[Introduction to meta-analytic methodology]

How can we evaluate developmental care practices prior to their implementation in a neonatal intensive care unit?

Developmental care is designed to allow optimal neurobehavioral development of the preterm infant. The Neonatal Individualized Developmental Care and Assessment Program (NIDCAP) associates various strategies and focuses on individualized, family-oriented care. Scientific evaluation of developmental care is needed. Randomized controlled trials are the basis for medical evaluation, but present some limitations for developmental care studies. Qualitative research and benchmarking could be of interest in this field of neonatal medicine.

Phosphodiesterase III inhibitors for heart failure

BACKGROUND

In the treatment of chronic heart failure, vasodilating agents, ACE inhibitors and beta-blockers have shown an increase of life expectancy. Another strategy is to increase the inotropic state of the myocardium : phosphodiesterase inhibitors (PDIs) act by increasing intra-cellular cyclic AMP, thereby increasing the concentration of intracellular calcium, and lead to a positive inotropic effect.

OBJECTIVES

This overview on summarised data aims to review the data from all randomised controlled trials of PDIs III versus placebo in symptomatic patients with chronic heart failure. The primary endpoint is total mortality. Secondary endpoints are considered such as cause-specific mortality, worsening of heart failure (requiring intervention), myocardial infarction, arrhythmias and vertigos. We also examine whether the therapeutic effect is consistent in the subgroups based on the use of concomitant vasodilators, the severity of heart failure, and the type of PDI derivative and/or molecule. This overview updates our previous meta-analysis published in 1994.

SEARCH STRATEGY

Randomised trials of PDIs versus placebo in heart failure were searched using MEDLINE (1966 to 2004 January), EMBASE (1980 to 2003 December), Cochrane CENTRAL trials (The Cochrane Library Issue 1, 2004) and McMaster CVD trials registries, and through an exhaustive handsearching of international abstracting publications (abstracts published in the last 22 years in the "European Heart Journal", the "Journal of the American College of Cardiology" and "Circulation").

SELECTION CRITERIA

All randomised controlled trials of PDIs versus placebo with a follow-up duration of more than three months.

DATA COLLECTION AND ANALYSIS

21 trials (8408 patients) were eligible for inclusion in the review. 4 specific PDI derivatives and 8 molecules of PDIs have been considered.

MAIN RESULTS

As compared with placebo, treatment with PDIs was found to be associated with a significant 17% increased mortality rate (The relative risk was 1.17 (95% confidence interval 1.06 to 1.30; p<0.001). In addition, PDIs significantly increase cardiac death, sudden death, arrhythmias and vertigos. Considering mortality from all causes, the deleterious effect of PDIs appears homogeneous whatever the concomitant use (or non-use) of vasodilating agents, the severity of heart failure, the derivative or the molecule of PDI used.

AUTHORS' CONCLUSIONS

Our results confirm that PDIs are responsible for an increase in mortality rate compared with placebo in patients suffering from chronic heart failure. Currently available results do not support the hypothesis that the increased mortality rate is due to additional vasodilator treatment. Consequently, the chronic use of PDIs should be avoided in heart failure patients.

EasyMA: a program for the meta-analysis of clinical trials

Meta-analysis of clinical trial data is an increasingly important method in clinical research, particularly in the field of therapeutic evaluation. This method uses some specific statistical techniques which are not all available on standard packages and therefore require specific developments. This paper describes a program designed to help medical researchers perform meta-analyses of clinical trial data with dichotomous outcomes. This program includes the various statistical methods of meta-analysis and enables cumulative meta-analysis and sub-groups to be performed. A robustness index can be determined and the results obtained in table and graphic formats. Data-editing and data-manipulating facilities are also possible. Much care has been taken to make the user interface as user-friendly as possible, so that the program is within the reach of all medical researchers.