMRs. One hundred eighty-four of these were rejected by the MRA (132 had an old history of TM, 24 were ruled out, and 28 were miscoded). The remaining 361 cases were reviewed by a neurologist. One hundred ninety-three did not meet inclusion criteria, and 87 had alternative diagnoses. Eighty-one cases (14.9%) were accepted as new, acute-onset idiopathic TM using the TM Consortium Working Group definition. Of these 81 cases, 67 had received an immunization within 9 months prior to onset, and were included in the analysis. One case was vaccinated 29–42 days prior to TM diagnosis, and this case was included in the 2- to 42-day analysis, but not the 5- to 28-day analysis.
Of 118 cases of ADEM diagnosed in the EMR, 29 were rejected by the MRA (13 had a history of ADEM, 8 were ruled out, and 8 were miscoded). Of the remaining 89 cases reviewed by the neurologist, 17 were indeterminate and 16 were other diagnoses, so that 56 (47.5%) were accepted both by the MRA and by the neurologist as an acute new diagnosis (54 Brighton level 1, 1 level 2, and 1 level 3). Forty-seven of these had received an immunization within 9 months prior to the onset of ADEM.
Tables 2 and 3 show the results of the case-centered analyses results for TM and ADEM using the 5- to 28-day exposure interval. Results for the 2- to 42-day exposure interval are presented in Supplementary Tables 1 and 2. For TM, there was no statistically significant increased risk of immunization in either the 5- to 28-day or the 2- to 42-day risk interval prior to onset.
Table 2.
Relative Riska of Transverse Myelitis in the 5- to 28-Day Risk Interval Following Vaccines, Compared to Remainder of the 9 Months After Vaccination—Vaccine Safety Datalink, 2007–2013
| Vaccinated Cases |
Comparison Risk Sets: Cases and Noncasesb |
||||||
|---|---|---|---|---|---|---|---|
| Vaccine Type | No. | % in Exposure Interval | No. | % in Exposure Interval | Adjusted Odds Ratio | (95% CI) | P Value |
| MMR | 1 | 0 | 188 | 12.8 | 0 | (.0–129.8) | .87 |
| DTaP/IPV/HBV | 1 | 0 | 150 | 22.7 | 0 | (.0–64.8) | .77 |
| MCV4 (Sanofi) | 3 | 33.3 | 10 221 | 7.2 | 5.2 | (.2–70.9) | .27 |
| Tdap | 16 | 6.3 | 121 768 | 8.8 | 0.7 | (.0–3.8) | .78 |
| Rotavirus 5 | 1 | 0 | 151 | 20.5 | 0 | (.0–73.5) | .79 |
| Zoster | 2 | 50 | 6079 | 11.1 | 7.7 | (.2–301.4) | .23 |
| PCV13 | 1 | 0 | 154 | 22.7 | 0 | (.0–64.6) | .77 |
| Varicella | 4 | 0 | 4725 | 7.5 | 0 | (.0–10.7) | .67 |
| PPSV23 | 5 | 0 | 27 940 | 8.1 | 0 | (.0–9.4) | .66 |
| HPV4 | 3 | 0 | 7273 | 9.7 | 0 | (.0–15.3) | .73 |
| Any MCV4 | 3 | 33.3 | 10 221 | 7.2 | 5.2 | (.2–70.9) | .27 |
| Any PCV | 2 | 0 | 218 | 21.6 | 0 | (.0–12.0) | .6 |
| HAV | 4 | 0 | 3057 | 9 | 0 | (.0–12.6) | .71 |
| HBV | 2 | 0 | 284 | 12 | 0 | (.0–30.0) | .79 |
| Hib | 1 | 0 | 157 | 22.3 | 0 | (.0–66.2) | .78 |
| IIV | 38 | 7.9 | 969 316 | 6.2 | 1 | (.2–4.0) | .98 |
| LAIV | 1 | 100 | 2640 | 29.7 | NE | (.1 to . . .) | .3 |
| Td | 2 | 0 | 3419 | 6.8 | 0 | (.0–34.1) | .82 |
| Any | 66 | 10.6 | 1 692 825 | 10.3 | 1.1 | (.4–2.4) | .87 |
Abbreviations: CI, confidence interval; DTaP, diphtheria, tetanus, acellular pertussis vaccine; HAV, hepatitis A vaccine; HBV, hepatitis B vaccine; Hib, Haemophilus influenzae type b vaccine; HPV4, quadrivalent human papillomavirus vaccine; IIV, inactivated influenza vaccine; IPV, inactivated polio vaccine; LAIV, live attenuated influenza vaccine; MCV4, quadrivalent meningococcal conjugate vaccine; MMR, measles, mumps, rubella vaccine; NE, not evaluable; PCV13, 13-valent pneumococcal conjugate vaccine; PPSV23, 23-valent pneumococcal polysaccharide vaccine; Td, tetanus and diphtheria vaccine; Tdap, adolescent and adult tetanus, reduced diphtheria, acellular pertussis vaccine.
