Tseng HF, Slezak JM, Quinn VP, Sy LS, Van Den Eeden SK, Jacobsen SJ. Pneumococcal Vaccination and Risk of Acute Myocardial Infarction and Stroke in Men. JAMA. 2010;303(17):1699-1706. doi:10.1001/jama.2010.529
Author Affiliations: Department of Research and Evaluation, Kaiser Permanente Southern California, Pasadena (Drs Tseng, Quinn, and Jacobsen and Mr Slezak and Ms Sy); and Division of Research, Kaiser Permanente Northern California, Oakland (Dr Van Den Eeden).
Context Multiple studies have shown that preventing influenza by vaccination reduces the risk of vascular events. However, the effect of pneumococcal polysaccharide vaccine on vascular events remains controversial.
Objective To examine the association between pneumococcal vaccination and risk of acute myocardial infarction (MI) and stroke among men.
Design, Setting, and Participants A prospective cohort study of Kaiser Permanente Northern and Southern California health plans with 84 170 participants aged 45 to 69 years from the California Men's Health Study who were recruited between January 2002 and December 2003, and followed up until December 31, 2007. The cohort was similar to the population of health plan members and men who responded to a general health survey in California on important demographic and clinical characteristics. Demographic and detailed lifestyle characteristics were collected from surveys. Vaccination records were obtained from the Kaiser Immunization Tracking System.
Main Outcome Measure Incidence of acute MI and stroke during the follow-up period in men who had no history of such conditions.
Results During follow-up, there were 1211 first MIs in 112 837 vaccinated person-years (10.73 per 1000 person-years) compared with 1494 first MI events in 246 170 unvaccinated person-years (6.07 per 1000 person-years). For stroke, there were 651 events in 122 821 vaccinated person-years (5.30 per 1000 person-years) compared with 483 events in 254 541 unvaccinated person-years (1.90 per 1000 person-years). With propensity score adjustment, we found no evidence for an association between pneumococcal vaccination and reduced risk of acute MI (adjusted hazard ratio [HR], 1.09; 95% confidence interval [CI], 0.98-1.21) or stroke (adjusted HR, 1.14; 95% CI, 1.00-1.31). An inverse association was also not found in men of different age and risk groups. The results appeared to be consistent, because using more specific International Classification of Diseases, Ninth Revision codes for the outcome definition did not change the estimations.
Conclusion Among a cohort of men aged 45 years or older, receipt of pneumococcal vaccine was not associated with subsequent reduced risk of acute MI and stroke.
Multiple studies have shown that vaccination against influenza can reduce the risk of recurrent myocardial infarction (MI), sudden cardiac death, cardiac hospital admissions, need for revascularization, and stroke.1- 5 A similar finding has been recently reported for pneumococcal polysaccharide vaccine.6 In the study by Lamontagne et al,6 the authors hypothesized that besides preventing bacterial infections, pneumococcal vaccination may protect against cardiovascular events by decreasing the extent of atherosclerosis. There were, however, several potential limitations of this study that raise questions about the validity of the results, including preferential inclusion of a healthier cohort, confounding from dietary factors, physical activity, and family history. Moreover, it is known that antibody levels decrease over time. This does not fit with the authors' finding that pneumococcal vaccination appeared to have a greater protective effect over time.
Beyond the study by Lamontagne et al,6 studies assessing the association between pneumococcal vaccination and vascular events are limited. To address this, we took advantage of information collected as part of an ongoing prospective cohort study, the California Men's Health Study (CMHS), to address the association between vaccination with pneumococcal vaccine and the risk of developing acute MI and stroke, taking into account known and potentially important confounders.
The CMHS is a multiethnic cohort of men who are members of the Kaiser Permanente Northern California and Kaiser Permanente Southern California health plans. Details of the study cohort design, recruitment methods, and data collection were described previously.7 In brief, 84 170 men, aged 45 to 69 years at baseline in January 2000, who had been members of the health plan for at least 1 year at recruitment, completed an extensive questionnaire. The CMHS cohort was similar to the population of health plan members and men who responded to a general health survey in California (California Health Interview Survey) on a variety of important demographic and clinical characteristics.7 Among the CMHS participants, from baseline through December 31, 2007, 3203 (3.8%) have died, 13 227 (15.7%) have terminated their membership at Kaiser Permanente, and 67 740 (80.5%) were active members.
The CMHS was reviewed and approved by the institutional review boards of Kaiser Permanente Northern California and Kaiser Permanente Southern California. Men were informed of the voluntary nature of their participation and that their decision about participation would not affect their medical care.
The comprehensive 24-page questionnaire solicited information on demographics, family history of cancer, health and lifestyle, existing health conditions, medication or drug use, physical activity, tobacco use, diet or supplement use, country of origin, duration of US residency, income, sexual orientation, and general health condition. Men of different race/ethnicity might have different uptake of the pneumococcal vaccine, as well as a different risk of developing MI, stroke, or both; therefore, race/ethnicity is a potential confounder in this case and was measured based on self-response to the first-stage survey questionnaire. Men were asked to identify all racial/ethnic groups that applied, including other (American Indian; Alaskan Native; Native Hawaiian; other Pacific Islanders; races specified by the participants other than white, black, Hispanic, and Asian; and mixed [multiple] races).
Diet was evaluated with a detailed semi-quantitative food-frequency questionnaire adapted from a questionnaire developed for the Women's Health Initiative and other studies and modified for men's health. Physical activity was assessed with questions adapted from the CARDIA Physical Activity History,7 which queried the men about the frequency and duration of their participation in specific types of moderate and vigorous recreational, household, and work-related activities. Tobacco smoking status was ascertained with the question, “Have you smoked at least 100 cigarettes in your lifetime?” For those participants who answered yes (ever-smokers), the questionnaire also asked the total number of years smoked and average number of cigarettes smoked per day. For former smokers, duration of quitting also was elicited.
Body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared. Participants who did not complete all pages of the food-frequency questionnaire (≥5 items per page) or who had total calculated energy intake of less than 800 kcal or more than 5000 kcal were excluded from the nutrient analyses. Vigorous activities were defined as participating in a minimum of 1260 metabolic equivalent task (MET) hours per week, on average, equivalent to at least 3.5 hours of activity, with a minimum MET level of 6. Moderate activities were defined as participating in a minimum of 630 MET hours of activity per week, on average, equivalent to at least 3.5 hours of activity, with a minimum MET level of 3.
Immunization records were tracked by the Kaiser Immunization Tracking System, which has been in place in the Kaiser Permanente California health plans since 1980. The Kaiser Immunization Tracking System contains information identifying the patient's unique medical record number, name, date of birth, and sex. It also contains information about the vaccination delivered, including the date of vaccination, type of vaccine, route and site of administration, person administering the vaccine, facility in which the vaccine was administered, dose, manufacturer, and lot number. Vaccines delivered outside the health plan or before enrollment are entered into the registry, with appropriate documentation to substantiate its delivery. All participants who had received at least 1 dose of pneumococcal polysaccharide vaccine before the baseline date or during the follow-up period were defined as the exposed group. For incident cases, only vaccines received 2 weeks or more before the incident event were considered exposed.
Acute MI and stroke in this cohort were identified from electronic health records. Medical encounters with the International Classification of Diseases, Ninth Revision (ICD-9) codes 410.xx (acute myocardial infarction) and 411.xx (other acute and subacute forms of ischemic heart diseases) were identified as acute MI cases, and ICD-9 codes 433.xx (occlusion and stenosis of precerebral arteries), 434.xx (occlusion of cerebral arteries), and 436.xx (acute, but ill-defined cerebrovascular diseases) were identified as stroke cases. Cases were also identified in mortality records by using the Group Cause of Death Codes (International Statistical Classification of Diseases, 10th Revision [ICD-10]). Death caused by acute MI (ICD-10 codes I21-I22) and other acute ischemic heart diseases (ICD-10 code I24) were included as acute MI cases, and deaths caused by cerebral infarction (ICD-10 code I63) and stroke, not specified as hemorrhage or infarction (ICD-10 code I64), were included as stroke cases. To identify cases, we performed computer matches on the unique Kaiser Permanente medical record number of incident acute MI and stroke cases between study baseline (2002-2003) and December 31, 2007, and the entire CMHS cohort. Incident events were defined as the first diagnosis of acute MI or stroke for each participant between 2002 and 2007, excluding those participants with a prior history of these outcomes.
The association between vaccination status and patient characteristics was assessed using the χ2 test for categorical factors or Kruskal-Wallis test for continuous factors. The association between vaccination and MI or stroke events was assessed using the Cox proportional hazards regression model, both in bivariate and multivariable models, the latter to adjust for the propensity score of vaccination with a pneumococcal vaccine estimated from a logistic regression model. The number of pneumococcal vaccines received was included in the models as a time-varying covariate. The number of vaccines received before study entry was included as the initial number of vaccines, and each vaccination during the follow-up period resulted in an increment of the number of vaccines administered. Follow-up started at the time of the baseline survey and ended at the time of acute MI or stroke diagnosis, termination of health plan membership, death, or December 31, 2007, whichever came first. Men who reported having had a previous acute MI (n = 6020) or stroke (n = 2849) at baseline were excluded from the regression analysis separately.
