ABSTRACT

Context.  A review by the Institute of Medicine found a possible relationship between rubella vaccination and chronic arthritis among women.

Objective. To evaluate the risk of persistent joint and neurologic symptoms in rubella seronegative women subsequently vaccinated with RA 27/3 rubella vaccine.

Design. Retrospective cohort study based on computerized laboratory data and medical record review. Records were reviewed for symptoms occurring within 2 years before and after the date of serological testing and to identify vaccinees. Possible cases were evaluated by a rheumatologist blinded to serological findings and vaccination status.

Setting. Large health maintenance organization in northern California.

Patients. Women aged 15-59 years serotested for rubella during 1990 with continuous health plan membership for 2 years before and after the date of their serological test. Seronegative women immunized within one year of serotesting (n=971) were defined as exposed. Primary comparison groups included all unvaccinated, seronegative women (n=924) and randomly selected seropositive, unvaccinated women (n=2421) matched to exposed subjects on serological test date and age (+3 years).

Main Outcome Measures. Prevalence and incidence of chronic joint and neurologic symptoms during 1 year follow-up period stratified by age and serological findings, immunization, and postpartum status.

Results. No significantly increased risk was associated with receipt of rubella vaccine for any outcome except for prevalence of carpal tunnel syndrome in vaccinated women at least 30 years old compared with seropositive, unvaccinated women (2.9% vs. 1.4%, p=.03). A total of 34 women had onset of conditions within the one year follow-up period; 9 of these were in the group of seronegative, immunized women, of whom 6 had onset of symptoms within 6 weeks of vaccination.  Among these 6 women, symptoms included transient arthritis or arthralgias (<6 weeks duration) in 4 women, arthralgia of indeterminate chronicity in 1 woman, and carpal tunnel syndrome in 1 woman. Postpartum women across all groups were less likely to be seen for nontraumatic arthropathies than nonpostpartum women (4.5% vs. 7.2%, p=.08 in vaccinated women; 4.8% vs. 8.1%, p=.09 in seronegative controls; and 4.8% vs. 10.0%, p=.01 in seropositive controls).

Conclusions. In this large retrospective cohort analysis there was no evidence of any increased risk of new onset chronic arthropathies or neurologic conditions in women receiving the RA 27/3 rubella vaccine. These data support the continued vaccination of rubella-susceptible women to reduce the risk of congenital rubella syndrome.

INTRODUCTION
The objective of rubella vaccination programs in the United States. is to prevent the consequences of rubella, particularly congenital rubella syndrome (CRS). The Advisory Committee on Immunization Practices (ACIP) of the Centers for Disease Control and Prevention (CDC), Atlanta, Ga,  recommends that all persons 12 months of age or older without evidence of rubella immunity, especially women of childbearing age, be vaccinated against rubella (1). The ACIP also recommends that all pregnant women not known to be immune be screened for evidence of rubella immunity. Those who are found to be susceptible should be vaccinated after delivery, before discharge from the hospital. The basis for this, in part, is that approximately 40% of CRS cases in the United States occur in infants born to mothers who had a previous live birth; such cases could be prevented by postpartum rubella immunization (2).

Acute joint complaints following vaccination of women with RA 27/3 rubella vaccine are common, occurring in approximately 25% of seronegative adult female vaccinees. It is generally believed that persistence or recurrence of these symptoms is rare (2-5). One group of investigators in Canada has reported that the incidence of persistent or recurrent frank arthritis following rubella vaccination is as high as 5%-11% in small studies of adult female vaccinees (6,7). This rate is higher than the background annual rate of chronic arthritis owing to all causes in women. Data from the National Health Interview Survey indicate that 2% of persons younger than 35 years and 11% of persons aged 35-44 years have consulted a physician within the last year for arthritis (8).

Other chronic illnesses that have been reported following rubella vaccination include paresthesias, painful limb syndrome, blurred vision, fibromyalgia, and fatigue (9,10). Support for a possible role of rubella vaccine virus in causing chronic arthritis comes from isolation of virus from peripheral blood lymphocytes and/or synovial fluid of such patients using special cocultivation techniques (6,11).

