Subtype-specific pattern of white blood cell differential in endogenous hypercortisolism
Mario Detomas 01, Barbara Altieri1, Irina ChifuDD1, Hanna Remde1, Xiang Zhou2, Laura-Sophie Landwehr1, Silviu Sbiera DD1, Matthias Kroiss DD1,3, Martin FassnachtOD1,4 and Timo Deutschbein1,5
1Division of Endocrinology and Diabetes, Department of Internal Medicine I, University Hospital, University of Würzburg, Würzburg, Germany, 2Division of Hematology, Department of Internal Medicine II, University Hospital Würzburg, Würzburg, Germany, 3Medizinische Klinik und Poliklinik IV, Klinikum der Universität München, Munich, Germany, 4Central Laboratory, University Hospital Würzburg, Würzburg, Germany, and 5Medicover Oldenburg MVZ, Oldenburg, Germany
Correspondence should be addressed to M Detomas
Abstract
Objective: Endogenous hypercortisolism predisposes to impaired immune function and infections. To date, however, it is unknown whether there is a subtype-specific pattern in white blood cell (WBC) and WBC differential (WBCD) count. Methods: A retrospective monocentric cohort study was carried out in patients with overt endogenous Cushing’s syndrome (CS) or adrenal incidentalomas and autonomous cortisol secretion (ACS), with WBC/WBCD analysis at initial diagnosis and after biochemical remission. Cut-offs were obtained by receiver-operating characteristics analysis. Results: In total, 253 patients were analyzed (Cushing’s disease (CD); n = 88; ectopic CS (ECS), n = 31; cortisol- producing adrenal adenomas (CPA), n = 40; ACS, n = 45; adrenocortical carcinomas (ACC), n = 49). Total leukocytes and neutrophils correlated positively with serum cortisol after 1-mg dexamethasone (r = 0.314 and r = 0.428), while a negative correlation was observed for lymphocytes and eosinophils (r = - 0.374 and r= - 0.380) (each P < 0.0001). Similar observations were made for 24 h-urinary free cortisol. CD and ECS differed in numbers of neutrophils and lymphocytes (P < 0.0001) and were well differentiated at a cut-off of 6.1 for the neutrophil/lymphocyte ratio (sensitivity 90.0%, specificity 89.4%, and areas under the curve (AUC) 0.918). For adrenocorticotropic hormone (ACTH)- independent CS, the best diagnostic outcome was obtained for the discrimination of CPA and ACC at a cut-off of 187.9 for the platelet/lymphocyte ratio (sensitivity 59.6%, specificity 80.6%, and AUC 0.713). For ECS, CPA, and CD, neutrophils decreased (delta -47.0, -29.7, and -26.2%) and lymphocytes increased (+123.2, +78.1, and +17.7%) already 3 months after remission.
Conclusion: Most immune cells correlate with the degree of hypercortisolism and differ among CS subtypes. WBCD changes are already identified 3 months after remission from endogenous hypercortisolism.
European Journal of Endocrinology (2022) 187, 439-449
Introduction
Clinically overt endogenous glucocorticoid excess is a rare condition with an incidence of 0.2-5.0 per million people per year (1, 2). The predominant subtype (accounting for about 80-85% of cases) is adrenocorticotropic hormone (ACTH)-dependent Cushing’s syndrome (CS). The vast majority of the latter is caused by an ACTH-secreting pituitary adenoma (so-called Cushing’s disease, CD), whereas ectopic
ACTH-secretion (e.g. arising from lung carcinoid tumors) is less common. In contrast, ACTH-independent CS is mainly attributed to cortisol-producing adrenal adenomas (CPA) or (less frequently) to adrenocortical carcinomas (ACC) (3). Furthermore, biochemical evidence of autonomous cortisol secretion (ACS) without clinical signs of overt CS is observed in up to 40% of patients with adrenal incidentalomas (4, 5).
Irrespective of the underlying etiology, one of the typical adverse effects of glucocorticoid excess is an increased risk of infections. The latter is caused by distinct hematological alterations such as leukocytosis (e.g. neutrophilia) with lymphopenia and eosinopenia (6, 7, 8, 9, 10). So far, however, it is not clear whether these alterations show a subtype-specific pattern, and how fast they normalize after biochemical remission.
Consequently, we performed a retrospective monocentric study involving a large cohort of patients with either endogenous CS or ACS in whom white blood cell differential (WBCD) was analyzed not only at the initial diagnosis of hypercortisolism but also after surgical cure. Furthermore, we compared the different CS subtypes according to their WBCD pattern, and to two serum inflammation-based scores that were already investigated in patients with CS and ACC (11, 12, 13, 14).
Subjects and methods
Subjects
A retrospective cohort of patients who were treated for either endogenous CS or ACS at the University Hospital of Würzburg was identified via chart review. All diagnoses were made according to established biochemical criteria (5, 15) between 2000 and 2021. Per definition, all patients with ACS suffered from an adrenal incidentaloma and showed biochemical evidence of cortisol autonomy (i.e. serum cortisol after the 1-mg dexamethasone suppression test (1-mg DST) of > 5 µg/dL) without clinical signs and symptoms of glucocorticoid excess.
Patients with a WBC count (either with or without a concomitant WBCD) at the time of the initial diagnosis of hypercortisolism were considered eligible. Patients were excluded in case of (i) concomitant infections; (ii) known autoimmune and/or hematological diseases; (iii) occurrence of serious cardiac events (e.g. myocardial infarction and acute heart failure) within 4 weeks before the first WBCD analysis; (iv) hepatic and/or renal comorbidities; (v) administration of chemotherapy for ACC or ECS before blood sampling.
The analysis included four time points, with blood taken at initial diagnosis and 3, 12, and at least 24 months after biochemical remission (i.e. after surgical removal of the causative tumor). With respect to the follow-up analysis, patients with supra-physiological doses of glucocorticoid replacement therapy (i.e. more than 30 mg hydrocortisone-equivalent per day), incomplete recovery from CS (as outlined by pathological biochemical tests),
or treatment with mitotane and/or chemotherapy were excluded.
