Drugged and Docile?

    This weekend, the New York Times uncovered new, seemingly damaging information about nursing homes, this time unrelated to their mishandling of the Covid-19 epidemic. “Phony Diagnoses Hide High Rates of Drugging at Nursing Homes,” is a detailed investigative essay by three reporters that reveals that nursing  homes regularly under-report the frequency with which they prescribe sedating, antipsychotic medications for residents with dementia. Such medications, while useful for controlling  paranoia and delusions (which may afflict people with dementia), have not been shown to be more generally helpful in controlling the behavioral symptoms of dementia. They are, however, associated with a two-fold increase in mortality and other adverse effects. As a result, nursing homes have been under pressure for years to limit their use of “chemical restraints,” medications that suppress agitation--including general sedatives and antipsychotics. 

            The campaign against the use of anti-psychotics in dementia began with the Nursing Home Reform Act of 1987 (OBRA-8)7, legislation asserting that residents have the right to be free from physical and chemical restraints that are “not required to treat specific medical symptoms.” Then, nearly a decade ago, CMS announced a new approach, the National Partnership to Improve Dementia Care and Reduce Unnecessary Antipsychotic Use in Nursing Homes. This strategy provided for training modules for nursing home staff on how to handle residents with dementia; the strategy also involved including as a “quality indicator” the proportion of long-stay nursing home residents receiving an antipsychotic medication. In 2015, this quality indicator was added to the list of measures that comprise the overall rating of nursing homes that CMS publishes on its website, Nursing Home Compare.

            The last 34 years have seen a marked decline in the use of antipsychotic medications in nursing home residents. The Nursing Home Reform Act of 1987 led to a 27 percent reduction in antipsychotic use. The introduction of psychoactive drug use as a quality indicator led to a further ten percent fall in prescribing. Today, CMS statistics assert that 15 percent of long-stay nursing home residents regularly receive antipsychotics for some problem other than Tourette’s syndrome, schizophrenia, or Huntington’s Chorea, for which antipsychotics are approved. The new report by the NY Times suggests that the correct figure is more like 21 percent, with the excess accounted for by bogus diagnoses of schizophrenia: the implication is that doctors want to control their demented patients by sedating them, but are discouraged from doing so by CMS regulations, so they get around the rules by falsely labeling their patients as schizophrenic.  

            The Times argues that “caring for dementia patients is time- and labor-intensive. Workers need to be trained to handle challenging behaviors like wandering and aggression. But many nursing homes are chronically understaffed and do not pay enough to retain employees, especially the nursing assistants who provide the bulk of residents’ daily care.” All true. The Times goes on to argue that nursing homes with poor staffing ratios (facilities that get a 1 or 2 star rating for the adequacy of staff: patient ratios) dole out more antipsychotic medications than those with better staffing ratios (facilities with 4 or 5 star ratings for staffing). 

While this graph strongly suggests an inverse relationship between staffing and antipsychotic use, what is equally striking is that all the facilities, regardless of staffing, administer antipsychotics to upwards of 15 percent of residents, without appropriate justification. What this suggests, contrary to the NYT implication, is that nursing homes find that no amount of training and no  number of staff members can consistently and reliably treat all the behavioral symptoms of dementia.

            The unfortunate reality is that we know very little about how to care for people with moderate to advanced dementia who exhibit problematic behaviors. The behaviors I am referring to are not merely inconvenient to staff, such as “wandering” off the nursing unit unattended. They include aggression, paranoia, and delusions. That means hitting or spitting; it means slugging or biting the well-intentioned caregiver who sought to bathe or feed a resident. These behaviors are disturbing to persons with dementia, their caregivers and, in a nursing home setting, other residents. 

            The Alzheimer’s Association, the American Psychiatric Association, and others have come up with guidelines for managing these symptoms. Their recommendations advise beginning with psychosocial interventions: first tryi to determine the cause of the behavior (perhaps the person slugged the aide who tried to give him a shower because the water was too cold) and then either address the root of the problem (for example, adjusting the temperature of the shower water) or engage in distraction (taking the individual to a “quiet room” with a box of trinkets and other treasures to examine). But then comes the caveat: “Unfortunately, large population-based trials rigorously supporting the evidence of benefit for non—pharmacological therapies are presently lacking.” The evidence for these approaches is largely anecdotal and commonsensical--as is true for anti-psychotic medication.

            What is the solution? The Times seems to suggest we should either penalize physician prescribers, fine nursing homes that over-prescribe, add further regulations, or all three. Another alternative is to substitute “Green Houses” for conventional nursing homes. The Green House is a model of nursing home care that seems to be successful in many domains where traditional nursing homes have failed abysmally, so there is some reason to believe it may be a good way to deal with dementia. The Green House project restricts homes to no more than 12 residents, employs a home-like focus, and cross-trains staff members to meet any and all resident needs rather than using a rigidly siloed model. Green House nursing homes may be the answer: they are associated with high rates of family satisfaction as well as a superior track record in controlling the Covid epidemic. I hope Green Houses have the solution, but I can find no published data on their ability to manage severe dementia. In fact, it’s far from clear Green Houses, which currently provide care to a total of 3000 people (out of a total nursing home population of 1.3 million), in 300 facilities (out of a total of 15,600 nursing homes nation-wide), actually care for people who have dementia and troublesome behavior. Their limited numbers, small size, and high cost (45 percent of residents pay out of pocket compared to 22 percent of residents in usual nursing homes) suggest they may be able to cherrypick their residents.

