Good riddance?

This week we suspended our tuberculin skin test program for healthcare workers due to the nationwide shortage of tuberculin. I say, Glory Hallelujah! I can't think of a more poorly performing test. Really, it's junk. And we've reached a point at our hospital that no healthcare worker will take isoniazid after a positive skin test until an interferon-gamma release assay has been done and is positive. However, there doesn't appear to be a good correlation between the skin test and the IGRA, so when the two tests are discordant, it's a coin toss as to which is the true result. To complicate matters even more, there's also a nationwide shortage of isoniazid. And don't even get me started on the use of isoniazid to treat latent TB (see here).

Photo:  Mayo Clinic

So just how evil is INH?

A recent issue of MMWR has a report of 17 cases of severe liver injury occurring in persons being treated with isoniazid (INH) for latent tuberculosis. Five of these patients required liver transplantation, and five died. Of note, all patients were monitored for toxicity in compliance with guidelines. These cases were detected via a passive surveillance system. I suspect that this is just the tip of the iceberg since as I have blogged before, I have personally seen 3 cases in the last few years (two died, one survived after liver transplantation). In addition, less than half of patients complete the 9-month recommended course of therapy. On the other hand, only a small percentage of those treated develop severe liver injury, since it is estimated that 300,000-400,000 are treated with the drug each year. Still, I continue to believe that this drug is simply too toxic to treat a latent infection which has on average only a 10% chance of converting to active disease. Certainly some patients are at much higher risk for developing active tuberculosis, but those patients are also often at higher risk of toxicity. For those clinicians who feel compelled to use it, I would recommend careful informed consent and monitoring of transaminases for all patients, not just those for whom testing is recommended by the ATS/CDC guideline.

Treatment of latent tuberculosis: efficacy vs effectiveness

I've blogged before about the poor options we have for treating latent tuberculosis. The standard regimen, 9 months of isoniazid (INH), is too long and too toxic for the treatment of tuberculosis that is not active. For patients that aren't immunosuppressed and not recent skin test converters, the risk for developing active tuberculosis is only about 5%. Though the data for efficacy are not as robust, the 4-month daily rifampin regimen is much easier and better tolerated, and I have become more comfortable with this regimen. Because of the referral nature of my practice, the patients referred to me for evaluation of latent tuberculosis often have elevated liver enzymes at the outset or may not want to wait 9 months to complete the standard regimen before starting a therapy they feel is important (e.g., TNF-alpha inhibitors or other immunosuppressants).

A new study reported in this month's Chest examines completion rates of latent TB treatment courses. The investigators found that less than half of persons started on a treatment course actually finish it. About 45% complete a 9-month course of INH, 55% complete a 6-month course of INH, and nearly 65% complete the 4-month rifampin regimen. This leads me to wonder: even if 9 months of INH were ultimately proven to be more efficacious, perhaps 4-months of rifampin is more effective. Efficacy refers to how well a therapy works under ideal conditions (i.e., you take all the doses as prescribed), whereas effectiveness refers to how well a therapy works under real-world conditions (i.e., how well does the drug work when compliance is factored in). Given the toxicities of INH and the longer duration of treatment, coupled with the results of this survey, it's not far-fetched to conclude that what is considered our most efficacious regimen for latent tuberculosis may be our least effective.

When the treatment is worse than the disease

The University of Missouri Health Care System is re-evaluating its current policy on healthcare workers with latent tuberculous infection, which currently encourages but does not mandate treatment. The impetus for the review is a healthcare worker who had tested positive for latent tuberculosis via positive skin testing many years ago but declined to be treated, and recently developed active disease, exposing over 200 persons to tuberculosis. I suspect that most hospitals have similar policies and am not aware of a hospital that either prohibits those with latent but untreated infection from working or mandates treatment of latent tuberculosis. By definition, latent tuberculosis is not contagious but every person with latent infection has a small chance of developing active infection, which is contagious.

Isoniazid (INH), the standard drug used to treat latent tuberculosis given daily for nine months, is not an innocuous drug. In just the past few years, I have seen two deaths (both persons in their early 40s) and another healthy man who required a liver transplant due to the liver toxicity caused by the drug. Thus, I believe the drug is too toxic to require latently infected healthcare workers to be treated. Like the vast majority of healthcare workers, I have a TB skin test placed yearly. But I have decided that should my skin test convert to positive, I would not take INH because I believe it is too dangerous. We sorely need safer drugs for this condition. And we need better tests to detect latent tuberculosis--those currently available have many false positives and false negatives.