5 approved drugs that don’t measure up

By Naveed Saleh, MD, MS, for MDLinx
Published January 17, 2020

Key Takeaways

When you prescribe a medication to a patient, your expectation is that the drug will work. The reality, however, is that most patients won’t benefit from most prescribed drugs. Of course, some patients will benefit—otherwise a drug wouldn’t be approved by the FDA in the first place.

It’s somewhat disconcerting how many people need to be treated in order for one case of benefit, a measure referred to as number needed to treat (NNT). Even so, NNT is an imperfect measure and often miscalculated.

Here’s a closer look at NNT, as well as five examples.

NNT defined

The concept of NNT was first introduced in the literature in 1988. It’s a measure of absolute effect that represents the benefit or harm of a medical treatment. Mathematically, NNT = 1/ARR, where ARR stands for the absolute risk reduction between two treatment options (ie, risk difference). NNT is rounded to the nearest whole number (because you can’t have a fraction of a person).

For instance, if a drug has an NNT of 10, it would take 10 people treated with the drug to observe 1 case of benefit or avoidance of 1 bad outcome. Keep in mind that NNT pertains to the results of a given comparison and not to a specific therapy per se.

NNTs are calculated differently based on study characteristics, including trial design and variables used to determine outcomes. As such, it’s not clear whether most study researchers calculate NNT properly.

In an attempt to better elucidate the verisimilitude of NNT, investigators of a study in BMC Medicine analyzed NNTs reported in 25 high-impact general or internal medicine journals.

“A considerable proportion of studies, particularly meta-analyses, applied methods that are not in line with basic methodological recommendations,” concluded the authors. “Despite their usefulness in assisting clinical decisions, NNTs are uninterpretable if incompletely reported, and they may be misleading if calculating methods are inadequate to study designs and variables under evaluation. Further research is needed to confirm the present findings.”

A group of physicians created a website called TheNNT, which details the NNTs for various drugs. Let’s look at a few examples of drugs that may not actually work for everyone.

Statins for heart disease

According to The NNT Group, in patients without known heart disease receiving daily statin therapy for 5 years, the NNT for heart attack prevention is 104. In other words, 104 people must take statins to prevent one additional person from having a heart attack. Moreover, the NNT for stroke prevention is 154.

In recent years, these numbers have served as a point of controversy and contention in the medical community, with some experts questioning whether people with no previous heart disease should even be prescribed statins.

This was the question at the center of a recent “statins war” between two top-tier journals: the BMJ and the Lancet. On one side, researchers argued that prescribing statins to patients with low risk of heart disease would increase the incidence of adverse effects without providing additional health benefits. Researchers on the other side, however, contended that concerns about adverse effects were exaggerated. They also noted that such misleading claims lead to nonadherence among patients and underprescription by clinicians.  

So, who’s right?

Although there continues to be some disagreement among experts, 2018 guidelines on blood cholesterol management from the American College of Cardiology (ACC) and the American Heart Association (AHA) recommends the use of statins to reduce LDL cholesterol levels and decrease the risk of atherosclerotic cardiovascular disease.

Warfarin for atrial fibrillation

Atrial fibrillation is a serious cardiovascular health condition that can result in embolic stroke and may prove fatal. Although oral anticoagulants such as warfarin can decrease the risk of atrial fibrillation, these drugs may actually increase the risk of hemorrhagic stroke and bleeds.

In patients without a history of previous stroke, the NNT for one case of stroke prevention with warfarin is 25. Furthermore, the NNT for the prevention of death from any cause is 42.

Due to concerns over the potential for severe bleeding events with warfarin, health experts are now recommending therapy with direct oral anticoagulants over warfarin in people with atrial fibrillation.

Thrombolytics for pulmonary embolism

“Traditionally, anticoagulation has been the mainstay of treatment,” according to the authors of a review on TheNNT. “In massive pulmonary embolism, however, it is important to restore pulmonary blood flow rapidly, which is often accomplished by the use of thrombolytic agents or surgical embolectomy. The medical literature has reported improvement in clot lysis, restoration of normal pulmonary circulation, a decrease in right heart strain, and improvement in long-term cardiac output and exercise tolerance with the use of thrombolytics. However, not all studies have demonstrated improvement in survival or other patient outcomes with thrombolytics, and thrombolytics carry an increased risk of bleeding.”

The NNT for prevention of one death via thrombolytics is 34, whereas the NNT for prevention of 1 case of recurrent pulmonary embolism is 50.

Systemic antibiotics after abscess incision and drainage

Every year, more than 3 million Americans present to the emergency department with a cutaneous abscess that requires incision and drainage, and this number is on the rise. The use of systemic antibiotics after incision and drainage, however, is controversial. The clinical cure rate without antibiotics is estimated to be 84%, while the cure rate with antibiotics is slightly increased at 92%. Researchers have suggested that this slightly increased cure rate be balanced against the harms linked to antibiotic use, including antibiotic resistance and adverse events such as gastrointestinal disturbances.

Nevertheless, the NNT to prevent 1 case of treatment failure is 14, and to prevent infection recurrence, 10.

Antihypertensives for cardiovascular disease

As we know, blood pressure medications help prevent stroke and heart attack. In people receiving antihypertensive therapy for 5 years, the NNT to prevent 1 death is 125. Additionally, the NNT to prevent heart attack is 100, and to prevent stroke, 67.

But what about people with mild hypertension (systolic blood pressure of 140-159 or diastolic blood pressure of 90-99) without preexisting cardiovascular disease? Do blood pressure medicines prevent cardiovascular events in these patients? 

The answer is no. After up to 5 years of follow up, people with mild hypertension taking blood pressure medicine experienced no differences in mortality, cardiovascular events, coronary artery disease, or stroke, according to the authors of a review on TheNNT. That is, there is no NNT—or, rather, the NNT is infinite. 

Not incidentally, 1 in 12 patients experienced side effects from the medication.

Issues with NNT

As mentioned, NNT is an imperfect metric and often miscalculated.

Importantly, NNT is binary and not continuous. For instance, NNT may consider either survival or death with the use of antihypertensives but not other less stark benefits from lowering blood pressure levels. Keep in mind that most types of treatment exert influence on more than one clinical outcome; not every disease straight-up kills. Moreover, NNT is a time-dependent calculation.

Ultimately, NNT is useful to consider, but it’s not gospel. Every patient’s treatment must be considered individually, with the understanding that any patient seen in your office may not react the same to a drug as members of a study sample.  Furthermore, NNT changes with baseline risk. Finally, even a drug with a high NNT may prove useful if it has few adverse effects and is cheap. Likewise, a drug with a low NNT may be inadvisable if dangerous and expensive.

Share with emailShare to FacebookShare to LinkedInShare to Twitter