SKRIFTLIG RAPPORT
Självständigt arbete, (30hp) Läkarprogrammet
Fracture Prediction in Patients with Cushing´s Disease by Estimate of Trabecular Bone Score
Hannes Brolén
Thesis Tutors: Britt Edén Engström, Andreas Kindmark Department of Endocrinology
Date: 2018-01-23
2
Contents
List of abbreviations ... 3
Abstract ... 4
Populärvetenskaplig sammanfattning ... 5
Introduction ... 6
Cushing’s disease ... 6
Glucocorticoid osteoporosis ... 7
Trabecular microarchitecture, GIO and DXA scans ... 10
Trabecular bone score ... 11
Study aim and hypothesis ... 12
Methods ... 13
The Swedish Pituitary Registry ... 13
Parameters ... 14
Statistics and ethical clearance ... 15
Results ... 16
Treatment ... 16
Bone health ... 17
Relevance of bone status ... 18
Bone health over time ... 19
Discussion ... 20
Acknowledgements ... 21
References ... 22
3
List of abbreviations
ACTH=adrenocorticotrope hormone ANOVA=analysis of variance
ap*cc*ll=anteroposterior*
craniocaudal*laterolateral BAE=bilateral adrenalectomy BMD=bone mineral density BMI=body mass index CD=Cushing’s disease CS=Cushing’s syndrome
DXA=dual energy X-ray absorptiometry EGFR=epidermal growth factor receptor FN=femur neck
GC=glucocorticoids
GHD=growth hormone deficiency GHST=growth hormone substitution therapy
GI=gonadal insufficiency
GIO=glucocorticoid induced osteoporosis GST=gonadal substitution therapy
HPT=hyperparathyroidism
HR-pQCT= high-resolution peripheral quantitative computed tomography
HSP90=heat shock protein 90 IGF1=insulin-like growth factor 1 IGFBP-rP1=insulin-like growth factor binding protein-related protein 1
IGFBP-rP2=insulin-like growth factor binding protein-related protein 2
INCA=Information Network for Cancer treatment
KC=ketoconazole
L1-L4=vertebral bodies including L1, L2, L3 and L4
LS=lumbar spine
M-CSF=macrophage colony- stimulating factor µCT=X-ray microtomography MTP=metyrapone
OPG=osteoprotegerin
OT=other osteogenic treatment PO=postmenopausal osteoporosis PTH=parathyroid hormone
RANK-L=receptor activator of nuclear factor kappa-B ligand
RCC=regional cancer center rr=relative risk
RT=radiotherapy S=serum
SD=standard deviation
SPR=Swedish pituitary registry SPSS=statistical package for the social sciences
TBS=trabecular bone score TI=thyroid insufficiency
TPS=transsphenoidal pituitary surgery
TR4=testicular orphan nuclear receptor 4
TS=T-score
TST=thyroid substitution therapy U=urinary
USP8=ubiquitin-specific protease 8 WB=whole body
4
Abstract
Cushing’s disease (CD) is a rare endocrinological disorder in which raised cortisol levels, among other complications, significantly hampers the skeletal integrity.
Although this group of patients has a fracture risk of between 15-40%, the current options for evaluating bone health are limited. Trabecular bone score (TBS) is a relatively new analysis of conventional DXA images; measuring material heterogeneity, it can estimate trabecular microarchitecture and independently predict fractures.
The aim of this study was to evaluate this measurement of bone health in a group of patients with CD. Retrospectively examining patient charts, the correlation between fractures and both TBS and conventional bone mineral density (BMD) measurements were analyzed, as well as general bone health characteristics of CD patients. What was found was that TBS and lumbar spine BMD significantly recovers with time from diagnosis while no significant correlation was found between the fracture risk and neither TBS nor conventional measures of bone density. Reasons why this might be include a small number of patients with available TBS data (n=22) as comparable studies have found TBS to correlate with fracture risk, albeit not in a CD group. Further, preferrably larger or prospective studies might reveal more about the use of this analysis.
5
Populärvetenskaplig sammanfattning
Cushings sjukdom är en ovanlig sjukdom där en liten hypofystumör orsakar att kroppen får väldigt höga nivåer av ett stresshormon, kortisol. Kortisol är vanligtvis ett livsnödvändigt hormon och har många hälsosamma effekter, men om nivåerna är alldeles för höga blir tillståndet skadligt. Många patienter utvecklar fetma, och många delar av kroppen bryts ner. Muskler blir svagare, huden blir skörare och skelettet får lättare frakturer. Dessa frakturer kan orsaka ett stort lidande för patienter som ofta även har andra problem av sin sjukdom.
Att försöka förutsäga om en patient kommer drabbas av en fraktur är ett gammalt problem inom vården. Det går bland annat att ta ut en liten bit av skelettet och studera den i mikroskop eller att göra avancerade röntgenundersökningar, men det vanligaste är att man gör en bentäthetsmätning, eller Dual X-Ray Absorptiometry (DXA) undersökning. Det är en enkel och relativt billig röntgenundersökning som visar hur väl mineraliserat eller förkalkat skelettet är, men dess förmåga att förutsäga frakturer är begränsad. Därför har det på senare tid utvecklats en ny analys av de bilder man får av en DXA-undersökning, Trabecular Bone Score (TBS). Denna värderar skelettets struktur och inte mängden kalcium, något man har sett vanligen försämras hos patienter med Cushing’s sjukdom.
Den här studien har undersökt 48 patienter med Cushings sjukdom som har fått sin diagnos i Uppsala/Örebro regionen i Sverige. Syftet var att jämföra hur vanligt det är med frakturer mellan de patienter som har låga TBS-värden och de som har höga värden, såväl som hur bra de klassiska metoderna för att förutsäga frakturer kan det.
Det undersöktes även hur DXA och TBS-mätningarna förändrades över tid.
Studien visar att skelettstrukturen hos patienter med Cushings sjukdom förbättras med tiden, alltså att den återhämtar sig när patienterna blivit botade. Däremot kunde ingen skillnad ses för frakturrisken mellan patienter med låga och höga TBS-värden, inte heller kunde man visa den roll för risken som den klassiska bentäthetsmätningen har. Detta beror sannolikt på att det fanns för få patienter i studien. I vetenskapliga undersökningar behövs en tillräckligt stor grupp människor för att kunna visa att något beror på en faktisk skillnad och inte bara på slumpen. För att få en bättre förståelse för om TBS faktiskt är ett användbart verktyg behövs fler, och rimligen större studier.
