Statistical evaluation of biomedical studies.

The aim of this chapter is to familiarize the reader with the basic information and common statistical analyses used in medical research. The chapter will aid in deciding what type of analyses best fit the study data and how each analysis differs. The chapter was written to be user-friendly from a medical research and statistical consultant perspective.

A randomized, double-blind, placebo-controlled, parallel-group study to assess the efficacy and safety of dual ACE/NEP inhibitor GW660511X in mild-to-moderate hypertensive patients.

This multicentre, double-blind, placebo-controlled, parallel-group study determined the efficacy and safety of GW660511 200 mg, a dual inhibitor of angiotensin-converting enzyme (ACE) and neutral endopeptidase (NEP), in mild-to-moderate hypertensive patients (diastolic blood pressure (DBP), > or =90 and < or =109 mm Hg; systolic blood pressure (SBP), > or =150 and < or =180 mm Hg). After a single-blind 2- to 4-week placebo run-in period, 123 patients (aged 18-65 years) were randomized to either placebo (n=62) or to active treatment (n=61) consisting of two consecutive 3-day dose titration periods of GW660511X 50 mg once daily and 100 mg once daily followed by GW660511X 200 mg once daily for 14 days. GW660511X 200 mg significantly lowered (baseline and placebo-corrected) both trough mean cuff SBP (-8.00 mm Hg, P=0.002) and DBP (-5.38 mm Hg, P=0.003). GW660511X 200 mg significantly reduced placebo-corrected mean 24-h and daytime but not night-time ambulatory SBP and DBP. Over the 0-24 h time period following GW660511X 200 mg, there were significant (P<0.001) reductions in serum ACE activity and significant (P<0.001) increases in plasma ANP concentration compared with placebo in terms of both peak and trough effects. In addition, treatment with GW660511X 200 mg significantly (P=0.003) increased (placebo-corrected, 1.52-fold) urinary excretion of cGMP over the 0-24 h interval. Treatment-related adverse events were experienced by 43% of the patients administered GW660511X 200 mg and 44% of those dosed with placebo with headache the most commonly reported. In conclusion, GW660511X 200 mg is an effective antihypertensive in mild-to-moderate hypertensive patients with potent effects on biological markers of ACE and NEP inhibition.

Association between myeloperoxidase levels and risk of coronary artery disease.

CONTEXT: Myeloperoxidase (MPO), a leukocyte enzyme that promotes oxidation of lipoproteins in atheroma, has been proposed as a possible mediator of atherosclerosis. OBJECTIVE: To determine the association between MPO levels and prevalence of coronary artery disease (CAD). DESIGN, SETTING, AND PATIENTS: Case-control study conducted from July to September 2000 in a US tertiary care referral center, including 158 patients with established CAD (cases) and 175 patients without angiographically significant CAD (controls). MAIN OUTCOME MEASURES: Association of MPO levels per milligram of neutrophil protein (leukocyte-MPO) and MPO levels per milliliter of blood (blood-MPO) with CAD risk. RESULTS: Leukocyte- and blood-MPO levels were both significantly greater in patients with CAD than in controls (P<.001). In multivariable models adjusting for traditional cardiovascular risk factors, Framingham risk score, and white blood cell counts, MPO levels were significantly associated with presence of CAD, with an OR of 11.9 (95% CI, 5.5-25.5) for the highest vs lowest quartiles of leukocyte-MPO and an OR of 20.4 (95% CI, 8.9-47.2) for the highest vs lowest quartiles of blood-MPO. CONCLUSIONS: Elevated levels of leukocyte- and blood-MPO are associated with the presence of CAD. These findings support a potential role for MPO as an inflammatory marker in CAD and may have implications for atherosclerosis diagnosis and risk assessment.

Effect of hydroxymethylglutaryl coenzyme a reductase inhibitors on the progression of calcific aortic stenosis.

