Low bone mass is a well-recognized complication of

Transcription

1 GASTROENTEROLOGY 2010;138: Low Bone Mass and Severity of Cholestasis Affect Fracture Risk in Patients With Primary Biliary Cirrhosis NÚRIA GUAÑABENS,*, DACIA CERDÁ,* ANA MONEGAL,* FRANCESCA PONS,* LLORENÇ CABALLERÍA,, PILAR PERIS,*, and ALBERT PARÉS, *Metabolic Bone Diseases Unit, Service of Rheumatology, Hospital Clínic, IDIBAPS, University of Barcelona, Barcelona; CIBERehd, Barcelona; and Liver Unit, Hospital Clínic, IDIBAPS, University of Barcelona, Barcelona, Spain See related article, Kuiper EMM et al, on page 530 in CGH. BACKGROUND & AIMS: The influence of osteoporosis and liver disease on fracture risk is not well characterized in patients with primary biliary cirrhosis (PBC). We studied a large series of women with PBC to assess the prevalence and risk factors for fractures and the fracture threshold. METH- ODS: In female patients with PBC (n 185; age, years; range years), age, duration of PBC, menopausal status, and histologic stage and severity of liver disease were assessed. Vertebral and non-vertebral fractures were recorded in 170 and 172 patients, respectively. Osteoporosis and osteopenia were diagnosed based on densitometry analysis. RESULTS: The prevalences of vertebral, non-vertebral, and overall fractures were 11.2%, 12.2%, 20.8%, respectively. Osteoporosis was significantly more frequent in patients with PBC than in normal women. Osteoporosis was associated with age, weight, height, histologic stage, severity, and duration of liver damage; fractures were associated with osteoporosis, menopause, age, and height but not with severity of PBC. Osteoporosis was a risk factor for vertebral fracture (odds ratio [OR], 8.48; 95% confidence interval [CI]: ). Lumbar T score 1.5 (OR, 8.27; 95% CI: ) and femoral neck T score 1.5 (OR, 6.83; 95% CI: ) were significant risk factors for vertebral fractures. CONCLUSIONS: Fractures, particularly vertebral fractures, are associated with osteoporosis, osteopenia, and T scores less than 1.5, whereas osteoporosis and osteopenia are associated with the severity of liver damage. Patients with T scores less than 1.5 might require additional monitoring and be considered for therapy to prevent fractures. Keywords: Metabolic Bone Disease; Bone Mineral Density; Osteoporosis; Osteopenia. Low bone mass is a well-recognized complication of primary biliary cirrhosis (PBC). 1 3 However, for years, it has been discussed whether PBC itself represents a further risk for osteoporosis or whether its prevalence is that which should be expected in a population of middleaged postmenopausal women. 4 In this respect, we have recently reported that osteoporosis is more prevalent in women with PBC than in the age-matched general population. 5 However, what is even more important is to know the characteristics of fragility fractures, the hallmark of osteoporosis, and particularly the underlying mechanisms of such a complication. There are few studies assessing fracture risk in women with PBC, mostly limited to short series and with debatable results. A questionnaire of fracture experience in PBC patients showed that fracture risk was not increased when compared with general population. 6 By contrast, Solayman- Dodaran et al found a modest increase in both the absolute and relative fracture risks in a series of 930 people with PBC, from the General Practice Research Database in the United Kingdom. 7 This later study, however, was not addressed to analyze the risk factors for fractures. Indeed, it has been described that age and severity of liver disease are the main risk factors for osteoporosis, defined by bone mineral density (BMD) measurements. 6 Although low BMD is an important predictor of fracture risk when large cohort studies have been assessed, 8 there are other independent risk factors for fractures, such as age, use of glucocorticoids, falls, or previous fracture. 9,10 In fact, half of postmenopausal women with incident fractures have BMD values above the diagnostic threshold of osteoporosis, defined as a T score of 2.5 standard deviation or more below the average value of young healthy women. 11 When non-vertebral fracture risk was assessed in postmenopausal women, independent risk factors such as age, height, vitamin D, or prevalent non-vertebral fractures, in addition to low BMD, were found. 12 Likewise, the gradient of risk for fracture, provided by BMD measurements in PBC patients, is uncertain. The aims of this study were to analyze the prevalence of fractures in a large series of PBC patients Abbreviations used in this paper: PBC, primary biliary cirrhosis; BMD, bone mineral density by the AGA Institute /$36.00 doi: /j.gastro

