Introduction

Cushing’s syndrome was first described in 1912 by Harvey Cushing [1]. A variety of clinical features and laboratory findings are attributed to glucocorticoid (GC) excess in Cushing’s syndrome with variable reported proportions, with obesity and weight being the most frequent features, but also diabetes mellitus, hypertension, hirsutism, osteoporosis, hypokalemia, and more [2–4]. The diagnosis is complicated due to high prevalence of similar clinical symptoms and signs in patients without the disorder [4].

One of the characteristic laboratory findings associated with Cushing’s syndrome is elevated white blood cells (WBC) counts, as described early in the 1940s. In a study by de la Blaze et al. [5] ten patients with Cushing’s disease were found to have higher leukocyte counts than normal subjects, with a tendency for higher percentage of polymorphonuclear (PMN) cells and lower percentage of lymphocytes.

However, there is insufficient data regarding the long term effects of GC’s excess on hematologic indices, apart from the observation of high leukocyte counts in patients with Cushing’s syndrome or following exogenous administration of GC’s. Moreover, no data is available regarding the true prevalence of this finding and alterations in other blood components as evaluated by complete blood counts.

We conducted a retrospective study on 26 patients with Cushing’s disease to describe the true prevalence of leukocytosis in Cushing’s disease, to compare alterations in blood cell components before and after treatment for hypercortisolism, and to correlate WBC counts with disease activity.

Methods

We retrospectively studied 26 consecutive cases diagnosed with Cushing’s disease and followed in the Endocrine Institute, Rabin Medical Center, Beilinson Hospital, Israel. Demographic information including age at diagnosis, gender, urinary free cortisol levels (UFC), complete blood count (CBC) (hemoglobin, WBC’s, neutrophils, lymphocytes and platelets) at baseline and following treatment were collected. Baseline parameters were collected from the year preceding diagnosis. The study was approved by the ethical institutional board at the Rabin Medical Center.

The diagnosis of Cushing’s disease was confirmed by at least two elevated UFC collections (≥2 times the upper limit of normal; ULN), high or normal range ACTH levels, and a pituitary lesion consistent with an adenoma seen on MRI (in 24/26 of the patients). Remission was defined as at least 2 measurements of UFC within the normal range achieved either by pituitary surgery, medical treatment, or both. Only then WBC’s measurements were performed. In patients who were not in full hormonal remission WBC’s data was collected during the lowest measurements of UFC.

CBC’s were measured using an automatic ABX MicrosCRP 200 analyzer (Clinical Laboratory International, Brussels, Belgium). The within-run CV for WBC was 2.7 % using this assay. UFC was measured using radioimmunoassay for the determination of cortisol in urine (Beckman Coulter) with high antibody specificity for cortisol and low cross reactivity against other naturally occurring steroids. The analytical sensitivity is 10 nM. The intra- and inter-assay CVs were 5.8 and 9.2 %, respectively. Reference levels for 24 h urine collection are 38–208 nmol/24 h.

Statistical analysis

Continuous data are presented as means and standard deviations. Continuous variables were compared using Student’s t test or Mann–Whitney U test, as appropriate. The relationships between UFC and blood count components were analyzed using Pearson’s correlation. Analysis was performed using SPSS software (Version 20; SPSS Inc., Chicago, IL). A p value of less than or equal to 0.05 was considered statistically significant.

Results

Patients’ characteristics

The study included 26 patients with Cushing’s disease, 19 females and 7 males, with a mean age at presentation of 39.8 ± 12.7 years; 16 patients had microadenomas, 8 had macroadenomas, and in two cases there was no visible adenoma on MRI (Table 1). All patients except two had transsphenoidal surgery (TSS). The two patients who did not have surgery were treated medically. Twenty patients achieved remission, including two patients who experienced transient remission. The mean baseline UFC levels was 5.4 ± 6.9 × ULN and dropped to 1 ± 1.1 × ULN following treatment.

Table 1 Clinical characteristics of 26 patients with Cushing’s disease included in the cohort, and 28 events (*) of hypercortisolemia (including 2 patients with disease relapse)
Full size table

Blood count changes

Mean WBC’s before treatment was 10,500 ± 2,600 cells/μl (normal <10,800 cells/μl) with 40 % having elevated WBC’s, and dropped to 8,400 ± 1,900 cells/μl (p < 0.05), with 11 % having elevated WBC’s after treatment. The mean neutrophil count was 7,600 ± 2,600 cells/μl before treatment and dropped to 5,300 ± 1,700 cells/μl after treatment (p < 0.05), and the mean lymphocyte count was 2,000 ± 600 cells/μl at baseline and increased to 2,300 ± 600 cells/μl following treatment (p < 0.05). Hemoglobin was 13.7 ± 1.2 g/dl at baseline and decreased to 12.8 ± 1.4 g/dl (p < 0.05) (Fig. 1).

