Factors influencing the indication of coronary angiography in patients presenting with chest pain unspecified: an analysis of two decades (1994-2014).

Guidelines for cardiac catheterization in patients with non-specific chest pain (NSCP) provide significant room for provider discretion, which has resulted in variability in the utilization of invasive coronary angiograms (CAs) and a high rate of normal angiograms. The overutilization of CAs in patients with NSCP and discharged without a diagnosis of coronary artery disease is an important issue in medical care quality. As a result, we sought to identify patient demographic, socioeconomic, and geographic factors that influenced the performance of a CA in patients with NSCP who were discharged without a diagnosis of coronary artery disease. We intended to establish reference data points for gauging the success of new initiatives for the evaluation of this patient population. In this 20-year retrospective cohort study (1994-2014), we examined 107 796 patients with NSCP from the Myocardial Infarction Data Acquisition System, a large statewide validated database that contains discharge data for all patients with cardiovascular disease admitted to every non-federal hospital in NJ. Patients were partitioned into two groups: those offered a CA (CA group; n = 12 541) and those that were not (No-CA group; n = 95 255). Geographic, demographic, and socioeconomic variables were compared between the two groups using multivariable logistic regression, which determined the predictive value of each categorical variable on the odds of receiving a CA. Whites were more likely than Blacks and other racial counterparts (19.7% vs. 5.6% and 16.5%, respectively; P < .001) to receive a CA. Geographically, patients who received a CA were more likely admitted to a large hospital compared to small- or medium-sized ones (12.5% vs. 8.9% and 9.7%, respectively; P < .05), a primary teaching institution rather than a teaching affiliate or community center (16.1 % vs. 14.3% and 9.1%, respectively; P < .001), and at a non-rural facility compared to a rural one (12.1% vs. 6.5%; P < .001). Lastly from a socioeconomic standpoint, patients with commercial insurance more often received a CA compared to those having Medicare or Medicaid/self-pay (13.7% vs. 9.5% and 6.0%, respectively; P < .001). The utilization of CA in patients with NSCP discharged without a diagnosis of coronary artery disease in NJ during the study period may be explained by differences in geographic, demographic, and socioeconomic factors. Patients with NSCP should be well scrutinized for CA eligibility, and reliable strategies are needed to reduce discretionary medical decisions and improve quality of care.

US Initiative to Eliminate Racial and Ethnic Disparities in Health: The Impact on the Outcomes of ST-Segment-Elevation Myocardial Infarction in New Jersey.

Background In 1998, President Clinton launched a federal initiative to eliminate racial and ethnic health disparities. The impact on the outcomes of ST-segment-elevation myocardial infarction has not been well studied. Methods and Results ST-segment-elevation myocardial infarction outcomes from 1994 to 2015 were studied in 7942 Black, 27 665 Hispanic, and 88 727 White patients with first admission of ST-segment-elevation myocardial infarction using the Myocardial Infarction Data Acquisition System. Logistic regressions were used to assess mortality adjusting for demographics, comorbidities, and interventional procedures. There was an overall rise from 1994 to 2015 in the use of percutaneous coronary interventions in all 3 groups. Before 1998, White patients received more percutaneous coronary interventions compared with Black and Hispanic patients (P<0.05). After 1998, the disparity in use of percutaneous coronary interventions in Black and Hispanic patients was greatly reduced compared with White patients, and the difference reversed in favor of Hispanic patients after 2005 (P<0.05). There was an overall downward trend of in-hospital mortality without evidence of disparity among Black, Hispanic, and White patients. A linear regression model was used with a change point in 1998. Before 1998, the slope of 1-year all-cause and cardiovascular mortality was not statistically significant. After 1998, the mortality showed negative slopes for all 3 groups, however, with lower overall crude mortality for Hispanic patients compared with Black and White patients (P<0.0001). Conclusions The initiative launched in 1998 may have contributed to a reduction in percutaneous coronary intervention usage disparity in patients with ST-segment-elevation myocardial infarction. Short- and long-term mortality decreased in all 3 groups, but more in the Hispanic population.

Monotherapy treatment with chlorthalidone or amlodipine in the systolic blood pressure intervention trial (SPRINT).

