Latest News

Levothyroxine dose for checkpoint inhibitor toxicity may be too high



Current dosing recommendations for thyroid replacement during immune checkpoint inhibitor therapy may overshoot the mark, both for patients with preexisting and de novo hypothyroidism.

Dr. Megan Kristan, a diabetes, endocrinology, and nutrition fellow at the University of Maryland, Baltimore Kari Oakes/MDedge News

Dr. Megan Kristan

The real-world data, presented by Megan Kristan, MD, at the annual meeting of the American Thyroid Association, refine recommendations for dosing by body weight for levothyroxine in patients receiving checkpoint inhibitor therapy.

Immune checkpoint inhibitors stand a good chance of turning the tide against melanoma, some lung cancers, and other malignancies that have long been considered lethal. However, as more patients are exposed to the therapies, endocrinologists are seeing a wave of thyroid abnormalities, and must decide when, and at what doses, to treat hypothyroidism, said Dr. Kristan, a diabetes, endocrinology, and nutrition fellow at the University of Maryland, Baltimore.

Six checkpoint inhibitors are currently approved to hit a variety of molecular targets, and the prevalence of thyroid toxicity and hypothyroidism across the drug class ranges from a reported 9% to 40%, said Dr. Kristan.

The acknowledged thyroid toxicity of these drugs led the American Society for Clinical Oncology (ASCO) to issue guidelines advising that oncologists obtain baseline thyroid function tests before initiating checkpoint inhibitors, and that values be rechecked frequently – every 4-6 weeks – during therapy.

The guidelines advise dosing levothyroxine at approximately 1.6 mcg/kg per day, based on ideal patient body weight. The recommendation is limited to patients without risk factors, and approximates full levothyroxine replacement.

However, some patients enter cancer treatment with hypothyroidism, and some develop it de novo after beginning checkpoint inhibitor therapy. It is not known how best to treat each group, said Dr. Kristan.

To help answer that question, she and her collaborators at Georgetown University Hospital, McLean, Va., made use of a database drawn from five hospitals to perform a retrospective chart review. They looked at 822 patients who had received checkpoint inhibitor therapy, and from those patients, they selected 118 who had a diagnosis of hypothyroidism, or who received a prescription for levothyroxine during the 8-year study period.

The investigators assembled all available relevant data for each patient, including thyroid function tests, levothyroxine dosing, type of cancer, and type of therapy. They sorted participants into those who had received a diagnosis of hypothyroidism before or after receiving the first dose of checkpoint inhibitor therapy.

At baseline, 81 patients had preexisting hypothyroidism and were receiving a mean levothyroxine dose of 88.2 mcg. After treatment, the mean dose was 94.3 mcg, a nonsignificant difference. The median dose for this group remained at 88 mcg through treatment.

For the 37 patients who developed hypothyroidism de novo during checkpoint inhibitor therapy, the final observed levothyroxine dose was 71.2 mcg.

The mean age of the patients at baseline was 69 years. About half were women, and 91% were white. Either nivolumab or pembrolizumab was used in 72% of patients, making them the most commonly used checkpoint inhibitors, though 90% of patients received combination therapy. Taken together, melanoma and lung cancer accounted for about two-thirds of the cancers seen.

For both groups, the on-treatment levothyroxine dose was considerably lower than the ASCO-recommended, weight-based dosing, which would have been 122.9 mcg for those with preexisting hypothyroidism and 115.7 mcg for those who developed hypothyroidism on treatment (P less than .001 for both).

Dr. Kristan noted that thyroid stimulating hormone (TSH) values for patients with pretreatment hypothyroidism peaked between weeks 12 and 20, though there was no preemptive adjustment of levothyroxine dosing.

For those who developed on-treatment hypothyroidism, TSH values peaked at a series of times, at about weeks 8, 16, and 32. These waves of TSH elevation, she said, support the 4- to 6-week follow-up interval recommended in the ASCO guidelines.

However, she said, patients with de novo hypothyroidism “should not be started on the 1.6-mcg/kg-a-day weight-based dosing.” The cohort with de novo hypothyroidism in Dr. Kristan’s analysis required a daily dose of about 1 mcg/kg, she said. These real-world results support the idea that many patients on checkpoint inhibitors retain some thyroid reserve.

Dr. Kristan said that based on these findings, she and her collaborators recommend monitoring thyroid function every 4-6 weeks for patients taking immune checkpoint inhibitors. Patients with preexisting thyroid disease should not have an empiric adjustment of levothyroxine dose on checkpoint inhibitor initiation. For patients who develop thyroiditis after starting therapy, initiating a dose at 1 mcg/kg per day of ideal body weight is a good place to start, and treatment response should be monitored.

The study was limited by its retrospective nature and the small sample size, acknowledged Dr. Kristan. In addition, there were confounding variables and different frequencies of testing across institutions, and antibody status was not available and may have affected the results. Testing was performable for all participants.

Dr. Kristan said that the analysis opens up areas for further study, such as which patient populations are at risk for developing thyroid toxicity, what baseline characteristics can help predict which patients develop toxicity, and whether particular checkpoint inhibitors are more likely to cause toxicity. In addition, she said, a subset of patients will develop hyperthyroidism on checkpoint inhibitor therapy, and little is known about how to treat that complication.

Dr. Kristan reported no conflicts of interest. The research she presented was completed during her residency at Georgetown University.

SOURCE: Kristan M et al. ATA 2019, Oral Abstract 25.

Next Article: