If You Take Galantamine, Take It!
Well in Real Life
Anti-Alzheimer’s agent is as effective in a
clinically relevant setting as in formal research studies
By Will Block
hy can’t people be perfect? Instead, we make all kinds of misteaks, such as forgetting our anniversary, cutting corners, losing our cool, believing politicians, driving too fast, not taking our supplements, etc. And what about institutions that don’t work well? Why can’t we have an orderly society where everything runs like a Swiss watch? Apart from Switzerland, that doesn’t happen anywhere except, perhaps, in prison—but not too many people are clamoring to get behind bars.
Life on the outside may be risky, and things often don’t go as planned, but we love it. We love our freedom, with all its excitements and dangers, triumphs and tragedies, blessings and curses. Freedom is turbulent and exhilarating—it’s all about living, not just being alive.
A free and open society could be described as naturalistic, meaning that things are pretty much allowed to take their natural course, with unforeseeable and often messy consequences. In a narrower context, this term was chosen by a team of Australian researchers to characterize a study they published recently (“A Naturalistic Study of Galantamine for Alzheimer’s Disease”). Their objective was to take a freewheeling approach to clinical trials, with fewer restrictions than the norm.
Formal Studies—Paradoxical and Difficult, but Not Bad
The conventional, formal approach is to choose the study population based on strict criteria that allow the inclusion of, or require the exclusion of, any given individual, so as to achieve two main objectives that are, alas, somewhat at odds with each other. The group chosen should be: (1) reasonably representative of the general population—in a word, heterogeneous, but also (2) as homogeneous as possible in terms of factors that would complicate the data analysis and make it harder to draw valid conclusions.
The formal approach, in other words, is to recruit individuals who represent the broader community but who are reasonably alike in terms of demographics, socioeconomic status, lifestyle factors, physical and mental health, and healthcare factors, such as drugs or supplements being taken. This allows the effects of the drug or supplement being tested to be more reliably distinguished from all the background “noise” created by differences in the varied aspects of life and living. (The ideal study population, in terms of minimizing the effects of most such variables, would be a homogeneous, regimented, institutionalized group, such as the inmates of a prison or insane asylum—but that raises other problems.)
With its built-in paradox, is the formal approach a bad one? Not at all—as should be obvious from the way I described it. It’s a reasonable compromise, designed to provide results that are as meaningful as possible under the always trying circumstances of clinical trials, which are difficult to design and execute well, not to mention interpret correctly.
People Behaving Badly . . .
The problem (and this too should be obvious) is that people in real life are not as “well-behaved,” in a sense, as those in the formal studies. For starters, they’re typically more diverse, in terms of all the factors mentioned above, than the study populations. Furthermore, they tend not to be as conscientious and cooperative about sticking to a diet or to a drug or supplement regimen as people who are enrolled in a trial, where researchers keep reminding the participants to be good and follow the protocol.
In real life, many people take their pills erratically or in the wrong amounts, or they just give up on them altogether. In medical practice (and in the supplement business), such “patient noncompliance” is a major problem: it can undermine or eliminate the therapeutic benefits predicted by the studies. And, of course, it can make doctors look less than competent when the therapy “didn’t work.” (With preventive measures, it’s usually impossible to tell in any case whether or not a substance is working, because there are no observable cues.)
. . . Are Part of the Naturalistic Approach
Because people in real life can be so problematic to researchers, the formal approach to clinical trials (which is hard enough) is usually adopted—but not by the team Down Under. They decided to take a naturalistic approach to see if the results obtained in a “clinically relevant setting” (to use their words) would agree with those of the many formal studies conducted on the plant alkaloid galantamine. Those studies have demonstrated convincingly that galantamine is a safe and effective treatment for mild to moderate Alzheimer’s disease, as well as for the common precursor condition known as mild cognitive impairment (MCI).
Let’s cut to the chase: the naturalistic results did agree, by and large, with those of the formal studies, thus providing yet another level of validation for what we already knew. The new study also highlighted some important differences between the two approaches and illustrated how a certain vexing problem caused by real-life behavior can be addressed.
