Aligning treatment to physiology: when to use chronotherapy, block and replace, or titration to treat hypercortisolism

Abstract

Medical treatment of Cushing syndrome (CS) with steroidogenesis inhibitors or ACTH-directed agents is typically given using a titration strategy to achieve “normal” exposure to cortisol, as determined by a 24-hour urine free cortisol (UFC) measurement. This approach generally achieves similar serum cortisol levels across the day and fails to provide a normal diurnal pattern of cortisol. Chronotherapy addresses this problem through weighted afternoon/evening dosing of cortisol-lowering drugs. Additionally, because of day-to-day variability of cortisol production, patients may be undertreated on days of higher cortisol production and overtreated on days of lower cortisol production. A “block and replace” strategy is an alternative approach in which consistent biochemical glucocorticoid deficiency is the goal, with added glucocorticoid replacement to provide physiologic diurnal exposure to cortisol.

Recently, the term mild autonomous cortisol secretion (MACS) has been used to describe patients without clinical features of CS who fail to suppress cortisol after 1 mg dexamethasone. MACS includes individuals with an adrenal mass(es) who may have hypertension, type 2 diabetes mellitus, and/or osteoporosis. The role of medical vs surgical treatment of MACS is currently debated.

We here review the advantages and disadvantages of each approach to medical treatment of CS and MACS, concluding that “block and replace” deserves wider consideration, especially in patients who have severe disease, and to avoid adrenal insufficiency in those with highly variable UFC, or are taking agents that may cause adrenal insufficiency. Conversely, in those with mild disease, chronotherapy normalizes nighttime cortisol exposure while reducing drug burden.

The diagnosis of Cushing syndrome (CS) requires both clinical features (weight gain, hypertension, hyperglycemia, fractures/decreased bone density, immune compromise, thrombosis, psychiatric and cognitive changes, and/or hypogonadism) and biochemical features of hypercortisolism (increased 24-hour urine free cortisol [UFC], increased bedtime salivary cortisol, or lack of normal regulation of cortisol production as shown by lack of cortisol suppression after 1 mg dexamethasone). Current diagnostic guidelines recommend documentation of at least 2 biochemical abnormalities and recognize that clinical features accumulate and intensify over time, leading to poor quality of life (QOL), and increased mortality [1].

Treatment guidelines recommend resection of the causal tissue(s) as the optimal treatment of CS [2]. Successful surgery unmasks hypothalamic-pituitary-adrenal (HPA) axis suppression caused by chronic hypercortisolism. The immediate post-surgical period is characterized by glucocorticoid deficiency requiring “physiologic” glucocorticoid replacement for a median of 15 to 30 months [3-5]. Although no glucocorticoid replacement regimen achieves truly physiologic cortisol rhythms, an appropriate postoperative dose does not suppress the HPA axis. This allows for recovery of physiologic ultradian and circadian cortisol rhythms and improvement or normalization of QOL and associated comorbidities [2, 6]. Bilateral adrenalectomy is the obvious exception to this recovery.

Despite the high remission rate of Cushing disease (CD) after pituitary microsurgery in experienced centers, 20% to 40% of patients experience recurrent hypercortisolism within 10 years. Additionally, complete resection may be impossible in CD with cavernous sinus invasion or in ectopic ACTH secretion with occult or metastatic tumor. Accordingly, effective pharmacotherapy for the treatment of neoplastic hypercortisolism is critical. Current options include reducing cortisol production by inhibiting zona fasciculata steroidogenesis (osilodrostat, levoketoconazole [both FDA approved; osilodrostat also EMA approved], metyrapone, and ketoconazole [both EMA approved]) or by decreasing ACTH production in CD (pasireotide [FDA and EMA approved] and cabergoline); or antagonizing cortisol binding to the glucocorticoid receptor (mifepristone [FDA approved for hyperglycemia in CS]) [7].

Recent publications and our own experience highlight problems with the currently available medical modalities and dose titration strategies; conversely, alternative dosing strategies suggest how to improve current practice. Collectively, we have treated more than 3000 patients with mild to life-threatening hypercortisolism as part of clinical care or treatment trials [8-16]. Thus, our perspective on ways to improve medical treatment of CS derives from a large historical experience that is informed further by recently published work.

