Key words
hypercortisolism - Cushing’s disease - circadian rhythm - adrenocorticotropin - pituitary
- adrenal
Introduction
Endogenous hypercortisolism, also referred to as Cushing’s syndrome (CS) is a rare
but severe disease affecting approximately 0.2–5.0 per million people annually [1]. The diagnosis of CS can be challenging due to the variety of clinical signs and
symptoms. Standard screening tests for CS include the 1 mg low dose dexamethasone
suppression test (LDDST), late-night salivary cortisol, and the cortisol levels in
a 24-hour urine sample (urine free cortisol, UFC) [2]. All tests are quite sensitive and specific [2]
[3]. However, overall test accuracy is dependent on assays and pre-analytic conditions.
Also, due to the rareness of the disease, pre-test probability is low, and therefore,
despite high sensitivity and specificity, there can be a high rate of false-positive
and even false-negative results [4]. In the past decades, new diagnostic methods such as plasma steroid profiling, face
recognition, and clinical scores have been developed to increase the accuracy of diagnosis
and to differentiate subtypes [4]. However, none of the mentioned methods has yet been established in clinical routine.
Of the standard screening tests, late-night salivary cortisol seems to be the best
test for early detection of the recurrence of Cushing’s disease [5]
[6]. It can be performed easily on an outpatient basis. There causes for false negative
or positive results are only a few, such as smoking, gum bleeding, contamination with
steroid-containing ointments, or disturbed cortisol rhythms due to jet lag or shift
working [4]. However, there is a lack of knowledge about the use of the daily salivary cortisol
profile in diagnosing and subtyping CS. In a study including 100 healthy subjects
(both children and adults) and 150 patients with CS, cut-offs for saliva cortisol
at different points of time were established but salivary profiles in different subtypes
were not analyzed [7]. In their study, Castro et al. observed that salivary cortisol in patients with
CS (33 subjects) was elevated at 9:00 a.m., 5 p.m. and 11 p.m. Sensitivity can be
improved by combining the 1 mg low dose dexamethasone suppression test and measuring
salivary cortisol [8]. Similar results were observed by Laudat et al. [9].
From a pathophysiological perspective, it is known that patients having CS suffer
from a disturbed diurnal cortisol rhythm. This would make salivary cortisol measurements
an ideal method for subtype diagnosis between the pituitary, adrenal, and ectopic
forms. Although late-night salivary cortisol measurement is a firmly established screening
tool, no study has systematically analyzed the value of a five-point diurnal profile
in diagnosing CS and identifying Cushing subtypes. Based on this background, the current
study aimed to investigate daily salivary cortisol profiles in patients with ACTH-dependent
and ACTH-independent CS.
Material and Methods
This study was part of the prospective German Cushing’s registry founded in 2012.
Since then, 520 patients with suspected or confirmed CS were prospectively included
in the registry. The structure of the registry has been described previously [10]
[11]. Ethical approval was obtained (195–09) and all patients gave written informed consent.
All patients underwent a standardized clinical examination and a standardized biochemical
screening including the 1 mg-low-dose-dexamethasone-suppression-test, collection of
24-hours urine (for UFC), and a salivary cortisol day-profile. For the latter, patients
were asked to collect saliva at 8:00 and 12:00 a.m. and 4:00, 8:00, and 11:00 p.m.
Patients were carefully instructed for the saliva collection. The laboratory analyses
were performed at the Endocrine Laboratory of the Medizinische Klinik und Poliklinik
IV, LMU Klinikum. Salivary cortisol was measured using an automated chemiluminescence
immunoassay (CLIA, IDS-iSYS, Immunodiagnostic Systems, Boldon). Reference intervals
were only established for late-night salivary cortisol (normal:<1.5 ng/mL, grey zone:
1.5–2.2 ng/mL); reference intervals for the other time points have not been established
so far. After confirmation of CS, further subtyping including CRH-Test, the 8 mg-high-dose-dexamethasone-suppression-test,
imaging (sellar magnetic resonance imaging (MRI) or abdominal computed tomography
(CT)) and – if needed – inferior petrosal sinus sampling was conducted.