Doctors’ Treatment Decisions Were Independent of Study
As their study group, the researchers used 345 people with mild to moderate Alzheimer’s disease who were recruited by 137 clinicians at 48 Australian hospitals over a 6-month period. Other than the diagnosis of Alzheimer’s, there were only minimal inclusion or exclusion criteria for participation. The patients had to live at home (not in an institution of any kind), they had to have supporters or caregivers with whom they were in regular contact, and they had to speak English well enough to be able to take the tests. They were excluded if they were hypersensitive to any ingredient in the galantamine formulation, if they were being treated or had been treated with a drug similar to galantamine, if they had severe liver or kidney impairment, or if they had any uncontrolled chronic condition other than dementia.*
Importantly for the naturalistic approach, the doctor’s decision to prescribe galantamine was made, in each case, before—and independently of—the decision to enroll the patient in the study. Thus, these patients were going to be getting galantamine in any case, regardless of the researchers’ opinion. The treatments did not begin, however, until the patients had been evaluated to establish their baseline profile; they were reevaluated at 3 months and again at 6 months. The prescribed galantamine dosages varied widely but averaged 10 mg/day at the outset and 15 mg/day at the 3-month and 6-month stages (the usual dosage in clinical practice, as in most of the formal studies, is 24 mg/day).
Lots of Dropouts . . .
A weakness of the study related to the naturalistic approach was that it was uncontrolled (no one received placebo) and open-label, i.e., all the patients knew that they were taking galantamine. This, of course, opened the door to the “placebo effect” (the power of mind over matter, causing improvement based on positive attitude alone), making it impossible to know how much of the improvement was due to the galantamine.
Of the 345 patients (average age 78) who were initially enrolled, 229 (66%) showed up for both the 3-month and 6-month follow-up visits, while 116 (34%) did not—they dropped out at some point. In medical parlance, the 229 finishers were the per protocol group, i.e., they abided by the treatment protocol (more or less—we don’t know how good they were about taking their galantamine faithfully as prescribed). The entire group of 345 constituted the intention-to-treat group, i.e., they were all intended to be treated and evaluated for the full 6-month duration, although 116 of them, ultimately, were not.
. . . Can Bias the Results
I know what you’re thinking: Why not throw out the incomplete data for those 116 dropouts and stick with the 229 who stuck it out? Let me answer that question with another question: What if those 116 were not representative of the overall average composition of the study population, because factors other than pure chance were involved in their dropping out? They would then have a composition that was skewed from the overall average of the intention-to-treat group. And if that were true, then the 229 in the per protocol group would also be skewed from the average, because if one group is skewed in one direction, then the other group must be skewed in the other direction in order for the overall average to have been what it was.
Thus, if the 116 dropouts were factored out of the analysis, the study would have a built-in bias toward the per protocol group—and any bias is bad, is it not? So it’s important that the incomplete data for the dropouts be taken into account somehow.
In another such test, 86% of
the PP group were somewhat
improved. These data showed that
galantamine was effective and that
dropping out was a bad idea.
Again I know what you’re thinking: If the incomplete data were included in the analysis, wouldn’t that tend to undermine the validity of the study by “diluting” good data with not-so-good data? Excellent question! The answer, generally speaking, is yes, although the degree to which this problem is significant depends on many factors. In any case, there are statistical techniques that can be used to get around the problem (with varying degrees of success). In this case, the Aussie researchers simply analyzed the “good” data for the per protocol (PP) group and, separately, the “good-plus-not-so-good” data for the intention-to-treat (ITT) group; they then compared the results. (Remember that the ITT group is the total number of patients that began the study, and the PP group is the number that finished it.)
Those Who Stuck with the Protocol Did Better
After 6 months, 70% of the PP group showed some improvement (compared with the baseline score) on a standardized test of cognitive function, whereas only 44% of the ITT group was improved. In another such test, 86% and 33% of the PP and ITT groups, respectively, were somewhat improved. These data showed that galantamine was effective (it’s unlikely that the placebo effect could have accounted for improvements of the magnitude seen) and that dropping out was a bad idea.