2 June 2010 BONE FRACTURES IN PRIMARY BILIARY CIRRHOSIS 2349 and their association with low bone mass by BMD measurements and with the severity and duration of liver damage. Patients and Methods Patients The study was carried out in 185 consecutive female patients aged (range, years) diagnosed with PBC using clinical, biochemical, immunologic, and histologic criteria. All patients received daily oral calcium supplements (1 g) and low doses of 25- hydroxyvitamin D (20 g). One hundred thirty-one patients were evaluated before treatment with ursodeoxycholic acid (15 mg/kg/day). Throughout the study period, 61 other patients were excluded because of the use of steroids or other drugs affecting bone metabolism (17 patients), other previous severe disease apart from PBC (13 patients), not consenting to participate in the study (11 patients), and unknown reasons (20 patients). The Ethical Committee at Hospital Clínic approved the study protocol, and informed consent was obtained from all subjects. Study Procedures Demographic and clinical characteristics were assessed in all patients. Age, menopausal status, body mass index (BMI) as well as duration, severity, and histologic stage of the liver disease were recorded. The severity of cholestasis and liver damage was evaluated by the serum concentrations of bilirubin, aspartate aminotransferase, -glutamyltransferase, and albumin and total alkaline phosphatase activity, as well as the prothrombin index. The histologic stage of PBC was evaluated according to Ludwig et al s criteria. 13 Bone disease evaluation was performed as follows: BMD at the lumbar spine (180 patients) and femoral neck (140 patients), spinal lateral x-rays (170 patients), and self-reported non-vertebral fractures. Measurements of BMD were performed by dual-energy x-ray absorptiometry (Lunar, Madison, WI), except in those patients included before January 1991, in whom BMD measurements were performed on Lunar machines by dual photon absorptiometry. In these latter patients, the BMD measurements were converted to dual energy x-ray data using an established formula. In patients with compression fractures, lumbar BMD measurements were determined on the intact vertebrae. The coefficients of variation in lumbar spine for healthy volunteers and for patients with osteoporosis were 0.8% and 1.3%, respectively. The coefficient of variation in the femoral neck was 2.3%. Osteopenia and osteoporosis were diagnosed using the World Health Organization thresholds: T score between 1.0 and 2.5 and 2.5, respectively. 14 Prevalence of osteoporosis in women with PBC was compared with that reported in Spanish women of matched age, classified according to the same densitometric criteria. 15 Standard lateral radiographs of the thoracic and lumbar spine were obtained to disclose vertebral fractures. Vertebral fracture was defined as a reduction of 20% or more in the anterior, middle, or posterior height of the vertebral body. All self-reported non-vertebral fractures were recorded, but fractures resulting from high-energy trauma were excluded from the analysis. Patients gathered during the 21-year duration of the study ( ) were divided into 3 assessment periods (n 62), (n 62), and (n 61) to determine changes on the prevalence of fractures, bone densitometry, demographics, and severity of liver disease among these periods. The number of patients excluded in the 3 periods was 21, 23, and 17, respectively. Statistical Analysis Data are expressed as mean standard error of the mean. The 2 test was used to analyze differences in noncontinuous variables, and the Student t test or 1-way analysis of variance was used to analyze differences in continuous variables. Continuous variables were transformed in dichotomous categories according to the median or