levels of serum protein in a large series of patients with Cushing's syndrome (CS) have been lacking. We have now evaluated, retro- spectively, the levels of ...
0021-972X/00/$03.00/0 The Journal of Clinical Endocrinology & Metabolism Copyright © 2000 by The Endocrine Society
Vol. 85, No. 9 Printed in U.S.A.
Alterations in Serum Protein Levels in Patients with Cushing’s Syndrome before and after Successful Treatment P. PUTIGNANO*, G. A. KALTSAS, M. KORBONITS, P. J. JENKINS, J. P. MONSON, G. M. BESSER, AND A. B. GROSSMAN Department of Endocrinology, St. Bartholomew’s Hospital, London EC1A 7BE, United Kingdom ABSTRACT Alteration in serum protein concentration is used commonly in clinical practice as a nonspecific indicator of underlying disease or to monitor disease activity. Although hypercortisolemia may affect protein metabolism directly or indirectly, data regarding alterations of levels of serum protein in a large series of patients with Cushing’s syndrome (CS) have been lacking. We have now evaluated, retrospectively, the levels of circulating serum albumin, globulins, total proteins, and the albumin to globulin ratio in 99 patients with endogenous CS before, immediately after, and 3, 6, and 12 months following successful treatment. Subjects with concomitant infections or other chronic diseases were excluded from the analysis. Although mean serum albumin and total protein levels were within the normal reference ranges, in general, they gradually increased
I
T IS WELL KNOWN that adrenocorticosteroids exert important, albeit complex, effects on protein metabolism (1, 2). However, although the catabolic activity of glucocorticoids on muscle and connective tissue is well documented in both animals (1–3) and humans (1, 2, 4, 5), few studies have investigated the effects of glucocorticoids on circulating hepatic proteins. Serum protein levels are among the most common biochemical markers measured routinely as screening tests for detecting underlying disease or for monitoring disease activity; alterations of their levels, although not specific, may be of diagnostic significance in patients with complex diseases such as Cushing’s syndrome (CS). A decrease in serum protein and albumin levels has been reported in a small number of patients with CS (6) and in a series of patients with major depression (7), a condition frequently associated with increased cortisol secretion. Glucocorticoids may exert a permissive physiological effect on hepatic protein synthesis (8, 9), whereas either acute or chronic overexposure may affect the net protein balance by increasing catabolic pathways (10). Taken together, these data suggest dose- and time-dependent changes of the glucocorticoids on hepatic protein metabolism. This could be clinically important because variations in the expected levels
Received November 19, 1999. Revision received March 12, 2000. Accepted June 7, 2000. Address correspondence and requests for reprints to: Prof. A. B. Grossman, Department of Endocrinology, St. Bartholomew’s Hospital, London EC1A 7BE, United Kingdom. E-mail: A.B.Grossman@ mds.qmw.ac.uk. * Present address: Second Chair of Endocrinology, University of Milan, IRCCS Ospedale San Luca, Istituto Auxologico Italiano, Milan, Italy.
