This review concluded that in people with left ventricular impairment there was some advantage of cardiac resynchronisation therapy (CRT) with cardioverter defibrillator compared to CRT alone. Further research was needed. Questions about the quality of included data and methods of synthesis mean that the authors' conclusions should be treated with caution. The stated need for further research appeared appropriate.

To assess whether cardiac resynchronisation therapy combined with cardioverter defibrillator (CRT-D) is better than cardiac resynchronisation therapy (CRT) alone in people with heart failure and a wide QRS complex.

MEDLINE, EMBASE, Science Citation Index and Cochrane Central Register of Controlled Trials (CENTRAL) were searched from 1970 to September 2010. Search terms were reported. Conference proceedings from four relevant associations and internet sources on cardiology were checked.

Studies of at least two weeks in people with impaired left ventricular systolic function (ejection fraction <35%) and a wide QRS complex (QRS duration >120ms) that compared CRT to CRT-D were eligible for inclusion. The primary outcome of interest was all-cause mortality (at the longest available follow-up, up to one year or more than one year). Sudden cardiac death and heart failure related mortality were reported.

In the included studies between 67% and 93% of participants were men and mean ages ranged from 64 to 69 years. Where stated between 40% and 68% had ischaemic heart disease, left ventricular ejection fraction (LVEF) ranged from 21% to 28%, 76% to 100% had New York Hearth Association (NYHA) heart failure Class III-IV and mean QRS intervals ranged from >120ms to 160ms. Most participants were reported to be on optimal pharmaceutical therapy.

Two reviewers independently assessed studies for inclusion.

Quality was assessed on concealment of treatment allocation (blinding of participants and investigators)scored as yes, no or unclear and the MOOSE guidelines. These were based on assessments of selection, performance, detection and attrition bias and each element was classified as low, moderate or high risk of bias. Details were not presented.

Two reviewers independently extracted data. Discrepancies were re-evaluated by one reviewer.

Data were extracted in order to calculate odds ratio (OR) and 95% confidence intervals (CI).

Two reviewers independently extracted data. Discrepancies were re-evaluated by one reviewer. Authors were contacted for missing information.

Pooled odds ratios and 95% confidence intervals were calculated using a random-effects model. Heterogeneity was assessed using Ι². Sensitivity analyses investigated effects of study design. Funnel plots and Egger's test were used to assess for publication bias.

Seven studies (3,404 participants, range 96 to 1,212) were included: one (1,212 participants) randomised controlled trial (RCT) and six (2,192 participants) observational studies. Mean follow-up ranged from at least seven months to 58 months

The RCT was at moderate risk of selection and attrition bias and low risk of detection and performance biases. One observational study was at high risk of selection bias and others were at moderate risk. One study was at high risk of attrition bias and others at low risk. One study was at moderate risk of performance bias and the others were at low risk. All were at low risk of detection bias. Tests showed no evidence of publication bias.

At longest follow up, compared to CRT, CRT-D was associated with a reduction in all-cause mortality (OR 0.59, 95% CI 0.43 to 0.81; Ι²=63.9%; six studies), sudden cardiac death (OR 0.20, 95% CI 0.07 to 0.59; Ι²=64.7%; four studies) and heart failure-related mortality (OR 0.60, 95% CI 0.47 to 0.76; Ι²=0%; four studies). There was a reduction in all-cause mortality after one year (Ι²=64.9%; three studies) but not during the first year (Ι²=6.7%; six studies). There was a reduction in sudden cardiac death after one year (Ι²=75.3%; three studies). There was insufficient data for further analyses related to length of follow-up.

Sensitivity analyses that excluded the RCT showed similar results to the main analyses except for heart failure-related mortality which showed no statistically significant difference between treatment groups.

Figures indicated that results for the RCT alone showed no statistically significant difference between treatment groups for all-cause mortality at one year and at longest follow up.

Evidence demonstrated some advantage of CRT-D over CRT, such as all-cause death rate after one-year follow-up and cardiac death, in patients with left ventricular impairment. These findings needed to be verified by larger, randomised, prospective trials with extended patient follow-up.

The aims of this review were clearly stated in terms of inclusion criteria. The search included unpublished studies. It was not clear whether language restrictions were applied so it was not possible to say whether language bias affected the review. Tests for publication bias were likely to be unreliable given the small number of studies. The methods of the review aimed to reduce risks of reviewer error and bias. Details of the methods used to assess quality were only partly reported so it was difficult to comment on the quality of included data.

It may not have been appropriate to combine data from studies of differing design. Considerable statistical heterogeneity existed for some analyses. Details of included participants were limited but event rates differed widely between studies and this suggested that there may have been clinical differences between studies. Most data came from observational studies (considered of lower quality as other factors may have affected outcomes). Sensitivity analyses looked at results for lower rather than higher quality studies. Contrary to the main results, those for the higher quality RCT showed no effect on all-cause mortality.

Questions about the quality of included data and methods of synthesis mean that the authors' conclusions should be treated with caution. The stated need for further research appeared appropriate.

Practice: The authors stated that they were unable to make any strong recommendations from evidence from recent studies.

Research: The authors stated that large multicentre RCTs with longer follow-up were needed to compare the effects of CRT-D to CRT in people with heart failure and assess which clinical subgroups of people with heart failure would benefit most.

Natural Science Foundation of China; Shanghai Rising-Star Program; Shanghai Med-X program.

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.