All patients provided written informed consent to at least one of two disease-specific clinical registries (European Network for the Study of Adrenal Tumors registry (ENSAT) and/or Network of Excellence for Neuroendocrine Tumors (NeoExNET) registry). The approval numbers from the local ethics committee of the University Hospital of Würzburg are 88/11 (for the ENSAT registry) and 85/12 (for the NeoExNET registry).
Methods
Hormonal analysis
The endocrine workup was carried out with commercially available analytical procedures - that is, the Immulite apparatus (Siemens) for the analysis of plasma ACTH and serum cortisol, a manual luminescence immunoassay (IBL) for the analysis of salivary cortisol, and a manual RIA (Immuntech) for the analysis of urinary free cortisol (UFC).
White blood cell count analysis
WBCD included the analysis of neutrophils, lymphocytes, eosinophils, monocytes, and basophils. Analysis of the WBC and WBCD count was performed with Beckman Coulter GenS hematology analyzer (until 2009), Sysmex XE-2100 (from 2009 until 2017), and Sysmex XN-9000 (from 2017 onward). The systems did not demonstrate clinically relevant differences in terms of measurement results and reference values. Leukocytes, neutrophils, lymphocytes, eosinophils, monocytes, basophils, and platelets were analyzed. These laboratory parameters were also used to calculate two inflammation-based scores already investigated in the context of CS and/or ACC (11, 12, 13, 14): (i) the neutrophil/lymphocyte ratio (NLR); (ii) the platelet/lymphocyte ratio (PLR).
Statistical analysis
Continuous variables were tested for Gaussian distribution with the Shapiro-Wilk test. Normally distributed data are presented as mean and 95% CIs, while not normally distributed data are shown as median and interquartile range. Categorical variables are expressed as numbers and percentages. Parametric and non-parametric data were analyzed with Student’s t-tests or ANOVA followed by Tukey post-hoc test and Mann-Whitney U test or Kruskal-Wallis test followed by Dunn’s post-hoc test,
respectively. Correlation between continuous variables (r) was determined by Pearson’s correlation coefficient for two variables with normal distribution or Spearman’s correlation coefficient for not normally distributed variables. Qualitative variables were analyzed with the chi-square (x2) test. Receiver operator characteristics (ROC) curve analyses were performed to establish optimal cut- off values and their associated sensitivities, specificities, and areas under the curve (AUC). The Youden’s index (J=sensitivity + specificity - 1) was used to identify the most appropriate cut-off for both NLR and PLR. To identify the change in WBCD after biochemical remission, a mean delta change (evaluated in percentage) from baseline was calculated. For this analysis, only individuals with matched baseline and follow-up WBCD levels were taken into account. Generally, only one decimal digit was considered. In order to allow a delta mean change calculation, baseline values of 0.0 were rounded at 0.01. A P-value < 0.05 was considered statistically significant. Statistical analysis was performed with SPSS version 26 (IBM Corporation) and GraphPad Prism version 8 (GraphPad Software).
Results
Baseline characteristics of the study cohort
The study population comprised 253 patients (CD: n=88 (34.7%); ECS: n=31 (12.3%); CPA: n=40 (15.8%); ACS: n =45 (17.8%); ACC: n =49 (19.4%)).
Table 1 summarizes the clinical and biochemical data obtained in ACTH-dependent as well as ACTH-independent hypercortisolism (per definition, patients with ACS do not present any clinical signs and symptoms suggestive for Cushing’s syndrome). Briefly, ECS patients showed the highest degree of glucocorticoid excess, whereas ACS patients had the lowest (as illustrated by the outcome of various screening tests). Out of the 31 patients with ECS, 21 (67.7%) had a histologically confirmed diagnosis (lung carcinoid tumors, n=7; neuroendocrine tumors, n=6; small cell lung cancer, n = 4; medullary thyroid cancer, n = 3; pheochromocytoma, n=1). No statistically significant differences between the histologically confirmed and occult ECS tumors were observed if sex, age at initial diagnosis, and various biochemical parameters (i.e., serum cortisol after the 1-mg DST, 24 h-UFC, late-night salivary cortisol, and ACTH) were taken into account.
Within the first 3 months after the initial diagnosis of CS, patients with ECS were characterized by the highest
rate of infections (35.5%) compared to the other subtypes. More details on the observed infections are provided in Supplementary Table 1 (see section on supplementary materials given at the end of this article), whereas Supplementary Table 2 summarizes information on major cardiovascular events and anticoagulation therapy.
WBC and WBCD at initial diagnosis: correlation with the degree of hypercortisolism
Figure 1 shows the results derived from the entire study cohort. As illustrated, a positive correlation between the 1-mg DST and the absolute number of leukocytes (r=0.314; P < 0.0001) and neutrophils (r=0.428; P < 0.0001) was found. On the other hand, lymphocytes (r =- 0.374; P < 0.0001), eosinophils (r =- 0.380; P< 0.0001), and basophils (r =- 0.193; P < 0.05) were negatively correlated, whereas monocytes were not correlated at all (r =- 0.047; P=0.528). Furthermore, a significant positive correlation between the 1-mg DST and the NLR (r=0.554, P < 0.0001) was detected, while the PLR only showed a moderate correlation (r=0.246, P < 0.005) (Supplementary Fig. 1).
A correlation between the degree of hypercortisolism and the WBCD was also observed for 24-h UFC and late- night salivary cortisol (Supplementary Table 3). Further analyses revealed no significant differences between the correlation coefficients if the entire ACS population (irrespective of their biochemical outcome) or only ACS patients with abnormal results of 24-h UFC and/or late- night salivary cortisol were taken into account.
WBC and WBCD at initial diagnosis: general overview
We separately analyzed the distinct patterns of WBC, WBCD, NLR, and PLR in the five study subgroups (Fig. 2 and Table 2). At the time of the initial diagnosis of the underlying disorder, all WBC types except for basophils differed among the CS subtypes.