after treatment with maximal values being reached at 12 months after normalization of hypercortisolemia (P ⬍ 0.0001 for both); there were no significant changes in serum globulin levels or in the albumin to globulin ratio. Patients with CS as a whole showed a weak but significant negative correlation between serum albumin and 0900 h cortisol level (r ⫽ ⫺0.303; P ⫽ 0.0035). In conclusion, our data suggest that CS is associated with a small but significant reduction in circulating serum protein levels, which are restored following treatment of hypercortisolemia, although these changes occur within the reference range. Thus, extreme alterations in serum total protein or albumin levels in patients with CS should alert physicians to the presence of concomitant pathology, and additional specific investigation should be undertaken to elucidate the cause. (J Clin Endocrinol Metab 85: 3309 –3312, 2000)
of circulating hepatic proteins in states of cortisol excess might be taken to indicate underlying disease and, thus, knowledge of their expected level in CS should aid appropriate investigation, diagnosis, and treatment. We undertook this retrospective analysis to investigate the effect of chronic endogenous hypercortisolemia on the basic indices of hepatic protein metabolism in a series of 99 patients with CS. Studies were performed before any therapy, after preoperative medical therapy with adrenolytic drugs, immediately after, and at fixed intervals following successful surgical treatment. Subjects and Methods Subjects We reviewed retrospectively the case records of all the 299 patients with CS seen at St. Bartholomew’s Hospital in the years 1966 –1997. CS was diagnosed on the basis of relevant clinical symptoms and signs, an elevated midnight sleeping serum cortisol, and a failure to suppress serum cortisol to less than 50 nmol/L on a low-dose dexamethasone suppression test (dexamethasone, 0.5 mg 6-h for 48 h) (11, 12). The source of the excessive cortisol production was confirmed with relevant imaging studies and, where appropriate, inferior petrosal sinus sampling (12, 13). Two hundred patients were excluded from the analysis because of the presence of the following potentially confounding conditions that might affect protein levels: acute or chronic infection, diabetes mellitus (fasting blood glucose, ⬎6.8 mmol/L), liver or kidney failure (with or without proteinuria), malignancy, autoimmune, chronic inflammatory, hematological or rheumatological disorders, or incomplete data. Patients receiving medication that could potentially affect protein metabolism were also excluded. Inclusion criteria were: a normal sedimentation rate, full blood count (including white blood cell count and differential), serum liver enzymes (alanine transferase, aspartate transaminase, and alkaline phosphatase), bilirubin, blood glucose, urea
3309
3310
JCE & M • 2000 Vol. 85 • No. 9
PUTIGNANO ET AL.
and creatinine levels, and a normal urinalysis and urinary protein excretion. A total of 99 patients (79 females) with CS fulfilled these criteria; eighty-one patients had pituitary-dependent CS [Cushing’s disease (CD)], whereas 10 patients had ACTH-independent cortisol excess (six adrenal adenomas and four nodular hyperplasias). Eight patients had an ectopic source of ACTH secretion (seven bronchial carcinoids and one unknown source). Patients were prepared for surgery, for at least a period of 6 weeks, with the cortisol-lowering agents metyrapone and/or ketoconazole to minimize surgical complications and promote tissue healing. A mean cortisol value (based on a five-point day curve) of 150 –300 nmol/L was taken as the target range, because this reflects restoration of a normal 24-h cortisol production rate (14). Following surgical removal of the relevant lesion, patients who were cured (cortisol levels ⬍50 nmo/L) were started on a standard glucocorticoid replacement therapy (prednisolone, either 5 mg on awakening and 2.5 mg at 1800 h; or hydrocortisone, 10 mg on awakening, 5 mg at 1200 h, and 5 mg at 1700 h) and reviewed regularly at 3-, 6-, and 12-month intervals to ensure remission and assess recovery of the hypothalamo-pituitaryadrenal axis (11, 14). Hormonal deficiencies were treated with appropriate hormonal replacement. Twenty-one patients with CD were not cured following surgery; these patients were rendered eucortisolemic with either further surgery (n ⫽ 15), external beam radiotherapy (n ⫽ 4), and/or adrenolytic medication (n ⫽ 2). Six patients who were not cured following initial surgery were eucortisolemic during subsequent follow-up. Serum albumin, globulin, and total protein levels were assessed at diagnosis (basal value), immediately preoperatively, and postoperatively (within 48 h of the operation), and thereafter at 3, 6, and 12 months. In addition, to evaluate whether these parameters may be further altered by concomitant chronic pathology, a group of 17 patients with CS and diabetes mellitus (11 females and 6 males; 15 with CD, 1 with ectopic source of ACTH secretion, and 1 with an adrenal adenoma) on long-term oral hypoglycemics were analyzed separately.