In 1991 the Institute of Medicine (IOM) reviewed the available scientific evidence and determined it to be consistent with, but not probative of, a causal relationship between rubella vaccination and chronic arthritis in women. The IOM recommended further investigations to clarify the relationship (12). With support from the Centers for Disease Control and Prevention's Vaccine Safety Datalink project (13), we undertook a retrospective cohort evaluation to assess whether the risk of persistent joint and neurologic symptoms is higher among rubella-seronegative women who are subsequently vaccinated with rubella vaccine than among unvaccinated seronegative and seropositive women.

Top of Page

METHODS

Study Population:
The study population was drawn from adult female members of the Northern California Kaiser Permanente Health Plan, Oakland, a large health maintenance organization (HMO) with a total membership of approximately 2.5 million and 35,000 pregnancies annually. Perinatal testing for rubella immunity is routine, and individuals may also be tested for employment or school purposes, at physical examinations, or to rule out rubella infection. In 1993, the regional laboratory database was used to select a retrospective cohort of women between the ages of 15 and 59 years who had serological testing by enzyme immunoassay (EIA) for rubella IgG antibody in 1990 and were health plan members for 2 continuous years before and after their serological test date. Continuous HMO membership for 2 years preceding and 2 years following the date of serological testing was required so that complete medical records for the surveillance period would be available. To obtain a sufficient number of older vaccinated subjects the sample was augmented by including women aged 30 - 59 years whose serological testing was performed during the first two months of 1991.

Medical records of all seronegative members of the cohort (EIA value < .8) were reviewed to identify women who received rubella vaccine within one year following serological testing. Vaccinated women were defined as exposed, and those who were not vaccinated were assigned to the "seronegative, unimmunized group," one of two primary comparison (control) groups. Women whose immunization status could not be determined or who were vaccinated between the first and second year after their serological tests were excluded from the study. The latter group was excluded to allow a full year of follow-up time in order to detect possible delayed onset of study related outcomes. Also excluded were women who had discordant serological tests during the two years prior to the 1990 serological testing date.

A second comparison (control) group included seropositive women (EIA value > 1.2) who were matched to exposed subjects on the date of serological testing (within 30 days) and were age-matched within 3-year age groups. Two seropositive women were selected for each seronegative, immunized woman younger than 30 years and three for those at least 30 years old. Vaccination status also was verified for the seropositive women and those who were vaccinated within two years prior to their rubella serological tests were excluded from the study as were those whose vaccine status was indeterminate or who had had a negative serological test within the previous two years. Replacements were selected for all potential seropositive women who were excluded.

Medical Record Reviews:
A new medical record is generated whenever a patient is seen for the first time at any of the 32 HMO facilities in the region. An interactive regional database provides information on patients' medical record locations, type of record (inpatient vs. outpatient), active or inactive status, and availability on any given date. For the initial medical record abstractions, all inpatient and outpatient records that were active at any time between January 1, 1988, and February 28, 1993, were reviewed. In addition to serological and immunization status the medical record analysts abstracted information on rubella disease history, reason for rubella serology, and the conditions under investigation. These included arthropathies, neuropathies, paresthesias, systemic lupus erythematosus, chronic fatigue syndrome, fibromyalgia, carpal tunnel syndrome, and tendinitis.

Reviewers were instructed to record arthropathies or joint complaints regardless of whether they were noted to be chronic, acute, traumatic, caused by overuse, or secondary to another disease. Conditions that were clearly traumatic in origin were flagged (e.g. sports injuries, accidents). No attempt was made at this level to classify joint conditions into diagnostic categories, beyond differentiating between traumatic and nontraumatic conditions. Joint symptoms for which a traumatic origin was uncertain were categorized as nontraumatic. Procedures such as joint taps, intra-articular injections, arthroscopies, and other joint surgery were captured as was any mention of prior history of juvenile rheumatoid arthritis, arthritis, neuritis, trauma to joints, sickle cell disease, or family history of arthritis. Strains, sprains, and other muscle and ligament problems were excluded as were episodes of nonspecific back and neck pain (arthritis not specified), influenza-related joint pain, fractures without joint involvement, and sickle cell trait.

