Osteoporos Int (2008) 19:941–949 DOI 10.1007/s00198-007-0514-x
ORIGINAL ARTICLE
Skeletal differences in bone mineral area and content before and after cure of endogenous Cushing’s syndrome L. Fütő & J. Tőke & A. Patócs & Á. Szappanos & I. Varga & E. Gláz & Z. Tulassay & K. Rácz & M. Tóth
Received: 3 August 2007 / Accepted: 26 October 2007 / Published online: 28 November 2007 # International Osteoporosis Foundation and National Osteoporosis Foundation 2007
Abstract Summary We examined bone densitometric data in a fouryear follow-up period before and after the cure of CS. Plasma cortisol concentrations were similar, but the duration of estimated glucocorticoid excess was longer in patients with prevalent bone fractures compared to those without fractures. After therapy of CS, bone area, BMC and BMD increased significantly at the LS and femur during follow-up, but they decreased at the forearm, suggesting redistribution of bone minerals from the peripheral to the axial skeleton. Introduction Only a few studies report the changes in bone mineral density (BMD) after the cure of Cushing’s syndrome (CS). Methods Forty-one patients with Cushing’s disease, 21 patients with adrenal CS and 6 patients with ectopic CS were prospectively enrolled. BMD, bone mineral content (BMC) and bone area were measured by DXA. Results No significant correlations were found between serum cortisol concentrations and baseline bone densitometric data. After successful therapy of CS, bone area and BMD increased significantly at the lumbar spine (LS) and L. Fütő Department of Internal Medicine, Ferenc Markhot Hospital, Eger, Hungary J. Tőke : Á. Szappanos : E. Gláz : Z. Tulassay : K. Rácz : M. Tóth (*) 2nd Department of Internal Medicine, Semmelweis University, Szentkirályi u. 46., Budapest, Hungary, H-1088 e-mail:
[email protected] A. Patócs : I. Varga Molecular Medicine Research Group, Hungarian Academy of Sciences and Semmelweis University, Budapest, Hungary
femur during follow-up, but they decreased at the forearm. The progressive increase in BMC at the LS had a significant negative correlation with the change of the BMC of radius in the first and second follow-up years. The change in the body mass index was an independent predictor for changes in BMC both at the LS and at the forearm at the second year of remission. Conclusions The regional differences and the time-dependent changes of BMC suggest that the source of marked increase in axial BMC after the cure of CS is, at least partly, due to the redistribution of bone minerals from the peripheral to the axial skeleton. Keywords Dual-energy X-ray absorptiometry . Endogenous Cushing’s syndrome . Fractures . Redistribution of bone minerals after cure
Introduction Endogenous Cushing’s syndrome (CS) is a well-known, but infrequent cause of secondary osteoporosis. Previous studies on bone mineral density (BMD) in patients with Cushing’s syndrome typically included a few dozens of patients [1–10]. In 2006 Tauchmanova et al. reported baseline BMD results of the largest cohort, including 37 patients with Cushing’s disease (CD), 31 patients with adrenal Cushing’s syndrome (ACS) and six patients with ectopic Cushing’s syndrome (ECS) [11]. However, it is still uncertain, whether these three forms of endogenous Cushing’s syndrome are associated with similar degree of bone loss. Cortisol hypersecretion itself leads to increased bone resorption and, perhaps more importantly, to suppressed bone formation by its direct effect on osteoblasts [12], and it may also cause hypogonadotropic hypogonadism.
DO00514; No of Pages
942
Osteoporos Int (2008) 19:941–949
Overproduction of adrenal androgens should be taken also into account in patients with ACTH excess and in mixed cortisol- and androgen-producing adrenal neoplasms. In addition, hypercortisolism may directly suppress the growth hormone-IGF-1 system, and GH-deficiency may be frequently detected in patients with pituitary macroadenomas. Also, both ACTH- and cortisol-producing tumours seem to be heterogeneous regarding their hormoneproducing activities, and patients may have different sensitivity to glucocorticoids. Both uni- and bilateral adrenal lesions may cause subclinical Cushing’s syndrome, defined as a syndrome with some hormonal characteristics of Cushing’s syndrome but without typical phenotypic features, signs and symptoms of classical Cushing’s syndrome. There are only a few reports on the short- and long-term changes in BMD at lumbar spine (LS) and proximal femur after the cure of endogenous hypercortisolism [1, 6, 10, 13– 15]. The majority of studies reported an increase in BMD of LS and proximal femur, but a great interindividual variability was also noted. Two recent longitudinal studies reported changes in total body BMD [10, 15]. However, a study reported in 1999 showed that not only the BMD and the bone mineral content (BMC), but the bone area of the entire skeleton was also decreased in patients with endogenous Cushing’s syndrome [16]. In one study a timedependent increase of BMC and BMD was associated with a significant increase in projected bone area of the LS but not of the femoral neck and of the total body in patients
cured from endogenous Cushing’s syndrome [15]. To our knowledge, there are no longitudinal studies on bone densitometric parameters of the peripheral skeleton after the cure of endogenous Cushing’s syndrome.