This post hoc analysis of the Systolic Blood Pressure Intervention Trial (SPRINT) examined the performance of chlorthalidone (C) versus amlodipine (A) monotherapies. ANOVA was used to analyze the differences in systolic blood pressure (SBP) response between C and A. Logistic regression was used to examine monotherapy failure (adding a second antihypertensive agent or switching to a different antihypertensive agent) rates. Four hundred ninety-one participants were treated with C monotherapy (n = 210, mean dose = 22 mg/day) or A monotherapy (n = 281, mean dose = 7 mg/day). There was a significant difference in mean SBP reduction between the C and A monotherapies at the third visit (higher reduction with A, adjusted p = .018). Unadjusted analysis showed a higher failure with C in the standard treatment group. Although the average SBP at failure was higher and above the 140 mm Hg cutoff that indicated monotherapy failure with A (142.60) compared with C (138.40), more participants on C failed despite having SBP below the 140 cutoff. This was probably due to decisions made by the investigative teams to change the antihypertensive regimen, because, in their opinion, the clinical picture required it. After adjusting for baseline characteristics, C had higher failure than A only in the standard treatment group (1.64 odds ratio [OR], 95% CI 1.06-2.56, p = .028). A sub-analysis including participants who had never used antihypertensive treatment before randomization had similar results (2.57 OR, 95% CI 1.34-5.02, p = .004). Overall, in SPRINT chlorthalidone was associated with higher monotherapy failure than amlodipine in the standard treatment group because of decisions of the investigative teams.

A Twenty-Year Analysis of Demographics, Surgical Management, and Outcomes of Aortic Stenosis in New Jersey.

We investigated the incidence and characteristics of 14,996 patients with aortic stenosis (AS) who were hospitalized in New Jersey between the years 1995 to 2015. The average age was 72, the majority were Caucasian males and common co-morbidities were hypertension, coronary artery disease and hypercholesterolemia. Hospital admission for AS declined between 1995 to 2007, to 10/100,000 patients, and increased to 15/100,000 patients in 2015 (p for trend <0.001). During the study period, the percentage of patients who received aortic valve replacement (AVR) increased (p <0.001). All-cause and cardiovascular mortality were higher among patients who did not undergo AVR at 1-year (HR 1.98 CI 1.75 to 2.23, p <0.001 and HR 1.82 CI 1.57 to 2.11, p <0.001, respectively) and 3-years (HR 2.16 CI 1.96 to 2.38, p <0.001 and HR 2.16 CI 1.90 to 2.45, p <0.001, respectively). The probability for readmission for AS was higher in patients who did not receive AVR compared to patients who had AVR at 1 year (HR 92.95 CI 57.85 to 149.35, p <0.001) and 3 years (HR 70.36 CI 47.18 to 104.95, p <0.001). These data imply that earlier diagnosis of AS and AVR when indicated will improve outcomes.

Constrictive Pericarditis after Open Heart Surgery: A 20-Year Case Controlled Study.

BACKGROUND: Constrictive pericarditis is a rare complication of open heart surgery (OHS), but little is known regarding the etiologic determinants, and prognostic factors. The purpose of this study was to investigate clinical predictors and long term prognosis of post-operative constrictive pericarditis (CP). METHODS: Using the Myocardial Infarction Data Acquisition System database, we analyzed records of 142,837 patients who were admitted for OHS in New Jersey hospitals between 1995 and 2015. Ninety-one patients were hospitalized with CP 30 days or longer after discharge from OHS. Differences in proportions were analyzed using Chi square tests. Controls were matched to cases for demographics, surgical procedure type, history of OHS, and propensity score. Cox proportional hazard models were used to evaluate the risk of all-cause death. Log-rank tests and Cox models were used to assess differences in the Kaplan-Meier survival curves with and without adjustments for comorbidities. RESULTS: Patients with CP were more likely to have history of valve disease (VD, p < 0.001), atrial fibrillation (AF, p = 0.024) renal disease (CKD, p = 0.028), hemodialysis (HD, p = 0.008), previous OHS (p < 0.001). Patients with CP compared to matched controls had a higher 7-year mortality (p < 0.001). This difference became statistically significant at 1-year after surgery. CONCLUSION: CP is a rare complication of OHS that occurs more frequently in patients with VD, AF, CKD, HD, multiple OHS, and it is associated with an unfavorable long-term prognosis. Given the large number of OHS performed every year, the results highlight the need for clinicians to recognize and properly manage this complication of OHS.

Impact of the COVID-19 pandemic on interventional cardiology fellowship training in the New York metropolitan area: A perspective from the United States epicenter.