WBC and WBCD at initial diagnosis: subtype- specific differences in ACTH-dependent Cushing’s syndrome
In patients with ACTH-dependent CS, ACTH showed a significantly positive correlation with neutrophils (r=0.298) and leukocytes (r=0.263) on the one hand and a significant negative correlation with lymphocytes
| All patients | ACTH-dependent hypercortisolism | ACTH-independent hypercortisolism | ||||||
|---|---|---|---|---|---|---|---|---|
| CD | ECS | P-value | CPA | ACS | ACC | P-value | ||
| Total (n) | 253 | 88 | 31 | 40 | 45 | 49 | ||
| Females | 202 (74.8%) | 71 (80.7%) | 19 (61.3%) | <0.05 | 33 (84.6%) | 30 (66.7%) | 37 (75.5%) | n.s. |
| Age at diagnosis (years)t | 51.1 (49.3-52.8) | 46.1 (43.2-49.1) | 55.7 (50.6-60.8) | n.s. | 48.9 (44.2-53.6) | 56.9 (53.3-60.7) | 49.8 (45.8-53.8) | n.s. |
| BMI (kg/m2) | 28.7 (8.6) | 30.0 (9.2) | 26.2 (6.5) | n.s. | 27.3 (11.4) | 28.3 (9.2) | 28.0 (6.8) | n.s. |
| HbA1c (%) | 6.4 (1.1) | 6.3 (0.9) | 6.9 (1.1) | n.s. | 6.4 (0.8) | 6.2 (1.7) | 6.8 (1.6) | n.s. |
| Antihypertensives per patient (n) | 2 (3.0) | 2 (3.0) | 2 (3.0) | n.s | 2 (2.75) | 1 (3.0) | 2 (1.25) | n.s. |
| Patients under | 57 (22.5%) | 18 (20.4%) | 11 (35.4%) | n.s. | 8 (20.0%) | 3 (6.7%) | 17 (34.7%) | <0.001 |
| antidepressants | ||||||||
| Patients with infections* | 22 (8.7%) | 6 (6.8%) | 11 (35.5%) | <0.005 | 2 (5.1%) | 1 (2.2%) | 2 (4.1%) | n.s. |
| ACTH (ng/L) | 14.5 (50.5) | 50.6 (60.2) | 93.1 (98) | <0.0001 | 5.0 (1.3) | 6.8 (7.0) | 5.0 (3.3) | <0.005 |
| Serum cortisol after | 11.9 (16.0) | 12.4 (14.3) | 28.5 (32.3) | <0.0001 | 19.8 (9.6) | 7.3 (3.6) | 19.8 (18.2) | <0.0001 |
| 1-mg DST (µg/dL) | ||||||||
| 24 h-urinary free | 168.5 (344.0) | 170.0 (265.5) | 555.7 (805.9) | <0.0001 | 258.3 (315.2) | 63.6 (47.8) | 279.5 (592.5) | <0.0001 |
| cortisol (µg/day) | ||||||||
| Late-night salivary cortisol (µg/dL) | 0.45 (1.0) | 0.5 (0.6) | 3.9 (4.1) | <0.0001 | 0.6 (0.5) | 0.2 (0.2) | 0.9 (2.7) | <0.0001 |
*Infections within the first 3 months from diagnosis; tMean (95% CI).
ACC, adrenocortical carcinoma; ACS, autonomous cortisol secretion; ACTH, adrenocorticotropin; BMI, body-mass-index; CD, Cushing’s disease; CPA: cortisol-producing adrenal adenoma; DST,
dexamethasone suppression test; ECS, ectopic Cushing’s syndrome; HbA1c, glycated hemoglobin; IQR, interquartile range; n.s., not significant.
25
r= 0.314
20
r = 0.428
Leukocytes (n*1000/uL)
p < 0.0001
Neutrophils (nº1000/uL)
p < 0.0001
20
15
15
0
10
5
5
0
0
20
40
60
80
100
0
0
20
40
60
80
100
Serum cortisol after 1-mg DST (pg/dL)
Serum cortisol after 1-mg DST (µg/dL)
Lymphocytes (n*1000/uL)
5
r =- 0.374
8
r =- 0.380
p < 0.0001
Eosinophils (nº1000/L)
p < 0.0001
4
.6
3
4
2
1
.2
··
0
0
20
40
60
80
100
.0
0
20
40
60
80
100
Serum cortisol after 1-mg DST (µg/dL)
Serum cortisol after 1-mg DST (µg/dL)
.3
r= - 0.193
2.5
Basophils (n*1000/uL)
r =- 0.047
p < 0.05
Monocytes (nº1000/uL)
p = 0.528
2.0
.2
1.5
.1
1.0
.5
…
.
.0
0
20
40
60
80
100
.0
0
20
40
60
80
100
Serum cortisol after 1-mg DST (µg/dL)
Serum cortisol after 1-mg DST (µg/dL)
Figure 1 Correlation between serum cortisol during the 1-mg dexamethasone suppression test and leukocytes, neutrophils, lymphocytes, eosinophils, basophils, and monocytes at the time of the initial diagnosis of hypercortisolism. 1-mg DST, 1-mg dexamethasone suppression test.
(r =- 0.303) and eosinophils (r =- 0.245) on the other (each P <0.005).
A significant subtype-specific difference in the number of neutrophils and lymphocytes (each P < 0.0001), and leukocytes and eosinophils (each P < 0.005), respectively, was identified. In detail, CD patients showed lower levels of leukocytes (median: 9.1 n*1000/uL vs 11.3 n*1000/uL) than ECS patients and neutrophils (median: 6.2 n*1000/ uL vs 9.1 n*1000/uL). On the contrary, lower levels of lymphocytes (median: 1.0 n*1000/uL vs 1.7 n*1000/µL) and eosinophils (median: 0.0 n*1000/uL vs 0.1 n*1000/uL) were observed in ECS. No significant difference was found for basophils and monocytes.
NLR was significantly lower in CD compared to ECS (P < 0.0001). ROC analysis revealed an optimal cut-off of 6.1. Applying this cut-off for subtype differentiation, a sensitivity of 90.0% and a specificity of 89.4% were obtained (AUC=0.918) (Fig. 3A). This AUC value was higher than those calculated for the 1-mg DST and for 24-h UFC (AUC: 0.828 vs AUC 0.741) (Fig. 4). Of note, a subset of 86 patients (67 CD, 19 ECS) underwent a stimulation test with Corticotropin-releasing hormone (CRH). With respect to subtype differentiation, the
p<0.005
p=0.999
P<0.0001
p=0.999
25
p<0.001
25
p=0.213
p<0.05
p<0.05
A
20
Leukocytes (n*1000/pl)
00
Neutrophils (n*1000/pl)
20
A
0°
A
15
AA
15
0
a
A
A
4
o
44
A
10
ARA
10
01
CA44
AAA
5
ADA
0
A
5
O
000
4
0
CD
ECS
CPA
ACS
ACC
0
CD
ECS
CPA
ACS
ACC
P<0.0001
p=0.090
P<0.005
p=0.999
5
p=0.999
P<0.0001
0.85
p<0.05
o
p<0.0001
.