The entity of mild autonomous cortisol secretion (MACS) has been studied for years under the moniker of “subclinical CS” in the setting of an incidentally found adrenal mass. From a clinical perspective, this condition may cause or exacerbate obesity, hypertension, and diabetes, but does not present any specific clinical signs (such as muscle wasting, bruising, or striae) and rarely progresses to overt CS [3, 17]. From a pathophysiological standpoint, the main characteristic of MACS is cortisol secretory autonomy (ie, independent of physiological regulation by ACTH), illustrated by an abnormal response to 1 mg of dexamethasone associated with a slight increase in cortisol secretion, as shown by the suppression of ACTH and DHEAS, while salivary cortisol at bedtime and UFC may be normal [17]. The available evidence suggests that adrenalectomy may improve cardiometabolic comorbidities, specifically hypertension, in MACS [18]. However, current recommendations acknowledge the limitations of these studies and suggest that treatment should be individualized, either with enhanced management of co-morbidities or adrenalectomy, finding that medical therapy has not been validated sufficiently [17].

Goals of medical treatment of CS

All treatments of CS aim to restore physiologic ultradian and circadian cortisol rhythms that provide normal cortisol exposure and lead to improvement or normalization of QOL and associated comorbidities. As mentioned above, this is most closely approximated by surgical resection of the causal tissue. As will be discussed, because current technology cannot evaluate whether tissue exposure to free cortisol is “normal”, we rely on surrogate biochemical markers. The alleviation or persistence of diminished QOL and comorbidities are unmistakable endpoints of success that should be used to judge the efficacy of medical treatments.

Current dose-finding strategies in CS

Dose-finding with a titration strategy is the most common approach. It aims to gradually increase the dose of medication until “normal” total daily cortisol exposure is achieved based on UFC normalization. One alternative, a block and replace (B&R) strategy, aims to achieve undetectable or very low serum cortisol values and to “add back” glucocorticoid replacement. This approach recapitulates the low-cortisol postoperative state. Here, in contrast to “adrenal insufficiency” (AI), characterized by clinical symptoms, we use the term “biochemical glucocorticoid deficiency”. This emphasizes that a B&R approach prevents the development of symptomatic adrenal insufficiency and adrenal crisis, assuming that adequate add-back glucocorticoid is taken and absorbed, and sick-day rules are followed. Importantly, although use of short-acting hydrocortisone or prednisolone does not mimic perfectly the circadian cortisol rhythm, it achieves physiologically appropriate low nocturnal glucocorticoid exposure.

A more recently introduced dosing alternative, termed “chronotherapy”, uses short-acting steroidogenesis inhibitors to achieve normal evening and post-bedtime cortisol levels [19]. Theoretically, this allows cortisol production to increase before wakening, avoiding AI and achieving normal 24-hour integrated cortisol exposure.

In contrast to these strategies, glucocorticoid receptor antagonists prevent cortisol action by competing for receptor binding and preventing the downstream effects of cortisol. Here, because drug action is not reflected in cortisol measurements, only improvement in cortisol-mediated effects such as diabetes, hypertension, and obesity can be used as (surrogate) endpoints for efficacy.

Factors that influence successful medical therapy

Three main factors influence whether target cortisol levels are achieved and the time required to achieve this goal. Higher initiation and adjustment doses reduce the number of dose increments needed, and the time to goal, as demonstrated for ketoconazole, metyrapone, and osilodrostat [7, 10, 12, 16]. The interval between dose increments is another time-related factor. If one assumes that a full effect is achieved after 5 half-lives of a biologically active drug, agents with shorter half-lives (ie, the steroidogenesis inhibitors) could be titrated as often as every 2 days [10]. None of the recent prospective trials tested this idea, as UFC was the primary endpoint, generally with dose increases at 2 to 3 week intervals. This variable interacts with the first one in that higher initiation and adjustment doses increase the risk of AI, and (theoretically) can only work well if the interval to efficacy assessment is short, to avoid AI and to rapidly up-titrate if needed. The final variable is the intrinsic maximal potency of the agent. In this regard, pasireotide and cabergoline seem least, while osilodrostat seems most potent. As ketoconazole analog dosing is generally capped to avoid hepatotoxicity, its maximal potency is unknown.