BACKGROUND: The healthcare burden posed by the coronavirus disease 2019 (COVID-19) pandemic in the New York Metropolitan area has necessitated the postponement of elective procedures resulting in a marked reduction in cardiac catheterization laboratory (CCL) volumes with a potential to impact interventional cardiology (IC) fellowship training. METHODS: We conducted a web-based survey sent electronically to 21 Accreditation Council for Graduate Medical Education accredited IC fellowship program directors (PDs) and their respective fellows. RESULTS: Fourteen programs (67%) responded to the survey and all acknowledged a significant decrease in CCL procedural volumes. More than half of the PDs reported part of their CCL being converted to inpatient units and IC fellows being redeployed to COVID-19 related duties. More than two-thirds of PDs believed that the COVID-19 pandemic would have a moderate (57%) or severe (14%) adverse impact on IC fellowship training, and 21% of the PDs expected their current fellows' average percutaneous coronary intervention (PCI) volume to be below 250. Of 25 IC fellow respondents, 95% expressed concern that the pandemic would have a moderate (72%) or severe (24%) adverse impact on their fellowship training, and nearly one-fourth of fellows reported performing fewer than 250 PCIs as of March 1st. Finally, roughly one-third of PDs and IC fellows felt that there should be consideration of an extension of fellowship training or a period of early career mentorship after fellowship. CONCLUSIONS: The COVID-19 pandemic has caused a significant reduction in CCL procedural volumes that is impacting IC fellowship training in the NY metropolitan area. These results should inform professional societies and accreditation bodies to offer tailored opportunities for remediation of affected trainees.

Trends in Hospitalization for Infective Endocarditis as a Reason for Admission or a Secondary Diagnosis.

We postulate that the trends for infective endocarditis (IE) are different for patients admitted for this condition compared with those admitted for a different reason with IE as a secondary diagnosis. Using the Myocardial Infarction Data Acquisition System (MIDAS) database, we analyzed 21,443 records of patients hospitalized with diagnosis of IE from 1994 to 2015. There were 9,191 patients hospitalized with IE as the primary diagnosis, and 12,252 patients with IE as a secondary diagnosis. Piecewise linear models were used to detect changes in trends. A bootstrap method was used to assess the statistical significance of the slopes and break point of each model. Differences in co-morbidities and microbiological patterns were analyzed. Trend analysis showed a significant decrease in IE as the primary diagnosis starting in the year 2004 (p <0.01). Hospitalizations with IE as a secondary diagnosis showed a linear increase in incidence (p <0.001), without any change points. In primary diagnosis IE, the proportion of streptococci as a causative microorganism was higher compared with staphylococci (p <0.001). On the contrary, in secondary diagnosis IE, the proportion of staphylococci was higher than streptococci (p <0.001). The proportion of gram-negative and other organism IE was similar in both groups. In conclusion, this study showed 2 divergent temporal trends in hospitalizations for IE as a primary or secondary diagnosis starting in 2004. The profile of the microorganisms reveals a steady higher proportion of staphylococcal infection in secondary diagnosis IE compared with streptococcal infection. Different strategies are needed for the prevention of IE.

Association of orthostatic hypertension with mortality in the Systolic Hypertension in the Elderly Program.

We examined the association of orthostatic hypertension with all-cause mortality in the active treatment and placebo randomized groups of the Systolic Hypertension in the Elderly Program (SHEP). SHEP was a multicenter, randomized, double-blind, placebo-controlled clinical trial of the effect of chlorthalidone-based antihypertensive treatment on the rate of occurrence of stroke among older persons with isolated systolic hypertension (ISH). Men and women aged 60 years and above with ISH defined by a systolic blood pressure (SBP) of 160 mm Hg or higher and diastolic blood pressure lower than 90 mm Hg were randomized to chlorthalidone-based stepped care therapy or matching placebo. Among 4736 SHEP participants, 4073 had a normal orthostatic response, 203 had orthostatic hypertension, and 438 had orthostatic hypotension. Compared with normal response, orthostatic hypertension was associated with higher all-cause mortality at 4.5 and 17 years in analyses adjusted for age, gender, treatment, SBP, and pulse pressure (PP, HR 1.87, 95% CI 1.30-2.69, p = 0.0007; HR 1.40, 95% CI 1.17-1.68, p = 0.0003, respectively). These associations remained significant after additional adjustment for risk factors and comorbidities (HR 1.43, 95% CI 0.99-0.08, p = 0.0566 at 4.5 years, and HR 1.27, 95% CI 1.06-1.53, p = 0.0096 at 17 years). The increased risk of all-cause mortality associated with orthostatic hypertension was observed in both the active and placebo groups without significant interaction between randomization group and the effect on mortality. Orthostatic hypertension is associated with future mortality risk, is easily detected, and can be used in refining cardiovascular risk assessment.

