My previous KevinMD article discussed a Cochrane Collaboration review that was completed in 2002 but was not published until 2006. This review examined three randomized controlled trials (RCTs) comparing anticoagulants with placebos or anti-inflammatory drugs for venous thromboembolism (VTE).

Along with my two co-authors, I found that across these three trials, anticoagulant-treated patients had six deaths out of 66 patients, compared to only one death out of 60 in the placebo group. We concluded that anticoagulation lacked evidence of efficacy and likely caused harm. Due to Cochrane’s five-year delay in publication, I submitted my version to George Lundberg, then chief editor of Medscape General Medicine. He published it in 2004, accompanied by a video editorial titled “Is the Current Standard of Medical Practice for Treating Venous Thromboembolism Simply Wrong?” (viewable only by subscribers to Medscape). In that editorial, Dr. Lundberg noted:

A patient has a disease. A physician invents a treatment that seems to make sense and tries it. If the patient improves, both the physician and the patient may credit the treatment. This is an anecdote and proves nothing. But, the treatment is tried again and the next patient improves. Soon, the physician believes in the treatment and tells colleagues about it. They try it and more patients recover. So the doctors write an article about their observations. What now should happen is called the controlled clinical trial with randomization, blinding, and statistical power, the gold standard for evaluating interventions. Unfortunately, what often does happen is simply a proliferation of use of the treatment before it has been adequately tested. When enough physicians in a given region use the treatment, it becomes labeled as the standard of practice. It may even become required of other physicians to use even though it has never been proven. Elsewhere in today’s MedGenMed, author David Cundiff argues from a lengthy literature that such a conundrum obtains for one of the most commonly practiced medical interventions, anticoagulation therapy for venous thromboembolism. Many physicians will disagree with Cundiff’s conclusions, some strenuously. But I believe that the data, as he presents them, are pretty impressive. I recommend you pay attention to this challenge of a medical sacred cow. That’s my opinion. I’m Dr. George Lundberg, Editor of MedGenMed.

Despite this, Cochrane editors eventually insisted on softening our conclusion to state that evidence was “inconclusive” and that further placebo-controlled trials would be “unethical” because the treatment was already “widely accepted.”

I now believe a retraction of that paper is necessary. A retraction might have garnered the media attention required to spark a true scientific debate and, potentially, save lives.

The global impact of anticoagulation

It is estimated that well over 50 million people worldwide are currently treated with anticoagulants. Atrial fibrillation alone affects roughly 37 to 40 million people globally, the majority of whom require long-term anticoagulation. Below is a breakdown of the estimated morbidity and mortality from anticoagulants globally for the 30 Food and Drug Administration medical indications:

• Major bleeding: 1.2 to 2.0 million events (2.0 percent to 3.5 percent of treated patients).
• Fatal bleeding: 150,000 to 300,000 deaths (0.3 percent to 0.6 percent of treated patients).
• Intracranial hemorrhage: 100,000 to 200,000 events (0.2 percent to 0.4 percent of treated patients).

Estimated impact of anticoagulants in the U.S. includes:

• Emergency department visits: ~235,000+ (Key indications: Atrial fibrillation, DVT/PE treatment).
• Inpatient adverse events: ~10 percent of all drug errors (Key indications: Post-surgical prophylaxis, VTE).
• Major bleeding events: 15 percent to 20 percent of users/year (Key indication: Long-term warfarin/DOAC use).
• Estimated mortality: ~3,000 to 6,000+ deaths (Key causes: Intracranial hemorrhage, GI bleeding).

Methodological errors in the literature

In response to Cochrane’s resistance, I systematically reviewed the evidence base of 57 Cochrane reviews covering 30 different medical indications for anticoagulation.

While the Cochrane editors acknowledged my feedback, their response was minimal: 12 months later, authors had replied to only 13 of the 57 letters. My analysis identified 207 instances of methodological errors and 18 instances of bias across these reviews. I concluded then, as I do now: The efficacy of anticoagulants for these 30 indications must be reconsidered by the scientific community. I will use the example of atrial fibrillation, the anticoagulation indication with the most harm to patients.

Atrial fibrillation: real-world versus researcher-monitored populations

With atrial fibrillation, accounting for most bleeding complications, a primary reason for the high rate of complications is the “age gap” between clinical trials and real-world patients. In the 1990s, the average age in major AF trials (like AFASAK or SPAF I) was roughly 67 years old. Today, the average hospitalized AF patient in the U.S. is 75 to 82 years old.

Why the rates differ:

• Exponential risk: Major hemorrhage risk roughly doubles for every decade of life.
• Selection bias: RCTs often exclude the “very old,” those with high fall risks, or patients with multiple comorbidities (the exact population most treated today).
• Monitoring intensity: Real-world “time in therapeutic range” (TTR) is often far lower than in controlled studies, leading to higher bleeding rates (4 to 7 percent in practice versus 1 to 3 percent in trials).

David K. Cundiff is a physician, author, and health care reform advocate whose work centers on transforming the U.S. health care system and addressing broader societal challenges, including climate change. He is the author of Grand Bargains: Fixing Health Care and the Economy, which proposes structural reforms to dramatically reduce health care costs while improving outcomes. His essay “Much Better Healthcare for Way Less Cost” explores accountable care cooperatives and community-based reform. Additional works include Money Driven Medicine – Tests and Treatments That Don’t Work and Whistleblower Doctor: The Politics and Economics of Pain and Dying.

From 1981 to 1998, Dr. Cundiff practiced, taught, and conducted clinical research in internal medicine and pain control at the Los Angeles County + USC Medical Center, where he directed the Cancer and AIDS Pain Service for nine years, and previously held an academic affiliation with Harbor-UCLA Medical Center. After exposing how systemic inefficiencies increased hospital utilization and revenue, he became a whistleblower, an experience documented in Whistleblower Doctor.

Outside his professional work, Dr. Cundiff values time with friends and family, including six grandchildren, and maintains his health through Hatha yoga, meditation, swimming, Zumba, biking, and a diet emphasizing minimally processed organic food.