Data characterizing the erythrocytic profile in patients with CS is still limited and conflicting; the literature mostly focuses on how these parameters could potentially differ by sex, but not by the CS etiology [9, 10]. In this retrospective population study, the largest to date and focusing just on patients with CD and aCS (excluding adrenal cancer and ectopic CS), hemoglobin and hematocrit were significantly higher in patients with CS compared to matched controls, regardless of the etiology of CS or sex (Fig. 1). We have also identified for the first time an influence of DM presence on selected erythrocytic parameters in patients with CS.
Glucocorticoids are known erythropoiesis stimulators, although the mechanisms are not entirely understood. In vitro, glucocorticoids enhance the response of erythroid progenitor cells to erythropoietin [21–23]. Studies have identified dexamethasone’s role in augmenting the expression of genes involved in erythropoiesis [24–26], upregulating macrophage activity in the stress erythropoiesis response [27], and promoting embryonic stem cell differentiation to erythroid precursor [28]. Activation of glucocorticoid receptors launches proliferation of erythroid progenitor cells to attain stress erythrocytosis, which has been observed in vitro [29, 30] and in vivo alike; in vivo, endogenous glucocorticoid release and receptor-binding are triggered by states of physiologic stress such as hemorrhage, hemolysis or hypoxia [11, 20, 25].
Varricchio et al. identified a unique class of CD34 + erythrocytes that expressed a constitutively active glucocorticoid receptor, resulting in increased erythrocytosis. Glucocorticoid excess prompts the erythropoiesis cascade onset, culminating in erythrocytosis; this drives the amplified production of the identified CD34 + erythroid progenitor cells exhibiting an altered, constitutively activated glucocorticoid receptor [11]. The positive feedback loop of hypercortisolism in CD leads to continuous self-amplification and resultant polycythemia until remission, often post-surgery, interrupts the cycle [12].
One could extrapolate that without excess glucocorticoid inducing these stimulators, hemoglobin and hematocrit levels would normalize to levels seen in controls when patients attain remission. However, we did not observe any significant difference, though a trend towards decreases in hemoglobin and hematocrit was noted post-remission in both CD and aCS when compared to patients who did not achieve remission status or had disease recurrence, within the same period. In contrast, patients with persistent or recurrent disease maintained elevated hematocrit, which was significant in CD (42.45% [p = 0.0049]), but not in aCS. Our data aligned with few other studies that also found statistically significant near-normalization of erythroid parameters among patients with CD post-remission, regardless of sex [9, 10].
Intriguingly, we found that the baseline levels of hemoglobin and hematocrit, identified as measurements 12 months before diagnosis, could potentially act as predictors for remission in CD (Fig. 2). Those patients who were in remission had baseline erythrocytic values of hemoglobin 13.62 g/dL and hematocrit 41.00%, both parameters significantly lower than those who had persistent hypercortisolemia, with starting values of 14.27 g/dL (p = 0.0204) and 44.08% (p = 0.0003). In contrast, the baseline erythrocytic profile did not differ in patients with aCS in function of remission status. We postulate that different disease stages in CD versus aCS at diagnosis could play a role. Hirsch et al. suggested that CD was previously more widely detected owing to a more distinct clinical presentation [31]. However, more aCS incidentalomas are also being identified in earlier stages, due to the advancement of abdominal imaging and guidelines to appropriately screen all tumors for hormonal hypersecretion [32, 33]. This aligns with a meta-analysis that has identified a shorter time to diagnose aCS compared to CD, although the result was not significant [34]. This approach can potentially reduce the exposure to excess cortisol over time [31, 35]. Despite these differences, decreases in erythrocytic parameters with treatment were seen in all patients, regardless of remission status.
Data on sex differences in erythrocytic profile patients with CS is still debatable. Our results align with those observed in previous studies for females [9, 10] who had significantly higher median hemoglobin and hematocrit (p < 0.0001 for both) at diagnosis compared with controls, regardless of etiology (Fig. 3).