the pathologic level (bilirubin higher than 1.2 mg/dl, and albumin lower than 38 g/l). The independent effect of significant variables associated with osteoporosis or fracture was assessed using stepwise logistic regression. The criteria for entering and removing variables were.05 and.1, respectively. A 2-tailed P value.05 was considered to indicate a significant difference. Results A total of 185 women diagnosed with PBC were included in the analysis. Demographic and clinical characteristics are shown in Table 1. The prevalence of vertebral, non-vertebral, and overall fractures was 11.2%, 12.2%, and 20.8%, respectively. Five patients (3%) had both non-vertebral and vertebral fractures. Non-vertebral fractures included mostly wrist fractures (8 patients) followed by lower limb fractures (4 patients). Only 1 patient had a hip fracture. The prevalence of osteoporosis (T score 2.5) at the lumbar spine was 30.6%, compared with 11.2% in the age-matched Spanish population (P.001). At the femoral neck, the prevalence of osteoporosis was 12.9% in PBC, compared with 4.3% in the general population (P.001). Overall, 37% of patients with PBC had osteoporosis, taking into account the lowest BMD value at the lumbar spine or femoral neck. The clinical, laboratory, and densitometric data of PBC patients with and without osteoporosis are summarized in Table 2. Osteoporosis was associated with age, menopause, weight, height, histologic stage, and severity and duration of liver disease. Thus, patients with osteoporo-

3 2350 GUAÑABENS ET AL GASTROENTEROLOGY Vol. 138, No. 7 Table 1. Demographic, Analytical, Histologic, and Bone Mineral Density Data of Patients With Primary Biliary Cirrhosis Normal range N 185 Age (y) Menopause (%) 151 (81.6) Height (cm) Weight (kg) BMI (kg/m 2 ) Duration PBC (y) Bilirubin (mg/dl) AP (u/l) ggt (u/l) ALT (u/l) Albumin (g/l) Prothrombin index (%) Histologic stage (%) I 47.1 II 29.7 III 15.6 IV 7.6 L2 L4 (g/cm 2 ) T score Z score Femoral BMD Neck (g/cm 2 ) T score Z score ALT, alanine aminotransferase; ggt, AP, alkaline phosphatase; -glutamyltransferase; BMI, body mass index. sis at the lumbar spine were older than patients without osteoporosis (P.001). Patients were mostly postmenopausal women (95.6% vs 79.3%, respectively, P.05), and the duration of PBC was significantly longer (P.01). In addition, PBC women with osteoporosis had higher levels of bilirubin (P.05) and lower values of both albumin (P.001) and prothrombin index (P.001). Histologic stage of liver disease was more advanced in patients with osteoporosis than in those without (P.001). Age above 56 years, duration of the disease longer than 4 years, and advanced histologic stage were the independent variables associated with overall osteoporosis and lumbar and femoral osteoporosis as well. The demographic, laboratory, and densitometric data in patients with and without fractures are summarized in Table 3. Age, height, menopausal status, and osteoporosis by BMD but not the severity of liver disease were associated with fractures. PBC patients with fractures were older (P.000) and shorter (P.05) than those without fractures. In addition, patients with fractures were mostly postmenopausal women (94% vs 77%, respectively, P.02) and had lower BMD at the lumbar spine (P.001) and femoral neck (P.001). However, fractures were not associated with duration or severity of liver disease. The severity of low bone mass as defined by a lumbar T score below 1.5 or the femoral neck T score Table 2. Clinical Features, Densitometric Data, and Fractures in Patients With PBC According to Osteoporosis Osteoporosis at lumbar spine Osteoporosis at femoral neck Yes No P value Yes No P value No Age (y) Menopause (%) NS Height (cm) NS Weight (kg) NS BMI (kg/m 2 ) NS Duration PBC (y) Bilirubin (mg/dl) a a AP (u/l) NS ggt (u/l) NS NS ALT (u/l) NS NS Albumin (g/l) Prothrombin index (%) NS Histological stage Stages I and II (%) 25 (52.1) 103 (85.8) (61.1) 97 (82.9).03 L2 L4 (g/cm 2 ) T score Z score Femoral BMD Neck (g/cm 2 ) T score Z score Overall fractures (%) Vertebral (%) Non-vertebral (%) NS ALT, alanine aminotransferase; AP, alkaline phosphatase; ggt, -glutamyltransferase; BMI, body mass index; NS, nonsignificant. a Log transformation.

