Anticholinergic agents and dementia

A new study came out this week suggesting use of a particular class of drug after age 65 was associated with dementia.  Here’s what you need to know.*

The study is retrospective, meaning it took people who developed the disease of interest and then looked backwards at their medications.  Retrospective studies are prone to a number of problems, the biggest one being that even young people with healthy memories are crap at giving you their drug history over the past 10 years, and this is a study of people with dementia.  The researchers dodged this by using an HMO database of the subjects complete medical history, which is a neat trick.  The second problem is that retrospective studies can easily end up being painting the bulls-eye after they’ve fired the arrow.  Mere chance dictates that if you track enough traits, any random subset of a population is likely to have something more in common with each other than with the rest of the population.  If you use the traditional bar of statistical significance (5% chance of results arising by chance), checking 20 traits gives you an expected value of 1 false positive.  To be fair, this study has a much higher significance level, and the effect was dose dependent, which is a very good sign that it’s legit.  The authors heavily imply they deliberately studied anticholinergics rather than shotguning it, but without preregistration there’s no way to be sure.

Anticholinergics come in two forms: antimuscarinics, and antinicotinic.  Short version: these work on different types of neuroreceptors, which live in different parts of the body and do different things . Every example drug they give is an antimuscarinic and of the classes of drugs they list, many have no antinicotinic members.  Even if they technically included antinicotinics in the analysis, they would be such a small portion of the sample that their effect could be overwhelmed.  So I don’t think you can apply this study to drugs like bupropion, which is an antinicotinic.

I don’t like the way they calculated total exposure at all.  Essentially they counted the normally recommended dose of any medication as One Standardized Daily Dose.  But those dosages vary wildly (even the examples they give span an order of magnitude), as do the particular drugs’ ability to cross the blood-brain barrier.  The drugs are prescribed for a huge variety of causes, and what’s sufficient to stop incontinence has nothing to do with what’s sufficient to slow Parkinson’s.  This oversight may cancel out with the fact that they created buckets of dosages rather than do a proper linear regression, in the sense that low-def pictures cancels out bad skin.

The obvious question is “but maybe the same thing that drove people to need anticholinergics increases the likelihood of dementia?”  This study has a much better retort for that than most, which is that anticholinergics were prescribed for a variety of causes, and it’s unlikely they all correlate with dementia.  I find that explanation extremely satisfying, except that they only evaluated the drugs as a single unit.  Antidepressants make of over 60% of the total SDDs taken.  The next most common is antihistamines at 17%.  But since more than 60% of the population took at least one SDD, it seems likely that those were taken intermittently, as opposed to the constant drip of antidepressants.  This leaves open the possibility that the entirety of the effect they attributed to anticholinergics was in fact caused by tricyclic antidepressants alone- and that the real culprit was depression.  The obvious controls were to evaluate the anticholinergics separately, and to compare rates of dementia among TCA treated patients with those treated with other antidepressants.

The subtler version of this question is “what if anticholinergics prolong life, giving you more time to develop dementia?”  I don’t see anything where they checked for that either way.  They did ask for people’s perception of their own health, and that was negative correlated with TSDD, but if TSDD is correlated with depression it’s hard to know how to interpret that.

For all those criticisms, this is an amazingly strong result for a medical study**.   No one study can prove anything (even if i think they had the data to do more than they did).  It definitely merits further investigation (ideally some with animal models, so we can do the causality experiments that would be super unethical in humans), and maybe even behavior change in the meantime, although a lot of the drugs studied are already obsolete or second line.  Plus it another piece of data that will help us figure out how to fight dementia, and that makes me really hopeful.

*Read: here’s what I learned.

**Yes, this should worry you.

Loratadine for Allergies?

The Decision Tree casually describes loratadine (brand name: Claritin) as barely better than placebo for treating allergies.  This is news to me because Claritin was absolutely critical to me graduating middle school.  If I forgot to take it in the morning my mom had to drop it off at school by lunch.  Without it I slept 16 hours a day,* woken up only by hives that itched so intensely they burned.  This isn’t actually relevant to me now because my allergies were taken care of my unprocessed honey and moving, but I couldn’t believe something once so important was essentially a sugar pill. So I investigated.