0.70
4
Lymphocytes (n*1000/pl)
4
00
Eosinophils (n*1000/pl)
o
00
0
.55
3
0
..
A
00
2
A
.40
00
1
O
0
2
₾
0000
60.25
O
CHCIDO
00
AAAAAA
A
1
0
00
U
0.10
CLIO
Andbones
O
O
0
4
AA
0
CI ..
MUITO
0
CPA
-0.05
CD
ECS
ACS
ACC
CD
ECS
CPA
ACS
ACC
p=0.999
p=0.999
p=0.999
p<0.05
0.25
p=0.167
p=0.147
2.0
p=0.085
p=0.999
0.2
o
M
-
o
A
Basophils (n*1000/pl)
Monocytes (n*1000/pl)
£1.5
o
80.15
..
A
00
O
A4
GIROD
000
0.1
MIXED
…
0000
000000006
1.0
000
000
…
A
CHICO
…
C
44
…
CIILD
…
AAAAA
$0.05
o
-
aBID
…
TTIm
-In
00005000
000
-
OLD
…
0,500
…
AAAAA
o
M
A
0.5
CIODOOD
—
000
AMA
…
0
a.
CXOD
AZA
o
..
0
AAA
ΔΔΔ
o
-0.05
CD
ECS
CPA
ACS
ACC
0.0
CD
ECS
CPA
ACS
ACC
diagnostic outcome of the NLR and the CRH stimulation test (with either cortisol or ACTH analysis) was comparable (Supplementary Fig. 2).
Although the PLR was significantly lower in CD compared to ECS (P < 0.005), its diagnostic accuracy was minor (sensitivity 70.0%, specificity 70.2%, AUC 0.727 at the optimal cut-off of 192.7).
| All patients, n | ACTH-dependent hypercortisolism | ACTH-independent hypercortisolism | ||||||
|---|---|---|---|---|---|---|---|---|
| CD | ECS | P-value | CPA | ACS | ACC | P-value | ||
| Leukocytes (n*1000/µL) | 253 | 9.1 (3.4) | 11.3 (4.8) | <0.005 | 9.8 (3.2) | 7.8 (3.2) | 9.3 (3.7) | <0.005 |
| Neutrophils (n*1000/µL) | 225 | 6.2 (2.7) | 9.1 (4.5) | <0.0001 | 7.4 (2.8) | 5.5 (3.1) | 6.8 (4.2) | <0.005 |
| Lymphocytes (n*1000/µL) | 225 | 1.7 (0.9) | 1.0 (0.5) | <0.0001 | 1.7 (0.8) | 1.9 (1.1) | 1.4 (1.0) | <0.001 |
| Eosinophils (n*1000/µL) | 225 | 0.1 (0.1) | 0.0 (0.1) | <0.005 | 0.1 (0.1) | 0.2 (0.1) | 0.0 (0.1) | <0.0001 |
| Basophils (n*1000/µL) | 225 | 0.00 (0.03) | 0.00 (0.00) | n.s. | 0.00 (0.01) | 0.00 (0.10) | 0.00 (0.01) | n.s. |
| Monocytes (n*1000/µL) | 225 | 0.7 (0.3) | 0.7 (0.4) | n.s. | 0.7 (0.4) | 0.7 (0.4) | 0.6 (0.4) | <0.05 |
| NLR | 225 | 3.5 (2.3) | 8.8 (4.7) | <0.0001 | 4.7 (2.8) | 2.4 (1.6) | 5.2 (5.7) | <0.0001 |
| PLR | 225 | 159.9 (85.6) | 244.6 (183.8) | <0.0001 | 142.7 (67.8) | 135.3 (56.7) | 210.7 (136.9) | <0.001 |
ACC, adrenocortical carcinoma; ACS, autonomous cortisol secretion; ACTH, adrenocorticotropin; CD, Cushing’s disease; CPA: cortisol-producing adrenal adenoma; ECS, ectopic Cushing’s syndrome; IQR, interquartile range; NLR, neutrophil/lymphocyte ratio; PLR, platelet/lymphocyte ratio.
WBC and WBCD at initial diagnosis: subtype- specific differences in ACTH-independent Cushing’s syndrome and autonomous cortisol secretion
In patients with CPA and ACC, only the number of monocytes and the PLR significantly differed between both entities (each P < 0.05) (Fig. 3B). For the PLR, an optimal cut-off of 187.9 was calculated for separation of CPA and ACC (sensitivity: 59.6%, specificity: 80.6%, AUC=0.713) (Fig. 3B).
Comparison of CPA and ACS revealed significant differences in the numbers of leukocytes (P < 0.0001), neutrophils (P < 0.005), eosinophils (P < 0.05) and in the NLR score (P < 0.005). Applying the optimal cut-off of 3.7 for NLR, a sensitivity of 61.3% and specificity of 81.3% were obtained (AUC=0.717) (Fig. 3C).
The comparison between ACC and ACS revealed significantly different numbers of leukocytes (P < 0.05), neutrophils (P < 0.05), lymphocytes (P < 0.0001), and eosinophils (P < 0.0001), as well as NLR and PLR (each P<0.0001). An optimal cut-off of 2.6 for the NLR resulted in a sensitivity of 91.5% and a specificity of 59.4% (AUC=0.797) (Fig. 3D), while for PLR a cut-off of 169.9 showed a sensitivity of 68.1% and a specificity of 78.1% (AUC=0.722).
Changes in WBC count and WBCD after remission from hypercortisolism
Data from follow-up at 3 months after surgical cure of hypercortisolemia were available in 127/253 (50.2%)
patients (CD: n=52/88 (59.1%); ECS: n=18/31 (58.1%); CPA: n =26/40 (65.0%); ACS: n = 31/45 (68.9%)).