            I am not advocating blanket use of psychotropic medications in nursing home residents. Nor am I endorsing mislabeling patients with psychiatric diagnoses they do not have. Antipsychotics should only be used in people with dementia in well-defined and relatively rare circumstances. Perhaps informed consent by the patient's health care proxy should be a prerequisite. But blaming nursing homes or pointing the finger at nursing home doctors or devising new regulations are not the best ways to serve people afflicted with dementia. What we need is to find better ways to alleviate the suffering of people with dementia. We should recognize that the doctors who prescribe antipsychotics are not necessarily lazy or devious, though any who are should be disciplined; many of them are simply desperate, desperate to provide relief to their patients. Medicare is poised to spend billions of dollars on aducanumab, a drug recently approved by the FDA for treatment of Alzheimer’s disease despite the paucity of evidence that it works and the abundance of evidence of its association with severe side effects. Perhaps we should instead devote resources to what is likely to be the more tractable problem, the symptomatic relief of Alzheimer’s disease and other dementias.

 

 

 

 

 

  

Deja Vu All Over Again?

 

The big news in geriatrics this week was the FDA  approval granted to a drug against Alzheimer's disease, the first new drug in 20 years. It reminded me of the day in 1986 when the initial report about what would be the very first FDA-approved drug against this disease appeared.

It was November 13, 1986 and I had been a practicing geriatrician for four years. My weekly copy of the New England Journal of Medicine had arrived right on time, as it did every Thursday. I scanned the table of contents and one article immediately jumped out at me. It had the suitably serious, scientific-sounding title “Oral tetrahydroaminoacridine in long-term treatment of senile dementia, Alzheimer type.”

 

During four years  of clinical practice—plus a year of geriatrics fellowship and three years doing an internal medicine residency—I had encountered patients with what we then called “SDAT,” (senile dementia of the Alzheimer’s type) and which we now simple call Alzheimer’s disease. The cognitive impairments of dementia were to me, to family members of the afflicted, and often to the affected themselves, among the saddest of the many disorders that develop among older individuals. Death, while also very sad, was part of the natural order of things, especially when it came after a long and rich life. But dementia in general and Alzheimer’s disease in particular was devastating because it attacked personality; some would say it assaulted personhood itself. While I would go on to spend much of my career thinking about how best to enable people with dementia (as well as those with physical frailty) to live meaningful lives despite their limitations, I recognized then and continue to believe today that the condition is a scourge that we should strive to prevent, eradicate, or at least ameliorate.

 

The medication described in the 1986 NEJM article would ultimately be approved by the FDA under the name of Tacrine for treatment of mild to moderate Alzheimer’s disease; it would be supplanted by its first cousin, the drug donepezil, brand name Aricept; and Aricept would top $2 billion in US sales by the time it came off patent in 2013. The story of the drug’s development says volumes about Americans’ desperation for a medical fix to Alzheimer’s, about big business, and about our regulatory system. Both the similarities and the differences between the Tacrine story and the tale of the new drug approved by the FDA for the treatment of Alzheimer’s are illuminating.

That 1986 study ostensibly showing that Tacrine led to improvements in cognition as well as to overall functioning was based on a whopping 17 patients, only 14 of whom actually completed the study. Its lead author was Dr. William Summers, a psychiatrist at UCLA medical center who had never before published anything of importance and had done very little research altogether. The scientific community immediately began questioning not only Summers’ credibility but also his methodology. Did the 17 patients actually have compelling evidence of an Alzheimer’s diagnosis? Did the “global assessment rating” used to measure outcomes translate into meaningful improvement?

 

After Summers filed for FDA approval for his drug,the FDA investigated Summers and his lab. The agency issued a rare “interim report” in 1991 in which it criticized Summers for the absence of documentation that the study was actually performed as  claimed in the NEJM  paper.  It questioned the randomization process and whether the physicians were, as asserted by Summers, blinded to what drug the patient was receiving. The best the FDA could say was that there was “no clear evidence of purposeful misrepresentation.”

 

In response, the public vilified the FDA, claiming the agency was “heartlessly impeding the relief of suffering.” David Kessler, the FDA director, received hate mail.