First stop, Wikipedia, which definitely backed my claim that Claritin treated sneezing, runny nose, itchy or burning eyes, hives, and other skin allergies.  But of 19 citations, 5 were unavailable to me (either they were books or in languages I don’t read), 13 were on topics other than clinical efficacy (e.g. side effects or mechanism), and 1 had a sample size of 192 and was a comparison against another anti-histamine, with no placebo or no-treatment group.

So I checked google scholar, where I found numerous minuscule studies (n = 14, 7 treatment groups) in which loratadine was better than placebo but worse than other drugs in the same class.**  If that’s true, why did loratadine get so much more attention?  I looked up the other drugs, and it turns out that some of them (cetirizine/Zyrtec) had similar efficacy but came out later, and went over the counter later as well.  Others (Terfenadine/Seldane) had much uglier side effect profiles (e.g. cardiac arrythmia if you eat a grapefruit).  So Claritin’s advantage seems to be being the first drug to market that treated the problem with minimal side effects.  I also wonder if Decision Tree‘s author (Thomas Goetz) was looking at a particular symptom set?  For example, loratadine appears to do well as a treatment for hives but there are better options for hay fever.

Some people suggest that having multiple drugs with similar response rates in the same class on the market is some sort of failure.  They are wrong and they should feel wrong.  First, these drugs were developed in parallel by different companies. While all the ones we heard of worked out, very few chemicals that pharma companies research become prescribable drugs, and they can’t predict which ones will do so ahead of time.  What if McNeil stopped researching Zyrtec because Bayer was researching Claritin, and Claritin made you grow arms out of your face?  We’d have lost years of allergy relief.  Second, the fact that they had similar average efficacy and side effects doesn’t mean they have the same effect in every person.  People are squishy and they don’t make sense, and differing reactions to drugs is one of the milder ways this manifests.

*No, fatigue is not a normal symptom of allergies, but I got it most springs and it went away with anti-histamines, which is good enough for a field diagnosis of allergies.

**I also found a lot of studies detailing the effects of loratadine in conjunction with another drug, mostly montelukast, and abstracts that reported loratadine’s efficacy relative to older antihistamines but without absolute numbers.

Review: The Decision Tree (by Thomas Goetz)

My trail of discovery to The Decision Tree was as follows:

  • Discover Iodine’s in-browser medical translator, become fan for life.
  • Watch Iodine CEO’s (Thomas Goetz) TED talk on the problems with how medical information is currently presented, and his solution.  Become very impressed.
  • Discover Goetz has a whole book on this stuff.  Order from library.

This was maybe not have been the best order to do it in.  Decision Tree is really, really good, but it lacks the specificity of the TED talk or Iodine’s recent work.  If I’d read it first, the other work (which was produced later)  would have been fulfilling the promise of the book.  But reading it last, I kept waiting for the other shoe to drop.  It is an amazing launching pad, but I went in expecting to see what had landed.

That may not even be fair.  Goetz points to a lot of specific things, like the Quantified Self movement, PatientsLikeMe.com, and actual research on how Dr. Internet affects people.  It’s just that none of these so singularly improve the signal to noise ratio the way Goetz’s work on presentation of test metrics did.  I guess what I’m saying is you should watch that TED talk.

Now that I’m over the fact that Decision Tree is not a 250 page TED talk, I can appreciate it for what it is, which is a reasonable 101 text on the concept of individuals monitoring and improving their own health.  It doesn’t give many specifics for either of those because the answers are so specific and so personal, but it does leave the reader better prepared to evaluate possible solutions they find.   That’s actually pretty hard to do, and really useful.  I could also see it as useful for medical professionals who are on the fence about patient-driven care.*  It is extremely helpful in explaining why over-testing is so dangerous, while respecting individuals’ right to data.  And if you’re not reading it during or immediately after a painful, stressful medical procedure, it’s actually a pretty light read. So if this book looks interesting to you I’d recommend it.