Methods All patients had complete biochemical and endocrine assessment, according to a standard protocol (11). Plasma cortisol was measured fluorometrically from 1966 –1982 and afterward by a specific RIA (15). The cortisol value used for correlation with protein levels was the mean of five measurements taken during a day curve (14). Plasma ACTH levels were measured by an in-house RIA (16). Albumin and total protein determinations were performed colorimetrically; until 1989, albumin was measured using bromocresol green dye binding (BCG) (Techmcan SMA reagents) via a Hitachi 717 analyzer (Roche Molecular Biochemicals, Lewis, East Sussex, UK); from 1989 –1995 albumin was measured using BCG (Roche Molecular Biochemicals) via a Hitachi 717 analyzer, and from 1995 onward using BCG (Instrumentation Laboratory, Warrington, Cheshire, UK) via IL 900 machinery, with coefficients of variation of 2.6% at 27g/L, 2.4% at 42g/L, and 1.9% at 52g/L, respectively. Total protein levels were measured using Technicon SMA reagents with SMA II machinery until 1989, from 1989 –1995 using the SYS 2 717/911 reagent kit (Roche Molecular Biochemicals), and from 1995 onward using the Synermed protein kit (Monitor Bioscience, Burgess Hill, West Sussex, UK) using IL 900 machinery, with a coefficient of variation of 1.4% at 42g/L, 1.5% at 67g/L, and 1.2% at 83g/L, respectively. Globulin levels were calculated as the difference between total protein and albumin concentrations. The normal ranges for albu-
min and globulin (mean ⫾ 2 sd, 35–50 g/L and 20 –25 g/L, respectively) were derived from 100 healthy adult blood donors (age, 15– 45 yr).
Statistical analysis Data are presented as the mean ⫾ sd. Statistical analysis was performed with StatView-4 software (Abacus Concepts, Berkeley, CA) by ANOVA, followed by Bonferroni/Dunn post hoc tests. Correlation analysis between basal serum cortisol and serum protein levels was performed by linear regression. Statistical significance was taken as P less than 0.05.
Results
Before surgery. The mean total serum protein level at initial presentation, although within the low part of the normal range, was significantly lower than values obtained at subsequent restoration of normal cortisol levels for the group as a whole (Table 1). Mean serum albumin level, although also within the normal range, was also significantly lower compared with values obtained after successful treatment (Table 1); there was no difference between total protein or albumin values or in the other routine biochemical indices obtained at initial presentation according to the etiology of the hypercortisolemia or the concomitant presence of diabetes mellitus (data not shown). Preoperative medical treatment with cortisol-lowering drugs (i.e. metyrapone and/or ketoconazole) resulted in a modest rise in mean serum albumin and globulin levels in the group as a whole, but this did not reach statistical significance (Fig. 1). Postoperatively. Immediately postoperatively all patients demonstrated a significant decrease in serum albumin levels (P ⫽ 0.0031), compared with values obtained after restoration of normal circulating cortisol levels, unassociated with any change in serum total protein or globulin levels; however, the mean serum albumin was not significantly different to that at initial presentation. Thereafter, there was a rise in both mean serum albumin and total protein levels at 3, 6, and 12 months, although there were no significant differences between any of these time points; maximum values were obtained at 12 months, and these were significantly higher compared with the basal values (P ⬍ 0.0001 and P ⬍ 0.002, respectively) (Table 1 and Fig. 1). There was a slight rise in serum globulin levels, which did not reach statistical significance. There was no difference in either total protein, albumin and/or globulin values, or in the any other routine biochemical indices according to the etiology of the hypercortisolemia, or between patients who obtained biochemical
TABLE 1. Protein levels in patients with Cushing’s syndrome before and after treatment Variable
Basal value
Pre-op
Post-op
TP Alb Glb A/G Rt
65.5 ⫾ 0.69 41.0 ⫾ 0.48 24.8 ⫾ 0.51 1.70 ⫾ 0.04
68.1 ⫾ 0.73 42 ⫾ 0.64d 25.9 ⫾ 0.61 1.66 ⫾ 0.04
65.1 ⫾ 0.76 39.3 ⫾ 0.58 25.8 ⫾ 0.52 1.57 ⫾ 0.04
3 months
6 months
69.1 ⫾ 0.77 42.6 ⫾ 0.59b 26.6 ⫾ 0.55 1.66 ⫾ 0.05
a,b
TP, Total protein; Alb, albumin; Glb, globulin; A/G Rt, albumin/globulin ratio. a P ⬍ 0.002 (vs. basal values). b P ⬍ 0.0001. c P ⬍ 0.0001. d P ⬍ 0.0033 (vs. post-op values).
