Table 5. Overall association between various characteristics of interest and mean BOT-2 score, with and without adjustment for age and sex (N = 37) and concentrations of each of the biochemical predictors (N = 32).

		Model 1	Model 2	Model 3
Variable	Crude estimates (95% CI)	Adj.* estimates (95% CI)	Adj.* estimates (95% CI)	Adj.* estimates (95% CI)
Age, Years, per 1 year increase	−0.44 (−1.42, 0.53)	0.43 (−0.56, 1.42)	0.52 (−0.42, 1.45)	0.42 (−0.47, 1.32)
Gender
Male	Reference	Reference	Reference	Reference
Female	−1.65 (−7.56, 4.26)	1.59 (−3.83, 7.02)	1.78 (−3.34, 6.91)	1.17 (−3.77, 6.11)
KABC-II score (#), per 1 unit increase	0.65 (0.37, 0.93)	0.61 (0.14, 1.07)	0.70 (0.24, 1.15)	0.82 (0.36, 1.28)
Albumin (µmol/L), per 10 µmol/l increase	0.37 (0.14, 0.61)	0.18 (−0.08, 0.44)	0.16 (−0.08, 0.41)	0.28 (0.02, 0.53)
Triglycerides (mmol/l), per 0.2 unit increase	−0.64 (−1.27, −0.01)	−0.19 (−0.78, 0.39)	−0.51 (−1.15, 0.12)	−0.54 (−1.14, 0.06)
Isoprostane III (ng/ml)	−3.16 (−25.64, 19.30)		20.77 (0.09, 41.44)
Isoprostane VI (ng/ml)	−6.52 (−19.60, 6.54)
8,12-iso-iPF2α-VI isoprostane (ng/ml)	−10.88 (−24.95, 3.20)			20.61 (4.03, 37.19)
R-square (%)		47.08%	54.81%	58.07%
BIC (lower is better)		227 (6 df)	225 (7 df)	223 (7 df)

All multivariable models include age and gender as control variables. (*)  =  Adjusted for individual predictors, (#)  =  higher score indicates greater cognition performance. Model 1  =  baseline multivariable model. Model 2  =  Model 1 supplemented with isoprostane-III. Model 3  =  Model 1 supplemented with 8,12-iso-iPF2α-VI isoprostane as predictor. The addition of isoprostane VI did not improve the baseline model.

After adjustment for each of the predictors including cognitive performance, only concentrations of albumin and 8,12-iso-iPF2α-VI isoprostane, among the biochemical measures, were significant predictors in the model explaining the largest variation (Model 3) in the BOT-2 scores; models built for all subjects irrespective of the konzo status.