Wednesday, June 4, 2014

Statins Linked to Diabetes

In an article in the British Medical Journal, researchers document that they have found a link between higher potency statins and new onset of Type II (adult onset) diabetes. This is stunning, at least to me.

The three drugs that were defined for the study as higher potency were these:

  • 10 or more mg rosuvastatin (Crestor)
  • 20 or more mg atorvastatin (Lipitor)
  • 40 or more mg simvastatin (Zocor)
In the US, these are all among the 50 most prescribed drugs. In fact, depending on which list you use, they might all be among the 10 most prescribed drugs. And, according to this study, they increase the likelihood of a patient developing diabetes.

That's not good news.

Often, I read data similar to this and really wonder about its credibility. The data is reported by people with no background in statistics and is published in a non-refereed journal. That is not the case here. The researchers working on behalf of the Canadian Network for Observational Drug Effect Studies include a variety of academics including some with measurement and statistical training. The descriptions of their methods and data sources are sound.

So, what did they find?

The study looked at patients who were prescribed higher potency versus lower potency statins for secondary prevention. In other words, these were patients (at least age 40) who had been hospitalized for a major cardiovascular event (heart attack, stroke, bypass, etc.) who had never previously been diagnosed with diabetes and who were newly prescribed a statin. The key metric that the researchers used is a rate ratio, a common term in epidemiology.

Simply stated, the rate ratio is the ratio of incidences. A rate ratio of 1.20, for example, would indicate that the incidence rate for the studied group is 20 percent higher than that for the control group. They reported data at the 95th confidence interval.

In the first two years post-statin intervention, the researchers found a rate ratio of 1.15 (95% confidence interval of  1.05 to 1.26). In the first four months of use, however, the rate ratio was 1.26 (95% CI of 1.07 to 1.47).

I repeat. This is stunning. the increase in the likelihood of a new diabetes diagnosis for higher potency statin users in the first four months post-intervention was 26%. Increases between four months and two years were much lower, but there were still more newly diagnosed or treated diabetics among higher potency statin users than lower potency users. And, both levels of statin users showed increases in diagnoses of diabetics.

From a practical standpoint, what does this mean?

It's a little bit difficult to tell. From a purely lay standpoint (I have no medical training), the researchers present no data that suggests if there is a similar effect for patients who have not had a major cardiovascular event. However, it is clear that the use of high potency statins is linked to the onset of diabetes. The data set is large and the confidence interval sufficiently high.

What I suspect (and I repeat that I have no medical training) is that people who have a predisposition toward diabetes are significantly more likely to become diabetic after taking statins, especially high potency statins. The researchers note that higher potency statins have not shown statistically significant better results for secondary intervention than lower potency statins.

Looking at this from a statistical and financial standpoint, I would note the following:
  • The three drugs considered higher potency seem to result a statistically significantly higher rate of diabetes onset.
  • Diabetes is one of the highest cost and most dangerous chronic medical conditions, at least in the US.
  • Diabetes, when found in a comorbid state (other chronic conditions exist), increases medical risk very meaningfully.
  • Patients with a predisposition to diabetes along with their physicians should very carefully weigh the risk and benefits of various statins and their dosages as compared to other forms of treatment.

No comments:

Post a Comment