 

Approval of the drug came, but only after the release in 1992 of a larger more carefully conducted study by the “Tacrine Collaborative.” The trial lasted for 12 weeks and was carried out at 23 centers involving 468 patients. The results, published in the Journal of the American Medical Association, showed a statistically significant improvement in cognition and in overall function, whether measured by physicians or caregivers. And so, the first drug was approved for treatment of Alzheimer’s disease. Sales soared.  But questions continued to plague use of the drug—a subsequent study, for example, testing the effectiveness of a higher dose of the drug,  found that the higher dose was more effective than lower doses—but more than 2/3 of the patients dropped out of the study. Tacrine was soon effectively replaced by donepezil (Aricept), another cholinesterase inhibitor that differed only from Tacrine in that it is taken once a day rather than twice and has fewer gastrointestinal side effects. Patients and families demanded these drugs, which were soon followed by chemically slightly different but no more effective agents such as Excelon; the drug companies advertised them widely and made a small fortune on their  sales; but clinicians remained skeptical. I, for one, believe that the cholinesterase inhibitors are next to useless. None of the numerous studies of the drugs carried out since 1986 have persuaded me otherwise.

 

Fast forward to 2021 and the approval by the FDA of aducanumab under the brand name of Aduhelm. This is a completely different kind of treatment. Tacrine and Aricept are cholinesterase-inhibitors: they work by increasing the level of the neurotransmitter, acetylcholine, which is dramatically reduced in Alzheimer’s. They are given orally and were never terribly expensive, although Aricept got cheaper when the generic version became available. Side effects, particularly in the case of Aricept, are modest and consist of nausea and very rarely of liver enzyme abnormalities.  The new drug, by contrast, is a monoclonal antibody that works to mop up deposits of beta-amyloid from the brain, the chemical widely thought to cause Alzheimer’s disease. It must be given intravenously once a month. The average yearly cost will be set at $56,000. Two years ago, its manufacturer, Biogen, stopped an ongoing clinical trial of the drug after interim analysis failed to demonstrate efficacy. The drug company then re-analyzed the data and claimed it was effective after all, but an FDA Advisory Panel, convened in  November, 2020 unanimously concluded there was insufficient evidence of significant benefit to proceed. 

 

Multiple other monoclonal antibodies targeting beta amyloid have also failed, leading some to suggest that by the time these drugs are given to patients, it’s already too late. Or maybe beta amyloid is a marker for the disease and not the cause of the disease. It is in this setting, that the FDA approval of aducanumab is something of a surprise.

What is particularly striking about today’s FDA approval is that it is not based on the clinical effectiveness of the drug. Rather, it is based on its ability to clear the brain of amyloid deposits. There is a long tradition of requiring that drugs cause improvement in clinically meaningful outcomes, not just in “surrogate markers.” Cholesterol-lowering drugs were approved based on their ability to prevent heart attacks, not just on their ability to lower blood cholesterol levels. Antihypertensive drugs were approved based on their effectiveness in reducing strokes, not just on their capacity to lower blood pressure. To be sure, the FDA is holding off on full, unconditional approval until it sees the results of a yet to be performed large clinical trial demonstrating long-term benefit of amyloid plaque reduction. In the meantime, anxious patients and their families will submit to monthly injections of a drug that has been shown to cause symptomatic brain swelling, manifested by nausea, vomiting, visual problems, headaches and sometimes small strokes, in 40 percent of cases. 

We are a long way from the trials and tribulations of tacrine, but in the end, is the tale of aducanumab any less disturbing? In both cases, there was enormous pressure by the public to approve a drug that offered hope, some hope, however slim, of ameliorating this terrible disease. In both cases, the FDA, after seemingly acting based on scientific considerations alone, seemed to succumb to external pressures. And in both cases, the pharmaceutical industry stood to gain enormously by release of the drug. The FDA currently has an acting director: President Biden has yet to name a new, permanent head. Maybe it's time for the FDA director to become a civil servant, selected by the FDA members, rather than a political appointee. At the very least, the new director, when chosen, should take a long hard look at decision-making within the organization.

Have We Been Barking Up the Wrong Tree?

More of my blog posts deal with dementia than with any other subject and the news about Alzheimer’s disease over the years has been largely dispiriting, so who would have thought that I would leap at the opportunity to write about a new diagnostic test. But with so much of the medical literature relentlessly focused on COVID-19, it’s reassuring to realize that research on other subjects is continuing. The new study does not report a treatment, let alone a cure for Alzheimer’s disease. Furthermore, the prospect of screening healthy individuals to determine their future risk of developing progressive cognitive impairment is ethically fraught. Nonetheless, in the current climate, this report is good news.

It’s good news, and not just because it indicates that not all medical scientists have retooled as corona virus researchers, though it does that. It’s good news, and not just because it means it will be possible to target intervention studies to high risk individuals will permit studies to be carried out on smaller numbers of people and over a shorter period of time, though it means that. It’s good news because it shines a bright light on a long-neglected character in the Alzheimer’s story, the tau protein. 

Back in 1906, when Alois Alzheimer peered into his microscope at tissue from the brain of a patient who had died of the disorder of cognition that would one day bear his name, he identified two unusual substances that he described as plaques and tangles. The plaques, which were located between neurons, would ultimately be found to be composed of a protein known as amyloid. The neurofibrillary tangles, which were located inside the nerve cell bodies, would eventually be identified as a protein called tau. These two substances have been recognized as the hallmarks of Alzheimer’s disease for over a century.