BACKGROUND: Recent studies have supported the hypothesis that calcific aortic stenosis is the product of an active inflammatory process, with similarities to atherosclerosis. We sought to determine whether therapy with hydroxymethylglutaryl coenzyme A reductase inhibitors (statins) might slow the progression of aortic stenosis. METHODS AND RESULTS: A retrospective study of 174 patients (mean age 68+/-12 years) with mild to moderate calcific aortic stenosis was conducted. Patients required normal left ventricular function, /=2 echocardiograms performed at least 12 months apart. Fifty-seven patients (33%) received treatment with a statin; the remaining 117 (67%) did not. The statin group was older and had a higher prevalence of hypertension, diabetes mellitus, and coronary disease. During a mean follow-up of 21 months, patients treated with statin had a smaller increase in peak and mean gradient and a smaller decrease in aortic valve area. When annualized, the decrease in aortic valve area for the nonstatin group was 0.11+/-0.18 cm(2) compared with 0.06+/-0.16 cm(2) for those treated with a statin (P=0.03). In multivariate analysis, statin usage was a significant independent predictor of a smaller decrease in valve area (P=0.01) and a lesser increase in peak gradient (P=0.02). CONCLUSIONS: Statin-treated patients, despite a higher risk profile for progression, had reduced aortic stenosis progression compared with those not treated with a statin. These data provide justification for a prospective randomized trial to substantiate whether statin therapy slows the progression of aortic stenosis.

Safety profile of tolterodine as used in general practice in England: results of prescription-event monitoring.

BACKGROUND: Unstable bladder symptoms are a common problem in general practice. Drug therapy with anticholinergic drugs is frequently used in the management of this condition. However such drugs are associated with a high incidence of anticholinergic adverse effects. Tolterodine is a competitive anticholinergic agent, selective for the bladder as opposed to the salivary glands. OBJECTIVE: To monitor the safety of tolterodine as used in general practice patients in England for the treatment of urinary frequency, urgency and incontinence. DESIGN: Prospective observational cohort study. PATIENTS AND PARTICIPANTS: 14,526 patients [mean age 62.7 (SD 16.4) years; 68.6% female]. METHODS: Patients prescribed tolterodine in general practice, soon after the release of the drug in the UK, were followed up for a minimum of 6 months using the technique of prescription-event monitoring (PEM). RESULTS: The most common adverse events reported were dry mouth, headache, malaise, constipation, dyspepsia, nausea and vomiting and pain in abdomen. We identified some uncommon events as possible adverse drug reactions--notably hallucinations, tachycardia and palpitations. The prevalence of these events was compared with that in patient cohorts for other drugs on the PEM database. The age- and sex-adjusted relative risk of hallucinations on tolterodine compared with 10 drugs of other therapeutic classes, and with terodiline, another drug indicated for bladder instability, was 4.85 [95% confidence interval (CI) 2.72 to 8.66] and 1.25 (95% CI 0.62 to 2.53), respectively. There was no significant difference for tachycardia/palpitation in this comparison. CONCLUSIONS: Tolterodine is well tolerated in general practice at the recommended daily dose. Hallucinations, tachycardia and palpitations are infrequently associated with the drug.

Is there an association between selective serotonin reuptake inhibitors and risk of abnormal bleeding? Results from a cohort study based on prescription event monitoring in England.