6
Introduction
In his original paper from 1932, Harvey Cushing himself noted that in 4 out of 8 autopsied patients with what would later be recognized as Cushing’s disease (CD), their bones could easily be cut with a knife or a pair of scissors1. Although not usually the primary of associated symptoms of CD, the increased risk of fractures is a heavy burden for people afflicted with the disease. In fact, it’s estimated that 15-40% of patients with the diagnosis suffer fractures related to the disease2,3.
Cushing’s disease
Cushing’s disease is a rare endocrinological disorder, with an incidence rate just under 2 per million, it is primarily diagnosed in females (2.8:1) and although it can afflict all ages the typical age of diagnosis is between 20 and 60 years4,5.
Caused by a pituitary adenoma producing adrenocorticotropic hormone (ACTH), in turn raising cortisol levels from the adrenals, CD causes a hypersecretion with abundant complications. The typical symptoms include a “cushingoid habitus” with fat redistribution, elevated blood pressure, a disturbed metabolic state including but not limited to hyperglycemia and diabetes6,7, generalized muscle loss8, cognitive impairment9, various skin afflictions10 and an increased risk of fractures.
The adenoma is generally of small size and thus rarely affects the optic chiasm located superior to the hypophysis, a typical complication of other pituitary tumors; the tumors are in most cases benign. The etiology of CD is still somewhat unclear but recent research has revealed monoclonal growth as a predominant feature11. Other studies have found that in 35-62% of CD cases a mutated ubiquitin-specific protease 8 (USP8) is involved12,13. This is a deubiquitinating protease, and the specific mutations that have been found increases USP8’s deubiquitinating effects on epidermal growth factor receptor (EGFR), a receptor involved in proliferation and, in the corticotroph cells of the hypophysis, stimulation of ACTH production. Supporting this hypothesis is that EGFR and related ACTH signaling substances is concurrently increased with the USP8 mutation(ibid). Other theories include the mechanisms and overexpression of
7
Testicular orphan nuclear receptor 4 (TR4) and Heat shock protein 90 (HSP90), both involved in the regulation of ACTH secretion14.
Glucocorticoid osteoporosis
The mechanisms contributing to glucocorticoid osteoporosis (GIO) are multiple.
Research in the recent couple of decades have both shown and put an emphasis on the direct glucocorticoid (GC) action on bone cells, particularly the osteoblast, osteocyte and osteoclast15.
Osteoblasts and osteocytes
In brief, osteoblasts, of mesenchymal origin, are responsible for bone apposition following resorption by osteoclasts, and leave further differentiated osteoblasts - osteocytes - to modulate the bone tissue in their trace. While it has been shown that physiological levels of glucocorticoids act on osteoblasts anabolically16, the state of hypercortisolemia has a profoundly hampering, catabolic effect on osteoblasts and osteocytes.
Numerous studies have shown that GCs induce apoptosis in osteoblasts and osteocytes17-19 and that markers of bone formation are reduced17,19-21, along with evidence that osteoblastogenesis is reduced in the presence of GC19. GC administration also lowers the levels of anabolic signal proteins produced by osteoblasts, inhibiting the transcription of IGF-122 and stimulating transcription of IGF- 1 binding proteins IGFBP-rP123 and IGFBP-rP224, further contributing to the inhibition of bone formation.
Osteoclasts
The effect on osteoclasts is double-edged in comparison. Of hematopoietic origin and arising from a fusion of monocytes, osteoclasts have two stages of GC response. In the initial stages of hypercortisolemia, osteoclasts are generally stimulated. Markers of bone resorption are increased17,20,21, a direct anti-apoptotic effect on osteoclasts can be observed25,26 and the numbers of osteoclasts increase, covering a larger part of bone surface27.
8
GC also mediates some of the osteoclastogenic effect via osteoblasts, stimulating said cells’ production of Receptor activator of nuclear factor kappa-B ligand (RANK-L), one of the core stimulants of osteoclast differentiation and activity28. Simultaneously, GC reduces the osteoblastic transcription of osteoprotegerin (OPG)(ibid), a decoy receptor for RANK-L, supplementing GC’s anti-apoptotic effect on osteoclasts. GC also increases the osteoblasts’ production of membrane bound Macrophage-Colony Stimulating Factor (M-CSF)29, further facilitating osteoclast differentiation.
It has also been shown that the pattern of resorption changes following administration of GC. When osteoclasts resorb bone under physiological conditions they reach a certain depth and then migrate to a proximal site and restart the resorption process. In a study by Søe et al.30, it was found that GC-treated osteoclasts, in vitro, differ from this process in certain ways; the resorption pits end up deeper, and the pattern of the pits changes from circular, individually separated, to continuous “trench like” pits. They also found that the difference in depth between the respective groups consists of the demineralized collagen matrix left behind by osteoclasts under physiological GC levels, which is almost entirely disintegrated in GC treated cultures. Another study showed that while bone formation markers correlated negatively, and resorption markers correlated positively with cortisol levels in patients with Cushing’s syndrome (CS), the physiological correlation between the markers themselves was non-existent (r=-0.14), indicating that the feedback between the cells may be compromised during the disease. When later cured of CS, a strong correlation (r=0.80) between the markers was restored in a similar manner to that of healthy controls31.
Cortisol does not simply stimulate osteoclasts, however. While initial resorption increases, the loss of bone mineral density (BMD) in hypercortisolemic patients primarily occurs during the first 6 months and then substantially decreases32. Reasons why this occurs may be related to the findings that although osteoclastic apoptosis is inhibited, production of new osteoclasts decreases19 and numbers of osteoclast progenitors are reduced following GC presentation26. Another theory is that as osteoblast and osteoclast numbers continuously diminish, the physiological as well as the GC-induced production of RANK-L and M-CSF also decline, further reducing the pool of mature osteoclasts to maintain resorption.
9 Parathyroid hormone and glucocorticoids
It was previously believed33,34 that one of the main mechanisms for osteoporosis in GC-treated patients was the resultant decrease in calcium absorption from the intestine35 and a simultaneous increase in renal calcium excretion36. This, establishing a secondary hyperparathyroidism (HPT), would be the cause of increased bone resorption as is normally seen with increased levels of parathyroid hormone (PTH). For various reasons, this theory has been disputed. The correlation itself is not unequivocal; while there are studies showing correlations between PTH and GC treatment, a fair number of studies have also been unable to find such a correlation37. One study of patients with CS found that while PTH levels were significantly higher in the CS group than the control group, there was no correlation between PTH and bone formation or resorption markers21. Another discrepancy between HPT and GIO demineralization is the localization of resorption. While GIO primarily targets cancellous bone(ibid) and thus causes vertebral or femoral fractures38, HPT generally demineralizes cortical bone39 with a relative emphasis on forearm fractures. Perhaps most contradictory is the effective treatment option of teriparatide40, indicating the preventive rather than complicating effect of PTH.
