Skip to main content

Table 2 Associations between established AKI risk factor covariates and maximum 48-h serum creatinine change from baseline using a linear model and each subpopulation of a two-component latent variable mixture model

From: Latent variable modeling improves AKI risk factor identification and AKI prediction compared to traditional methods

Risk factor Linear model Latent variable mixture model
Subpopulation 1 Subpopulation 2
Age (per 10 years) 0.037 (−0.044, 0.118) 0.040 (−0.021, 0.101) 0.042 (0.016, 0.068)**
BMI (per 5 kg/m2) 0.040 (0.013, 0.068)** 0.107 (0.081, 0.132)*** 0.012 (0.005, 0.019)***
History of hypertension 0.005 (−0.046, 0.056) 0.080 (0.002, 0.158)* −0.017 (−0.031, −0.003)*
History of diabetes −0.024 (−0.082, 0.035) −0.123 (−0.177, −0.068)*** −0.007 (−0.021, 0.008)
Baseline pulse pressure (per 10 mmHg) 0.003 (−0.009, 0.015) −0.019 (−0.035, −0.003)* 0.004 (7.2e-5, 0.008)*
Baseline SCr (per mg/dL) 0.203 (−0.309, 0.715) 0.054 (−0.217, 0.326) 0.158 (−0.023, 0.339)
Baseline SCr:age interaction −0.001 (−0.008, 0.007) 0.002 (−0.003, 0.007) −0.001 (−0.003, 0.002)
Baseline eGFR (per 30 mL/min/1.73 m2) 0.081 (−0.012, 0.174) 0.045 (−0.066, 0.156) 0.099 (0.051, 0.147)***
Baseline hematocrit (per %) −0.010 (−0.016, −0.005)*** −0.034 (−0.040, −0.028)*** −0.003 (−0.005, −0.001)***
Cardiopulmonary bypass time (per hour) 0.006 (−0.018, 0.030) −0.072 (−0.108, −0.036)*** 0.012 (2.0e-4, 0.024) **
Aortic cross clamp time (per hour) 0.036 (0.001, 0.072)* 0.156 (0.120, 0.192)*** −0.006 (−0.018, 0.006)
Intraoperative hydroxyethyl starch volume (per L) 0.200 (0.000, 0.400) 0.300 (0.100, 0.500)** 0.000 (−0.056, 0.094)
Intraoperative urine output (per L) −0.100 (−0.200, −0.048)** −0.300 (−0.400, −0.200)*** −0.100 (−0.094, −0.016)***
Mean intraoperative MAP adjusted for baseline MAP (per 10 mmHg) 0.023 (0.003, 0.042)* 0.064 (0.042, 0.086)*** 0.005 (0.001, 0.009)*
Maximum intraoperative lactate (per mmol/L) 0.004 (−0.021, 0.028) −0.009 (−0.033, 0.016) 0.013 (0.007, 0.019)***
Length of surgery (per hour) 0.034 (0.009, 0.059)** 0.113 (0.086, 0.140)*** 0.027 (0.021, 0.034)***
  1. *p < 0.05, **p < 0.01, ***p < 0.001; BMI body mass index, eGFR estimated glomerular filtration rate using CKD-Epi formula, SCr serum creatinine concentration, MAP mean arterial blood pressure
  2. For example, an increase of ten years in age is associated with a 0.037 increase in 48-h postoperative change in serum creatinine concentration (ΔSCr) in the linear model, and a past medical history of hypertension was associated with a 0.080 increased in 48-h ΔSCr in the subpopulation 1 component model. Ninety-five percent confidence intervals are listed after each covariate coefficient estimate