Criteria abstracted from The Users' Guides to the Medical Literature series in JAMA
Outcome of severe sepsis in pediatric oncology patients.
Fiser RT, West NK, Bush AJ, Sillos EM, Schmidt JE, Tamburro RF.
Pediatr Crit Care Med. 2005 Sep;6(5):531-536..[abstract]Reviewed By: Danny Castro DO, Baylor College of Medicine, Texas Children
Review posted July 07, 2008
- What is being studied?:
- The study objective:
This study sought to determine the outcome of pediatric oncology patients with severe sepsis at two different time periods.
- The study design:
This is a retrospective analysis of a heterogeneous group of pediatric and young adult oncology patients over a 13 year period from a single ICU at a pediatric oncology hospital.
- Are the results in the study valid?
- Was there a representative and well-defined sample of patients at a similar point in the course of the disease?
The patient sample was selected by screening of an ICU admission log and was then followed by a chart review. The “screening” process is not clearly described and there is a theoretical possibility that potential subjects could have been missed.
Patients were subsequently chosen if they had a primary diagnosis of cancer. Patients were defined as having severe sepsis or septic shock according to the definitions outlined by the 1992 ACCP/SCCM Consensus Conference Committee (1). Inclusion criteria were:
- Newborn to 25 years of age
- Admitted to the ICU during the period of 1/1/90 to 12/31/02
- Primary diagnosis of cancer
- Diagnosis of severe sepsis (based on fever, clinically suspected or microbiologically proven infection AND at -least 2 of the following: tachycardia, tachypnea, peripheral perfusion abnormalities, altered mental status, hypotension or acidosis at the time of ICU admission
- Receipt of IVF boluses of > 30ml/kg (total) for hypoperfusion or receipt of a dopamine infusion of > 5mcg/kg/min for inotropic support
No exclusion criteria were described
Subjects were at variable points in the course of their oncologic disease. Additionally, various oncologic diseases can carry various prognoses depending on multiple variables, including age. The authors’ inclusion criteria included a very wide age range. Thus, grouping various oncologic diseases leads to an infinite number of prognoses which could affect the measured outcome. Furthermore, multiple therapeutic regimens with various degrees of toxicity may also affect the measured outcome. Therefore, even if at similar points in their sepsis, subjects may be at different points in their oncologic disease which could affect their outcome.
Otherwise, the patients involved in this study were a fair representative sample of the underlying population and were well-defined, within reason.
- Was follow-up sufficiently long and complete?
The authors described survival at ICU discharge and at 6 months after discharge from the ICU. Follow up at time of discharge from the ICU was 100%. Only four admissions were not available for 6 month follow up at the time of analysis. Therefore, it seems the number of subjects lost to follow-up would not have a large effect on the calculated outcome. Information regarding the demographics and/or characteristics of the admissions lost to follow up were not provided.
- Were objective and unbiased outcome criteria used?
Yes. The measured outcome event of this study was survival at ICU discharge and 6 months after ICU discharge.
- Was there adjustment for important prognostic factors?
These authors identified a subgroup of patients who progressed to septic shock and required mechanical ventilation (MV) and inotropic support. The authors analyzed this subgroup, via a univariate analysis, to determine factors associated with ICU mortality. They compared 13 variables between survivors and non-survivors. The factors they included in their analyses were 1) months from oncologic diagnoses, which could be considered as a surrogate indicator of disease course or cumulative toxicity from therapy, 2) use of multiple inotropes, which could be assumed to be a surrogate for persistent hemodynamic instability, 3) PRISM III score, which could, presumably, be a surrogate for admission acuity and 4) admission era, which could be of significance since the 13 year period was prior to any published consensus on the management of sepsis. Additional factors included in their analyses were the percentage of patients who underwent bone marrow transplant (BMT), presence of fungal sepsis and positive blood cultures, presence of hepatic/renal failure plus max. organ system failure (OSF) score, age/sex and presence of ANC < 500.
Those variables with a P value of less than 0.1, in the univariate analysis, were then used in a multivariate, logistic regression model to determine variables significantly associated with mortality. After the multivariate analysis, the variables that proved to be of significance in predicting ICU mortality in this subgroup of patients were patients who had underwent BMT, presence of fungal sepsis, use of multiple inotropes and PRISM III score.
It would have been interesting to have mechanical ventilation and/or ALI/ARDS included as an independent variable since these entities are know to have an impact on mortality. However, the authors had a limited number of subjects in their subgroup, therefore, to limit error, the number of variables that could be applied was also limited. I presume they chose the ones that were of greatest interest, however, the fact that everyone in the subgroup was mechanically ventilated could already have accounted for the greatest effect on outcome. Thus, displaying it as an independent variable would have been of interest.
Additionally, duration of neutropenia, presence/absence of indwelling catheter and length of ICU stay would have been other factors to consider. Kutko et al. did a retrospective study, of which 71% of their admissions were patients with oncologic illness and 19% of those were post-BMT. The presence of neutropenia was not found to be a predictor of mortality, but, duration of neutropenia was also not included. Length of ICU stay for the survivors was half the duration of non-survivors.
- What are the results?
