Many targeted cancer tablets that block blood vessel growth, such as sunitinib, axitinib, lenvatinib and cabozantinib, can make the thyroid underactive. It causes tiredness, feeling cold, constipation and weight gain, which are easy to blame on the cancer. A blood test (TSH) before and during treatment finds it, and a daily levothyroxine tablet treats it. Your own team's plan comes first.
VEGFR-targeted kinase inhibitors cause hypothyroidism, sometimes after a brief thyrotoxic phase, probably through loss of thyroid capillaries. Label rates: hypothyroidism in 19% on axitinib (TSH rising to 10 or more in 32% of those under 5 beforehand), thyroid dysfunction in 19% on cabozantinib, hypothyroidism in 21% to 24% on lenvatinib, 8% on tivozanib, 4% to 7% on pazopanib and 17% to 21% on fruquintinib (31.5% in the EU text, median onset 56 days). Reviews put it at 32% to 85% on sunitinib, with onset from 4 to 94 weeks. In people already taking levothyroxine, including after thyroidectomy, these medicines can raise the dose needed: on vandetanib 49% needed more against 17% on placebo.
Most labels ask for thyroid function before treatment and during it, and differ on how often: at least monthly for lenvatinib (US); every 3 months for sunitinib (EU); at baseline, 2 to 4 weeks, 8 to 12 weeks and then every 3 months for vandetanib; 'periodically' for axitinib and tivozanib; at baseline, during treatment and as clinically indicated for pazopanib; for cabozantinib (US), an assessment for signs before and during treatment with testing 'as clinically indicated'. The sorafenib US label asks for monthly TSH only in differentiated thyroid cancer, where it impairs TSH suppression; the regorafenib and fruquintinib US labels list hypothyroidism without a monitoring instruction, and the EU regorafenib text asks only for biochemical and metabolic monitoring in general. Two endocrine reviews give tighter schedules: TSH monthly for 4 months then every 2 to 3 months (Hamnvik 2011), or monthly for 6 months, on day 1 of the cycle for intermittent regimens, then every 2 to 3 months (French expert opinion 2018, abstract only).
Treatment: levothyroxine for a TSH above 10 mIU/L or a low free T4 (Hamnvik), or for symptomatic hypothyroidism or a TSH above 10 mIU/L (French opinion). They differ below that: Hamnvik finds a low-dose trial reasonable for a TSH of 5 to 10 because symptoms are hard to tell from those of cancer treatment; the French opinion treats only symptomatic hypothyroidism or a TSH above 10. The ATA guideline covers dosing: full or partial replacement titrated to TSH, rechecked 4 to 6 weeks after any change. Starting doses: 1.6 micrograms per kg a day is called reasonable without cardiac disease, and a lower dose such as 50 micrograms a day for the first few weeks in coronary artery disease (Hamnvik); NICE gives similar figures for primary hypothyroidism in general (see variants). The French opinion says the kinase inhibitor should not be stopped because of thyroid dysfunction, and no label read has a dose rule specific to it.
Why: The EU sunitinib text recommends it for all patients, and the US labels for sunitinib, lenvatinib, axitinib, tivozanib, pazopanib and vandetanib also ask for a thyroid test before treatment.
“Baseline laboratory measurement of thyroid function is recommended in all patients.”
Why: The tivozanib label and the EU sunitinib text ask for existing thyroid disease to be treated before the medicine is started; the axitinib label and the EU lenvatinib text ask for a normal thyroid state to be maintained during treatment.
“Treat hypothyroidism and hyperthyroidism to maintain euthyroid state before and during treatment with FOTIVDA.”
Why: The review calls TSH the single best screening test for thyroid dysfunction, and says free thyroxine should be measured in everyone with an abnormal TSH and in people with pituitary disease.
“The single best screening test for thyroid dysfunction is a TSH assay, except in patients with pituitary disease”
Why: In the French expert opinion about half of cases are subclinical, meaning they show on a blood test before they are felt, and the EU sunitinib text says hypothyroidism occurs early as well as late in treatment.