Gonadal hormones and glucocorticoids
Another potential mechanism of GIO is that of the gonadal hormones’ effect on skeletal status. There are numerous studies in support of GC’s inhibition on the levels of and release of sex hormones41-43 and an anabolic effect of such hormones on bone, especially of estrogen, is well established44. However, this theory has been questioned as well, due in part to the different mechanisms of action. Hypogonadism increases osteoblastogenesis and osteoclastogenesis, as confirmed in a study comparing orchidectomized mice with GC treated mice and the combination of the two. The same study found that the addition of orchidectomy to GC administration did not contribute to bone loss, but rather that the GC treatment overrode the expected results from hypogonadism45. These findings, along with the increasing amount of data on GC’s direct actions on bone cells, have contributed to a quite diminished emphasis on this potential mechanism in GIO compared to its apparent relevance 20 years ago.
10
Trabecular microarchitecture, GIO and DXA scans
A central aspect of GIO is its preferential demineralization of the axial skeleton19,21. This is often hypothesized to be caused by the trabecular bone’s considerably larger surface compared to that of cortical bone, facilitating rapid osteoclast access and bone remodeling. This predisposition towards trabecular deterioration has important implications for the skeletal microarchitecture of patients with CD. Trabecular microarchitecture, an individual determinant of bone strength and fracture risk46, is a term describing how well preserved the cancellous bone is by measure of its structure rather than its mineral content. The skeletal plates and rods of trabecular bone, dispersing mechanical energy between themselves when physical stress is applied, can be perforated or even resorbed in the event of excessive resorption in proportion to the bone that is formed by osteoblasts, resulting in a porous and irregular bone structure with less mechanical integrity.
This is especially relevant in light of the current clinical praxis of predicting fracture risk in individual patients. With the development of Dual-energy X-ray Absorptiometry (DXA) scans along with other similar measures of BMD, an opportunity arose to non- invasively and with low radiation exposure analyze the mineral content of bone. This ultimately resulted in the WHO classifications of osteopenia and osteoporosis, at -1 and -2.5 standard deviations (SD) of BMD, also referred to as T-scores of -1 and -2.5, respectively47. While useful, the main concern with using BMD as a predictor of fracture risk is that there is a significant overlap in BMD values between the patients who suffer a fracture and those who do not. One study of postmenopausal women (n=8065) found that while osteoporotic women had a 4.3 times higher relative risk compared to non- osteoporotic women, 54% of women with a hip fracture within 1 year were not osteoporotic as measured by total hip BMD48. To adjust for this, BMD has since been supplemented by FRAX and other clinical tools, accounting for multiple additional risk factors, but a calculation based on microarchitecture is still largely absent in these predictions49.
Directly examining trabecular microarchitecture in a clinical setting is problematic, as this either requires sampling a bone biopsy for histomorphometric analysis, or the use of advanced radiological equipment such as high-resolution peripheral computed tomography (HR-pQCT) or micro-CT (µCT), rarely available for clinical use.
11 Trabecular bone score
To this purpose, Trabecular bone score (TBS) has been developed. TBS is a gray- level textural metric calculated from DXA images of the lumbar spine. The calculation is based on experimental variograms of the two-dimensional images gathered from a DXA examination, which allows the software to quantify the average difference of gray level variations at a set distance in the DXA image. This means that a homogenous image with lesser variations in gray level intensity produces a high TBS value while a heterogenous image with large variations between pixels produces a low TBS value.
In a somewhat poetic analogy by Silva et al.50, a TBS measurement can be compared to an aerial view of a forest. While individual trabeculae cannot be discerned from a DXA scan, like the trees from a flying airplane, areas of disrupted microarchitecture, like clearings in the woods, can readily be. The result is not an actual measurement of the physical microarchitecture, but a proxy developed to correlate with parameters as measured by 3D microarchitectural evaluations.
TBS has in numerous studies been shown to, independently from BMD, predict fractures51. In the largest study to date (n=29,407), it was found that although weakly correlated (r=0.33), BMD and TBS could predict fractures independently of each other, and when used in the same model the combination was significantly better at predicting fractures than either variable alone52.
As mentioned, Cushing’s disease carries with it a high risk of fractures. One study found that in patients with CD, 78% had suffered a vertebral fracture, including asymptomatic, radiologically diagnosed fractures, and 40% had a symptomatic fracture
53. A large study of exogenously GC treated patients (n=244,235) also found that the relative risk is high for vertebral fractures (rr=2.6), while hip fracture and any non- vertebral fracture carry a still significant but relatively modest risk (rr=1.33 and rr=1.61, respectively). Typical of GIO, forearm fractures were the least impacted, although significantly increased, with a relative risk of 1.09. Notably, with reservation for the impact of underlying disease, daily doses as low as 2.5mg of prednisolone presented a significant fracture risk38. Accurately predicting fracture risk would thus enable
12
preventive measures to a greater extent, and TBS shows to be a potential benefactor in this regard.
The predictive capability of BMD is even lower for GIO patients than for patients with postmenopausal osteoporosis (PO), as it has been found that they suffer fractures at higher BMD values54,55. It has also been found that patients with GIO have a higher rate of microarchitectural deterioration compared to PO patients, with lower trabecular thickness, wall thickness and higher trabecular separation56. The same study found, in patients with high doses of GC, reduced trabecular number, increased marrow star volume and interconnectivity index compared to both PO and GIO patients treated with lower GC doses, all indicative of poor trabecular microarchitecture. Other studies of GIO have corroborated these results, having found reduced trabecular thickness and increased trabecular spacing17, and one study, while unable to find a significance of BMD, had a moderately strong, significant correlation between microarchitectural parameters and fracture status57. It should however be mentioned that results vary between studies, and contrasting findings have been presented58. Nonetheless, TBS has shown a strong correlation with microarchitectural parameters such as connectivity density, trabecular number and trabecular separation, with TBS explaining between 67.2% to 79.5% of the mentioned variables’ variance59,60. This suggests that TBS can give an accurate estimation of bone health, aside from the mineral content of bone.
Study aim and hypothesis
The purpose of this study was to examine, in a group of patients diagnosed with Cushing’s disease, the predictive value of TBS in estimating risk of fractures and how this compares with the current measurement of choice, BMD. The hypothesis was that a low TBS as measured at diagnosis would show a significantly higher risk of suffering fractures, and that this would be more accurate than measurement by BMD.
13
Methods
Patient data was gathered from individuals registered in the Swedish Pituitary Registry (SPR), living in the Uppsala-Örebro region. The data was gathered from patient charts and referrals available at Akademiska Sjukhuset, Uppsala, Sweden, as well as directly from SPR.