The propensity score was created using a logistic regression model predicting the probability of receiving at least 1 dose of pneumococcal vaccine during the study period. Quintiles of the propensity score were used to adjust for likelihood of vaccination in the models of MI and stroke outcomes. The variables included in the model were age, race/ethnicity, region (northern vs southern California Kaiser Permanente), household income, education, BMI, cigarette smoking, physical activity level, sedentary for more than 6.5 hours per day outside of work, alcohol consumption, number of influenza vaccines received, total calorie intake, fat intake, fruit and vegetable consumption, history of diabetes, history of high blood pressure, history of high cholesterol, history of peripheral artery disease, history of other heart diseases, history of stroke, history of acute MI, and the log scale transformed number of outpatient visits during the 5 years before baseline. Cigarette smoking was modeled by a set of variables that were created for a lung cancer study.8 The set of variables included smoking status (never, current, quit for ≤5 years, quit for 6-9 years, and quit for ≥10 years), duration of smoking (≤10, 11-20, 21-30, 31-40, and ≥41 years), and mean number of cigarettes smoked per day (≤10, 11-20, and ≥21 cigarettes) in all analyses. This model is detailed in the eTable.
To examine whether the association between pneumococcal vaccination and acute MI and stroke would vary among participants of different ages, risk groups, and number of influenza vaccination records, adjusted hazard ratios (HRs) and 95% confidence intervals (CIs) estimated by Cox proportional hazards regression models were presented separately for participants younger than 65 years and 65 years or older; 3 presumable high-risk groups, including (1) current smokers, (2) patients with history of diabetes, and (3) patients with history of hypertension, and 1 low-risk group, which included participants younger than 55 years, with no smoking history, no history of diabetes, and no history of hypertension; and participants receiving 0, 1 to 10, and more than 10 influenza vaccines. In addition, to examine whether the results were sensitive to the ICD-9 codes used for identifying outcomes, effects were also estimated using the more specific acute MI code 410.xx and stroke codes 433.xx and 434.xx.
Given the sample size and risk of developing MI and stroke in the unvaccinated group (2.0% and 0.7%, respectively), we were able to detect true relative risk of 0.9 for MI and 0.8 for stroke in the overall cohort, with 80% power and type I error rate at .05. The significance level was set at .05 based on a 2-sided test. SAS version 9.1 (SAS Institute Inc, Cary, North Carolina) was used for all statistical analyses.
The distribution of baseline characteristics by vaccination status is shown in Table 1. A total of 2705 incident acute MI (3.2%), including 46 mortality cases with no medical encounter records (incidence rate, 7.53 events per 1000 person-years; 95% CI, 7.25-7.82 per 1000 person-years) and 1134 stroke cases (1.3%), including 20 mortality cases with no medical encounter records (incidence rate, 3.01 events per 1000 person-years; 95% CI, 2.83-3.19 per 1000 person-years) were identified. The mean (SD) length of follow-up was 4.7 (1.36) years. The unvaccinated group had relatively shorter length of follow-up. The development of acute MI and stroke was positively associated with increased number of pneumococcal vaccines (for MI: 2.0% in the unvaccinated group vs 4.6%, 5.5%, 5.3% in the vaccinated group for 1, 2, and ≥3 vaccine doses; and for stroke: 0.7% in the unvaccinated group vs 2.1%, 2.8%, 3.0% in the vaccinated group for 1, 2, and ≥3 vaccine doses). Participants who were vaccinated were significantly older than participants who were not vaccinated. Region, race/ethnicity, household income, education, and BMI also were associated significantly with vaccination status.
Table 2 shows potential lifestyle risk factors for acute MI and stroke by number of pneumococcal vaccines received at any time. The distribution of cigarette smoking variables, alcohol consumption, daily energy intake, total fat intake, daily vegetables and fruit consumption, physical activity level, and daily sedentary hours were all significantly different among participants with varying number of vaccine doses. The unvaccinated group included a higher proportion of nonsmokers (46.7% vs 37.2% in the vaccinated group) and men who engaged in vigorous activities (18.0% vs 14.1% in the vaccinated group).
