• Open Access

Efficacy and safety of Ginkgo biloba extract EGb 761® in mild cognitive impairment with neuropsychiatric symptoms: a randomized, placebo-controlled, double-blind, multi-center trial


  • The GIMCIPlus Study Group: Gavrilova SI, Mental Health Research Center of the Russian Academy of Medical Sciences, Moscow; Vilyanov VB, City Hospital No. 2 " V. I. Razumovskiy" , Saratov; Sholomov II, City Clinical Hospital No. 9, Saratov; Voskresenskaya ON, Saratov State Medical University of Roszdrav, Saratov; Shiryaev OY, Voronezh Regional Clinical Psychoneurological Policlinic, Voronezh; Chefranova ZY, Belgorod Regional Clinical Hospital, Belgorod; and Gubarev YD, Belgorod State University Chair of Nervous Diseases, Belgorod.



The study was conducted to explore the effects of EGb 761® (Dr. Willmar Schwabe GmbH & Co. KG, Karlsruhe, Germany) on neuropsychiatric symptoms (NPS) and cognition in patients with mild cognitive impairment (MCI).


One hundred and sixty patients with MCI who scored at least 6 on the 12-item Neuropsychiatric Inventory (NPI) were enrolled in this double-blind, multi-center trial and randomized to receive 240 mg EGb 761 daily or placebo for a period of 24 weeks. Effects on NPS were assessed using the NPI, the state sub-score of the State-Trait Anxiety Inventory and the Geriatric Depression Scale. Further outcome measures were the Trail-Making Test (A/B) for cognition and global ratings of change. Statistical analyses followed the intention-to-treat principle.


The NPI composite score decreased by 7.0 ± 4.5 (mean, standard deviation) points in the EGb 761-treated group and by 5.5 ± 5.2 in the placebo group (p = 0.001). Improvement by at least 4 points was found in 78.8% of patients treated with EGb 761 and in 55.7% of those receiving placebo (p = 0.002). Superiority of EGb 761 over placebo (p < 0.05) was also found for the State-Trait Anxiety Inventory score, the informants' global impression of change, and both Trail-Making Test scores. There were statistical trends favoring EGb 761 in the Geriatric Depression Scale and the patients' global impression of change. Adverse events (all non-serious) were reported by 37 patients taking EGb 761 and 36 patients receiving placebo.


EGb 761 improved NPS and cognitive performance in patients with MCI. The drug was safe and well tolerated. Copyright © 2014 John Wiley & Sons, Ltd.


Mild cognitive impairment (MCI) defines a condition characterized by subjective complaints of deficiencies in cognitive performance and objective evidence of cognitive impairment while diagnostic criteria for dementia are not met (Petersen et al., 1999; Zaudig, 2002). Longitudinal studies have reported that a number of patients diagnosed as having MCI will become demented, some will remain at the MCI stage for years, and others will return to normal (Ebly et al., 1995; Wolf et al., 1998; Morris et al., 2001; Larrieu et al., 2002). Hence, MCI does not necessarily represent a transitional stage between normal aging and dementia, but a condition associated with an elevated risk of developing dementia (Smith, 2002).

Memory impairment, in particular when evidenced by a delayed recall task, is generally considered to be the first sign of incipient dementia (Elias et al., 2000; Collie et al., 2001). It may precede the stage of overt dementia by 15–20 years (Elias et al., 2000; Kawas et al., 2003).

Neuropsychiatric symptoms (NPS) are frequently observed in MCI subjects. Lyketsos and co-workers (2002) reported total prevalences of NPS in the past month of 43% for subjects with MCI, 16% for the general population, and 75% for dementia patients. Depression (20%), apathy (15%), irritability (15%), and sleep disturbances (14%) were the symptoms found most frequently. Similar rates were reported from another population-based study (Chan et al., 2003) with prevalences of 47% for any mental psychiatric or behavioral symptom. Forsell and co-workers (2003) reported a significantly elevated frequency in anxiety syndromes among subjects with MCI (10%).