“Monitor thyroid function prior to initiating LENVIMA and at least monthly during treatment.”
Why: The review proposes this for sunitinib, sorafenib, axitinib and related medicines from the reported time course, and says its recommendations are logical but not evidence based.
“TSH before treatment and then every month for the first 4 mo, then every 2–3 mo.”
Why: The French opinion gives this timing in its abstract. The 2011 review cites a report of TSH rising during sunitinib treatment and recovering in the two-week break, so the timing of the test can change the result.
“During TKI treatment, TSH should be assayed monthly, or at end of off-period (i.e., day 1 of new cycle after interruption), for the first 6 months, then every 2-3 months or in case of clinical signs of dysthyroidism.”
Why: The symptom lists are those of the EU lenvatinib patient leaflet. The 2011 review notes that symptoms of hypothyroidism may be wrongly put down to the cancer or its treatment, and can lead to the dose of the cancer medicine being reduced.
“Monitor patients closely for signs and symptoms of thyroid dysfunction, including hypothyroidism, hyperthyroidism, and thyroiditis, during treatment with SUTENT.”
Why: In the French opinion a passing thyrotoxicosis comes before hypothyroidism in 20% to 40% of cases and may be overlooked. The US sunitinib label also reports hyperthyroidism, in some people followed by hypothyroidism.
“Treatment is symptomatic in case of thyrotoxicosis, and levothyroxine replacement therapy is initiated in case of symptomatic hypothyroidism or TSH>10mIU/L.”
Why: The review says treatment should be started at this level, and cites one study in which at least half of the people who started levothyroxine for hypothyroidism on sunitinib had less fatigue. The French opinion uses the same TSH threshold.
“In patients with TSH greater than 10 mIU/L or with low free T4 levels, treatment should be initiated.”
Why: The review reasons that in someone on cancer treatment it is hard to tell whether mild hypothyroidism is truly free of symptoms. The French opinion starts levothyroxine only for symptomatic hypothyroidism or a TSH above 10 mIU/L, so the sources differ here (see variants).
“Therefore, it seems reasonable to start low-dose levothyroxine therapy in these patients as a therapeutic trial, especially if they are receiving an agent associated with a high risk of hypothyroidism.”
Why: The guideline says levothyroxine reaches a steady level by 4 to 6 weeks. It reports that many experts start low and go slow in older people and those with cardiovascular disease, for fear of setting off cardiac events; that advice is in its discussion and is not part of the graded recommendation. Starting doses are in the variants.
“Thyroid hormone therapy should be initiated as an initial full replacement or as partial replacement with gradual increments in the dose titrated upward using serum thyrotropin as the goal.”
American Thyroid Association (Thyroid; Jonklaas et al.) 2014 · Strong recommendation. Moderate quality of evidence.
Why: The guideline says these medicines affect how thyroxine is metabolised, and reports that people with no thyroid have needed substantially higher doses. In the French opinion 20% to 60% of people already on levothyroxine need more. Schedules for rechecking are in the variants.
“Serum thyrotropin should also be reassessed in patients who are started on agents such as tyrosine kinase inhibitors that affect thyroxine metabolism and thyroxine or triiodothyronine deiodination.”
American Thyroid Association (Thyroid; Jonklaas et al.) 2014 · Strong recommendation. Low-quality evidence.
Why: The French expert opinion says this directly, in its abstract. No label read gives a dose rule specific to thyroid dysfunction, although the cabozantinib label refers from its thyroid section to its general dose-modification rules for adverse reactions.
“TKIs should not be discontinued because of onset of thyroid dysfunction.”
Why: The thyroid recovers in some people and not in others, and it cannot be predicted which.
“At TKI termination, remission of hypothyroidism is possible but unpredictable, and progressive discontinuation of levothyroxine may be considered under monitoring.”
How care is usually carried out, read from the guidelines and standards cited. Your own team's protocol and their judgement come first; this page is for understanding it, not for doing it yourself.