*Goetz is unreservedly pro- patient led decision making and research.  I am too, until I remember a lot of the anti-vaxxers have put an enormous amount of research into their idiotic, dangerous, anti-social position.  I don’t know how to preserve the rights of me + my friends to know our own data and correct our doctors’ mistakes while preserving the rights of children to not die of entirely preventable diseases.

The Kitten Pain Scale

I very briefly flirted with Quantified Self and then jumped off the bandwagon because it was making my personal signal:noise ratio worse.  But my neuroendodontist* has given me several drugs, and he wants to know how they work.  Allow me to give you a brief list of things that make measuring this difficult

  • Treatments are all on varying schedules- some daily, some daily with a build up in blood stream leading to cumulative effects, some as needed to treat acute pain, some on my own schedule but hopefully having longer running effects.  Some are topical and some are systemic.
  • I have several home treatments like tea and castor oil.  I’m not going to not take them in order to get more accurate assessments of the drugs, both because ow and because pain begets pain.
  • Taking treatments as needed + regression to the mean = overestimate of efficacy.
  • Pain is affected by a lot of non drug things: sleep, stress, temperature, how ambitious I got with food, amount of talking, number of times cat stepped on my face in the night, etc.
  • We are hoping some of these drugs will work by disrupting negative feedback loops (e.g. pain -> muscle tension -> pain), which means the effect could last days past when I take in.  In the particular case of doxepin it might have semi-permanent effects.
  • Or I could develop a tolerance to a drug and my response to a particular drug will attenuate.  That is in fact one reason I was given so many choices as to medication: to let me rotate them.
  • We have no idea how these drugs will interact with each other in me.  We barely have an idea how the interact in people in general.
  • If I believe something will help my pain will lessen as soon as I take it, long before it could actually be effective.  Not because I’m irrational, but because my brain reinforces the self-care with endorphins, which lessen pain.
  • At the same time, having more pain than I expected to feels worse than the exact same pain level if it was anticipated.
  • Side effects: also a thing.

“I think I feel better when I take this one” was not going to cut it.

Then there was the question of how to measure pain.  Ignoring the inherent subjectivity of pain, neuralgia is a weird beast.  I already hate the 1-10 pain scale because pain has threshold effects and is exponential.  I could create a single pain number at the end of the day, but my pain is not constant: it spikes and recedes, sometimes for reasons, sometimes not.  What I would ideally like to track is area under the curve of pain**, but that requires polling, which would create horrible observer effects.  If I ask myself if I’m in pain every 15 minutes, I will increase my total pain level.  I could poll less often, but the spikes are random and short enough that this was not going to be accurate enough to evaluate the treatments.  I could count pain spikes, but that ignores duration.  Determining duration requires polling, so we’re back where we started.  I could deliberately poke a sore spot and see how bad the resulting pain is, but

  1. Ow
  2. A treatment that doesn’t affect sensitivity but does keep me from spontaneously feeling pain because the nerve is bored is a success.  If we wanted me to be numb we would do that.

It’s just really hard to measure something when your goal is for it to be unnoticeable, and measuring it creates it.

So I came at it from the other side.  What happens when pain is unnoticeable?  I enjoy life more and I get more things done.  Could I measure that?  Probably.  They have the bonus of being what I actually care about- if something left me technically in pain but it no longer affected my ability to enjoy or accomplish things, that would be a huge success.  If something took away the pain but left me miserable or asleep, it is not solving my actual problem.**

So one metric is “how much I get done in a day”.  Initially this will be the first number between 1 and 10 that I think of when I ask the question at the end of the day, but I’m hoping to develop a more rigorous metric later.  You’d think enjoyment of life couldn’t ever be rigorously measured, since it’s so heavily influenced by what is available to me in a given day, but I say that brave men can make it so.  And so I introduce to you: the kitten pain scale.  Kitten videos vary a little in quality, but I think my enjoyment of any single video reflects my internal state more than it does the video. Three times a day (shortly after waking up, shortly before screen bed time, and sometime mid-day that can vary with my schedule but must be selected ahead of time to avoid biasing the data), I will watch a cute kittens video and record how much I enjoy it.  The less pain I am in the more I should enjoy the video.  This will give me a (relatively) standardized measure of pain without risking inducing it.

