|Home | About | Journals | Submit | Contact Us | Français|
Some breast cancer survivors report cognitive difficulties greater than 1 year after chemotherapy. Acetylcholinesterase inhibitors (AChEI) may improve cognitive impairment. We conducted a randomized, placebo-controlled, pilot study to assess the feasibility of using the AChEI, donepezil, to improve subjective and objective measures of cognitive function in breast cancer survivors.
Women who received adjuvant chemotherapy 1–5 years prior with current cognitive dysfunction symptoms were randomized to 5 mg of donepezil/day vs placebo for 6 weeks and if tolerated 10 mg/day for 18 weeks for a total of 24 weeks. A battery of validated measures of attention, memory, language, visuomotor skills, processing speed, executive function, and motor dexterity and speed was administered at baseline and at 24 and 36 weeks. Subjective cognitive function, fatigue, sleep, mood, and health-related quality of life were evaluated at baseline and at 12, 24, and 36 weeks.
Sixty-two patients were enrolled, 76 % completed the study, self-reported compliance was 98 %, and toxicities were minimal. At the end of treatment, the donepezil group performed significantly better than the control group on two parameters of memory—the Hopkins Verbal Learning Test -Revised (HVLT-R) Total Recall (p=0.033) and HVLT-R Discrimination (p=0.036). There were no significant differences on other cognitive variables or in subjective cognitive function or quality of life.
Accrual to this feasibility trial was robust, retention was good, compliance was excellent, and toxicities were minimal.
Randomized clinical trials in breast cancer survivors to improve cognitive dysfunction are feasible. A phase III trial testing the efficacy of donepezil is warranted given these pilot results.
There are 13.7 million cancer survivors in the USA, and those affected by breast cancer make up 22 % of that group . Women who receive adjuvant chemotherapy for breast cancer sometimes report a long-term consequence of cognitive dysfunction . The reports of cognitive impairment are highest during chemotherapy and decline as time post-chemotherapy increases. However, for some, cognitive effects remain. Breast cancer survivors report long-term (>1 year post-chemotherapy) cognitive difficulties that began during chemotherapy [2–7]. Furthermore, there may be some patients who have no evidence of acute cognitive dysfunctions during chemotherapy but have delayed decline 1 year subsequent to chemotherapy . The cognitive dysfunction is apparent across key domains, including concentration, memory, processing speed, and executive functions [9–11]. Accordingly, cognitive impairment associated with cancer and chemotherapy can adversely affect occupational performance, interpersonal relationships, and quality of life.
The specific mechanisms of chemotherapy or cancer-induced injury to the central nervous system remain unclear. Potential pathways of injury are direct (neurotoxicity), including impaired neurogenesis [10, 11]. Cognitive impairment could also be through indirect mechanisms such as treatment-induced metabolic and hormonal abnormalities, inflammatory cytokine activation, medical co-morbidities, fatigue, injury to other body organs, or micro-vessel disease [12, 13]. In brain imaging studies, changes in cerebral metabolism and blood flow have been noted  along with cerebral atrophy . Furthermore, cognition may be affected secondarily through factors such as fatigue, sleep disturbance, anemia, and mood [16–19].
Neurotransmitter modulators have improved cognitive function in other cancer populations. Donepezil, a reversible acetylcholinesterase inhibitor used widely to treat symptoms associated with Alzheimer’s disease and vascular dementia [20–23], directly affects neuronal function by increasing the bioavailability of acetylcholinesterase and increasing cerebral perfusion . Shaw et al. performed a phase II open-label study and reported that irradiated brain tumor survivors who had completed a course of≥ 30 Gy of brain irradiation ≥6 months prior to enrollment and were treated with 5 mg/day of donepezil for 6 weeks followed by 10 mg/day for 18 weeks showed improvement in cognitive symptoms, cognitive functioning (attention/concentration, verbal and figural memory, and verbal fluency), mood, fatigue, and quality of life . A recent phase III randomized, placebo-controlled trial of donepezil (5 mg/day×6 weeks followed by 10 mg/day× 18 weeks) in the same population of brain tumor late survivors revealed a modest benefit for memory and motor dexterity and speed. Among patients with the poorest pre-treatment cognitive performance, the effect was stronger . Given the encouraging results with brain tumor patients and the significant burden of cognitive dysfunction in breast cancer survivors, we conducted a small randomized pilot study to evaluate the feasibility of using donepezil as a treatment for cancer-associated cognitive dysfunction in breast cancer survivors. To allow time for the recovery of cognitive function after receiving chemotherapy, we chose to study women 1–5 years post-chemotherapy who reported having significant cognitive symptoms.