Fallibility of current biomarkers of treatment success

Dose titration with UFC normalization is required by current FDA guidance for development of medical treatments of CS [20] and is endorsed in guideline and consensus statements [2, 21]. Unfortunately, UFC measurement is a suboptimal surrogate marker for clinical and biochemical normalization and under- or over-treatment.

Post-hoc analyses, chart review, and Delphi-review panel reports suggest that clinical improvement with medical treatment is inferior to that seen with surgical results, even when UFC is normal [11, 22-28]. Additionally, since most patients with MACS (and 8% with CD [15] have normal UFC, it cannot serve as a marker of treatment success for all patients.

In contrast to a bedtime cortisol value, UFC often cannot detect recurrent or early CS; thus, it cannot detect mild undertreatment with cortisol-lowering medication [28, 29]. This failure to detect undertreatment likely occurs because abnormal serum cortisol levels are not sufficiently high to exceed corticosteroid-binding globulin (CBG) capacity and increase renal excretion. Furthermore, UFC cannot detect AI [30], a risk with any medical treatment, and provides no information on the cortisol diurnal rhythm.

UFC has a large day-to-day variability in patients with CD; in one study of 152 patients, the intra-patient coefficient of variation for the mean of 4 samples was 52% (95% CI: 48 to 56) [31]. Therefore, an invariant dose titration schedule may result in eucortisolism, AI, or hypercortisolism on different days, and may contribute to the poor association between “normal” UFC and clinical normalization. Finally, UFC relies on the patient's ability to collect a complete sample, which is burdensome for all patients and may be impossible for those with cognitive impairments.

From a safety perspective, UFC results are usually not available for several days, which postpones each potential UFC-based dose titration and increases the risk of prolonged hypercortisolism.

Successful B&R dose-finding requires near-nil endogenous cortisol production. In this scenario, UFC continues to have the inherent disadvantages noted above but does detect inadequate dosing in both B&R and chronotherapy. Additionally, hydrocortisone add-back in B&R must be individualized with divided doses, ideally based on weight or body surface area [32], to avoid nonphysiologic excursions that exceed CBG capacity, increase UFC, and falsely suggest over-replacement [33].

Alternative biomarkers of treatment success

Morning serum cortisol is an attractive additional endpoint to monitor the efficacy of B&R and chronotherapy dosing. Results are often available on the same or next day, allowing for rapid dose adjustment. A morning cortisol can detect biochemical glucocorticoid deficiency or predict its occurrence if values are dropping, allowing for prompt institution of glucocorticoid replacement in B&R. Excessively high values during evening chronotherapy may require use of a longer-acting agent or institution of early morning doses.

What are the target morning cortisol values for each dosing strategy? When using a titration or chronotherapy strategy, a target morning cortisol value at the higher end of the normal 24-hour mean serum cortisol concentration avoids AI, because some diurnal variability usually persists in CS. In healthy young men, the mean (SD) daily cortisol level is ∼7.9 (2.1) mcg/dL (218 [58] nmol/L) measured by immunoassay [34]. Using this approach, a target titration morning cortisol level would be 8-12 mcg/dL (220-331 nmol/L). However, because a mid-range value might represent the nadir of a pulse, under-treatment (ie, persistent hypercortisolism) should be excluded by adjunctive UFC measurement. With the block and replace strategy, a low or undetectable serum cortisol value (prior to glucocorticoid dosing) represents adequate steroidogenesis inhibition, while higher levels indicate under-treatment.

Bedtime salivary cortisol is an important surrogate marker for all dosing regimens, based on an association of abnormal (increased) values with more severe comorbidities despite normal UFC [19, 23, 24, 28]. It follows that optimal improvement of clinical features may require normalization of the physiologic sleep-entrained cortisol nadir.