OBJECTIVE: To investigate whether an association between the use of selective serotonin reuptake inhibitor (SSRI) antidepressants and abnormal bleeding is demonstrated in a large population study. METHODS: An observational cohort study using cohorts from the Drug Safety Research Unit's prescription event monitoring database was performed. RESULTS: Analysis of combined haemorrhagic event rates calculated for the first 6 months of treatment for four SSRIs showed no significant difference between the rate for abnormal bleeding in the first month after starting treatment compared with months 2-6 [difference in rates 0.63 per 1000 patient months of treatment, 99% confidence interval (CI) -0.4, 1.67]. Comparison of the rates for the exposed combined SSRI cohort with the unexposed non-psychiatric drug cohort for the first month [relative risk (RR) 1.38, 95% CI 0.82, 2.34] and months 2-6 (RR 1.17, 95% CI 0.81, 1.68) showed no significant differences after adjustment for age and gender. However, there was a tendency towards highest risk with the combined SSRI cohort and lowest with the baseline cohort. CONCLUSION: This study provides weak evidence to support the hypothesis of a link between SSRIs and precipitation of bleeding events at a population level. The 95% CI is consistent with a possible risk of bleeding associated with SSRI users versus non-psychiatric drug users in the first month. Fuller consideration of confounding would be possible using follow-up of identified cases in a nested case-control study.

Comparison of cardiovascular risk and lipid profiles in patients undergoing aortic valve surgery versus those undergoing coronary artery bypass grafting.

BACKGROUND AND AIM OF THE STUDY: Hyperlipidemia is a risk factor for the progression of coronary artery disease, and possibly also valvular aortic stenosis. Thus, patients with aortic stenosis, coronary disease (or both) might be expected to have more abnormal lipid profiles than those without these two conditions. METHODS: The lipid profiles of patient subsets undergoing aortic valve replacement (AVR) with or without concomitant coronary artery bypass grafting (CABG), as well as those undergoing isolated CABG, between 1987 and 1997 were analyzed retrospectively. Four surgical groups were identified: AVR for aortic regurgitation (n = 370); AVR for predominant aortic stenosis (n = 1,072); AVR for aortic stenosis (AS) with CABG (n = 914); and isolated CABG (n = 11,156). The complete fasting lipid profiles of patients were collected, analyzed by group, and compared. RESULTS: Analysis by Spearman's correlation showed that total cholesterol levels, triglycerides and low-density lipoproteins (LDL-C) were modestly, yet significantly, increased in each successive group, while high-density lipoproteins were decreased. AS patients undergoing isolated AVR had significantly higher total cholesterol (215 versus 201 mg/dl; p <0.0001), triglycerides (125 versus 104 mg/dl; p <0.0001) and LDL-C (139 versus 132 mg/dl; p = 0.003) than those undergoing AVR for aortic regurgitation. Total cholesterol >200 mg/dl was significantly associated with AS, even after adjusting for differences in age, sex, diabetes mellitus and hypertension, with an odds ratio of 1.5 (95% confidence interval, 1.2-2.0; p = 0.001). CONCLUSION: Progressively abnormal lipid profiles are associated with AS and coronary disease in patients undergoing AVR. This evidence helps to extend the link between dyslipidemia and AS in a large consecutive series of patients.

Profiles of hepatic and dysrhythmic cardiovascular events following use of fluoroquinolone antibacterials: experience from large cohorts from the Drug Safety Research Unit Prescription-Event Monitoring database.