4 Table 3. Clinical Features and Laboratory Histologic and Densitometric Data in Patients With PBC According to the Presence of Fractures Overall fracture Vertebral fracture Nonvertebral fracture Yes No P value Yes No P value Yes No P value N Age, y Menopause (%) 34 (94.4) 106 (76.8) NS Height (cm) NS Weight (kg) NS NS NS BMI (kg/m 2 ) NS NS NS Duration PBC (y) NS NS NS Bilirubin (mg/dl) NS NS NS AP (u/l) NS NS NS ggt (u/l) NS NS ALT (u/l) NS NS NS Albumin (g/l) NS NS NS Prothrombin index (%) NS NS NS Histologic stage (%) I II III IV NS NS L2 L4 (g/cm 2 ) T score Z score % Osteoporosis NS % T score NS Femoral BMD Neck (g/cm 2 ) T score Z score NS NS % Osteoporosis % T score ALT, alanine aminotransferase; AP, alkaline phosphatase; ggt, -glutamyltransferase; BMI, body mass index; NS, nonsignificant. June 2010 BONE FRACTURES IN PRIMARY BILIARY CIRRHOSIS 2351

5 2352 GUAÑABENS ET AL GASTROENTEROLOGY Vol. 138, No. 7 Table 4. Clinical, Analytical, Histologic, and Densitometric Data and Prevalence of Fractures in Patients With PBC According to the Period of Assessment F value P value No Age (y) NS Menopause (%) NS Height (cm) NS Weight (kg) BMI (kg/m 2 ) Duration PBC (y) Bilirubin (mg/dl) AP (u/l) ggt (u/l) ALT (u/l) Albumin (g/l) Prothrombin index (%) NS Histologic stage (%) 2 I II III IV L2 L4 (g/cm 2 ) NS T score NS Z score NS % Osteoporosis NS Femoral BMD Neck (g/cm 2 ) T score Z score NS % Osteoporosis NS Vertebral fracture (%) NS Non-vertebral fracture (%) NS ALT, alanine aminotransferase; AP, alkaline phosphatase; ggt, -glutamyltransferase; BMI, body mass index; NS, nonsignificant. below 1.5 were the independent variables associated with fractures. Moreover, age, height, duration, severity of cholestasis, and histologic stage were significantly different in patients with a lumbar T score below 1.5 than in those with a lumbar T score above this level when analyzing patients without fractures (data not shown). Similar results were observed when analyzing only the patients who were not formerly treated with ursodeoxycholic acid. There were differences in the patients among the 3 study periods (Table 4) because severity of liver disease was significantly milder over the study period, whereas there was a trend to lower severity of bone disease observed at the lumbar spine but significantly at the femoral level. Consequently, the prevalence of lumbar osteoporosis decreased from 37.1% in the first study period to 20.7% in the latest period. Femoral osteoporosis also decreased, being 17.9% in the first period and 12.1% in the latter period. No significant changes in age and menopausal status were observed among the 3 periods, but there was a trend to older age in the third study period. Previous treatment with ursodeoxycholic acid had no significant effect on the prevalence of osteoporosis and fractures nor on the severity of liver disease among the 3 study periods. Osteoporosis (T score 2.5) was associated with both overall and vertebral fractures as was low bone mass by means of a T score lower than 1.5. Furthermore, 17 out of 19 patients with vertebral fractures had low bone mass with a T score 1.5 at the lumbar spine, and 13 patients had a T score 1.5 at the femoral neck. Thus, a lumbar or a femoral neck T score below 1.5 identified the patients with high risk of fractures, particularly vertebral fractures (Figures 1 and 2) (Table 5). Non-vertebral fractures were associated with lower BMD and age. Femoral neck osteoporosis (T score 2.5) was the only risk factor for non-vertebral fractures besides age higher than 56 years. Discussion This study shows that osteoporosis in PBC, defined by bone mineral density measurements, is especially related to severity and duration of liver disease and that fractures are strongly associated with low bone mass. Furthermore, the close association between vertebral fractures and a T score lower than 1.5 at the lumbar spine provides a cut-off value for BMD measurements that indicates a higher risk for vertebral fracture in PBC patients.

6 June 2010 BONE FRACTURES IN PRIMARY BILIARY CIRRHOSIS 2353 Figure 1. Lumbar T score of patients with and without vertebral and non-vertebral fractures. A T score 1.5 (dashed line) is associated with high risk for vertebral fracture. In fact, when overall risk factors for fracture are taken into account, our findings are not unexpected. Advanced age and osteoporosis are known major risk factors for fracture. 9,16,17 In addition, risk factors such as previous fracture or other covariates with age, such as likelihood of falling significantly affects fracture risk. 14 An interesting contribution of our study, however, is that it defines a cut-off value that identifies PBC women at higher risk Figure 2. Femoral neck T score of patients with and without vertebral and non-vertebral fractures. A T score 1.5 (dashed line) is associated with high risk for vertebral fracture.