As outlined in Fig. 5 and Supplementary Table 4, a remarkable decrease in leukocytes from baseline to 3 months after remission was observed for patients with ECS (mean delta: - 30.6%), CPA (mean delta: - 19.9%), and CD (mean delta: - 14.4%). A similar effect was seen for the neutrophils (mean delta: ECS, -47.0%; CPA, -29.6%; CD, -26.2%). In contrast, patients with ACS did not show a substantial change in either leukocytes or neutrophils (mean delta: +1.0% and +1.1%) (Fig. 5).
Lymphocytes increased during short term follow-up, with the most pronounced effect observed in patients with ECS (mean delta: ECS, +123.2%; CPA, +78.1%; CD, +17.7%). Eosinophils showed the highest increase in ECS (mean delta: +757.1%), followed by CD (mean delta: +381.1%) and CPA (mean delta: +339.1%) (Fig. 5 and Supplementary Table 4).
Follow-up data at 12 and >24 months were available in 95/253 (37.5%) patients (CD: n=46/88 (52.3%); ECS: n =10/31 (32.3%); CPA: n =22/40 (55.5%); ACS: n =17/45 (37.7%)), and in 84/253 (33.2%) patients (CD: n=40/88 (45.5%); ECS: n =11/31 (35.5%) CPA: n =19/40 (47.5%); ACS,: n=14/45 (31.1%)), respectively. Patients with formerly overt CS showed a decrease in leukocytes and neutrophils over time, whereas patients with former ACS did not. In contrast, the number of lymphocytes constantly increased in CD and ECS patients, while patients with CPA showed a significant increase until the first postoperative follow-up at 3 months, which was then followed by a progressive decrease over time.
A
B
20
·
Neutrophil/lymphocyte ratio
500
A
..
Platelet/lymphocyte ratio
A
15
400
C
..
AUC= 0.92
AA4
AUC= 0.71
.
300
0
A
10
00
…
40
9
8
4ª 4º
.
200
0
.1
00
.
188
A
0
8
5
4
000 00 000
.
4
0
100
4ª
A
0000000060
·
0000000000000
4
0
C
CD
ECS
0
D
CPA
ACC
20
20
Neutrophil/lymphocyte ratio
Neutrophil/lymphocyte ratio
0
A
15
AUC= 0.72
15
A
AUC= 0.80
0
AA
DAA
10
10
.
-
AZ
0
4
.
.
4
5
O
.
5
A4 444
.7
4
2.6
AAAAAAAA
0
0
0
CPA
ACS
0
ACS
ACC
Figure 3 Diagnostic outcome of two inflammation-based scores (NLR and PLR) for the differentiation of study subgroups. Results are presented along with the corresponding ROC curves. The dotted lines in the scatter plots illustrate the optimal cut-off for the NLR and PLR. ACC, adrenocortical carcinoma; ACS, autonomous cortisol secretion; AUC, area under the curve; CD, Cushing’s disease; CPA, cortisol-producing adrenal adenoma; ECS, ectopic Cushing’s syndrome; NLR, neutrophil/lymphocyte ratio; PLR, platelet/lymphocyte ratio; ROC, receiver-operating characteristics.
As expected from the preoperative analysis and from the absence of a correlation with the degree of hypercortisolism, no substantial changes in the number of monocytes were detected. The mean delta-% change of the WBCD parameters is separately reported for each study subgroup in Fig. 5 and Supplementary Table 4.
Discussion
Impaired immune function and altered immune cells due to a glucocorticoid excess predispose CS patients to infectious diseases. Although some of these changes have already been reported more than 70 years before (16), few studies with limited cohorts have provided additional knowledge since then (6, 7, 8, 9, 13, 14). We here report the results of a large monocentric analysis, highlighting the correlation of the WBCD with the degree of hypercortisolism. Moreover, the changes from initial diagnosis to (short- and long- term) follow-up and the potency of both WBCD analysis and inflammation-based scores for subtype differentiation of endogenous CS are expressed.
Elevated neutrophils in context of both CD (8, 9) and CPA (13, 16) as well as an impaired neutrophil function (as illustrated by a reduced endothelial adhesion, a delayed extravasation, and an impaired chemotaxis) (6, 7) have already been described. So far, however, different CS subtypes were not compared among each other. In the present study, we were able to show that patients suffering from ECS did not only have the highest cortisol levels but also the highest neutrophil count among all CS subtypes.
Lymphopenia is another hematological alteration related to CS (9). Glucocorticoids affect the production and action of T helper cells (Th1 and Th2), which are a crucial component of adaptive immunity (7). Furthermore, it has been shown that patients with CS have a decreased proliferation of B lymphocytes (6). In our study, we found a direct negative correlation between serum cortisol levels after the 1-mg DST or 24-h UFC on the one hand and the number of lymphocytes on the other. Once again, ECS patients had the most pronounced lymphopenia.