In addition, the proportion of participants having a history of diabetes (6.2% vs 20.6%), hypertension (30.3% vs 45.2%), high cholesterol (35.8% vs 46.5%), congestive heart failure (1.6% vs 3.9%), peripheral artery disease (3.9% vs 7.7%), and other heart diseases (14.1% vs 23.4%) were significantly lower in the unvaccinated group (Table 3). The mean numbers of outpatient visits and inpatient visits 5 years before follow-up were also significantly lower in the unvaccinated group.
During follow-up, there were 1211 first MIs in 112 837 vaccinated person-years (10.73 per 1000 person-years) compared with 1494 first MI events in 246 170 unvaccinated person-years (6.07 per 1000 person-years). For stroke, there were 651 events in 122 821 vaccinated person-years (5.30 per 1000 person-years) compared with 483 events in 254 541 unvaccinated person-years (1.90 per 1000 person-years). With adjustment for the propensity score, we found no evidence for an association between pneumococcal vaccination and reduced risk of acute MI (adjusted HR, 1.09; 95% CI, 0.98-1.21) or stroke (adjusted HR, 1.14; 95% CI, 1.00-1.31) (Table 4). In addition, association was not observed either in the current smokers, men with history of diabetes, men with history of hypertension, or men of the low-risk group. The results appeared to be robust because using more specific codes for acute MI and stroke did not change the estimates. Adjusted HRs were 1.24 (95% CI, 1.04-1.48) if ICD-9 code 410.xx was used to define acute MI and 1.17 (95% CI, 0.95-1.45) if ICD-9 codes 433.xx and 434.xx were used to define stroke.
Our cohort study assessed the potential association between pneumococcal vaccine and the reduced risk of acute MI and stroke in an ethnically and socioeconomically diverse male population. In contrast with the findings of a case-control study by Lamontagne et al,6 our data did not support the protective role of pneumococcal vaccine against acute MI and stroke. Two major methodological distinctions may explain the difference. First, many possible confounding factors were not considered in the case-control study because it was limited to data found in administrative databases.9- 11 In our study, dietary factors, disease history, and lifestyle factors such as cigarette smoking and physical activity level were comprehensively ascertained and adjusted for in the analyses. Second, the controls chosen by Lamontagne et al were criticized as being likely to be a healthier group.10 The cohort design protects against selection bias and the prospective ascertainment of relevant exposure factors either from questionnaire data or electronic clinical records protects against biases related to recall. After controlling for potential confounding factors and eliminating possible selection bias, the protective role of pnemococcal vaccine against acute MI and stroke was not observed among men older than 45 years. The results of several sensitivity analyses, including subgroup analyses of men of different age and risk groups or analysis using more specific codes to identify outcomes, also showed no evidence of an inverse association.
Pneumococcal vaccine is recommended to adults aged 65 years or older and to persons who have certain underlying medical conditions that may increase the risk for pneumococcal infection. Such conditions include chronic cardiovascular diseases (eg, congestive heart failure or cardiomyopathy), chronic pulmonary diseases (eg, chronic obstructive pulmonary disease or emphysema), or chronic liver diseases (eg, cirrhosis), and diabetes mellitus.12 Therefore, it is not surprising to find that the vaccinated population was generally older and had higher prevalences of chronic conditions. The observation that the unadjusted HRs for pneumococcal vaccination were significantly more than 1 for both acute MI and stroke is consistent with the clinical guidelines. The modestly increased risk of MI or stroke in men younger than 65 years is probably due to residual confounding, because no evidence suggests that the vaccine would increase the risk of vascular events, particularly in men in this age category. With additional adjustment of risk factors of MI or stroke, such as smoking, diabetes, and history of heart diseases in the Cox proportional hazards regression model in the younger than 65 years age group, the HR associated with vaccination reduced to 1.11 (95% CI, 0.98-1.29) for MI and 1.24 (95% CI, 1.05-1.50) for stroke. Such adjustment does not affect the estimate in the 65 years or older age group.
In a recent systematic review and meta-analysis by Huss et al,13 the authors found a high degree of heterogeneity between trials in the efficacy of pneumococcal polysaccharide vaccine in the prevention of a range of clinical outcomes. They found little evidence of protection against pneumonia among elderly persons or adults with chronic respiratory illness, for whom the pneumococcal vaccine is recommended in many industrialized countries. Trials of higher quality (ie, those with a double-blind design and adequate concealment of allocation) generally showed little evidence of a protective vaccine effect, regardless of the study population and setting. Although these findings differed from a recent Cochrane review,14 which found strong evidence supporting the vaccine's efficacy against invasive pneumococcal disease and reported a combined odds ratio (OR) of 0.26 (95% CI, 0.15-0.46), the authors suggested that the inconsistency could be largely explained by the inclusion of 2 earlier studies with inadequate randomization and whose participants had limited access to care and diagnostic procedures in the Cochrane review.13 If pneumococcal polysaccharide vaccine is not as effective as previously believed in preventing infection and its complications, any putative effect of preventing infection-triggered acute MI or stroke becomes unlikely.