This is still not what you would call a rigorous study.  An individual choosing what to take among known options never will be.  But I seriously think the kitten pain scale could be a contender to replace the stupid frowny faces.  My first draft is available here.  Right now it’s set to measure over the course of a day, because that’s the scale I expect from these meds, but you can add bonus measurements at set times after taking meds if you like.

Possible additions: cups of tea drunk in day.  Right now that seems like too much work to measure, but when tea is available it’s a pretty good indicator of how much pain I’m in.

*I am still angry that I know what that is, much refer to one using possessive case.  But given that, I am extremely grateful I live within biking distance of a world class research facility in the discipline.  Even if the physical facility could be a case study in how economic insulation leads to bad user experience.

**This is why none of my treatment options are opioids.  Strong ones technically reduce pain, but they also leave me miserable.  The fact that some people take them for fun is all the proof of human variability I could ever need.

Adventures in Dentistry and Neurology

I forget if I mentioned it, but I had nerve damage from the first dental surgery, way back in June.  Everything else healed up more or less all right, but that one kept hurting.  Actually it felt like two damages- one that was healing, albeit slowly, and one that was staying static or getting worse.  The prospect of living with that pain for the rest of my life was really daunting.  Medical marijuana, which had been so helpful at first, was having more side effects with fewer desirable effects every day.  It eventually became clear my surgeon had no idea what was going on or how to fix it so I went to a neuroendodontist, a subspeciality I really wish I wasn’t already familiar with.

toenailectomy looks awful but feels like nothing at all.  A neuroendodontal exam is the exact opposite.  It looks like some guy very gingerly touching around your mouth, but he is not only deliberately provoking pain, he needs you to pay attention to the pain and report on in it excruciating detail, while you remind yourself that inaccurate reporting leads to inaccurate diagnoses.

For all that pain, I actually got very good news.  Even though it feels like I have two distinct damages, it’s actually only one, and it is healing.  Nothing is guaranteed in neurology but existing data is consistent with this eventually healing itself.  And in the meantime, he gave me new and different medicines.  We’ll see what the side effects are, but at the very least I have options to rotate through.

Adventures in Podiatry and Neurology

WARNING: THIS ONE IS GRAPHIC EVEN BY MY STANDARDS.  NEEDLES, PAIN, AND TOENAILS.

Recently I learned toenails aren’t supposed to be under the skin of your foot and hurt constantly; this is an ingrown toenail and it’s a solvable problem.  By “recently” I mean a year and a half ago, but a little pain when I flexed my toes in a shoe did not seem as important as the pain in my mouth or my inability to digest food, so I only got around to seeing a podiatrist now.  If you develop an ingrown toenail there are home treatments to coax it better, but if you’ve always had it the cure is a little more drastic: they cut off the bit of the nail that has grown under the skin and cauterize the nail bed so it never grows back. If you are curious, here’s a video of the actual medical procedure:

The worst part is the lidocaine injection. There’s a topical anesthetic, but they root the extremely thin nail around under your skin in order to find the nerves and inject directly over them. The podiatrist will describe it as slightly painful, but they are lying, and it will make you doubt them when they promise the rest of the procedure is painless. That part turned out to be true: with enough lidocaine you genuinely can’t feel them slip the scissors/pliers under the nail bed, or the burny stuff*, unless you are a freak who processes -cain very quickly, in which case they will give you more and it will stop hurting.  But the anesthetic injection was pretty brutal.

That is not actually the interesting part. In between the lidocaine and the scissors/pliers, they test your numbness with what looked like a large blunt toothpick. My podiatrist, which more flourish then was strictly necessary, brought it down from a great height onto my toe.

I screamed.