The Swedish Pituitary Registry
SPR was initiated in 1991 by the Swedish Pituitary Group and is based in the Regional Cancer Center (RCC) of Stockholm/Gotland. This group enlists representatives from endocrinology, neurosurgery, oncology, pathology, ophthalmology and neuroradiology in all 6 health care regions in Sweden. SPR functions as a national quality registry with the purpose to evaluate given therapies and ensure equal treatment to patients with pituitary tumors. Data is submitted to the national IT-platform for cancer registries INCA (Information Network for Cancer treatment), from which data was extracted for this study.
SPR gathers data on patients with pituitary tumors regarding gender, age at diagnosis, MRI/CT findings, visual symptoms, pituitary hormone deficiencies and pituitary hormone overproduction at diagnosis and at subsequent controls. Also included are histopathological findings, methods of treatment (surgery, radiotherapy and pharmacological treatment) and hormone substitution therapy. Since 2012 additional data including therapy related complications, hormone concentrations in patients with hormone producing tumors, sick leave, EQ5D (EuroQol five dimensions questionnaire) and BMI have been gathered in the registry.
As of Dec 31st, 2016, 6375 patients with pituitary tumours were included in the registry of which 364 were diagnosed with Cushing’s disease61. It is estimated that for patients diagnosed with Cushing’s disease in Sweden SPR approximates almost complete registration coverage62
The Uppsala-Örebro region comprises approximately 20% of Sweden’s total population and seven hospitals in the region register patients to SPR61. As Akademiska
14
Sjukhuset in Uppsala is the only hospital in the region providing neurosurgery of pituitary tumours and certain care of patients with CD, most patients in the region see an endocrinologist in Uppsala during the course of their disease. It was thus expected that sufficient individual data should be available from patient files located at Akademiska Sjukhuset in Uppsala.
Data was extracted from SPR and patient charts of both care given in Uppsala as well as any available referred charts from incoming referrals when patients had been diagnosed elsewhere. The studied period lapsed between Jan 1st, 1981 to Dec 31st, 2016.
Parameters
Gender, age at diagnosis and date of diagnosis were directly extracted from SPR. The other parameters were documented from patient charts and referrals.
Fractures and DXA scans
Fractures were recorded from 10 years before the date of diagnosis to Dec 31st, 2016.
Clinically or radiologically diagnosed fractures were included as well as clinically significant fractures reported by the patient, while subclinical fractures reported by the patient but not verified otherwise (i.e. rib fractures) were not included in the analysis.
Bone mineral density (BMD) at the lumbar level (L1-L4 average), total hip and total body BMD, and TBS (L1-L4 average) as calculated by TBS iNsight v. 2.1.2.0 were extracted from GE Lunar iDXA, running EnCore v. 16. TBS categories were defined as Normal microarchitecture, Partially degraded microarchitecture and Degraded microarchitecture (TBS≥1350, 1350>TBS≥1200 and 1200>TBS, respectively).
Measurements were recorded with regard to number of days relative to the patient’s date of diagnosis as well as in their sequential order relative to diagnosis. When a time reference was not specified as a date, the respective analysis was based on the middle point of the reference (e.g. “2000” was analyzed as July 1st, 2000). For analysis of correlations of DXA measurements and fracture risk, fractures from 4 years preceding to 5 years following diagnosis were included in the subsequent analysis. The 4-year cut off was based on the estimation that the time from onset of symptoms to diagnosis averages 3.8 years63 and the 5-year cut off on a plausible time frame of CD-related
15
bone fragility, albeit arbitrary, as no similar studies were found for post-diagnosis fracture risk.
Treatment
Treatment was included in the analysis subdivided by type, including transsphenoidal pituitary surgery, bilateral adrenalectomy and any form of pituitary radiation therapy noted by date of treatment. Pharmacological treatment was further subdivided into ketoconazole or metyrapone and documented as ever present or not. Biochemical control at the latest available point of time was noted as present or not and consequently used to determine what treatment led to cure in each case.
Other radiological and biochemical markers
Morning cortisol levels, ACTH and 24-hour collection of urinary cortisol was each registered from time of diagnosis and divided by the upper reference intervals at their respective local laboratory. Tumor volume as estimated by MRI or CT was recorded.
Given dimensions were calculated for volume by the formula anteroposterior*craniocaudal*laterolateral divided by 2. If one dimension was missing, the missing value was assumed to be the average of those reported. If only one dimension was available, the missing values were assumed to be equal to the reported one. If no tumor was radiologically visible but a pituitary adenoma was verified otherwise, the tumor volume was documented as 1 mm
³ .
Deficiencies in gonadal hormones, thyroid hormones, growth hormone, and any substitution therapies for named deficiencies were noted as either present or not present. Treatment with calcium, Vitamin D and other osteogenic treatment including bisphosphonates, estrogen, teriparatide, calcitonin and denosumab was documented as ever present or not.Statistics and ethical clearance
Statistical calculations were performed using SPSS version 24. The statistical methods that were used were linear regression analysis and multiple regression analysis.
Ethical clearance from an ethics committee was not requested for this work.
Approval for a student thesis was granted by Akademiska Sjukhuset.
16
Results
71 patients were registered between 1991 and 2016 in SPR. Exclusion criteria were lack of sufficient data, defined as no available DXA-scan (n=22) and misdiagnosis (n=1) with a total of 48 patients included in the analysis.
37 patients were female and 11 were male with a sex ratio of 3.4:1. The mean age at diagnosis was 48. The biochemical marker most markedly increased at diagnosis was 24-hour U-cortisol (Table 1).
Treatment
Transsphenoidal pituitary surgery ultimately cured 61% of the patients, with radiotherapy and bilateral adrenalectomy curing 11% and 15%, respectively. 9% of patients were not cured at the time of death and 4% of patients were not cured at the end of 2016. The mean time from diagnosis to the treatment that would ultimately lead to cure was 21 months (Table 2).
Table 2. Number of patients receiving a treatment, time from diagnosis to the treatment that would ultimately lead to cure and the number of times a treatment ultimately cured a patient. TPS=transsphenoidal pituitary surgery, BAE=bilateral adrenalectomy, KC=Ketoconazole, MTP=Metyrapone, RT=radiotherapy, N/a=not cured. Months are mean, with ±1SD.
17
73%
23%
4%
4. BMD category, whole body
46%
46%
8%
3. BMD category, femur
18%
55%
27%
1. TBS category
34%
50%
16%
2. BMD category, lumbar spine
Noted pituitary insufficiencies relevant to bone health shows that 20 of 48 patients had at least one form of pituitary insufficiency, including lack of gonadal, thyroid and growth hormone, with thyroid and gonadal insufficiency being the most common (33% and 25%, respectively). A majority of patients received some form of fracture risk reducing drug. Besides calcium and vitamin D, alendronic acid was the most common therapy with 15 patients (31%) having ever used, and 5 patients (10%) were noted as having ever used an osteogenic drug besides the three previously mentioned.
