Pills and Tablets
© SebastianRushworth.com
Last year I spent a couple of months working as a physician in a geriatric hospital, i.e. a hospital that specializes in taking care of elderly people. One thing that struck me particularly was the large number of medications each patient was on. I don't think it would be much of an exaggeration to say that the average patient had ten or more medications that they were taking on a daily basis.

This condition, of taking many different medications, is known as polypharmacy. It is a big problem. And a growing problem. In the US, 31% of older adults were taking 5 or more medications per year in 2006. Five years later, that number had increased to 36% . In a Swedish population study, 17% of adults were taking five or more drugs per day in 2006. This had increased to 19% in 2014. Among people over 90 years old, 80% were taking five or more drugs per day.

But, you might say, surely they needed those drugs? After all, they were prescribed the drugs by a doctor for a reason. And older people generally have more underlying conditions, and therefore need more drugs.

All of this is true, to a point, but the reality is more complicated. The first problem concerns how scientific studies of drugs are carried out. Most of the time, studies look at a group that has a single condition, for example heart disease, and exclude people with co-morbidities like lung disease or diabetes. Additionally, studies usually exclude people over a certain age.

These two things together mean that there is very little data on what happens when you give medications to people with several underlying conditions, and there is also very little data on what happens when you give medications to the elderly. So, on balance, we often don't know whether we're helping or harming the patient. We extrapolate from data that really applies to a different patient - a younger, healthier patient.

The second problem is that studies on medications are generally done in patients with functioning kidneys and a functioning liver. The kidneys and liver are the organs that metabolize and excrete drugs from the body. If people have kidney disease or liver disease, they will not be able to metabolize and excrete medications as effectively as healthy people, which means there is an increased risk of getting toxic levels of the medication in the body. And while a doctor might check a patient's kidney function when putting them on a medication, there is often no follow-up to ensure that the medication is still appropriate as their kidney function inevitably declines with age.

The third problem is that patients, and especially elderly patients, see lots of different doctors. And one doctor usually doesn't want to stop a medication that another doctor has started. The reason for this is that the second doctor generally doesn't have full information on why the drug was started, and so doesn't want to take the patient off a medication that they might "need". This leads to patients accumulating medications over their years of life.

When I look in patient charts, it's not uncommon for them to be taking medications that they've been taking for ten or more years. Oftentimes, these drugs have since fallen out of favor because new data has come to light showing that the harms outweigh the benefits. Often, the patients themselves have no idea why they're taking a drug, they just keep taking it obediently because at some point a doctor told them they should be on it.

One example of this phenomenon is baby aspirin (low dose aspirin). A lot of older adults without any underlying heart conditions take this drug to prevent heart attacks, even though evidence that has come out in recent years has shown that the harms clearly outweigh the benefits. As a doctor, it's often impossible to find the original chart explaining why the patient was put on a drug, especially if they were put on it several years ago. So you keep them on it, because that's easier than taking the "risk" of taking them off the drug.

The fourth problem is that there are very few studies looking at what happens when you take combinations of drugs, even though that is commonly the reality. There are plenty of dangerous synergies that happen when people are taking multiple drugs. I'll give two examples:

A lot of elderly people have atrial fibrillation. This is a disorder of the heart that is connected with an increased risk of stroke. You can decrease the risk by taking a blood thinner, like warfarin (a.k.a. coumadin). The problem is that lots of people with atrial fibrillation also have anxiety issues, and so they are put on sertraline (a.k.a zoloft). On it's own, sertraline has a very mild anti-coagulant effect, but in combination with other drugs there can be a significantly increased risk of bleeding. Say the patient also has some joint pain, and so they are also prescribed paracetamol (a.k.a. tylenol, acetaminophen). On it's own, paracetamol has no effect on bleeding risk, but in combination with warfarin there is a significantly increased risk of bleeding.

Each of these three drugs on its own might be ok, but in combination, there is a heavily increased risk of bleeding. This can take the form of an intracranial hemorrhage, or a bleeding stomach ulcer, which can both be rapidly fatal. And in reality, drugs are used in combination more often than not. There are few if any studies showing at which point different combinations become more harmful than helpful.

Here's another example: Many elderly people take one or more blood pressure medications. These decrease the body's ability to regulate the flow of blood to the brain, and so slightly increase the risk of fainting and falling. Many elderly people also have trouble sleeping, for which they are often prescribed bensodiazepines. Benzos make you a bit groggy, and thereby also increase your risk of falling, especially if you have to get up to pee in the middle of the night (as many older people do). On top of this, many older people have pain issues for which they are prescribed opioids, which also increase the risk of falling. Each of these drugs on their own might not be too bad, but taken together they heavily increase the risk of falls, which are a leading cause of death and disability in the elderly.

Not only is there not a lot of research on these dangerous synergies, but doctors are often unaware of them. I am willing to bet that most doctors don't know that it's a bad idea to combine warfarin and paracetamol, since that is a combination I see frequently in patients. Obviously, the more drugs you're taking, the bigger the risk of dangerous interactions. Like I wrote above, it's not uncommon for elderly people to be taking ten drugs or more. I don't even want to think about how many dangerous interactions they are being exposed to with those kinds of cocktails.