Most labels that address the problem ask for thyroid function before treatment and during it, with thyroid dysfunction treated (to a euthyroid state in the axitinib, tivozanib and EU lenvatinib texts), and none gives a dose rule specific to thyroid dysfunction; the cabozantinib label refers from its thyroid section to its general dose-modification rules. They differ on frequency: at least monthly (lenvatinib, US label); every 3 months (sunitinib, EU text; the US label says 'periodically'); baseline, 2 to 4 weeks, 8 to 12 weeks, then every 3 months (vandetanib); 'periodically' (axitinib, tivozanib; lenvatinib in the EU text); at baseline, during treatment and as clinically indicated (pazopanib); an assessment for signs before and during treatment, with testing 'as clinically indicated' (cabozantinib, US). Two endocrine sources advise a fixed, tighter schedule for all of them: monthly for 4 months then every 2 to 3 months (Hamnvik 2011, which calls its schedule logical but not evidence based), or monthly for 6 months then every 2 to 3 months (French Endocrine Society expert opinion 2018, read as abstract only). Both start levothyroxine for a TSH above 10 mIU/L (Hamnvik also for a low free T4, the French opinion also for symptomatic hypothyroidism); below 10 they differ, Hamnvik finding a low-dose trial reasonable and the French opinion treating only if symptomatic. The ATA hypothyroidism guideline gives the only graded statements: reassess TSH when a kinase inhibitor is started in someone on levothyroxine (strong recommendation, low-quality evidence) and titrate levothyroxine to TSH with a recheck 4 to 6 weeks after any change (strong recommendation, moderate quality). No ESMO, ASCO or European Thyroid Association guideline on this was found.
Eisai, via DailyMed (US National Library of Medicine) 2026 · Pfizer, via DailyMed (US National Library of Medicine) 2026 · European Medicines Agency 2026 · Genzyme (Sanofi), via DailyMed (US National Library of Medicine) 2026 · Pfizer, via DailyMed (US National Library of Medicine) 2026 · Novartis, via DailyMed (US National Library of Medicine) 2025 · AVEO, via DailyMed (US National Library of Medicine) 2026 · Exelixis, via DailyMed (US National Library of Medicine) 2025 · European Medicines Agency 2026 · Journal of the National Cancer Institute (Hamnvik, Larsen and Marqusee); funded by the US National Institutes of Health 2011 · Annales d'Endocrinologie (Drui, Illouz, Do Cao and Caron; French Endocrine Society expert consensus) 2018 · American Thyroid Association (Thyroid; Jonklaas et al.) 2014
Where it differs: Lenvatinib: before treatment and at least monthly (US); before and periodically (EU). Sunitinib: baseline and periodically (US); baseline and every 3 months (EU). Vandetanib: TSH at baseline, 2 to 4 weeks, 8 to 12 weeks and every 3 months. Axitinib, tivozanib: before and periodically. Pazopanib: at baseline, during treatment and as clinically indicated. Cabozantinib (US): assess for signs before and during treatment, with testing as clinically indicated. Sorafenib (US): TSH monthly in people with differentiated thyroid cancer, with thyroid replacement adjusted; no instruction for other cancers. Regorafenib and fruquintinib (US): hypothyroidism listed as an adverse reaction, no monitoring instruction. Regorafenib (EU): monitoring of biochemical and metabolic parameters is recommended in general, with raised TSH named among the abnormalities.
Evidence: Regulatory texts; the intervals are not from trials of monitoring. Rates in the labels: axitinib 19% (sorafenib 8% in the same trial); cabozantinib 19% thyroid dysfunction; lenvatinib 21% to 24% grade 1 or 2 hypothyroidism; tivozanib 8%; pazopanib 7% in the kidney cancer trial and 5% in the sarcoma trial; fruquintinib 21% and 17% in its two US trial tables and 31.5% in the EU text; regorafenib 18% against 6% on placebo in the GIST trial (among those with a normal TSH and no thyroid supplement at baseline) and 6.4% in the liver cancer trial (US), with about 4.6% needing hormone replacement across the phase 3 trials (EU).
