The aim of our study was to determine whether pre-emptive statin therapy was associated with improved outcome of infective endocarditis (IE). We conducted a nationwide, population-based, propensity score-matched cohort study with the Taiwan’s National Health Insurance Research Database. All patients with IE between January 2000 and December 2010 were enrolled. The primary outcome was in-hospital mortality. The secondary outcome included all-cause mortality within the first 3 months, 6 months, and one year after the diagnosis of IE. Among 13,584 patients with IE, we applied propensity score-matching on a 1:4 ratio, in which 370 statin users were matched to 1,480 statin non-users. Compared with statin non-users, statin users had a significantly lower risk of in-hospital mortality (adjusted hazard ratio [aHR] 0.65, 95% confidence interval [CI], 0.49–0.86). The reduction in mortality from IE remained significant for follow-up 3 months (aHR 0.68, 95% CI, 0.53–0.88), 6 months (aHR 0.73, 95% CI, 0.58–0.91), and 12 months (aHR 0.68, 95% CI, 0.55–0.84). Statin therapy was associated with a reduced risk of ICU admission rates, shock events, the need for mechanical ventilation, but not significantly with the need for heart valvular replacement surgery. In conclusion, our study found that statin therapy is associated with a reduced risk of in-hospital and subsequent mortality of IE.
A retrospective cohort of 283 patients with a diagnosis of IE conducted by Anavekar and colleagues first reported prior statin therapy were associated with reduced subsequent embolic events, but not significantly with reduced risks of mortality. However, the independent association of statin use on mortality may be limited by small sample size, inclusion and the relatively short follow-up period. Thus, whether subjects complicated with IE before exposure of statin therapy gain the benefits of mortality at different time points in clinical practice need to be elucidated. Above-mentioned observations prompted us to test the hypothesis whether prior statin therapy could improve the outcome of IE. To reduce the impact of potential confounding elements stemming from inadequate controls for comorbidities, we conducted a nationwide population-based, propensity score-matched study of patients with IE by using Taiwan’s National Health Insurance Research Database (NHIRD). The aim of this study was to evaluate whether statins influence subsequent adverse consequence or outcome in patients first admitted to the hospital with IE who were taking statins before hospital admission.
Method
In the current study, we used the National Health insurance Research database. In Taiwan, the National Health insurance (NHI) program was launched since 1995, which covers 99% of the population of 23 million people now. The NHI is a mandatory universal health insurance program, offering comprehensive medical care coverage, including outpatient, inpatient, emergency, dental, traditional Chinese medicine services, and prescription drugs. In 1999, the Bureau of National Health Insurance began to release patient data in electronic form. The multiple NHI databases provide comprehensive utilization and enrollment information for all patients under the NHI program. The diseases were coded according to the International Classification of Disease, Ninth Revision, Clinical Modification (ICD-9-CM) diagnosis codes, 2001 edition. The accuracy of diagnoses in the NHIRD has been validated for several diseases. All information that would potentially expose a specific individual patient to be identified has been encrypted. The confidentiality of the data abides by the data regulations of the Bureau of National Health Insurance and the National Health Research Institute.
We designed a nationwide population-based, observational retrospective cohort study in Taiwan to determine the association between the prior statin use and the mortality in patients with IE. The study cohort comprised of all patients who had hospitalization with the diagnosis of IE and receiving antibiotic treatments between January 2000 and December 2010. Patients entered the cohort on the first day of hospitalization of IE and were followed up to death, lost to follow-up, or until December 31, 2011, whichever came first.
Baseline demographic data included age, sex, income, and urbanization. Charlson Comorbidity Index (CCI) score was used to determine overall systemic health. Other systemic diseases not included in the CCI and concomitant medications associated with anti-inflammatory effects, patient’s physical condition and underlying immune condition were extracted were also examined. We also identified patients who had a history of receiving heart valulvar replacement surgery. The statin users were defined as those who had received the continued prescription of statins for ≥ 30 days before the index date.
In order to reduce selection bias, we performed 1:4 case control matching analysis. For each statin user, we identified one such control patient with the most similar demographic characteristics, which were matched according to propensity score (±0.1 score) for the likelihood of statin use that was calculated from baseline covariates by using multivariate logistic regression analysis.
The primary outcome was in-hospital mortality. The secondary outcome was all-cause mortality within the first 3 months, 6 months, and one year of diagnosis of IE. Other adverse consequences including intensive care unit (ICU) admission rates, shock events, the need for mechanical ventilation and heart valvular replacement surgery during hospitalization were also included in our analysis.