Although it is tempting to speculate that an increase of immune cells represents a protective factor for infections in patients with overt CS, patients with neutrophilia and lymphopenia showed the highest prevalence of infections within 3 months from the initial diagnosis of CS. Our
1-mg dexamethasone suppression test
24h-urinary free cortisol
Neutrophil/lymphocyte ratio
1.0
1,0
1,0
0.8
0,8
0,8-
0.6-
0,6
0,6
Sensitivity
Sensitivity
Sensitivity
0,4-
0,4
0,4-
0,2
0,2
0,2
0.0
0,0
0,0
0,2
0,4
0,6
0,8
1,0
0,0
0,0
0,2
0,4
0,6
0,8
1,0
0,0
0,2
0,4
0,6
0,0
1,0
1-specificity
1-specificity
1-specificity
AUC=0.83
AUC=0.74
AUC=0.92
0 %-change from preoperatory levels
Leukocytes
0 % -change from preoperatory levels
Neutrophils
40
100
20
50
0
-20
0
-40
-50
-60
-80
3 months
CD 12 months
CD >24 months
ECS 3 months
ECS 12 months
ECS >24 months
CPA 3 months
CPA 12 months
CPA >24 months
ACS 3 months
ACS 12 months
ACS >24 months
100
3 months
CD 12 months
CD >24 months
ECS 3 months
ECS 12 months
ECS >24 months
CPA 3 months
CPA 12 months
CPA >24 months
ACS 3 months
ACS 12 months
ACS >24 months
Po % -change from preoperatory levels
Lymphocytes
% -change from preoperatory levels
Eosinophils
300
2500
2000
200
1500
100
1000
500
0
0
100
3 months
CD 12 months
CD >24 months
ECS 3 months
ECS 12 months
ECS >24 months
CPA 3 months
CPA 12 months
CPA >24 months
ACS 3 months
ACS 12 months
ACS >24 months
-500
CD 3 months
CD 12 months
CD >24 months
ECS 3 months
ECS 12 months
ECS >24 months
CPA 3 months
CPA 12 months
CPA >24 months
ACS 3 months
ACS 12 months
ACS >24 months
0 %-change from preoperatory levels
Basophils
% -change from preoperatory levels
Monocytes
600
150
100
400
50
200
0
-50
0
-100
200
3 months
CD 12 months
CD >24 months
ECS 3 months
ECS 12 months
ECS >24 months
CPA 3 months
CPA 12 months
CPA >24 months
ACS 3 months
ACS 12 months
ACS >24 months
-150
CD 3 months
CD 12 months
CD >24 months
ECS 3 months
ECS 12 months
ECS >24 months
CPA 3 months
CPA 12 months
CPA >24 months
ACS 3 months
ACS 12 months
ACS >24 months
Figure 5 Mean delta change (%) 3 months, 12 months, and more than 24 months after remission from hypercortisolism. For each analysis, only patients with available matched samples (i.e. from baseline and the respective time point during follow-up) were taken into account. For eosinophils and basophils, levels of 0.0 at baseline were artificially set to a baseline level of 0.01, thereby enabling the delta change analysis. ACS, autonomous cortisol secretion; ACTH, adrenocorticotropic hormone; AUC, area under the curve; CD, Cushing’s disease; CPA, cortisol-producing adrenal adenoma; ECS, ectopic Cushing’s syndrome.
observed rate (35.5%) is in line with previously reported data (ranging from 23 to 51%) (17, 18, 19).
Since both neutrophilia and lymphopenia were more accentuated in ECS than in CD, we further evaluated the diagnostic value of the inflammation-based score NLR. As recently reported, a similar approach was performed in a pediatric CS population (20). Of interest, the NLR was previously described as a predictor of (i) mortality in the general population (21); (ii) critical illness in patients with COVID-19 pneumonia (22); (iii) prognosis and survival in several types of cancer, including ACC (11, 12). Furthermore, an increased NLR was observed in a variety of clinical conditions like bacterial or fungal infection, acute stroke, myocardial infarction, atherosclerosis, and systemic inflammatory response (23). Considering that (i) the normal NLR in an adult, non-geriatric population
should be in the range between 0.78 and 3.92 (24, 25); (ii) all of our ECS patients with an infection before the initial diagnosis of CS had an elevated NLR (as illustrated in Supplementary Table 1), we assume that neutrophilia and lymphopenia were major drivers for the higher rate of infections in these patients.
Applying our newly generated cut-off of 6.1 for the NLR in patients with ACTH-dependent CS, remarkably high sensitivity (90.0%) and specificity (89.4%) for the identification of ECS were obtained. This phenomenon may be explained by the fact that (i) ECS patients are usually characterized by a more pronounced glucocorticoid excess compared to CD patients (26, 27); (ii) the WBCD is influenced by the degree of hypercortisolism (of note, both aspects were also observed in the current analysis). However, we were able to show that the NLR is not entirely
dependent on the degree of glucocorticoid excess (Fig. 3) and that the WBCD has a better discriminatory power for subtype differentiation than a dynamic testing procedure commonly applied for this purpose (i.e. the CRH stimulation test). Nevertheless, these results need to be confirmed in a prospective study with a larger patient population.
Of note, our results are in contrast to those of a recently reported pediatric population (20), where the NLR did not relevantly differ between the ACTH-dependent CS subtypes. However, it has to be pointed out that, for example, differences in the levels of lymphocyte number and action occur with age, probably explaining discrepant findings between children and adults (28, 29, 30, 31).
Interestingly, NLR was already used to discriminate between CPA and non-functioning adrenal adenomas (13). Although a significant difference between the two groups was identified, the obtained AUC was relatively small (0.729). With respect to our cohort, NLR was also different among CPA and ACS (P <0.005), thereby confirming what was already reported elsewhere (13). Of note, if the NLR was used to differentiate between ACC and ACS, a high sensitivity (91.5%) along with low specificity (59.4%) was observed. We were also able to show that the PLR ratio was significantly different between CPA and ACC patients. PLR was also described as a possible predictor of disease-specific survival in ACC (11). Of note, however, the newly obtained optimal PLR cut-off of 187.9 was characterized by moderate specificity (80.6%) but low sensitivity (59.6%), illustrating its limitations in clinical routine.
If patients with CPA were compared to those with ACC, the WBDC pattern differed only in the monocytes. The latter are circulating precursors of macrophages and play a fundamental role in inflammatory activation, innate and adaptive immunity, and tissue repair. In patients with CS, low (7, 32) as well as high (13, 18) levels of monocytes were previously reported. In our study, however, monocytes were the only immune cells that did not correlate at all with the degree of hypercortisolism. This difference with respect to former data is possibly explained by the fact that former studies comprised much smaller cohorts (ranging from 26 to 42 patients) than our current large study population (n=225) and were therefore probably more prone to outliers.
A smaller study already reported a significant decline of leukocytes already 4 weeks after surgery for CD, followed by a continuous decrease over the next years (8). Here, we were able to show that leukocytes and neutrophils substantially decreased not only in CD but also in all CS subtypes. Furthermore, we observed an increase in lymphocytes and eosinophils after surgical cure of hypercortisolism. Such a
‘lymphocytes-rebound’ was already described by others as a possible trigger for autoimmune disorders occurring after remission of CS (33, 34).
Although this study included a large population of patients with different entities and degrees of endogenous hypercortisolism, its retrospective nature surely represents a major limitation. Furthermore, the size of the subtype groups is still limited, and a relevant number of patients were lost to follow-up. On the other hand, our single- center study is characterized by a homogenous approach to patients’ diagnostic workup, treatment, and data collection. Additionally, it comprises a robust number of patients and subtypes.
