Abstract
Objectives: To determine whether there is an excess of respiratory tract infections in the 5-week, 3-month, and 12-month periods before MS symptom onset and if there is an association between MS and a history of infectious mononucleosis (IM).
Background: The etiology of MS remains unknown, but infection is frequently suggested as a putative etiologic agent. Epidemiologic studies have produced inconsistent evidence for an etiologic role of respiratory tract infections (RTI) and IM in MS.
Methods: The authors performed a case-control study using the General Practice Research Database from the United Kingdom. There were 225 subjects with definite or probable MS, and 900 controls matched for age, sex, and physician practice. Using computerized patient records, the authors compared the mean rates of RTI per patient in the 5-week, 3-month, and 12-month periods before the date of onset of the first symptoms compatible with MS (index date). They also compared histories of IM.
Results: In all periods, an increased frequency of RTI was associated with a significantly increased risk of MS. A history of IM was associated with greater than five times the risk of MS (OR = 5.5 [95% CI 1.5 to 19.7]).
Conclusions: These results support an association between a history of IM and subsequent MS. Respiratory tract infections may precipitate disease onset.
MS is a CNS demyelinating disease of unknown etiology. A multifactorial etiology is proposed in which multiple environmental factors act on a genetically susceptible individual to cause disease. Support for an environmental etiology derives from geographic and temporal variation in incidence and prevalence, and from the apparent effects of migration on disease risk.1
Infection is frequently touted as a putative etiologic agent. Animal models of virally mediated CNS demyelination exist, although the mechanisms are unknown.2 Infectious mononucleosis (IM) and other viral infections are known to be associated with neurologic complications including encephalitis, transverse myelitis, and postinfectious encephalomyelitis.3,4
Some argue that infections play a role in both initiating and perpetuating MS, and investigators have demonstrated increased risks of relapse following viral infections.5,6 Epidemiologic studies have produced inconsistent evidence for an etiologic role of respiratory tract infections (RTI) and IM in MS.7-15
Using data from the United Kingdom General Practice Research Database (GPRD), we investigated the association between infections and MS. The computer-recorded information in this population-based database regarding diagnoses and drug exposures has been validated for research purposes.16-18 We examined whether 1) there is an excess of RTI in the 5-week, 3-month, and 12-month periods before MS symptom onset and 2) there is an association between MS and a history of IM.
Methods.
This is a population-based case-control study using the UK GPRD. The British National Health Service provides free primary care by a general practitioner (GP) to every resident in the United Kingdom.16 All persons are registered with a single GP who holds the medical record of each individual on his or her register, including the results of all secondary care. Demographic information, drug prescriptions, and clinical events, including symptoms, signs, diagnoses, investigations, and hospital care, are recorded using computers provided to GPs participating in the database. Medical diagnoses are recorded using a modification of the Oxford Medical Information system classification (OXMIS), and there is a unique patient identifying code for each individual in a participating practice that can be decoded only by the GP.
The database represents approximately 8% of the UK population and the distribution by age and sex is comparable to that of the UK population.16 The database population is also representative in terms of birth and death rates and incidence of common diseases.16
Between January 1, 1990 and June 1, 1997, 625 practices contributed information regarding approximately 4 million subjects. This represents the source population for this study. The study population consists of subjects age 20 to 60 years with permanent registration status or death at June 1, 1997, and with a registration date before January 1, 1991.
Cases.
Using computerized patient records, the database was searched for all subjects with an incident diagnosis of MS, optic neuritis, transverse myelitis, or demyelination that had led to at least one specialist referral or hospitalization. The computerized records of those subjects were reviewed to eliminate persons for whom the initial symptoms consistent with the outcome were already present at the date of study entry. A panel of three neurologists blinded to exposure status reviewed computerized case notes and classified cases as either definite or probable MS according to the following criteria, designed specifically for this study.
Classification as definite MS required: 1) a diagnosis of multiple (disseminated) sclerosis with specialist or hospital documentation, and relevant neurologic symptoms/items in the database before the date of diagnosis; or 2) a diagnosis of demyelination, optic neuritis, or myelitis leading to specialist or hospital documentation, and relevant neurologic symptoms/items in the database before diagnosis. Another documented episode of demyelination was required after the initial date to lead to a diagnosis of definite MS, although not written as such in the database.