12 months
70.4 ⫾ 0.81 43.5 ⫾ 0.52b 26.9 ⫾ 0.7 1.65 ⫾ 0.04
b,c
Normal range
69.9 ⫾ 0.75 44 ⫾ 0.74a,c 26.5 ⫾ 0.56 1.68 ⫾ 0.05 b,c
60 –75 g/L 35–50 g/L 20 –25 g/L 1.5–2.2
SERUM PROTEIN LEVELS IN CS
FIG. 1. Alterations of albumin levels in patients with CS at presentation, preoperatively, immediately postoperatively, and during the follow-up.
cure and patients who were not cured but obtained normalization of circulating cortisol levels (data not shown). There was a strong positive correlation between the mean cortisol value obtained from the five-point day curve and the 0900 h value (r ⫽ ⫹0.8, P ⬍ 0.001). There was a weak but significant negative correlation between both serum albumin (Fig. 2) and serum total proteins and 0900 h serum cortisol for the group as a whole (r ⫽ ⫺0.303, P ⫽ 0.0035 and r ⫽ ⫺0.372, P ⫽ 0.0003, respectively). There was no correlation between serum globulin and cortisol. Discussion
A number of studies in the rat have focused on the effects of corticosteroids on the major protein reservoir, skeletal muscle, generally concluding that glucocorticoid excess is associated with a fall in muscle protein, suppressing protein synthesis and increasing catabolism (1–3). However, the effects of hypercortisolemia on splanchnic protein metabolism are more controversial; glucocorticoids can exert differing effects on muscle and visceral proteins, promoting proteolysis in muscle and decreasing protein synthesis in the gut (1–3). In terms of circulating serum hepatic proteins, in experimental animal models glucocorticoids acutely stimulate hepatic protein synthesis, especially acute phase proteins (8, 9) and may enhance immunoglobulin production (17); in contrast, direct (6) and indirect (7) evidence in humans suggest that chronic cortisol excess is associated with decreased total serum protein and albumin concentrations. In this series of patients with CS, we found significantly lower serum albumin and total protein levels compared with those recorded at different intervals after restoration of normal serum cortisol levels. The absence of a significant change in the overall globulin synthesis cannot exclude alterations in globulin electrophoretic fractions that may be of particular clinical significance, because patients with CS may be immunocompromised (18) and routine inflammatory indices may be misleading. These changes were small in magnitude, normalized rapidly with medical or surgical therapy, and would not, in most cases, have produced levels outside of the normal range for our population. Therefore, as the magnitude of the changes is relatively minor, it is likely to be
3311
FIG. 2. Univariate linear regression analysis of serum albumin plotted against cortisol levels in patients with CS: F, patients with CD; E, patients with ectopic ACTH syndrome; Œ, patients with ACTHindependent CS. Shaded area reflects the reference range values for serum albumin (35–50 g/L).