This relative risk is estimated by the odds ratio, using the case-centered method, as described in the “Methods” section.
Comparative risk sets: cases and noncases: All Vaccine Safety Datalink (VSD) members similar to cases (matched on age, sex, vaccine type, and VSD site) on the onset date of the case. No. represents the number of people in all the risk sets. If there were no cases within 9 months of a vaccine, we could not perform a case-centered analysis, so these are not included in the table.
Table 3.
Relative Riska of Acute Disseminated Encephalomyelitis in the 5- to 28-Day Risk Interval Following Vaccines, Compared to Remainder of the 9 Months After Vaccination—Vaccine Safety Datalink, 2007–2013
| Cases |
Comparison Risk Sets: Informative Cases and Noncasesb |
||||||
|---|---|---|---|---|---|---|---|
| Vaccine Type | No. | % in Exposure Interval | No. | % in Exposure Interval | Adjusted Odds Ratio | (95% CI) | P Value |
| MMR | 6 | 33.3 | 46 482 | 10.1 | 5 | (.6–29.9) | .12 |
| IPV | 1 | 0 | 6025 | 6.6 | 0 | (.0–270.1) | .93 |
| PCV7 | 3 | 0 | 39 788 | 4.5 | 0 | (.0–48.1) | .87 |
| MCV4 (Sanofi) | 3 | 0 | 14 731 | 9 | 0 | (.0–18.0) | .76 |
| Tdap | 3 | 66.7 | 9972 | 11.5 | 15.8 | (1.2–471.6) | .04 |
| DTaP/IP/Hib | 2 | 0 | 5387 | 7.9 | 0 | (.0–40.7) | .85 |
| Zoster | 1 | 0 | 3226 | 9.9 | 0 | (.0–173.1) | .9 |
| DTaP/IP | 4 | 25 | 29 751 | 8.9 | 4.1 | (.1–44.0) | .32 |
| PCV13 | 3 | 33.3 | 7047 | 22.8 | 3.6 | (.1–95.4) | .49 |
| DTaP | 1 | 0 | 6153 | 6.6 | 0 | (.0–270.4) | .93 |
| Varicella | 6 | 33.3 | 44 953 | 10.5 | 4.3 | (.5–25.4) | .14 |
| MPSV4 | 1 | 100 | 3 | 33.3 | NE | (.1 to . . .) | .33 |
| PPSV23 | 3 | 0 | 13 359 | 10.3 | 0 | (.0–14.5) | .71 |
| HPV4 | 6 | 16.7 | 56 846 | 12.1 | 1.5 | (.1–10.7) | .7 |
| MMR-Varicella | 1 | 0 | 2046 | 6.1 | 0 | (.0–292.0) | .94 |
| Any MCV4 | 3 | 0 | 16 157 | 9.1 | 0 | (.0–17.5) | .75 |
| Any PCV | 6 | 16.7 | 46 862 | 7.3 | 2.8 | (.1–21.2) | .4 |
| H1N1 | 6 | 0 | 56 953 | 1.3 | 0 | (.0–102.7) | .95 |
| HAV | 8 | 12.5 | 60 878 | 7.5 | 1.9 | (.1–13.0) | .54 |
| HBV | 2 | 0 | 2112 | 2.4 | 0 | (.0–128.7) | .95 |
| Hib | 3 | 0 | 18 285 | 6.8 | 0 | (.0–25.3) | .8 |
| IIV | 21 | 19 | 281 468 | 7 | 1.5 | (.4–5.0) | .53 |
| LAIV | 4 | 0 | 2650 | 12.6 | 0 | (.0–4.9) | .34 |
| Td | 1 | 0 | 155 | 1.9 | 0 | (.0–962.7) | .98 |
| Any | 47 | 17 | 813 498 | 8.8 | 1.7 | (.7–3.8) | .19 |
Abbreviations: CI, confidence interval; DTaP, diphtheria, tetanus, acellular pertussis vaccine; DTaP/IP, diphtheria, tetanus, acellular pertussis, inactivated polio; DTaP/IP/Hib, diphtheria, tetanus, acellular pertussis, inactivated polio, Haemophilus influenza type B; HAV, hepatitis A vaccine; HBV, hepatitis B vaccine; Hib, Haemophilus influenzae type b vaccine; HPV4, quadrivalent human papillomavirus vaccine; IIV, inactivated influenza vaccine; IPV, inactivated polio vaccine; LAIV, live attenuated influenza vaccine; MCV4, quadrivalent meningococcal conjugate vaccine; MMR, measles, mumps, rubella vaccine; MPSV4, quadrivalent meningococcal polysaccharide vaccine; NE, not evaluable; PCV7, 7-valent pneumococcal conjugate vaccine; PCV13, 13-valent pneumococcal conjugate vaccine; PPSV23, 23-valent pneumococcal polysaccharide vaccine; Td, tetanus and diphtheria vaccine; Tdap, adolescent and adult tetanus, reduced diphtheria, acellular pertussis vaccine.
This relative risk is estimated by the odds ratio, using the case-centered method, as described in the “Methods” section.
Comparison risk sets: cases and noncases: All Vaccine Safety Datalink (VSD) members similar to cases (matched on age, sex, vaccine type, and VSD site) on the onset date of the case. No. represents the number of people in all the risk sets. If there were no cases within 9 months of a vaccine, we could not perform a case-centered analysis, so these are not included in the table.
For ADEM, the Tdap (adolescent and adult tetanus, reduced diphtheria, acellular pertussis) vaccine was associated with a statistically significant increase in risk in the 5- to 28-day exposure interval (OR, 15.8 [95% CI, 1.2–471.6]; P = .04), but not the longer 2- to 42-day interval.
We calculated the excess risk (risk difference) for each vaccine and outcome in the 5- to 28-day exposure interval (Tables 4 and 5). The tables show the calculated attributable risk per 1 million doses of each vaccine. The excess risk for Tdap was calculated to be 0.385 (95% CI, −.04 to 1.16).
Table 4.
Mantel–Haenszel Estimate of Risk Difference (Excess Risk) of Transverse Myelitis in the 5–28 Days After Immunization, per 1 Million Doses of Each Vaccine—Vaccine Safety Datalink, 2007–2013
| Vaccine | Risk Difference per Million Doses | (95% CI) | P Value | Total Risk Cases |
|---|---|---|---|---|
| MMR | −0.161 | (−.21 to 1.76) | .693 | 0 |
| DTaP/HBV/IP | −0.393 | (−.97 to 1.19) | .750 | 0 |
| MCV4 (Sanofi) | 0.709 | (−.41 to 3.25) | .152 | 1 |
| Tdap | −0.107 | (−.41 to .55) | .667 | 1 |
| Rotavirus 5 | −0.459 | (−1.33 to 1.33) | .739 | 0 |
| Zoster | 1.446 | (−.72 to 6.39) | .137 | 1 |
| PCV13 | −0.356 | (−.61 to 1.54) | .750 | 0 |
| Varicella | −0.268 | (−.34 to .85) | .810 | 0 |
| PPSV23 | −0.369 | (−.46 to 1.12) | .815 | 0 |
| HPV4 | −0.210 | (−.29 to .81) | .782 | 0 |
| Any MCV4 | 0.682 | (−.39 to 3.11) | .152 | 1 |
| Any PCV | −0.364 | (−.61 to .66) | .836 | 0 |
| HepA | −0.173 | (−.22 to .66) | .793 | 0 |
| HepB | −0.334 | (−.48 to 1.81) | .745 | 0 |
| Hib | −0.218 | (−.38 to .94) | .748 | 0 |
| IIV | −0.004 | (−.39 to .49) | .507 | 3 |
| LAIV | 1.725 | (−.87 to 5.91) | .118 | 1 |
| Td | −0.542 | (−.66 to 3.63) | .741 | 0 |
| Any | 0.014 | (−.20 to .28) | .454 | 7 |
Comparison is within 9 months of the same vaccine. This analysis requires at least 1 case in the exposure or comparison interval.