For patients who suffer from MCI-related symptoms and who exhibit signs or symptoms that indicate an increased risk for progression to dementia, close attention is advised, and treatment should be considered, as outlined recently (Förstl et al., 2008). Those with a high risk for progression to dementia are subjects with “amnestic MCI” (aMCI), where impairment of the short-term memory is the most salient symptom (e.g., Ahmed et al., 2008). In addition, the presence of NPS is both a burden to the patient and an indicator of increased risk for dementia (Robert et al., 2006; Palmer et al., 2007; Teng et al., 2007).

In pre-clinical studies, Ginkgo biloba extract EGb 761® (Dr. Willmar Schwabe GmbH & Co. KG, Karlsruhe, Germany) has been found to improve mitochondrial function and neuronal energy metabolism (Abdel-Kader et al., 2007), to promote neuroplasticity (Müller et al., 2012) and hippocampal neurogenesis (Tchantchou et al., 2007), and to enhance microperfusion by decreasing blood viscosity (Erdinçler et al., 1996). In clinical trials, EGb 761 improved cognition, NPS, activities of daily living, and overall clinical impression in patients with dementia (Weinmann et al., 2011; Bachinskaya et al., 2011; Herrschaft et al., 2012). Improvement of learning, memory, and executive functions have also been reported from studies in healthy volunteers and in patients with mild aging-associated impairment of cognitive function (Kaschel et al., 2009; Grass-Kapanke et al., 2011; Kaschel, 2011)

The objective of the present study was to explore whether G. biloba extract EGb 761 can alleviate NPS associated with aMCI and whether the profile of effects is similar to that found in patients with dementia. A second goal was to demonstrate that EGb 761 improves cognitive performance in patients with MCI as defined by international consensus criteria.

Patients and methods

The trial was conducted in accordance with the ethical principles laid down in the Declaration of Helsinki (as revised in 1983), the ICH Harmonized Tripartite Guideline for Good Clinical Practice (ICH, 1996), and local laws. Written informed consent was obtained from all patients enrolled after complete and extensive description of the study and before any trial-related procedures were conducted. The trial was approved by the ethics committee under the Federal Service on Surveillance in Healthcare and Social Development of the Russian Federation, Moscow.

Patient selection

Patients were enrolled by the psychiatric or neurologic outpatient departments of seven academic hospitals in Russia. Male and female outpatients aged 55 years or older were eligible if they were diagnosed with aMCI in accordance with international consensus criteria (Winblad et al., 2004, Figs. 2 and 3). Specifically, cognitive complaints expressed by the patient were corroborated by an informant (partner, close relative, or friend); the total score on the revised version of the cognitive part of the Cambridge Mental Disorders of the Elderly Examination (CAMCOG) (Huppert et al., 1995) was below the 10th percentile for age, sex, and educational level; the CAMCOG combined memory score was 20 or less for patients younger than 80 years of age and 17 or less for patients aged 80 or older; a decline from a former level of functioning was demonstrated either by informant report or by cognitive testing; basic activities of daily living were preserved with a mean score lower than 4 on the short form of the Informant Questionnaire on Cognitive Decline in the Elderly (Jorm, 1994); overall cognition was not severely impaired as evidenced by a score of at least 24 on the mini mental state examination (Folstein et al., 1975); DSM-IV criteria for dementia were not fulfilled; and the cognitive complaints had been present for at least 6 months. A total score of 6 or higher on the 12-item Neuropsychiatric Inventory (NPI) (Cummings, 1997) was required, with at least one of the item scores for “depression/dysphoria,” “anxiety,” “apathy/indifference,” “agitation,” or “irritability/lability” being at least 4. A magnetic resonance imaging scan that was taken no more than 1 year prior to screening had to be consistent with the diagnosis of MCI (i.e., mild to moderate cortical and overall atrophy, white matter lesions and lacunas were permitted, but signs of normal pressure hydrocephalus, intracranial hemorrhage/hematoma, brain tumor, or other structural brain disease were exclusionary), major depression had to be ruled out (i.e., in cases of conspicuous depression scores, the DSM-IV diagnostic criteria for major depression had to be checked), and adequate Russian language skills were required to understand and respond to all interview questions and undergo neuropsychological testing without the need of an interpreter. Moreover, an informant who was frequently in contact with the patient was required to provide information about the patient's history of cognitive problems, functional abilities, and behavioral symptoms.