For years, the roles of amyloid and tau were hotly debated. Some researchers felt that amyloid was the result of Alzheimer’s; others were confident it was the cause. Some scientists were more interested in studying amyloid; others directed their efforts towards tau. But over the course of the last 25 years, amyloid has gained the upper hand. Study after study has sought to improve cognition in Alzheimer’s disease by ridding the brain of amyloid-laden plaques—and each time, the approach failed. 

A great deal of excitement was engendered by immunotherapy back in 2001: the idea was to stimulate the body to create antibodies against amyloid with what was essentially a vaccine—but the study had to be stopped because a subset of patients developed meningitis. Then there was enthusiasm about the use of monoclonal antibodies. Several such antibodies have made it to phase 3 trials in which their efficacy was compared to placebo. In 2014, two studies of Bapineuzumab showed no benefit. In 2018, Solznezumab was tried for individuals with mild Alzheimer’s and it was unsuccessful. In the same year, additional negative results were reported for Verubecestat in people with mild to moderate Alzheimer’s. 

All these negative studies don’t exonerate amyloid. Maybe the trials are initiated too late in the course of these disease’s development. Maybe the dose is too low. But with anti-amyloid strategies repeatedly striking out, I can’t help but wonder, as have others who know much more about the science than I do, that we’re looking at the wrong target.

Which is why the new study that focuses on tau is exciting. The authors found that their tau antibody test was able to diagnose Alzheimer’s disease as well as or better than more invasive existing tests—when they used the test in patients all of whom had some kind of neurodegenerative disease. That is, the test did well in answering the question: is this person being tested more likely to have Alzheimer’s or, say, Parkinson’s? That’s a very different question from: is this person normal or does he have Alzheimer’s? Not only was the population in which the test was studied composed exclusively of patients with some neurologic condition, not only did the population include a much larger proportion of people with Alzheimer’s than would be found in the general population, but the subjects were far from ethnically or racially diverse. So, it’s a long way from the article in JAMA to a widely useful diagnostic test.

Despite the test’s preliminary nature, it is a compelling piece of evidence that tau should get more attention. Two weeks before the on-line publication of the JAMA study, a small Swiss pharmaceutical company, AC Immune, announced that together with Johnson & Johnson, it was launching a trial of a vaccine designed to stimulate the body to produce antibodies against tau—leading its stock price to soar by 18.9 percent. Just a few weeks earlier, the giant Swiss pharmaceutical company, Roche, announced it, too, was investing in the development of an anti-tau vaccine. 

It’s too early to say whether the attack on tau will fizzle, much like the previous attacks on amyloid. But maybe, just maybe, it will be a rousing success.

Never Say Old

As a brief follow up to her important 2011 book, Never Say Die: The Myth and Marketing of the New Old Age, the writer Susan Jacoby published an editorial in the NY Times this past week. 

Provocatively entitled,  “We’re Getting Old, but We’re Not Doing Anything About It,” she points out that the presence of 5 septuagenarian presidential candidates (Biden, Sanders, Warren, Bloomberg, and Trump) falsely encourages the population to think that 75 really is the new 50 and that aging doesn’t bring with it the risk of frailty, dementia, and disability. Jacoby, in her usual incisive way, argues this view is mistaken—and dangerous.

In fact, while many people do remain vigorous well into their seventies, and a smaller but not inconsequential percentage continue to be robust well into their eighties (and an even smaller percentage into their nineties), the risk of developing one or more medical conditions that get in the way of independence and well-being rises steadily with age. The much vaunted “compression of morbidity,” the dream that we will all remain totally intact until precipitously, preferably during sleep, we die, hasn’t quite materialized. Jacoby cites a study reporting that one in seven people who are 65 years old today can expect to experience 5 or more years of disability before they die. Dementia rates are falling—but since the size and proportion of older people is increasing, the absolute number of people with dementia is projected to rise steadily. According to the Alzheimer’s Association, there are 5.6 million people over 65 with dementia today, there will be 7.8 million in 2025—and 13.8 million in 2050. Frailty remains a major problem: a 2015 population study in the US found that while only 10.7 percent of people age 65-69 were frail, among the 75-79-year-olds the rate was 20.1 percent, and for those aged 85-89, it was 37.9 percent.

So, what are the implications of this reality? Jacoby offers three: provide more support for caregivers; encourage employers to allow vigorous older adults to continue working; and address various medical ethical issues such as physician assisted suicide. This is a good beginning—and about as much as one can reasonably put into a NY Times op-ed—but there are many others.

Here are a few additions: In the arena of housing, we need far more housing that is handicapped accessible. Doorways should be wide enough for wheelchairs, buildings should be on a single level or have elevators. In the domain of urban planning, communities need to be walkable with extensive and accessible public transportation. In terms of medical care, the overriding issue is not care of the dying, but how best to care for the living. We need far more attention to function: to maintaining and fostering basic abilities such as mobility and higher-level abilities such as cooking and shopping. To achieve this end, physicians need to assess function, they need to know how to diagnose frailty, and they need to know how to promote and support function. Physicians also need to determine the old person’s goals of care and to work with patients and family caregivers to develop an approach to treatment that is consistent with those goals.

