Assessment of a Draft Streamlined Ventilator-Associated Pneumonia Definition: Data Collection Burden, Agreement with Current National Healthcare Safety Network Pneumonia Definitions, and Clinical Relevance

Background:   Ventilator-associated pneumonia (VAP) surveillance is hampered by current VAP definitions’ lack of validity and reliability. We assessed a draft streamlined VAP (s-VAP) surveillance definition, which uses objective criteria based on readily available clinical data.

Objective: 1) To determine agreement between s-VAP and National Healthcare Safety Network (NHSN)-defined VAP and compare case determination times (CDTs); 2) To compare outcomes of patients with s-VAP or NHSN VAP to outcomes of patients not meeting the definitions.

Methods: Criteria for s-VAP were: 1) mechanically ventilated for ≥ 4 days; plus 2) sustained (≥2 days) FiO₂ increase ≥15 points or PEEP increase ≥2 cmH₂O; plus 3) ≥1 inflammatory sign (leukopenia/leukocytosis, fever/hypothermia, or quantitatively-defined purulent sputum). Medical record review was performed retrospectively for a sample of mechanically-ventilated patients in hospitals reporting NHSN VAP in 2008-2009. To enrich the sample for NHSN VAP cases, patients were randomly selected from 2 strata: those reported to have NHSN VAP, and those who were not reported to have NHSN VAP.  Two experienced infection preventionists (IPs) reviewed each record, one using s-VAP and one using NHSN VAP, collecting data on demographics, definition criteria, hospitalization dates, vital status, and CDTs.  Descriptive statistics were calculated. The kappa (κ) statistic was used to assess agreement. Proportion of patients who died and hospital length of stay were compared using odds ratios and Wilcoxon rank-sum tests (SAS version 9.2).

Results:   Records of 260 patients from 6 hospitals were reviewed. Agreement was low, with 36% of patients meeting one definition but not the other (κ 0.13). CDTs were available for 141 patients with matching NHSN VAP and s-VAP determinations; the total time to make s-VAP determinations was 17.2 hours, compared to 21.6 hours for NHSN VAP.  Among patients mechanically ventilated for >4 days, s-VAP case status and NHSN VAP case status were significantly associated with worse outcomes (Table).

 

Outcome	s-VAP					NHSN VAP
	Case (n=50)	Non-case (n=161)	RR (95% CI)	p-value		Case (n=85)	Non-case (n=126)	RR (95% CI)	p-value
Death	18 (36%)	27 (17%)	1.34 (1.05-1.73)	--		22 (26%)	23 (18%)	1.21 (0.89-1.65)	--
Hospital length of stay (median days)	36	20	--	<0.05		36	18	--	<0.05

Conclusions:   In this preliminary assessment, there was poor agreement between s-VAP and NHSN VAP determinations. Surveillance using s-VAP took less time and appears of equal clinical relevance, compared to NHSN VAP.  S-VAP may provide a suitable framework for developing a clinically relevant surveillance definition that is less subjective and time consuming than current NHSN VAP definitions.