Metrifonate treatment of the cognitive deficits of Alzheimer's disease

Alzheimer's disease (AD) is a progressive neurodegenerative disease that becomes increasingly frequent with advancing age until it affects as many as 45% of individuals 85 years of age and older.¹ AD imposes a dramatic emotional and financial burden on patients, caregivers, and society and is currently estimated to cost more than $100 billion annually in the United States in direct and indirect costs.²

Recently, there has been considerable progress in deciphering the pathophysiology of AD. Several genetic mutations and risk factors have been identified that share the property of inducing a cascade of events leading to the production of intracellular neurofibrillary tangles and extracellular amyloid-containing neuritic plaques and ultimately to changes in synaptic density and cell death.³ Among the cell populations affected are neurons of the neocortex and groups of cells that are involved in the production of critically important neurotransmitters. The nucleus basalis of Meynert, a source nucleus for choline acetyltransferase used in the synthesis of acetylcholine, is affected early in the course of AD, and its atrophy is regarded as the cause of the well-documented cholinergic deficiency.^4,5

That the cholinergic deficit contributes to the cognitive abnormalities of AD is supported by correlations between the severity of the chemical deficiency and the degree of cognitive disturbance by the similarities of the intellectual abnormalities of AD and those produced by administration of anticholinergic agents and by the unusual sensitivity that AD patients exhibit for anticholinergic agent-induced cognitive impairment.^6-8 The hypothesized contribution of the cholinergic abnormalities to the dementia of AD led to studies of the role of cholinergic compounds in alleviating the cognitive abnormalities of the disease. Acetylcholine precursors, cholinesterase inhibitors, acetylcholine release facilitators, and cholinergic receptor agonists have been administered to AD patients in attempts to improve cognition,⁹ with variable results. Two cholinesterase inhibitors available in the United States for the treatment of AD are tacrine and donepezil.^10-12 Other compounds are currently under investigation in the search for agents with improved efficacy and safety profiles.

Metrifonate is a long-acting acetylcholinesterase inhibitor currently under investigation for the treatment of the cognitive, behavioral, and functional deficits of AD. Unlike tacrine and donepezil, which are metabolized by the hepatic cytochrome P450 enzyme system, metrifonate is hydrolyzed nonenzymatically to 2,2-dimethyl dichlorovinyl phosphate (DDVP), a molecule that provides a sustained inhibition of acetylcholinesterase by binding stably to the catalytic site of the enzyme.¹³ Preclinical studies have shown that metrifonate raises brain acetylcholine levels and has beneficial effects on learning and memory in rodents.^14,15 It also has been found in preliminary clinical investigations to benefit cognition in patients with AD.¹⁶ We hypothesized that metrifonate would improve cognitive function in patients with AD. We report the results of a double-blind, parallel group, placebo-controlled, dose-finding study of the efficacy and safety of metrifonate in AD patients.

Methods. Study design. This multicenter, randomized, double-blind, placebo-controlled, dose-finding study compared the efficacy and safety of three doses of metrifonate with those of placebo. The study included a 2-week screening period, a 12-week double-blind treatment phase(2-week loading phase followed by a 10-week maintenance phase), and follow-up visits at 8 and 16 weeks after completion of the double-blind portion of the study. Patients completing double-blind treatment had the option of entering a 40-week open-label extension study and were not required to complete the 8- or 16-week follow-up visits.

AD patients. Patients were evaluated by clinical interview, mental status assessment, neurologic examination, neuroimaging, and laboratory studies. All 480 eligible participants met the criteria of the NINCDS-ADRDA¹⁷ for probable AD. Patients were required to have Mini-Mental State Examination (MMSE)¹⁸ scores between 10 and 26, modified Ischemia Scale¹⁹ scores of less than 4, and Global Deterioration Scale²⁰ stages of 3 to 6. Patients had to weigh between 98 and 207 pounds (45 to 94 kg). All participating patients had caregivers with whom they were in daily contact.