Fig. 1
figure 1

Mean blood cell counts in patients with Cushing’s disease at diagnosis and following treatment; the values are the average of a number of blood counts both before and after surgical and/or medical treatment. WBC white blood cells (×1,000 cells/μl), NEUT neutrophils (×1,000 cells/μl), LYMPH lymphocytes (×1,000 cells/μl), HGB hemoglobin (g/dl). *p < 0.05

Full size image

There was no correlation between pre-treatment UFC and baseline WBC’s levels (Fig. 2). However, there was a significant positive correlation between decrease in UFC secretion (ΔUFC) and change in WBC’s (ΔWBC) following treatment (r = 0.67, p < 0.01, 2-tailed). It is of note that this strong correlation is mostly due to one patient (Fig. 2, right column) who had very high UFC levels before treatment (up to 34 × ULN) together with high WBC counts, and who went into remission following treatment with normalization of his WBC levels. Upon further analysis of this case, there were 14 pre-treatment CBC’s showing consistently high WBC levels (>15,000/μl), and 7 post-therapy CBC’s showing consistently normal WBC levels (<10,000/μl) upon normalization of UFC levels. According to this data, the patient’s change in WBC levels was definitely attributable to the change in UFC.

Fig. 2
figure 2

Individual relationships between baseline UFC and WBC counts among patients with Cushing’s disease. WBC white blood cells (×1,000 cells/μl), UFC (×ULN)

Full size image

Subgroup analysis

We performed sub-group analysis for the cohort of patients with elevated baseline WBC count (defined as ≥10,800 WBC’s/μl; n = 11) and patients with normal baseline WBC count (n = 15). There was no difference in age, gender, or adenoma size between the two groups, and the mean baseline UFC was not significantly different.

Mean WBC’s in the group with elevated baseline count was 12,970 ± 2,260 cells/μl and dropped to 9,620 ± 2,000 cells/μl (p < 0.05) following successful treatment (Fig. 3a). Baseline neutrophil count was 9,700 ± 2,600 cells/μl and dropped to 6,300 ± 1,600 cells/μl (p < 0.05). There were no significant alterations in lymphocytes and platelets counts or hemoglobin concentrations (Fig. 3a).

Fig. 3
figure 3

Mean blood cell counts in patients with elevated baseline WBC’s (>10,800 cells/μl) (a) and normal baseline WBC’s (b), at diagnosis and following treatment; the values are the average of a number of blood counts both before and after treatment. WBC white blood cells (×1,000 cells/μl), NEUT neutrophils (×1,000 cells/μl), LYMPH lymphocytes (×1,000 cells/μl), HGB hemoglobin (gr/dl). *p < 0.05

Full size image

In the group with normal baseline counts, the mean WBC’s was 9,000 ± 1,400 cells/μl and dropped to 7,700 ± 1,300 cells/μl (p < 0.05); neutrophil count was 6,200 ± 1,500 cells/μl and dropped to 4,600 ± 1,400 cells/μl (p < 0.05); the lymphocyte count raised from 1,800 ± 580 cells/μl at presentation to 2,280 ± 580 cells/μl after treatment (p < 0.05); hemoglobin was 13.7 ± 1.2 g/dl and decreased to 12.8 ± 1.4 g/dl (p < 0.05) following treatment (Fig. 3b). Platelets count didn’t change significantly.

In a separate analysis of patients who were in remission (n = 20) and those who did not achieve cortisol normalization (n = 8, including 2 patients with disease relapse), the baseline UFC was 4 ± 3.8 × ULN and 8 ± 10 × ULN, respectively; the mean lowest UFC after treatment was 0.6 ± 0.3 × ULN and 2 ± 1.4 × ULN, respectively. There was no significant difference in mean WBC’s at baseline between patients who achieved remission and those who did not, nor a difference in the magnitude of the post treatment change.

Discussion

We studied the effects of high cortisol levels on the different blood cell components in patients with Cushing’s disease. Leukocytosis was present in 40 % of our cohort patients, and there was a significant decrease of 20 % in WBC’s and 30 % in neutrophil counts upon UFC decrease following surgical and/or medical treatments. Moreover, there was a significant correlation between the decrease in cortisol secretion and change in WBC’s counts. These CBC alterations in patients with Cushing’s disease could potentially serve as another marker for hormonal response to treatment, and remission during follow-up. However, many factors including infections, stress, medications, physiological fluctuations and others can affect WBC counts making this measurement a non-accurate tool for assessing treatment efficacy or disease relapse.