Frequency of Stroke After Percutaneous Coronary Intervention or Coronary Artery Bypass Grafting (from an Eleven-Year Statewide Analysis).

We compared stroke rates associated with coronary artery bypass grafting (CABG), both on-pump and off-pump, and percutaneous coronary intervention (PCI) with both drug-eluting stent (DES) and bare-metal stent (BMS) and the impact on 30-day and 1-year all-cause mortality. The Myocardial Infarction Data Acquisition System database was used to study patients who had on-pump CABG (n = 47,254), off-pump CABG (n = 19,118), and PCI with BMS (n = 46,641), and DES (n = 115,942) in New Jersey from 2002 to 2012. Multiple logistic and Cox proportional hazard models were used to compare the risk of stroke and mortality. Adjustments were made for demographics, year of hospitalization, and co-morbidities. The rate of postprocedural stroke was lowest with DES (0.5%), followed by BMS (0.6%), off-pump CABG (1.3%), and on-pump CABG (1.8%). After adjustment, on-pump CABG had a higher risk of stroke compared with off-pump (odds ratio 1.36, 95% CI 1.18 to 1.56, p <0.0001). DES had lower risk of stroke compared with off-pump CABG (odds ratio 0.64, 95% CI 0.55 to 0.74, p <0.0001). There was a significant excess risk of 1-year mortality due to the interaction between stroke and procedure type (on-pump vs off-pump CABG and PCI with DES vs BMS; p value for interaction = 0.02). In conclusion, in this retrospective analysis of nonrandomized data from a statewide database, PCI with DES was associated with the lowest rate of postprocedural stroke, and off-pump CABG had a lower rate of postprocedural stroke than on-pump CABG; there was an excess 1-year mortality risk with on-pump versus off-pump CABG and with DES versus BMS in patients with stroke.

Baseline blood pressure control in Hispanics: characteristics of Hispanics in the Systolic Blood Pressure Intervention Trial.

The Systolic Blood Pressure Intervention Trial (SPRINT) tested whether a systolic blood pressure (SBP) value <120 mm Hg reduces adverse clinical outcomes compared with the goal of <140 mm Hg. Here the authors describe the baseline characteristics of Hispanic participants in SPRINT. Nondiabetic hypertensive patients 50 years and older with SBP 130-180 mm Hg taking zero to four blood pressure (BP) medications were enrolled from the mainland United States and Puerto Rico. Cross-sectional, bivariate analysis was employed comparing sociodemographic and clinical factors in Hispanics vs non-Hispanics. Multivariable logistic regression models restricted to Hispanics were used to identify factors associated with achieved BP control (SBP <140 mm Hg and diastolic BP <90 mm Hg) at baseline. Eleven percent (n=984) of SPRINT participants were Hispanic; 56% (n=549) of Hispanics were living in Puerto Rico and the remainder were living on the US mainland. Hispanics overall were younger, more often female, less likely to live alone, and more likely to have lower education and be uninsured, although just as likely to be employed compared with non-Hispanics. BP control was not different between Hispanics vs non-Hispanics at baseline. However, a significantly higher percentage of Hispanics on the US mainland (compared with Hispanics in Puerto Rico) had controlled BP. BP control was independently associated with cardiovascular disease history and functional status among Hispanics, specifically those living in Puerto Rico, whereas functional status was the only independent predictor of BP control identified among mainland Hispanics. These findings highlight the diversity of the SPRINT population. It remains to be seen whether factors identified among Hispanics impact intervention goals and subsequent clinical outcomes.

Ischemic Stroke Rate Increases in Young Adults: Evidence for a Generational Effect?