The new year will begin in just two days. Among our resolutions for the coming year should be a commitment to making America a better place for us as we get old—and I agree with Jacoby that we should abandon the foolish euphemism “older people” and stick with plain “old.” This should be a resolution we actually follow.

Rescinding the DNR order--for anti-amyloid drugs

The pharmaceutical firm Biogen issued a stunning press release this week: it would seek FDA approval for its anti-Alzheimer’s drug, Aducanumab—the very same drug the company had pronounced a failure just last March, after preliminary analyses indicated the drug was very unlikely to achieve its objectives. Dennis Selkoe, a prominent Alzheimer’s researcher at the Brigham and Women’s Hospital in Boston, wrote in STAT (a health-oriented news website owned by the Boston Globe) that he believed that “aducanumab is the breakthrough we have been waiting for.” But is it?

Aducanumab is by no means the first drug studied that targets beta amyloid, a protein widely believed to play an important role in the development of Alzheimer’s disease. An article in Nature Reviews (Drug Discovery), published shortly after the first analysis of the Aducanumab data, put the drug in perspective: a monoclonal antibody that binds to beta amyloid, it is the fourth such drug to make it to a Stage III clinical trial. And all three of its predecessors, as well as another five anti-amyloid drugs that have a different mechanism of action, were abysmal failures leading Nature Reviews to proclaim "Anti-Amyloid Failures Stack Up as Alzheimer's Antibody Flops." Maybe what this new analysis of the Aducanumab data has done is to reverse the Do-Not-Resuscitate label previously attached to monoclonal antibodies and instead put them on life support. So, what, exactly, does the new data analysis show? 

The details will be released to the FDA when Biogen files with the FDA to approve the drug. Thus far what we know is what Biogen has included in its press release. Biogen conducted two randomized, double blind, controlled studies, giving the monoclonal antibody (presumably by injection, like other monoclonal antibodies) or placebo to a total of 3285 patients and following them, some for 78 weeks. One study, sporting the acronym ENGAGE, produced negative results even upon re-analysis, although a subset of ENGAGE patients who received a high dose of the drug appeared to respond. The other study, named EMERGE, showed a reduction in the rate of clinical decline in treated patients compared to those receiving placebo. In other words, the study patients did not improve nor did they plateau; rather, they got worse more slowly. Unfortunately, the press release doesn’t report the absolute magnitude of the change, opting instead to say that the scores on the Mini-Mental State Examination (MMSE), a commonly used measure of cognition, fell 15 percent less with treatment than with placebo. Similarly, the rating on the ADAS-Cog13, another measure of cognitive function, declined 27 percent less in those on the drug. Finally, and perhaps most encouragingly, the decline in the ability to perform basic daily tasks (as measured by the Alzheimer’s Disease Cooperative Study Activities of Daily Living Inventory) was 40 percent lower in those receiving Aducanumab than in their counterparts who received placebo. 

These are not spectacularly impressive results—the patients continued to worsen and the difference between treated individuals and untreated individuals, at least on measures such as the MMSE, was very small. But what we don’t know is whether the slowed rate of decline will persist over time or whether, as with drugs that are currently on the market such as donepezil (Aricept), there is a one-time effect. We don’t know whether the effect would be greater if the medication were used earlier in the progression of Alzheimer’s disease, perhaps in people with mild cognitive impairment. We don’t know whether the drug will prove to be more effective in those with “pure” Alzheimer’s disease, that is, without concurrent vascular changes. 

What we can predict is that if the drug is approved by the FDA, Biogen, whose stock price soared after the press release, will aggressively market the drug. An estimated 5.8 million Americans have Alzheimer’s disease, according to Alzheimer’s Association statistics released in 2019: since the time from diagnosis to death is 5-7 years, roughly one-fifth of the affected individuals or over a million people likely have early Alzheimer’s and are potential Aducanumab consumers. This could be a dream come true for the pharmaceutical company--a blockbuster specialty drug. (Blockbuster drugs are prescribed to at least one million people a year and specialty drugs are very expensive biologicals such as monoclonal antibodies). And if the sales price for the drug is anything like that of other newly marketed monoclonal antibodies, it may be upwards of $1000 per month (Dupilumab, brand name Dupixent, a new monoclonal antibody used in the treatment of asthma and/or nasal polyps, has a list price of $37,000 per year). 

Biogen stockholders will benefit enormously; how much people with Alzheimer’s disease will benefit is less clear. But on a more positive note, as Dennis Selkoe commented in STAT, for the first time we have evidence that it is possible both to design molecules that target amyloid and to administer them to real people with minimal toxicity and some evidence of benefit. That truly is good news.

Clearing the Air

I’ve been reading “The Uninhabitable Earth: Life After Warming,” an extensively researched account of where we are headed that begins with the warning: “It is much, much worse than you think.” When I got to the chapter depressingly entitled “Unbreathable Air,” I encountered the following shocking sentence: “Pollution has been linked with increased mental illness in children and the likelihood of dementia in adults.” Now I’ve seen all kinds of things associated with dementia: head trauma, assorted medications (anticholinergic drugs, anti-anxiety drugs, and anti-ulcer drugs), aluminum. Some of those links have become well-established over time, such as head trauma. Some of have been totally debunked, such as aluminum. Others are questionable and I’ve written about them on this blog (drugs). But air pollution? This was a new one to me. 