Women with a history of invasive breast cancer who completed adjuvant chemotherapy 1 to 5 years previously and who reported significant cognitive symptoms were eligible to enroll. Documentation of prior chemotherapy (>4 cycles of cytotoxic chemotherapy) was required. Women on hormonal therapy must have been on the same agent for at least 3 months with plans to continue it for the duration of the study. Prospective participants were pre-screened for self-reported cognitive problems with the FACT-Cognition (Version 3) . Women were required to have self-reported cognitive problems. Women who scored 63 or higher on the FACT-Cog Perceived Cognitive Impairment sub-scale (indicating little dysfunction) were ineligible. Patients whose performance status was>60 (Karnofsky Scale) or 0–2 (ECOG score) were eligible. Women were excluded if they had ever been diagnosed with metastatic disease, dementia, or a cerebrovascular accident. The use of any non-study cognition-enhancing drugs was not allowed during study or up to 4 weeks prior to enrollment. Additional exclusion criteria included (1) ongoing ketoconazole or quinidine, (2) hypersensitivity to donepezil, (3) use of investigational medications within the prior 30 days, (4) traumatic brain injury, multiple sclerosis, or recent myocardial infarction, (5) history of schizophrenia, psychosis or substance abuse, (6) untreated current severe depression (treated depression was permitted if treatment was stable), (7) acute severe fatigue, chronic fatigue syndrome, or fibromyalgia, and (8) history of hepatic or renal dysfunction. The study (NCT 01466270) was approved by the Institutional Review Board (IRB) at Wake Forest University School of Medicine (Winston Salem, NC) (WFUSM) and by the IRBs at the participating sites. The study accrued patients at WFUSM and at 15 Community Clinical Oncology Programs (CCOPs) affiliated with the NCI-approved Wake Forest University CCOP Research Base.
This was a randomized, double blind, placebo-controlled clinical trial in which eligible participants were stratified by menopausal status (pre- vs peri-/post-) and time since chemotherapy (12–36 months vs >36 months) and assigned with equal probability (using variably sized permuted block randomization) to receive 5 mg of donepezil/day for 6 weeks which was escalated to 10 mg/day for 18 weeks if well tolerated or matching placebo. Study drug (donepezil, Aricept®) and placebo were distributed to the study sites by Biologics Inc., Raleigh, NC, USA. Treatment was discontinued at 24 weeks. Study measures were collected at baseline and at 12 weeks by mail and telephone (some measures), 24 weeks (end of treatment), and 36 weeks by centrally trained and certified examiners following a standardized testing protocol. Adherence was measured by self-reported pill diaries. Toxicity was assessed by study personnel at clinician encounters and by telephone at 12 weeks. Toxicity was graded by the Common Toxicity Criteria v4.03.