However, our results differed for males with CS in other studies [9, 10]. Higher values of hemoglobin (p = 0.0341) and hematocrit (p = 0.0128) were seen in male patients versus controls, while Ambrogio et al. [9] and Detomas et al. [10] identified lower mean hemoglobin and hematocrit in male patients with CS, compared to the control group. We hypothesize that the differences are due to both the inclusion of patients with ectopic neuroendocrine tumors and adrenal cancer, which represented over a third of the patients with CS in Detomas et al. study [10], and the dissimilar control populations. Our control group was composed of subjects that were age, sex, BMI and socioeconomic status-matched to our patient population, compared to a database of over 3000 healthy subjects [9] or age and sex-matched controls with inactive pituitary or adrenal tumors [10]. As patients and controls were also matched by BMI in our study, mean BMI of control group was 30.15 ± 6.98 kg/m2, comparable to patients with CS (31.15 ± 7.76 kg/m2); obesity alone could explain the tendency towards anemia (due to iron deficiency or chronic disease) in the control group [36, 37]. Interestingly, in Detomas et al. study, BMI was less than 30 for both patients with CS and controls (29.0 vs. 28.0 kg/m2, respectively) [10]. Importantly, our controls were also matched by socioeconomic status which can infer a relatively similar nutritional status, also important in erythropoiesis [38–40].
We also analyzed the effects of DM on the erythrocytic parameters in patients with CS. Anemia has been reported in 14 to 45% of patients with DM of various ethnic populations, generally attributed to the disease’s detrimental effects on the kidneys [41, 42]. In our study, 28.46% of CS patients who had DM as one of their comorbidities shared a similar value of hemoglobin to control subjects, but elevated relative hematocrit (p = 0.0419). Meanwhile, patients with CS without DM had significantly greater values of these erythrocytic parameters compared to control subjects (p < 0.0001). This demonstrates how DM and CS had opposing effects on tugging erythrocytic parameters toward anemia or polycythemia, respectively.
We cannot exclude the potential role of hypogonadism on erythroid parameters, but our study had limited data for testosterone levels and physical findings of hypogonadism. Lower testosterone levels are thought to exert a more significant inhibitory impact on erythropoiesis compared to the stimulatory effect of cortisol [10]. Ambrogio et al. suggested that the anemia at baseline among males with CD was due to associated hypogonadism, diagnosed in 81% of their patients [9]. Additionally, glucocorticoid receptor resistance has been described as an eventual response to chronic excessive receptor activation, which may also contribute to anemia in males with CS [11, 43].
Our study has several strengths including the largest dataset built from a CS patient population (397 patients), nearly doubling the previously largest study [10], both increasing the statistical power and expanding overall data points for analyses: disease etiology, remission status, sex, and diagnosis of DM. Furthermore, we had available a large well-matched population, including for BMI and with all parameters for erythrocytes. Importantly, in contrast with other studies, we have included just patients with CD and aCS, thus excluding potential confounding factors from patients with ectopic ACTH-secreting neuroendocrine tumors or adrenal carcinoma. Notable limitations of our study include its retrospective approach though selection bias is smaller in large population databases), missing data for some patients regarding disease etiology, remission status, gonadal function, patients lost to follow-up, and lack of regularly scheduled blood tests for patients with CS in remission.
In conclusion, patients with CS of both pituitary and adrenal origins have elevations in hemoglobin and hematocrit; this has been confirmed in both sexes in our study. Presence of DM in patients with CS had significant effect on hematocrit, but not hemoglobin. Interestingly, we also found that patients with CD who had lower baseline hemoglobin were more likely to attain remission. An implied link between disease remission and hemoglobin/hematocrit levels may be considered, as patients with persistent or recurrent disease retained elevated levels, while patients in remission had levels that mirrored the control group. A standardized timeline for blood testing starting at diagnosis, treatment initiation, and regular intervals thereafter would be beneficial. Future studies are indicated to corroborate our findings and explore changes in hemoglobin and hematocrit as potential predictors and/or markers for remission.