Outcomes Associated With Timing of Neurologic Dysfunction Onset Relative to Pediatric Sepsis Recognition

Abstract

Objectives: To compare outcomes associated with timing-early versus late-of any neurologic dysfunction during pediatric sepsis.

Design: Secondary analysis of a cross-sectional point prevalence study.

Setting: A total of 128 PICUs in 26 countries.

Patients: Less than 18 years with severe sepsis on 5 separate days (2013-2014).

Interventions: None.

Measurements and main results: Patients were categorized as having either no neurologic dysfunction or neurologic dysfunction (i.e., present at or after sepsis recognition), which was defined as Glasgow Coma Scale score less than 5 and/or fixed dilated pupils. Our primary outcome was death or new moderate disability (i.e., Pediatric Overall [or Cerebral] Performance Category score ≥3 and change ≥1 from baseline) at hospital discharge, and 87 of 567 severe sepsis patients (15%) had neurologic dysfunction within 7 days of sepsis recognition (61 at sepsis recognition and 26 after sepsis recognition). Primary site of infection varied based on presence of neurologic dysfunction. Death or new moderate disability occurred in 161 of 480 (34%) without neurologic dysfunction, 45 of 61 (74%) with neurologic dysfunction at sepsis recognition, and 21 of 26 (81%) with neurologic dysfunction after sepsis recognition (p < 0.001 across all groups). On multivariable analysis, in comparison with those without neurologic dysfunction, neurologic dysfunction whether at sepsis recognition or after was associated with increased odds of death or new moderate disability (adjusted odds ratio, 4.9 [95% CI, 2.3-10.1] and 10.7 [95% CI, 3.8-30.5], respectively). We failed to identify a difference between these adjusted odds ratios of death or new moderate disability that would indicate a differential risk of outcome based on timing of neurologic dysfunction (p = 0.20).

Conclusions: In this severe sepsis international cohort, the presence of neurologic dysfunction during sepsis is associated with worse outcomes at hospital discharge. The impact of early versus late onset of neurologic dysfunction in sepsis on outcome remains unknown, and further work is needed to better understand timing of neurologic dysfunction onset in pediatric sepsis.

Figure 1.. Baseline Pediatric Overall Performance Category (POPC) and Pediatric Cerebral Performance Category (PCPC) Scores by presence and timing of neurologic dysfunction.

The distribution of POPC scores was different between the three groups (p=0.03, Fisher’s exact test), but there was no difference in the distribution when comparing neurologic dysfunction by timing of onset (p=0.38, Fisher’s exact test). The difference in PCPC score distribution was not different between the three groups (p=0.05, Fischer’s exact test).

Figure 2.. Proportion of patients who experienced death or moderate disability based on day of neurologic dysfunction recognition.

Graphical representation of outcomes based on timing of neurologic dysfunction onset in the first seven days of severe sepsis. Neurologic dysfunction at sepsis recognition is represented as day 1 of neurologic dysfunction onset (n=61). Neurologic dysfunction after sepsis recognition was stratified into three groups based on onset after sepsis recognition: days 2-3 (n=17), days 4-5 (n=6) and days 6-7 (n=3). The proportion of patients who died or had moderate disability was not different based on the specific day of sepsis recognition (p=0.22, Fisher’s exact test).

Figure 3.. Proportion of outcomes following pediatric sepsis based on presence and timing of neurologic dysfunction.

All outcomes were determined at hospital discharge. Moderate disability was defined as POPC or PCPC score ≥3 and ≥ 1 increase from baseline and mild disability was any increase in POPC or PCPC score from baseline that does not meet criteria for moderate disability. Survivors without an increase in POPC or PCPC score were labeled as no new disability. The bars are labeled with the percentage of patients within each group that experienced the outcome. The total size for each group is displayed. The proportion of patients with each of these outcomes was different between the three groups (p<0.001, Fisher’s exact test) but was no different based on the timing of neurologic dysfunction during sepsis (p=0.96, Fisher’s exact test).