Most likely, I figured, it would prove to be another spurious association. Probably, I thought, there was some other factor that was associated with both air pollution and dementia. The alleged connection would be like the link between washing machines and colon cancer—a favorite example of a “confounder” from my medical school epidemiology class. People who own washing machines, it turns out, do have a higher rate of colon cancer than people who don’t. But they also vary in where they live and what they eat, which is far more important than their possessing a washing machine. Surely air pollution was likewise a marker for something that did matter. But then, as I read on in Wallace’s book, I came to an even more dramatic statement: “An enormous study in Taiwan found that, for every single unit of additional air pollution, the relative risk of Alzheimer’s doubled.” This I had to look into.

The “enormous study in Taiwan” was published in a minor but respectable journal, the Journal of Alzheimer’s Disease in 2015. It was large: it was a cohort study of 95,690 adults aged 65 and older followed prospectively for 10 years beginning in 2001. Not only was it large, but it was a random sample drawn from Taiwan’s National Insurance Research Database comprised of 23 million people, or 99 percent of the entire Taiwanese population. Moreover, Taiwan has 70 EPA monitoring stations distributed over the island, allowing it to have reasonably accurate measures of both ozone exposure and small (less than 2.5 micrometers) particulate measure. Finally, the population is fairly stable over time, allowing for fairly good estimates of exposure based on home address. The conclusion? The risk of newly diagnosed Alzheimer’s disease (adjusting relevant co-morbidities such as stroke, hypertension, and diabetes) rose steadily with the rate of exposure to ozone or small particulate matter—going up, for example by 211 percent for each 10.91 ppb increase in ozone.

Taiwan isn’t the only place where a relationship between air pollution and dementia has been discovered. In 2017, a similar study entitled “Exposure to Ambient Air Pollution and the Incidence of Dementia: A Cohort Study,” appeared in Environmental International. Carried out in Ontario, Canada and involving a cohort of just over two million adults, this analysis attributed just over six percent of all dementia cases to air pollution. 

Neither study is conclusive, but they’re awfully suggestive. I wondered if there had been any further work on this subject since Wallace wrote his book. Lo and behold, a systematic review was just published by Peters et al from Australia, also in the Journal of Alzheimer’s Disease. These authors found thirteen reasonably well-conducted studies bearing on the question. They concluded that small particulate matter (containing nitrogen) and carbon monoxide are both associated with an increased risk of dementia. 

These reports are very disturbing in light of the Trump administration’s systematic assault on air pollution regulation. According to an article just published in the New York Times, 85 environmental rules are being rolled back, including 24 in the arena of air pollution. Of these 24, 10 have already been undone and another 14 are “in process.” 

We already know that climate change will have an enormous impact on health, principally through its multitudinous indirect effects—for example, by causing drought, which in turn affects agricultural productivity, which in turn results in death. Now there may be another health risk to add to the list of adverse effects of environmental harm. Dementia is such an enormous public health problem that even measures that only slightly affect the risk of developing this devastating condition may be worthwhile. But the good news is that air pollution is an area where we can intervene. We even know how to. The last thing we should be doing is unraveling the progress we have made. So, speak up, tell your senators and representatives to act, and vote wisely to decrease the pollution that threatens us all.

The Dirt on Drugs

Several months ago, in writing about Katy Butler’s generally admirable new book, The Art of Dying Well,I questioned her uncritical acceptance of claims that certain drugs, the anticholinergics, caused dementia. That drugs such as antihistamines and some antidepressants cause delirium, or acute confusion, is well-established. But dementia? A large, well-conceived study just published in JAMA Internal Medicine provides additional evidence that they may well result in dementia. Previous studies were tainted because they tended to be small, observational studies that lumped many different anticholinergic drugs together, some of which are known to be far more potent than others. So how does the new study fare by comparison?

The recent analysis is still an observational study—it is impossible to randomize patients into those who receive anticholinergics and those who do not and then follow them for years—but it is very large, it analyzes different classes of anticholinergics separately, and it focuses on the cumulative anticholinergic exposure over a period of many years. While not perfect, it is probably the best we are going to get and it provides strong, although not definitive, evidence that anticholinergics are a risk factor for dementia.

This “nested case control study” allowed the authors to identify just under 59,000 people over age 55 with dementia from a British data base of 30 million. For each case, they found 5 controls matched along four discrete dimensions. The main variable of interest was the total, cumulative anticholinergic exposure which they calculated according to a well-established protocol that allows standardization across drugs with different dosage regimens. The authors also decided in advance to study the association between the total standardized daily dose (TSDD) for each type of anticholinergic: antihistamines, antidepressants, antipsychotics, antiparkinsonian drugs were considered both separately and collectively. Finally, the study repeated the analysis for patients diagnosed with Alzheimer’s disease, those diagnosed with vascular dementia, and those with some other form of dementia. So what did they find?

