O biloba improves cognitive functioning in patients with Alzheimer’s disease, with effect sizes similar to that obtained with other anti-dementia drugs such as acetylcholinesterase inhibitors [27]. An alternative explanation for the present findings showing lesser long-term cognitive decline in subjects reporting using EGb761H than in those reporting use of piracetam or inhibitor Neither drug could be related to differences in psychotropic use observed between the study groups. Indeed, our results showed that use of EGb761H was associated with significantly lower consumption of psychotropic drugs including antidepressants, benzodiazepines and antipsychotics. Given the well-characterised adverse effects of chronic psychotropic drug use on cognitive function [54?5], it was possible that the beneficial effect observed of EGb761H on cognitive decline was indirectly due to less psychotropic drugs consumption. However, when psychotropic drug use was taken into account as a possible confounding factor in the statistical model, the differences in the rate of cognitive decline betweenTable 3. Comparison of change in cognitive outcomes over twenty years in the PAQUID cohort in subjects receiving EGb761H (n = 589) or piracetam (n = 149) compared to the `neither treatment’ group (n = 2874) (mixed linear effects model).Unadjusted for psychotropic drug use Cognitive score Mini Mental State Evaluation Variables Time Piracetam EGb761H Isaacs Sets Test (30 sec) Time Piracetam EGb761H Benton Visual Retention Test Time Piracetam EGb761H bAdjusted for psychotropic drug use b1 20.302 20.592 0.461 20.258 21.468 0.271 20.078 20.470 20.014 SE 0.013 0.202 0.085 0.019 0.516 0.227 0.004 0.184 0.SE 0.013 0.211 0.089 0.020 0.523 0.231 0.005 0.194 0.p,.0001 0.0057 ,.0001 ,.0001 0.0077 0.3561 ,.0001 0.0242 0.p,.0001 0.0034 ,.0001 ,.0001 0.0045 0.2328 ,.0001 0.0106 0.20.315 20.584 0.482 20.290 21.395 0.213 20.081 20.438 20.1 Covariates: age, gender, educational level, MMSE score at inclusion, depressive symptomatology and memory complaints. doi:10.1371/journal.pone.0052755.tGinkgo Biloba and Long-Term Cognitive DeclineFigure 2. Estimated change in MMSE score over the Epigenetic Reader Domain twenty-year follow-up period in the three treatment groups. Legend: —- Neither treatment (n = 2874 at inclusion). 
????EGb761H (n = 589 at inclusion). ???Piracetam (n = 149 at inclusion). doi:10.1371/journal.pone.0052755.ggroups persisted, suggesting that the slower decline of MMSE scores in the EGb761H group cannot be explained 15755315 by differences in psychotropic drug consumption. Regarding, the finding of stronger decline in the group using piracetam, due 1326631 to the small number of participants in this group, it is difficult to draw conclusions. The study of the relationship between cognitive decline and piracetam consumption was not the principal objective of our study, and was included in this study to have a point of comparison with a group of participants using a nootropic drug prescribed for the same condition as EGb761H. However, this result is somewhat intriguing and it would be important to replicate this result in another prospective study to draw reliable conclusions.In conclusion, even though some points remain unclear, in particular the reason for the stronger decline observed in the piracetam group, or the question of a possible dose-effect of EGb761H that could not be presently tested, this study reports a beneficial effect of EGb761H on long-term cognitive decline as assessed by the MMSE in non-demented el.O biloba improves cognitive functioning in patients with Alzheimer’s disease, with effect sizes similar to that obtained with other anti-dementia drugs such as acetylcholinesterase inhibitors [27]. An alternative explanation for the present findings showing lesser long-term cognitive decline in subjects reporting using EGb761H than in those reporting use of piracetam or neither drug could be related to differences in psychotropic use observed between the study groups. Indeed, our results showed that use of EGb761H was associated with significantly lower consumption of psychotropic drugs including antidepressants, benzodiazepines and antipsychotics. Given the well-characterised adverse effects of chronic psychotropic drug use on cognitive function [54?5], it was possible that the beneficial effect observed of EGb761H on cognitive decline was indirectly due to less psychotropic drugs consumption. However, when psychotropic drug use was taken into account as a possible confounding factor in the statistical model, the differences in the rate of cognitive decline betweenTable 3. Comparison of change in cognitive outcomes over twenty years in the PAQUID cohort in subjects receiving EGb761H (n = 589) or piracetam (n = 149) compared to the `neither treatment’ group (n = 2874) (mixed linear effects model).Unadjusted for psychotropic drug use Cognitive score Mini Mental State Evaluation Variables Time Piracetam EGb761H Isaacs Sets Test (30 sec) Time Piracetam EGb761H Benton Visual Retention Test Time Piracetam EGb761H bAdjusted for psychotropic drug use b1 20.302 20.592 0.461 20.258 21.468 0.271 20.078 20.470 20.014 SE 0.013 0.202 0.085 0.019 0.516 0.227 0.004 0.184 0.SE 0.013 0.211 0.089 0.020 0.523 0.231 0.005 0.194 0.p,.0001 0.0057 ,.0001 ,.0001 0.0077 0.3561 ,.0001 0.0242 0.p,.0001 0.0034 ,.0001 ,.0001 0.0045 0.2328 ,.0001 0.0106 0.20.315 20.584 0.482 20.290 21.395 0.213 20.081 20.438 20.1 Covariates: age, gender, educational level, MMSE score at inclusion, depressive symptomatology and memory complaints. doi:10.1371/journal.pone.0052755.tGinkgo Biloba and Long-Term Cognitive DeclineFigure 2. Estimated change in MMSE score over the twenty-year follow-up period 
in the three treatment groups. Legend: —- Neither treatment (n = 2874 at inclusion). ????EGb761H (n = 589 at inclusion). ???Piracetam (n = 149 at inclusion). doi:10.1371/journal.pone.0052755.ggroups persisted, suggesting that the slower decline of MMSE scores in the EGb761H group cannot be explained 15755315 by differences in psychotropic drug consumption. Regarding, the finding of stronger decline in the group using piracetam, due 1326631 to the small number of participants in this group, it is difficult to draw conclusions. The study of the relationship between cognitive decline and piracetam consumption was not the principal objective of our study, and was included in this study to have a point of comparison with a group of participants using a nootropic drug prescribed for the same condition as EGb761H. However, this result is somewhat intriguing and it would be important to replicate this result in another prospective study to draw reliable conclusions.In conclusion, even though some points remain unclear, in particular the reason for the stronger decline observed in the piracetam group, or the question of a possible dose-effect of EGb761H that could not be presently tested, this study reports a beneficial effect of EGb761H on long-term cognitive decline as assessed by the MMSE in non-demented el.