Bone health
31% of CD patients suffered a fracture during the period of analysis, measuring 4 years preceding to 5 years following diagnosis. Patients generally had poor bone health at diagnosis, especially regarding TBS, lumbar spine and femur measurements (figures 1-4 and table 4).
Figures 1-4. Percent of patients in each bone health category at diagnosis. Figure 1: Green, yellow and orange colors represent normal microarchitecture (TBS ≥1.350), partially degraded microarchitecture (TBS <1.350 but
≥1.200) and degraded microarchitecture (TBS <1.200), respectively. Figures 2-4: Green, yellow and orange colors represent normal bone density (T-Score >-1.0), osteopenia (T-Score ≤-1.0 but >-2.5) and osteoporosis (T-Score
≤-2.5), respectively.
18
0 1 2 3 4 5 6 7 8
-9 -8 -7 -6 -5 -4 -3 -2 -1 1 2 3 4 5 6 7 … 17
Fracture incidents
Years relative to diagnosis
Fracture incidents (i.e. an incident resulting in one or more fractures) of the entire study period (10 years previous to a patient’s date of diagnosis to Dec 31st, 2016) were centered within 2-6 years previous to diagnosis and 5 years following diagnosis (figure 5).
Relevance of bone status
This study could not find a significant correlation between TBS and the risk of suffering a fracture during the period of study. A linear regression of both TBS and TBS-status by category at diagnosis had positive, not statistically significant correlations with risk of suffering a fracture during the period of 4 years preceding and 5 years following diagnosis. Lumbar spine T-score, femur neck T-score and whole body T-score, all at diagnosis, showed negative, not statistically significant correlations with fracture risk during the studied period. A multiple regression analysis of TBS combined with either lumbar spine T-score or femur neck T-score and fractures as the dependent variable showed higher R² but no statistical significance by ANOVA (table 5).
Fractures
(n=48)
T-Score, lumbar spine
(n=38)
T-Score, femur neck
(n=37)
T-Score, Whole body
(n=26)
TBS
(n=22)
Male (n=11) 7 -1.0, ±1.67 -1.3, ±0.98 -.7, ±0.60 1.196, ±0.141
Female (n=37) 8 -1.4, ±1.10 -1.0, ±0.95 -.2, ±1.22 1.250, ±0.087
Total (n=48) 15 -1.3, ±1.24 -1.0, ±0.95 -.3, ±1.11 1.241, ±0.097
19
Table 5. Correlations between DXA-measurements at diagnosis with fracture risk. TBS=trabecular bone score, LS=lumbar spine, FN=femur neck, WB=whole body, TS=T-Score
Bone health over time
TBS-measurements from all available DXA examinations in the study (n=44) shows a significant, positive correlation between TBS and number of days from diagnosis (figure 6). A linear regression analysis results in an R² of .128 and a β-coefficient of .358 (p=0.017). Noteworthy for this analysis is that one outlier was omitted.
This patient’s TBS-data was from 13, 15, 18 and 22 years after diagnosis and a successful transsphenoidal hypophysectomy had been performed 6 months after diagnosis with no recurring CD. It was thus estimated that this patient, with pathological TBS-measurements and generally normal cortisol levels, skewed the data in a manner that was not related to his earlier CD diagnosis.
A similar linear regression of all lumbar spine BMD measurements (n=84) also results in a statistically significant positive correlation but with a roughly halved r², .066, and a lower β-coefficient, .257 (p=0.018). Similar femur neck and whole body BMD linear regressions were not significant (p>0.5, r²<.01).
TBS TBS by category
T-Score, LS
T-Score, FN
T-Score, WB
TBS &
TS, LS
TBS &
TS, FN
R² 0.062 .151 0.048 0.090 0.031 0.070 0.161
Β-coefficient 0.250 .389 -0.219 -0.300 -0.176 Odds ratio 1.28 1.48 0.80 0.74 0.84
p-value 0.25 0.07 0.18 0.07 0.37 0.50 0.21
20
Discussion
This study has attempted to bring further light on a method of measurement that is still gaining traction. The concept of looking at bone microarchitecture is a relatively new practice and field of research, and TBS holds some promise as a cheap, non-invasive method of approximating the microarchitecture of the lumbar spine.
Interestingly a significant correlation between both TBS and lumbar spine BMD with time from diagnosis was found, with a stronger effect on TBS. This indicates that the trabecular microarchitecture recovers over time, with normalized cortisol levels the physiological response to mechanical load would supposedly resume normal function and the bone cells regain their activity. This would however require further studies, both to reproduce this result and to determine the specific mechanisms underlying the recovery.
Notably, lumbar spine BMD improved significantly (p=0.018) while femur neck and whole body BMD showed no signs of recovery (p>0.5, r²<0.01). A possible explanation of this can be found in the drawbacks of the retrospective approach. The examinations from later points in time would predominantly be made on those who had clinical osteoporosis. This translates as a low BMD rather than a low TBS, which is rarely used as a clinical measure or indication of further treatment and evaluation. Thus, those with a low BMD would likely more often go through numerous DXA-examinations than those with a low TBS as this has a lesser impact on repeated examinations, and this might mask a general BMD recovery that could possibly be found in a study covering a larger segment of the patient population.
This study was not able to verify a correlation between TBS at time of diagnosis and the fracture risk a patient with Cushing’s disease is exposed to. In fact, the trend was the opposite, contradicting existing research - with a positive correlation between TBS and fracture risk. This correlation was however not statistically significant and especially in light of previous research, it seems likely that it was simply a misrepresenting result due to lack of power. The BMD correlations were however negative, while still not significant, and it’s possible that this result which is more in line with previous research was related to group size.
21
While all patients in the study had at some point gone through a DXA, not all of them had done so at the time of diagnosis, from which the correlation analyses were calculated. Even further, TBS data at diagnosis could only be obtained from 22 patients, while T-Scores were available from 38, 37 and 26 patients (lumbar spine, femur and whole body, respectively). As none of these results were significant however, not much can be concluded from the analysis.
Other limitations of this study concern the methodology. Numerous patients only received care in Uppsala for a few years and the time frame for reporting fractures, other complications and therapies could be quite short. The registered amount of fractures in this study was still in line with previous research, but a prospective or a retrospective study with direct inquiries from patients could bring better documentation of those with little locally available data, who might be inadequately represented in a study such as this.
