A Country Doctor Writes:

I happened to read about the pharmacodynamics of parenteral versus oral furosemide when I came across a unique interaction between this commonest of diuretics and risperidone: Elderly dementia patients on risperidone have twice their expected mortality if also given furosemide. I knew that all atypical antipsychotics can double mortality in elderly dementia patients, but was unaware of the additional risperidone-furosemide risk. Epocrates only has a nonspecific warning to monitor blood pressure when prescribing both drugs.

This is only today’s example of an interaction I didn’t have at my fingertips. I very often check Epocrates on my iPhone for interactions before prescribing, because – quite frankly – my EMR always gives me an entire screen of fine print idiotic kindergarten warnings nobody ever has time to read in a real clinical situation. (In my case provided by the otherwise decent makers of UpToDate.)

I keep coming back in my thoughts to, and blogging about, drug interactions. And every time I run into one that surprised me or caused harm, I think of the inherent, exponential risks of polypharmacy and the virtues of oligopharmacy.

One conclusion I have come to is that too often the benefit of our prescribed medication is actually too small to justify the drug. The way drugs are approved today is pretty much that they have to bring a 20% or so advantage over placebo for a certain outcome. Other than the drug versus placebo, all other factors are ignored or “controlled for”, which is easier said than done.

But this whole premise seems wrong to me: If pill A is 20% better than placebo at lowering blood pressure, but salt restriction, weight loss, exercise and stress reduction are twice as powerful as pill A, why are we so stuck on prescribing pill A? If a Mediterranean diet lowers cardiovascular risk as much as atorvastatin, why isn’t that a blockbuster/no-brainer intervention?

The health of our nation is not great, in spite of all the pills at our disposal. And the more pills we prescribe, the more we risk interactions: antidepressants and cholesterol pills with blood thinners, gout medicines with cholesterol pills, mood stabilizers with cardiac medications and on and on and on.

May we all take a step back and look at the big picture of what we are doing and where we are heading.

Donald W Light from the Harvard Center of Ethics wrote in 2014:

Few people know that new prescription drugs have a 1 in 5 chance of causing serious reactions after they have been approved. That is why expert physicians recommend not taking new drugs for at least five years unless patients have first tried better-established options, and have the need to do so.

Few know that systematic reviews of hospital charts found that even properly prescribed drugs (aside from misprescribing, overdosing, or self-prescribing) cause about 1.9 million hospitalizations a year. Another 840,000 hospitalized patients are given drugs that cause serious adverse reactions for a total of 2.74 million serious adverse drug reactions. About 128,000 people die from drugs prescribed to them. This makes prescription drugs a major health risk, ranking 4th with stroke as a leading cause of death. The European Commission estimates that adverse reactions from prescription drugs cause 200,000 deaths; so together, about 328,000 patients in the U.S. and Europe die from prescription drugs each year. The FDA does not acknowledge these facts and instead gathers a small fraction of the cases.

There are obviously more recent statistics out there, but this piece struck me because it was published in a forum about ethics. Think about that for a moment: We are subjecting our patients to known and unknown risks of harm with every prescription we issue.

Once upon a time there were jokes circulating about putting Prozac (fluoxetine) in the drinking water.

The idea was that the modern antidepressants were indicated for most people living in today’s society, and that these drugs were completely safe to use.

Now, these same drugs have so many warnings that they have become increasingly difficult to use in treating the common maladies of depression and anxiety.

The earliest scuttlebutt about Prozac causing suicidal thinking never did go away completely. We were assured that the suicide rate itself did not go up, only thinking about it. Some of that was explained by powerful antidepressants potentially unmasking bipolar illness and causing manic episodes. But a few years later, Paxil (paroxetine) was reported to cause suicidality in adolescents.

Just in the past few months, I was reminded several times about the intricacies of prescribing SSRIs.

A middle aged man on the blood thinner warfarin suddenly developed nosebleeds when his INR shot above the therapeutic range. His psychiatrist had doubled his Prozac two weeks earlier.

An anxious and depressed elderly woman wanted to try an antidepressant, and I gave her a low dose of Lexapro (escitalopram). She told me the following week that she felt tired and listless. I ordered some routine blood taste and her sodium level came back precipitously low. We stopped her Lexapro and she felt better again within a week.

A tall, thin elderly man with Parkinson’s disease and depression wanted to try something for his mood. I searched the literature and prescribed the SSRI with the most data on use in his condition. The following week he felt lightheaded. His blood pressure and pulse were both lower than his baseline, and I ended up cutting his beta blocker in half twice before his vital signs normalized; I had never heard or seen this interaction before.

A middle aged woman with a history of pulmonary emboli had been difficult to regulate on warfarin, and during a recent hospitalization, the doctors had changed her over to the novel anticoagulant Eliquis (apixaban), which doesn’t require any monitoring. When it was time to refill her Zoloft (sertraline), my computer flagged me with a bold, red warning that apixaban and sertraline are a dangerous combination resulting in increased risk for bleeding.

