Antoni Basinski

Depression and risk of sudden cardiac death after acute myocardial infarction: testing for the confounding effects of fatigue.

OBJECTIVES This study examined the impact of depressive symptoms and social support on 2-year sudden cardiac death (SCD) risk, controlling for fatigue symptoms. METHODS Myocardial infarction (MI) patients (N = 671) participating in the Canadian Amiodarone Myocardial Infarction Arrhythmia Trial completed measures of depression, hostility, and social support. RESULTS After controlling for significant biological predictors, psychosocial predictors of increased SCD risk in the survival analysis were greater social network contacts (RR = 1.04; 95% CI = 1.01-1.06; p < .007), lower social participation (RR = 0.98; 95% CI = 0.96-1.00; p < .05), and, in placebo-treated patients, elevated depressive symptoms (RR = 2.45; 95% CI = 1.14-5.35; p < .02). Fatigue was associated with SCD (RR = 1.31; 95% CI = 1.11-1.53; p < .001), and, when included in the model, diminished the influence of depression (RR = 1.73; 95% CI = 0.75-3.98; p = .20). When the cognitive-affective depressive symptoms were examined separately from somatic symptoms, there was a trend for an association between cognitive-affective symptoms and SCD in placebo-treated patients after controlling for fatigue (RR = 1.09; 95% CI = 0.99-1.19, p < .06). CONCLUSIONS Symptoms of depression and fatigue overlap in patients with MI. The trend for the cognitive-affective symptoms of depression to be associated with SCD risk, even after controlling for dyspnea/fatigue, suggests that the association between depression and mortality after AMI cannot be entirely explained as a confound of cardiac-related fatigue. The independent contribution of social participation suggests a role of both depressive symptomatology and social factors in influencing mortality risk after MI.

Assessment of priority for coronary revascularisation procedures

Abstract To develop guidelines for ranking the urgency with which patients with angiographically proven coronary disease need revascularisation procedures, factors that a panel of cardiac specialists agreed were likely to affect urgency were incorporated into 438 fictitious case-histories. Each panelist then rated the cases on a 7-point scale based on maximum acceptable waiting time for surgery; 1 on the scale represented emergency surgery and 7 delays of up to 6 months. For only 1% of cases was there agreement on a single rating by at least 12/16 panelists. Results of this ranking exercise were used by the panel to draw up triage guidelines. The three main urgency determinants were severity and stability of symptoms of angina, coronary anatomy from angiographic studies, and results of non-invasive tests for risk of ischaemia. Together these three factors generally gave an urgency rating for any given case to within less than 0·25 scale points of the value predicted with all factors. A numerical scoring system was derived to permit rapid estimation of the panels recommended ratings.

Aspirin and fibrinolysis in acute myocardial infarction: meta-analytic evidence for synergy.

A meta-analysis of randomized clinical trials of fibrinolysis was performed, examining the interaction between aspirin and fibrinolysis in treating patients with acute myocardial infarction. Reductions in the odds of death up to 35 days were assessed for patients receiving tissue plasminogen activator or streptokinase up to 6 hours after the onset of symptoms. No significant difference in effectiveness between tissue plasminogen activator and streptokinase was demonstrated. The overall reduction in odds of death due to fibrinolytic therapy was 28%. However, there was a significant difference between the odds reduction of 24% when fibrinolysis is compared to placebo, and 40% when fibrinolysis and aspirin combined are compared to aspirin alone (p = 0.02). This difference indicates that there exists a synergistic interaction between coronary fibrinolysis and aspirin rather than independence of their beneficial effects, as is generally believed. These results illustrate the perils of assessing drug efficacy, even in an overview of all relevant trials, without consideration of identifiable sources of heterogeneity such as the interaction between the treatment of interest and co-interventions. They also demonstrate the potential application of logistic regression diagnostic techniques to meta-analyses.

Primary prevention drug therapy: can it meet patients' requirements for reduced risk?