Patients and methods Patients and endocrine investigations Sixty-eight consecutive patients (51 women, 17 men; age, 15–74 years) referred to the 2nd Department of Medicine, Semmelweis University between 1997 and 2006 for various forms of endogenous hypercortisolism were prospectively enrolled. Of the 68 patients, 41 had CD caused by ACTHproducing pituitary adenomas, 21 had ACS due to uni- or bilateral adrenal tumours, and six had ECS (Table 1). All patients had the characteristic signs and symptoms of active hypercortisolism. Patients with subclinical hypercortisolism associated with incidentally discovered adrenal masses were not included in this study. None of the patients had an active metabolic bone disease other than CS. Initial hormonal investigations included measurements of plasma cortisol concentrations in the morning (between 0800 and 0900 hours), during midnight (between 2300 and 2400 hours) and after an overnight low-dose dexamethasone suppression test (LDDST; 1 mg dexamethasone was given orally at 2400 h, and blood was drawn for plasma
Table 1 Clinical, anthropometrical and hormonal characteristics of patients with endogenous Cushing’s syndrome Cushing’s disease Adrenal Cushing’s syndrome Ectopic Cushing’s syndrome Number of patients Female / Male Number of postmenopausal women Mean age, years (min - max) Estimated duration of disease years; mean ± S.D. Body weight kg; mean ± S.D. Body mass index kg/m2; mean ± S.D. Plasma cortisol at 0800 h μg/dl; mean ± S.D. Plasma cortisol at 2400 hμg/dl; mean ± S.D. Plasma cortisol after LDDST μg/dl; mean ± S.D. Number (%) of patients with IGT/DM Pituitary microadenoma, No. of patients Pituitary macroadenoma, No of patients Adrenal tumor diameter, mm; mean ± S.D. Number of patients with vertebral/ nonvertebral/any fractures Number of fractures: vertebral non-vertebral total Number of females / males with at least 1 DXA follow-up
41 33/8 7 34.9 (15–65) 3.7±2.79 80.1±19.5 30.2±6.25 23.5±11.6 19.2±12.7 18.6±11.4 30 (73.2 %) 30 11 NA 7/11/16 18 15 33 14 / 3
21 16/5 5 46.0 (29–74) 3.8±3.65 81.5±13.9 31.3±4.46 20.8±9.0 15.9±7.3 16.0±7.0 11 (52.4 %) NA NA 46.6±29.6 3/6/8 5 10 15 7/1
LDDST: low dose dexamethasone suppression test; IGT: impaired glucose tolerance; DM: diabetes mellitus NA: not applicable
6 2/4 0 42.2 (31–65) 1.2±1.08 80.7±15.5 27.4±5.14 30.7±7.6 25.1±9.3 33.0±19.8 5 (83.3 %) NA NA NA 0/0/0 0 0 0 0/1
Osteoporos Int (2008) 19:941–949
cortisol measurement the next morning between 0800 and 0900 hours). Basal plasma ACTH concentration was measured between 0800 and 0900 hours. The diagnosis of CD was confirmed by hormonal findings. Magnetic resonance imaging of the pituitary gland was performed in all patients. Sixteen patients with ambiguous pituitary imaging underwent inferior petrosal sinus sampling for exclusion of ECS. Of the 41 patients with ACTH-producing pituitary adenoma, 38 underwent transsphenoidal pituitary tumour resection. Twenty-eight of the 38 operated patients had successful pituitary tumour removal resulting in a cure of hypercortisolism that lasted for at least 12 months after surgery. Four patients had a late relapse of CD between 12 and 66 months after the first surgical intervention. Eleven and two patients had a second and a third surgical intervention, respectively. Three patients underwent pituitary irradiation. Seven of the 31 women were postmenopausal at the time of diagnosis. The diagnosis of ACS was based on suppressed plasma ACTH concentrations and adrenal imaging studies (computed tomography or magnetic resonance imaging). Of the 21 patients with ACS, 18 had unilateral and three had bilateral adrenal tumors. Eighteen of the 21 patients underwent adrenalectomy; histology revealed adrenocortical carcinoma in four patients and a benign adrenocortical lesion in 18 patients. Of the 6 patients with ECS, four had bronchial and two had pancreatic neuroendocrine tumours. One patient with bronchial carcinoid tumour was followed by dual-energy Xray absorptiometry (DXA) after surgical cure. Two patients with pancreatic carcinoids were not included in the followup because of an incurable, hormonally active metastatic disease in one patient and an early death of the other patient. The cure of CS after surgical intervention was confirmed by low plasma cortisol concentrations or by a normal suppression of plasma cortisol after the LDDST in the postoperative period and during follow-up. Postoperative and follow-up endocrine evaluation in patients with CD included also a detailed evaluation for hypopituitarism with measurements of plasma TSH, free T4, LH, FSH, as well as testosterone in men and estradiol in women. In these patients, the GH-IGF-l axis was also assessed by measurements of plasma GH during an insulin tolerance or arginine tests and by determination of plasma IGF-l levels. Thirtytwo patients had postoperative hypoadrenalism treated with oral hydrocortisone substitution with an initial daily dose of 15–40 mg, which was tapered off gradually and stopped between 2 and 4 months postoperatively. Postoperative evaluation indicated hypogonadotropic hypogonadism in 5 patients and GH-deficiency in three patients. Of the five patients with hypogonadotropic hypogonadism, one woman was treated with estradiol.