Then I realized it didn’t hurt at all. My brain had combined the memory of the painful needles and the visual information about incoming sensation and preemptively sent a scream response before it noticed I couldn’t feel anything. I never had quite that strong a reaction again, but there was an extremely weird dissonance as I watched something I knew should hurt, yet got only vague reports of pressure from the area.

This works in reverse too.  Phantom limb syndrome is a condition in which people missing a limb (even one they never had) experience excruciating pain where their brain thinks that limb should be.  One of the only effective treatments is mirror therapy, where a mirror is used to simulate the appearance of the missing limb, and somehow the brain goes “oh, I guess it’s fine.”  This clip from House is not quite as accurate as the matrixectomy one (mirror therapy rarely involves kidnapping), but the science is sound.

The lesson here is that even something that feels incredibly simple and real, like pain, is in fact an artifact of post-processing on several different inputs.

*Dr. Internet says phenol but I could have sworn it started with an M. In my defense, he gave me the proper name after the needles bit and I was fuzzy.

Depression as a false negative

Slate Star Codex points out that rates of suicide and depression are weirdly terrible metrics for how good a society is.  I wonder if some of that is a definitional effect.  Depression is more or less defined as occurring for no reason.  If you have a reason for sleeping poorly and feeling unable to do everyday things (e.g. fibromyalgia), you’re diagnosed with that instead.  As society gets worse, people who were chemically destined to be depressed are given reasons to be sad, and so stop contributing to the depression statistics.

This is related to but slightly distinct from the idea that depressed people are less likely to commit suicide when conditions are objectively miserable than when they are good because bad conditions leave room for hope in a way good conditions don’t.  That is about individuals specific reaction to their depression.  My hypothesis is about how the number “% depressed” is measured.

Of course, my suggestion doesn’t account for increased suicide rate.  The expectations hypothesis does account for that.  One other factor I think may be in the mix is coping mechanisms.  Before the AIDS cocktail, someone noticed AIDS patients actually got better when co-infected with another virus.  The reason turned out to be interferon, an intercellular signal to ramp up anti-viral defenses.  HIV didn’t trigger it, or didn’t trigger it enough, but when another virus did the resulting interferon protected from HIV as well as the original virus.  Maybe external bad events trigger coping mechanisms in a way depression doesn’t, and they incidentally fight depression.  This could be true even if  “coping mechanism” just means disassociating until things get better.

The FDA strikes a blow for HIV virus everywhere

Part of the FDAs job is making sure the pill you swallow is the pill you think you are swallowing.  It would be very bad if you thought you were taking penicillin and instead took sugar, or Prozac.  This is a government task even a libertarian could love.

But.

Sometimes enforcing a definition is harder than penicillin vs. Prozac.  For example, condoms.  The FDA has definition of a condom, and if your product doesn’t meet it, you can’t sell it, or at least you can’t call it a condom.  Some of this is good.  You want condoms tested for structural integrity, and you don’t want flat pieces of latex being sold as sexual health devices.

But.

The FDAs definition of a condom does not appear to account for human variation.  Condoms must be at least 160 mm in length.   Condoms are not allowed to be wider than 54 mm (-> diameter of 170 mm).   I’m not sure exactly how closely condom dimensions need to match penis dimensions, but that is longer than ~70% of penises and (using a condom width to girth conversion I don’t understand) narrow than ~18% of penises (source).  Yes, condoms are stretchy and will fit over your leg if you try- but they’re more prone to break, and your leg isn’t kept rigid via blood flow.  A too loose condom will just fall off, but a too long condom bunches, which also causes breaks and reduces blood flow.

I understand why we had to put up with this 100 years ago, but condoms used to be custom made, and some genuises in England are bringing that back.  But thanks to our FDA and its scrupulous definition of condoms, it is illegal to sell them in the states.  And it would definitely be illegal to use the cheap package forwarding service they link to, unless you are an EU citizen.

What doctors can learn from day care workers

So if being fat is bad for people, than doctors should tell them not to be fat, right?  Or at least tell them to eat vegetables and hit the elliptical, right?