A battery of validated and standardized measures of attention, memory, language, visuomotor skills, processing speed, executive function, and motor dexterity and speed was completed by the participants at baseline and at 24 and 36 weeks. Verbal learning and memory were assessed with the Hopkins Verbal Learning Test-Revised (HVLT-R) . HVLT consists of a 12-item word list which is read to subjects on three successive learning trials. Free recall scores are recorded for each learning trial. After a 20-min interval during which subjects complete other non-interfering tasks and questionnaires, they were asked to recall the target words. Lastly, a yes/no recognition task is then presented in which subjects were asked to identify all target words by responding “yes” and to reject 12 non-target words by responding “no” to minimize confounding by practice effects . Test–retest reliability of the HVLT is quite good (0.74). HVLT-R variables included learning [total recall (TR)=sum of three learning trials; score range 0–36], memory [delayed recall (DR)=trial 4; score range 0–12], recognition [true positives (TP); score range 0–12], [discrimination index (Discrim)=true positives minus false positives; score range −12–12] and [%savings (%S)=[(DR/highest of learning trials 2 and 3) × 100]; score range ≥0]. The modified Rey–Osterreith Complex Figure (mROCF)  assessed visuo-motor skills (mROCF-Copy; score range 0–24), immediate recall (RF-IR; score range 0–24), and delayed recall (RF-DR; score range 0–24). The Trail Making Test—Parts A & B  assessed attention (TMT-A) and executive function set shifting (TMT-B). Verbal fluency was assessed with the Controlled Oral Word Association (COWA) . Concentration and working memory were measured with the Digit Span (DS) test [forward (DSF); score range 0–16 and backward (DSB); score range 0–14], a subtest of the Wechsler Adult Intelligence Scale-III ; a DS-total score was the sum of DSF and DSB. Motor speed and dexterity were measured with the Grooved Pegboard  for the dominant hand and the non-dominant hand. For most tests, a higher score is better with the exception of when the score represents time and a lower score is better (Trail Making Test and Grooved Pegboard).
Subjective cognitive functioning and symptoms were evaluated with the Functional Assessment of Cancer Therapy—Cognition (Version 3) at baseline and at 12 (by mail and telephone), 24, and 36 weeks. The FACT-Cog (Version 3) is a validated self-report measure of perceived cognitive function and impact on quality of life over the prior 7 days [27, 34]. It consists of four subscales—Perceived Cognitive Impairments (PCI; range: 0–80 ), Perceived Cognitive Abilities (PCA; 0 –36), Impact on Quality of Life (IQL; 0–16), and Comments from Others (CFO; 0–16). Higher subscale scores indicate better functioning. We enrolled only women with a FACT-Cog PCI score<63, a score that reflected moderate to severe problems [35, 36]. Fatigue was assessed with the FACIT-Fatigue Subscale and the PROMIS 7-item Fatigue Scale. Sleepiness was measured with the Epworth Sleepiness Scale, and mood was assessed with the Beck Depression Inventory and Beck Anxiety Inventory. Health-related quality of life (HRQL) was measured by the Rand-36 Health Survey [37–41]. All of the written questionnaires were administered at the four time points.
Chi-square, Fisher exact, and Wilcoxon rank-sum tests were used to assess unadjusted group differences in categorical and continuous variables. Mixed effects repeated-measures analysis of covariance (RMANCOVA) was used to assess treatment differences in the cognitive function and HRQL measures and to obtain least squares (LS) estimates of the measures over time. Differences (donepezil minus placebo) in LS means are also provided, so for those outcomes for which higher scores reflect better performance, positive differences indicate a benefit for donepezil. An unstructured covariance matrix was used to model the correlation in outcomes over time. Models were fitted for each outcome which just included treatment group, time, strata, and the baseline level of the outcomes. A sample size of 60 participants was selected to provide reasonable estimates of retention (±16 % with 95 % confidence) and treatment group differences for the continuous outcomes (± 0.6 SDs with 95 % confidence assuming 20 % dropout). This study was not powered to detect treatment differences in any of the outcomes. In this paper, we present the results for the first 24 weeks, which was the period during which the participants received treatment.
Sixty-two patients were enrolled between July 24, 2012 and January 24, 2013. The women enrolled had a median age of 56 years, were primarily white (90 %), were married (71 %), and had high school education or more (98 %) (Table 1). Most of the women were peri- or postmenopausal (95 %) and were on hormonal therapy (68 %), predominantly an aromatase inhibitor. Characteristics were similar for the two groups.