The main finding was that the adjusted odds ratio associated with low anticholinergic exposure was 1.06, while the ratio associated with the highest degree of anticholinergic exposure was 1.49, a highly significant difference. Also interesting was the observation that only certain groups of anticholinergics increased the risk of developing dementia. In particular, antidepressants, antipsychotics, antiepileptics, and drugs used to treat Parkinson’s and incontinence were the main offenders (and antihistamines, which geriatric physicians inveigh against as potential causes of delirium, had no effect). Curiously, the effects were more dramatic in patients with vascular dementia than in those with Alzheimer’s disease, a novel finding. Finally, the strongest association was found in people diagnosed before age 80.

As the authors are quick to point out, associations do not demonstrate causality. They can’t. Researchers simply cannot exclude the possibility that some of the drug use in fact reflected early, preclinical effects of as yet undiagnosed dementia. For example, many people with dementia are depressed; it is entirely possible that the depression manifests itself before a formal diagnosis is made, at a time when patients are beginning to detect subtle but disturbing changes in their memory and problem-solving ability. But if anticholinergics are causative, then they may well be responsible for as much as 10 percent of all dementia. The evidence is sufficiently suggestive and the magnitude of the danger sufficiently great that it’s time to be very wary of these drugs, especially in people under age 80. 

Treatment for Alzheimer's or False Alarm?

The most promising treatment of Alzheimer’s disease to date uses monoclonal antibodies to rid the brain of amyloid plaques. These plaques are widely thought to be crucial in the development of Alzheimer’s disease, the form of dementia that afflicts over 5 million Americans—and, unless we find a way to cure or prevent it, will afflict millions more in the coming years. Researchers hope that these drugs will prove to be “disease-modifying,” that they will disrupt the cascade of events that produces the symptoms of the disease. 

So far, the studies of monoclonal antibodies in this setting have been very disappointing: Bapinezumab: 3 failed trials. Solanezumab: 3 failed trials. Crenezumab: 1 failed trial. In the past year, three major pharmaceutical companies, Eli Lilly, AstraZeneca, and Merck, all stopped development of the monoclonal antibodies they were testing against Alzheimer’s disease. Another company, Pfizer, closed its neurodegenerative disease research section entirely. Now we learn that a Phase II trial of the as yet unnamed agent “BAN2401” actually “looks promising.” How promising?

Like its cousins, BAN2401 is a “humanized monoclonal antibody.” That means an antibody that was developed in mice and modified so it wouldn’t be rejected as alien by the human immune system. Now mice don’t normally get Alzheimer’s disease but certain strains have been genetically engineered so that they do. Using the genetically engineered mouse model, a monoclonal antibody was created that binds to amyloid. Today, this chemical is called BAN2401; when it enters the next stage of assessment, it will presumably acquire a name that ends in “zumab” to indicate it is just such a humanized monoclonal antibody. So, what do we know about this nameless entity?

We know that BAN2401 made it through a Phase I trial that tested its safety, tolerability, and pharmacokinetics in 80 people with mild to moderate dementia. We know it then entered Phase II of testing in people aged 50-90 with evidence of amyloid on their PET scans and either mild Alzheimer’s disease or mild cognitive impairment (a condition that is not dementia but often evolves into dementia). Five different dose regimens were administered intravenously and compared to placebo in terms of their effects on cognition and on brain amyloid deposits on PET scan. We know that when the preliminary results were announced last December, no cognitive benefit was reported. But now, after 18 months of study, the final results are in. They were presented at the Alzheimer’s Association International Conference in Chicago this past week.

What we don’t know is how this study will fare when subjected to peer review. We don’t know what the final paper will look like. But here’s what I can glean from the press release by Eisai, the company that collaborated with Biogen to develop and test the drug, and the reports by the news media:

--The amount of amyloid in the brain (as measured by PET scan) decreased with all the doses tested. The more drug, the greater the change. The findings were statistically significant. 

--There was no discernible effect on cognitive function except in the 161 people who were treated with the highest dose of BAN2401. In those 161 people, cognitive function deteriorated over time, just as it did in all the other groups, but it deteriorated less.

--Cognitive function was measured in a way that differs from the way it has been measured in other studies of treatments for Alzheimer’s disease. The researchers used a composite measure made up of several scales, each of which has been individually validated as a way to assess mental status, but only one of which is regularly used to report the outcome of clinical trials of Alzheimer’s drugs. 

--The patients tested included a mix of people with mild dementia and people with mild cognitive impairment (who don’t actually have dementia but aren’t entirely normal either) and the study had no way to determine whether the effectiveness of BAN2401 was any different in the two subgroups.

What does all this mean? The PET scan changes indicate that the monoclonal antibodies were successful in destroying amyloid deposits in the brain. That doesn’t prove that whatever effect the drug had was due to its amyloid-busting, but it’s suggestive. It also indicates that to the extent that the drug was a failure, it wasn’t a failure because it didn’t destroy amyloid. 

