L model, only APOEe4 status remained a significant predictor of the volumes of WMH (Table 3). The model performed well (Omnibus test of model coefficients p,0.05), and the model fit was good (Generalised linear models, Pearson Chi Square p = 0.179). Only age remained a significant predictor of DWMH 1326631 scores (Table 4). The model performed well (Omnibus test of model coefficients p = 0.010), and the model fit was good (Hosmer and Lemeshow test p = 0.492). We also performed multiple logistic regression analyses (stepwise and forced entry) controlling for scanning site and including variables known from previous studies to be associated with WMH (age, hypertension, diabetes mellitus), in addition to OH or systolic or Chebulagic acid web diastolic BP drops. In these analyses, both with respect to the volumetry group and the semi-quantitative group, only age remained a significant predictor of WMH load (data not shown). However, in some of the models the predictor “MRI centre” achieved borderline significance (p = 0.048?.050). When analysing the patients with DLB/PDD separately, we found no significant correlations between Scheltens DWMH scores and systolic or diastolic BP drops. Similarly, there were no significant differences between those in the highest and lowest Scheltens DWMH score quartiles with respect to the other variables in Table 2 (data not shown). In bivariate logistic regression analyses, diastolic BP drop, age and APOEe4 status achieved the lowest p-values (0.124, 0.117 and 0.094, respectively). Due to the rather small subsample, in combination with missing values for the relevant variables, it was not statistically feasible to perform multiple logistic regression analyses using these variables.Statistical AnalysesA total of 82 patients had MRI scans that could be analysed volumetrically (volumetry group), and 139 had scans that could be rated semi-quantitatively (the semi-quantitative group) according to the Scheltens scale. The scans of 61 patients were analysed with both methods, yielding a correlation coefficient (Spearman’s rho) of 0.791 (p,0.001) between the scores of the two methods. MannWhitney U-test, Chi-square, Spearman rank order or Fisher’s exact test were used as appropriate. None of the continuous variables had a normal distribution, according to the KolmogorovSmirnov test. Potential predictor variables having p-values ,0.25 in bivariate logistic regression analyses were included in stepwise multiple logistic regression analyses, with the response variable defined as being in the highest quartile of total WMH volume ratios or Scheltens deep WMH (DWMH) scores vs. the lowest quartile, respectively. P-values ,0.05 (get PS 1145 two-tailed) were considered statistically significant. All statistical tests were performed using PASW Statistics 18, release 18.0.1.ResultsWhen comparing the baseline characteristics of patients undergoing WMH volume analysis with those who were not included in the study, the only significant difference was a higher proportion with Alzheimer’s disease among the participants (volumetry group: Pearson Chi square 14.558, df 1, p,0.001, semi-quantitative group: Pearson Chi square 8.162, df 1, p = 0.006 (Table 1)). In the volumetry group, the only significant difference with respect to relevant clinical characteristics between patients in the highest and lowest WMH quartiles was a lower proportion in the former group with at least one APOEe4 allele (Table 2). In the semi-quantitative group, patients in the highest DWMH score qu.L model, only APOEe4 status remained a significant predictor of the volumes of WMH (Table 3). The model performed well (Omnibus test of model coefficients p,0.05), and the model fit was good (Generalised linear models, Pearson Chi Square p = 0.179). Only age remained a significant predictor of DWMH 1326631 scores (Table 4). The model performed well (Omnibus test of model coefficients p = 0.010), and the model fit was good (Hosmer and Lemeshow test p = 0.492). We also performed multiple logistic regression analyses (stepwise and forced entry) controlling for scanning site and including variables known from previous studies to be associated with WMH (age, hypertension, diabetes mellitus), in addition to OH or systolic or diastolic BP drops. In these analyses, both with respect to the volumetry group and the semi-quantitative group, only age remained a significant predictor of WMH load (data not shown). However, in some of the models the predictor “MRI centre” achieved borderline significance (p = 0.048?.050). When analysing the patients with DLB/PDD separately, we found no significant correlations between Scheltens DWMH scores and systolic or diastolic BP drops. Similarly, there were no significant differences between those in the highest and lowest Scheltens DWMH score quartiles with respect to the other variables in Table 2 (data not shown). In bivariate logistic regression analyses, diastolic BP drop, age and APOEe4 status achieved the lowest p-values (0.124, 0.117 and 0.094, respectively). Due to the rather small subsample, in combination with missing values for the relevant variables, it was not statistically feasible to perform multiple logistic regression analyses using these variables.Statistical AnalysesA total of 82 patients had MRI scans that could be analysed volumetrically (volumetry group), and 139 had scans that could be rated semi-quantitatively (the semi-quantitative group) according to the Scheltens scale. The scans of 61 patients were analysed with both methods, yielding a correlation coefficient (Spearman’s rho) of 0.791 (p,0.001) between the scores of the two methods. MannWhitney U-test, Chi-square, Spearman rank order or Fisher’s exact test were used as appropriate. None of the continuous variables had a normal distribution, according to the KolmogorovSmirnov test. Potential predictor variables having p-values ,0.25 in bivariate logistic regression analyses were included in stepwise multiple logistic regression analyses, with the response variable defined as being in the highest quartile of total WMH volume ratios or Scheltens deep WMH (DWMH) scores vs. the lowest quartile, respectively. P-values ,0.05 (two-tailed) were considered statistically significant. All statistical tests were performed using PASW Statistics 18, release 18.0.1.ResultsWhen comparing the baseline characteristics of patients undergoing WMH volume analysis with those who were not included in the study, the only significant difference was a higher proportion with Alzheimer’s disease among the participants (volumetry group: Pearson Chi square 14.558, df 1, p,0.001, semi-quantitative group: Pearson Chi square 8.162, df 1, p = 0.006 (Table 1)). In the volumetry group, the only significant difference with respect to relevant clinical characteristics between patients in the highest and lowest WMH quartiles was a lower proportion in the former group with at least one APOEe4 allele (Table 2). In the semi-quantitative group, patients in the highest DWMH score qu.