OBJECTIVE: To investigate how frequently serious dysrhythmic cardiovascular, and hepatotoxic events are reported during routine clinical use of fluoroquinolones (quinolones) in general practice. DESIGN: Cohorts prescribed quinolones (cohort sizes: ciprofloxacin 11 477; enoxacin 2790; ofloxacin 11 033 and norfloxacin 11 110; mean age in each cohort of 48.6 to 57.0 years) were selected from the Drug Safety Research Unit's Prescription-Event Monitoring (PEM) database. The monitoring periods were November 1988 to January 1989 for ciprofloxacin; April 1989 to January 1991 for enoxacin; May 1991 to December 1991 for ofloxacin and October 1990 to October 1991 for norfloxacin. Data collected over the total PEM surveillance period on selected gastrointestinal events were extracted and reviewed to identify possible hepatic events, together with selected cardiovascular events associated with dysrhythmias. For each quinolone, times to onset of the event and patient-months of observation (denominator values) were calculated. The analysis was based on two observation periods: rate of event during the first 7 days following dispensing of a prescription for each drug (W(1)), and rate of event during the second to sixth week inclusive (W(2)). RESULTS: Scrutiny of original green forms revealed no evidence of drug-induced hepatic dysfunction within 42 days of drug administration for any of the quinolones monitored. No evidence was found of drug-induced dysrhythmic events associated with enoxacin within 42 days of drug administration. Of the other quinolones, 'atrial fibrillation' was reported most often within 42 days following ciprofloxacin administration, with no change in event rate over that time, crude relative risk (CRR)[W(1)/W(2)] 1.0 [95% confidence interval (CI) 0.02 to 8.92]. Other less serious events associated with dysrhythmia were reported with varying incidence within 42 days of quinolone administration. The crude rate of palpitation did not change significantly over that time for ciprofloxacin, ofloxacin and norfloxacin: CRR 0.83 (95% CI 0.02 to 6.86), 2.00 (95% CI 0.19 to 12.20) and 4.99 (95% CI 0.06 to 391.94), respectively. Syncope and tachycardia were also reported for ofloxacin [CRR 9.99 (95% CI 0.52 to 589.49 for both events)] and ciprofloxacin [1.0 (95% CI 0.02, 8.92)] and 2.50 (95% CI 0.04, 47.96) for syncope and tachycardia, respectively]. CONCLUSION: It cannot be ruled out that some rare hepatic and dysrhythmic events associated with quinolones may be drug related. The primary purpose of PEM is signal generation. Compared with the other quinolones, ciprofloxacin was associated with the highest number of reports of dysrhythmic cardiovascular events occurring within 42 days of administration. This requires further investigation by other types of epidemiological study.

HDL cholesterol level predicts survival in men after coronary artery bypass graft surgery: 20-year experience from The Cleveland Clinic Foundation.

BACKGROUND: HDL cholesterol (HDL-C) is an important independent predictor of atherosclerosis, yet the role that HDL-C may play in the prediction of long-term survival after CABG remains unclear. The risk associated with a low HDL-C level in post-CABG men has not been delineated in relation to traditional surgical variables such as the use of arterial conduits, left ventricular function, and extent of disease. METHODS AND RESULTS: We performed a prospective, observational study of 432 men who underwent CABG between 1978 and 1979 in whom preoperative HDL-C values were available. Baseline lipid and lipoprotein values, history of diabetes mellitus and hypertension, left ventricular ejection fraction, extent of disease, and use of internal thoracic arteries were recorded. Hazard ratios (HRs) were determined in the patients with and without a low HDL-C level, which was defined as the lowest HDL-C quartile (HDL-C 35 mg/dL) were 50% more likely to survive at 15 years than were patients with low HDL-C level (35 mg/dL were 50% more likely to survive without a subsequent myocardial infarction or revascularization (HR 1.42, P:=0.02). CONCLUSIONS: HDL-C is an important predictor of survival in post-CABG patients. In this study of >8500 patient-years of follow-up, HDL-C was the most important metabolic predictor of post-CABG survival. One third fewer patients survive at 15 years if their HDL-C levels are

Preoperative triglycerides predict post-coronary artery bypass graft survival in diabetic patients: a sex analysis.