Cortisol exerts its main effects through binding to the nuclear glucocorticoid receptor, inducing a cascade of intra-cellular signaling and leading to protein synthesis [6]. Cortisol is secreted in response to stress and pain signaling and due to cytokine stimulation, resulting in hypothalamic–pituitary–adrenal activation. Cortisol binds with high affinity to the glucocorticoid receptor and this complex inhibits inflammation both through genomic and nongenomic mechanisms. The rise in cortisol levels in the absence of active inflammation, like in Cushing’s syndrome, can cause immune suppression associated with susceptibility to severe infections [7]. This is associated with lymphopenia seen after administration of adrenocorticotrophic hormone (ACTH), as reported many years ago [8].

Glucocorticoid-induced leukocytosis has been attributed to several mechanisms including increased release of PMN’s from bone marrow to the circulation [9], delayed neutrophil apoptosis in the circulation [10], and reduced cell transfer from blood vessels into tissues [11]. Another mechanism playing a role is an influx of PMN’s from the intravascular marginated pools of PMN’s [9]. These cells are found in the lungs, liver, spleen and bone marrow, in close contact with the endothelial cells of small blood vessels. Stress signals like cortisol and catecholamines can activate mature granulocytes in these marginated pools to undergo rapid demargination, resulting in acute elevation of circulating leukocytes. Cortisol results also in prolongation of their intravascular half-life and delays their disappearance from the circulation [9]. Some of the effects of GC’s on blood cells are recognized as their clinical use is very common, but their long-term effects were not carefully investigated, and most published studies on their hematological effects were designed to show acute effects. Administration of dexamethasone intravenously or orally to healthy subjects results in an increase of neutrophil count and lymphopenia, the increment is in a linear relation to the dose introduced intravenously but not orally [12]. The long-term effect of GC’s on blood counts shown in our Cushing’s patients argues against the demargination effect theory of GC’s on granulocytes that was shown shortly after a single dose of dexamethasone [9, 12], and supports other studies displaying direct effects of stimulatory factors on bone marrow release of new PMN’s [13].

Importantly, patients with unexplained persistent leukocytosis are often studied extensively for various hematological diseases, sometimes without success. These patients should be referred also for cortisol measurements, and Cushing’s syndrome should be part of the regular work-up for patients with unexplained leukocytosis, as our series reveals.

The action of GC’s on lymphoid cells is well known and they are used successfully in the treatment of leukemia and lymphoma with high lymphocyte counts [14]. In vitro studies demonstrated that GC’s induce apoptosis of human mature T lymphoid cells [15]. In our cohort lymphocyte counts increased following cortisol normalization, in line with the known effects of elevated cortisol on human lymphocytes. Effects of GC’s on platelets are not well recognized and in some reports patients with elevated cortisol did not show significant alterations in their platelet counts [16], whereas other series showed increased numbers of platelets in Cushing’s patients [17]. Although it was previously thought that platelet cells lack GC’s receptors as they lack a nucleus, it is now well recognized that GC’s affect platelet function through binding to the GC receptors expressed in human platelets [18], activating nongenomic processes.

The role of GC’s administration in aplastic anemia and other forms of non-immune hemolytic anemia implies a direct effect on bone marrow erythropoiesis [19]. GC’s exert their effects on the erythroid line by enhanced response of red blood cells to erythropoietin or directly by stimulating cell proliferation [13]. In our cohort, hemoglobin levels significantly decreased following treatment to normalize cortisol secretion, in agreement with these observations.

Like other features of Cushing’s syndrome such as overweight, diabetes, hypertension, osteoporosis, plethora or edema which severity is not linearly related to serum cortisol or UFC levels, we were not able to show any correlation between pre-treatment UFC levels and baseline WBC’s counts (Fig. 2). This can be partially explained by GC’s receptor polymorphism which may result in relative resistance or sensitivity to GC’s [20], leading to a different in vivo response to cortisol. This is also a well-known phenomenon in the response to GC’s treatment in autoimmune diseases [21].

Our study has several limitations. First, it is a retrospective study, and the blood tests were done routinely and not for the study. Also, the number of blood tests available for each subject varied between patients. Second, a relatively small number of patients are included in the study due to the rarity of the disease, and that may cause the lack of correlation between baseline WBC counts and disease severity. Third, the study includes only patients with Cushing’s disease, so any conclusion can’t be instantly drawn to patients with ACTH-independent Cushing’s syndrome, although previous studies showed that exogenous administration of GC’s exert the same effects on blood cell counts.

In conclusion, elevated WBC’s is common in patients with Cushing’s disease. During remission or hormonal control of the disease there is significant decrease in neutrophil counts and in hemoglobin concentration together with a rise in lymphocyte numbers. These parameters may help to identify a trend towards improvement or deterioration in cortisol control. A larger study may clarify the effects of these changes on disease prognosis and other features of Cushing’s syndrome.