Sixty-two patients were accrued from 16 sites in 6 months for an accrual rate of ten patients per month. Forty-seven women (76 %) completed the 24-week treatment phase, 71 % of the women receiving donepezil and 81 % of the women receiving the control (p=.374). Of the six control patients who dropped out, two were due to sickness or toxicity, one refused, and three were lost to follow-up. Of the nine donepezil patients who dropped out, five were due to sickness or toxicity, three refused, and one was an MD decision. Self-reported adherence to the study drug while on therapy was 98 % for both arms. The common toxicities experienced by patients on both arms were mostly grade 1 or 2 with the exception of grade 3 insomnia reported in two patients in the donepezil arm and one case of grade 3 muscle cramps on the placebo arm.
The measures of objective and subjective cognitive function are summarized in Table 2. On average, our sample showed poorer performance compared to non-cancer comparison groups on all cognitive measures except the mROCF (data not shown). However, the range of scores for each measure was broad and included some scores that were better than the mean for the comparators. There were no significant group differences for any cognitive parameter at baseline. Scores on most cognitive measures improved from baseline to week 24 for both groups perhaps due to test familiarity (practice effects) or statistical regression to the mean. Results of the mixed effects models are shown in Table 3. The donepezil group performed significantly better than the control group on two parameters of memory from the HVLT-R: total recall (p=0.033) and discrimination index (p=0.036). There were no other significant group differences after 24 weeks of treatment with donepezil.
We looked at sub-groups within our study to discern if there were differences by hormonal therapy. The women who were concurrently on hormonal therapy with an aromatase inhibitor performed worse on all the HVLT-R components compared to the larger group (not on hormonal therapy or tamoxifen) though not significantly so. Furthermore, this group declined in their performance on the HVLT-R over time. In contrast, all other groups improved over time, although modestly. Given the heterogeneity of our sample with respect to cognitive functioning at baseline, we performed an exploratory analysis of a treatment by baseline cognitive function interaction for all cognitive measures. A significant interaction effect was found for executive function (TMT B, p= 0.007); those with lower baseline executive function randomized to the donepezil arm showed greater improvement in executive function at 24 weeks than the controls.
Quality of life measures are also summarized in Table 2. The quality of life scores overall were lower than what is found in women without cancer (data not shown). There were few significant differences between the donepezil and the placebo groups in the quality of life measures over the course of the study. There were no significant differences between the donepezil and placebo groups at week 24 for either the FACT-Cog total score or the four sub-scale scores (p>0.05). At 12 weeks, participants in the donepezil arm reported greater anxiety than those in the placebo arm (p=0.049), but there were no significant differences between the arms at 24 weeks (p=0.205). No group differences were observed in fatigue, sleep, or depression between the two groups of participants or on any subscale of the SF-36 at baseline and at 12 or 24 weeks. It is of note that the physical component score (PCS) and the mental component scores (MCS) of the SF-36 were poorer than what has been found in women without a history of cancer . There were also no significant group differences in any of the HRQL measures when the analyses were stratified by time since completion of chemotherapy for breast cancer (<30 months, 30–60 months).
This randomized, placebo-controlled clinical trial assessed the feasibility of using 5–10 mg/day of donepezil for 24 weeks to improve cognitive function in breast cancer survivors who self-reported moderate cognitive dysfunction 1 to 5 years following chemotherapy. Enrollment was rapid, retention was good, and the drug was well tolerated. Estimates of variability and treatment differences for various cognitive function and quality of life outcomes are provided.
Although this study was not powered to detect differences in outcomes, there appears to be an indication that donepezil improved verbal memory in survivors who entered the study with poorer cognitive functioning. At completion of the intervention, there were modest but significant differences between the groups for verbal learning and memory, with better performance in the donepezil group compared to controls. Thus, our small clinical trial demonstrated that this treatment approach was feasible and potentially effective in breast cancer survivors following chemotherapy. This is the first clinical trial of an acetylcholine esterase inhibitor to demonstrate a benefit for objectively measured cognitive functioning [42, 43] in patients with chemotherapy-related cognitive dysfunction. Other agents that have been tested with some success have wake-promoting or stimulant properties .
