The role of intraoperative parameters on predicting laparoscopic abdominal surgery associated acute kidney injury

BMC Nephrology

Table 5 Risk reclassification using intraoperative parameters and clinical predictors compared with clinical predictors alone

Model with clinical predictors alone^a	Model with clinical predictors + intraoperative parameters^b			Direction of reclassification
Model with clinical predictors alone^a				Increase risk	Decrease risk
Non-AKI (N = 41)	< 30% risk	30–60% risk	> 60% risk
< 30% risk	12 (75)	4 (25)	0
30–60% risk	9 (37.5)	14 (58.3)	1 (4.2)	5 (12.2)	9 (22.0)
> 60% risk	0	0	1
AKI (N = 23)
< 30% risk	1	3	0
30–60% risk	5 (27.8)	10 (55.6)	3 (16.7)	6 (26.1)	5 (22.0)
> 60% risk	0	0	1

^aClinical predictors alone including age, diabetic status, eGFR and urological surgery status
^bIntraoperative parameters including operation time, inflation time and exposure index
Data in shaded boxes (diagonals) represent similar risk classification between clinical model and clinical plus intraoperative parameters
Numbers to the right of diagonals represent the patients who were reclassified as increased risk by adding intraoperative parameters to the clinical model. Numbers to the left of the diagonals represent patients who were reclassified as lower risk when intraoperative parameters are added to the clinical model
Parentheses represent the percentage of reclassification in each risk category. The net reclassification improvement (NRI) for Non-AKI and AKI was calculated from the difference in proportions moving up and down among AKI and Non-AKI (NRI = 14.1% (95% confidence interval 0.0 to 48.6%, P = 0.409)). The relative integrated discrimination improvement (IDI) was measured for the increment in the predicted probabilities for the subset experiencing Non-AKI and the decrement for the subset experiencing AKI (relative IDI = 1.8% (95% CI 1.1 to 16.3%, P = 0.025))

ISSN: 1471-2369