Going into this study with an estimated 71 patients to evaluate would also have been served a benefit by checking the available data early on. With 22 TBS measures at diagnosis, it is a reasonable assumption that sufficient power is not present. By expanding the study to adjacent regions or other methodological approaches a greater chance at finding significant results might have been achieved.
Conclusion
In conclusion, this study sheds some light on the progression of bone health over time but was insufficient to provide evidence whether TBS is a useful tool in fracture risk evaluations regarding patients newly diagnosed with Cushing’s disease. With fracture risk difficult to determine but often critical to prevent, TBS is still a relatively new tool and more research can hopefully gather further knowledge in this field.
Acknowledgements
I would like to thank Britt Edén Engström and Andreas Kindmark for their continuous support, help with study design and aid with any issues along the way. Without them this thesis would not have seen the light of day.
22
References
1. Cushing H, Press JH. The Basophil Adenomas of the Pituitary Body and Their Clinical Manifestations (pituitary Basophilism) [Internet]. 1932. Available from:
https://books.google.com/books/about/The_Basophil_Adenomas_of_the_Pituitary_B.html?hl=&id=04X xPgAACAAJ
2. van Varsseveld NC, van Bunderen CC, Franken AAM, Koppeschaar HPF, van der Lely AJ, Drent ML.
Fractures in pituitary adenoma patients from the Dutch National Registry of Growth Hormone Treatment in Adults. Pituitary [Internet] 2016;19(4):381–90. Available from:
http://dx.doi.org/10.1007/s11102-016-0716-3
3. Belaya ZE, Hans D, Rozhinskaya LY, et al. The risk factors for fractures and trabecular bone-score value in patients with endogenous Cushing’s syndrome. Arch Osteoporos [Internet] 2015;10(1):44.
Available from: http://dx.doi.org/10.1007/s11657-015-0244-1
4. Tjornstrand A, Gunnarsson K, Evert M, et al. The incidence rate of pituitary adenomas in western Sweden for the period 2001-2011. Eur J Endocrinol [Internet] 2014;171(4):519–26. Available from:
http://dx.doi.org/10.1530/eje-14-0144
5. Raappana A, Koivukangas J, Ebeling T, Pirilä T. Incidence of pituitary adenomas in Northern Finland in 1992-2007. J Clin Endocrinol Metab [Internet] 2010;95(9):4268–75. Available from:
http://dx.doi.org/10.1210/jc.2010-0537
6. Di Dalmazi G, Quinkler M, Deutschbein T, et al. Cortisol-related metabolic alterations assessed by mass spectrometry assay in patients with Cushing’s syndrome. Eur J Endocrinol [Internet] 2017;EJE – 17–0109. Available from: http://dx.doi.org/10.1530/eje-17-0109
7. Feelders RA, Pulgar SJ, Kempel A, Pereira AM. MANAGEMENT OF ENDOCRINE DISEASE: The burden of Cushing’s disease: clinical and health-related quality of life aspects. Eur J Endocrinol [Internet] 2012;167(3):311–26. Available from: http://dx.doi.org/10.1530/eje-11-1095
8. Schakman O, Kalista S, Barbé C, Loumaye A, Thissen JP. Glucocorticoid-induced skeletal muscle atrophy. Int J Biochem Cell Biol [Internet] 2013;45(10):2163–72. Available from:
http://dx.doi.org/10.1016/j.biocel.2013.05.036
9. Sonino N, Fava GA. Psychiatric Disorders Associated with Cushing??s Syndrome. CNS Drugs [Internet] 2001;15(5):361–73. Available from: http://dx.doi.org/10.2165/00023210-200115050-00003
10. Stratakis CA, Mastorakos G, Mitsiades NS, Mitsiades CS, Chrousos GP. Skin manifestations of Cushing disease in children and adolescents before and after the resolution of hypercortisolemia.
Pediatr Dermatol [Internet] 1998;15(4):253–8. Available from:
https://www.ncbi.nlm.nih.gov/pubmed/9720685
11. Biller BM, Alexander JM, Zervas NT, Hedley-Whyte ET, Arnold A, Klibanski A. Clonal origins of adrenocorticotropin-secreting pituitary tissue in Cushing’s disease. J Clin Endocrinol Metab [Internet]
1992;75(5):1303–9. Available from: http://dx.doi.org/10.1210/jcem.75.5.1358909
12. Reincke M, Sbiera S, Hayakawa A, et al. Mutations in the deubiquitinase gene USP8 cause Cushing’s disease. Nat Genet [Internet] 2014;47(1):31–8. Available from: http://dx.doi.org/10.1038/ng.3166
23
13. Ma Z-Y, Song Z-J, Chen J-H, et al. Recurrent gain-of-function USP8 mutations in Cushing’s disease.
Cell Res [Internet] 2015;25(3):306–17. Available from: http://dx.doi.org/10.1038/cr.2015.20
14. Daniel E, Newell-Price J. Recent advances in understanding Cushing disease: resistance to glucocorticoid negative feedback and somatic USP8 mutations. F1000Res [Internet] 2017;6:613.
Available from: http://dx.doi.org/10.12688/f1000research.10968.1
15. Komori T. Glucocorticoid Signaling and Bone Biology. Horm Metab Res [Internet] 2016;48(11):755–63.
Available from: http://dx.doi.org/10.1055/s-0042-110571
16. Sher LB, Woitge HW, Adams DJ, et al. Transgenic expression of 11beta-hydroxysteroid
dehydrogenase type 2 in osteoblasts reveals an anabolic role for endogenous glucocorticoids in bone.