Final diagnosis and subtypes of CS were based on unambiguous surgical confirmation
of the underlying histopathological entity. The exclusion of CS was based on normal
screening tests and clinical observation, as described earlier [12]. The definition of autonomous cortisol secretion followed the European Society of
Endocrinology guideline definition [13].
Finally, CS was confirmed in 111 patients and excluded in 290 patients. In 39 patients
with an adrenal incidentaloma, the final diagnosis was autonomous cortisol secretion.
In the remaining patients, either other endocrine diseases, such as primary hyperaldosteronism,
were diagnosed or patients were lost-to-follow-up before diagnosis was confirmed or
excluded. In patients with the exclusion of CS, primary symptoms varied, such as obesity,
hypertension, or hirsutism. Patients with cyclic CS were also excluded from the cohort
(N=3). Of the prospectively observed patients with a confirmed exclusion or confirmation
of CS, salivary day profiles were available for 428 patients. These patients formed
the final study cohort.
Statistical analysis
For statistical analysis, SPSS 26 (IBM) was used. Results of standard biochemical
screening including UFC, LDDST, serum cortisol, and ACTH are shown as median and ranges.
As data did not show normal distribution, differences between groups were tested by
non-parametric tests (Mann-Whitney U test and Kruskal-Wallis test). P-values<0.05
were considered of statistical significance. Discriminant analysis with stepwise variable
selection (saliva cortisol at different time points) was used to test which combination
of saliva cortisol helps to discriminate the subtypes. The results from these analyses
were compared with those of single late-night salivary cortisol determination.
Results
Patient characteristics
The biochemical characteristics of the patients are shown in [Table 1] with 68% of patients having pituitary CS, 24% adrenal, and 8% were diagnosed with
ectopic CS.
Table 1Biochemical characteristics of the patients included in the study. (CS: Cushing’s
syndrome, LDDST = 1mg-dexamethasone-suppression-test. UFC: urinary free cortisol).
|
CS confirmed (N=111)
|
Autonomous cortisol secretion (N=39)
|
CS excluded (N=278)
|
p-values (all groups)
|
|
Sex
|
74% females
|
80% females
|
77% females
|
–
|
|
Age (years)
|
46 (±) 14
|
63 (±) 8
|
39 (±) 17
|
p≤0.001
|
|
LDDST (µg/dL), Normal:<1.8 µg/dL
|
13.9 (7.1–22.4)
|
3.0 (2.2–5.5)
|
1.0 (0.8–1.3)
|
p≤0.001
|
|
UFC (µg/24 h), Normal:<150 µg/24 h
|
439 (233–787)
|
157 (90–231)
|
114 (77–188)
|
p≤0.001
|
|
ACTH (pg/mL), Normal: 4–50 pg/mL
|
44 (10–77)
-
pituitary: 59 (36–83)
-
ectopic: 104 (52–221)
-
adrenal: 3 (2–5)
|
6.0 (3.0–9.0)
|
13 (9–19)
|
p≤0.001
|
|
Morning Serum Cortisol (µg/dL), Normal: 1.8–24 µg/dL
|
21.8 (15.6–27.5)
|
8.9 (6.4–12.3)
|
8.9 (6.6–12.9)
|
p≤0.001
|
Salivary profiles in patients with or without Cushing’s syndrome
Mean salivary cortisol concentrations were significantly higher in patients with CS
at all five time points compared to patients in whom CS had been excluded ([Table 2]). Overall, ranges were greater in patients with CS than in the control group ([Fig. 1]). The use of the entire profile improved the diagnostic accuracy slightly compared
to a single late-night salivary cortisol (85% of cases correctly classified using
only late-night salivary cortisol vs. 91% correctly classified using the entire profile).
If additionally, the difference between morning and evening salivary cortisol was
considered, accuracy improved further (92% correctly classified). Combining salivary
cortisol profiles with UFC and low-dose-dexamethasone-suppression-test did not improve
accuracy further (92% correctly classified) and was only slightly better than the
combination of all three screening tests (91% correctly classified).
Fig. 1
a and b: Cortisol day profiles in all subgroups. CS = Cushing’s syndrome. ° * = statistical
outliers.
