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

Hypoalbuminaemia in critically ill children: incidence, prognosis, and influence on the anion gap.

Durward A, Mayer A, Skellett S, Taylor D, Hanna S, Tibby SM, Murdoch IA.

Arch Dis Child. 2003 May;88(5):419-22. [abstract]

Reviewed by Richard Lambert MD, University of Florida Shands Children's Hospital at Gainesville, FL

Review posted August 25, 2004

I. What is being studied?

Study objective:

This study had two aims:

1. document the incidence and prognostic value of hypoalbuminemia in pediatric patients over 24 hours following admission to the intensive care unit; and
2. determine the influence of correction for hypoalbuminemia on the diagnosis of a "raised anion gap" metabolic acidosis.

Study design

Prospective, descriptive study of 134 critically ill children admitted to a pediatric intensive care unit (PICU) over a 6 month period.

The patients included:

Patients less than 16 yrs of age with the presence of indwelling arterial lines were eligible for this study.

The patients excluded:

Postoperative cardiac surgery patients, prior administration of hyperoncotic albumin solution (20%), or concomitant use of total parenteral nutrition caused exclusion from this study. In addition, there were no patients with burns, nephrotic/nephritic syndrome, or protein losing enteropathy, which are all associated with significant hypoalbuminemia in critically ill patients.

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? Was follow-up sufficiently long and complete?

Yes. The group of patients was representative of medical and surgical admissions to tertiary PICU's. The investigators analyzed the data for all patients upon admission and 24 hours after admission, then followed them until discharge from their unit or death if that occurred first.

2. Were all potential predictors included?

No. The authors set out to analyze the influence of hypoalbuminemia on standard acid base analysis using anion gap methodology. In addition to the electrolytes needed to calculate anion gap, namely Na+, K+, Cl-, and the TCO2 (total concentration of CO2), they also collected blood lactate levels, and arterial blood gas samples. They then corrected for hypoalbuminemia, when present, using Figge's equation: AGcorr = AG + 0.25 X (normal serum albumin (40g/l) - observed albumin). This equation was further validated using a population of patients from their PICU and found to be accurate.

This was sufficient to accomplish their stated aims, but may not have been sufficient enough for a comprehensive evaluation of hypoalbuminemia in critically ill patients without analyzing the role of physiochemical changes in response to a low serum albumin. This method, more recently known as the Fencl-Stewart method, is getting an increasing amount of attention in the literature and may prove to be a very useful predictor for outcome in patients with complex metabolic abnormalities. Of note, the authors do mention that this was of interest to them but was beyond the scope of their study goals.

Additional predictors of mortality that could have been evaluated with this population include the updated risk of mortality scoring models like PRISM III (Pediatric RISk of Mortality) or PIM 2 (Pediatric Index of Mortality) as well as the recently validated model for measuring severity of outcome based on multi-organ dysfunction, the PELOD (PEdiatric Logistic Organ Dysfunction) scoring system. As discussed below, using a single predictor (i.e., serum albumin levels) for risk of mortality in complex clinical situations will likely not provide information comprehensive enough to make a thorough assessment of a patients' overall risk for mortality.

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

No. They only tested the prediction of mortality and anion gap using albumin levels with univariate analyses.

4. Were outcome variables clearly and objectively defined?

Yes. The outcome parameters were length of ventilation, length of PICU stay, survival to discharge and death. They were objective and unlikely to be open to subjective interpretation.

III. What are the results?

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

Patients were divided into two groups based on admission serum albumin concentrations, then each of these groups were further divided based on serum albumin concentrations at 24 hours after admission. Those with levels < 33 g/l were classified as hypoalbuminemic and those with levels > 33 g/l were normal. Patients whose levels were < 20 g/l were further classified as having extreme hypoalbuminemia.

They found that hypoalbuminemia (< 33 g/l) was more frequent at 24 hours than at admission; 76% (98/134) vs. 56.7% (76/134). There was no difference in mortality between the two groups (8/76 for hypoalbuminemic group and 7/58 for normal albumin group) and survival to PICU discharge was also similar. The relative risk (RR) for mortality between the groups was 0.9 with a 95% confidence interval (CI) of 0.3 to 2.3. There also was no difference in mortality when comparing those with extreme hypoalbuminemia to those with albumin levels > 20 g/l. Mean albumin concentrations were similar between survivors and non-survivors. The only variable with discriminatory value was length of stay days in PICU expressed as median (interquartile range); 4.9 (2.8-11.5) in hypoalbuminemic group versus 3.6 (1.9 to 5.7) in normal albumin group, p = 0.006.

They then compared the blood chemistry and acid-base data for patients with and without hypoalbuminemia on admission and at 24 hours. The major difference found was that in patients with metabolic acidosis (HCO3 < 22 mmol/l n=103), the incidence of a raised anion gap (> 18 meq/l) increased from 37.9% to 72.8% when corrected for hypoalbuminemia. Therefore, failure to correct for low albumin resulted in a consistent underestimation of the anion gap.

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

Poorly. The discriminatory ability of admission albumin concentrations for detecting PICU non-survival was weak with an area under the ROC curve of 0.50 (95% CI 0.32 to 0.62), and although the anion gap was grossly underestimated by not adjusting for hypoalbuminemia, the clinical significance of this was not discussed.

3. How confident are you in the estimates of the risk?

Not confident. As discussed above, there was no statistical difference in risk for mortality between the groups; RR = 0.9. The 95% CI for RR is wide (0.3 to 2.3), suggesting a lack of accuracy to the already insignificant risk. It should be noted that the 'true' RR as exemplified by the 95% CI ranges from the low value of 0.3, which could be interpreted as meaning hypoalbuminemia is actually associated with a reduced risk of death, to a high value of 2.3, suggesting hypoalbuminemia is a potentially meaningful indicator of increased risk for mortality.

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?

If the authors had found a predictor with good discriminatory power, they may have further tested it in an independent sample of patients. However, this study does not validate hypoalbuminemia as a single predictor of outcome in critically ill children, and therefore no independent sampling was warranted.

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

Yes.

3. Does the tool improve your clinical decisions?

Even though the authors were not successful in showing an increased risk of mortality in patients with hypoalbuminemia, I believe that their recognition and validation of serum albumin concentrations during routine calculation of anion gap may have clinical significance. Additionally, the authors are correct in pointing out that the clinical significance of underestimating the anion gap in hypoalbuminemic patients and its impact on the treatment of metabolic acidosis warrant further investigation. Clearly, single biochemical parameters are not likely to have long-standing predictive value in patients representative of modern day PICU's due to the increasing complexity and often dynamic condition regularly encountered in these critically ill children.

Rather, I would advocate for taking the next step and incorporating the calculation of hypoalbuminemia with the method of accounting for unmeasured anions in serum as done by Balasubramanyan, et al. (1) over 5 years ago, which revealed that mortality was more strongly associated with the presence of hidden unmeasured anions (area under the receiver operating characteristic curve (ROC) = 0.79; p = 0.0002 compared with lactate (ROC area = 0.63; p = 0.05), BE (ROC area = 0.53; p = 0.32), or anion gap (ROC area = 0.64; p = 0.08) in their patient sample (255 patients admitted to a University-based PICU). To date, there is no set consensus in either the adult or pediatric literature concerning the role of individual markers of acidosis in predicting mortality.

4. Are the results useful for reassuring or counseling patients?

No.

References:

1. Balasubramanyan N, Havens PL, Hoffman GM. Unmeasured anions identified by the Fencl-Stewart method predict mortality better than base excess, anion gap, and lactate in patients in the pediatric intensive care unit. Crit Care Med 1999; 27(8): 1577-81. [abstract]

 

Comments

Back to the EB Journal Club Index