OBJECTIVE: Hypertriglyceridemia is commonly observed in association with diabetes. Despite cross-sectional studies and isolated longitudinal analyses in patients without coronary artery disease, the suggestion that triglyceride levels are relevant to subsequent cardiovascular events in the setting of diabetes remains controversial. This study evaluates the predictive value of serum triglyceride levels on mortality in post-coronary artery bypass graft (CABG) diabetic patients with subsequent analysis by sex. RESEARCH DESIGN AND METHODS: This longitudinal observational study involving a large metropolitan hospital consists of 1,172 diabetic post-CABG patients (792 men and 380 women) with lipid data collected between the years 1982 and 1992. Cox proportional hazards regression models were used to estimate the risk of mortality and cardiac events associated with triglyceride levels in the highest quartile (> 2.90 mmol/l for men and > 3.12 mmol/l for women). RESULTS: Elevated preoperative serum triglyceride values in post-CABG subjects with diabetes were correlated with increased overall mortality (hazard ratio [HR] 1.26, 95% CI 1.00-1.59). The greatest influence of triglyceride levels was observed on overall (1.89, 1.30-2.73) and event-free survival (1.49, 1.06-2.08) in women. High triglyceride values were also modestly related to risk of cardiac events in diabetic men (1.28, 0.99-1.66). CONCLUSIONS: These data suggest that increased preoperative triglyceride levels predict increased late mortality and cardiac event risk in diabetic post-CABG patients, more strongly in women than in men.

How deadly is the "deadly quartet"? A post-CABG evaluation.

OBJECTIVES: The aim of the study was to determine the value of a cluster of metabolic risk factors in predicting mortality after coronary artery bypass surgery (CABG). BACKGROUND: The "deadly quartet" of metabolic risk factors (i.e., obesity, diabetes, hypertension, and hypertriglyceridemia) has been associated with coronary heart disease in healthy population studies. The expected influence of the cluster on survival in secondary prevention remains untested overall as well as by gender. METHODS: Patients with lipid profiles undergoing primary isolated CABG (n = 6,428) between 1987 and 1992 were followed a median of eight years. Cox models were used to evaluate all-cause mortality. Metabolic risk factors were incorporated as the sum of deadly quartet risk factors present in each patient (0 to 4). The role of gender as it relates to survival and metabolic risk clusters was also examined. RESULTS: The sum of deadly quartet risk factors showed a significant relationship to mortality as the hazard ratio increased from 1.64 (confidence interval [CI] = 1.34-2.01) for one risk factor to 3.95 (2.73-5.69) for four risk factors. Annualized mortality ranged from 1% per year in patients with no risk factors to 3.3% per year in patients with all four risk factors. Within gender, the hazard ratio associated with four risk factors was 2.58 for men and 13.39 for women. The expected clustering of risk factors was 8% compared to the observed clustering of 10% in men and 21% in women. CONCLUSIONS: This cohort showed risk factor clustering beyond that expected due to chance, particularly in women. Even after revascularization, survival is diminished for patients with members of a family of metabolic risk factors at the time of surgery.

Relation of serum triglyceride levels to survival after coronary artery bypass grafting.

We performed a prospective observational study on 6,602 subjects (94% for 5 years and 34% for 10 to 15 years) who underwent coronary artery bypass graft surgery (CABG) between 1982 and 1992. We examined whether triglyceride concentrations adjusted for other factors (total cholesterol, history of diabetes mellitus, systemic hypertension, left ventricular function, number of coronary arteries significantly narrowed, and use of the internal thoracic arteries) explained total and event-free survival. These analyses were duplicated within gender (1,354 women and 5,248 men). This approach allowed a determination of any gender-related disparities in lipid predictors. Triglycerides in the highest quartile were associated with an increased risk of mortality of 20% (confidence interval [CI] 1.0 to 1.4). Similar risk was seen for event-free survival. Although there was no evidence of gender differences in adjusted survival (p = 0.33), a gender by triglyceride interaction (p = 0.004) indicated that the response to high triglycerides as related to survival did differ by gender. Specifically, women had a dramatically higher risk (hazard ratio [HR] 1.5, CI 1.1 to 2.1) than men (HR 1.1, CI 0.9 to 1. 3). Both men and women did have triglyceride-associated risk with regard to event-free survival (HR in men 1.2, CI 1.1 to 1.4; HR in women 1.4, CI 1.1 to 1.8). Examination of high-density lipoprotein cholesterol in a subcohort did not eliminate the observed triglyceride effects. Thus, triglyceride baseline values are primary determinants (similar to baseline left ventricular function or extent of coronary disease) for long-term total and event-free mortality after CABG in women but not in men.