Table 2 Daily salivary profiles of patients with or without Cushing’s syndrome (CS).
|
8:00 a.m.
|
12:00 a.m.
|
4:00 p.m.
|
8:00 p.m.
|
11:00 p.m.
|
|
CS confirmed (N=111)
|
10.3 (5.2–18.4)
|
9.2 (5.1–15.9)
|
8.6 (5.0–13.9)
|
8.7 (5.0–13.6)
|
7.6 (4.1–11.8)
|
|
CS excluded (N=278)
|
6.5 (4.4–9.7)
|
3.0 (2.0–5.0)
|
2.4 (1.7–3.4)
|
1.3 (0.9–2.1)
|
1.0 (0.7–1.7)
|
|
Autonomous cortisol secretion (N=39)
|
7.6 (4.2–10.7)
|
3.6 (2.6–6.0)
|
3.2 (2.2–5.0)
|
2.4 (1.7–4.0)
|
2.6 (1.6–3.6)
|
|
p-values (all groups)
|
p≤0.001
|
p≤0.001
|
p≤0.001
|
p≤0.001
|
p≤0.001
|
|
CS vs. CS Excluded
|
p≤0.001
|
p≤0.001
|
p≤0.001
|
p≤0.001
|
p≤0.001
|
|
CS vs. Autonomous cortisol secretion
|
p=0.04
|
p≤0.001
|
p≤0.001
|
p≤0.001
|
p=0.001
|
|
Excluded vs. Autonomous
|
p=1.0
|
p=0.2
|
p=0.01
|
p≤0.001
|
p≤0.001
|
On average, late-night salivary cortisol levels were 85% (pituitary CS 83%, adrenal
CS 97%, ectopic CS 61%) of the morning cortisol. Subjects in whom CS was excluded,
had low relative late-night cortisol levels (26% of the 8:00 a.m. salivary cortisol
concentration); whereas patients with autonomous cortisol secretion had significantly
lower salivary cortisol levels compared to patients with CS at all time points but
salivary cortisol in the afternoon and evening was significantly higher than in subjects
with the exclusion of CS (p≤0.001), and relative late-night salivary cortisol levels
were 51% of the morning levels.
Salivary profiles in the different subtypes
Salivary cortisol was higher in patients with pituitary CS and ectopic CS compared
to those with adrenal CS. The difference between patients with ectopic and adrenal
CS was significant in the morning (pituitary vs. adrenal CS) and at all time points
except for noon.
Patients with pituitary and adrenal disease typically had constantly high cortisol
levels during the entire day. In patients with ectopic CS, cortisol levels were especially
high without any circadian rhythm; compared to patients with pituitary CS, salivary
cortisol was significantly higher in the morning ([Table 3]). Using the entire salivary profile or combining it with the other screening tests
did not help to classify the subtypes correctly (data not shown).
Table 3 Daily salivary profiles in patients with different subtypes of Cushing’s syndrome
(CS), Cushing’s disease (CD).
|
8:00 a.m.
|
12:00 a.m.
|
4:00 p.m.
|
8:00 p.m.
|
11:00 p.m.
|
|
Pituitary CS (N=75)
|
10.5 (5.2–18.8)
|
9.3 (4.9–16.5)
|
8.6 (4.8–13.8)
|
8.8 (5.1–14.3)
|
7.9 (4.1–11.9)
|
|
Adrenal CS (N=27)
|
7.1 (4.3–11.4)
|
9.0 (4.8–12.7)
|
7.5 (4.4–12.1)
|
6.5 (4.2–11.6)
|
6.1 (3.2–10.3)
|
|
Ectopic CS (N=9)
|
24.8 (14.1–39.4)
|
17.2 (6.5–37.9)
|
19.5 (6.5–37.9)
|
20.0 (5.8–40.0)
|
15.4 (7.7–22.1)
|
|
p-values (All groups)
|
p=0.001
|
p=0.1
|
p=0.04
|
p=0.05
|
p=0.03
|
|
p-values (CD vs. Adrenal)
|
p=0.04
|
–
|
p=1.0
|
p=0.5
|
p=0.4
|
|
p-values (CD vs. Ectopic)
|
p=0.04
|
–
|
p=0.07
|
p=0.2
|
p=0.1
|
|
p-values (Adrenal vs. Ectopic)
|
p=0.001
|
–
|
p=0.04
|
p=0.04
|
p=0.02
|
Discussion and conclusions
Discussion and conclusions
Interpretation of screening results for the detection or exclusion of endogenous hypercortisolism
can be challenging as elevated cortisol levels can be physiological in some cases
[2]
[14]. Likewise, the lack of specificity of clinical symptoms of CS that overlap with
much more prevalent conditions such as obesity, hypertension, and osteoporosis impede
efficient case finding [4]. Different approaches have been tested in the past to improve diagnostic accuracy,
such as face classification [15], diagnostic scores [16]
[17], plasma steroid profiling [18], and measuring hair cortisol [19]. However, none of these approaches have been introduced into clinical practice.
