In Which I Go Under The Knife And Learn About Medicine In The Real World

I’m not a doctor but I thought I knew something about anticoagulation. Over the course of a career covering cardiology I’ve written countless stories about heparin, warfarin, the low molecular weight heparins, and the new oral anticoagulants.

So when I had bilateral total knee replacement a few weeks ago I thought I knew what to expect, at least in this one particular area. (I was completely out of my depth for all the rest, of course.) I wasn’t surprised when the day after surgery I was given a shot of Lovenox, and I wasn’t surprised a day later when I was switched to Xarelto (rivaroxaban). (I even enjoyed asking my doctor why he chose rivaroxaban over apixaban, since I knew this would have been the preference of nearly all my cardiologist friends.)

In the first few days my treatment seemed in accord with the trials I had covered. UpToDate has a useful summary. It recommends “the routine use of pharmacologic prophylaxis (Grade 1A). Reasonable choices include: LMW heparin at a usual high-risk dose, fondaparinux, a vitamin K antagonist (target INR 2.5, range: 2.0 to 3.0), dabigatran, or rivaroxaban. All have been approved for this indication.” The UpToDate authors also “suggest” that the regimen extend out to 35 days after TKR.

So I thought I knew what to expect for the next month of so. I assumed that I would stay on the rivaroxaban for at least a few weeks. But on the next day I received a big surprise. After only one day on rivaroxaban I was switched to aspirin, one dose a day each of 81 mg and 325 mg. I asked about this and was told that it was because I had been very aggressive about getting up and moving around, so the surgeon was less concerned that I would develop a clot.

I was confused. I shot off an email to Sanjay Kaul, who served on many of the FDA panels that reviewed the new oral anticoagulants. His response was that it looked like an “ad hoc treatment protocol for thromboprophylaxis.” He observed that in his experience orthopedic surgeons are not strict followers of the guidelines. “They all follow their own best protocol!”

As a medical journalist who has always been a big supporter of evidence-based medicine I was slightly suprised and concerned about this turn of events. (On the other hand, I was also quietly pleased, because the low-risk designation seemed to validate my own gung-ho approach to recovery.)

Things got even more peculiar when I was transferred to a rehab facility a few days later. The doctor and nurse practitioner at the rehab hospital were concerned that the aspirin was not strong enough. They wanted to put me back on Lovenox. Before agreeing to that I asked them to consult with the surgeon. I don’t know the details but I imagine the surgeon pulled rank since I remained on the aspirin. However, the medical staff at my rehab hospital were still worried that I might develop a clot, so they sent me for a Doppler ultrasound. That test was negative and I remained on the aspirin at the rehab hospital and have remained on it for nearly 2 weeks since I’ve been home.

Did I receive the best treatment? It’s impossible to know. If I make it another week without developing a clot then I certainly had a good outcome. Or maybe I’ll just be lucky. In any case, one anecdote is not data, and it would be dangerous to draw from story any broad conclusion about optimal thromboprophylaxis after TKR.

Evidence-Based Medicine?

But what does my story say about the role of evidence-based medicine in the real world?

Think of this: when it comes to anticoagulation we probably have as large an evidence base as anything in medicine, with the possible exception of statins. But if my case is in any way typical, many physicians feel entirely comfortable employing their own treatment strategies based on their personal experiences and preferences.

Are they wrong to do this? It strikes me that we don’t even know how to assess this. I suspect that for most physicians their individual choices work out reasonably well for most of their paitients— as, I hope, is true in my case— but we simply don’t know. And since there is a wide variety of opinions, it is highly likely that some doctors make good choices and other doctors make poor choices. The only problem, as the old joke has it, we have no way of knowing which doctor is in which group.

Here’s one clear limitation: the vast majority of the clinical trial data was accumulated for the singular purpose of getting individual drugs approved. The trials all follow rigorous designs in order to meet FDA standards and generally provide a useful answer to the specific situation addressed by the trial.

Let’s leave aside for now all the objections that have been raised about the unreliability of clinical trials and the myriad ways these trials can be manipulated. And let’s also leave aside the important economic issues. Overall, I think it’s fair to say that the development over the last generation of the low-molecular weight heparins and the new oral anticoagulants represents an important and welcome advance.

But it’s also true that although the evidence to support approval of the drugs is strong, there’s very little evidence to support how to use the drugs in real practice. There are few comparative trials, and little motivation or financing to perform them. Imagine, for instance, that my surgeon’s aspirin regimen was really as effective as the alternatives. Who would conceivably pay the hundreds of millions of dollars it would take to pay for the trial. And that’s just one trial, for one indication. Undoubtedly there are dozens of potential thromboprophylaxis strategies worth testing against each other. And this is just one small corner of medicine. The unanswered questions are endless.

There will never be enough traditional clinical trials to answer these questions. Observational studies have even more limitation. The only solution that I am aware of is the randomized registry trial. Here is how two editorialists described one of the first such trials a few years ago:

“The TASTE investigators designed a large-scale trial to answer an important clinical question and carried it out at remarkably low cost by building on the platform of an already-existing high-quality observational registry. With this clever design, which leveraged clinical information that was already being gathered for the registry and for other preexisting databases, the investigators were able to quickly identify potential participants, to enroll thousands of patients in little time…, to avoid filling out long case-report forms, to obtain accurate follow-up with minimal effort, and to report their findings, all for less than the amount of a typical modular R01 grant…”