Following the initial screening a second review was performed by the medical record analysts to identify individuals whose records met criteria for further evaluation by a rheumatologist. The medical record analysts conducting these secondary reviews were not blinded to serological or immunization status because of  logistical considerations and the resources that would have been required to mask portions of the large numbers of records indicating vaccination and serological status. Records of patients without evidence of joint complaints or other conditions of interest occurring during the study period, those who had only a single episode of trauma to a joint, or those who had tendinitis only were not referred to the rheumatologist. Cases not meeting these exclusion criteria were further evaluated to eliminate any with conditions resulting from trauma or overuse with a known cause or those that were clearly mechanical or degenerative in nature.

Records of patients whose complaints could not be ruled out temporally (none during the study period) or attributed to a known etiology were referred to a rheumatologist who was blinded to serological and immunization status. The rheumatologist was asked to identify potential immunization-related cases based on information in the medical record and to assign these cases to diagnostic categories according to criteria set by the American Rheumatism Association (14). Symptoms were considered to be potentially attributable to rubella vaccination unless in the rheumatologist's judgment they were mechanical, degenerative, or non-inflammatory in nature; were due to trauma or overuse; were pregnancy related; or were secondary to another diagnosis. If the rheumatologist was unable to rule out an individual based on history, symptoms, or date of onset (prior to 1990), the patient was accepted as a possible vaccine-related case.

When sufficient information was available in the medical record the rheumatologist also specified or estimated an onset date and, for diagnoses not considered chronic by definition, she attempted to assess duration. Chronic cases were defined as those with symptoms persisting for more than six weeks.

Top of Page

Analysis:
For the primary analysis, seronegative, subsequently immunized women (exposed subjects) and seropositive, unimmunized women (age-matched controls) were followed for one year from the immunization date of the matched subject. Seronegative, unimmunized women (unmatched controls) were followed for one year from their 1990 serological dates dates. Estimates of prevalence among these groups were based on all study-related conditions identified by the medical record analysts at the initial record review, which occurred within the one year follow-up period regardless of their date of onset or etiology. Incidence rates of possible vaccine-related outcomes were calculated using patients identified by the rheumatologist as potential cases with onset of symptoms occurring during the one year follow-up period. The denominators for the one year follow-up period prevalence rates were the entire study groups. For incidence rates, the denominators were the study groups minus persons with the condition present at or before the start of follow-up.

The prevalence data were also analyzed controlling for pregnancy status. Women who had been vaccinated postpartum were identified by accessing the HMO's automated regional hospitalization database. Any rubella-vaccinated woman in the study population with a childbirth-related hospitalization during the nine months following the 1990 serological testing date who was immunized within six weeks following the hospitalization or whose abstracted reason for serological testing was "pregnancy" was defined as having been vaccinated postpartum and was followed up for one year from the vaccination date. Women in the two unvaccinated comparison groups (seronegative and seropositive) who were defined as pregnant based on the hospitalization database were followed for one year from the hospital admission date for childbirth. Those unvaccinated women whose reason for serological testing was "pregnancy" but who did not have a hospitalization for childbirth within the HMO were excluded from this analysis since a delivery date could not be established. Also excluded were women older than 44 years.

Sample sizes and the number of matched controls selected were based on calculations that would allow detection of an elevated risk for persistent arthritis of 2.5 in vaccinated women younger than 30 years and 1.5 in those at least 30 years with 95% confidence and 80% power. It was assumed that the cumulative incidence in unexposed women is 2% in the younger age group and as high as 10% in the older group. The P values were calculated using a two-tailed Chi Square test unless the expected cell count for a comparison was less than five, in which case the Fisher exact test was used. For the analysis by postpartum status, the Cochran-Mantel-Haenszel statistic was used to test for association adjusting for age. The alpha of .05 was not adjusted for multiple comparisons. (See discussion in comment section).