Sedation with "non-sedating" antihistamines: four prescription-event monitoring studies in general practice.

OBJECTIVES: To investigate the frequency with which sedation was reported in post-marketing surveillance studies of four second generation antihistamines: loratadine, cetirizine, fexofenadine, and acrivastine. DESIGN: Prescription-event monitoring studies. SETTING: Prescriptions were obtained for each cohort in the immediate post-marketing period. SUBJECTS: Event data were obtained for a total of 43 363 patients. MAIN OUTCOME MEASURES: Reporting of sedation or drowsiness. RESULTS: The odds ratios (adjusted for age and sex) for the incidence of sedation were 0.63 (95% confidence interval 0.36 to 1.11; P=0.1) for fexofenadine; 2.79 (1.69 to 4.58; P<0.0001) for acrivastine, and 3.53 (2.07 to 5.42; P<0.0001) for cetirizine compared with loratadine. No increased risk of accident or injury was evident with any of the four drugs. CONCLUSIONS: Although the risk of sedation was low with all four drugs, fexofenadine and loratadine may be more appropriate for people working in safety critical jobs.

Homocysteine and lipoprotein(a) interact to increase CAD risk in young men and women.

A biochemical link between homocysteine (tHcy) and lipoprotein(a) [Lp(a)] related to fibrin binding has been proposed. This hypothesis has not been specifically examined in human subjects. We sought to determine in a clinical setting whether these risk factors would interact to increase coronary artery disease (CAD) risk. We performed a cross-sectional analysis of 750 men and 403 women referred to a preventive cardiology clinic at the Cleveland Clinic Foundation, in whom baseline tHcy and Lp(a) data were available. Logistic regression after adjusting for standard cardiovascular risk factors was used to estimate the relative risk of CAD in patients with an Lp(a) >/=30 mg/dL and a tHcy >/=17 micromol/L. Neither isolated high tHcy (odds ratio [OR]=1.06, P=0.89) nor isolated high Lp(a) (OR=1.15, P=0.60) appeared to be associated with CAD in women. However, strong evidence of an association was seen when both risk factors were present (OR=4.83, P=0.003). Moreover, this increased risk showed evidence of an interactive effect beyond that attributable to either additive or multiplicative effects of tHcy and Lp(a) (P=0.03). In contrast, both elevated tHcy (OR=1.93, P=0. 05) and elevated Lp(a) (OR=1.87, P=0.01) showed evidence of being independent risk factors for CAD in men. The presence of both risk factors in men did not appear to confer additional risk (OR=2.00, P=0.09), even though ORs as high as 12.4 were observed within specific age intervals. Consistent with prior studies, tHcy and Lp(a) are risk factors, either independently or in concert, for CAD in this clinical population. More significantly, we found evidence that when both risk factors were present in women, the associated risk was greater than what would be expected if the 2 risks were simply acting independently. The absence of such an interactive effect in men may be due to the confounding effects of age manifested as "survivor bias." These clinical findings provide insights into the potential roles of both tHcy and Lp(a) in the pathogenesis of atherosclerosis.

Adverse effects associated with the use of donepezil in general practice in England.

Donepezil is an acetylcholinesterase inhibitor indicated for the symptomatic treatment of mild to moderate Alzheimer's disease. It is reported to have a relatively favourable side-effect profile. We report here on a pharmacovigilance study carried out post-marketing in England. An observational cohort study using the technique of Prescription-Event Monitoring was carried out. Some 1762 patients (mean age 72.9 years; 42% male) were followed up for 6 months minimum. The commonest adverse events were nausea, diarrhoea, malaise, dizziness and insomnia. Aggression, agitation and abnormal dreams were uncommonly associated with the drug. There were no cardiac rhythm disturbances or liver disorders causally associated. The commonest adverse drug reactions are already reported in the product information. Given the relatively small size of this cohort, the signals of abnormal dreams and psychiatric disturbance as possible adverse drug reactions need further investigation in carefully planned studies.

