This protocol was approved by the Kobe University Hospital ethical committee (IRB No. 1351). Written informed consent was obtained from all patients. Six patients with CS (three with cortisol-secreting adrenocortical adenoma (A); n = 3 and three with ACTH-secreting pituitary adenoma (P); n = 3, Table 1) who had been hospitalized at Kobe University Hospital during the period from 2013 to 2016 were recruited in this study. All patients presented with symptoms of overt CS including moon faces, central obesity, and buffalo humps. The diagnosis of CS was based on general diagnostic criteria as previously described (12, 13). This study was performed during medical therapy to control hypercortisolism before surgery in most of the patients. Five of the six patients received metyrapone treatment for preoperative control of hypercortisolemia. The remaining patient (No. 4) received metyrapone for medical control of a residual ACTH-secreting pituitary adenoma. In patients who underwent surgery, the diagnosis of cortisol-secreting adrenocortical adenoma or ACTH-secreting pituitary adenoma was histologically confirmed.

Subjects rinsed mouth thoroughly with water 10 min before the sample is collected. They tilted the head forward, allowing the saliva to pool on the floor of the mouth, then passed the saliva into a polypropylene tube. Samples were centrifuged (1,500 × g, 15 min), the supernatant transferred into new tubes, and stored at −20°C until measurement. Sample collection was not set within 60 min after eating a major meal.

The salivary cortisol collection was performed five times in a day: at 6:00 (wake up time), 8:00 (before breakfast), 12:00 (before lunch), 18:00 (before dinner), and 22:00 (before sleep). A daily mean value (MV) and daily area under the curve (AUC) of salivary cortisol levels were calculated (MV: average of five measurements of salivary cortisol levels and AUC: integral of five measurements of salivary cortisol levels).

The samples were collected before metyrapone treatment, during metyrapone treatment, and when the dose of metyrapone was increased. The date of sampling was set for 2 days after the start or increase in the dose of metyrapone. The initial dose of metyrapone was 500–750 mg per day, and the maintenance dose was 500–2,000 mg per day. The administration timing was 14:00 and 22:00 for patients receiving the drug twice per day (500 mg per day); 6:00, 14:00, and 22:00 for patients receiving the drug three times per day (750–1,250 mg per day); and patients receiving the drug at 6:00, 12:00, 18:00, and 24:00 for four times per day (1,000–2,000 mg per day). The dosage and timing for each patient were determined based on the evaluation of the morning serum cortisol levels, the UFC levels, and the well-being of each patient. The target levels of morning serum cortisol and UFC were below the upper normal limit. Salivary cortisol measurements were performed, and its utility was analyzed retrospectively. In patient No. 6, salivary samples before treatment were not obtained. Therefore, only the changes in MV and AUC during treatment were analyzed in this patient.

Patient No. 4 was hospitalized again due to worsening of hypercortisolemia 1 year after the first hospitalization. To test whether the salivary cortisol is useful, we collected and measured multiple salivary samples from this patient during the treatment and adjusted the dose and timing of metyrapone administration as a prospective case study.

The serum cortisol levels were measured by an enzyme immunoassay (reference range: 6.4–21.0 µg/dL, EIA; TOSOH, Tokyo, Japan), and the daily UFC levels were measured by radioimmunoassay (reference range: 11.2–80.3 μg/day, RIA; TFB, Tokyo, Japan). Salivary cortisol levels were measured by enzyme immunoassay (Salimetrics Inc., State College, PA, USA).

All statistical analyses were performed using SAS 11.2 (Statistical Analysis Software release 11.2; SAS Institute Inc., Cary, NC, USA). Data are shown as mean ± SD. The paired t-test was used to compare the values of mean from two related samples. The correlation between non-parametric data was assessed with Spearman’s rank correlation. The cutoff for statistical significance was set at p < 0.05.