Well, maybe.  But sometime around age two humans realize that they are independent beings who do not physically have to do everything an authority tells them to do.  Unfortunately, most doctors’ patients are over the age of two, and those that aren’t have their own issues.

vomiting baby
They’re gross

Telling people to do things they already know they need to do has mixed results.  Scott Alexander suggests alcoholism could be decreased by as much as 13% if doctors would spend five minutes telling alcoholics it was bad for them.  What this doesn’t capture is how lectures change the doctor-patient relationship.  It is very difficult to give a non-judgmental lecture when your billing model gives you 10 minutes per patient.  Patients might avoid or delay visits for problems- alcohol related or not- in order to avoid the lecture.  This is a pretty big issue with overweight patients, and apparently without upside: patients lectured by their doctors are more likely to attempt weight loss but no more likely to achieve it.

In this TED talk, Thomas Goetz talks about a study of dental patients (no cite, unfortunately) that found that scaring them had no effect, but patients’ belief in their own ability to floss had a large one.  It’s impossible for me to separate my personal experience from this data.  Multiple dentists and hygienists told me my pain was my fault for terrible oral hygiene, and if I brushed and flossed it would go away.  This turned out to be untrue on a couple of levels.  The pain was caused by structural damage and internal infection, which may have been made incrementally worse by oral plaques but wasn’t caused by it.  And I was actually brushing pretty regularly, it just wasn’t do anything.  Then I started treating a completely unrelated digestive problem, and suddenly my teeth were cleaner.  I didn’t even tell my dentist anything had changed, she asked spontaneously.  So I guess, yeah, patients belief in their own ability to effect change matters, and if they don’t believe it, maybe consider that they’re correct and investigate why.

But let’s go one step farther.  Crum and Langer did an interesting experiment on two groups of hotel maids.  Both were told exercise is good for you.  One was given additional information about the intensity level of the work they did all day, and told just by going to work they were exceeding the surgeon general’s RDA of exercise.  Four weeks later, the informed group was slightly thinner (they even checked body fat %.  I am so pleased) and had lower blood pressure .  Not astoundingly lower(10 points on diastolic BP), but it was only four weeks, and a pamphlet is even less work than a doctor lecture.

This suggests that one of the more helpful things public health officials can do is reinforce the good things people are already doing.  You did a stretch?  Hurray for you.  Check parking lot twice before accepting a far out spot?  Still counts.  It would not shock me if part of the health improvements attributed to standing desks turned out to be simply a halo effect of feeling like you made a healthy choice.  Which coincidentally is how you turn a two year old into a civilized human being.

Multifactor analysis

Elodie Under Glass (no relation) has a guest post up at Captain Awkward about dealing with family members with disabilities/when you have a disability. You should read the whole thing and the comments, because it lives up to CA’s high standards, but here’s the thread I want to talk about: one LW owns a house that is unable to accommodate her disabled father comfortably. It’s three stories, no elevator, no bathroom on the ground floor, and she doesn’t have a good bed to offer him. He has MS. Some people viewed her choice of house as a failure to accommodate her father properly. Others pointed out that it’s quite possible she couldn’t afford a disability-friendly house, or the renovations necessary to make it so. One commenter went so far as to say:

The people replying to you rightfully take issue with your snide implication that the LW should have just bought a smaller house- which has no correlation to price of house.

Which is clearly bullshit.  It is true that size is not the only determinant of the price of a house.  It may even be true that in a particular set of houses (one that sampled over a wide geography, time, condition, and set of amenities), price and size are not particularly predictive of each other.  But if you hold those others factors constant, size is strongly positively correlated with price. So while I do believe people should get off the LW’s case about buying a house that couldn’t accommodate her father, for lots of reasons, I think saying there’s nothing she could have done to accommodate him is wrong.

This is actually a pretty good metaphor for weight, if houses actively fucked with you to maintain their price within a set range. Build an addition to your house?  Black mold. Hog rendering plant built upwind?  Enjoy your newly refinished basement.  Eat less fewer calories?  Enjoy catching twice as many colds this year.  Eat more calories?  Never stop fidgeting.  Weight is not beyond our influence, but neither is it completely in our control.