The usual titration schedule, with evenly divided doses over the day, provides a similar drug effect over 24 hours (Fig. 1). Because people with CS have high bedtime cortisol, doses that avoid morning AI are unlikely to normalize the bedtime value. Hence, evening chronotherapy may be needed, as has been explored in MACS [35]. Similarly, osilodrostat chronotherapy, giving the entire daily dose in the evening, restored the physiologic sleep-entrained cortisol nadir and improved QOL and sleep [19]. While these studies only used evening doses, titration could be modified by “weighted chronotherapy” in which higher doses are given late in the day. In theory, with titration, only drugs with short half-lives, such as ketoconazole, metyrapone, or osilodrostat, would achieve this desirable outcome. By contrast, B&R achieves low nocturnal glucocorticoid exposure with morning administration of a short-acting glucocorticoid such as hydrocortisone (or prednisolone).

Figure 1.

Medical treatment approaches in Cushing syndrome. The graphs show cortisol patterns in healthy individuals who sleep at night (A), in Cushing syndrome (CS) patients (B), and in CS patients receiving medical therapy by titration (C), chronotherapy (D), block of cortisol production (E), and “block and replace” with hydrocortisone add-back. (A) The dashed line approximates the physiologic diurnal variation in serum cortisol, which is released in a pulsatile manner. A nadir occurs shortly after sleep onset. The solid line represents mean cortisol. (B) In Cushing syndrome, although cortisol pulsatility persists, the sleep-entrained nadir is diminished or lost. (C) In a titration strategy, medication is dosed evenly during 24-hours (arrows), reducing the mean cortisol (solid line) by a similar amount after each dose throughout the day. (D) In patients with milder disease, with a morning cortisol within target range (8 to 12 mcg/dL), chronotherapy can be used with weighted dosing, aiming to only reduce evening and overnight cortisol. The largest dose (wider arrow) is given in the early evening to obtain a physiologic bedtime nadir of cortisol. A smaller later dose allows for increasing cortisol levels at the usual time. (E–F) In a block-and-replace strategy, larger doses (E, arrows) are given throughout the day to achieve consistent biochemical glucocorticoid deficiency. Glucocorticoid replacement, eg, hydrocortisone (F, solid arrows), is given in a larger dose upon awakening and a smaller dose in the afternoon to approximate the physiologic cortisol rhythm (solid line).

The risk of adrenal insufficiency (AI)

Assuming patient compliance with all dosing strategies, the risk of AI is greatest using a titration approach. It is difficult to compare AI rates for various medications because they are defined differently in their clinical trials. In general, symptoms consistent with the diagnosis occurred frequently, though investigator-designated frank AI was less common. Among trials of levoketoconazole, metyrapone, osilodrostat, and mifepristone, rates of AI/nausea were 10%/29% [36], 24%/48% [9], 54%/45% [13, 37], and 4%/48% [11], respectively. Fatigue, nausea, and AI were among the most frequent events in the FDA Adverse Event Reporting System for osilodrostat [38]. The high rates of nausea in each study, and the possibility that this was a component of AI or a precursor to adrenal crisis, underline the importance of patient education about AI and sick day rules.

The high rate of investigator-assessed AI or milder hypocortisolism during osilodrostat administration in the LINC3 study, 67/137 (54%) patients, led to study discontinuation in 5 and initiation of glucocorticoid therapy (ie, B&R) in 31 participants [13]. This study included randomization to a placebo after UFC normalization. During the 8-week placebo phase, 10 of 34 participants continued to have normal UFC. During the extension phase, 26 of 98 participants unexpectedly developed AI while taking a stable dose. Thus, despite a half-life of 4 hours, osilodrostat had persistent effects after discontinuation or enhanced effects during stable dosing in about 30% of patients.

Similar findings were reported in other studies. In a meta-analysis of 37 patients with ACTH-dependent CS receiving osilodrostat, 12 of 22 who started with a titration strategy were switched to B&R; ultimately, 73% received B&R [39]. Another retrospective study of 33 patients with ectopic ACTH secretion reported that 6 received titration only while 11 others began with or switched to B&R [12]. In other reports, AI persisted after treatment discontinuation [40, 41] and occurred despite stable dosing [39, 40]. A “cytotoxic” effect of osilodrostat has been postulated, based on the shrinkage of hyperplastic adrenal glands in eleven reported patients [39, 42]. Other factors, such as ethnicity, may play a role: in a pooled analysis of the LINC3 and LINC4 studies, AI was more common in Asian (59%) than non-Asian (40%) participants [43]. Based on this collective experience, a B&R approach is now recommended by some authors to avoid AI during osilodrostat treatment [44, 45].