Descriptive statistics were used to describe the baseline characteristics of our cohort. Baseline characteristics of the two groups were compared using Pearson χ 2 tests for categorical variables; the independent t-test for parametric continuous variables. The propensity score for the likelihood of statin use was calculated by multivariate logistic regression analysis, conditional on the baseline covariates in Table 1 . Cox regression models were used to calculate the adjusted hazard ratio (HR) and 95% confidence intervals (CI) for the association between statin use and mortality. We also used multivariate logistic regression to calculate adjusted odds ratios (aOR) and 95% CI for the association between statin use and ICU admission rates, shock events, the need for mechanical ventilation and heart valvular replacement surgery. Adjustments were made for clinically relevant variables and for those that showed a statistically significant difference between the two groups at baseline. Finally, we performed subgroup analyses based on age, sex, CCI score, cerebrovascular disease, using antiplatelet drugs, and receiving heart valvular replacement surgery. Tests of interactions were performed for those subgroups by the likelihood ratio test. Microsoft SQL Server 2012 (Microsoft Corp., Redmond, Washington, USA) was used for data linkage, processing, and sampling. The propensity score and adjusted absolute incidence rate was calculated with SAS version 12.0 (SAS Campus Drive, Cary, North Carolina, USA). All other statistical analyses were conducted using STATA statistical software (version 12.0; StataCorp., Texas, USA). Statistical significance was defined as a p value of <0.05.
| Characteristic | Before Propensity Score-Matched | Propensity Score-Matched | ||||
|---|---|---|---|---|---|---|
| Statin Users | Non-Users | p Value | Statin Users | Matched Non-Users | p Value | |
| Patient (no.) | 370 | 13,214 | 370 | 1,480 | ||
| Mean age (SD), year | 64.9 (12.7) | 56.6 (19.2) | <0.001 | 64.9 (12.7) | 65.0 (13.7) | 0.879 |
| Male | 193 (52%) | 8,651 (66%) | <0.001 | 193 (52%) | 762 (52%) | 0.816 |
| Income | <0.001 | 0.998 | ||||
| Dependent | 128 (35%) | 3,285 (25%) | 128 (35%) | 509 (34%) | ||
| NT$ <19,100 | 72 (19%) | 4,366 (33%) | 72 (19%) | 294 (20%) | ||
| NT$ 19,100–42,000 | 154 (42%) | 5,104 (39%) | 154 (42%) | 615 (42%) | ||
| >NT$ 42,000 | 16 (4.3%) | 459 (3.5%) | 16 (4.3%) | 62 (4.2%) | ||
| Urbanization ∗ | 0.003 | 0.976 | ||||
| Level 1 | 211 (57%) | 6557 (50%) | 211 (57%) | 832 (56%) | ||
| Level 2 | 115 (31%) | 5290 (40%) | 115 (31%) | 466 (32%) | ||
| Level 3 | 34 (9.2%) | 1,151 (8.7%) | 34 (9.2%) | 136 (9.2%) | ||
| Level 4 (rural area) | 10 (2.7%) | 216 (1.6%) | 10 (2.7%) | 46 (3.1%) | ||
| Charlson comorbidity score † | <0.001 | 0.890 | ||||
| 0 score | 3 (0.8%) | 1,595 (12%) | 3 (0.8%) | 18 (1.2%) | ||
| 1 score | 13 (3.5%) | 1,960 (15%) | 13 (3.5%) | 61 (4.1%) | ||
| 2 score | 30 (8.1%) | 1,759 (13%) | 30 (8.1%) | 116 (7.8%) | ||
| 3 score | 36 (9.7%) | 1,531 (12%) | 36 (9.7%) | 128 (8.6%) | ||