The clinical characteristics of the patients are described in Table 1. The patient ages ranged from 22 to 64 years. Five of six were female. The basal ACTH level was undetectable in all of the patients with an adrenocortical adenoma and was 32.6–179.0 pg/mL in the patients with an ACTH-secreting pituitary adenoma. The range of morning serum cortisol levels and UFC levels were 19.2–38.6 µg/dL and 97.3–1,980.0 μg/day, respectively. The serum cortisol level after a low-dose dexamethasone suppression test was 14.0–38.6 µg/dL. These data were consistent with an active state of CS.

Both morning serum cortisol and UFC levels were decreased within the normal range after initiation of treatment in all patients except for Patient No. 5 (Figure 1). Patient No. 5 was referred to neurosurgery before hormonal management achieved target levels.

In Patient No. 1, both the UFC and morning serum cortisol levels decreased after initiation of metyrapone treatment in a dose-dependent manner (Figure 1A). However, multiple measurements of salivary cortisol levels revealed relatively lower cortisol levels in the morning and higher levels in the daytime and night, indicating that the diurnal cortisol rhythm was still impaired in this patient (Figure 1A). Similar results were observed in Patients No. 3 and 4 (Figures 1C,D). In Patient No. 2 (Figure 1B), the serum cortisol and UFC levels did not change with 750 mg of metyrapone, while the MV and AUC of salivary cortisol levels decreased. Furthermore, when the dose was increased at night by 500 mg, morning cortisol levels were suppressed in both serum and saliva. However, the night salivary cortisol did not decrease. These results suggest that not only the dose but also the timing should be considered to reconstitute the diurnal rhythm.

In Patient No. 5, despite a marked reduction in UFC levels, the morning serum cortisol levels did not change with 1,250 mg of metyrapone. We then increased the dose to 2,000 mg, resulting in decrease in both morning serum cortisol and UFC levels, though still not up to target levels. This suggests that a further increase in the dose might be necessary. Salivary cortisol levels revealed that, even with the dose of metyrapone 2,000 mg, the CS remained uncontrolled, and obvious impairment of the diurnal cortisol rhythm remained (Figure 1E).

In Patient No. 6, UFC levels and morning serum cortisol levels were both normalized after metyrapone treatment. However, the diurnal rhythm was not fully reconstituted (Figure 1F).

Patient No. 4 was readmitted to the hospital 1 year after the first admission due to the worsening in hypercortisolemia despite a dose of 1,250 mg of metyrapone. During the second admission, although the dose of metyrapone was increased to 2,250 mg, the UFC level rather increased (Figure 2). According to the diurnal salivary cortisol level, it was observed that when administered four times a day, the efficacy did not last until next administration. Therefore, we shortened the dosing interval and increased the frequency from four (6:00, 12:00, 18:00, and 24:00) to five times per day (6:00, 10:00, 14:00, 18:00, and 22:00). As a result, the diurnal salivary cortisol pattern was obviously improved, despite the slight increase in the total daily dose (Figure 2). The UFC level also decreased from 211 to 110 μg/day, indicating that appropriate optimization had been obtained. Furthermore, insomnia due to hypercortisolemia was markedly improved. In this series, there were no adverse events that were related to metyrapone treatment.

Both the MV and AUC of salivary cortisol levels decreased after initiation of metyrapone treatment (Table 2). Metyrapone treatment suppressed diurnal salivary cortisol levels in a dose-dependent manner (Figures 1A–F). The effect of metyrapone on salivary cortisol levels was observed 2 h after administration of the drug (Figure 1G). Changes in both the MV and AUC of salivary cortisol levels were significantly correlated with changes in UFC levels (r_s = 0.75 and 0.70, p = 0.007 and 0.016, respectively), whereas changes in the morning serum cortisol levels were not (r_s = 0.42, p = 0.200; Table 3). On the other hand, changes in both the MV and AUC of salivary cortisol levels were not correlated with those in midnight salivary cortisol levels (Table 4).