Fig 2 - uploaded by Santiago Perez-Lloret
Content may be subject to copyright.
Contexts in source publication
Context 1
... were prospective cohorts, and the rest used a cross-sectional design. Twenty-two studies used the Consensus to diagnose OH; 6 and 3 studies used the MDS criteria to define PDD and PD-MCI, respectively. A random-effect analysis found a significant association between OH and PDD or PD-MCI (OR, 95% CI: 3.31, 2.16-5.08; k = 18, n = 2251; p < 0.01; Fig. 2). Heterogeneity was high (I 2 = 66%). The association between OH and PDD (4.64, 2.68-8.02; k = 13, n = 1194; p < 0.01) was stronger than with PD-MCI (1.82, 0.92-3.58; k = 5, n = 1056; p = NS), as revealed by a test for subgroup differences (chisq = 4.43, p = 0.04; Fig. 2). A Harbord test did not suggest any significant publication bias ...
Context 2
... and PDD or PD-MCI (OR, 95% CI: 3.31, 2.16-5.08; k = 18, n = 2251; p < 0.01; Fig. 2). Heterogeneity was high (I 2 = 66%). The association between OH and PDD (4.64, 2.68-8.02; k = 13, n = 1194; p < 0.01) was stronger than with PD-MCI (1.82, 0.92-3.58; k = 5, n = 1056; p = NS), as revealed by a test for subgroup differences (chisq = 4.43, p = 0.04; Fig. 2). A Harbord test did not suggest any significant publication bias (t = 1.45, df = 16, p = 0.16). The funnel plot is available in the Online Supplemental Material (Figure ...
