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

Outcome of out-of-hospital cardiac or respiratory arrest in children.

Schindler MB, Bohn D, Cox PN, McCrindle BW, Jarvis A, Edmonds J, Barker, G.

N Engl J Med 1996;333:1473-9. [abstract]

Reviewed by Adrienne Randolph

Review posted March 17, 1997

I. What is being studied?:

The study objective:

To determine the survival rate and predictors of survival among children after out-of-hospital cardiac arrest.

The study design:

Retrospective cohort.

The outcomes assessed:

Survival to hospital discharge and neurologic status at one year.

II. Are the results in the study valid?

Primary questions:

1. Was there a representative and well-defined sample of patients at a similar point in the course of the disease?

Yes. The 101 patients reported were children, ages 3 days to 18 years (median 2 years), who suffered cardiac or respiratory arrest outside of the hospital. Upon arrival in the emergency room, all patients had no spontaneous respiration and 80 were pulseless. Patients were identified retrospectively by reviewing all deaths that occurred in the emergency room and all admissions to the intensive care unit after CPR for January 1986 through June 1993.

Because reporting of deaths occurring in the hospital is usually very accurate, it is unlikely any deaths in the emergency room were missed. The completeness of identification of patients surviving CPR in the emergency room is dependent upon how accurate that is recorded. The authors do not state whether the emergency room kept a log of all patients who received CPR, if a log was kept by the intensive care unit upon admission, or if they identified patients from a diagnosis of CPR in their discharge abstracts.

2. Was follow-up sufficiently long and complete?

Yes. Surviving patients were followed for 12 months. Most patients who will regain neurologic function will do so within the first 12 months (1). No patients were lost to follow-up.

Secondary questions:

3. Were objective and unbiased outcome criteria used?

Yes. The outcomes were survival, and neurologic status of survivors at 12 months. A full neurologic assessment took place in the outpatient clinic at 12 months. The outcomes of survivors were classified according to the Pediatric Cerebral and Overall Performance Category Scores, which are validated scales describing the functional outcomes of children (2).

4. Was there adjustment for important prognostic factors?

Yes. Using logistic regression analysis, the authors first tested potential prognostic factors individually, and then tested those factors significant at the individual level using multiple logistic regression to see if they were still independent predictors. Age, weight, sex initial arterial blood gas values, interval between the arrest and arrival to the hospital, the duration of resuscitation in the emergency department, the number of doses of epinephrine and bicarbonate used in the resuscitation, the presence of preexisting illness, and the presence of asystole in the emergency department were all tested for predictiveness.

III. What are the results?

1. How large is the likelihood of the outcome event(s) in a specified period of time?

Of the 101 patients, 80 had a cardiac and respiratory arrest and 21 never lost their pulse (respiratory arrest only). 37 patients died in the emergency room and 64 were transferred to the ICU with 15 surviving to hospital discharge and 13 alive at 12 months after discharge. Of the 21 patients with a respiratory arrest only, 9 survived to hospital discarge (43%) and 5 were neurologically normal. Of the 80 patients with a cardiac arrest only, only 6 patients (8%) survived to hospital discharge and all 6 had moderate-to-severe neurologic sequelae.

Age, weight, sex, and initial arterial blood gas values were not significant predictors of survival to hospital discharge. Patients who survived were more likely to have had a pulse present upon arrival and a significantly shorter interval between the arrest and arrival at the hospital. They also had a shorter duration of resuscitation in the emergency department and received significantly lower doses of epinephrine and bicarbonate. They were also more likely to have a preexisting illness with a higher body temperature. However, when all of these individually significant factors were tested together using multiple regression, the only factors which were still independent predictors of survival were the use of only basic CPR (or no CPR) before arrival at the hospital (odds ratio 0.032) and a short duration of resuscitation in the emergency department (odds ratio 0.058).

2. How precise are the estimates of likelihood?

Confidence intervals are not reported for the main outcomes of survival, mortality, and neurologic function. For the 80 patients with cardiac arrest there were no patients with a normal or near normal outcome. Given this evidence, how confident can you be in the results that no patients arriving pulseless in the emergency room would have a good neurologic outcome? Using the rule of 3's, the upper limit of the 95% confidence interval around the likelihood of a normal or near normal neurologic outcome for patients with out-of-hospital cardiac arrest is 3/80 x 100 or 3.8%. You can be 95% certain that the likelihood is 3.8% or less of encountering a survivor of an out-of-hospital cardiac arrest with a normal or near-normal neurologic outcome.

For patients who suffered a respiratory arrest, the confidence interval around the likelihood of a positive outcome can be calculated using an equation. There were 5/21 patients (23.8%) suffering an out-of-hospital respiratory arrest who were neurologically normal and the confidence interval around this estimate is 5.6% to 42%.

The confidence intervals for the prognostic factor found to still be important independent predictors in the multivariate analysis are reported. The use of only basic CPR or no CPR had a 95% confidence interval around the odds ratio of 0.004 to 0.25 and a short duration of resuscitation in the emergency room had a confidence interval around the odds ratio of 0.01 to 0.33. [Read more about odds ratios at the NHS Research and Development Centre for Evidence-Based Medicine]

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

1. Were the study patients and their management similar to my own?

The characteristics of the children in the study represented a large variety of disorders and diagnoses leading to the arrest. 24/80 patients with a cardiac arrest were diagnosed with SIDS which means that they had no underlying problems which could be identified and 9/80 were victims of near drowning. Trauma patients made up 21% of the population. Arrythmias, cancer, seizures, asthma, cerebral hemorrhage, and drug overdose were other diagnostic categories underlying the arrest. Therefore, the sample was representative of a spectrum of underlying disorders in pediatric patients. The patient management took place in a large pediatric hospital with a trained pediatric emergency staff. You will need to estimate how similar their resuscitation practices may be to your own center and if this could make the outcomes potentially better or worse than those of your center.

2. Will the results lead directly to selecting or avoiding therapy?

It is clear the for patients with a primary respiratory arrest who are quickly resuscitated, the outcome is quite promising. For patients with a cardiac arrest outside of the hospital, the sample size is insufficient to rule out the likelihood of a normal survivor. In fact, with 80 patients there still could be a 3.8% probability of a normal survivor if you needed to be 95% certain. To be certain enough of the results to the point that therapy could be withdrawn at an early stage would require a larger sample size from a prospective multicenter study or, at the minimum, from a meta-analysis of multiple smaller studies.

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

Yes. For patients with a primary a respiratory arrest, there is still hope of a normal outcome and their neurologic exam needs to be closely monitored. For patients with an out-of-hospital cardiac arrest who have a neurologic exam consistent with minimal function, the likelihood of recovery of normal neurologic status is slim.

References

1. Kriel RL, Krach LE, Jones-Saete C. Outcome of children with prolonged unconsciousness and vegetative states. Pediatr Neurol 1993; 9:362-368. [abstract]

2. Fiser DH. Assessing the outcome of pediatric intensive care. J Pediatr 1992; 121: 68-74. [abstract]

 

Comments

Back to the EB Journal Club Index