| ≥4 score | 288 (78%) | 6,369 (48%) | 288 (78%) | 1,157 (78%) | ||
| Concomitant medications | ||||||
| Aspirin | 155 (42%) | 1,796 (14%) | <0.001 | 155 (42%) | 622 (42%) | 0.962 |
| Clopidogrel | 30 (8.1%) | 201 (1.5%) | <0.001 | 30 (8.1%) | 94 (6.4%) | 0.227 |
| Ticolpidine | 11 (3.0%) | 81 (0.6%) | <0.001 | 11 (3.0%) | 43 (2.9%) | 0.945 |
| Cilostazole | 6 (1.6%) | 66 (0.5%) | 0.003 | 6 (1.6%) | 30 (2.0%) | 0.614 |
| Warfarin | 26 (7.0%) | 609 (4.6%) | <0.001 | 26 (7.0%) | 116 (7.8%) | 0.600 |
| Dipyridamole | 33 (8.9%) | 560 (4.2%) | <0.001 | 33 (8.9%) | 123 (8.3%) | 0.707 |
| ACE inhibitor or ARB | 132 (36%) | 1,617 (12%) | <0.001 | 132 (36%) | 517 (35%) | 0.789 |
| Beta blocker | 60 (16%) | 1,174 (8.9%) | <0.001 | 60 (16%) | 237 (16%) | 0.924 |
| Calcium-channel blocker | 152 (41%) | 1,917 (15%) | <0.001 | 152 (41%) | 586 (40%) | 0.601 |
| Hypoglycemic drug | 150 (41%) | 1,345 (10%) | <0.001 | 150 (41%) | 592 (40%) | 0.850 |
| Coexisting conditions | ||||||
| Cerebrovascular disease | 188 (51%) | 4,570 (35%) | <0.001 | 188 (51%) | 727 (49%) | 0.561 |
| Receiving heart valvular replacement surgery | 26 (7.0%) | 673 (5.1%) | 0.097 | 26 (7.0%) | 95 (6.4%) | 0.672 |
| Hypertension | 339 (92%) | 7,144 (54%) | <0.001 | 339 (92%) | 1,356 (92%) | 1.000 |
| Myocardial infarction | 74 (20%) | 911 (6.9%) | <0.001 | 74 (20%) | 269 (18%) | 0.419 |
| Coronary artery disease | 269 (73%) | 5,082 (38%) | <0.001 | 269 (73%) | 1,078 (73%) | 0.958 |
| Chronic pulmonary disease | 209 (56%) | 5,891 (45%) | <0.001 | 209 (56%) | 811 (55%) | 0.559 |
| Asthma | 109 (29%) | 2,617 (20%) | <0.001 | 109 (29%) | 424 (29%) | 0.758 |
| Heart failure | 186 (50%) | 4,746 (36%) | <0.001 | 186 (50%) | 745 (50%) | 0.981 |
| Valvular heart disease | 161 (44%) | 5,806 (44%) | 0.871 | 161 (44%) | 630 (43%) | 0.742 |
| Atrial fibrillation | 51 (14%) | 1,561 (12%) | 0.248 | 51 (14%) | 233 (16%) | 0.350 |
| Peripheral vascular disease | 91 (25%) | 1,253 (9.5%) | <0.001 | 91 (25%) | 336 (23%) | 0.440 |
| Dyslipidemia | 322 (87%) | 3,434 (26%) | <0.001 | 322 (87%) | 1,316 (89%) | 0.307 |
| DM | 271 (73%) | 4,390 (33%) | <0.001 | 271 (73%) | 1,087 (73%) | 0.937 |
| DM, complicated with organ failure | 154 (42%) | 1,741 (13%) | <0.001 | 154 (42%) | 640 (43%) | 0.573 |
| Rheumatoid disease | 37 (10%) | 870 (6.6%) | 0.009 | 37 (10%) | 159 (11%) | 0.678 |
| Cancer | 64 (17%) | 2,113 (16%) | 0.499 | 64 (17%) | 251 (17%) | 0.877 |
| AIDS | 1 (0.3%) | 165 (1.2%) | 0.091 | 1 (0.3%) | 3 (0.2%) | 0.802 |
| Chronic liver disease | 113 (31%) | 4,068 (31%) | 0.920 | 113 (31%) | 477 (32%) | 0.533 |
| Chronic renal disease | 171 (46%) | 3,487 (26%) | <0.001 | 171 (46%) | 717 (48%) | 0.443 |
| Chronic dialysis | 12 (3.2%) | 176 (1.3%) | 0.002 | 12 (3.2%) | 58 (3.9%) | 0.542 |
| Drug abuse | 11 (3.0%) | 1,284 (9.7%) | <0.001 | 11 (3.0%) | 34 (2.3%) | 0.450 |
| Propensity score (SD) | 0.113 (0.048–0.227) | 0.004 (0.002–0.018) | <0.001 | 0.113 (0.048–0.227) | 0.113 (0.048–0.227) | 0.999 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