The study was approved by the Kaiser Foundation Research Institute's institutional review board.

RESULTS:
Overall, 14,460 inpatient and outpatient records of 4,884 women were reviewed. Table 1 describes the study population status based on the initial medical record reviews. The number of women immunized was similar in the two seronegative age groups, although proportionately somewhat higher in the younger group (55.4% vs. 47.6%).

The prevalence of conditions during the one year follow-up period, as identified at the initial record reviews by the medical record analysts, is shown in Table 2. These rates indicate the proportion of patients with visits at any time during the follow-up period and are regardless of onset prior to or during the study period. Prevalence did not differ significantly across the three groups with the exception of vaccinated women at least 30 years old who were approximately twice as likely to have been diagnosed with carpal tunnel syndrome as those in either of the comparison groups. The proportion of arthropathies, both traumatic and nontraumatic, was similar in the immunized and unimmunized seronegative women in both age groups and in the older seropositive group. Among the younger women, non-traumatic arthropathies tended to be reported more frequently by those in the seropositive, unimmunized group compared with those in the seronegative, immunized group (4.8% vs. 2.9%; p =.08). The younger seropositive matched controls were more likely to have been seen for traumatic arthropathies, paresthesias, and tendinitis (4.1% vs. 2.0%, P= .04;  2.0% vs. 0.6%, P= .04; and 2.0% vs. 0.4%, P= .02, respectively). Significantly more tendinitis as well as fibromyalgia was also observed in the seropositive, unimmunized group when all ages were pooled. Age adjustment made no appreciable difference in these results.

Incidence rates of new onset conditions by diagnostic category as assigned by the rheumatologist are reported in Table 3. The medical records of 732 patients were referred to the rheumatologist for further evaluation. Initially 113 of these were identified by the rheumatologist as having conditions that were potentially related to rubella vaccination; that is, no other cause of the symptom or complaint could be readily ascertained. Of these 113 possible cases, 34 had onset of symptoms within the one year follow-up period: 9 were in the seronegative, immunized group; 4 were in the seronegative, unimmunized group; and 21 were in the seropositive, unimmunized group. Among the rubella-vaccinated women, six had developed symptoms within six weeks of immunization. One of these was diagnosed with carpal tunnel syndrome and the remaining five cases were categorized as "other rheumatic conditions." The majority of patients falling into this category were diagnosed with arthralgias. Of the five conditions in rubella-vaccinated women, four were defined as acute (less than six weeks in duration) and one as indeterminate. Among the age-matched seropositive controls diagnosed with "other rheumatic conditions", one patient's condition was acute; six were chronic. Chronicity was indeterminate for the seven remaining cases in unimmunized women. The rates presented in Table 3 are without regard to chronicity; however, with the exception of the "other rheumatic conditions," all diagnostic classifications are chronic by definition. No statistically significant differences in rates were found for any diagnosis between the vaccine-exposed women and either comparison group in any age category.

Ten percent of the 3345 women in the two comparison groups were excluded from the analysis controlling for pregnancy status because their delivery dates could not be established, and an additional 306 women older than 44 years were excluded. In comparing the one year follow-up period prevalence of the various outcomes of interest between postpartum and nonpostpartum women significant differences were seen, both within and across the three groups. However, no consistent patterns emerged indicating an association of any condition with either pregnancy and/or rubella vaccination, with the exception of differences according to postpartum status of older women (Table 4). In each of the three older groups a greater percentage of the nonpostpartum patients were seen for both nontraumatic and traumatic arthropathies during the follow-up period; however, after adjusting for age these differences were significant only in the seropositive control group for nontraumatic arthropathies. Among the younger women, traumatic arthropathies were reported more often among nonpostpartum than postpartum patients. With the exception of traumatic arthropathies in nonpostpartum vaccinated vs. nonpostpartum seropositive patients who were younger than 30 years, no appreciable differences were seen for arthropathies of any type in comparisons between the vaccine-exposed and control groups.