The list of warnings for this class of drug goes on, including effects on pregnancy, heart arrhythmias, seizures, glaucoma, liver disease and diabetes.

Given that their effectiveness is reported to be only marginally better than placebo, I have become increasingly more cautious about prescribing them.

Just this year, though, in Molecular Psychiatry, a new analysis of old data suggests that previous studies had used an ineffective rating system for depression, and that SSRIs are more powerful than we thought.

I remain skeptical. Once bitten, twice shy.

I, like most primary care physicians, have many patients on chronic “blood thinners”. Warfarin, essentially the same chemical as rat poison, is the most common drug we use, and it can be difficult to manage. Because its effects are counteracted by vitamin K, simple dietary changes like eating fewer or more greens can change the effects of warfarin. There are also many drug interactions to keep in mind.

Because of these interactions we never assume that patients can stay on the same dose of warfarin indefinitely. Some people’s numbers vary enough to warrant testing a few times per week. Our clinic’s minimum standard is that even stable patients get a blood test once a month to monitor the medication’s effect.

We measure the “prothrombin time”, or how many seconds the blood takes to clot, and “INR”, International Normalized Ratio, which is, roughly speaking, how long a patient’s blood takes to clot compared to an untreated person’s blood. We typically strive for an INR of 2 to 3, which is 2 or 3 times the normal, untreated, clotting time.

Antibiotics are among the most common drugs that interact with warfarin. Only a handful of antibiotics are safe in this regard. Penicillins, cephalosporins and nitofurantoin are choices we don’t worry about. Azithromycin sometimes interferes, and common urinary antibiotics like sulfa and ciprofloxacin almost always interfere to some degree.

Florence Fitch, an elderly patient of mine with atrial fibrillation, had a urinary tract infection and had seven days of ciprofloxacin prescribed by another doctor. She ended up in the hospital with an intestinal hemorrhage and needed two units of blood.

Today I saw Gwen Hubert. She has high cholesterol and atrial fibrillation. She must have been on warfarin and simvastatin for ten years. Her numbers were always quite stable. When I saw Gwen the last time, she had complained of fairly significant muscle aches. Her cholesterol was perfect and her creatine phosphokinase (CPK) test didn’t show any sign of muscle damage. Still, even when there is no damage, people on simvastatin as well as all the other statins can have bothersome muscle aches.

At our last visit, Gwen and I agreed that she would not take the simvastatin for three to four weeks. If there was no difference, she was to start her cholesterol pill again and see me a month or so later.

She had had an INR drawn the other day and her level was high enough that we had called her to tell her to skip a day of warfarin and start taking a lower dose after that.

Gwen was concerned when I saw her.

“I’ve never had a high INR before. Do you think starting the simvastatin again caused a problem with my warfarin?”

I looked at her flowsheet. About the time we stopped her simvastatin her INR had dropped. I hadn’t thought much of it and just increased her warfarin dose a little. The following week her number was higher, but still not in range, so we had her increase her dose some more. That took care of it. Then, when she started the simvastatin again, her INR went up to 4.

“I haven’t seen simvastatin do that before, but I’ll look it up.”

Our usual drug interaction website didn’t respond. The first result on my Google search was an abstract of an article from Oslo, Norway, published in 2007:

An 82-year-old white female was admitted to the hospital because of an international normalized ratio (INR) value greater than 8, which was detected at a routine follow-up visit to monitor warfarin therapy. Four weeks earlier her lipid-lowering therapy had been switched from atorvastatin 10 mg daily to simvastatin 10 mg daily. She had been treated with 2.5 mg of warfarin daily for almost 30 years due to episodes of deep venous thrombosis and lung embolism. Her INR had been stable within the treatment range (2.0-3.5) for more than 2 years before the INR increase. Upon hospitalization, she was given 5 mg of vitamin K orally. A few hours later she lost the feeling and movement of her right arm and a computed tomography scan showed major bleeding in the left cerebral hemisphere. She died the following day.

DISCUSSION: One study has shown a lack of interaction between warfarin and atorvastatin. In comparison, 3 studies have shown significant increases (10-30%) in warfarin effect and/or reductions in dose requirement after starting concomitant simvastatin treatment. The interaction mechanism between simvastatin and warfarin is not known but is possibly associated with reduced elimination of warfarin. Use of the Naranjo probability scale showed that the likelihood of warfarin-induced INR increase following the switch to simvastatin was probable.

CONCLUSIONS: Atorvastatin and simvastatin appear to differ in their potential to interact with warfarin. Clinicians should be aware of the interaction risk when starting simvastatin treatment in patients on warfarin therapy. 

In Gwen’s case, restarting a drug she had been on for over a decade could have had the same deadly effect.

I graduated near the top of my class from the second oldest university in the world, and one of the best medical schools on the planet. I have 44 years of post graduate experience in a part of the United States, where there is a severe lack of specialists and where a small cadre of primary care doctors see more and do more than in many other parts of this country.

If and when I look at my personal report card in my new Epic EMR I find that I am average or worse. What that means is that I’m not real quick in responding to routine messages and other such things.

