Is early ventricular dysfunction or dilatation associated with lower mortality rate in adult severe sepsis and septic shock? A meta-analysis

Abstract

Introduction: Reversible myocardial depression occurs early in severe sepsis and septic shock. The question of whether or not early ventricular depression or dilatation is associated with lower mortality in these patients remains controversial. Most studies on this topic were small in size and hence lacked statistical power to answer the question. This meta-analysis attempted to answer the question by increasing the sample size via pooling relevant studies together.

Methods: PubMed, Embase (and Medline) databases and conference abstracts were searched to July 2012 for primary studies using well-defined criteria. Two authors independently screened and selected studies. Eligible studies were appraised using defined criteria. Additional information was sought the corresponding authors if necessary. Study results were pooled using random effects models. Standardized mean differences (SMD) between survivor and non-survivor groups were used as the main effect measures.

Results: A total of 62 citations were found. Fourteen studies were included in the analysis. The most apparent differences between the studies were sample sizes and exclusion criteria. All studies, except four pre-1992 studies, adopted the Consensus definition of sepsis. Altogether, there were >700 patients available for analysis of the left ventricle and >400 for the right ventricle. There were no significant differences in left ventricular ejection fractions, right ventricular ejection fractions, and right ventricular dimensions between the survivor and non-survivor groups. When indexed against body surface area or body height, the survivors and non-survivors had similar left ventricular dimensions. However, the survivors had larger non-indexed left ventricular dimensions.

Conclusion: This meta-analysis failed to find any evidence to support the view that the survivors from severe sepsis or septic shock had lower ejection fractions. However, non-indexed left ventricular dimensions were mildly increased in the survivor group but the indexed dimensions were similar between the groups. Both survivors and non-survivors had similar right ventricular dimensions.

Figure 1

Workflow of studies identification.

Figure 2

Standardized mean difference (SMD) for LV function in survivor and non-survivors. Forest plot showing the SMDs for different studies and the overall SMD. Negative or positive SMDs imply smaller or larger LVEF or LVFAC in the survivors. TEE, transesophageal echocardiography; TTE, transthoracic echocardiography.

Figure 3

Funnel plot for small-study effects. Funnel plot showing possible publication bias due to small-study effects. (A) Funnel plot for LV function studies. (B) Funnel plot for LV dimension study.

Figure 4

Standardized mean difference (SMD) for LV dimension in survivor and non-survivors. Forest plot showing the SMDs for different studies and the overall SMD. Studies were divided into two groups: Indexed, where LV dimension were indexed to body surface area or height, and not indexed. Subtotal SMD for each group are also shown. Negative or positive SMDs imply smaller or larger LV dimension in the survivors.

Figure 5

Standardized mean difference (SMD) for RV function and dimension in survivor and non-survivors. Forest plot showing the SMDs for different studies and the overall SMD. Upper panel, Forest plot for RV function; lower panel, Forest plot for RV dimension. Negative or positive SMDs imply smaller or larger LV dimensions in the survivors.