Exclusionary criteria were designed to ensure that study participants had AD as the cause of their dementia. Patients meeting the criteria of the DSM-III-R²¹ for major depressive disorder, bipolar disorder, schizophrenia, substance use disorder, or mental retardation were excluded. CT or MRI within the past 12 months was required to exclude patients with vascular dementia, hydrocephalus, and intracranial mass lesions. Thyroid hormone (thyroid-stimulating hormone, tri-iodothyronine resin uptake test (T3RU), thyroxine) levels, serum vitamin B12 levels, and syphilis serology were evaluated to exclude patients with non-Alzheimer dementias; patients with a history of traumatic brain injury or another neurologic disease (e.g., Parkinson's disease, Huntington's disease, seizure disorder) were also excluded. Patients with significant medical problems(poorly controlled diabetes; systolic blood pressure >180 mm Hg or<100 mm Hg; cancer within the past 5 years; asthma or chronic obstructive pulmonary disease; or clinically significant hepatic, renal, cardiac or pulmonary insufficiency) were not allowed to participate in the study. Also excluded were patients with cardiac conduction defects, significant bradycardia (<50 bpm), recent myocardial infarction (within the previous 4 months), or clinically significant arrhythmias. Patients taking psychotropic agents, antidementia drugs (including tacrine, donepezil, and hydergine), cholinergic or anticholinergic agents, centrally active antihypertensives, anticonvulsants, antacids, or cimetidine were not allowed into the study. Also excluded were patients who had taken any investigational drug within the previous 30 days. Physicians entered patients into the study; monitors reviewed all patients and excluded those not meeting all inclusion and exclusion criteria.

The study was fully explained to all patients and their caregivers and informed consent for participation was given by the patient and his or her legal representative or family caregiver.

Metrifonate dosing. The 480 enrolled patients were assigned to receive placebo (n = 120) or a relatively low (n = 121), mid (n = 120), or high dose (n = 119) of metrifonate according to a computer-generated randomization code. The investigators were blinded to random code assignment.

The double-blinded study medication was blister packaged. Identical three-part labels were affixed to each package. Two parts of the label reflected pertinent study information (study, center, patient, and visit numbers). The third part of the label concealed the identity of the treatment.

All metrifonate-treated patients received a loading dose for 2 weeks followed by a maintenance dose for 10 weeks. A loading dose was used because in its absence, steady-state RBC acetylcholinesterase inhibition is not reached for 6 to 8 weeks.²² Those patients in the low-dose group received a loading dose of 0.5 mg/kg (25 to 45 mg) and a maintenance dose of 0.2 mg/kg (10 to 20 mg). Patients in the mid-dose group received a loading dose of 0.9 mg/kg (45 to 80 mg) and a maintenance dose of 0.3 mg/kg (15 to 25 mg). Patients in the high-dose group received a loading dose of 2.0 mg/kg (100 to 180 mg) and a maintenance dose of 0.65 mg/kg (30 to 60 mg). The loading and maintenance doses were selected for this study to achieve steady-state RBC acetylcholinesterase inhibition levels of 30%, 50%, and 70%, respectively. Metrifonate was administered orally once daily before breakfast.

Outcome measures. Primary efficacy measures were the Alzheimer's Disease Assessment Scale-Cognitive subscale(ADAS-Cog)²³ and the Clinician's Interview-Based Impression of Change with Caregiver Input (CIBIC-Plus).²⁴ Secondary outcome measures included the MMSE,¹⁸ the Clinician's Interview-Based Impression of Severity with Caregiver Input(CIBIS-Plus), the Geriatric Evaluation by Relative's Rating Instrument(GERRI),²⁵ the Instrumental Activities of Daily Living(IADL), and the Physical Self Maintenance Scale (PSMS).²⁶ All investigators were trained in the administration of the study instruments, and monitors visited all sites regularly to ensure that data were collected correctly.

Serum levels of metrifonate and DDVP and RBC acetylcholinesterase activity at weeks 2, 8, and 12 were analyzed according to standard methods.