Eisai, via DailyMed (US National Library of Medicine) 2026 · European Medicines Agency 2026 · Pfizer, via DailyMed (US National Library of Medicine) 2026 · European Medicines Agency 2026 · Genzyme (Sanofi), via DailyMed (US National Library of Medicine) 2026 · Pfizer, via DailyMed (US National Library of Medicine) 2026 · Novartis, via DailyMed (US National Library of Medicine) 2025 · AVEO, via DailyMed (US National Library of Medicine) 2026 · Exelixis, via DailyMed (US National Library of Medicine) 2025 · Bayer, via DailyMed (US National Library of Medicine) 2024 · Takeda, via DailyMed (US National Library of Medicine) 2026 · European Medicines Agency 2026 · Bayer, via DailyMed (US National Library of Medicine) 2026 · European Medicines Agency 2026
Where it differs: TSH before treatment, monthly for the first 4 months, then every 2 to 3 months (Hamnvik 2011, for sunitinib, sorafenib, dasatinib, nilotinib, axitinib and cediranib). TSH monthly for the first 6 months, timed to day 1 of the cycle on intermittent regimens, then every 2 to 3 months; and for people already on levothyroxine, monthly for 3 months then every 3 months (French expert opinion 2018).
Evidence: Expert opinion from the reported time course: in the review, hypothyroidism on sunitinib occurred in 32% to 85% across studies, starting 4 to 94 weeks into treatment. Hamnvik 2011 states that no official screening recommendations existed and that its own are not evidence based.
Journal of the National Cancer Institute (Hamnvik, Larsen and Marqusee); funded by the US National Institutes of Health 2011 · Annales d'Endocrinologie (Drui, Illouz, Do Cao and Caron; French Endocrine Society expert consensus) 2018
Where it differs: A trial of low-dose levothyroxine even without clear symptoms (Hamnvik 2011), against levothyroxine only for symptomatic hypothyroidism or a TSH above 10 mIU/L (French expert opinion 2018).
Evidence: Opinion on both sides. Hamnvik notes that treating subclinical hypothyroidism is generally discouraged in healthy people for lack of evidence of benefit, and argues that in cancer patients symptoms are hard to separate from those of treatment or the disease. Neither cites a trial of treating a TSH of 5 to 10 mIU/L in people on kinase inhibitors. For primary hypothyroidism in general, NICE says to consider a 6-month trial of levothyroxine in adults under 65 with symptoms and a TSH above the reference range but below 10 on 2 occasions 3 months apart.
Journal of the National Cancer Institute (Hamnvik, Larsen and Marqusee); funded by the US National Institutes of Health 2011 · Annales d'Endocrinologie (Drui, Illouz, Do Cao and Caron; French Endocrine Society expert consensus) 2018 · National Institute for Health and Care Excellence 2019
Where it differs: Hamnvik 2011, writing about hypothyroidism from cancer medicines: a starting dose of 1.6 micrograms per kg a day is reasonable for people without underlying cardiac disease, and in coronary artery disease a lower dose such as 50 micrograms a day may be used for the first few weeks. ATA 2014 (discussion, not a graded recommendation): full replacement, roughly 1.6 micrograms per kg, when TSH is markedly raised, and lower doses such as 25 to 50 micrograms in milder hypothyroidism, for example a TSH of 10 mIU/L or less. NICE NG145, for primary hypothyroidism in general: consider 1.6 micrograms per kg a day, rounded to the nearest 25 micrograms, for adults under 65 with no history of cardiovascular disease, and consider 25 to 50 micrograms a day with titration for adults aged 65 and over and those with a history of cardiovascular disease.
Evidence: None of the three figures comes from a study of hypothyroidism caused by a kinase inhibitor. The ATA discussion cites one randomised trial in which older people free of cardiovascular disease were safely started on the full dose. NICE does not grade its recommendations; both of these use 'consider'.