Abbreviations: CI, confidence interval; DTaP/HBV/IP, diphtheria, tetanus, acellular pertussis, hepatitis B, inactivated polio; HAV, hepatitis A vaccine; HBV, hepatitis B vaccine; Hib, Haemophilus influenzae type b vaccine; HPV4, quadrivalent human papillomavirus vaccine; IIV, inactivated influenza vaccine; LAIV, live attenuated influenza vaccine; MCV4, quadrivalent meningococcal conjugate vaccine; MMR, measles, mumps, rubella vaccine; PCV13, 13-valent pneumococcal conjugate vaccine; PPSV23, 23-valent pneumococcal polysaccharide vaccine; Td, tetanus and diphtheria vaccine; Tdap, adolescent and adult tetanus, reduced diphtheria, acellular pertussis vaccine.
Table 5.
Mantel–Haenszel Estimate of Risk Difference (Excess Risk) of Acute Disseminated Encephalomyelitis in the 5–28 Days After Immunization, per 1 Million Doses of Each Vaccine—Vaccine Safety Datalink, 2007–2013
| Vaccine | Risk Difference per Million Doses | (95% CI) | P Value | Total Risk Cases |
|---|---|---|---|---|
| MMR | 1.807 | (−.55 to 6.05) | .084 | 2 |
| IPV | −0.148 | (−.19 to 3.23) | .642 | 0 |
| PCV7 | −0.201 | (−.37 to 1.47) | .698 | 0 |
| MCV4 (Sanofi) | −0.255 | (−.33 to 1.24) | .766 | 0 |
| Tdap | 0.385 | (−.04 to 1.16) | .042 | 2 |
| DTaP/IPV/Hib | −1.751 | (−3.72 to 9.26) | .712 | 0 |
| Zoster | −0.185 | (−.24 to 2.65) | .672 | 0 |
| DTaP/IP | 3.616 | (−2.45 to 17.12) | .17 | 1 |
| PCV13 | 0.797 | (−.69 to 3.93) | .196 | 1 |
| DTaP | −0.079 | (−.10 to 1.75) | .641 | 0 |
| Varicella | 0.95 | (−.35 to 3.30) | .097 | 2 |
| MPSV4 | 470.628 | (−214.74 to 1940.23) | .124 | 1 |
| PPSV23 | −0.295 | (−.37 to 1.17) | .789 | 0 |
| HPV4 | 0.189 | (−.68 to 2.00) | .377 | 1 |
| MMR-Varicella | −0.226 | (−.31 to 5.10) | .637 | 0 |
| Any MCV4 | −0.245 | (−.32 to 1.19) | .766 | 0 |
| Any PCV | 0.285 | (−.47 to 1.83) | .282 | 1 |
| H1N1 | −0.071 | (−.16 to 1.11) | .627 | 0 |
| HAV | 0.233 | (−.38 to 1.64) | .289 | 1 |
| HBV | −0.081 | (−.12 to 1.96) | .627 | 0 |
| Hib | −0.207 | (−.37 to 0.95) | .742 | 0 |
| IIV | 0.137 | (−.26 to .63) | .263 | 4 |
| LAIV | −2.018 | (−3.86 to 1.41) | 0.9 | 0 |
| Any | 0.141 | (−.07 to .41) | .107 | 8 |
Comparison is within 9 months of same vaccine. This analysis requires at least 1 case in the exposure or comparison interval.