Atumelnant, an orally administered nonpeptide small molecule melanocortin 2 receptor (MC2R) antagonist (also referred to as ACTH antagonist), is under development in a Phase 1b/2a trial. Preliminary data from that study showed that 6 participants with ACTH-dependent CS promptly developed biochemical glucocorticoid deficiency after a median of 2 once-daily doses, and received hydrocortisone add-back treatment [46].

Successful B&R induces “biochemical glucocorticoid deficiency” that will manifest as clinical AI if patients do not receive appropriate glucocorticoid replacement, including a lack of increase in hydrocortisone dosage during intercurrent “sick day” events. Because of this ubiquitous risk, any cognitive limitation for adherence or lack of understanding of AI may be a relative contraindication to B&R. This also applies to hydrocortisone use during sick days in patients treated by a titrated dosage regimen. In our experience, patients who have never taken glucocorticoids may not comply with their use on sick days.

Unmet needs and future research directions in medical treatment

One fundamental need stems from the lack of published data on the safety and efficacy of B&R treatment. Prospective trials may be ideal, but even retrospective data would be very helpful. Current clinical practice and regulatory requirements for the use of UFC as a primary endpoint likely account for this deficit.

The holy grail of CS treatment is to achieve normal physiologic exposure to cortisol. Readily available technology evaluates integrated daily exposure (UFC) and a few timepoints (morning, evening, or multiple timed serum or salivary levels). Theoretically, frequent measurement of free interstitial fluid cortisol levels over 24 hours provides the most clinically relevant biomarker. However, both interstitial fluid and frequent serum measurements are labor-intensive, expensive, and only available in research centers. Happily, many investigators are developing real-time cortisol sensors analogous to continuous glucose monitors [47], which may eventually improve efficacy endpoints. While not widely available, measurement of hair cortisol detects over-treatment with hydrocortisone and under-treatment with medical therapy when UFC is normal [28]. Increased availability of these technologies may help to fine-tune therapeutic approaches. In the meantime, reconsideration of UFC as the only endpoint of pharmaceutical registration trials, and incorporation of additional serum and/or salivary timed endpoints, is needed in clinical trials.

A related drawback is the lack of information about the cortisol pharmacodynamic (PD) profile of the medications used. Regrettably, measurements of pre and postdose cortisol are rarely obtained. A single osilodrostat PD study showed a maximal effect 1 hour after dosing. Although a dose-related effect was detectable at 12 hours, the group size was too small to determine the reason(s) for a very high PD variability [48]. Based on other lines of evidence referenced herein, the cortisol-lowering effects of osilodrostat may be much longer than its 4 hour half-life [49]. There also is a hypothetical mismatch between dosing interval and PD efficacy of ketoconazole and levoketoconazole, given their half-lives of 3 to 10 hours and 4 to 6 hours, with a risk of escape of efficacy with twice daily dosing. Incorporation of PK and PD data into future studies will help to inform optimal use of these agents.

A more physiologic cortisol delivery system is needed for B&R approaches. While the main advantage of B&R over titration is the ability to normalize evening cortisol, its main drawback is its inability to restore waking physiological values, which can result in an impaired QOL. From this perspective, the delayed-release hydrocortisone preparations are the most advanced approach to this problem. Subcutaneous cortisol delivery via an insulin pump may better approximate physiologic levels, but PK effects have not been adequately studied. Ideally, a system that delivers pulses is needed to achieve normal cortisol input into gene transcription [50].

The foregoing deficits relate to optimization of biomarker measurements. These lacunae in knowledge need to be filled so that important clinical outcomes can be assessed. For example, there are few data on the relationship between normalization of bedtime/nocturnal cortisol levels and improved clinical features and quality of life, outcomes that determine patient satisfaction and reduce mortality. However, a prospective comparative study is probably unfeasible.

