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

Oxygen transport in critically ill infants after congenital heart operations.

Rossi AF, Seiden HS, Gross RP, Griepp RB.

Ann Thorac Surg. 1999;67(3):739-44. [abstract] [full-text] [discussion on the Ann Thoracic Surg website]

Reviewed by Amr Shahin MD, King Khaled National Guard Hospital, Jedda, Saudi Arabia

Review posted December 27, 2003

I. What is being studied?

Study objective:

Examine whether variables of oxygen transport can serve as outcome predictors for critically ill infants following operations for congenital heart disease.

Study design

A retrospective review of the records of infants that were less than 1 year old (range 1 day to 1 year) when they underwent heart surgery during a period of 16 months at the authors' institution. The above patients were included if they had simultaneous assessment of arterial blood gas (ABG) and mixed venous saturation (SvO2) immediately following the operation (on admission to the PICU) as well as 6 and 24 hours thereafter. Calculations of arterial pH, base deficit, arterio-venous oxygen saturation difference (AVDO2) and oxygen extraction ratio (OER) were obtained. The above data was analyzed and a comparison of the trends as it related to patient survival was done. Patient survival was defined as that to hospital discharge.

II. Are the results of the study valid?

Note: These questions follow from Randolph AG et al. Understanding articles describing clinical prediction tools. Crit Care Med 1998;26:1603-1612. [abstract]

1. Was a representative group of patients completely followed up?

A consecutive group of patients (n=49) that represents the target population (infants after congenital heart operations) was studied. There were 69 other infants less than one year of age who had cardiac surgery who were not included, presumably because they did not have the appropriate measurements performed. There is likely a systematic difference between those patients who were studied compared to the larger group. It may also be argued that the variability of cardiac anatomical lesions and procedures does not lend homogeneity to the population studied. Is it appropriate to pool a VSD repair with a Stage I Norwood procedure?

2. Was follow-up sufficiently long and complete?

Follow up was presumably complete for the 49 patients recruited retrospectively for the study, although this is not explicitly stated. There is no mention of whether mortality was ICU mortality or in-hospital mortality. These could be considered the derivation set. There was no validation set.

3. Were all potential predictors included?

No. The purpose of the study was to examine the role of assessment of direct and indirect variables of oxygen transport as a predictor of outcome. The variables studied and analyzed were arterial pH, arterial base excess, arterial oxygen saturation, arterio-venous oxygen saturation difference, and the oxygen extraction ratio. No other clinical variables such as preoperative well being, intraoperative variables, anesthesia, cardiac muscle function parameters, inotropic requirement, intercurrent infections if any, requirement for blood product transfusion, and the state of inanition were evaluated. The authors are not attempting to create a predictive tool, but rather to examine whether variables of oxygen transport could serve as a component of a prediction tool model, or maybe even an independent predictor of outcome. The greater the number of relevant variables included, the more likely the investigators will generate a satisfactory predictive tool.

4. Did the investigators test the independent contribution of each predictor variable?

Derived and measured oxygen transport variables would collectively constitute one predictor variable studied by the authors, namely apparent adequacy of oxygen delivery.

5. Were outcome variables clearly and objectively defined?

Yes, these were survival and the lack thereof.

III. What are the results?

There were 49 patients studied, of which there were 39 survivors and 10 nonsurvivors. The age at operation was significantly less in nonsurvivors than survivors (0.44 months versus 3.2 months, p < 0.01). The main results of the study are as follows: No significant differences were found between survivors and nonsurvivors both at admission and at 24 hours for all variables of oxygen transport studied (arterial pH, base excess, AVDO2 difference, and OER).

At 6 hours, arterial pH was higher in survivors (7.48 versus 7.34, p <0.001) as was base excess (2.9 versus -4.3, p < 0.01). The mean AVDO2 at 6 hours was lower in survivors than nonsurvivors (34 versus 43, p < 0.05), as was the OER calculated at 6 hours (0.28 versus 0.53, p < 0.001). Significantly there was higher survival in patients with OER < 0.5 at all times and there was a significant association between nonsurvival and higher OER. Specifically, the mortality of patients with an OER > 0.5 at 6 hours was 70%.

This may in fact be what the clinicians caring for the patients did. The abnormal values at 6 hours were corrected by 24 hours, however, many of these patients still died. Likely the values were corrected at 24 hours by active interventions between 6 and 24 hours.

1. What is(are) the prediction tool(s)?

No real prediction tool was developed. Evidence of insufficient oxygen transport at 6 hours postoperatively appears to portend a higher risk of dying. More specifically the authors suggest the use of the OER at 6 hours of at least 0.5 as a predictor of a higher risk of mortality. The findings of the study are interesting but they do not satisfy the criteria needed in an effective comprehensive outcome predictive tool. Evidence of insufficient oxygen transport can qualify as one of the prognostic variables that based on common sense and also the results of the study could be incorporated into a model that tests multiple other variables. This of course after selecting a larger representative group and the properly layering risk groups, and statistically and exhaustively testing the independent contribution of each variable. Specifically, it would have been useful to prospectively separate out the Stage I HLHS repair patients given their known significantly increased risk of mortality postoperatively (at most centers).

2. How well does the model categorize patients into different levels of risk?

It doesn't. The authors suggest that clinicians can use abnormal oxygen transport values in patients after heart surgery as predictors of outcome. They also propose that assessment of such variables can guide the management process by identifying patients at higher risk requiring more vigorous intervention perhaps before tissue dysoxia sets in. However, this is purely speculative as an intervention study would be required to prove this suggestion.

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

1. Does the tool maintain its prediction power in a new sample of patients?

A validation set would be needed.

2. Are your patients similar to those patients used in deriving and validating the tool(s)?

Yes, although as noted, the heterogeneous and skewed sample (HLHS Stage I and Tetralogy of Fallot represented approximately half the sample) might not be representative of every center's population.

3. Does the tool improve your clinical decisions?

Yes. Yet only inasmuch as the study validates our current common sense practice and knowledge of pathophysiology - patients with impaired oxygen delivery do worse than those with intact oxygen delivery. No real prediction tool was developed.

 

Comments

Back to the EB Journal Club Index