The standard screening still consists of a careful history, clinical examination,
and the use of three standard screening tests: measuring cortisol in a 24-hour collection
of urine, the late-night salivary cortisol, and the 1 mg low dose dexamethasone suppression
test [2]. Salivary cortisol has also been proposed for surveillance of medical therapy in
recurrent or persistent disease [20]
[21].
Main results and implications for clinical practice
The main findings of our study are two-fold: first, a five-point diurnal saliva cortisol
profile improved the diagnostic accuracy of saliva cortisol measurement as a screening
tool. Secondly, although the diurnal profiles between different subtypes of CS vary
statistically significantly, there is no clinically meaningful application in the
differential diagnosis because of strong overlap and variability. The high number
of subjects and the vigorous methodology give our study the adequate basis for a firm
clinical recommendation. Although saliva sampling can be easily conducted by the patient
in an outpatient setting, based on our findings, we do not recommend diurnal cortisol
profiles as a standard diagnostic tool for CS or a subtyping tool.
Surprisingly, patients with pituitary CS retain a circadian rhythm of cortisol secretion,
which is shifted upwards, while patients with adrenal CS are characterized by a more
“rigid” cortisol secretion pattern with reduced fluctuations. This is in line with
the results of a study using plasma cortisol levels in patients with adrenal CS and
Cushing’s disease [22]. In patients with ectopic CS, rhythm is completely lost. In clinical practice, distinguishing
between patients with pituitary and ectopic CS is a challenging task. The identification
of microadenomas in pituitary MRI is far from perfect [23], the inferior petrosal sinus invasive sampling and other tests like the high-dose
dexamethasone- suppression test or the CRH-test lack specificity [24]. In these cases, salivary cortisol profiles might give an additional hint towards
one or the other diagnosis.
Based on adrenal incidentalomas, patients with autonomous cortisol secretion have
a quite different profile compared to patients with adrenal CS. Circadian rhythm is
mostly intact. In comparison to control subjects, their afternoon and evening salivary
cortisol levels are elevated. The mode of treatment of patients with autonomous cortisol
secretion is still a matter of debate. Our data might confirm the approach to treat
these patients with chronotherapy using medical therapy in the evening to lower cortisol
levels [25] taking into account the importance of a circadian rhythm [26]
[27]. Of note, large prospective studies are needed to analyze the benefits of such an
approach.
Besides, salivary cortisol profiles might also be used to tailor medical therapy according
to a more physiological cortisol rhythm, as measuring salivary cortisol – besides
urinary free cortisol – in patients with medical therapy is associated with improved
quality of life [20]
[21].
Limitations and strengths
Despite its thorough study design, this study has a few limitations including the
monocentric design. Also, due to a low prevalence, the number of patients with ectopic
CS was low. This study also has several important strengths; we analyzed a relatively
large group of well-examined patients with a high number of patients with pituitary
and adrenal CS. In addition, all salivary cortisol measurements were performed in
the same laboratory.
Conclusion
Salivary cortisol day-profiles differ significantly between patients with and without
CS and between different subtypes. They can be used to improve diagnostic accuracy
and prevent false positive or false negative screening results, but the value is limited.