7 2354 GUAÑABENS ET AL GASTROENTEROLOGY Vol. 138, No. 7 Table 5. Risk for Fractures According to the BMD Measurements at Lumbar Spine and Femoral Neck Vertebral fracture Non-vertebral fracture Overall fractures Yes No OR 95% CI P value Yes No OR 95% CI P value Yes No OR 95% CI P value No T score (57.9%) 39 (26.4%) (42.1%) 42 (28.0%) NS 15 (44.1%) 35 (25.7%) T score (89.5%) 75 (50.7%) (63.2%) 80 (53.2%) NS 25 (73.5%) 67 (49.3%) Femoral neck BMD No T score (46.7%) 11 (9.2%) (38.9%) 11 (9.3%) (36.7%) 7 (6.6%) T score (86.7%) 58 (48.7%) (72.2%) 58 (49.2%) (76.7%) 48 (45.2%) No Osteoporosis lumbar 15 (78.9%) 46 (30.7%) (52.4%) 50 (33.1%) (62.8%) 39 (28.5%) or femoral CI, confidence interval; OR, odds ratio; NS, nonsignificant. for fractures and particularly for vertebral fracture. This approach has not been previously tested in a large series of PBC patients, although we and others have suggested this BMD threshold for considering intervention. 1,3 Indeed, fragility fractures may occur in postmenopausal women without osteoporosis, 18 and, as suggested in this study, liver disease contributes to low bone mass, and through this mechanism to an increased risk of fracture. Up to 78% of patients with fractures in our series had osteoporosis, but we would like to stress that PBC patients without this diagnostic category are also at risk. Vertebral fractures have been scarcely studied in PBC patients because most studies have focused on bone loss, 4,19 21 and reports providing prevalence of these fractures are not large enough to evaluate this end point. Despite this fact, it has been occasionally described that prevalence of vertebral fractures ranges from 5.7% to 13% in middle-aged women with PBC 5,21 and about 13% in patients with end-stage cholestatic liver disease. 22 Studies assessing total fractures may reflect variability in the case series, particularly in the milder presentation of the disease. Accordingly, our study indicates that the prevalence of vertebral fractures decreases in parallel with the lower severity of cholestasis, which probably results in better bone mass as observed among the 3 periods of study (Table 4). Furthermore, the assessment of vertebral fracture in a number of studies is not well specified, the criteria for fracture differs among the studies, and many of these kinds of fractures do not come to clinical attention. 16 When fracture risk has been assessed in the records of a large PBC cohort, vertebral fractures have not been considered in the analysis because of some ascertainment bias due to the variability in the symptoms. 7 In addition, vertebral/coccyx fractures were only self-reported in 2 out of 85 PBC patients (4%) included in a questionnaire survey of fracture experience. 6 By contrast, our study provides a figure based on x-rays of the spine in a large series of consecutive PBC patients, and, at present, this is the best method for assessing the presence of vertebral fractures, with no satisfactory alternative currently available. 9 The pattern of non-vertebral fractures in PBC patients is in accordance with what was expected. Thus, the mean age of our patients was 55.7 years, ranging from 28 to 79 years. It is known that hip fractures occur mainly among the elderly population, and the mean age of hip fracture is around 80 for women. 16 This fact may explain that hip fractures were practically absent in our series. By contrast, the incidence of wrist fractures seems to start to increase at younger ages, between 40 and 65, and thereafter there is a plateau in most populations. 17 Accordingly, in our series, wrist fractures were the most frequent non-vertebral fractures, followed by lower limb fractures. Solayman-Dodaran et al found in their cohort of 930 people with PBC 25 incident hip fractures and 44 ulna/ radius (wrist) fractures. The mean age of their cohort at

8 June 2010 BONE FRACTURES IN PRIMARY BILIARY CIRRHOSIS 2355 start of follow-up was 60.1 years, with 23% of patients being older than 70 years. 7 Thus, the difference in the rates of hip fractures may be a result of the advanced age of their patients. Reinforcing this hypothesis, it is interesting to consider that fractures of the distal forearm and proximal humerus account for more fractures than the hip up until the age of 70 years in the general population. 23 When assessing the mechanisms for fractures, we found that they are related to low bone mass. Older age, menopause, lower weight and height, as well as advanced histologic stage and severity and duration of liver damage are risk factors for such a condition in this and other studies. 5,21 Indeed, bone disease in PBC is strongly related to liver disease. Accordingly, Menon et al stated that the severity of liver disease contributes significantly to the severity of bone disease in PBC. 21 Although the pathogenesis of hepatic osteodystrophy is still unclear, it is established that it mainly results from decreased bone formation. 2,24 27 In this context, the role of retained substances of cholestasis, including bilirubin and biliary salts, seems reasonable, although some reports do not support this hypothesis. 