Abbreviations: CI, confidence interval; DTaP, diphtheria, tetanus, acellular pertussis vaccine; DTaP/IP, diphtheria, tetanus, acellular pertussis, inactivated polio; DTaP/IP/Hib, diphtheria, tetanus, acellular pertussis, inactivated polio, Haemophilus influenza type B; HAV, hepatitis A vaccine; HBV, hepatitis B vaccine; Hib, Haemophilus influenzae type b vaccine; HPV4, quadrivalent human papillomavirus vaccine; IIV, inactivated influenza vaccine; IPV, inactivated polio vaccine; LAIV, live attenuated influenza vaccine; MCV4, quadrivalent meningococcal conjugate vaccine; MMR, measles, mumps, rubella vaccine; MPSV4, quadrivalent meningococcal polysaccharide vaccine; PCV7, 7-valent pneumococcal conjugate vaccine; PCV13, 13-valent pneumococcal conjugate vaccine; PPSV23, 23-valent pneumococcal polysaccharide vaccine; Tdap, adolescent and adult tetanus, reduced diphtheria, acellular pertussis vaccine.
DISCUSSION
Our study examined the risk of TM and ADEM following vaccines of any kind. Our case-centered method used only cases and comparison individuals who had previously been vaccinated. This controlled for many of the differences between vaccinated and unvaccinated individuals. In addition, this method, by anchoring each case and comparison to a specific calendar date, controls exceedingly well for time-varying confounding, such as seasonality.
For TM, we found no evidence of a safety concern, or any association with subsequent illness. If there is any association, it is <1 per million doses for vaccines other than live zoster and live attenuated influenza vaccines, and <2 per million doses of these 2 vaccines. For ADEM, we found a possible association with the Tdap vaccine in the 5- to 28-day risk interval (OR, 15.8 [95% CI, 1.2–471.6]; P = .04). However, there are some caveats to this finding. First, the number of exposed cases (2) was very small. Of the 2 cases, 1 was a healthy adult who was routinely vaccinated 11 days prior to symptoms. The second case was vaccinated with both the meningococcal polysaccharide vaccine (not recommended for his age group at that time) and Tdap. The case vaccinated in the comparison period was immunized routinely with Tdap, quadrivalent human papillomavirus vaccine (HPV4), quadrivalent meningococcal conjugate vaccine (MCV4), and hepatitis A vaccine at the same time. Statistically, we would not have been surprised by 1 exposed case, and a second exposed case is what gives us a statistically significant result. The paucity of cases leads to a wide CI for the OR, the lower bound of which is close to 1.0. In addition, we performed a great number of statistical comparisons, and we did not adjust for multiple testing; therefore, this result could be due to chance alone. However, if we accept this finding at face value that Tdap does increase the risk of ADEM, we estimate that the attributable risk of ADEM with Tdap vaccination is only 0.4 cases per 1 million doses of vaccine. Because so many vaccines were given, the CI for the risk difference estimate is narrow, and from the upper bound of the 95% CI, we can be confident that the excess risk is not likely to exceed 1.16 cases of ADEM per million doses of Tdap. For other vaccines, the excess risk is even smaller or nonexistent.
These results are similar to what we have found previously in assessment of the association between immunization and other acute demyelinating diseases [29, 32] and are very reassuring.
Strengths and limitations of the study are as follows: Review of all TM and ADEM cases minimized misclassification. We did not analyze combinations of vaccines. The method depends on selection of an appropriate risk (exposure) interval, so if the interval is misspecified, an increased risk could have been missed. We purposely did not adjust for multiple observations in order to increase the sensitivity of our results; nonetheless, we identified only 1 statistically significant association.