There is another problem, which contributes to polypharmacy, and that is that many drugs are hard to stop once you start. Opioids (like oxycontin) and bensodiazepines (like valium) are the most obvious examples - both are highly addictive. Other drugs, like proton pump inhibitors (for example omeprazole), that are used to treat gastritis and acid reflux, are hard to stop because the body tries to compensate for the fact that its ability to produce stomach acid is blocked, so the moment you take the pedal off the brakes, the body overcompensates and produces lots of stomach acid, and the patient gets more symptoms than ever. Rather than wait these symptoms out, many people go straight back on the proton pump inhibitor, and stay on it for life.

The problem of polypharmacy has been getting increasing attention in recent years, especially among physicians specializing in treating the elderly population. The proposed solution to the problem is deprescribing, which is, simply put, the opposite of prescribing. Instead of adding to the list of medications a patient takes, you remove drugs.

A meta-analysis was published in the British Journal of Clinical Pharmacology in 2016 seeking to quantify the extent to which deprescribing can decrease mortality. It included 10 randomized controlled trials with a total of 3,151 patients. The average age of the patients was 74 years. The average follow-up period was 10 years, and the authors reported no conflicts of interest.

One oddity about this review was that it included both interventions in which patients actually had the number of drugs they were taking decreased (8 trials with a total of 1,906 patients), and also interventions where health care providers were educated about the benefits of deprescribing, and how to go about it (2 trials with a total of 1,245 patients). To me, the first is a deprescribing intervention, and the second is an educational intervention, and it's very weird to put both together under the umbrella of a deprescribing intervention.

Luckily, the two types of intervention were analyzed separately in terms of effect on mortality, so we have that information. But they were lumped together in all other analyses (for example when looking at quality of life, or when determining if the effect size varies between age groups), so unfortunately, we don't have access to that information.

Why is this a problem? Because it's pretty obvious that an indirect intervention where you're talking to doctors about deprescribing is going to have a much weaker effect than a direct intervention where you're actually deprescribing drugs from patients. So any analysis which lumps together education about deprescribing with actual deprescribing interventions is going to show a weaker effect than is actually there. I've tried to understand what the investigators were thinking, but can't come up with a good answer. The only thing I can come up with that makes any sense was that they wanted the review to fail to find benefit. But more likely they just weren't thinking things through properly.

Anyway, let's look at the results.

When looking at the overall results for the ten randomized studies, there was an 18% reduction in the odds of dying. However the difference was not statistically significant. But, like I said, this overall result includes the studies which were only educating health care providers about deprescribing.

So let's look instead at what the results were when only studies that actually implemented deprescribing in patients directly are included. Overall, in that situation, there was a 38% reduction in the odds of dying over the ten year follow-up period! That difference was highly statistically significant.

For the sake of completeness, we will look also at what the results are if we just take the studies that educated health care providers. In that scenario, there was a 20% increase in the odds of dying. However, the increase was not statistically significant. Basically, teaching stuff to doctors either does nothing at all, or makes them more likely to kill their patients!

So what can we conclude from all this?

Deprescribing has a big impact on mortality. A 38% reduction in the odds of death is pretty huge. Especially since all you have to do to get that benefit is take fewer pills. For comparison, statins decrease the odds of death by 14% according to a meta-analysis by the Cochrane collaboration, and statins are often sold as a hugely life saving intervention. Well, the effect of deprescribing is almost three times as big.

On the other hand, just teaching doctors about deprescribing does not seem to work. To me, this makes perfect sense. Most doctors are working under significant time constraints, and so do not have the time to go through patients medication lists in detail, to decide which medications can safely be removed. Additionally, doctors face an unrelenting barrage of lobbying from pharmaceutical companies to prescribe more medications, not less.

So, the best thing you can do as an individual patient is to get educated about your medications, and to avoid being put on too many medications in the first place. Always ask your doctor why you are being put on a certain medication, what the size of the benefit that can be expected is, what potential harms there are, and how long you're supposed to take the drug. If you're not satisfied with the answers you get, it is perfectly within your rights to say no. It is also perfectly within your rights to say that you want to go home and do some research before deciding whether to take the drug or not. This is especially important when it comes to drugs that you are supposed to take continuously for the rest of your life, and drugs that are hard to stop once your start.

The simple fact is that a drug might well be beneficial when given on its own, but harmful when combined with several other drugs. I have a hypothesis that, on balance, when the number of drugs you take goes beyond five, the negative effects of each additional drug start to outweigh the positive effects, and that the negative effects increase exponentially with each additional drug, as the risk of interactions and dangerous synergies increases.

It is well known among doctors and nurses who care for patients in the last days and weeks of life that their symptoms often improve when the decision is taken to stop giving life prolonging treatment, and they are taken off most of their meds.

Here's an idea for a randomized controlled trial that I would love to see done: Take a group of elderly patients who are all on 10+ drugs. Rank the drugs based on how important you think they are for each patient. In the intervention group, get rid of all drugs except the five that you have rated as most important (obviously you would have to replace the drugs you've gotten rid of with placebos, in order to maintain blinding). The control group continues on as before. See how many are alive in each group five years later.

My guess is that the intervention group would live significantly longer, with a significantly higher quality of life, than the control group. No pharmaceutical company would ever fund that study, for obvious reasons. It would need to be funded by the government, or by a philanthropist. So if you have several million dollars lying around and want to make a huge contribution to human health... You're welcome for the idea.

You might also enjoy my article about whether statins save lives or my article about how aggressively high blood pressure should be managed.

I am rolling out a ton of new science-backed content over the coming months, including:

- Analyses of the benefits and risks of all common supplements and medications
- The keys to a longer, healthier life (possibly quite different from what you may have heard)
- A long-term follow-up of the health consequences of the covid pandemic and global lockdown.