The incidence of premature termination, treatment-emergent events, mortality, and laboratory abnormalities and changes in vital signs, ECGs, and neurologic examinations were evaluated to determine the safety of metrifonate. Adverse events were rated as mild, moderate, or severe based on the degree of discomfort and disability induced by the event. Selected adverse events-those whose incidence occurred 5% more often in at least one metrifonate treatment group compared with the placebo group-were reported.

Statistical analyses. The target sample size specified in the protocol was 116 patients randomized per treatment group. The sample size was determined as that required to provide 80% dual outcome power, where dual outcome power refers to the probability of observing a significant metrifonate versus placebo comparison with respect to both primary efficacy variables. Computer simulations of the ANOVA methodology for the efficacy analyses were used to estimate the required sample size; these were based on clinically meaningful treatment effects of 3.00 points on the ADAS-Cog scale and 0.40 points on the CIBIC-Plus scale and assumed SDs of 5.72 for the ADAS-Cog and 0.74 for the CIBIC-Plus changes. The SDs were based on estimates from Bayer Corporation-sponsored studies of nimodipine efficacy in the treatment of dementia.

Patients were included in the efficacy and safety analyses according to criteria specified in the study protocol. Patients were included in the safety analysis if they took at least one dose of study medication and had any postbaseline safety assessments. All randomized patients were valid for the safety analysis. Patients were valid for the intent-to-treat (ITT) analysis of efficacy if in addition to being valid for the safety analysis, they also had baseline and any postbaseline efficacy data collected. Only two randomized patients were not valid for the ITT analysis. Patients were valid for the efficacy analysis if they were valid for the ITT analysis, continued to meet inclusion and exclusion criteria, and received double-blind treatment for at least 14 days. Of the 480 randomized patients, 463 were valid for the efficacy analysis.

The types of protocol deviations that were relevant to the interpretation of the efficacy and safety assessments included a failure to satisfy entry criteria (13/480) and the administration of an incorrect treatment or dose(5/480). The methods of managing assessments missing because of premature study discontinuation included observed cases, the last observation carried forward for analyses based on the ITT patient population, and the last valid observation carried forward for the analyses of the valid-for-efficacy patient population.

All efficacy variables were analyzed as a change from baseline with the exception of the CIBIC-Plus, for which the value itself was analyzed. These variables were analyzed using two-tailed statistical tests conducted at the 5% significance level. An ANOVA model with terms for study center, treatment, and treatment-by-center interaction was used; if the treatment-by-center interaction was not significant, it was removed from the model. The efficacy analysis consisted of three pairwise comparisons of the metrifonate groups to placebo; a pairwise comparison was considered to be statistically significant provided that the overall F-test of treatment effects from the ANOVA model was also significant.

With respect to the variables analyzed using the ANOVA model, treatment groups were summarized using least squares means and SEs obtained from the model. For the other variables, treatment groups were compared using descriptive statistics (n, mean, SD, minimum and maximum for continuous variables, cell counts, and percents for categoric variables).

Safety was evaluated by comparing treatment groups with respect to the incidence rates of adverse events, laboratory abnormalities, ECG findings, neurologic examination findings, concomitant medication use, and premature discontinuation from the study. Changes from baseline in vital signs (blood pressure and pulse rate), weight, ECG cardiac cycle measurements, and ECG heart rate were also assessed.

Blood concentrations of metrifonate and DDVP in samples collected 1 hour after dosing were summarized by visit and treatment group. The percent of RBC acetylcholinesterase inhibition was compared across treatment groups and also related to the efficacy and safety variables.

Results. The patients enrolled in this study did not differ appreciably among the four treatment groups in age, gender or race distribution, weight, height, and dementia history (table 1). Patient discontinuation from the study was very modest in all four treatment groups, with 89 to 96% of patients completing the study (seetable 3). Most patients who terminated treatment before study completion did so because of adverse events. Missing data resulted in exclusion of few patients from the final analysis; 99 to 100% of patients were available for the ITT analysis and 94 to 99% of patients were valid for the efficacy analysis.