How should we interpret the cognitive changes? It’s hard to know, given that the researchers used a non-standard means of measuring clinical decline (a test they called ADCOMS). We cannot, for example, say that BAN2401 is better than one of the other monoclonal antibodies that was deemed a failure since nobody looked at the effect of the other drug on the “ADCOMS.” What we can say is that a 30 percent change in the rate of decline may be statistically significant but clinically, it’s a lot less impressive. Everyone  deteriorated, including those given the highest dose regimen; they just deteriorated a little less. How much of a difference actually occurred depends on the absolute decline: suppose the score went from 100 at baseline to 80 at the end of the study in the controls (I don’t know what the actual numbers are, so this is hypothetical). A 30 percent difference in the rate of decline means that the score in the treated group dropped from 100 to 86.  Is that clinically meaningful? Probably not.

So, no, BAN2401 is not likely to likely to dramatically change the course of Alzheimer’s disease. I hope the FDA does not allow the drug’s manufacturers catapult it into a clinical trial prematurely, before it goes through Phase 3 testing. But it has taught us a few valuable lessons. 

We now know that monoclonal antibodies can be designed that destroy amyloid plaques in the brain. We know that only the highest tolerated dose of the drug has any chance of being clinically useful. We should also realize the importance of studying patients with early dementia and those with MCI separately—it’s possible that using monoclonal antibodies once dementia has set in is too late, and I suppose that it’s also possible that using them before dementia has developed is too early. Finally, the study reminds us of the importance of a uniform methodology in conducting this kind of work. If the test that was previously regarded as the gold standard, the ADAS-Cog, is not the right test to use, then researchers need to agree on that and decide collectively what test to use instead. 

ABC News reported the drug as “giving patients hope” and showing “big promise.” Fortune magazine asserted that “world leaders want to end Alzheimer’s by 2025; a new drug breakthrough means we just might.” Investors are closer to the mark: Eisai’s stock price fell 10 percent after the data were reported.

As Simple as ABCD: A Better Care Directive

Kudos to Barak Gaster, a general internist at the University of Washington, who has developed a dementia-specific advance directive. In this document, he starts out by explaining what dementia is and why patients and families might not want every possible medical treatment throughout the course of this long illness. Then he simply and clearly characterizes each of the three main stages of dementia: mild, moderate, and severe (or early, mid-stage, and late) and asks what the goal of care would be in each instance. 

I like the framework because it's basically what I have used for the past twenty years in articles and books. Gaster uses the same formulation that I propose—prolonging life, maximizing comfort, and something in between that is related to maintaining function and dignity—but adds one more option, life prolonging therapy minus attempted CPR. That’s it. No long list of potential treatments, ranging from ventilators to iv’s as with typical instructional directives (or, indeed, with the POLST, which is an instructional directive turned into a medical order). No complicated list of symptoms for each of the three stages of dementia, a list that leaves people reeling and uncertain about just what stage a given person is actually in.

            

Here are the parts of the document:

    1. What is dementia? Three short paragraphs indicating that Alzheimer’s disease and other dementias are progressive, ultimately fatal diseases. Indicates average time from onset to death (says 8 years, though some of the data I’ve seen indicate 5-7 is more accurate). Highlights features of last stage.

     2. Why is it important to express your wishes? Emphasizes that patients lose capacity and families benefit from receiving guidance.

     3. What kind of guidance can you give? Explains that tests and procedures are more difficult to tolerate as dementia progresses. Also emphasizes that many people would not want life-prolongation as their thinking and ability to derive meaning from life deteriorate.

     4. Stages of dementia:

     

     Stage I (mild): lose ability to remember recent events; 

     routine tasks become difficult (ie cooking); other tasks can

     be more dangerous (ie driving).

     

     Stage II (moderate): lose ability to have conversations and

     to understand what is going on around them. Require full 

     time assistance with dressing (I would quibble over this one;

     people do need some help with their “activities of daily

     living,” one of which is dressing, but I don’t think singling out

    dressing and saying “full time assistance” is needed is quite

    accurate.

    

    Stage III (severe): no longer able to recognize family, may become angry and agitated (one caveat here: agitation often starts in moderate dementia); need round the clock help with all daily activities.

    Even the description of the goals of care are well done: there is some explanation of why a person might choose a particular goal. For example, if the goal of care is “to only receive care in the place where I am living,” the text clarifies that the reason for not wanting to go to an emergency room or be hospitalized is to avoid the “possible risks and trauma that can come from being in the hospital.”

    For people who find images helpful to elucidate the words, I suggest viewing the videos from acpdecisions.org addressing dementia. Angelo Volandes, president of ACPdecisions, has filmed a series of dementia-related videos based on the script I offered: one addresses the stages of dementia, and another details the goals of care across the spectrum of the disease (suggesting that many people favor different goals as the illness becomes more advanced).

    

    The proposed new "advance directive for dementia" is a

    valuable contribution to the tools available to help individuals plan for the future. I hope it will be used widely--and that individuals who are physically frail will also engage in a similar type of advance care planning. It's high time we recognized that thinking about the future is not just for those who have no future.