Top of Page

COMMENT
We found no evidence of a causal relationship between immunization with the RA 27/3 strain of rubella vaccine and persistent joint symptoms, neuropathies, or any other chronic condition studied in this retrospective cohort analysis. For certain individual conditions within these categories, however, the sample size may not have been large enough to detect an elevated vaccine-associated risk. A diagnosis of rheumatoid arthritis was found in the medical record of one of the 4,884 women in the study population, which is similar to incidence rates reported in other studies (15,16). Furthermore, this particular case was diagnosed in a member of the seropositive, unimmunized comparison group. These results contrast with the early findings of some investigators studying the HPV-77 (DK) and HPV-77 (DEV) strains of rubella vaccine (17-19), as well as with those of Tingle and colleagues investigating the RA 27/3 strain (6,7,9,11,12,20) used since 1979, but they are similar to more recent findings by Tingle (21) and Slater et al (22) . Although we found a higher rate of rheumatic conditions other than rheumatoid arthritis in the older exposed subjects, all but one of these patients had symptoms that were of limited duration and were defined as acute by the rheumatologist. If vaccine-related, these conditions are consistent with the well-documented short-term adverse events associated with rubella immunization.

We found little other evidence of the commonly reported transient joint complaints following vaccination that have been documented elsewhere. This may have been a function of the study design, which did not include active prospective follow-up of recently vaccinated patients but relied instead on retrospective medical record review of patients engaged in routine utilization of medical care who may be unlikely to seek treatment for mild, anticipated symptoms of short duration. Women are routinely informed in advance of the possible adverse effects of vaccination including arthralgias and are told that these symptoms are self-limited. These women may be less likely to report the expected symptoms if they occur.

The prevalence rates of arthritis or arthropathies for our study population were very similar to the reported background rate among the female health plan members from whom we drew our sample. In a 1993 Kaiser Permanente Member Health Survey conducted within the HMO that sampled 34,000 health plan members aged 20 years and older, 2.2% of female respondents aged 30 years or younger and 8.7% of those aged 31 to 59 years reported that they had had arthritis during the past year (Nancy Gordon, ScD, written communication, June 1997).

The higher prevalence of carpal tunnel syndrome among the older vaccinated women in our study is puzzling. Cases of carpal tunnel syndrome have been reported following both wild rubella infection (23) and vaccination with the HPV-77 (DK) (24) and RA 27/3 (9) strains, but this condition has not been clearly associated with rubella vaccination. Higher rates of carpal tunnel syndrome were also seen in the analysis of incidence rates for this group although this difference was not statistically significant. Of the two cases identified, however, only one had onset of symptoms within six weeks of vaccination. The second patient did not develop symptoms until four months after vaccination, an interval that attenuates the likelihood of a causative relationship with immunization. The findings with respect to carpal tunnel syndrome and the greater proportions of traumatic arthropathies, paresthesias, and tendinitis in the younger seropositive, unimmunized group are unexplained but may simply be an artifact of the multiple comparisons engendered by this type of analysis. The nature of the significant associations seen in this analysis as well as the relatively low levels of significance found suggest that this may be the case.

Up to 75% of cases of rheumatoid arthritis go into remission during pregnancy (25-28). In addition, some studies have shown an increased risk for onset of rheumatoid arthritis during the postpartum period, suggesting that hormonal influences and timing of vaccination (9,29-31) may be factors in the development of arthropathies following rubella immunization. It has also been suggested based on the few cases observed to date that persistent arthritis subsequent to rubella vaccination may be more likely to occur in women who have altered immune function (6), as may be the case during the early postpartum period when it is also common practice to vaccinate rubella seronegative individuals. More than half our study population was vaccinated within six weeks of delivery. For these reasons we further stratified the data on arthritis based on pregnancy status, although this reduced the power to detect differences among the six sub-groups. Due to these power limitations the rates reported are not for new-onset conditions alone but rather for all conditions identified during the one year follow-up period. Nevertheless, the very similar rates of arthropathies in the younger postpartum vs. nonpostpartum vaccinated subjects and the somewhat lower rates observed in postpartum older women are reassuring and would tend to support that rubella immunization following childbirth is not a risk factor for developing arthritis.