BACKGROUND: The incidence rates of ischemic stroke and ST-segment elevation myocardial infarction (STEMI) have decreased significantly in the United States since 1950. However, there is evidence of flattening of this trend or increasing rates for stroke in patients younger than 50 years. The objective of this study was to examine the changes in incidence rates of stroke and STEMI using an age-period-cohort model with statewide data from New Jersey. METHODS AND RESULTS: We obtained stroke and STEMI data for the years 1995-2014 from the Myocardial Infarction Data Acquisition System, a database of hospital discharges in New Jersey. Rates by age for the time periods 1994-1999, 2000-2004, 2005-2009, and 2010-2014 were obtained using census estimates as denominators for each age group and period. The rate of stroke more than doubled in patients aged 35 to 39 years from 1995-1999 to 2010-2014 (rate ratio [RR], 2.47; 95% CI, 2.07-2.96 [P<0.0001]). We also found increased rates of stroke in those aged 40 to 44, 45 to 49, and 50 to 54 years. Strokes rates in those older than 55 years decreased during these time periods. Those born from 1945-1954 had lower age-adjusted rates of stroke than those born both in the prior 20 years and in the following 20 years. STEMI rates, in contrast, decreased in all age groups and in each successive birth cohort. CONCLUSIONS: There appears to be a significant birth cohort effect in the risk of stroke, where patients born from 1945-1954 have lower age-adjusted rates of stroke compared with those born in earlier and later years.

Role of Coronary Angiography in the Assessment of Cardiovascular Risk in Kidney Transplant Candidates.

Cardiovascular disease is the leading cause of death among those with renal insufficiency, those requiring dialysis, and in recipients of kidney transplants reflecting the greatly increased cardiovascular burden that these patients carry. The best method by which to assess cardiovascular risk in such patients is not well established. In the present study, 1,225 patients seeking a kidney transplant, over a 30-month period, underwent cardiovascular evaluation. Two hundred twenty-five patients, who met selected criteria, underwent coronary angiography that revealed significant coronary artery disease (CAD) in 47%. Those found to have significant disease underwent revascularization. Among the patients found to have significant CAD, 74% had undergone a nuclear stress test before angiography and 65% of these stress tests were negative for ischemia. The positive predictive value of a nuclear stress test in this patient population was 0.43 and the negative predictive value was 0.47. During a 30-month period, 28 patients who underwent coronary angiography received an allograft. None of these patients died, experienced a myocardial infarction, or lost their allograft. The annual mortality rate of those who remained on the waiting list was well below the national average. In conclusion, our results indicate that, in renal failure patients, noninvasive testing fails to detect the majority of significant CAD, that selected criteria may identify patients with a high likelihood of CAD, and that revascularization reduces mortality both for those on the waiting list and for those who receive an allograft.

Usefulness of a normal coronary angiogram in patients aged ≥65 years to foretell survival.

A normal coronary angiogram (CA) has been reported to confer a good prognosis. However, how this applies to patients aged ≥65 years is not well known. From 1986 to 1996, 11,625 patients aged ≥65 underwent coronary angiography. We identified 271 patients with either normal (NORM, n = 160) CA or <30% diameter stenosis disease (NEAR-NORM, n = 111). Using the Myocardial Infarction Data Acquisition System, we examined the probability of survival and the risk of developing an ischemic event or undergoing a revascularization procedure during an average of 15.1 ± 6.2 years (range 0.5 to 25.8 years). Matched actuarial subjects were used to compare survival to the general population. The incidence of an ischemic event was low (2.0 events per 100 persons/year for the NORM and 2.8 patients per 100 persons/year for the NEAR-NORM group, p = NS). Rates of revascularization were higher in the NEAR-NORM group compared to the NORM group (1 per 100 persons/year vs 0.5 per 100 persons/year, p = 0.04). During the 25.8-year follow-up, there were 77 deaths (48.4%) for the NORM and 64 (57.1%) for the NEAR-NORM group (χ2 = 1.7, NS). The NORM group survived 6,789 days, 1,517 more days than the actuarial subjects (95% confidence interval [CI] 1,072 to 1,956; p <0.0001) and the NEAR-NORM group survived 5,922 days, 875 more days (95% CI 368 to 1,376; p <0.005). In conclusion, patients with normal or near-normal CA at age ≥65 years have a low rate of myocardial ischemic events and have significantly longer survival than matched subjects from the general population.

Gender disparity in the use of drug-eluting stents during percutaneous coronary intervention for acute myocardial infarction.