In conclusion, we here provide the data from a large cohort with different entities and degrees of endogenous hypercortisolism, thereby illustrating how the WBCD parameters are directly influenced by elevated cortisol levels. Furthermore, we identified how the WBCD differs among the different entities and could be potentially useful in the discrimination of the different entities. Finally, we observed how postoperatively most of the WBC gradually normalize, starting already at 3 months after remission.
Supplementary materials
This is linked to the online version of the paper at https://doi.org/10.1530/ EJE-22-0211.
Declaration of interest
Martin Fassnacht is a senior editor of the European Journal of Endocrinology and was not involved in the peer review or editorial process for this paper on which he is listed as an author. The other authors declare no competing interest.
Funding
This work was supported by the DFG German Research Foundation Project 314061271-TRR 205 (to M K and M F) and the European Reference Network on Rare Endocrine Conditions (Endo-ERN).
Author contribution statement
M D designed the research. M D, B A, and T D performed the statistical analyses and drafted the manuscript. All authors collected samples and clinical data from patients, contributed to writing the manuscript, and approved the final version to be published.
Acknowledgements
The authors would like to thank Yvonne Möhres and Kristina Hoffmann for data acquisition and for facilitating the assessment of patients.
References
1 Lacroix A, Feelders RA, Stratakis CA & Nieman LK. Cushing’s syndrome. Lancet 2015 386 913-927. (https://doi.org/10.1016/S0140- 6736(14)61375-1)
2 Ragnarsson O, Olsson DS, Chantzichristos D, Papakokkinou E, Dahlqvist P, Segerstedt E, Olsson T, Petersson M, Berinder K, Bensing S et al. The incidence of Cushing’s disease: a nationwide Swedish study. Pituitary 2019 22 179-186. (https://doi.org/10.1007/ s11102-019-00951-1)
3 Detomas M, Altieri B, Schlotelburg W, Appenzeller S, Schlaffer S, Coras R, Schirbel A, Wild V, Kroiss M, Sbiera S et al. Case report: consecutive adrenal Cushing’s syndrome and Cushing’s disease in a patient with somatic CTNNB1, USP8, and NR3C1 mutations. Frontiers in Endocrinology 2021 12 731579. (https://doi.org/10.3389/ fendo.2021.731579)
4 Ebbehoj A, Li D, Kaur RJ, Zhang C, Singh S, Li T, Atkinson E, Achenbach S, Khosla S, Arlt W et al. Epidemiology of adrenal tumours in Olmsted County, Minnesota, USA: a population-based cohort study. Lancet: Diabetes and Endocrinology 2020 8 894-902. (https://doi. org/10.1016/S2213-8587(20)30314-4)
5 Fassnacht M, Arlt W, Bancos I, Dralle H, Newell-Price J, Sahdev A, Tabarin A, Terzolo M, Tsagarakis S & Dekkers OM. Management of adrenal incidentalomas: European Society of Endocrinology Clinical Practice Guideline in collaboration with the European Network for the study of adrenal tumors. European Journal of Endocrinology 2016 175 G1-G34. (https://doi.org/10.1530/EJE-16-0467)
6 Hasenmajer V, Sbardella E, Sciarra F, Minnetti M, Isidori AM & Venneri MA. The immune system in Cushing’s syndrome. Trends in Endocrinology and Metabolism 2020 31 655-669. (https://doi. org/10.1016/j.tem.2020.04.004)
7 Pivonello R, Isidori AM, De Martino MC, Newell-Price J, Biller BM & Colao A. Complications of Cushing’s syndrome: state of the art. Lancet: Diabetes and Endocrinology 2016 4 611-629. (https://doi. org/10.1016/S2213-8587(16)00086-3)
8 Ambrogio AG, De Martin M, Ascoli P, Cavagnini F & Pecori Giraldi F. Gender-dependent changes in haematological parameters in patients with Cushing’s disease before and after remission. European Journal of Endocrinology 2014 170 393-400. (https://doi.org/10.1530/EJE-13- 0824)
9 Masri-Iraqi H, Robenshtok E, Tzvetov G, Manistersky Y & Shimon I. Elevated white blood cell counts in Cushing’s disease: association with hypercortisolism. Pituitary 2014 17 436-440. (https://doi.org/10.1007/ s11102-013-0522-0)
10 Chifu I, Detomas M, Dischinger U, Kimpel O, Megerle F, Hahner S, Fassnacht M & Altieri B. Management of patients with glucocorticoid- related diseases and COVID-19. Frontiers in Endocrinology 2021 12 705214. (https://doi.org/10.3389/fendo.2021.705214)
11 Bagante F, Tran TB, Postlewait LM, Maithel SK, Wang TS, Evans DB, Hatzaras I, Shenoy R, Phay JE, Keplinger K et al. Neutrophil- lymphocyte and platelet-lymphocyte ratio as predictors of disease specific survival after resection of adrenocortical carcinoma. Journal of Surgical Oncology 2015 112 164-172. (https://doi.org/10.1002/ jso.23982)
12 Mochizuki T, Kawahara T, Takamoto D, Makiyama K, Hattori Y, Teranishi JI, Miyoshi Y, Yumura Y, Yao M & Uemura H. The neutrophil- to-lymphocyte ratio (NLR) predicts adrenocortical carcinoma and is correlated with the prognosis. BMC Urology 2017 17 49. (https://doi. org/10.1186/s12894-017-0240-4)
13 Wang W, Wang J, Shen C, Zhu S, Gao Y & Zhang J. Neutrophil- lymphocyte ratio as an initial screening biomarker for differential diagnosis of Cushing’s syndrome from nonfunctional adenoma in patients with an adrenal mass. BioMed Research International 2021 2021 6635594. (https://doi.org/10.1155/2021/6635594)
14 Marques P, de Vries F, Dekkers OM, van Furth WR, Korbonits M, Biermasz NR & Pereira AM. Pre-operative serum inflammation-based scores in patients with pituitary adenomas. Pituitary 2021 24 334-350. (https://doi.org/10.1007/s11102-020-01112-5)