In conclusion, TM and ADEM are rarely, if ever, associated with vaccines. If there is an association of ADEM with Tdap, the excess risk is not likely to exceed 1.16 cases of ADEM per million doses of Tdap administered.
Supplementary Material
Acknowledgments
Financial support. This work was supported by a task order from the CDC (contract number 200-2012-53581/0006).
Potential conflicts of interest. R. B. and N. P. K. report research grants for unrelated studies from GSK, Sanofi Pasteur, Merck, Pfizer, Protein Sciences, and MedImmune. A. L. N. Reports research grants for unrelated studies from Merck, Pfizer, and MedImmune. F. D. and J. G. are employees of the CDC. All other authors report no potential conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.
Footnotes
Disclaimer. The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention (CDC).
Supplementary Data
Supplementary materials are available at http://cid.oxfordjournals.org. Consisting of data provided by the author to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the author, so questions or comments should be addressed to the author.
References
- 1.Chen RT, Pless R, Destefano F. Epidemiology of autoimmune reactions induced byvaccination. J Autoimmun 2001; 16:309–18. [DOI] [PubMed] [Google Scholar]
- 2.Agmon-Levin N, Paz Z, Israeli E, Shoenfeld Y. Vaccines and autoimmunity. NatRev Rheumatol 2009; 5:648–52. [DOI] [PubMed] [Google Scholar]
- 3.Schattner A Consequence or coincidence? The occurrence, pathogenesis and significance of autoimmune manifestations after viral vaccines. Vaccine 2005; 23:3876–86. [DOI] [PubMed] [Google Scholar]
- 4.Kawasaki A, Purvin VA, Tang R. Bilateral anterior ischemic optic neuropathy following influenza vaccination. J Neuroophthalmol 1998; 18:56–9. [PubMed] [Google Scholar]
- 5.Payne DC, Rose CE Jr, Kerrison J, Aranas A, Duderstadt S, McNeil MM. Anthraxvaccination and risk of optic neuritis in the United States military, 1998–2003. Arch Neurol 2006; 63:871–5. [DOI] [PubMed] [Google Scholar]
- 6.Huynh W, Cordato DJ, Kehdi E, Masters LT, Dedousis C. Post-vaccination encephalomyelitis: literature review and illustrative case. J Clin Neurosci 2008; 15:1315–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Frohman EM, Wingerchuk DM. Clinical practice. Transverse myelitis. N Engl JMed 2010; 363:564–72. [DOI] [PubMed] [Google Scholar]
- 8.Vial T, Descotes J. Autoimmune diseases and vaccinations. Eur J Dermatol 2004; 14:86–90. [PubMed] [Google Scholar]
- 9.Toplak N, Kveder T, Trampus-Bakija A, Subelj V, Cucnik S, Avcin T. Autoimmuneresponse following annual influenza vaccination in 92 apparently healthy adults. Autoimmun Rev 2008; 8:134–8. [DOI] [PubMed] [Google Scholar]
- 10.Ram M, Shoenfeld Y. Hepatitis B: infection, vaccination and autoimmunity. IsrMed Assoc J 2008; 10:61–4. [PubMed] [Google Scholar]
- 11.Mikaeloff Y, Caridade G, Suissa S, Tardieu M. Hepatitis B vaccine and the risk ofCNS inflammatory demyelination in childhood. Neurology 2009; 72:873–80. [DOI] [PubMed] [Google Scholar]
- 12.DeStefano F, Verstraeten T, Chen RT. Hepatitis B vaccine and risk of multiple sclerosis. Expert Rev Vaccines 2002; 1:461–6. [DOI] [PubMed] [Google Scholar]
- 13.DeStefano F, Verstraeten T, Jackson LA, et al. Vaccinations and risk of central nervous system demyelinating diseases in adults. Arch Neurol 2003; 60:504–9. [DOI] [PubMed] [Google Scholar]
- 14.Chao C, Klein NP, Velicer CM, et al. Surveillance of autoimmune conditions following routine use of quadrivalent human papillomavirus vaccine. J Intern Med 2011; 271:193–203. [DOI] [PubMed] [Google Scholar]