clinically significant only at the highest levels of serum cortisol. In general, while we did not systematically relate serum cortisol levels to clinical features, we have noted an approximate correlation between the integrated levels of cortisol (related to the mean cortisol levels and the length of history) and the clinical manifestations of CS. It is, therefore, likely that patients with the most severe manifestations of CS would be those showing lowest levels of serum protein and serum albumin. The endocrine regulation of human protein metabolism is complex, being determined by a number of hormones such as adrenal steroids, insulin, thyroid hormones, GH, insulinlike growth factor I, sex steroids, glucagon, and catecholamines (1, 2). Because glucocorticoids participate in modulating this hormonal pattern, acute or chronic hypercortisolemia may affect albumin or globulin synthesis both directly and/or through interactions with other hormones. In our series, basally there was no difference in serum protein levels between diabetic and nondiabetic patients, therefore, probably excluding any major role of insulin resistance in affecting protein metabolism in CS. However, patients with CS exhibit subnormal GH levels and responses to dynamic tests (19), and GH deficiency might be considered as a contributor to the impaired nitrogen balance because it has previously been demonstrated that GH has powerful protein anabolic effects in patients with CS (20). Nevertheless, it is likely that in vivo the GH/insulin-like growth factor I axis exerts only very minor effects on hepatic protein synthesis (21). Similarly, although alterations of either thyroid or gonadal hormones have been shown to influence net protein metabolism (1), it is unlikely that may have affected our results as anterior hormonal pituitary deficiencies were sought for and, where present, adequately treated with hormonal replacement. It is also possible that extra-hormonal factors might have contributed to decreasing albumin and total protein levels in our patients. These include possible subtle protein-calorie malnutrition, immunosuppression, and changes in extracel-
3312
PUTIGNANO ET AL.
lular volume, which are often present in CS. Thus, Lotsikas et al. (22) have recently demonstrated that cure of CS is associated with a decrease in interleukin 10 production, and these or other cytokines may be involved in changes in hepatic albumin synthesis. Whereas it has been suggested that normalization of serum cortisol levels in CS may cause some degree of hemodilution, this would have been associated with a fall in protein and albumin levels, rather than the rise that we observed. Furthermore, measurement of changes of total exchangeable sodium in adult patients with CS suggests that the role of sodium retention is minor (23). It is, therefore, unlikely that the changes we have recorded may be secondary to alterations in blood volume. In conclusion, we have observed in our patients with chronic hypercortisolememia a small but significant reduction in both serum albumin and total protein levels, although both remained within the normal range in the great majority of patients. Only in patients with markedly elevated serum cortisol levels was the reduction of serum proteins of any magnitude. The fall in both albumin and total protein was restored with treatment of the hypercortisolemia. Although it is probable that these changes are a direct consequence of the elevated circulating cortisol levels, their precise mechanism is currently undefined. The minor nature of the changes reported suggests that any significant abnormality in serum albumin or globulin levels in a patient with CS requires specific investigation to exclude another concomitant underlying pathology. Acknowledgments We are grateful to Peter Brownie and Dr. Francesca Pecori Giraldi for skillful statistical assistance.
References 1. Umpleby AM, Russell-Jones DL. 1996 The hormonal control of protein metabolism. Protein metabolism. Baillieres Clin Endocrinol Metab. 10:551–570. 2. De Feo P. 1996 Hormonal regulation of human protein metabolism. Eur J Endocrinol. 135:7–18. 3. Hanoune J, Chambant AM, Josipowicz A. 1972 The glucose effect and cortisone upon rat liver and muscle protein metabolism. Arch Biochem Biophys. 148:180 –186. 4. Khaleeli AA, Edwards RHT, Gohil K, et al. 1983 Corticosteroid myopathy: a clinical and pathological study. Clin Endocrinol (Oxf). 18:155–166.
JCE & M • 2000 Vol. 85 • No. 9
5. Simmons PS, Miles JM, Gerich JE, Haymond MW. 1984 Increased proteolysis. An effect of increases in plasma cortisol within the physiological range. J Clin Invest. 73:412– 420. 6. Jeejeebhoy KN, Ho J, Mehia R, et al. 1977 Effects of hormones on the synthesis of 1 (acute phase) glycoprotein in isolated rat hepatocytes. Biochem J. 168:347–352. 7. Van Gool J, Boers W, Sala M, et al. 1984 Glucocorticoids and catecholamines as mediators of acute-phase proteins, especially rat-macrofoetoprotein. Biochem J. 220:125–132. 8. Heim WG, Ellenson SR. 1965 Involvement of the adrenal cortex in the appearance of rat slow 2 globulin. Nature. 208:1330 –1331. 9. Sato T, Tajiri J, Shimada T, Hiramatsu R, Umeda T. 1984 Abnormal blood chemistry data in Cushing’s syndrome: comparison with those for fatty liver. Endocrinol Jpn. 31:705–710. 10. Maes M, Wauters A, Neels H, et al. 1995 Total serum protein and serum protein fractions in depression: relationships to depressive symptoms and glucocorticoid activity. J Affect Disord. 34:61– 69. 11. Trainer PJ, Besser GM. 1995 The Barts endocrine protocols, ed 1. Edinburgh: Churchill Livingstone. 84 – 87. 12. Newell-Price J, Trainer P, Besser M, Grossman A. 1998 The diagnosis and differential diagnosis of Cushing’s syndrome and pseudo-Cushing’s states. Endocr Rev. 19:647– 672. 13. Kaltsas GA, Giannulis MG, Newell-Price JE, et al. 1999 A critical analysis of the value of simultaneous inferior petrosal sampling in Cushing’s disease and the occult ectopic adrenocorticotropin syndrome. J Clin Endocrinol Metab. 84:487– 492. 14. Trainer PJ, Eastment C, Grossman AB, Wheeler MJ, Perry L, Besser GM. 1993 The relationship between cortisol production rate and serial serum cortisol estimation in patients on medical therapy for Cushing’s syndrome. Clin Endocrinol (Oxf). 39:441– 443. 15. Cunnah D, Jessop DS, Besser GM, Rees LH. 1987 Measurement of circulating corticotrophin-releasing factor in man. J Endocrinol. 113:123–131. 16. Rees LH, Cook DM, Kendall JW, et al. 1971 A radioimmunoassay for rat plasma ACTH. Endocrinology. 89:254 –261. 17. Grayson Y, Dooley NJ, Koski IR, et al. 1981 Immunoglobulin production induced in vitro by glucocorticoid hormones: T cell-dependent stimulation of immunoglobulin production without B cell proliferation in cultures of human peripheral blood. J Clin Invest. 68:1539 –1547. 18. Yanovski JA, Cutler Jr GB. 1994 Glucocorticoid action and the clinical features of Cushing’s syndrome. Endocrinol Metab Clin North Am. 23:487–509. 19. Leal-Cerro A, Pumar A, Garcia-Garcia E, Dieguez C, Casanueva FF. 1994 Inhibition of growth hormone release after administration of GHRH and GHRP-6 in patients with Cushing’s syndrome. Clin Endocrinol (Oxf). 41:649 – 654. 20. Bowes SB, Umpleby M, Cummings MH, et al. 1997 The effect of recombinant human growth hormone on glucose and leucine metabolism in Cushing’s syndrome. J Clin Endocrinol Metab. 82:243–246. 21. Hoffman MD, Pallasser R, Duncan M, Nguyen VT, Ho KYK. 1998 How is whole body protein turnover perturbed in growth hormone-deficient adults. J Clin Endocrinol Metab. 83:4333– 4349. 22. Lotsikas AJ, Elenkov IJ, Link A, Papanicolaou DA. Cure of Cushing’s disease is associated with an increase in IL-12 and a decrease in IL-10 cytokine production. Presented at the 80th Annual Meeting of The Endocrine Society, New Orleans, LA, 1998; 337. 23. Ritchie CM, Sheridan B, Frazer R, et al. 1990 Studies on the pathogenesis of hypertension in Cushing’s disease, and acromegaly. Q J Med. 280:855– 867.