Endocrinology [Internet] 2004;145(2):922–9. Available from: http://dx.doi.org/10.1210/en.2003-0655
17. Lane NE, Yao W, Balooch M, et al. Glucocorticoid-Treated Mice Have Localized Changes in
Trabecular Bone Material Properties and Osteocyte Lacunar Size That Are Not Observed in Placebo- Treated or Estrogen-Deficient Mice. J Bone Miner Res [Internet] 2005;21(3):466–76. Available from:
http://dx.doi.org/10.1359/jbmr.051103
18. O’Brien CA, Jia D, Plotkin LI, et al. Glucocorticoids act directly on osteoblasts and osteocytes to induce their apoptosis and reduce bone formation and strength. Endocrinology [Internet]
2004;145(4):1835–41. Available from: http://dx.doi.org/10.1210/en.2003-0990
19. Weinstein RS, Jilka RL, Parfitt AM, Manolagas SC. Inhibition of osteoblastogenesis and promotion of apoptosis of osteoblasts and osteocytes by glucocorticoids. Potential mechanisms of their deleterious effects on bone. J Clin Invest [Internet] 1998;102(2):274–82. Available from:
http://dx.doi.org/10.1172/jci2799
20. Di Somma C, Pivonello R, Loche S, et al. Severe impairment of bone mass and turnover in Cushing’s disease: comparison between childhood-onset and adulthood-onset disease. Clin Endocrinol [Internet]
2002;56(2):153–8. Available from: http://dx.doi.org/10.1046/j.0300-0664.2001.01454.doc.x
21. Chiodini I. Alterations of Bone Turnover and Bone Mass at Different Skeletal Sites due to Pure Glucocorticoid Excess: Study in Eumenorrheic Patients with Cushing’s Syndrome. J Clin Endocrinol Metab [Internet] 1998;83(6):1863–7. Available from: http://dx.doi.org/10.1210/jc.83.6.1863
22. Delany AM, Canalis E. Transcriptional repression of insulin-like growth factor I by glucocorticoids in rat bone cells. Endocrinology [Internet] 1995;136(11):4776–81. Available from:
http://dx.doi.org/10.1210/endo.136.11.7588206
23. Pereira RC, Blanquaert F, Canalis E. Cortisol enhances the expression of mac25/insulin-like growth factor-binding protein-related protein-1 in cultured osteoblasts. Endocrinology [Internet]
1999;140(1):228–32. Available from: http://dx.doi.org/10.1210/endo.140.1.6411
24. Pereira RC, Durant D, Canalis E. Transcriptional regulation of connective tissue growth factor by cortisol in osteoblasts. Am J Physiol Endocrinol Metab [Internet] 2000;279(3):E570–6. Available from:
https://www.ncbi.nlm.nih.gov/pubmed/10950824
25. Jia D, O’Brien CA, Stewart SA, Manolagas SC, Weinstein RS. Glucocorticoids act directly on osteoclasts to increase their life span and reduce bone density. Endocrinology [Internet]
2006;147(12):5592–9. Available from: http://dx.doi.org/10.1210/en.2006-0459
24
26. Weinstein RS, Chen J-R, Powers CC, et al. Promotion of osteoclast survival and antagonism of bisphosphonate-induced osteoclast apoptosis by glucocorticoids. J Clin Invest [Internet]
2002;109(8):1041–8. Available from: http://dx.doi.org/10.1172/JCI14538
27. Sato AY, Cregor M, Delgado-Calle J, et al. Protection From Glucocorticoid-Induced Osteoporosis by Anti-Catabolic Signaling in the Absence of Sost/Sclerostin. J Bone Miner Res [Internet]
2016;31(10):1791–802. Available from: http://dx.doi.org/10.1002/jbmr.2869
28. Hofbauer LC, Gori F, Riggs BL, et al. Stimulation of osteoprotegerin ligand and inhibition of
osteoprotegerin production by glucocorticoids in human osteoblastic lineage cells: potential paracrine mechanisms of glucocorticoid-induced osteoporosis. Endocrinology [Internet] 1999;140(10):4382–9.
Available from: http://dx.doi.org/10.1210/endo.140.10.7034
29. Rubin J, Biskobing DM, Jadhav L, et al. Dexamethasone promotes expression of membrane-bound macrophage colony-stimulating factor in murine osteoblast-like cells. Endocrinology [Internet]
1998;139(3):1006–12. Available from: http://dx.doi.org/10.1210/endo.139.3.5778
30. Søe K, Delaissé J-M. Glucocorticoids maintain human osteoclasts in the active mode of their resorption cycle. J Bone Miner Res [Internet] 2010;25(10):2184–92. Available from:
http://dx.doi.org/10.1002/jbmr.113
31. Szappanos⁎ A, Toke J, Lippai D, et al. Bone turnover in patients with endogenous Cushing’s syndrome before and after successful treatment. Bone [Internet] 2009;44:S352. Available from:
http://dx.doi.org/10.1016/j.bone.2009.03.173
32. LoCascio V, Bonucci E, Imbimbo B, et al. Bone loss in response to long-term glucocorticoid therapy.
Bone Miner [Internet] 1990;8(1):39–51. Available from: http://dx.doi.org/10.1016/0169-6009(91)90139- q
33. Hodgson SF. Corticosteroid-induced osteoporosis. Endocrinol Metab Clin North Am [Internet]
1990;19(1):95–111. Available from: https://www.ncbi.nlm.nih.gov/pubmed/2192870
34. Adachi JD, Ioannidis G. Calcium and vitamin D therapy in corticosteroid-induced bone loss: what is the evidence? Calcif Tissue Int [Internet] 1999;65(4):332–6. Available from:
https://www.ncbi.nlm.nih.gov/pubmed/10485987
35. Huybers S, Naber THJ, Bindels RJM, Hoenderop JGJ. Prednisolone-induced Ca2+ malabsorption is caused by diminished expression of the epithelial Ca2+ channel TRPV6. Am J Physiol Gastrointest Liver Physiol [Internet] 2007;292(1):G92–7. Available from: http://dx.doi.org/10.1152/ajpgi.00317.2006
36. Suzuki Y, Ichikawa Y, Saito E, Homma M. Importance of increased urinary calcium excretion in the development of secondary hyperparathyroidism of patients under glucocorticoid therapy. Metabolism [Internet] 1983;32(2):151–6. Available from: https://www.ncbi.nlm.nih.gov/pubmed/6298567
37. Rubin MR, Bilezikian JP. Clinical review 151: The role of parathyroid hormone in the pathogenesis of glucocorticoid-induced osteoporosis: a re-examination of the evidence. J Clin Endocrinol Metab [Internet] 2002;87(9):4033–41. Available from: http://dx.doi.org/10.1210/jc.2002-012101
38. Van Staa TP, Leufkens HGM, Abenhaim L, Zhang B, Cooper C. Use of Oral Corticosteroids and Risk of Fractures. J Bone Miner Res [Internet] 2000;15(6):993–1000. Available from:
http://dx.doi.org/10.1359/jbmr.2000.15.6.993
25
39. Moosgaard B, Christensen SE, Vestergaard P, Heickendorff L, Christiansen P, Mosekilde L. Vitamin D metabolites and skeletal consequences in primary hyperparathyroidism. Clin Endocrinol [Internet]