Patients with any type of neurological disorder, namely stroke with sequelae within the last 3 months or hemorrhagic stroke within the last 12 months before enrolment or dementia; current or recurrent psychiatric disorder; severe or insufficiently controlled cardiovascular disorder or insulin-dependent diabetes mellitus; severe hepatic or renal dysfunction, clinically significant anemia, thyroid dysfunction, or vitamin deficiency; gastrointestinal disorders with uncertain absorption; alcohol or substance abuse; active malignant disease; or severe and insufficiently corrected impairment in vision or hearing were excluded from the study. Participation in another clinical drug trial within 4 weeks before enrolment, previous participation in any trial of a G. biloba product, female patients of child-bearing age and no contraception, and intake of any medication that could interfere with the efficacy assessment were also grounds for exclusion.

Randomization and treatment

Randomization in blocks, with equal probabilities for active drug and placebo and stratified by centers, was performed by the sponsor's biometrics group using a validated computer program. Drug and placebo tablets were indistinguishable in appearance, packaging, and labeling. Each box carried a unique drug number that was generated and assigned to active drug or placebo by the randomization program. The randomization code was stored in a locked compartment at the sponsor's offices inaccessible to all persons involved in the trial; the block length was not disclosed to investigators or clinical staff. This procedure ensured both double blinding and concealment of allocation sequence.

After the screening visit, a run-in period of up to 4 weeks was permitted to complete electrocardiogram and magnetic resonance imaging (MRI) examinations and obtain all information required. When eligibility was confirmed, patients underwent the baseline assessment; follow-up assessments were performed after 12 and 24 weeks of treatment. During the treatment period, patients were instructed to take one tablet of the study preparation, active drug or placebo, every morning. Active treatment consisted of tablets containing 240 mg of EGb 761. EGb 761 is a dry extract from G. biloba leaves (35–67:1); extraction solvent: acetone 60% (w/w). The extract is adjusted to 22.0–27.0% ginkgo flavonoids calculated as ginkgo flavone glycosides and 5.0–7.0% terpene lactones consisting of 2.8–3.4% ginkgolides A, B, and C and 2.6–3.2% bilobalide and contains less than 5 ppm ginkgolic acids. Treatment adherence was checked for all patients by inquiry and pill count.

Outcome measures

Because of the exploratory nature of the trial, no primary outcome measure was selected. The 12-item NPI (Cummings, 1997) was used to assess the severity of NPS overall. Briefly, frequency and severity of each symptom were rated in an interview with an informant who had close contact with the patient. The composite score was calculated by multiplying frequency and severity ratings item by item and summing up all 12 item scores. The caregiver distress rating was omitted, because MCI patients usually do not need care. A decrease by at least four points is considered as a clinically relevant improvement (Mega et al., 1999). The Geriatric Depression Scale (GDS) (Sheikh and Yesavage, 1986) and the state subscale of the State-Trait Anxiety Inventory (STAI-X1) (Spielberger et al., 1970) were employed to assess depression and anxiety symptoms more comprehensively. Both self-rating scales are valid for use in healthy persons or persons with mild sub-syndromal problems as well as in patients with full-blown depressive or anxiety disorder. The Clinical Global Impression (CGI) (NIMH, 1976) scale was used for global assessment of changes. Taking into account that the majority of patients with MCI usually show rather mild symptoms, no dramatic changes were expected. Hence, the CGI ratings of change were performed by the informants and by the patients themselves, because both were expected to be more sensitive in detecting subtle changes than the physicians who only saw the patients on three visits during the course of the study. Only ratings of “much improved” and “very much improved” were counted as improvement. The cognitive domain was covered by the Trail-Making Test (TMT) (Reitan, 1958), forms A and B, to measure attention, psychomotor speed, and executive functioning.

During the investigators' meeting, investigators and investigational staff who administered cognitive tests and rating scales were trained by a geriatrician (N. B.) and a psychologist (R. K.), both experienced in the field of clinical and neuropsychological assessment. Training involved comprehensive instructions on all procedures as well as demonstrations (NPI and GDS) and supervised test administration (CAMCOG).

Drug safety was assessed by measuring vital signs and inquiry about adverse events (AEs) at all visits as well as by repeating physical examination, lab tests, and electrocardiogram at the final visit (week 24 or early termination).

Sample size estimation and statistical analysis

Because the data were to be analyzed in an exploratory manner, no formal sample size calculation was conducted. With a sample size of 2 × 80 patients, the power is at least 80% at type I error rate of α = 0.05 to detect a statistically significant difference between placebo and active treatment groups, with a two-sided t-test if the treatment effect is about Δ/σ = 1/2.

For each of the efficacy variables, the EGb 761 group was compared with the placebo group with methods of descriptive data analysis. Additionally, descriptive p-values were calculated using t-test and chi-square tests for metrical and categorical data, respectively. Efficacy analysis was primarily based on the full analysis data set including all patients who received randomized study treatment at least once and having at least one measurement of any efficacy parameter during the randomized treatment period. Missing values were replaced by the last observation carried forward method on an item-for-item basis. Safety variables were assessed for the safety population, which included all patients randomized to study treatment and who took study medication at least once. AEs were summarized by means of appropriate frequency tables on the basis of coded items and taking severity and relationship to study drug into account.


Patient disposition and baseline data

Of 194 patients screened for the study, 160 (82%) were eligible and randomized. Patient characteristics at baseline are shown in Table 1. Five patients terminated the study prematurely, three of whom had no post-baseline data for efficacy measures. One of these was excluded from the efficacy analysis, and two were considered “informative dropouts” (reasons were lack of efficacy and possibly drug-related AE, respectively) and were included in both the full analysis set (FAS) and the per-protocol set (PPS). The PPS consisted of 146 patients (EGb 761, 72 patients; placebo, 74 patients) for whom no relevant protocol violations were documented. Patient disposition is depicted in Figure 1. All patients were free of psychotropic medication at baseline. Three patients in the active treatment group and two patients in the placebo group had taken psychotropic drugs earlier but had discontinued more than 9 weeks before baseline.

Table 1. Baseline characteristics of enrolled patients, absolute numbers, and percent (gender) or mean ± standard deviation (other); two-sided p-value of t-test for continuous variables and chi-square test for categorical variables
 EGb 761®Placebop-value
N = 80N = 79
  1. CAMCOG, Cambridge Mental Disorders of the Elderly Examination; MMSE, mini mental state examination; NPI, Neuropsychiatric Inventory; STAI-X1, State-Trait Anxiety Inventory; TMT, Trail-Making Test.

Female [n (%)]58 (73)66 (84)0.0929
Age (years)65 ± 763 ± 70.1180
Weight (kg)73 ± 973 ± 110.5853
Height (cm)168 ± 8166 ± 90.2033
School (years)9.7 ± 0.99.7 ± 0.81.0000
Academic (years)4.9 ± 1.55.0 ± 1.30.6541
Duration of cognitive complaints (months)42 ± 3136 ± 260.1883
CAMCOG total score76.3 ± 4.777.0 ± 4.50.3281
CAMCOG combined memory score18.5 ± 1.318.3 ± 1.50.3701
MMSE25.6 ± 1.325.7 ± 1.50.4408
NPI composite score11.5 ± 3.511.6 ± 3.70.7798
STAI-X1 (state) score46.1 ± 7.145.0 ± 6.50.2932
GDS total score6.1 ± 2.75.8 ± 2.70.5766
TMT-A [s]77.4 ± 28.676.6 ± 27.30.8518
TMT-B [s]165.6 ± 46.6161.8 ± 44.70.6029
Figure 1.

Patient disposition.


On average, patients in both treatment groups improved in all behavioral and cognitive outcome measures, with consistently larger improvements found in the EGb 761 group (Table 2). Superiority of EGb 761 over placebo was statistically significant for the NPI composite score and response rates (Figures 2 and 3), the STAI-X1 score (Figure 4), TMT A and B (Figure 5), and the informant's global assessment. Results were nearly identical for FAS and PPS; therefore, only the results for the FAS are presented.

Table 2. Changes from baseline to week 24 in efficacy measures, full analysis set; mean ± standard deviation (changes in scores) or percent (response/improvement rates); two-sided p-value of analysis of covariance with covariate “baseline score” and factors “center” and “treatment group” (changes in scores) or chi-square test (response/improvement rates)
 EGb 761®Placebop-value
N = 80N = 79
  1. NPI, Neuropsychiatric Inventory; STAI-X1, State-Trait Anxiety Inventory; GDS, Geriatric Depression Scale; TMT, Trail-Making Test; CGIC, Clinical Global Impression of Change.

