Assessing benefits and risks is a daily activity for us all, with the outcome dependent on the perspective of the assessor. In the context of medical therapies and depending on the disease in question, the individual patient and physician may have some “risk tolerance,” whereas for regulators and manufacturers the interests of the general population are foremost concerns, with more cautious approaches predominating. These discrepant approaches are brought into sharp focus in the case of rare or uncommon diseases, especially those associated with high morbidity and mortality. Cushing's disease is an excellent example of such a situation where effective therapy is essential, but where the balance of risk and benefit must be carefully judged. Although transsphenoidal surgery by an expert remains the mainstay of initial attempts at definitive management (1), the long-term outcome of this procedure means that there remain significant numbers of patients with uncontrolled disease and a requirement for effective medical and other therapy. Adrenal steroidogenesis inhibitors have been the mainstay of medical therapy for Cushing's syndrome for many years (2). Worldwide, the imidazole fungicide ketoconazole has been used most widely “off label” for the treatment of the hypercortisolemia of Cushing's syndrome. There are better antifungal agents, and cases of severe hepatotoxicity and deaths associated with the use of oral ketoconazole have led to regulatory authorities on both sides of the Atlantic issuing warnings and the major manufacturer ceasing production. In the case of fungal infection, this benefit-to-risk assessment seems reasonable, but is this the case for Cushing's syndrome or disease? One of the problems with ketoconazole is the paucity of published data concerning its use in the context of Cushing's disease. Prospective clinical trials are to be highly valued, but with few recent exceptions these are not found in the Cushing's literature (3). Furthermore, given the cost and complexity of such trials, they are highly unlikely to be conducted in the future with old drugs used “off label” in this uncommon disease. It is in such studies that data are collected rigorously for both efficacy and side effects, and ketoconazole use in Cushing's syndrome has never been subjected to this rigor of assessment, which leaves the endocrine community reliant on retrospective data. Nevertheless, the paper by Castinetti et al (4) in this issue of the JCEM provides important insights. Reported are real-world “effectiveness” data on the use of ketoconazole as a sole medical agent for the treatment of hypercortisolemia in 200 patients with Cushing's disease, the largest number ever reported as being treated with this agent, observed over a 17-year period in 14 French tertiary centers. This is a period far longer than acceptable in the setting of a clinical trial, with multiple assays between centers and over time and with nonstandardized follow-up and clinical assessment, the inherent constraints of a retrospective and multicenter study. Although there may be an inherent reporting bias favoring a positive effect of an intervention in such a study design, reports such as these remain highly useful to assess how well a therapy actually works in practice and to allow more informed judgment of associated risks. This is especially the case for Cushing's disease where large prospective studies, complete patient acquisition registries, and retrospective analyses are few. Overall, on ketoconazole therapy 75% of patients experienced a > 50% fall in urinary free cortisol (UFC). The outcomes were further analyzed in three main instances: 1) when treatment was administered before transsphenoidal surgery; 2) when used as primary therapy because patients were either unfit for or refused pituitary surgery, or secondary therapy after pituitary surgery; and 3) when used long-term for more than 2 years. When used before transsphenoidal surgery in 40 patients for a mean period of 4 months, ketoconazole resulted in controlled UFC in 49%, partial control in 36%, and no control in 15%. Clinical improvements in diabetes, hypertension, and hypokalemia were noted despite the short duration of therapy with a mean dose of 585 mg/d. When considering a longer duration of therapy, of 160 patients treated for a mean of 24 months as either primary or secondary therapy, 50% were controlled, 23% were partially controlled, and 27% were not controlled. These data suggest that around one-third of the patients on long-term follow-up were inadequately treated, and that other medical therapies or other strategies might have been considered. The fact that physicians were titrating the dose upward is evidenced by an increase in the average dose of ketoconazole from 542 to 779 mg/d over this period. Nevertheless, 32 patients in the study who were tolerant of the drug but inadequately controlled were not on doses as high as 1200 mg/d, often felt to be around the maximal daily dose. It is important to note that this purported “maximum” has not been formally established. This is interesting because it suggests that despite ketoconazole being used in tertiary centers familiar with its use, just over 15% of patients might have received inadequate treatment due to physician factors, despite their physicians being experts. This emphasizes the fact that clinical protocols harmonizing best practice for medical therapies for Cushing's disease would likely lead to more effective therapy, and that treatment in nonspecialist centers is still more likely to result in less effective outcomes. Even longer treatment duration was observed in 51 patients on ketoconazole therapy for more than 2 years, and it is not surprising that the rate of control was higher in this cohort (64.5%) because they will have been selected to stay on therapy due to tolerability and efficacy. In general, control in men on treatment was not as good as women, despite similar doses and levels of UFC, and the reasons for this are unclear. Together, these data show ketoconazole to be effective in controlling hypercortisolemia in the majority over long periods with associated clinical improvement, but with some patients not receiving adequate dosing and probably missing out on the full benefit. Although the benefits appear good, what are the risks? Approximately 20% of patients had to stop treatment because of tolerability issues. Adrenal insufficiency, arguably a good effect if recognized and treated, occurred in a few. The major concern, however, is hepatic toxicity. Elevated liver transaminases are frequently found when patients are on oral ketoconazole. Castinetti et al (4) found that 30 patients had elevations in liver transaminases 5-fold normal, and four patients had elevations between 5- and 10-fold (18% of all patients treated). In all, the liver transaminases increased either on instigating therapy or on dose increases, with no further worsening after 4 weeks. There was no dose-effect relationship. Within 2 to 4 weeks of either cessation of therapy or dose reduction, the values returned to normal in all. In one patient, the values exceeded 40 times the upper limit of normal and recovered after cessation of treatment to normal after a period of over 3 months. In this patient, there was a history of high alcohol intake and use of acetaminophen. What do these data tell us? Severe liver injury is a major concern when using ketoconazole, but in this series all derangements of liver biochemistry were reversible. It is possible that the close monitoring by endocrinologists, well versed in this potential side effect of ketoconazole, mitigated against more severe outcomes. The recommendations by the authors for weekly liver biochemistry for the first 4 weeks after dose initiation and escalation seem reasonable, and even greater care needs to be taken in patients who might have other reasons for hepatic toxicity. Therefore, for a patient with Cushing's disease needing medical therapy, the potential benefits of ketoconazole therapy are likely to outweigh the risks, but careful monitoring is mandatory, and harm is certainly possible. Other effective therapies for Cushing's disease are available including metyrapone, pasireotide, cabergoline, mifepristone, and etomidate, with newer therapies in development such as LCI699, and these have been the subject of a recent editorial in the JCEM (5). Therefore, there is a choice that can be made in selection of therapy. It is likely that clinicians and patients will continue to want access to ketoconazole as therapy for Cushing's syndrome, accepting the associated risks. It seems equally likely, however, that regulatory warnings and the lack of supply from the manufacturer will be the major issues facing its use, which limits options for patients and clinicians dealing with this challenging disease. Disclosure Summary: J.N.-P. has received consultancy and research income from Novartis Pharmaceuticals and HRA Pharma. urinary free cortisol.
No takes yet. Share an insight, caveat, or question.
John Newell‐Price (2014) studied this question.
Synapse has enriched 5 related papers. Consider them for comparative context: