Criteria abstracted from The Users' Guide to Medical Literature, from the Health Information Research Unit and Clinical Epidemiology and Biostatistics, McMaster University

Highlighted lines and questions below provide links to the pertinent description of criteria in The EBM User's Guide, now available at the Canadian Centres for Health Evidence

Article Reviewed:

Please visit the new Evidence Based Journal Club Reviews

Goal-directed medical therapy and point-of-care testing improve outcomes after congenital heart surgery.

Rossi AF, Khan DM, Hannan R, Bolivar J, Zaidenweber M, Burke R.

Intensive Care Med. 2005 Jan;31(1):98-104. [abstract]

Reviewed by Myke Drayer MD and Rick Harrison MD, Mattel Children's Hospital, University of California, Los Angeles

Review posted May 24, 2005

I. What is being studied?

Study objective:

To assess if goal-directed therapy based on serial lactate values obtained from a point-of-care testing device reduced mortality in patients after congenital heart surgery.

Study design

This study was a prospective study of all neonatal and pediatric patients undergoing congenital heart surgery from July 2001 through September 2003 (group B, n = 710) utilizing a goal-directed therapy in the postoperative period. This group was compared to a historical cohort from June 1995 through June 2001 (group A, n = 1,656).

Outcomes assessed:

Mortality was compared between the study group and the historical cohort. In addition, these groups were stratified by age and surgical risk using the RACHS-1 (Risk Adjustment in Congenital Heart Surgery) system and mortality was compared between the sub-groups.

II. Are the results of the study valid?

1. Are the outcome measures accurate and comprehensive?

Death is obviously an accurate and objective outcome measure. However, the time period in which death occurred was not defined. It appears to be in-hospital post-operative mortality, but this was not specified. Another significant outcome the authors failed to mention was rescue therapy including ECMO. Since a large database was used, interrater reliability and other measures of the data's veracity should have been mentioned.

2. Were the comparison groups similar with respect to important determinants of outcome, other than the one of interest, and were residual differences adjusted for in the analysis?

There were many differences between the study and comparison groups that may have influenced outcome. First, the study group had longer cardiopulmonary bypass and aortic cross-clamp times and was younger and smaller. One would expect this to increase mortality in the study group, so this was not a significant problem. The authors should have reported whether there was a significant difference in number of patients with single ventricle lesions between groups.

However, the major variations between the two groups arise when one takes into account the use of a historical cohort and the fact that the study spanned an 8 year period. Although all of the patients were cared for in the same intensive care unit, they may not have all received the same care. The authors claim that surgical technique did not change over the study period, but there could have been variation between surgeons. In contrast, they admit that improvements in technology and equipment with regards to surgical care may have led to lower mortality in the study group. The actual frequency of serum lactate measurements should have been reported for both groups.

Another large area of variability is the care provided in the intensive care unit. The authors admit that leadership in the ICU did change during the study period and this could have had an impact on mortality. Also, the protocol allowed for the physician to escalate care in a way they thought would best improve oxygen delivery. This obviously allows for a tremendous amount of variability in care, especially since care was provided by different attendings as well as "physician extenders." Other variables that may have impacted post-operative care are differences in nursing care, respiratory therapy, medications, type of resuscitation fluid administered, and different uses of the same medication over time. Intensity of interventions between groups (e.g., TISS score) should have been compared between groups. It is difficult to fathom that all clinical decisions were driven by the lactate alone.

Another source of variability relates to the risk of the surgical procedure, but the authors stratified comparison groups by age and also by RACHS-1 score, which stratifies cardiac lesion according to surgical risk. Mortality was compared between these subgroups in order to account for prognosis. Additional factors that may influence prognosis that were not addressed include emergent vs. planned operations, pre-operative medications and changes in cardioplegia.

III. What are the results?

1. What are the recommendations?

The mortality for all patients in the study group was significantly lower than in the comparison group (1.8% v 3.7%, p < 0.02). Although not provided in the article, one can calculate an absolute risk reduction of 1.9% (95% CI; 0.4-2.9%) and relative risk reduction of 51% (95% CI; 11-72%).

When subdivided by age, there was no difference in patients over one year of age, but there was a significant difference in infants (p = 0.008) and neonates (p = 0.001). When subdivided by surgical risk (RACHS-1 scores), there was no difference in low risk patients (RACHS-1 score 1 and 2), but there was a significant difference in medium (scores 3 and 4) and high risk (scores 5 and 6) patients. Only p values were provided with regards to the stratified group and precise values were not available. Using an approximation from the figures, relative risk was estimated for each group. With regards to age, relative risk of 0.32 was calculated for the infant group (mortality rates of ~3% vs. ~1%) and relative risk of 0.27 was calculated for the neonate group (~13% vs ~3.5%). With regards to surgical risk, relative risk of 0.5 was calculated for RACHS-1 scores 3 and 4 (mortality rates of ~6% vs. ~ 3%) and relative risk of 0.3 was calculated for groups 5 and 6 (~30% vs. ~8%). Although this was not a randomized control trial, it would have been beneficial had the authors provided meaningful metrics of effect size, like relative risk and absolute risk reduction as well as confidence intervals.

In conclusion, the combination of point-of-care testing and a goal-directed therapy protocol using serial blood lactate levels is associated with decreased mortality especially in younger patients and those undergoing high-risk surgeries.

IV. Will the results help me in caring for my patients?

1. How will the recommendations help you?

Without a reliable measure of cardiac function and oxygen delivery, serial lactates are likely a good gauge of the effectiveness of medical therapy. Point-of-care testing may be beneficial due to easy access and rapid turn around time, especially in the immediate post-operative period when many changes occur. However, many variables in this study make it difficult to trust that this protocol alone could explain the decrease in mortality after initiation of the protocol. It is possible that the institution of goal-directed therapy standardized care in this busy CVICU and served as the main cause for the improvement in mortality witnessed.

 

Comments

Back to the EB Journal Club Index