Our findings are consistent with those reported by Smeeth et al,1 which showed within-person comparisons using the case-series method to study the risks of MI and stroke after common vaccinations and naturally occurring infections. The authors concluded that acute infections are associated with a transient increase in the risk of vascular events. However, influenza, tetanus, and pneumococcal vaccinations do not produce a detectable increase in the risk of vascular events. Combined with the findings from our study, it appears that both short- and long-term risks of acute MI and stroke are not affected by pneumococcal vaccination.
Pneumococcal vaccination was also found to have no effect on MI in another case-control study.15 In the study by Meyers et al, the authors administered a standardized questionnaire to 335 patients with MI and 199 patients with fractures. The groups significantly differed by sex, age, BMI, smoking status, family history of heart disease, personal history of cardiovascular disease, and number of self-reported upper respiratory tract infections. Pneumococcal vaccine had been administered to 32% of patients with MI and 39% of patients with fractures, with an adjusted OR of 0.89 (95% CI, 0.60-1.33). The authors concluded that pneumococcal vaccine is not associated with a reduced risk of MI.
The health care delivery systems provide a well-developed infrastructure for conducting a prospective cohort study. The majority of health care for Kaiser Permanente members is delivered within an integrated system of Kaiser Permanente owned and operated medical centers and outpatient facilities. Minimal co-payments are a strong incentive to receive care within the system. Care provided outside the system either through contractual arrangements or for emergency care are captured in outside referral and claims reimbursement data systems. Importantly, for outside clinicians to be reimbursed by the health plan for covered emergent care, claims have to be submitted with documentation of the episode of care, and this information is entered into the administrative data systems. Thus, the capture of care delivered to members is reasonably comprehensive.
One limitation of our study is that we relied on diagnostic codes in the electronic record for acute MI and stroke and on electronic data for exposure. However, although misclassification might exist, it is mostly likely minimal and nondifferential. It seems unlikely that misclassification would introduce a systematic bias on the basis of either exposure or outcome. In addition, the association between this vaccine and acute MI and stroke in women or in very old populations was not addressed in our study.
In conclusion, among a cohort of men aged 45 years or older, receipt of pneumococcal vaccine was not associated with subsequent reduced risk of acute MI and stroke, after accounting for baseline differences in those participants receiving vaccine vs not receiving vaccine.
Corresponding Author: Hung Fu Tseng, PhD, Department of Research and Evaluation, Southern California Permanente Medical Group, Kaiser Permanente, 100 S Los Robles Ave, Second Floor, Pasadena, CA 91101 (firstname.lastname@example.org).
Author Contributions: Dr Tseng had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Tseng, Quinn.
Acquisition of data: Slezak, Quinn, Van Den Eeden.
Analysis and interpretation of data: Tseng, Slezak, Quinn, Sy, Van Den Eeden, Jacobsen.
Drafting of the manuscript: Tseng, Quinn.
Critical revision of the manuscript for important intellectual content: Tseng, Slezak, Quinn, Sy, Van Den Eeden, Jacobsen.
Statistical analysis: Tseng, Slezak.
Obtained funding: Quinn.
Administrative, technical, or material support: Tseng, Quinn, Van Den Eeden, Jacobsen.
Study supervision: Tseng, Jacobsen.
Financial Disclosures: Dr Tseng reported receiving research funding for other vaccine studies from Merck. Mr Slezak reported receiving research funding for other vaccine studies from Merck. Ms Sy reported receiving research funding for other vaccine studies from Merck and was an employee of Merck between 2004 and 2007. Dr Van Den Eeden reported receiving funding for a non−vaccine-related study from GlaxoSmithKline. Dr Jacobsen reported receiving research funding for other vaccine studies and served as an unpaid consultant for Merck Research Laboratories. Dr Quinn reported no financial disclosures.
Funding/Support: This study was funded by California Cancer Research Program and Kaiser Permanente.
Role of the Sponsors: California Cancer Research Program and Kaiser Permanente had no role in the design and conduct of the study, in the collection, analysis, and interpretation of the data, or in the preparation, review, or approval of the manuscript.