943
Follow-up bone densitometric measurements were evaluated until the initiation of antiresorptive medication or GH therapy in two patients. Of the 26 cured patients participating in the follow-up bone densitometric arm of this study, 17, eight and one patients previously had CD, CS and ECS, respectively. Of the 17 patients with previous CD (14 women and 3 men; 13 patients with micro- and four patients with macroadenomas), 12 patients showed normal pituitary functions during postoperative evaluation. Postoperative growth hormone deficiency, hypogonadotropic hypogonadism and diabetes insipidus were present in five, two and one patients, respectively. All patients cured from CS and ECS had normal pituitary functions postoperatively. Bone mineral density measurements Bone densitometric measurements at the L1-L4 lumbar spine (LS), left total femur (TF), femoral neck (FN), trochanteric and intertrochanteric femoral subregions, total radius (TR), distal one-third of the radius (R1/3) and the ultradistal (UDR) subregion of the nondominant radius were performed by DXA using a Hologic QDR 4500C instrument (Hologic, Waltham, MA, USA). For analysis, software version 9.03 was used. BMD measurements were expressed as g/cm2. Z-scores were calculated according to the manufacturer’s reference curves (as the number of standard deviations (SD) from age- and sex-matched healthy controls). NHANES III normative data were used as a reference database for femoral bone density measurements [17]. Quality control was maintained by daily scanning of Hologic Anthropometric Spine Phantom. The coefficient of variation of BMD measurements over a period of 4 years was 0.35 %. DXA measurements were performed during the active phase of CS and, thereafter, yearly following curative surgery until the time of relapse of endogenous hypercortisolism or until the start of an antiresorptive treatment or GH substitution. At the time of each bone densitometry, the body weight and height were determined to calculate body mass index (BMI). Patients receiving pituitary irradiation were excluded from follow-up evaluation. During the active phase of the disease, all patients with a body weight less then 125 kg had bone densitometric measurements at the LS, TF, femoral neck, trochanteric and intertrochanteric femoral subregions, whereas 37 of the 41 patients with CD, 20 of the 21 patients with ACS and all of the 6 patients with ECS had measurements also at the forearm. One patient with a body weight above 125 kg had baseline measurement only at the forearm. During followup, the bone densitometric changes were evaluated not only for BMD, but also for BMC and projected bone area (“area”). In addition to follow-up LS and femoral measure-
944
Osteoporos Int (2008) 19:941–949
ments, all patients were regularly followed with radius DXA measurements. After the surgical cure, 26 patients with at least one follow-up DXA measurements had a total of 47 follow-up DXA measurements. The mean follow-up time was 36.8 months (12–96 months). The changes in the area, BMC, BMD and BMD z-score values at the LS, TF, FN, TR, R1/3 and UDR were compared to corresponding baseline values. Assessment of fractures In all patients, a detailed history of peripheral fractures was obtained, and previous medical documentations were also searched for vertebral fractures. Each patient with dorsal or lumbar back pain underwent dorsolumbar X-ray examination. Fractures occurring within 5 years before the diagnosis of CS were considered as possible consequence of Cushing’s syndrome. Statistical analysis Statistical analysis was performed with SPSS 13.0 software package (SPSS Inc., Chicago, IL, USA). Normality of data was confirmed with Shapiro–Wilk’s test. Results are presented as means±S.D. Clinical, anthropometric and hormonal data in subgroups of patients as well as baseline bone densitometric and clinical data in patients with and without fractures were compared using Student’s independentsamples t-test and Mann–Whitney’s U-test according to the data distribution. Categorical variables were compared using Pearson’s chi-square test. Student’s one-sample t-test was used to ascertain whether the changes in bone densitometric parameters at follow-up were significantly different from zero. Standard linear regression and Spearman’s correlation analysis were used to correlate the changes in BMD and BMC values at LS with that measured in other regions during follow-up. Multivariate comparisons were made using multiple linear regressions to adjust for potential confounders such as age, gender and changes in BMI.
