To estimate the effect of angiotensin converting enzyme (ACE) inhibitors on the risk of sudden cardiac death following myocardial infarction: a meta-analysis of randomised clinical trials.

MEDLINE was searched from January 1978 to August 1997. Reference lists of identified studies and review articles were examined.

Study designs of evaluations included in the review

Randomised placebo-controlled trials that evaluated ACE inhibitors in patients following acute myocardial infarction were included if they fulfilled the following criteria: study duration was at least 6 weeks; blinded follow-up was for at least 6 weeks; and total mortality, cardiovascular mortality and sudden cardiac death rates were reported or could be obtained from the investigators. Duration of follow-up ranged from 6 weeks to 42 months.

Specific interventions included in the review

The following ACE inhibitors were compared to placebo: captopril; zofenopril; enalapril; and trandolapril. Co-interventions included beta-blockers; aspirin; calcium channel blockers; and thrombolytic therapy at the time of myocardial infarction.

Participants included in the review

Patients who had an acute myocardial infarction in the previous 14 days were studied. Characteristics of participants (where stated) were as follows: mean age ranged from 54 to 67 years; percentage male ranged from 72% to 91%, diabetes mellitus rates ranged from 7% to 22%; ejection fraction ranged from less than 0.31 or less to 0.54; and hypertension rates ranged from 16% to 43%.

Outcomes assessed in the review

Sudden cardiac death rates (SCD), all cardiovascular death rates and total mortality were assessed. Sudden death was defined as 'sudden unexpected collapse without documentation of arrythmia or collapse due to intractable ventricular tachycardia/fibrillation' or accepted as reported where trials had used a prespecified definition of unexpected death within one hour of symptom onset.

How were decisions on the relevance of primary studies made?

The authors do not state how the papers were selected for the review, or how many of the authors performed the selection.

The authors do not state that they assessed validity.

The authors do not state how the data were extracted for the review, or how many of the authors performed the data extraction.

How were the studies combined?

Separate analyses were performed for SCD, all cardiac deaths and total mortality. A summary odds ratio (OR) was calculated for each outcome using both the DerSimonian and Laird random-effects model and the Mantel-Haenzel fixed-effect model.

How were differences between studies investigated?

Cochran's test was used to assess heterogeneity with respect to each end point, and among subsets of trials with follow-up < 6 months duration and those with > = 6 months duration and for sub- groups with < = 6 months duration and > 6 months duration. Sensitivity analysis was performed by reanalysing the data with the addition of two trials that had initially been excluded and by omitting each included trial in turn. Logistic regression was used to evaluate the influence of the baseline characteristics of each trial and simple linear regression used to assess the relationship of trial duration and the effects of ACE inhibitors on SCD.

Fifteen RCTs were included (15,104 patients).

There were 2,356 deaths, of which 302 (13%) were non cardiovascular and 2,054 (87%) were cardiovascular. 900 (44%) of the cardiovascular deaths were considered to be SCD.

Total mortality: there were fewer deaths in the ACE inhibitor assigned patients with OR = 0.83 (95% CI: 0.71, 0.97). Homogeneity for mortality was of borderline significance (P = 0.09). Trials < 6 months were homogeneous. Trials > = 6 months duration were heterogeneous.

Cardiovascular deaths: there were fewer deaths in the ACE inhibitor assigned patients with OR = 0.82 (95% CI: 0.69, 0.97). Trials were not homogeneous with respect to cardiovascular mortality (P = 0.05). Trials lasting < 6 months were homogeneous. Trials lasting > = 6 months were heterogeneous.

Non cardiovascular deaths: no difference between treatment groups with OR = 0.87 (95% CI: 0.69, 1.09).

SCD: there were fewer deaths in the ACE inhibitor assigned patients with OR = 0.80 (95% CI: 0.70, 0.92). Logistic regression model that included dichotomous variable for each trial showed reduced SDC in ACE inhibitor assigned patients with OR = 0.80 (95% CI: 0.70, 0.92; P <=0.001). Study duration was not correlated with the observed OR for SCD. Trials were homogeneous with respect to SCD (P = 0.90). Homogeneity was present for trials < 6 months duration and for those >= 6 months duration.

Trials <= 6 months and trials > 6 months duration were homogeneous for all end points.

The following variables were not significant when added to the logistic regression model: beta-blockers; aspirin; calcium channel blockers; thrombolytic therapy; and the prevalence of diabetes and hypertension.

Sensitivity analysis: a similar protective effect of ACE inhibitors was noted after including each of two excluded trials (one was not placebo controlled and lasted only 42 days and in the other the mode of death was not reported and trial lasted only 35 days). Removing each trial in turn produced very similar point estimates.

This analysis is consistent with prior reports showing that ACE inhibitor treatment decreased the risk of death following a recent myocardial infarction by reducing cardiovascular mortality. Moreover this analysis suggests that an estimated 20% reduction in SCD risk with ACE inhibitors is an important component of this survival benefit.

The aims and inclusion criteria were clearly stated. Reasons were given for exclusion of identified studies. Outcomes were defined. Some relevant details of the included studies were presented in tabular format. Heterogeneity was assessed statistically and considerable investigation undertaken of the influence of various factors on the results. The discussion included consideration of potential publication bias and the possible mechanisms for reduction in SCD. By limiting the literature search to one database some other relevant studies may have been omitted. It is not clear whether language restrictions were applied and no details were given of the keywords. No details were given of the methods used to select primary studies or extract data. Validity was not assessed.

The authors' conclusions were supported by the evidence presented though validity assessment and a more comprehensive literature search would have added to that evidence.

Practice: The authors do not report any clinical implications of the review.

Research: The authors do not report any research implications of the review.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.