The objective was to identify, in primary prevention, patients whose “required risk reduction” (ReqRR) is greater than the “achievable risk reduction” (ARR) that cholesterol-lowering or antihypertensive medication could provide. Individualized estimates of 10-year coronary heart disease or stroke risk were derived for 66 hypercholesterolemic (HC) and 64 hypertensive (HT) patients without symptomatic cardiovascular disease. These estimates were used in trade-off tasks identifying each individual’s ReqRR. Then individual ARRs were estimated (in HC patients by assuming total cholesterol/high density lipoprotein ratio reductions to 5.0; in HT patients by assuming systolic blood pressure reductions to 120 mmHg). 12 (18%) HC and 12 (19%) HT subjects would refuse medication regardless of the risk reduction offered. Of the remaining patients, 15/54 (28%; 95% C.I.: 16-40%) HC and 19/52 (37%; 95% C.I.: 24-51%) HT subjects were “over-requirers,” in that their ReqRR/ARR ratio was 1.5. There may be a notable proportion of patients whose ReqRR is considerably greater than what is achievable, implying that decision aids may help individuals clarify preferences about accepting/refusing medication for the primary prevention of cardiovascular disease.

Determinants of left ventricular mass in early hypertension

One hundred seventy-six unmedicated mildly hypertensive subjects (113 men, 63 women) underwent M-mode echocardiography to determine left ventricular mass (LVM) and relative wall thickness (RWT), 24-h ambulatory blood pressure monitoring, and completed standardized questionnaires measuring marital and job stress. Subjects were aged 46 +/- 9 years old; 45.4% had daytime diastolic blood pressure < 90 mm Hg; 96.1% of LVM results were in the normal range. We found that neither marital distress nor job strain was a determinant of LVM. However, a segmental regression approach revealed inflection points of 131 mm Hg systolic daytime blood pressure and 83 and 87 mm Hg nighttime diastolic blood pressure in the relation between LVM and RWT, respectively, and ambulatory BP. In addition, we found that the variability of LVM was best explained by indexing LVM by height, rather than body surface area.

Mismatch of coronary risk and treatment intensity under the national cholesterol education program guidelines

Objective:To assess the match between multifactorial risk of coronary heart disease (CHD) and treatment intensity under the National Cholesterol Education Program (NCEP) guidelines for primary prevention of CHD.Methods:The multiple logistic regression equation from the Framingham Study was used to derive predicted risks for development of CHD over eight years of follow-up for different age — gender groupings, with serum total cholesterol (TC) values chosen in light of the NCEP cutoff points for both TC and low-density-lipoprotein cholesterol levels. Additional risk factors — hypertension, glucose intolerance, and smoking—were considered in combination for each of these values.Results:Controlling for the effects of age and gender, there is little difference in the ranges of absolute CHD risks for persons who would receive interventions of differing intensities (i.e., general dietary advice, dietary treatment, or drug therapy). Those who are candidates for drug treatment because of serum lipids alone are often at low levels of risk for the development of CHD when compared with those of the same age with lower TC values who have other risk factors. Discrepancies in CHD risk are wider still when age is also allowed to vary. Furthermore, in every age grouping, women with high TC levels (e.g., 6.9 mmol/L) and two other risk factors are eligible for drug treatment but have a CHD risk that is no higher, and often much lower, than that of males with one other risk factor and TC levels of 4.8 mmol/L or 5.7 mmol/L who are candidates for dietary advice or dietary therapy, respectively.Conclusions:Inconsistencies exist in the NCEP guidelines such that persons at low risk for the development of CHD are offered more intensive interventions than are others who actually are at much higher risks, and vice versa. Women in particular tend to be overtreated, relative to men. These findings point out the difficulties of promulgating guidelines that will appropriately match risk to preventive interventions in a complex multifactorial disease.

FURTHER THOUGHTS ON THE IMPORTANCE OF MODELS IN THE ASSESSMENT OF CLINICAL EVIDENCE

The assessment of clinical evidence relies on a meaningful and concise summary of clinical data. Any summary is likely to be meaningful only if it is biologically and statistically appropriate. If the analysis is “explanatory” rather than “pragmatic” [l], the summary may serve both to elucidate biological mechanisms and to draw statistical conclusions. Hlatky and Whittemore [2] have commented on the importance of statistical models in the identification of synergy between two therapeutic agents, a concept explored in previous publications [3,4], They have focused attention on the assessment of synergy between aspirin and fibrinolysis in acute myocardial infarction which we reported, originally for the ISIS-2 [5,6] trial data and subsequently in a metaanalysis [7]. These analyses demonstrated that the effect of fibrinolysis vs placebo when both are administered in the presence of aspirin is greater than that of fibrinolysis vs placebo when both are administered in the absence of aspirin. Logistic regression was utilized to analyze the aggregate trial data. Although it may be argued that the definition of interaction and synergy is arbitrary in the absence of external information to guide the analysis, it is rare that the biological, clinical, and statistical milieu within which the observed