Results Clinical and biochemical findings Patients with CD were younger (mean, 34.9 years; range, 15–65 years) than those with ACS (mean, 46.0 years; range, 29–74 years; p=0.002). The female/male ratio was high in CD (33/8) and ACS (16/5) but not in ECS (2/4). The estimated duration of hypercortisolism, body weight, body mass index and the proportion of patients with impaired glucose tolerance or diabetes mellitus were similar in each patient group (Table 1). There were no statistically significant differences between the corresponding plasma cortisol concentrations of the three subgroups (Table 1). Baseline bone densitometry Bone mineral density expressed as mean z-score value was decreased in all three subgroups of endogenous CS at all measured skeletal sites, with the exception of UDR. There were no statistically significant differences in mean z-score values between the three subgroups (Table 2). In each group, the lowest z-score value was registered at the LS. Fractures Twenty-four of the 68 patients (35%) had 23 vertebral and 25 nonvertebral fractures within 5 years before the establishment of the diagnosis of endogenous CS (48 fractures during 340 patient years). There were no statistically significant differences in the frequencies of vertebral and nonvertebral fractures between patients with CD and those with ACS. Patients with ECS had neither peripheral nor vertebral fractures. Correlations between baseline clinical findings and bone densitometry There were no significant correlations between serum cortisol concentrations at 0800 h, 2400 h, and after LDDST
Table 2 Mean baseline bone mineral density z-score values (±SD) at different skeletal regions in patients with endogenous Cushing’s syndrome Number of patients
Cushing’s disease n=41
Cushing’s syndrome n=21
Ectopic ACTH syndrome n=6
All patients n=68
Lumbar spine Total femur femoral neck trochanteric region intertrochanteric region Total radius ultradistal distal 1/3
−1.30±1.14 −0.76±0.99 −0.83±1.05 −0.76±1.05 −0.72±0.91 −1.02±1.02 +0.54±1.08 −0.75±1.02
−1.29±1.00 −0.70±0.97 −0.95±0.80 −0.73±0.77 −0.57±1.05 −1.08±1.40 +0.41±1.10 −0.71±1.75
−0.73±1.51 −0.16±0.70 +0.39±0.76 −0.04±1.00 −0.20±0.68 −0.64±1.46 +0.84±1.00 −0.56±1.67
−1.24±1.13 −0.69±0.96 −0.83±0.95 −0.68±0.97 −0.62±0.94 −1.00±1.18 +0.53±1.07 −0.72±1.33
(40) (38) (38) (38) (38) (35) (35) (35)
(21) (21) (21) (21) (21) (20) (20) (20)
Number of patients with measurements at specified sites are in parentheses
(6) (6) (6) (6) (6) (6) (6) (6)
(67) (65) (65) (65) (65) (61) (61) (61)
Osteoporos Int (2008) 19:941–949
945
and baseline bone densitometric data in any of patient group. Significant positive correlations were found between body mass index and BMD z-scores at LS (r=0.24; p= 0.049), TF (r=0.33; p=0.008), all femoral subregions (neck, r=0.32; p=0.009; trochanter, r=0.254; p=0.041; intertrochanter, r=0.333; p=0.007) and UDR (r=0.38; p= 0.003), but not at other subregions of the radius (data not shown). Association between bone fractures and clinical, biochemical findings and bone densitometry Patients with a history of bone fractures were older than those without a history of bone fractures (44.6±15.0 years vs. 35.2±11.5 years, respectively; p=0.007). There were no significant differences in plasma cortisol concentrations at 0800, 2400 h, and after the LDDST between patients with and without a history of fractures. The duration of estimated glucocorticoid excess was longer in patients with prevalent bone fractures compared to those without fractures (4.91±
4.25 years vs. 2.83±1.86 years, respectively; p=0.034). However there were no correlations between the estimated duration of glucocorticoid excess and BMD z-scores at any site. Patients with prevalent fractures showed significantly lower BMD z-scores at TR (−1.57±0.94 vs. −0.63±1.24; p=0.002), R1/3 (−1.30±1.37 vs. −0.38±1.24; p=0.011) and intertrochanteric subregion (−0.93±0.72 vs. −0.46± 0.97; p=0.045), but not at LS, other femoral regions and UDR compared to patients without a history of bone fractures. Body weight and body mass index had no effect on the occurrence of bone fractures. There was no statistically significant difference in the number of fractures of premenopausal and postmenopausal women. Changes in bone densitometric parameters in patients with surgically cured CS during follow-up As shown in Table 3, progressive increases in BMC and BMD were found during the first four years of postoperative follow-up at the LS and TF, which reached statistical
Table 3 Bone densitometric parameters and body weight of 26 patients with Cushing’s syndrome before and after surgical cure
LS area (cm2) LS BMC (g) LS BMD (g/cm2) LS z-score (S.D.) TF area (cm2) TF BMC (g) TF BMD (g/cm2) TF z-score (S.D.) FN area (cm2) FN BMC (g) FN BMD (g/cm2) FN z-score (S.D.) TR area (cm2) TR BMC (g) TR BMD (g/cm2) TR z-score (S.D.) UDR area (cm2) UDR BMC (g) UDR BMD (g/cm2) UDR z-score (S.D.) R1/3 area (cm2) R1/3 BMC (g) R1/3 BMD (g/cm2) R1/3 z-score (S.D.) Body weight (kg)
Baseline
Change to baseline (mean±S.D.)
(n=26)
First year (n=21)
Second year (n=11)
Third year (n=9)
Fourth year (n=6)
57.03±7.76 50.16±12.63 0.87±0.12 −1.49±1.11 34.94±5.46 29.67±8.64 0.84±0.15 −0.90±1.22 5.16±0.51 3.79±0.72 0.73±0.12 −1.02±1.12 25.17±5.27 13.30±4.78 0.52±0.09 −0.98±1.35 5.91±0.72 2.66±0.66 0.45±0.07 +0.48±1.15 4.95±0.93 3.26±0.99 0.65±0.11 −0.65±1.43 82.96±19.21
+1.25±2.06* +4.20±4.32*** +0.05±0.56*** +0.48±0.50*** +0.59±1.03* +0.78±1.81 +0.01±0.04 +0.16±0.31* +0.05±0.25 +0.04±0.35 +0.00±0.05 +0.11±0.44 −0.47±0.88* −0.77±0.69*** −0.02±0.02*** −0.35±0.36*** −0.20±0.42 * −0.23±0.22*** −0.02±0.02*** −0.43±0.44*** −0.06±0.12* −0.13±0.12 *** −0.02±0.02 ** −0.27±0.38 ** −4.19±3.77 ***
+3.74±1.43*** +10.71±5.30*** +0.11±0.06*** +1.18±0.49*** +0.88±0.93* +2.65±2.07** +0.05±0.03*** +0.42±0.23 ** +0.08±0.40 +0.29±0.27* +0.04±0.05* +0.32±0.32* −0.38±0.77 −0.82±0.85* −0.02±0.02** −0.37±0.37* −0.04±0.38 −0.21±0.19** −0.03±0.03** −0.54±0.49** −0.02±0.16 −0.13±0.16 ** −0.20±0.02 * −0.26±0.37 −4.85±2.82 **
+3.74±1.44*** +9.71±7.35** +0.13±0.10** +1.22±0.93** +0.98±1.52 +2.44±2.45* +0.04±0.05* +0.51±0.30** +0.19±0.35 +0.38±0.26** +0.04±0.04* +0.64±0.31** −0.24±0.66 −0.34±0.49 −0.01±0.02 −0.10±0.29 +0.02±0.46 −0.12±0.23 −0.02±0.03 −0.30±0.53 −0.01±0.17 −0.03±0.10 −0.01±0.02 −0.10±0.41 −3.97±2.70 **
+3.79±2.26** +12.75±6.07** +0.16±0.08** +1.42±0.73** 0.00±1.21 +3.56±2.20* +0.10±0.06* +0.80±0.34** +0.06±0.24 +0.30±0.37 +0.04±0.07 +0.56±0.61 +0.08±0.6 −0.22±1.16 −0.01±0.04 −0.10±0.57 +0.04±0.15 +0.00±0.15 −0.00±0.02 +0.04±0.46 0.00±0.15 −0.04±0.24 −0.01±0.03 −0.07±0.49 −3.20±3.34 *
Data are given as means ± S.D. LS: lumbar spine; TF: total femur, TR: total radius, UDR: ultradistal radius, R1/3: distal one-third of the radius *p