OBJECTIVE: We examined gender disparity in the use of drug-eluting stents (DES) versus bare metal stents (BMS) during percutaneous coronary intervention (PCI) for acute myocardial infarction (AMI), and gender disparity in all-cause mortality after coronary stent implantation for AMI. BACKGROUND: Gender disparities in AMI managements have been well documented, but it is unclear whether these disparities are seen in the type of coronary stent implantation for AMI and outcomes. METHODS: Hospital discharge data from January 1, 2003 through December 31, 2010 in New Jersey from the Myocardial Infarction Data Acquisition System were used to identify 40,215 patients (12,878 women and 27,337 men) with coronary stent implantation for AMI. The in-hospital, short term (30 days) and long term (1 and 5 year) all-cause mortality rates, unadjusted and adjusted for demographics and comorbidities, were determined. RESULTS: Women were older than men and had a higher prevalence of co-morbidities. Men had higher prevalence of prior coronary revascularizations. After adjustment for co-morbidities, there was no significant gender difference in the use of DES versus BMS for AMI, except in 2003 and 2006 where women were found to be more likely to receive a DES versus a BMS. After adjustment, women had higher odds of in-hospital deaths but no difference in short and long-term all-cause mortality rates. CONCLUSIONS: There was no significant gender difference in the proportion of DES implantation versus BMS for AMI in contemporary years. Women treated with either BMS or DES for AMI had higher in-hospital death than men.

Incidence and trends of heart failure admissions after coronary artery bypass grafting surgery.

AIMS: Factors related to hospitalization for heart failure (HF) following coronary artery bypass grafting (CABG) surgery were studied. METHODS AND RESULTS: Patients (n = 65 377) undergoing CABG surgery in New Jersey from 1998 to 2007 were identified from the state cardiac surgery database; subsequent hospitalizations for HF were assessed using the Myocardial Infarction Data Acquisition System database. Patients were classified based on pre-operative ejection fraction (EF). Multivariate models were used to identify factors related to HF admission and mortality. Post-CABG HF admission rates among patients with pre-operative EF <35% increased over the 10-year period (P = 0.02), but no significant trend was seen among patients with EF ≥35%. Independent factors associated with post-CABG HF admission within 2 years were: EF, age, female gender, Black race, smoking, diabetes, renal disease, hypertension, and cerebrovascular disease. Pre-operative use of beta-blockers increased over the years (P < 0.0001) and reduced the risk of admission for HF by 13%, with greater benefit in patients with lower EF. Mortality remained unchanged from 1998 to 2007, averaging 1.8% in-hospital and 5.1% and 7.2% at 1-year and 2-year follow-up, respectively. CONCLUSIONS: Pre-operative EF is a strong predictor of HF admission within 2 years after CABG surgery. The use of beta-blockers decreased HF admission after CABG, especially in patients with EF <35%. Despite the more pronounced benefit and increasing use of beta-blockers in patients with a low EF, HF admission rates in this group of patients are rising. This suggests that more comprehensive management of factors associated with HF is necessary.

Continuation of mortality reduction after the end of randomized therapy in clinical trials of lipid-lowering therapy.

BACKGROUND: Long-term follow-up of clinical trials with lipid-lowering medications has suggested a continuation of event reduction after study completion. OBJECTIVE: To evaluate the persistence of the benefit of lipid-lowering therapy in decreasing mortality after the end of clinical trials, when all patients were advised to take the same open-label lipid-lowering therapy. METHODS: Through searches of MEDLINE, the Cochrane Library, the Central Register of Controlled Trials, Web of Science, and ClinicalTrials.gov until June 2010 we identified randomized clinical trials of lipid-lowering agents with a second report describing results after the end of the trial. RESULTS: Among the 459 trials reviewed, only 8 including 44,255 patients and 8144 deaths qualified for the meta-analysis. All-cause and cardiovascular mortality were lower in the active intervention group during the first phase (0.84, 95% confidence interval [CI] 0.76-0.93; P = .0006 and 0.72, 95% CI 0.63-0.82, P < .0001, respectively) when 71 ± 23% of the patients randomized to receive active therapy actually received it compared with 13 ± 5% of patients who received active therapy although they were randomized to placebo (P = .0001). The lower mortality among those initially randomized to active therapy persisted during the second phase (odds ratio 0.90, 95% CI 0.84-0.97, P = .0035, and 0.82 95% CI 0.73-0.93, P = .0014), when patients in both randomized groups received active therapy in the same proportions (5 ± 2% for both groups). Numerous sensitivity analyses support the conclusions of the paper. CONCLUSION: The decrease in mortality with lipid-lowering therapy in clinical trials persists after discontinuation of randomized therapy when patients in the treatment and placebo groups receive active therapy.