Carcinogenicity of hydrochlorothiazide, voriconazole, and tacrolimus

In November, 2024, a Working Group of 22 scientists from 14 countries met at the International Agency for Research on Cancer (IARC) in Lyon, France, to finalise their evaluation of the carcinogenicity of hydrochlorothiazide, voriconazole, and tacrolimus.

Hydrochlorothiazide was classified as “carcinogenic to humans” (Group 1) based on “sufficient” evidence for cancer in humans. There was also “sufficient” evidence for cancer in experimental animals and “limited” mechanistic evidence. Voriconazole was classified as “carcinogenic to humans” (Group 1) based on “sufficient” evidence for cancer in humans. There was also “strong” mechanistic evidence in exposed humans and human primary cells. Tacrolimus was classified as “carcinogenic to humans” (Group 1) based on “sufficient” evidence for cancer in humans, and the combination of “sufficient” evidence for cancer in experimental animals and “strong” mechanistic evidence in exposed humans. These assessments will be published in Volume 137 of the IARC Monographs.¹

The main exposure to hydrochlorothiazide, voriconazole, and tacrolimus is from medication; exposure via the environment or occupation is expected to be orders of magnitude lower than from medications.

Hydrochlorothiazide is the primary prescription-only oral thiazide diuretic used worldwide to treat essential hypertension and peripheral oedema. Although partly replaced by other antihypertensives, hydrochlorothiazide remains a commonly prescribed drug. In most countries, single pill combinations containing hydrochlorothiazide plus angiotensin receptor blockers have become the most-used hydrochlorothiazide-containing drugs. Hydrochlorothiazide is phototoxic due to the molecule’s interaction with ultraviolet radiation (UVR). In humans, it is readily absorbed through the gastrointestinal tract. Two metabolites have been identified, but not the metabolising enzymes. Excretion, primarily through the kidney, can be influenced by organ function, genetic variability, and sex.

There was “sufficient” evidence in humans that hydrochlorothiazide causes squamous cell carcinoma (SCC) of the skin and cancer of the lip. A meta-analysis conducted by the Working Group showed consistent and precise elevated risks of skin SCC in ever-users compared with: never users (eight studies, including a Danish population-based study²); users of other thiazides (three studies); and users of angiotensin-converting enzyme inhibitors or other non-diuretic antihypertensive medications (three studies, including a large US study³). A strong, consistent elevation in risk of cancer of the lip was noted among ever-users compared with never-users in the Working Group meta-analysis. Evidence of dose–response effects was noted both for skin SCC² and for lip cancer.⁴ The evidence was deemed “limited” for basal cell carcinoma and melanoma of skin, Merkel cell carcinoma, and malignant adnexal skin tumours, because of weaker associations or fewer studies.

The “sufficient” evidence of cancer in experimental animals was derived from one study that complied with Good Laboratory Practice (GLP) in male and female B6C3F1 mice:⁵ hydrochlorothiazide caused hepatocellular adenoma, and hepatocellular adenoma or carcinoma (combined) in males. In addition, an increase in the incidence of pheochromocytoma of the adrenal gland was observed in female F344 rats.⁶ The Working Group considered pheochromocytoma a relevant indicator of carcinogenicity, based on the rarity of these tumours, the observation of hyperplastic and neoplastic changes in the rat adrenal gland at an unusually young age, and the ambiguous behaviour of these tumours with respect to their malignancy, as described in the WHO Classification of Tumours.⁷ There was “limited” evidence that hydrochlorothiazide exhibits key characteristics of carcinogens in exposed humans and in experimental systems. In several instances, effects were observed in the presence of UVR at levels similar to typical sunlight exposure.

Voriconazole is a broad-spectrum triazole antifungal medication. It is used as curative treatment or prophylaxis for invasive aspergillosis and other serious fungal infections, which are especially common in transplant recipients. It is approved in most countries for use in the adult and paediatric population (age ≥2 years). Voriconazole is mainly available as oral and intravenous formulations. The drug is well absorbed after oral administration and is widely distributed to all tissues. It is extensively metabolised in the liver. The major circulating metabolite, voriconazole N-oxide, is devoid of antifungal activity but is phototoxic and slowly eliminated. Genetic variability of the metabolic enzyme CYP2C19, which results in a wide range of metabolic phenotypes, correlates with skin toxicity.

There was “sufficient” evidence that voriconazole causes skin SCC in humans. Several studies among transplant recipients showed a positive association between use of voriconazole and the incidence of cutaneous SCC.⁸ There were concerns about confounding by indication due to intensity of immunosuppression in the transplant recipients. However, using triangulation, the totality of evidence of cancer in humans strongly supports a causal interpretation. These lines of evidence included a monotonic increase in risk with increasing voriconazole exposure in the most robust study,⁸ evidence derived from gene–drug interaction that informs risk specific to the phototoxic metabolite voriconazole N-oxide,⁹ and the occurrence of SCC even in children exposed to the agent.

The mechanistic evidence for voriconazole was “strong” with UVR exposure for the key characteristics of carcinogens “induces oxidative stress” and “alters cell proliferation, cell death, or nutrient supply”. The phototoxic metabolite voriconazole N-oxide consistently enhances UVR-induced reactive oxygen species generation and oxidative damage to DNA in human primary keratinocytes and in a human skin equivalent model. This was coherent with an informative report involving many voriconazole-treated patients,¹⁰ in which actinic keratosis (a pre-cancerous lesion involving abnormal cell proliferation) developed in the specific areas of skin exposed to sunlight and, in some instances, regressed after treatment cessation. The evidence regarding cancer in experimental animals was “inadequate”.

Tacrolimus is a calcineurin inhibitor used as an immunosuppressive medication. It is administered as oral and intravenous formulations, alone or in combination with other drugs, to reduce the risk of rejection of solid organ transplants in adult and paediatric patients, and for prevention of graft-versus-host disease in stem cell transplant recipients. Topical tacrolimus is indicated as second-line therapy for treatment of atopic dermatitis and vitiligo.

There was “sufficient” evidence that tacrolimus causes non-Hodgkin lymphoma (NHL) and post-transplant lymphoproliferative disorder in humans. The most compelling evidence came from six studies¹¹ conducted in transplant-recipient populations, for which the comparator groups include individuals using other immunosuppressive medications, including ciclosporin, which is classified in Group 1 and causes NHL.¹² Therefore, the moderate, although imprecise, increase in risk of SCC across studies, and data from Working Group meta-analyses, supported a causal conclusion. The evidence for a causal association between tacrolimus and leukaemia and skin SCC was deemed “limited” due to fewer studies or weaker associations.

The “sufficient” evidence for cancer in experimental animals derived from one GLP study in male and female B6C3F1 mice:¹³ tacrolimus caused pleomorphic lymphoma in both sexes and undifferentiated lymphoma in females. There was “strong” mechanistic evidence that tacrolimus exhibits key characteristics of carcinogens. Tacrolimus was immunosuppressive in exposed humans, in human primary cells, and in experimental systems. There was evidence of immunosuppression in tacrolimus-treated transplant recipients and patients with autoimmune disease. Tacrolimus altered many of the immune cells and mechanisms involved in immune system control of tumour growth. The evidence was coherent in primary cells and experimental systems. Tacrolimus was genotoxic and induced oxidative stress in experimental systems.