A clinical follow-up study of the 34 patients identified as having potential vaccine- induced conditions is currently underway. These patients will be examined by a rheumatologist to establish a clinical diagnosis. Blood samples will be obtained to test peripheral blood leukocytes for the presence of rubella virus. Specimens that are positive for rubella will be analyzed by molecular genetic techniques to determine if the virus is derived from a vaccine strain.

This is the largest study to date to address the issue of rubella vaccine-associated arthropathy. The overall incidence rate of the main outcome under study (chronic arthropathy) was 1.6 per 1000 person years based on seven cases identified in the rheumatoid arthritis and "other rheumatic conditions" categories. All of these chronic cases occurred in seropositive, unvaccinated women yielding a relative risk of zero and 95% confidence intervals of 0.0 - 1.5 for the vaccinated group. The upper bound of this interval is lower than both the overall rate observed and the rate in the unvaccinated comparison group (2.9 per 1000 person years) thus permitting us to rule out that there is an increased risk of chronic arthropathy following rubella vaccination.

In conclusion, in this large retrospective cohort analysis there was no evidence of any increased risk of new onset chronic arthropathy within one year following vaccination in women receiving the RA 27/3 rubella vaccine. However, age is a risk factor for most of the outcomes studied regardless of serological or immunization status. Public health programs promoting rubella vaccination, including those for women at risk, have had an important role in reducing the individual and societal impact of congenital rubella syndrome. Our findings support the continued vaccination of rubella-susceptible women of childbearing age as a safe and effective means of preventing congenital rubella syndrome.

Top of Page

REFERENCES:
¹ CDC. Rubella prevention. Recommendations of the Immunization Practices Advisory Committee, MMWR, 1990; 39:10-12.

² Kaplan KM, Cochi SL, Edmonds LD, Zell ER, Preblud SR. A profile of mothers giving birth to infants with congenital rubella syndrome. Am J Dis Child 1990;144:118-123.

³ Polk BF, Modlin JF, White JA, DeGirolami PC. A controlled comparison of joint reactions among women receiving one of two rubella vaccines. Am J Epidemiol. 1982;115:19-25.

⁴ Peltola H, Heinonen OP. 1986. Frequency of true adverse reactions to measles-mumps-rubella vaccine: a double-blind placebo-controlled trial in twins. Lancet. 1986;1:939-942.

⁵ Valensin PE, Rossolini GM, Cusi MG, Zanchi A, Cellesi C, Rollolini A. Specific antibody patterns over a two-year period after rubella immunization with RA 27/3 live attenuated vaccine. Vaccine. 1987;5:289-294.

⁶ Tingle AJ, Allen M, Petty RE, Kettyls GD, Chantler JK. Rubella-associated arthritis. 1. Comparative study of joint manifestation associated with natural rubella infection and RA 27/3 rubella immunization. Ann Rheum Dis. 1986;45:110-114.

⁷ Tingle AJ, Yang T, Allen M, Kettyls GD, Larke B, Schulzer M. Prospective immunological assessment of arthritis induced by rubella vaccine. Infect Immun. 1983;40:22-28.

⁸ CDC. Prevalence and impact of arthritis among women - United States, 1989-1991. MMWR 1995;44:329-335.

⁹ Tingle AJ, Chantler JK, Pot KH, Paty DW, Ford DK. Postpartum rubella immunization: association with development of prolonged arthritis, neurological sequelae, and chronic rubella viremia. J Infect Dis. 1985; 152:606-612.

¹⁰ Morton-Kute L. Rubella vaccine and facial paresthesias (letter). Ann Intern Med.1985; 102:563.

¹¹ Tingle AJ, Pot CH, Chantler JK. Prolonged arthritis, viraemia, hypogammaglobulinaemia, and failed seroconversion following rubella immunisation. Lancet. 1984;1:1475-76.