Journal of the National Cancer Institute (Hamnvik, Larsen and Marqusee); funded by the US National Institutes of Health 2011 · American Thyroid Association (Thyroid; Jonklaas et al.) 2014 · National Institute for Health and Care Excellence 2019
Where it differs: No new diagnosis to make: TSH is rechecked every 4 weeks after starting the kinase inhibitor and the levothyroxine dose raised as needed, then every 2 months once stable (Hamnvik 2011, written for imatinib, sorafenib and motesanib). In thyroid cancer the EU lenvatinib text says thyroid hormone should be adjusted to the patient's own TSH target, and the lenvatinib and sorafenib labels say these medicines impair TSH suppression.
Evidence: Vandetanib label: in the medullary thyroid cancer trial, where 90% had had a thyroidectomy, 49% needed more thyroid replacement against 17% on placebo. Lenvatinib label: among thyroid cancer patients with a normal TSH at baseline, TSH rose above 0.5 mU/L in 57%. Sorafenib label: TSH rose above 0.5 mU/L in 41% against 16% on placebo. The French opinion puts the share of levothyroxine users needing a higher dose at 20% to 60%. The ATA guideline grades reassessing TSH as a strong recommendation on low-quality evidence.
Journal of the National Cancer Institute (Hamnvik, Larsen and Marqusee); funded by the US National Institutes of Health 2011 · Genzyme (Sanofi), via DailyMed (US National Library of Medicine) 2026 · Eisai, via DailyMed (US National Library of Medicine) 2026 · European Medicines Agency 2026 · Bayer, via DailyMed (US National Library of Medicine) 2024 · Annales d'Endocrinologie (Drui, Illouz, Do Cao and Caron; French Endocrine Society expert consensus) 2018 · American Thyroid Association (Thyroid; Jonklaas et al.) 2014
| Recommendation | Grade, as printed | Source |
|---|---|---|
| Reassess TSH when a tyrosine kinase inhibitor is started in a person on levothyroxine | Strong recommendation. Low-quality evidence. | American Thyroid Association (Thyroid; Jonklaas et al.) 2014 |
| Start levothyroxine as full or partial replacement titrated to TSH, with TSH 4 to 6 weeks after a dose change “Dose adjustments should be made when there are large changes in body weight, with aging, and with pregnancy, with thyrotropin assessment 4–6 weeks after any dosage change.” | Strong recommendation. Moderate quality of evidence. | American Thyroid Association (Thyroid; Jonklaas et al.) 2014 |
| Thyroid function at baseline and every 3 months on sunitinib (EU) “During sunitinib treatment, routine monitoring of thyroid function should be performed every 3 months.” | Regulatory product information; not graded | European Medicines Agency 2026 |
| Thyroid function at baseline, periodically and as clinically indicated on sunitinib (US) “Monitor thyroid function at baseline, periodically during treatment and as clinically indicated.” | Regulatory label; not graded | Pfizer, via DailyMed (US National Library of Medicine) 2026 |
| Thyroid function before and periodically throughout axitinib, treating to a euthyroid state “Monitor thyroid function before initiation of, and periodically throughout, treatment with INLYTA. Treat hypothyroidism and hyperthyroidism according to standard medical practice to maintain euthyroid state.” | Regulatory label; not graded | Pfizer, via DailyMed (US National Library of Medicine) 2026 |
| Thyroid tests at baseline, during pazopanib and as clinically indicated “Monitor thyroid tests at baseline, during treatment and as clinically indicated and manage hypothyroidism as appropriate.” | Regulatory label; not graded | Novartis, via DailyMed (US National Library of Medicine) 2025 |
| Cabozantinib: assess for signs, test as clinically indicated “Thyroid function testing and management of dysfunction should be performed as clinically indicated” | Regulatory label; not graded | Exelixis, via DailyMed (US National Library of Medicine) 2025 |