15 Nieman LK, Biller BM, Findling JW, Newell-Price J, Savage MO, Stewart PM & Montori VM. The diagnosis of Cushing’s syndrome:
an Endocrine Society Clinical Practice Guideline. Journal of Clinical Endocrinology and Metabolism 2008 93 1526-1540. (https://doi. org/10.1210/jc.2008-0125)
16 De La Balze FA, Reifenstein EC, Jr & Albright F. Differential blood counts in certain adrenal cortical disorders (Cushing’s syndrome, Addison’s disease and panhypopituitarism). Journal of Clinical Endocrinology and Metabolism 1946 6 312-319. (https://doi.org/10.1210/ jcem-6-4-312)
17 Ilias I, Torpy DJ, Pacak K, Mullen N, Wesley RA & Nieman LK. Cushing’s syndrome due to ectopic corticotropin secretion: twenty years’ experience at the National Institutes of Health. Journal of Clinical Endocrinology and Metabolism 2005 90 4955-4962. (https:// doi.org/10.1210/jc.2004-2527)
18 Sarlis NJ, Chanock SJ & Nieman LK. Cortisolemic indices predict severe infections in Cushing syndrome due to ectopic production of adrenocorticotropin. Journal of Clinical Endocrinology and Metabolism 2000 85 42-47. (https://doi.org/10.1210/jcem.85.1.6294)
19 Ejaz S, Vassilopoulou-Sellin R, Busaidy NL, Hu MI, Waguespack SG, Jimenez C, Ying AK, Cabanillas M, Abbara M & Habra MA. Cushing syndrome secondary to ectopic adrenocorticotropic hormone secretion: the University of Texas MD Anderson Cancer Center Experience. Cancer 2011 117 4381-4389. (https://doi.org/10.1002/ cncr.26029)
20 Wurth R, Rescigno M, Flippo C, Stratakis CA & Tatsi C. Inflammatory biomarkers in the evaluation of pediatric endogenous Cushing syndrome. European Journal of Endocrinology 2022 186 503-510. (https://doi.org/10.1530/EJE-21-1199)
21 Song M, Graubard BI, Rabkin CS & Engels EA. Neutrophil-to- lymphocyte ratio and mortality in the United States general population. Scientific Reports 2021 11 464. (https://doi.org/10.1038/ s41598-020-79431-7)
22 Liu J, Liu Y, Xiang P, Pu L, Xiong H, Li C, Zhang M, Tan J, Xu Y, Song R et al. Neutrophil-to-lymphocyte ratio predicts critical illness patients with 2019 coronavirus disease in the early stage. Journal of Translational Medicine 2020 18 206. (https://doi.org/10.1186/s12967- 020-02374-0)
23 Buonacera A, Stancanelli B, Colaci M & Malatino L. Neutrophil to lymphocyte ratio: an emerging marker of the relationships between the immune system and diseases. International Journal of Molecular Sciences 2022 23 3636. (https://doi.org/10.3390/ijms23073636)
24 Fest J, Ruiter R, Ikram MA, Voortman T, van Eijck CHJ & Stricker BH. Reference values for white blood-cell-based inflammatory markers in the Rotterdam Study: a population-based prospective cohort study. Scientific Reports 2018 8 10566. (https://doi.org/10.1038/s41598-018- 28646-w)
25 Forget P, Khalifa C, Defour JP, Latinne D, Van Pel MC & De Kock M. What is the normal value of the neutrophil-to-lymphocyte ratio? BMC Research Notes 2017 10 12. (https://doi.org/10.1186/s13104-016-2335-5)
26 Arnaldi G, Angeli A, Atkinson AB, Bertagna X, Cavagnini F, Chrousos GP, Fava GA, Findling JW, Gaillard RC, Grossman AB et al. Diagnosis and complications of Cushing’s syndrome: a consensus statement. Journal of Clinical Endocrinology and Metabolism 2003 88 5593-5602. (https://doi.org/10.1210/jc.2003-030871)
27 Newell-Price J, Morris DG, Drake WM, Korbonits M, Monson JP, Besser GM & Grossman AB. Optimal response criteria for the human CRH test in the differential diagnosis of ACTH-dependent Cushing’s syndrome. Journal of Clinical Endocrinology and Metabolism 2002 87 1640-1645. (https://doi.org/10.1210/jcem.87.4.8357)
28 Holcar M, Goropevsek A, Ihan A & Avcin T. Age-related differences in percentages of regulatory and effector T lymphocytes and their subsets in healthy individuals and characteristic STAT1/STAT5 signalling response in helper T lymphocytes. Journal of Immunology Research 2015 2015 352934. (https://doi.org/10.1155/2015/352934)
29 Hulstaert F, Hannet I, Deneys V, Munhyeshuli V, Reichert T, De Bruyere M & Strauss K. Age-related changes in human blood
lymphocyte subpopulations. II. Varying kinetics of percentage and absolute count measurements. Clinical Immunology and Immunopathology 1994 70 152-158. (https://doi.org/10.1006/ clin.1994.1023)
30 Erkeller-Yuksel FM, Deneys V, Yuksel B, Hannet I, Hulstaert F, Hamilton C, Mackinnon H, Stokes LT, Munhyeshuli V & Vanlangendonck F. Age-related changes in human blood lymphocyte subpopulations. Journal of Pediatrics 1992 120 216-222. (https://doi. org/10.1016/s0022-3476(05)80430-5)
31 Linton PJ & Dorshkind K. Age-related changes in lymphocyte development and function. Nature Immunology 2004 5 133-139. (https://doi.org/10.1038/ni1033)
32 Fareau GG & Vassilopoulou-Sellin R. Hypercortisolemia and infection. Infectious Disease Clinics of North America 2007 21 639-657, viii. (https://doi.org/10.1016/j.idc.2007.06.001)
33 da Mota F, Murray C & Ezzat S. Overt immune dysfunction after Cushing’s syndrome remission: a consecutive case series and review of the literature. Journal of Clinical Endocrinology and Metabolism 2011 96 E1670-E1674. (https://doi.org/10.1210/jc.2011-1317)
34 Takasu N, Komiya I, Nagasawa Y, Asawa T & Yamada T. Exacerbation of autoimmune thyroid dysfunction after unilateral adrenalectomy in patients with Cushing’s syndrome due to an adrenocortical adenoma. New England Journal of Medicine 1990 322 1708-1712. (https://doi. org/10.1056/NEJM199006143222404)
Received 12 March 2022 Revised version received 13 June 2022 Accepted 6 July 2022