Primary outcome measures. Significant effects of metrifonate were demonstrated in both the ITT analysis and the valid-for-efficacy analysis. Table 2 shows the results of these analyses. The ITT analysis revealed highly statistically significant treatment effects on the ADAS-Cog and CIBIC-Plus in patients receiving high-dose metrifonate. Valid-for-efficacy analyses also demonstrated highly statistically significant metrifonate effects on the ADAS-Cog and CIBIC-Plus for the mid and high doses (see table 2). The effects of metrifonate on the ADAS-Cog and the CIBIC-Plus scores were evident at the end of the loading phase and were maintained throughout the remaining treatment period(figures 1 and 2).

Patients administered high-dose metrifonate demonstrated a 2.38-point improvement on the ADAS-Cog by week 12 of therapy, whereas patients receiving placebo showed a 0.56-point decline (see figures 1 and 2). Thus, after 3 months of therapy, a 2.94-point difference (95% CI, 1.61 to 4.27; p = 0.0001) in the ADAS-Cog scores of the placebo and high-dose metrifonate groups was observed. At this time, the patients receiving high-dose metrifonate also exhibited a 0.35-point improvement (95% CI, 0.15 to 0.54; p = 0.0007) on the CIBIC-Plus relative to the placebo patients. Patients receiving lower drug doses had scores intermediate between the placebo and the high-dose groups on both performance scales.

Secondary outcome measures. The secondary efficacy variables examined in this study included the MMSE, PSMS, CIBIS-Plus, GERRI, and IADL. For the ITT patients, the placebo versus metrifonate differences in the mean change in the MMSE score at week 12 were 1.11 for the low-dose group (95% CI, 0.39 to 1.84; p = 0.0029), 0.63 for the mid-dose group (95% CI, -0.10 to 1.35; p = 0.0905), and 1.37 for the high-dose group (95% CI, 0.64 to 2.10; p = 0.003). The low- and high-dose metrifonate effects were statistically different from those observed with placebo treatment.

Evaluation of the PSMS item scores revealed trends toward an improvement in the ability of patients to perform certain activities of daily living as quantified by this instrument. Housekeeping, laundry, and dressing exhibited trends toward an improvement from baseline or a dose-related improvement in function. However, the differences from the control group did not reach statistical significance. Similarly, with respect to the remaining secondary efficacy variables, the separation between placebo and metrifonate treatment groups tended to be minimal and typically did not favor any metrifonate treatment group consistently.

Follow-up efficacy assessments. Summary statistics were generated for the ADAS-Cog and CIBIC-Plus scores of patients who did not enter the open-label extension study and who returned for either the 8- or 16-week follow-up visit (data not shown). Although generally the metrifonate-treated patients tended to perform comparably with or better than the placebo-treated patients on the ADAS-Cog and the CIBIC-Plus scales, no reliable conclusions could be drawn because of the small number of patients with follow-up visits (n = 26).

Adverse events. Most adverse events were mild in intensity and were transient in nature (table 3). Adverse events in disfavor of metrifonate (i.e., selected adverse events), defined as those for which the rate in any metrifonate group was greater than the placebo rate by more than 5%, included abdominal pain, diarrhea, flatulence, nausea, and leg cramps (see table 3). These events also tended to be mild and transient regardless of the metrifonate dose and were most common early in the study. Severe adverse events were infrequent, ranging from 1 to 3% in the active treatment groups, and were most common early in treatment(see table 3). Adverse events leading to study withdrawal ranged from 4% in the placebo group to 7% in the mid- and high-dose metrifonate groups (see table 3).