¹² Howson CP, Katz M, Johnston RB, Fineberg, HV. Chronic arthritis after rubella vaccination. Clin Infect Dis. 1992;15:307-12.

¹³ Chen RT, Glasser JW, Rhodes PH, Thompson RS, Mullooly JP, Black SB, Shinefield HR, Vadheim CM, Marcy SM, Ward JI, Wise RP, Wassilak SG, Hadler SC, and the Vaccine Safety Datalink Team. Vaccine Safety Datalink project: a new tool for improving vaccine safety monitoring in the United States. Pediatrics. 1997; 99:765-773.

¹⁴ Schumacher HR, Klippel JH, Koopman WJ. Primer on the Rheumatic Diseases. Atlanta, GA: Arthritis Foundation; 1993: 328 - 332.

¹⁵ Dugowson CE, Koepsell TD, Voigt LF, Bley L, Nelson JL, Daling JR. Rheumatoid arthritis in women: incidence rates in Group Health Cooperative, Seattle, Washington, 1987-1989. Arthritis Rheum. 1991; 34:1502- 07.

¹⁶ Linos A, Worthington JW, O'Fallon MW, Kurland LT. The epidemiology of rheumatoid arthritis in Rochester, Minnesota: a study of incidence, prevalence, and mortality. Am J Epidemiol. 1980; 111: 87-98.

¹⁷ Lerman SJ, Nankervis GA, Heggie AD, Gold E. Immunologic response, virus excretion, and joint reactions with rubella vaccine: a study of adolescent girls and young women given live attenuated virus vaccine (HPV-77:DE-5). Ann Intern Med. 1971;74:67-73.

¹⁸ Spruance SL, Metcalf R, Smith CB, Griffiths MM, Ward JR. Chronic arthropathy associated with rubella vaccination. Arthritis Rheum. 1977;20:741-7.

¹⁹ Thompson GR, Weiss JJ, Shillis JL, Brackett RG. Intermittent arthritis following rubella vaccination: a three-year follow-up. Am J Dis Child. 1973;125:526-30.

²⁰ Mitchell LA, Tingle AJ, Shukin R, Sangeorzan JA, McCune J, Braun DK. Chronic rubella vaccine-associated arthropathy. Arch Intern Med. 1993; 153:2268-74.

²¹ Report of an international meeting on rubella vaccines and vaccination, 9 August 1993, Glasgow, United Kingdom. J Infect Dis. 1994;170:507-509.

²² Slater PE, Tirtsa B, Fogel A, Ehrenfeld M, Shai E. Absence of an association between rubella vaccination and arthritis in underimmune postpartum women. Vaccine. 1995; 13:1529-1532.

²³ Institute of Medicine. Adverse Effects of Pertussis and Rubella Vaccines. Washington, D.C.:National Academy Press; 1991.

²⁴ Thompson GR, Ferreyra A, Brackett RG. Acute arthritis complicating rubella vaccination. Arthritis Rheum. 1971; 14:19-26.

²⁵ Da Silva JAP, Spector TD. The role of pregnancy in the course and aetiology of rheumatoid arthritis. Clin Rheum. 1992; 11:189-194.

²⁶ Neely NT, Persellin RH. Activity of rheumatoid arthritis during pregnancy. Texas Med. 1977;73: 59-63.

²⁷ Ostensen M, Aune B, Husby G. Effect of pregnancy and hormonal changes on the activity of rheumatoid arthritis. Scand J Rheum. 1983;12:69-72.

²⁸ Ostensen M, Husby G. A prospective clinical study of the effect of pregnancy on rheumatoid arthritis and ankylosing spondylitis. Arthritis Rheum. 1983;26:1155-9.

²⁹ Lansink M., De Boer A, Dijkmans BAC, Vandenbroucke JP, Hazes JMW. The onset of rheumatoid arthritis in relation to pregnancy and childbirth. Clin Exp Rheum. 1993;11:171-174.

³⁰ Persellin RH. The effect of pregnancy on rheumatoid arthritis. Bull Rheum Dis. 1977;27: 922-27.