Cases were classified as probable MS if a diagnosis of MS leading to specialist or hospital documentation was identified, but without relevant neurologic symptoms/items before the date of diagnosis.
Index date.
The index date for cases was defined as the date of initial symptoms. There is often a delay between the presentation of initial symptoms and the diagnosis of MS, and the choice of index date was important to ensure an appropriate temporal relationship between exposure and disease. The same panel of three neurologists blinded to exposure status used computerized chart information to identify the date of the first symptoms compatible with MS.
Controls.
One to six controls matched on sex and age (within 2 years) were identified for each case from the same database. Matching on physician practice was performed to control for residence and the probability of diagnosis. The index date for controls was similar to that of the cases.
Exclusion criteria.
The exclusion criteria were applied to both cases and controls. Subjects were excluded if they had any of the following: prevalent central demyelinating disease with onset before January 1, 1991; neoplasm before the index date; immunosuppressive therapy before the index date; medical illnesses with etiologies postulated to be infection or immune-mediated (for example, rheumatoid arthritis, vasculitis, Guillain-Barré syndrome). Subjects were also excluded if medical information was incomplete in the year before the index date or there was no prescription for medication in the year before January 1, 1991. The latter two criteria ensured availability of complete exposure data during the periods of interest.
Exposure.
A GP diagnosis of IM occurring in the database at any time before the index date was considered an exposure. This was documented either prospectively (on occurrence) or retrospectively (as medical history on registration in the database, but always before the index date).
Complete exposure information was available for the 12 months before the index date. Given the sometimes limited information in the database for attributing symptoms such as cough to either upper or lower RTI, and the clinical difficulties of distinguishing between the two, all respiratory infections were grouped together. Computerized diagnoses recorded as “cold,” “coryza,” “upper respiratory tract infection,” “sinusitis,” “pharyngitis,” “laryngitis,” “rhinitis,” “sore throat,” “chest cold,” “pneumonia,” “pleurisy,” “acute bronchitis,” “tracheitis,” or “cough” were coded as RTI using a pre-established protocol. Of interest were the 5-week and 3- and 12-month periods before the index date. The 5-week period was selected because of studies by Sibley and others demonstrating increased relapse rates 5 weeks post-RTI in patients with established MS.5,6 The 3- and 12-month periods were selected somewhat arbitrarily to examine longer term patterns and effects of exposure.
Statistical analysis.
For the purposes of this analysis, only cases with four to six matched controls available were used. Four controls were randomly selected from the pool available for each case. Odds ratios, confidence intervals, and significance testing were determined by conditional logistic regression using SAS statistical software (Cary, NC). Conditional logistic regression analyzed subjects by strata, each comprised of a case and its matched controls, thus adjusting for confounding due to age, sex, and practice.
Results.
There were 225 cases and 900 controls. The mean age of all subjects was 40.3 ± 9.2 years. Table 1 shows the age distribution. The ratio of men to women was 1:2.3. Eighty-eight percent of the cases had definite MS; the remainder had probable MS. The frequency of other major medical and neurologic diagnoses was similar between cases and controls.
Age and sex distribution
Using conditional logistic regression, the OR for a history of IM was 5.5, with a 95% CI of 1.5 to 19.7. All subjects with a history of IM were or were matched to definite cases. The mean age at occurrence of IM was 23.3 ± 9.1 years (n = 11) (table 2). Dichotomizing IM exposure as present if it occurred late (after age 17 years) and absent if it occurred early or not at all, the OR increased to 6.0 (95% CI 1.4 to 25.4).
Mean number of respiratory tract infections (RTI) and mean age at occurrence of infectious mononucleosis (IM)
Table 2 shows the average number of RTI per person during the exposure periods of interest. Results of conditional logistic regression for cases of definite MS for the periods of interest are shown in table 3. In all periods an increased frequency of RTI was associated with a significantly increased risk of MS. Exposure to RTI in the 3 months before the index date was associated with more than double the risk of MS. Analyses including definite and probable cases yielded similar results (see table 3).
Conditional odds ratios (95% CI) for respiratory infections
Discussion.