A dose-related reduction in heart rate was observed in the metrifonate-treated patients. The maximum mean changes from baseline represented reductions of 0.5 bpm for the placebo group, 1.6 bpm for the low-dose metrifonate group, 4.6 bpm for the mid-dose metrifonate group, and 7.4 bpm for the high-dose metrifonate group. These changes usually occurred at the completion of the loading phase and tended to attenuate with the continuation into the maintenance phase. Three patients discontinued the study because of the occurrence of asymptomatic bradycardia during the loading phase of the study.

There were no significant laboratory changes associated with metrifonate therapy. Liver function tests were carefully monitored; no significant changes in hepatic enzyme levels were noted.

Acetylcholinesterase inhibition. Treatment with metrifonate produced the expected dose-related inhibition of RBC cholinesterase(table 4). No inhibition was observed in placebo-treated patients, whereas the mean inhibition levels achieved in the metrifonate-treated patients at week 12 were 34.5% at the low dose, 52.2% at mid dose, and 72.5% at the high dose.

The degree of improvement in cognitive and global functioning was associated with the level of cholinesterase inhibition. Generally, the greater the degree of acetylcholinesterase inhibition, the greater the improvement of the ADAS-Cog and CIBIC-Plus scores. In this regard, ITT patients with 15% or less acetylcholinesterase inhibition at week 12 of therapy demonstrated a mean 0.13-point worsening in the ADAS-Cog score, whereas those patients with an enzyme inhibition level of at least 80% showed a mean 1.51-point improvement in this score. Similarly, ITT patients with 15% or less acetylcholinesterase inhibition exhibited a mean CIBIC-Plus score of 4.13, whereas those with 80% or more acetylcholinesterase inhibition scored a mean of 3.71 (i.e., a superior score) on this scale (figure 3).

Discussion. Metrifonate is recommended by the World Health Organization for the treatment of schistosomiasis and has been used in the tropics as an antihelminthic since 1962.²⁷ The identification of metrifonate as a cholinesterase inhibitor, together with the recognition of the cholinergic deficit in AD, led to the exploration of metrifonate as a candidate for AD therapy.¹⁶

In this dose-finding study, metrifonate was efficacious in the treatment of the cognitive deficits of AD when administered in once-daily doses of 0.3 to 0.65 mg/kg (15 to 60 mg). For the ITT patient population, 0.65 mg/kg (30 to 60 mg) metrifonate improved the performance of AD patients on two primary outcome measures, the ADAS-Cog and the CIBIC-Plus. For the valid-for-efficacy population, 0.3 to 0.65 mg/kg (15 to 60 mg) of metrifonate produced a statistically significant treatment difference on these outcome measures. Metrifonate improved the cognitive and global functioning at 2 weeks of therapy after the completion of the loading phase of the study. These effects of metrifonate were sustained during the maintenance phase, such that at 12 weeks of treatment, significant differences in the mean ADAS-Cog and CIBIC-Plus scores between metrifonate- and placebo-treated patients were observed. Similarly, 0.65 mg/kg (30 to 60 mg) metrifonate significantly improved the performance of patients on a secondary efficacy variable, the MMSE, at 12 weeks of treatment.

At doses of 30-60 mg/d, metrifonate produced mild to moderate improvement in cognition with an efficacy comparable with that of tacrine, donepezil, and ENA-713.^10-12,28 The mean treatment difference of 2.94 points on the ADAS-Cog in favor of metrifonate translates into an approximate 5-month delay in the progression of the cognitive decline for the group receiving metrifonate. This average response reflects a broad range of responses for individual patients. No changes in activities of daily living or personal self-maintenance corresponding to these cognition changes were identified with the scales used in this study.

Abdominal pain, diarrhea, flatulence, nausea, and leg cramps were the most commonly reported side effects of metrifonate. Abdominal pain, diarrhea, flatulence, and nausea may reflect cholinergic overactivation in the gastrointestinal system. Leg cramps most likely reflect the overstimulation of nicotinic receptors localized at the neuromuscular junction. Bradycardia, presumably related to the vagotonic effect of acetylcholinesterase inhibitors, emerged as a dose-related ECG finding; three patients discontinued the study because of this asymptomatic bradycardia. At the time of discontinuation, these three patients were in the loading phase of the study receiving the highest dose of metrifonate. This finding, together with the observation that selected adverse events were most common early in the study, suggest that not using a loading dose may reduce such events.