NPI composite score−7.0 ± 4.5−5.5 ± 5.20.0010
NPI improvement ≥ 4 points78.8%55.7%0.0019
STAI-X1 (state) score−6.6 ± 7.9−3.8 ± 7.30.0271
GDS total score−3.0 ± 2.6−2.3 ± 3.20.0658
TMT-A [s]−22.4 ± 23.0−18.1 ± 25.20.0447
TMT-B [s]−44.7 ± 36.8−32.2 ± 39.00.0112
CGIC improvement—informant rating61.3%41.8%0.0140
CGIC improvement—patient rating60.1%46.8%0.0961
Figure 2.

Change in Neuropsychiatric Inventory composite score from baseline to week 24, means, and 95% confidence intervals, two-sided p-value of analysis of covariance.

Figure 3.

Response rates (%) in Neuropsychiatric Inventory composite score at week 24 (response defined as improvement by ≥4 points), two-sided p-value of chi-square test.

Figure 4.

Change in State-Trait Anxiety Inventory X1 score from baseline to week 24, means, and 95% confidence intervals, two-sided p-value of analysis of covariance.

Figure 5.

Change in Trail-Making Test B score from baseline to week 24, means, and 95% confidence intervals, two-sided p-value of analysis of covariance.


During the treatment period, 51 AEs were observed in 37 patients treated with EGb 761 and 46 AEs in 36 patients taking placebo. AEs reported for at least 5% of patients in either group were headache (EGb 761, six patients; placebo, nine patients), increased blood pressure (EGb 761, six patients; placebo, seven patients), respiratory tract infection (EGb 761, seven patients; placebo, three patients), and dyspepsia/epigastric discomfort (EGb 761, four patients; placebo, one patient). For 18 AEs observed under EGb 761 treatment and 16 AEs reported under placebo, a causal relationship with treatment could not be ruled out under double-blind conditions. There were no serious AEs in either treatment group.


Patients with aMCI and clinically significant NPS were enrolled in this randomized, placebo-controlled, exploratory trial. During the 24-week course of treatment, improvements in NPS and cognitive abilities that were significantly and consistently more pronounced in patients taking 240 mg per day of the G. biloba extract EGb 761 than in patients taking placebo were found. Improvements were recognized more easily by informants who had close contact with the patients than by the patients themselves.

Previous studies reported efficacy of EGb 761 at a daily dose of 240 mg in the symptomatic treatment of dementia (Alzheimer's disease (AD), vascular dementia, or mixed type) associated with NPS (Napryeyenko et al., 2007; Ihl et al., 2011; Herrschaft et al., 2012). Specific effects on NPS could be demonstrated in these studies (Scripnikov et al., 2007; Bachinskaya et al., 2011; Nacu et al., 2011). In recent studies, EGb 761 improved several cognitive domains including sustained attention, visual and verbal memory in subjects with very mild cognitive impairment (Grass-Kapanke et al., 2011), and prospective memory in healthy volunteers (Kaschel, 2011). So far, our findings are in line with those from previous trials. It has been suggested that aMCI may be an early stage of AD (Gainotti et al., 2008; Jessen et al., 2010).

The current study is the first trial of EGb 761 in patients with aMCI diagnosed according to consensus criteria (Winblad et al., 2004). However, our findings may not apply to some patients with aMCI diagnosed by other criteria owing to the lack of full congruence between the different diagnostic MCI criteria.

Anxiety was one of the symptoms that consistently and most markedly improved in dementia patients treated with EGb 761 (Scripnikov et al., 2007; Bachinskaya et al., 2011; Nacu et al., 2011). A previous study in patients with generalized anxiety disorder or adjustment disorder with anxious mood demonstrated significant dose-dependent effects of EGb 761 versus placebo in both physician- and patient-rated psychiatric assessments (Woelk et al., 2007). The clinical relevance of anxiety symptoms is reflected by the prevalence rates reported in MCI, which range from 10% in a population-based study (Lyketsos et al., 2002) to 52% in patients referred to a memory clinic (Gallagher et al., 2011). The presence of anxiety in MCI was found to be associated with a higher risk of dementia (Palmer et al., 2007) and earlier conversion to dementia (Gallagher et al., 2011). In a recent study, MCI patients with anxiety were more likely to have abnormal CSF levels of Aβ1–42 and total tau, suggesting that anxiety may indicate underlying AD pathology (Ramakers et al., 2013).