I almost get the feeling that clinical competency in primary care doctors is so much taken for granted and so commoditized that the only thing that matters is how fast we are at getting patients out the door and clerical tasks completed.

Metadata, as I call the majority of items in my EMR report card, is the typically invisible information on a CD, for example: Recording date, copyright holder, maybe the album cover photo for displaying on your device. But imagine if this was all you got, if the piece of music wasn’t there. That’s what it feels like sometimes in primary care: The clinical work we do is invisible while the metadata blocks the view of what really matters.

Would you pick a brain surgeon, if you needed one, based on their speed in the operating room or at the computer? I suspect neither.

There are a few clinical measures also in these report cards, but they are set as if they are absolute, immutable laws. And do they really reflect the quality of our care?

Quality measures in medicine are basically substitutes for the real information we wish we had.

We measure surrogate outcomes, like hemoglobin A1c and blood pressures. But the person with perfect numbers could suffer a fall one day from hypoglycemia or hypotension and break a hip. The diabetic could get urosepsis from their Jardiance pill that causes them to excrete more glucose in their urine. And the perfect blood pressure person could develop kidney failure or a life threatening arrhythmia from spironolactone or valsartan, or lupus from their hydralazine. In studies, such things are tracked and judgements are made that, for a majority of patients, the benefit of a particular treatment outweighs the risk. But, in practice, we don’t know who they are. If our patient is almost “at target” and the next drug we add could be the one that will harm that particular patient, is it worth doing? Polypharmacy is a dirty word some of the time, except when patients have several medical problems and are supposed to meet several numeric targets.

There is nothing in my report card about whether my diagnoses are correct, whether my workups are appropriate, whether my patients trust me or even like me, and what my clinical batting average is when it comes to actual outcomes.

Medicine is full of decisions, judgement calls, that should be made together by physicians and informed patients. The word patient-centered has become meaningless, because doctors who let their patients in on clinical decisions and respect their refusal to do certain tests risk getting bad quality scores.

Mary Madore is no stranger to pain, but it seems that over our last couple of visits, I have eliminated or mitigated all her painful conditions. She didn’t make good on it, but the other day she proclaimed “I could kiss you”.

Like many other people who have had their gallbladder removed, she has had terrible yellow diarrhea and abdominal pain, not just after her surgery, but even decades before. I explained simplistically “you seem to have too much random bile in your gut” and prescribed her cholestyramine pills to bind her bile. That worked like a charm.

Then we had her gout (read my basic treatise on gout here). She gets emotionally imbalanced on prednisone, so I tried her on colchicine for her smoldering, chronic gout. It worked, but I didn’t see my way to starting allopurinol for her, since it often aggravates gout in the beginning and patients often need prednisone then. It is taking two colchicine pills a day to control her symptoms, and we had to first get her off her blood pressure medication, carvedilol, because of the interaction between it and colchicine, but it’s working.

Not only that, but she told me her fibromyalgia pain is half of what it used to be and her chronic, debilitating back pain is 90% better.

This sent me googling the literature.

My search results were mixed:

Intravenous colchicine has shown some, albeit temporary, benefit for back pain. Oral colchicine has not shown benefit in double blind studies.

There is no good evidence that it helps fibromyalgia. In fact, there are reports that long term use can cause fibromyalgia or something like it.

Of course, we know colchicine is a powerful anti inflammatory drug that can treat many conditions. Its use dates back to ancient Egypt 1,500 BC. A short list from an excellent review article includes the following:

• Gout

• Pseudogout

• Pericarditis

• Bechet’s Disease

• Actinic keratosis

• Familial Mediterranean Fever

• Secondary Amyloidosis

• Epidermolysis bullosa acquisita

• Leukocytic vasculitis

• Sweet’s syndrome

• Recurrent aphtous stomatitis

• Chronic urticaria

• Granuloma Annulare

• Hennoch Shönlein Purpura

Colchicine was briefly studied as a potential Covid drug with dismal results.

Colchicine appears to lower overall cancer incidence and cardiovascular disease. An older study by Crittenden found that the myocardial infarction prevalence in colchicine treated gout patients was 1.2% versus 2.6% in patients on other treatments. A more recent study found a 49% lower risk of primary cardiovascular events in colchicine users and also a 73% reduction in all cause mortality.

This is better than the 25% risk reduction of the latest fad drug, Jardiance, which reduces blood sugar by lowering the renal threshold for peeing out sugar. It is now recommended by cardiologists for reducing cardiovascular deaths even in non-diabetics. Colchicine is more in the benefit range of atorvastatin, Lipitor, which was at one point the best selling drug in this country.

Will colchicine ever climb to those heights? Not likely, because it is old, generic and now cheap again. But there is an irony here: Colchicine never underwent the rigorous approval process of modern times because it was such an old drug. The FDA disallowed several such old medications and an enterprising drug company offered to do the testing in return for an exclusive patent of this ancient drug. They got what they wanted and the price went up astronomically. But now colchicine is generic again and there is no money to be made from these promising newer cardiovascular and cancer reducing properties.