Metrifonate was found to produce a dose-related inhibition of RBC cholinesterase. Brain and erythrocyte acetylcholinesterase inhibition are highly correlated²⁹; consequently, erythrocyte acetylcholinesterase inhibition can be used as a marker of the metrifonate effects in the brain. In the current study, the level of RBC cholinesterase inhibition was directly related to the degree of improvement in cognitive and global functioning in the AD patients.

Metrifonate possesses pharmacologic properties distinct from those of other cholinesterase inhibitors. Metrifonate is rapidly and almost completely absorbed and undergoes little protein binding (<15%). It has a serum half life of approximately 2 hours and is metabolized almost exclusively through biotransformation (only 1 to 3% is excreted unchanged in the liver).^30-32 The biotransformation of metrifonate occurs independently of the hepatic cytochrome P450 enzyme system. Metrifonate has no intrinsic anticholinesterase activity. It is slowly and nonenzymatically transformed to DDVP, which competitively inhibits acetylcholinesterase, leading to an increase in brain acetylcholine within 1 hour of oral metrifonate administration.^32-34 The competitive mechanism underlying DDVP action suggests that the compound may preferentially target those brain areas deficient in acetylcholine. DDVP is also capable of inhibiting butyrylcholinesterase, a cholinesterase found within the senile plaques characteristic of AD and that, like acetylcholinesterase, hydrolyzes acetylcholine to its inactive metabolites.^35,36 Therefore, the inhibition of butyrylcholinesterase by DDVP may contribute to the observed increases in acetylcholine levels.

The metrifonate-mediated cholinesterase inhibition is long lasting, recovering in concert with new enzyme synthesis.³² Enzyme reactivation is possible, however, either spontaneously over time or, if required, by oxime therapy.^37,38 The long duration of enzyme inhibition mediated by metrifonate makes it suitable for once per day administration, and advantage for the treatment of AD patients who must be reminded to take medications and whose behavioral symptoms (i.e., agitation, aggression, depression) may make multiple daily dosing regimen difficult or impossible.³⁹

In this study, metrifonate was found to be safe and efficacious for the treatment of the cognitive deficits of AD. However, because of the characteristics of the AD patients included in this and all other late-phase-development clinical trials, such studies have certain limitations. Generally, late-phase-development studies do not mirror clinical practice; patients typically are screened carefully and those individuals with important medical and neurologic comorbidity, neuropsychiatric symptoms, or more advanced AD are excluded from the study. This study design is reflected in the performance measures of placebo-treated AD patients, who often show less deterioration during the course of the clinical trial than would be expected from epidemiologic data.

Advances in understanding the pathophysiology of AD support the feasibility of cholinomimetic agents in the treatment of AD. Cholinomimetic agents, such as cholinesterase inhibitors, improve symptoms of the disease. Although such agents enhance cognitive and global functioning, they have little impact on the course of the underlying disease. New therapies, including antiamyloid agents, antioxidants, calcium channel blocking agents, anti-inflammatory drugs, and estrogens, may slow the progression of AD.³ However, although these agents may retard the rate of decline, they will not improve existing symptoms. Thus, combinations of cholinergic therapy with drugs that slow the progression of the illness will likely comprise future therapeutic regimens for AD, and a continued role for cholinergic enhancement therapy in AD can be anticipated. Metrifonate is a promising agent for the cholinergic treatment of AD. Metrifonate exerted beneficial effects in patients with AD when administered in the relatively low doses used in this study. This prodrug has promise in the treatment of AD, and the potential increased utility of higher doses will be explored.

Acknowledgment

We acknowledge the contribution of Bianca B. Ruzicka, PhD, in the preparation of this manuscript.

Appendix