Front line workers in long-term care: the effect of educational interventions and stabilization of staffing ratios on turnover and absenteeism.

OBJECTIVE: Front line workers in long-term care play a crucial role in helping residents achieve their highest possible functional level. High turnover (rates in excess of 100% are common) and absenteeism threaten the ability of long-term care providers to meet this goal. The purpose of this study was to determine the effect of a formalized certified nursing assistant (CNA) education program and stabilization of staffing ratios on turnover and absenteeism. DESIGN AND SETTING: This study was a 12-month prospective, nonrandomized trial involving two long-term care facilities providing traditional intermediate and skilled care serving as study sites and a similar facility serving as a control. For historical comparisons, each facility served as its own control using data from the year before the interventions. INTERVENTIONS: An in-house educational program based on the State of North Carolina core curriculum for CNAs was instituted in each study site. During the study period, efforts were made to achieve stable staffing ratios of 1 CNA:8 residents for days, 1:10 for evenings, and 1:15 for nights. Traditional quality of care indicators and resident/surrogate satisfaction were tracked during the study period. RESULTS: Both study facilities showed a decline in turnover, with the decline reaching statistically significant levels at Facility B (134% to 41%, P = 0.0001). Absenteeism rates did not change significantly during the study period. Resident/surrogate satisfaction with nursing care was improved significantly at Facility B (P = 0.02). CONCLUSION: A formal education program in conjunction with stabilization of staffing ratios can result in lower turnover rates for CNA's and improved resident/surrogate satisfaction.

Lipoprotein(a) associated with coronary artery disease in older women: age and gender analysis.

BACKGROUND: Lipoprotein (a) has been associated with increased coronary artery disease (CAD) risk in men, but relatively little data exists in women. While age influences the cardiovascular risk associated with Lp(a) in men, little is known about this phenomenon in women. The impact of gender on Lp(a) has not been fully studied in an ongoing clinical practice. METHODS AND RESULTS: Baseline Lp(a) values were measured in 918 CAD and 829 non-CAD patients (603 females, 1144 males) entering an outpatient prevention clinic. The age-specific association of elevated Lp(a) (> 30 mg/dl) with CAD was examined after adjustment for traditional risk factors. Lp(a) was a significant risk factor (OR = 1.9, CI, 1.4-2.6) in men and women (OR = 1.9, CI 1.3-2.9). In men age < or = 55 years the odds ratio for increased cardiovascular risk in high vs low Lp(a) was 2.5 (CI 1.6-3.9). In men < or = 55, CAD increased from 32 to 61% as Lp(a) progressively rose from < or = 5 to > or = 45 mg/dl (P value for trend < 0.001). No significant increase was observed in men > 55 years (OR = 1.3, CI 0.9-2.1). In women < or = 55 years, the risk of CAD increased from 22 to 35% (OR 1.6, CI 0.8-3.2), and increased from 38 to 63% in women > 55 (OR 2.1, CI 1.3-3.5). Further, of high-risk patients (men < or = 55 and women > 55 years) with an Lp(a) in the range of 20-44 mg/dl (third quartile), younger men showed a greater incidence of CAD (51%) than older women (43%). Both genders revealed substantial risk when the Lp(a) values were above 45 mg/dl. (OR = 3.7, CI = 2.0-6.8 in younger men; OR = 3.3, CI = 1.6-6.6 in older women). CONCLUSIONS: In this cross sectional study of both men and women, elevated Lp(a) was associated with a significantly increased risk of CAD in men and women. While we corroborate previous reports on the lack of association in older men, the determination of an enhanced Lp(a)-related risk in older women was new and unanticipated. Further, in this population of high risk patients, substantial cardiovascular risk appeared to be represented by higher concentrations of Lp(a) in women than observed in men.