28 Both bilirubin and biliary acids have been involved in the low bone formation associated with cholestasis, through an alteration in the proliferation and survival of osteoblasts In addition, other alterations such as increased production of a fibronectin isoform containing the oncofetal domain during liver disease may participate in the decrease of bone formation. 32 Other conditions in patients with PBC, including increased osteoclast formation, low vitamin D levels, calcium malabsorption, and nutritional status 2,33,34 may be contributing factors to the full picture of bone disease in PBC. Treatment with ursodeoxycholic acid before bone mass and fracture assessment had no significant consequences on the results because the prevalence of osteoporosis and fractures was alike when analyzing the overall series and the group of patients who did not receive this treatment. Previous ursodeoxycholic acid therapy had no influence on the 3 study periods either, although there were more patients under ursodeoxycholic in the last period. The potential long-term favorable effects of ursodeoxycholic acid therapy on bone mass resulting from the improvement in liver disease cannot be gathered from the current report. This information can only be verified in long-term sequential studies assessing the effect of ursodeoxycholic acid on bone mass evolution and the incidence of both osteoporosis and fractures. Finally, one of the most relevant findings of this study is the identification of a threshold in BMD measurements that captures most vertebral fractures. This finding has a practical utility when assessing bone disease in patients with PBC. Thus, a value in the T score below 1.5 in the lumbar spine, assessed by densitometry, in addition to other risk factors for fracture such as older age or postmenopausal status, will help to make the decision to consider therapy. Data from this study cannot be automatically extrapolated to other liver conditions and to all patients with advanced liver disease, particularly in those prior to organ transplantation because the study was carried out only in women with cholestatic disease. However, this is the unique assessment showing the close association between a T score below 1.5 and fracture risk, thus suggesting that patients with such BMD values should be considered for treatments addressed to prevent further bone loss and fractures. In conclusion, our results provide the prevalence of vertebral and peripheral fractures in a large series of middle-aged PBC women and demonstrate that the risk for fracture is associated with low bone mass and indirectly, through this condition, with the severity and duration of liver disease. The risk of fracture was inversely correlated with the T score, and, therefore, the patients with T scores below 1.5 may deserve more monitoring and be considered as candidates for therapy to prevent fractures. References 1. Parés A, Guañabens N. Treatment of bone disorders in liver disease. J Hepatol 2006;45: Guañabens N, Parés A, Mariñoso L, et al. Factors influencing the development of metabolic bone disease in primary biliary cirrhosis. Am J Gastroenterol 1990;85: Zein CO, Jorgensen RA, Clarke B, et al. Alendronate improves bone mineral density in primary biliary cirrhosis: a randomized placebo-contolled trial. Hepatology 2005;42: Newton J, Francis R, Prince M, et al. Osteoporosis in primary biliary cirrhosis revisited. Gut 2001;49: Guañabens N, Parés A, Ros I, et al. Severity of cholestasis and advanced histological stage but not menopausal status are the major risk factors for osteoporosis in primary biliary cirrhosis. J Hepatol 2005;42: Boulton-Jones JR, Fenn RMF, West J, et al. Fracture risk of women with primary biliary cirrhosis: no increase compared with general population controls. Aliment Pharmacol Ther 2004;20: Solayman-Dodaran M, Card TR, Aithal GP, et al. Fracture risk in people with primary biliary cirrhosis: a population-based cohort study. Gastroenterology 2006;131: Johnell O, Kanis JA, Oden A, et al. Predictive value of BMD for hip and other fractures. J Bone Miner Res 2005;20: Brown JP, Fortier M. Canadian Consensus Conference on Osteoporosis, 2006 Update. JOGC 2006;172:S Cheung AM, Detsky AS. Osteoporosis and fractures. Missing the bridge? JAMA 2008;299: Garnero P, Delmas PD. Contribution of bone mineral density and bone turnover markers to the estimation of risk of osteoporotic fracture in postmenopausal women. Musculoskel Neuron Interact 2004;4: Roux C, Briot K, Horlait S, et al. Assessment of non-vertebral fracture risk in postmenopausal women. Ann Rheum Dis 2007; 66: Ludwig J, Dickson ER, McDonald SA. Staging of chronic nonsuppurative destructive cholangitis (syndrome of primary biliary cirrhosis). Virchows Arch A Path Anat Histol 1978;379: World Health Organization Study Group: assessment of fracture risk and its application to screening for postmenopausal osteoporosis. WHO technical report series 843. Geneva: WHO, 1994: 2 25.