Steroidogenesis inhibitors cause dose-dependent reduction (ketoconazole and levoketoconazole) or ACTH-driven accumulation of precursors (metyrapone and osilodrostat) with possible risks of male hypogonadism or hypertension and female hirsutism, respectively. To the best of our knowledge, no study has compared adverse events in patients undergoing titration vs B&R. As some adverse effects during titration may relate to low cortisol levels (eg, nausea), these might decrease with B&R.

Finally, there are few studies, even retrospective ones, on the incidence of AI in patients treated with titration vs B&R strategies. This deficit is compounded by confusion regarding the definition of AI, and whether its components (fatigue, nausea, diarrhea, etc.) are intrinsic reactions to the drugs, or whether they represent AI or the glucocorticoid withdrawal syndrome when this occurs shortly after a drop in cortisol levels. Unfortunately, there are few data on adverse events in healthy volunteers at the doses used in patients. This might be helpful, as healthy volunteers might mount an ACTH response and thus only experience events inherent to the medication itself, unrelated to AI. Data on this important outcome will be invaluable in establishing the safety of each approach.

Recommendations for clinical care

Regardless of the dosing strategy and choice of biochemical target, clinicians should continue to treat CS comorbidities. Ideally, evaluation of cortisol exposure includes measurement of UFC and both morning [30] and bedtime serum/salivary levels (or a morning-to-bedtime “day curve”) to confirm that a purportedly effective dose is reached.

The choice of specific agent(s) must consider the patient's wishes and ability to adhere to the dosing and monitoring schedule, as well as interactions with other medications and the potential to worsen existing conditions or create new problems (hirsutism, hepatotoxicity, hypokalemia, hypogonadism, QT prolongation, diabetes, or mania), as elegantly reviewed recently [7]. The choice of dose schedule strategy, whether titration, B&R, or chronotherapy, should consider the advantages and disadvantages of each, as summarized in Table 1.

Table 1.

Advantages and disadvantages of titration, block & replace (B&R), and chronotherapy dosing strategies, and scenarios in which one might be chosen preferentially over another

Variable	Titration	B&R	Chronotherapy
Requirement for patient education	All patients require education about the adverse effects of the drug and adrenal insufficiency (AI).	All patients require education about the adverse effects of the drug and AI. Lack of demonstrated understanding and/or adherence to regimen is a contraindication because of AI risk.	All patients require education about the adverse effects of the drug and AI. AI is less likely when a short-acting steroidogenesis inhibitor is given only at night.
Time to cortisol control	Usually longer, because urine free cortisol (UFC) is used for dose adjustment, unless a high starting dose is chosen.	Usually shorter, especially if a high starting dose is chosen and AM cortisol is used for dose adjustment.	Varies, depending on dose; often longer than B&R because salivary cortisol +/− UFC values are used to adjust dose.
Adequacy of cortisol control	Post-hoc analyses, chart review, and Delphi-review panel reports suggest that clinical improvement with medical treatment is inferior to that seen with surgical results, even when UFC is normal [11, 22-28].	With near-nil AM cortisol, overall cortisol exposure is likely “controlled”. However, one study suggests that conventional replacement is not physiologic [28].	Insufficient data; UFC cannot be used if it is already normal.
Normalization of LNSC and nocturnal cortisol levels	Yes, with a risk of adrenal insufficiency in those with a relatively invariant diurnal rhythm, unless chronotherapy is used.	Yes, with a risk of insufficient afternoon/early evening cortisol exposure and fatigue, unless the replacement glucocorticoid is optimized. May not achieve if the hydrocortisone dose is too high or given late in the day.	Achievable.
Improved QOL	Some studies show improvement [9, 22, 24]	One study with mixed B&R and titration showed improvement [24].	Suggestion of improvement in one study [19].
Cost	For any given patient, likely to be lower compared to B&R.	Highest - complete blockade requires higher doses, with the additional cost of glucocorticoid replacement.	Lower than titration if only afternoon and/or evening doses are given. May be higher than titration if additional nighttime doses are needed to achieve normal bedtime salivary cortisol levels.
When to consider use	Mild hypercortisolism, relatively preserved diurnal rhythm, no urgency to control cortisol.	Severe hypercortisolism, urgency to control cortisol, large variability in UFC, use of osilodrostat or atumelnant, inability to monitor (eg, during pandemic).	Mild hypercortisolism, mild autonomous cortisol secretion (MACS).