2008;68(5):707–15. Available from: http://dx.doi.org/10.1111/j.1365-2265.2007.03109.x
40. Teriparatide or Alendronate in Glucocorticoid-Induced Osteoporosis. N Engl J Med [Internet]
2008;358(12):1302–4. Available from: http://dx.doi.org/10.1056/nejmc073408
41. Sakakura M, Takebe K, Nakagawa S. Inhibition of Luteinizing Hormone Secretion Induced by Synthetic LRH by Long-term Treatment with Glucocorticoids in Human Subjects. J Clin Endocrinol Metab [Internet] 1975;40(5):774–9. Available from: http://dx.doi.org/10.1210/jcem-40-5-774
42. MacAdams MR, White RH, Chipps BE. Reduction of serum testosterone levels during chronic glucocorticoid therapy. Ann Intern Med [Internet] 1986;104(5):648–51. Available from:
https://www.ncbi.nlm.nih.gov/pubmed/3083749
43. Crilly R, Cawood M, Marshall DH, Nordin BE. Hormonal status in normal, osteoporotic and
corticosteroid-treated postmenopausal women. J R Soc Med [Internet] 1978;71(10):733–6. Available from: https://www.ncbi.nlm.nih.gov/pubmed/712726
44. Compston JE. Sex steroids and bone. Physiol Rev [Internet] 2001;81(1):419–47. Available from:
https://www.ncbi.nlm.nih.gov/pubmed/11152762
45. Weinstein RS, Jia D, Powers CC, et al. The skeletal effects of glucocorticoid excess override those of orchidectomy in mice. Endocrinology [Internet] 2004;145(4):1980–7. Available from:
http://dx.doi.org/10.1210/en.2003-1133
46. Bouxsein ML, Seeman E. Quantifying the material and structural determinants of bone strength. Best Pract Res Clin Rheumatol [Internet] 2009;23(6):741–53. Available from:
http://dx.doi.org/10.1016/j.berh.2009.09.008
47. Bonjour P, Clark P, Cooper C. WHO Scientific Group on the Assesment of Osteoportosis at Primary Health Care Level [Internet]. 2004 [cited 2017 Jul 19];Available from:
http://www.who.int/chp/topics/Osteoporosis.pdf?ua=1
48. Wainwright SA, Marshall LM, Ensrud KE, et al. Hip fracture in women without osteoporosis. J Clin Endocrinol Metab [Internet] 2005;90(5):2787–93. Available from: http://dx.doi.org/10.1210/jc.2004- 1568
49. El-Hajj Fuleihan G, Chakhtoura M, Cauley JA, Chamoun N. Worldwide Fracture Prediction. J Clin Densitom [Internet] 2017;20(3):397-424. Available from: http://dx.doi.org/10.1016/j.jocd.2017.06.008
50. Silva BC, Leslie WD, Resch H, et al. Trabecular Bone Score: A Noninvasive Analytical Method Based Upon the DXA Image. J Bone Miner Res [Internet] 2014;29(3):518–30. Available from:
http://dx.doi.org/10.1002/jbmr.2176
51. Harvey NC, Glüer CC, Binkley N, et al. Trabecular bone score (TBS) as a new complementary approach for osteoporosis evaluation in clinical practice. Bone [Internet] 2015;78:216–24. Available from: http://dx.doi.org/10.1016/j.bone.2015.05.016
52. Hans D, Goertzen AL, Krieg M-A, Leslie WD. Bone microarchitecture assessed by TBS predicts osteoporotic fractures independent of bone density: the Manitoba study. J Bone Miner Res [Internet]
2011;26(11):2762–9. Available from: http://dx.doi.org/10.1002/jbmr.499
26
53. Tauchmanovà L, Pivonello R, Di Somma C, et al. Bone demineralization and vertebral fractures in endogenous cortisol excess: role of disease etiology and gonadal status. J Clin Endocrinol Metab [Internet] 2006;91(5):1779–84. Available from: http://dx.doi.org/10.1210/jc.2005-0582
54. Luengo M, Picado C, Del Rio L, Guañabens N, Montserrat JM, Setoain J. Vertebral fractures in steroid dependent asthma and involutional osteoporosis: a comparative study. Thorax [Internet]
1991;46(11):803–6. Available from: https://www.ncbi.nlm.nih.gov/pubmed/1771602
55. Van Staa TP, Laan RF, Barton IP, Cohen S, Reid DM, Cooper C. Bone density threshold and other predictors of vertebral fracture in patients receiving oral glucocorticoid therapy. Arthritis Rheum [Internet] 2003;48(11):3224–9. Available from: http://dx.doi.org/10.1002/art.11283
56. Dalle Carbonare L, Arlot ME, Chavassieux PM, Roux JP, Portero NR, Meunier PJ. Comparison of trabecular bone microarchitecture and remodeling in glucocorticoid-induced and postmenopausal osteoporosis. J Bone Miner Res [Internet] 2001;16(1):97–103. Available from:
http://dx.doi.org/10.1359/jbmr.2001.16.1.97
57. Graeff C, Marin F, Petto H, et al. High resolution quantitative computed tomography-based assessment of trabecular microstructure and strength estimates by finite-element analysis of the spine, but not DXA, reflects vertebral fracture status in men with glucocorticoid-induced osteoporosis.
Bone [Internet] 2013;52(2):568–77. Available from: http://dx.doi.org/10.1016/j.bone.2012.10.036
58. Aaron JE, Francis RM, Peacock M, Makins NB. Contrasting microanatomy of idiopathic and
corticosteroid-induced osteoporosis. Clin Orthop Relat Res [Internet] 1989;(243):294–305. Available from: https://www.ncbi.nlm.nih.gov/pubmed/2721071
59. Muschitz C, Kocijan R, Haschka J, et al. TBS reflects trabecular microarchitecture in premenopausal women and men with idiopathic osteoporosis and low-traumatic fractures. Bone [Internet]
2015;79:259–66. Available from: http://dx.doi.org/10.1016/j.bone.2015.06.007
60. Silva BC, Walker MD, Abraham A, et al. Trabecular bone score is associated with volumetric bone density and microarchitecture as assessed by central QCT and HRpQCT in Chinese American and white women. J Clin Densitom [Internet] 2013;16(4):554–61. Available from:
http://dx.doi.org/10.1016/j.jocd.2013.07.001
61. Ekman B, Dahlgren C, Forsgren M, Engström BE. Årsrapport Svenska Hypofysregistret 2016 [Internet]. [cited 2018 Jan 11];Available from:
https://www.cancercentrum.se/globalassets/cancerdiagnoser/hypofystumorer/hypofys_arsrapport_201 6.pdf
62. Svenska Hypofysregistret [Internet]. [cited 2017 Jul 18];Available from:
http://www.socialstyrelsen.se/register/registerservice/nationellakvalitetsregister/svenskahypofysregistr et
63. Kreitschmann-Andermahr I, Psaras T, Tsiogka M, et al. From first symptoms to final diagnosis of Cushing’s disease: experiences of 176 patients. Eur J Endocrinol [Internet] 2015;172(6):285-289.
Available from: http://dx.doi.3org/10.1530/eje-14-0766e