This study’s strengths center on the use of a population-based database. Use of incident cases reduced the risks of selection and selective survival biases. Respiratory infections were recorded prospectively after entry in the database. IM exposure was recorded retrospectively, on entry to the database, but before the identification of subjects as cases or controls, reducing the risk of differential recall. Although the diagnoses of IM and RTI were based on GP diagnoses and were not validated by comparison of computer records with the original medical charts, such computer-recorded information from the GPRD has previously been validated for research purposes for a range of diagnoses.17,18
The herpesviruses are known for their ubiquity, neurotropism, and tendency to produce latent, recurrent infections.19 Acute Epstein-Barr virus (EBV) infection leads to lifelong, latent infection of B-lymphocytes. Infection is often mild or inapparent in young children, whereas infection later in life may present as IM. Acute infection is known to be associated with neurologic complications.3 Our data show an association between a prior diagnosis of IM and a subsequent MS in at least some individuals, and that association depends partly on the age at infection. This is consistent with results of published case-control and historical cohort studies, wherein OR and relative risks ranged from 2.8 to 17.0.10,12-14
We found an even higher risk of MS when IM occurred after age 17 years, but Martyn et al. found the opposite.13 An increased risk was reported in seropositive subjects with a history of IM (OR 2.9) and this risk increased further if IM occurred before age 17 years (OR 7.9, 95% CI 1.7 to 37.9).13 However, their cases included patients with acutely relapsing MS, optic neuritis, and other isolated demyelinating lesions. Past exposure was assessed during an investigator-administered interview. Such use of recalled data and measurements by an unblinded investigator are potential sources of bias affecting their results. Another study failed to detect a significant difference in the reported age at occurrence of IM.8 The clinical significance of our results (as well as those of other studies) is unclear because of the small number of cases of IM.
Case-control and historical cohort studies investigated the association between RTI in early life (childhood or adolescence) and subsequent MS.7-9,15 Even the best case-control studies enrolled prevalent cases and used questionnaires for exposure ascertainment, with the risk of selective survivor and recall biases.8,9,15 None demonstrated an association with lower RTI (pneumonia or bronchitis).7-9 Results for upper RTI were variable, with increased frequency of tonsillitis or sinusitis reported along with absence of association.8,9,15
Although RTI are among the most frequent reasons for consulting a GP, patients seek consultation for only a fraction of their RTI.20 This and inaccuracies in diagnosis are possible sources of error in our measurements of exposure. Given that patients were matched by physician practice, and exposures were recorded before symptom onset, these errors should be similar in both groups. Such nondifferential misclassification bias would tend to reduce, rather than increase, the chances of demonstrating an association. The assumption that patients who later develop MS are as likely as other patients to report RTI is another potential source of bias. Nonspecific symptoms of MS, such as fatigue, may precede neurologic symptoms and could lead to more frequent GP visits, thus increasing the opportunity for diagnosis of RTI. Symptoms of RTI might be less tolerated in the pre-MS group, again prompting more frequent visits. The possibility that pre-MS subjects with RTI are over-represented in the database could induce a selection bias. We found significantly increased risks of MS associated with an increased frequency of RTI during the 5-week, 3-month, and 12-month periods before disease onset. This increase in risk was most pronounced in the 3-month interval. Given the increased risk in the 5-week interval before disease onset, RTI might “precipitate” disease onset in a manner akin to their known association with exacerbations.5,6
Studies have established an association between MS exacerbations and viral infections, particularly RTI.5,6 One prospective study followed 41 subjects enrolled in a randomized, placebo-controlled trial of interferon-β1a for a total of 2516 patient-weeks.5 Each patient recorded any infections or MS symptoms in a diary reviewed at each patient visit. “At-risk” periods in this study ranged from 2 weeks prior to 2 weeks after infection. Reported relative risks of exacerbation during “at-risk” periods were 2.0 in this study and 2.8 in another prospective study.5,6
Our study shows an association between a prior diagnosis of IM and subsequent MS. RTI may precipitate disease onset. Further study to better characterize the nature of these associations is required.
Acknowledgments
Supported by the French Department of Health (Direction Génèrale de la Santé et Reseau National de Santé Publique).
- Received November 2, 1999.
- Accepted in final form March 13, 2000.
References
Letters: Rapid online correspondence
REQUIREMENTS
If you are uploading a letter concerning an article:
You must have updated your disclosures within six months: http://submit.neurology.org
Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.
If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.
Submission specifications:
- Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
- Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
- Submit only on articles published within 6 months of issue date.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- Submitted comments are subject to editing and editor review prior to posting.
You May Also be Interested in
Dr. Jessica Ailani and Dr. Zubair Ahmed