We recommend B&R in patients with a higher risk of AI, such as when treating with osilodrostat [7, 45] or the novel ACTH antagonist atumelnant [46]. This approach also deserves consideration in patients with high UFC variability. To evaluate the latter need, measurement of multiple UFC concentrations during the initial evaluation of CS can help guide treatment. B&R also has advantages if monitoring is difficult, whether because of distance, affordability, or other restrictions (as illustrated by its safe use during the COVID-19 pandemic [51]). Additionally, in patients with moderate to severe clinical manifestations, rapid dose escalation does not risk AI. A mid-range initial dose of a steroidogenesis inhibitor with adjustments every 3 to 5 days based on morning cortisol levels is effective in this setting [10, 12, 52, 53].

Concerns with B&R treatment focus on the possibility of AI and a higher dose requirement (compared to titration), with potential increases in adverse events and cost. Ideally, the cortisol-suppressing agent should be sufficiently powerful with a relatively long duration of action and a minimum number of doses. Although routine use of glucocorticoid replacement greatly reduces the risk of AI compared to that associated with dose titration, comprehensive patient education regarding sick day management and compliance is mandatory with either approach. By analogy with bilateral adrenalectomy, theoretically, B&R may increase the risk of corticotrope tumor progression. Until this is better understood, periodic pituitary MRI may be warranted during long-term B&R therapy.

A titration scheme may be both fully successful and safer than B&R in patients with mild disease, in whom the bedtime cortisol is not severely deranged, and who have relatively invariant day-to-day UFC values. In these patients, there is less concern for swift control of hypercortisolism, and slow titration has the potential advantage of decreasing the risk of AI. In patients with mild disease, it may also be possible to achieve the aforementioned goals using chronotherapy with only afternoon and/or evening dose administration, or with increased evening doses of short-acting steroidogenesis inhibitors. These weighted approaches have the advantage of less risk for AI and potentially lower cost, as the drug is preferentially given when cortisol is abnormally high.

By contrast, patients with more than mild hypercortisolism usually require around-the-clock dosing to lower cortisol throughout the day. Since titration often results in an abnormal (ie, high) evening cortisol level, an additional evening dose of a short-acting agent (using ketoconazole or metyrapone) or moving a portion or all of the morning osilodrostat dose to the evening may improve or normalize the bedtime cortisol level. Subjective improvement in sleep and demonstration of a normal bedtime salivary cortisol are useful endpoints for chronotherapy. Here, measurement of both UFC and a morning serum cortisol is needed to exclude continued hypercortisolism, or AI, respectively.

In conclusion, B&R and chronotherapy strategies for the management of hypercortisolism deserve wider consideration. Comparison studies in similar patient cohorts are needed to establish optimal care of CS patients unable to undergo surgery. Finally, we acknowledge that it is difficult to provide an algorithm of precisely how to treat these rare patients, as the infrastructure (assays and their turnaround times; available medications) and experience (balancing risk and benefit; rapidity of dose escalation) vary by institution and investigator. We hope that this perspective is helpful and encourage our colleagues to reach out to those with more experience for guidance or to refer a challenging patient.

Funding

This work was supported by ZIA DK075122 signifies the intramural research funding provided to Dr. Nieman by NIDDK.

Disclosures

HE declares no conflicts of interest. JWF has served as a consultant for Corcept Therapeutics, Recordati Rare Diseases, Inc., Crinetics Pharmaceuticals, and Neurocrine Biosciences. AT has received honorarium for conferences and boards from Recordati Rare Diseases, Inc., Esteve, Pfizer, Lundbeck and Crinetics Pharmaceuticals. LKN has received funding to the NIH from Crinetics Pharmaceuticals under a Clinical Trials Agreement.

Data availability

Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.