- 15.Langer-Gould A, Qian L, Tartof SY, et al. Vaccines and the risk of multiple sclerosis and other central nervous system demyelinating diseases. JAMA Neurol 2014; 71:1506–13. [DOI] [PubMed] [Google Scholar]
- 16.Stratton K, Ford A, Rusch E, Clayton EW, eds; Institute of Medicine Adverse effects of vaccines: evidence and causality. Washington, DC: National Academies Press, 2012. [PubMed] [Google Scholar]
- 17.Koenig HC, Sutherland A, Izurieta HS, McGonagle D. Application of the immunological disease continuum to study autoimmune and other inflammatory events after vaccination. Vaccine 2011; 29:913–9. [DOI] [PubMed] [Google Scholar]
- 18.Fireman B, Lee J, Lewis N, Bembom O, van der Laan M, Baxter R. Influenza vaccination and mortality: differentiating vaccine effects from bias. Am J Epidemiol 2009; 170:650–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Baxter R, Lee J, Fireman B. Evidence of bias in studies of influenza vaccine effectiveness in elderly patients. J Infect Dis 2010; 201:186–9. [DOI] [PubMed] [Google Scholar]
- 20.Rowhani-Rahbar A, Klein NP, Baxter R, et al. Biologically plausible andevidence-based risk intervals in immunization safety research. Vaccine 2012; 31:271–7. [DOI] [PubMed] [Google Scholar]
- 21.Berman M, Feldman S, Alter M, Zilber N, Kahana E. Acute transverse myelitis: incidence and etiologic considerations. Neurology 1981; 31:966–71. [DOI] [PubMed] [Google Scholar]
- 22.Knebusch M, Strassburg HM, Reiners K. Acute transverse myelitis in childhood: nine cases and review of the literature. Dev Med Child Neurol 1998; 40:631–9. [DOI] [PubMed] [Google Scholar]
- 23.Fenichel GM. Neurological complications of immunization. Ann Neurol 1982; 12:119–28. [DOI] [PubMed] [Google Scholar]
- 24.Holt S, Hudgins D, Krishnan KR, Critchley EM. Diffuse myelitis associated withrubella vaccination. Br Med J 1976; 2:1037–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Sejvar JJ, Kohl KS, Bilynsky R, et al. Encephalitis, myelitis, and acutedisseminated encephalomyelitis (ADEM): case definitions and guidelines for collection, analysis, and presentation of immunization safety data. Vaccine 2007; 25:5771–92. [DOI] [PubMed] [Google Scholar]
- 26.Transverse Myelitis Consortium Working Group. Proposed diagnostic criteria andnosology of acute transverse myelitis. Neurology 2002; 59:499–505. [DOI] [PubMed] [Google Scholar]
- 27.Baker MA, Lieu TA, Li L, et al. A vaccine study design selection framework for thepostlicensure rapid immunization safety monitoring program. Am J Epidemiol 2015; 181:608–18. [DOI] [PubMed] [Google Scholar]
- 28.Baxter R, Lewis N, Bohrer P, Harrington T, Aukes L, Klein NP. Sudden onset sensorineural hearing loss after immunization—a case-centered analysis. Otolaryngol Head Neck Surg 2016; 155:81–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Baxter R, Lewis N, Fireman B, DeStefano F, Gee J, Klein NP. Case centered analysisof optic neuritis following vaccines. Clin Infect Dis 2016; 63:79–81. [DOI] [PubMed] [Google Scholar]
- 30.Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospectivestudies of disease. J Natl Cancer Inst 1959; 22:719–48. [PubMed] [Google Scholar]
- 31.Klingenberg B A new and improved confidence interval for the Mantel-Haenszel risk difference. Stat Med 2014; 33:2968–83. [DOI] [PubMed] [Google Scholar]
- 32.Baxter R, Bakshi N, Fireman B, et al. Lack of association of Guillain-Barre syndrome with vaccinations. Clin Infect Dis 2013; 57:197–204. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.