| Vandetanib: TSH at baseline, 2 to 4 weeks, 8 to 12 weeks, then every 3 months “Obtain Thyroid-stimulating hormone (TSH) at baseline, at 2 to 4 weeks and 8 to 12 weeks after starting treatment with CAPRELSA, and every 3 months thereafter.” | Regulatory label; not graded | Genzyme (Sanofi), via DailyMed (US National Library of Medicine) 2026 |
| Sorafenib: TSH monthly in differentiated thyroid cancer “Monitor TSH levels monthly and adjust thyroid replacement medication as needed in patients with DTC.” | Regulatory label; not graded | Bayer, via DailyMed (US National Library of Medicine) 2024 |
| Lenvatinib (EU): monitor before and periodically; in thyroid cancer adjust thyroid hormone to the TSH target “Hypothyroidism should be treated according to standard medical practice to maintain euthyroid state.” | Regulatory product information; not graded | European Medicines Agency 2026 |
| Fruquintinib: hypothyroidism in 31.5%, median onset 56 days, none leading to dose reduction or discontinuation (EU) “No events led to dose reduction or discontinuation.” | Regulatory product information; not graded | European Medicines Agency 2026 |
| The review's screening schedule is reasoned, not evidence based “Although these recommendations are not evidence based, they are logical given the reported complications and provide a framework for providers to identify and treat thyroid dysfunction before the appearance of symptoms.” | Narrative review; not graded | Journal of the National Cancer Institute (Hamnvik, Larsen and Marqusee); funded by the US National Institutes of Health 2011 |
| People on levothyroxine before a kinase inhibitor: TSH every 4 weeks, then every 2 months once stable “Once the TSH and dose of levothyroxine are stable, the frequency of monitoring can be decreased to every 2 months.” | Narrative review; not graded | Journal of the National Cancer Institute (Hamnvik, Larsen and Marqusee); funded by the US National Institutes of Health 2011 |
| Levothyroxine for symptomatic hypothyroidism or TSH above 10 mIU/L; the kinase inhibitor is not stopped | Expert opinion (French Endocrine Society), read as abstract; not graded | Annales d'Endocrinologie (Drui, Illouz, Do Cao and Caron; French Endocrine Society expert consensus) 2018 |
| Levothyroxine starting dose for adults under 65 with primary hypothyroidism and no cardiovascular disease “Consider starting levothyroxine at a dosage of 1.6 micrograms per kilogram of body weight per day (rounded to the nearest 25 micrograms) for adults under 65 with primary hypothyroidism and no history of cardiovascular disease.” | NICE guideline; not graded ('consider') | National Institute for Health and Care Excellence 2019 |
| Levothyroxine starting dose for adults aged 65 and over or with cardiovascular disease “Consider starting levothyroxine at a dosage of 25 to 50 micrograms per day with titration for adults aged 65 and over and adults with a history of cardiovascular disease.” | NICE guideline; not graded ('consider') | National Institute for Health and Care Excellence 2019 |
| Starting dose in hypothyroidism from cancer medicines “In patients with coronary artery disease, a lower dose such as 50 μg/d may be used for the first few weeks.” | Narrative review; not graded | Journal of the National Cancer Institute (Hamnvik, Larsen and Marqusee); funded by the US National Institutes of Health 2011 |
| Regorafenib (EU): biochemical and metabolic monitoring in general; no thyroid-specific schedule “Overall approximately 4.6% of patients treated with Stivarga developed hypothyroidism requiring hormonal replacement treatment.” | Regulatory product information; not graded | European Medicines Agency 2026 |
Shares Axitinib, Sorafenib, Sunitinib, Cabozantinib.
Shares Axitinib, Pazopanib, Sunitinib, Cabozantinib.
Shares Axitinib, Cabozantinib, Lenvatinib.
Shares Axitinib, Lenvatinib.
Shares Levothyroxine, Thyroid inflammation and underactive thyroid from checkpoint inhibitors.