- How large is the likelihood of the outcome event(s) in a specified period of time?
Overall, 83% of pediatric oncology admissions with severe sepsis, upon admission to the ICU, were alive upon ICU discharge. Those admissions who were post-BMT, had a lower survival rate (70%) as compared to those who were not post-BMT (88%). Additionally, admissions who progressed to septic shock and required mechanical ventilation and inotropic support had a lower survival rate (36%) to ICU discharge. Again, those who were post-BMT had a lower survival rate (26%) as compared to those who were not post-BMT (44%).
Six months after ICU discharge, 73% of pediatric oncology admissions with severe sepsis were alive.
Also, at 6 months after ICU discharge, 39% of patients who were post-BMT were alive and 69% of patients who were not post-BMT were alive.
The subgroup, who progressed to septic shock and required mechanical ventilation and inotropic support, was further analyzed in a multivariate analysis to determine which factors would be associated with ICU mortality. The factors associated with mortality were (in descending order of odds ratio):
- fungal sepsis (OR 10.7, [95% CI 1.2, 94.4])
- use of multiple inotropes (OR 4.1, [95% CI 1.4, 11.8])
- BMT (OR 2.9, [95% CI 1.1, 7.4])
- PRISM III score (OR 1.1, [95% CI 1.0, 1.2])
- How precise are the estimates of likelihood?
Having fungal sepsis had a significantly higher odds of predicting mortality, however, the confidence interval was extremely wide (95% CI 1.2, 94.4). Thus, indicating less precision in their estimate and possibly requiring a greater sample size in order to obtain a more precise estimate.
The PRISM score is a predictor of mortality in the PICU and the OR above indicates the odds of dying for each one point increase in PRISM score. The association of the PRISM III score with mortality had the narrowest confidence interval (95% CI 1.0, 1.2) and, therefore, was very precise. But, practically speaking, this estimate does not indicate that those with higher PRISM III scores are any more likely to perish than those with lower PRISM III scores. Of note, the study period encompassed a time when PRISM scoring was not routinely employed clinically, therefore, PRISM III scores were retrospectively assigned for patients in this study.
The use of multiple inotropes and being a post-BMT patient had relatively narrow confidence intervals and, therefore, are fairly precise estimates of predicting mortality. However, similar to the other variables found to be significant, they have confidence intervals with the low end of their ranges being slightly greater than one.
In summary, with the exception of the PRISM III score, these variables have wide confidence intervals and thus indicate a lack of precision.
- Will the results help me in caring for my patients?
- Were the study patients and their management similar to my own?
St. Judes Children’s Research Center is a freestanding, tertiary hospital that specializes in treating children with “catastrophic” diseases, within the pediatric subspecialties of allergy/immunology, infectious disease and oncology. This probably is reflective of other large programs with pediatric oncology and bone marrow transplant programs. Although no demographic information regarding the various oncologic diagnoses of the subjects were provided, the presumption is that a similar variety of oncologic patients are present at many institutions and are admitted to ICU’s for sepsis.
Management of subjects during the study period was likely very different from how patients with sepsis are now managed at most institutions. The study period ended just prior to release of published consensus guidelines on the management of severe sepsis and septic shock. Thus, the management of severe sepsis and septic shock have since evolved, as well as, other areas in the management of severe sepsis and septic shock, e.g., lung protective strategies, hyperglycemic control, etc.
- Will the results lead directly to selecting or avoiding therapy?
No. The results of this study will not directly lead to selecting or avoiding therapy. Although there is always room for improvement, the overall survival rates in this population of patients are within an acceptable range. Despite the relatively lower survival rate in the BMT subgroup, as opposed to the non-BMT subgroup, it is still within an acceptable range where limiting care would not be routine.
- Are the results useful for reassuring or counseling patients?
The results regarding ICU survival are useful in that the overall mortality (83%) in these particular study subjects are in line with current estimates (2,3). Therefore, implying that they carry the same prognosis as any other pediatric patient with severe sepsis. However, when comparing those patients who underwent BMT versus those who did not, the BMT patients carried a significantly lower survival rate (70% vs. 88%). This is useful in that it may help provide patients/parents with realistic goals and expectations of outcome.
Furthermore, in the subgroup who progressed to septic shock and required both mechanical ventilation and inotropic support, survival was 36% with BMT patients carrying a significantly lower survival rate than non-BMT patients (26% vs. 44%). Again, the results indicate a lesser likelihood of survival in those who have progressed to septic shock and are both mechanically ventilated and receiving inotropic support. This may help family members make informed decisions regarding care and/or continuation of care.
- Bone RC, et al. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis: The ACCP/SCCM Consensus Conference Committee. American College of Chest Physicians/Society of Critical Care Medicine. Chest. 1992; 101: 1644-1655
- Kutko MC, et al. Mortality rates in pediatric septic shock with and without multiple organ system failure. Pediatr Crit Care Med. 2003; 4(3): 333-337
- Angus DC, et al. Epidemiology of severe sepsis in the United States: analysis of incidence, outcome, and associated costs of care. Crit Care Med. 2001; 29(7): 1303-1310
Last Updated: July 07, 2008