Tamoxifen has been shown to be protective for cognitive performance tasks that are dependent on cholinergic pathways and hippocampal function . Therefore, women on tamoxifen may be less likely to benefit from an acetyl-cholinesterase inhibitor such as donepezil. Most of the women on our study were on hormonal therapy with an aromatase inhibitor, and therefore we had insufficient number of women to evaluate that difference. When the subset of participants on hormonal therapy were examined separately, the improvement on HVLT-R—Total Recall score with donepezil persisted (p= 0.048).
Although we enrolled only women who reported moderate to severe cognitive problems, their performance on the cognitive measures ranged from quite impaired to normal using non-cancer comparison groups. This incongruence was consistent with prior reports of discordance between self-reported symptoms and objective measures of cognitive dysfunction [9, 45–48]. The heterogeneity with respect to cognitive functioning at baseline may help to explain the overall modest treatment effects we observed with donepezil. When we examined the interactions between treatment and baseline cognitive functioning, a stronger treatment effect was found among women with poorer cognitive functioning at study entry.
Improvement in learning and memory may be critical to improving everyday problem solving for a breast cancer survivor. Memory performance, measured with the HVLT-R, was predictive of independent functioning over 5 years in a geriatric non-cancer cohort . Although significant treatment group differences in memory tasks did occur, we did not find significant differences in HRQL among women on donepezil compared with women receiving placebo. Previous reports have shown greater improvement in HRQL in the first 1–3 years post-chemotherapy than in later years . The mean time since chemotherapy in the present study was 30 months, and most study participants were<3 years post-treatment (65 % in donepezil, 65 % in placebo). The only significant difference in qualitative measures identified was that women taking donepezil reported higher anxiety at 12 weeks than the placebo group, but this effect was not present at 24 weeks. Anxiety is a potential side effect of donepezil, so this could have been related to the study drug. It is unclear why this difference occurred midway through the study treatment.
Limitations of the present study include the small sample size. This was designed to be a feasibility study, and it was not powered to adequately test donepezil’s efficacy. A follow-up study is planned based on these results. The women were enrolled subsequent to completion of chemotherapy 1–5 years previously and therefore have late cognitive dysfunction. We do not know the trajectory of their impairment. It is presumed that among the women who enrolled in this study, there were women who suffered acute cognitive dysfunction and continued to decline in the subsequent years, and there are also women who had late cognitive dysfunction that may lead to chronicity. These two groups may differ in the mechanism of impairment. We assessed adherence using a self-report pill diary; we do not have more objective measures of adherence. We do not have data on the screened populations as we do not require participating sites to track and report screening information as these are busy clinical community sites and our protocols are time and labor intensive. It is a limitation of the study that we cannot characterize better the entire prospective participant population which constrains generalization. Strengths of the study include the use of a well-validated neurocognitive test battery that assessed multiple cognitive domains, inclusion of community-based clinical oncology sites, excellent participant adherence to study requirements, and overall retention.
Complaints of cognitive dysfunction related to cancer and its treatments are common. To date, few effective treatments have been identified. This study suggests that breast cancer survivors experiencing cognitive problems several years following completion of chemotherapy may be helped by treatment with donepezil. Moreover, if treatment is targeted for survivors with greater cognitive deficits, the benefit may be greater. Other treatments, both pharmacological and non-pharmacological, are needed for this large and growing patient population.
The Wake Forest Community Clinical Oncology Program Research Base is funded by NCI/DCP grant 2U10CA081851.
Portions of these results were presented at 50th Annual Meeting of the American Society of Clinical Oncology, Chicago, IL, USA, May 29–June 2, 2014 (9516) and the 2014 International Cognition and Cancer Taskforce, Seattle, WA, USA, February 10–13, 2014 (22).
Conflict of interest There are no conflicts of interest to disclose.