CIBIC-plus scores (see preceding page) for 207 patients with mild to moderate Alzheimer’s disease, after treatment with galantamine for 3 months and 6 months. Patients with scores of 1 through 4 were defined as responders. (Adapted from Ref. 1)
Another way to rate the patients’ condition is via a standard protocol called the Clinician’s Interview-Based Impression of Change plus Caregiver Input (CIBIC-plus), which is self-explanatory. With CIBIC-plus, patients are categorized as “responders” if their condition has not worsened (which it typically does without medication) or if it has improved to some degree (as is common, for a period of up to 1 year, with galantamine). In this study, 86% of the PP group were considered to be responders after 6 months, vs. 54% of the ITT group (see chart).
The Majority Did Well in Their Daily Lives
In other measures of the patients’ progress (or, to be more realistic, the slowing of their deterioration), the results were also encouraging. Over the 6-month period, the majority showed no increase in their degree of dependency on others, nor did they show significant decline in their ability to perform normal activities of daily living, such as using the telephone, traveling more than 1 km (0.6 mile) outside the home, taking medications, and managing money.
The same was true with an 11-item behavioral assessment scale devised for this study: at 6 months, the majority had either remained stable or had improved somewhat in their behavior patterns (aggression, personality changes, agitation, hallucinations, sleep disturbances, irritability, apathy, aberrant motor behavior, disinhibition, depression, and inertia).
What Price a Year?
There you have it: galantamine works well in real life—for 6 months, anyway—and that’s good to know. Based on what we know from the formal studies, it’s likely that the peak response occurs somewhere between 6 and 12 months, after which the inexorable decline resumes—but from a somewhat higher level. Thus, the patients’ galantamine treatment will have bought for them and their loved ones the most precious of commodities: time, perhaps a year’s worth. How can one put a price on the value of a year in the life of a loved one?
Turmeric Fights Dementia
Galantamine is not the only supplement that’s effective against Alzheimer’s disease. In a recent report on the relationship between curry consumption and cognitive function in elderly Asians, the authors stated, “In view of its efficacy and remarkably low toxicity, curcumin shows promise for the prevention of Alzheimer’s disease. . . . We present here tentative evidence that increased consumption of curry is associated with better cognitive performance in nondemented subjects.”
A spice commonly found in curry powder (particularly the yellow variety) is turmeric (Curcuma longa), which has powerful anti-inflammatory and antioxidant properties. Turmeric is a staple throughout much of India and Southeast Asia, where since ancient times it has been called “the spice of life.” It’s widely used in the treatment of arthritis and other inflammatory disorders.
The agents principally responsible for turmeric’s health benefits are curcumin and related compounds called curcuminoids. They are believed to help prevent or ameliorate Alzheimer’s disease, in part by counteracting the effects of the destructive protein amyloid-beta, which forms neuritic plaques in the victims’ brains. Amyloid-beta causes severe oxidative and inflammatory damage and kills neurons.
The epidemiological study cited above involved 1010 elderly (average age 69) residents of Singapore who were of Chinese, Malay, and Indian extraction. It was undertaken to provide confirmation of the extensive experimental evidence demonstrating the positive relationship between curry consumption (especially yellow curry) and cognitive performance.* In that goal, it succeeded. Particularly noteworthy was the observation that even relatively low or moderate levels of curry consumption were associated with a reduced risk for dementia, as indicated by tests of cognitive function; higher levels of consumption were associated with higher test scores.
- Ng TP, Chiam PC, Lee T, Chua HC, Lim L, Kua EH. Curry consumption and cognitive function in the elderly. Am J Epidemiol 2006;164:
- Brodaty H, Woodward M, Boundy K, Barnes N, Allen G, for the NATURE Investigators. A naturalistic study of galantamine for Alzheimer’s disease. CNS Drugs 2006;20(11):935-43.
Will Block is the publisher and editorial director of Life Enhancement magazine.