SESSION TITLE: Biomarkers in Severe Sepsis and Septic Shock
SESSION TYPE: Original Investigation Slide
PRESENTED ON: Wednesday, October 28, 2015 at 02:45 PM - 04:15 PM
PURPOSE: Novel biomarkers of acute kidney injury (AKI) often perform poorly in patients with sepsis, possibly due to important differences in pathogenesis. In this study, we examined the accuracy of the urinary cell cycle arrest biomarkers, tissue inhibitor of metalloproteinase-2 (TIMP-2) and IGF-binding protein 7 (IGFBP7), in risk stratifying AKI in sepsis patients. We also examined whether non-renal organ failures degraded the performance of the biomarker test.
METHODS: We performed a secondary analysis of data collected in two international prospective observational cohorts (SAPPHIRE and TOPAZ) of 1,164 critically ill patients enrolled within 24 hours of ICU admission. Patients with sepsis at admission were analyzed. AKI was adjudicated using KDIGO criteria (stage 2-3) within 12 hours. Two [TIMP-2]·[IGFBP7] cutoffs (0.3 for high sensitivity and 2.0 for high specificity) were evaluated. We used receiver operating characteristic curve (ROC) analyses and multivariate logistic regression models to determine if the biomarker result improved AKI classification, also taking into consideration non-renal organ dysfunction.
RESULTS: Two-hundred thirty-two sepsis patients were identified with mean (SD) age of 62 (17) years, of which 17.2% patients met the AKI endpoint. Median (IQR) [TIMP-2]·[IGFBP7] was 2.1 (1.1-4.3) vs. 0.4 (0.2-1.0) (ng/mL)2/1000 in AKI vs. non-AKI, p<0.001. Relative risk (95% CI) for AKI by [TIMP-2]·[IGFBP7] >0.3 and >2.0 (ng/mL)2/1000 relative to ≤0.3 (ng/mL)2/1000 were 8.9 (3.1-20.9) and 19.7 (7.2-48.9). The area under the ROC curves (AUC) for predicting AKI were 0.84 (0.77-0.90) for [TIMP-2]·[IGFBP7] and 0.74 (0.65-0.83) for creatinine. A multivariate clinical model that included statistically significant variables from univariate analysis for AKI (non-renal APACHE III, non-renal SOFA, body mass index, chronic liver disease, congestive heart disease, and enrollment creatinine) showed AUC of 0.83 (0.75-0.91). The addition of [TIMP-2]·[IGFBP7] to the model increased AUC to 0.93 (0.88-0.97), p<0.01. Log10([TIMP-2]·[IGFBP7]) had an adjusted odds ratio (OR) of 38.2 (9.5-153.2; p<0.001) for AKI; whereas, non-renal APACHE III and non-renal SOFA had adjusted OR of 1.03 (1.01-1.05; p=0.01) and 1.02 (0.83-1.25; p=0.89), respectively.
CONCLUSIONS: Urinary [TIMP-2]·[IGFBP7] performs well in predicting AKI in sepsis patients, even in the presence of non-renal organ dysfunctions.
CLINICAL IMPLICATIONS: [TIMP-2]·[IGFBP7] can be used for early identification of septic patients at high risk of AKI.
DISCLOSURE: Bryant Nguyen: Grant monies (from industry related sources): Astute Medical Michelle Gong: Grant monies (from industry related sources): Astute Medical Kianoush Banaei Kashani: Grant monies (from industry related sources): Astute Medical Azra Bihorac: Grant monies (from industry related sources): Astute Medical Lakhmir Chawla: Grant monies (from industry related sources): Astute Medical Jing Shi: Grant monies (from industry related sources): Astute Medical John Kellum: Grant monies (from industry related sources): Astute Medical
No Product/Research Disclosure Information