9 2356 GUAÑABENS ET AL GASTROENTEROLOGY Vol. 138, No Diaz Curiel M, Garcia JJ, Carrasco JL, et al. Prevalencia de osteoporosis determinada por densitometría en la población española. Med Clin (Barc) 2001;116: Jonhell O, Kanis J. Epidemiology of osteoporotic fractures. Osteoporos Int 2005;16(Suppl 2):S3 S Harvey N, Dennison E, Cooper C. Epidemiology of osteoporotic fractures. In: Rosen CJ, ed. Primer on the metabolic bone diseases and disorders of mineral metabolism. 7th ed. Washington, DC: American Society for Bone and Mineral Research, 2006: Siris E, Miller PD, Barrett-Connor E, et al. Identification and fracture outcomes of undiagnosed low bone mineral density in postmenopausal women. Results from the National Osteoporosis Risk Assessment. JAMA 2001;286: Pereira SP, Bray GP, Pitt PI, et al. Non-invasive assessment of bone density in primary biliary cirrhosis. Eur J Gastroenterol Hepatol 1999;11: Bagur A, Mautalen C, Findor J, et al. Risk factors for the development of vertebral and total skeleton osteoporosis in patients with primary biliary cirrhosis. Calcif Tissue Int 1998;63: Menon KVN, Angulo P, Weston S, et al. Bone disease in primary biliary cirrhosis. J Hepatol 2001;35: Guichelar MMJ, Malinchoc M, Sibonga J, et al. Bone metabolism in advanced cholestatic liver disease: analysis by bone histomorphometry. Hepatology 2002;36: Delmas PD, Marin F, Marcus R, et al. Beyond hip: importance of other nonspinal fractures. Am J Med 2007;120: Hodgson SF, Dickson ER, Wahner HW, et al. Bone loss and reduced osteoblast function in primary biliary cirrhosis. Ann Intern Med 1985;103: Hodgson SF, Dickson ER, Eastell R, et al. Rates of cancellous bone remodeling and turnover in osteopenia associated with primary biliary cirrhosis. Bone 1993;14: Guichelaar MMJ, Malinchoc M, Sibonga JD, et al. Bone histomorphometric changes after liver transplantation for chronic cholestatic liver disease. J Bone Miner Res 2003;18: Ackerman Z, Weinreb M, Amir G, et al. Bone mineral metabolism and histomorphometry in rats with cholestatic liver disease. Liver 2002;22: Smith DLH, Shire NJ, Watts NB, et al. Hyperbilirubinemia is not a major contributing factor to altered bone mineral density in patients with chronic liver disease. J Clin Densitom 2006;9: Janes CH, Dickson ER, Okazaki R, et al. Role of hyperbilirubinemia in the impairment of osteoblast proliferation associated with cholestatic jaundice. J Clin Invest 1995;95: Ruiz-Gaspá S, Enjuanes A, Guañabens N, et al. Lithocholic acid down-regulates vitamin D effects in human osteoblasts. Eur J Clin Invest 2010;40: Ruiz-Gaspà S, Martinez-Ferrer A, Enjuanes A, et al. High bilirubin reduces cell survival and differentiation of primary human osteoblasts. Is this effect applicable to the serum of chronic cholestatic patients? J Bone Miner Res 24 (Suppl 1). Available at: aid a928a897-3e9d-4dea-a c532b7802. Accessed September 17, Kawelke N, Bentmann A, Hackl N, et al. Isoform of fibronectin mediates bone loss in patients with primary biliary cirrhosis by suppressing bone formation. J Bone Miner Res 2008;23: Olivier BJ, Schoenmaker T, Mebius RE, et al. Increased osteoclast formation and activity by peripheral blood mononuclear cells in chronic liver disease patients with osteopenia. Hepatology 2008;47: Collier JD, Ninkovic M, Compston JE. Guidelines on the management of osteoporosis associated with chronic liver disease. Gut 2002;50(Suppl 1):1 19. Received November 16, Accepted February 9, Reprint requests Address requests for reprints to: Núria Guañabens, Service of Rheumatology, Hospital Clínic, Villarroel 170, Barcelona, Spain. nguanabens@ub.edu. Conflicts of interest The authors disclose no conflicts. Funding Supported, in part, by Centro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBERehd), Instituto de Salud Carlos III, and FIS-08/0105, Ministerio de Ciencia e Innovación, Spain.