The symptoms of depression were also consistently and markedly alleviated in dementia patients treated with EGb 761 in previous studies. In the present study, a slight numerical superiority over placebo was found, but this did not reach statistical significance. However, considering that the mean GDS scores of our patients were only slightly higher than the mean score of healthy subjects, which was 5.75 ± 4.34 in the validation study by Yesavage et al. (1983), little improvement would be expected. Considering the prevalence rates reported for depression in MCI, which were as high as 20% (Lyketsos et al., 2002) and 27% (Geda et al., 2008) in population-based studies, and co-morbid depression disorders in 29.3% of outpatients visiting a memory clinic (Di Iulio et al., 2010), the low mean level of depression found in our study may be uncommon. However, patients screened for the current study were excluded if they had any co-morbid DSM-IV mental disorder or concurrent antidepressant drug treatment. This might represent a selection bias against patients with marked depression, but low depression scores in incident AD were also reported from an unselected population. In a cohort study of a geographically defined population, median scores of 1 on the Center for Epidemiologic Studies Depression Scale and 2 on the Hamilton Depression Rating Scale were found in subjects with incident AD (Wilson et al., 2010).

As expected from a cognitively activating drug, EGb 761 treatment improved performance in cognitive tests, tapping attention, working memory, visuospatial orientation, and executive function. Because patients with MCI, by definition, do not have significant impairment at work or in social life, it is difficult to gauge the immediate relevance of the observed improvements. Using the informant-rated Functional Assessment Questionnaire, the Alzheimer's Disease Neuroimaging Initiative found significant associations between the TMT-A and the number and severity of functional deficits in patients with aMCI (Brown et al., 2011). Taking into account that many patients with MCI are still employed, the observed improvement in TMT performance may have an influence on their ability to cope with demands of working life. Moreover, low scores in executive function were reported to be associated with shorter time to progression to AD (Blacker et al., 2007), and accelerated decline in executive function was observed within 2–3 years before the diagnosis of dementia (Grober et al., 2008).

Whereas both patients and informants were equally likely to rate global change as favorable in the EGb 761-treated group, more patients than caregivers perceived changes under placebo intake as favorable. Looking at single rating scales, we observed improvements in the placebo group in both informant- (NPI) and patient-rated scales (STAI-X1 and GDS). Intensified medical care in the context of study participation together with hope derived from study medication may have improved the patients' mood irrespective of the treatment received. Yet, as depression scores were low at baseline, it is unlikely that the changes in mood had a sufficient influence on the patients' overall perception of change to account for the higher rates of improvement under placebo. Anxiety levels were outside the normal range, but placebo effects on anxiety were relatively small and therefore unlikely to explain the favorable CGI ratings. It may, however, be possible that in patients affected by various NPS, the global rating might have been dominated by subjectively perceived overall improvements in such symptoms, whereas the informants, not personally affected by the disorder, might have paid more attention to the patients' minor cognitive failures.


With NPS and executive function, EGb 761 treatment significantly improved symptoms of MCI, which are known to be associated with faster cognitive decline and a higher risk of developing AD. As EGb 761 is safe and well tolerated, it represents a promising treatment option in aMCI as defined by international consensus criteria.

Conflict of interest

S. I. G. was the principal investigator of the study and received investigator fees from the sponsor; U. W. P., R. K., and N. B. received honoraria as advisors to the sponsor; R. H. is an employee of the sponsor receiving a fixed salary; J. W. M. W. has no conflict of interest. The study was designed and the data were collected and analyzed under the sponsor's responsibility as defined by Good Clinical Practice guidelines; the data were interpreted, the manuscript was written, and the decision to submit it for publication was taken jointly and unanimously by the authors.

Key points

  • Twenty-four week treatment with EGb 761 at a daily dose of 240 mg alleviated NPS in patients with MCI.
  • EGb 761 improved cognitive performance, including executive functioning.
  • The drug was safe and well tolerated.


This work was sponsored by Dr. Willmar Schwabe, GmbH & Co. KG, Karlsruhe, Germany.