The term hepatic osteodystrophy encompasses bone disease associated with chronic liver disease and includes osteoporosis and more rarely osteomalacia. Osteoporosis is clinically important as, unlike osteomalacia, it is frequently identified in patients with cirrhosis, leading to spinal fractures that often go undetected clinically but nonetheless lead to significant patient morbidity. The detection of osteoporosis in these patients therefore requires a high index of clinical suspicion for diagnosis as about a third of spinal fractures are asymptomatic and so will be identified only radiologically.1 In contrast, femoral neck fractures are uncommon in patients with cirrhosis as they occur about a decade later than spinal fractures, beyond the life expectancy of most patients with chronic liver disease.
Osteomalacia rarely occurs in adult patients with chronic liver disease despite a low serum vitamin D level being reported in up to two-thirds of patients with cirrhosis. In contrast, osteoporosis, which increases the risk of vertebral fractures, occurs in 12%-55% of patients with cirrhosis. Although the prevalence is probably falling, as shown by a fall from 57%-26% in patients with biliary disease requiring liver transplantation over the last 2 decades, it still accounts for significant patient morbidity. Bone density also falls in the first 3 months after liver transplantation, and pretransplant fractures are predictive of posttransplant fractures. Many of the known risk factors for postmenopausal osteoporosis exist in the cirrhotic population, such as excess alcohol intake, steroid use, poor nutrition, and hypogonadism. There is also an increased risk of osteoporosis in patients without cirrhosis, particularly those with hemochromatosis and biliary disease. The diagnosis is made with bone density measurements. The effective treatment is largely based on evidence from postmenopausal osteoporosis as there have been only a few small clinical trials of patients with chronic liver disease. Bisphosphonates are the mainstay of treatment; they have been shown to be effective in biliary disease and are well tolerated. (HEPATOLOGY 2007.)
Osteomalacia, although historically reported in up to 64% of individuals with primary biliary cirrhosis (PBC) in the 1970s, is now rarely seen in adult patients with chronic liver disease.2, 3 The change in prevalence probably reflects differences in the diagnostic criteria, selection bias, and possibly improved nutrition. Bone biopsy is the gold standard for identifying osteomalacia as not all patients with low serum vitamin D levels will have osteomalacia. In a recent study from the Mayo Clinic of 33 patients with end-stage PBC and primary sclerosing cholangitis (PSC) undergoing liver transplantation, no cases of osteomalacia were identified on bone biopsy.4 Given that the liver is involved in bile salt production, the absorption of vitamin D, and the subsequent 25-hydroxylation of vitamin D, it might be expected that osteomalacia would be common in patients with chronic liver disease. In fact, the intestinal absorption of cholecalciferol and 25-hydroxycholecalciferol is affected only in the presence of severe cholestasis, by which patients are jaundiced. Subsequent hepatic 25-hydroxylation of vitamin D3 has not been studied in humans, but in cirrhotic rats, this process is not impaired (reviewed in ref. 2).
About two-thirds of patients with cirrhosis and 96% of patients awaiting liver transplantation do have low vitamin D levels in the absence of osteomalacia,5 and a low vitamin D level is also associated with reduced bone mineral density (BMD), high bone turnover, and increased risk of hip fracture in the elderly without liver disease.
The risk of osteoporosis increases with age.1 The peak bone mass is reached in the third decade; after the age of 40, it declines in both sexes but more rapidly in women, for whom the decline accelerates after menopause. The risk of fracture is determined not only by the bone density but also by the trabecular architecture, geometry, bone turnover, and propensity for falls. Despite this, BMD as measured by dual-energy x-ray absorptiometry is currently the best predictor of fracture risk.
The risk of fracture increases 2-fold for every standard deviation below the mean BMD of normal controls.1, 2 Osteoporosis as defined by a BMD measured with dual-energy x-ray absorptiometry of less than 2.5 standard deviations below the normal peak bone mass (T score of −2.5). Although the risk of fractures rises steeply with falling BMD, other clinical risk factors for fracture have been identified that are independent of BMD1–3 (Table 1). The most important of these is a previous fracture. Once a vertebral fracture has occurred, the risk of a further vertebral fracture increases 10-fold, and the risk of a subsequent hip fracture increases 2.3-fold.1
|Previous fragility fracture|
|Oral glucocorticoid therapy (>5 mg for 3 months)|
|Body mass index (<19 kg/m2)|
|Alcohol intake (>3 units/day)|
|Maternal history of hip fracture|
Known risk factors for osteoporosis are frequently found in patients with chronic liver disease (Table 1). These include poor nutrition, excess alcohol intake, hypogonadism, and corticosteroid use. In addition, most patients with PBC are postmenopausal women. Whether or not cholestasis and cirrhosis are independent risk factors for chronic liver disease is currently unclear as studies to date have not been sufficiently powered to address this important question.
Excess alcohol intake is an independent risk factor for osteoporosis and is associated with a 2.8-fold increase in hip fractures. In a study of 76 men drinking more than 216 g/day for more than 24 years, only 22% had abnormal liver histology, and 30% had vertebral compression fractures, although only 4% were symptomatic. Cumulative alcohol intake in men has also been inversely related to BMD.6 Reduced serum testosterone levels occurring in both actively drinking alcoholics and patients with cirrhosis probably also contribute to osteoporosis. Low vitamin D levels have also been reported in a third of alcoholics with a low BMD, and vitamin D supplementation has been shown in one study to improve BMD at the wrist in some patients.7
Longstanding hypogonadism in men is associated with reduced bone remodeling and decreased bone formation, which is reversible with adequate hormonal replacement. Hypogonadism as measured by free testosterone levels has been reported to occur in up to 75% of patients with cirrhosis needing transplantation but is unrelated to the etiology of the cirrhosis.4, 8 An assessment of hypogonadism in the presence of cirrhosis can be difficult as total serum testosterone levels may overestimate free testosterone because of an increase in testosterone binding globulin levels. If the total testosterone rather than the free testosterone is measured in patients with cirrhosis, it should be expressed with respect to the testosterone sex hormone binding globulin.
Pathophysiology of Osteoporosis
Biological mechanisms underlying osteoporosis in chronic liver disease are complex and are poorly understood. Candidate genes affecting bone mass include the vitamin D receptor, collagen 1 alpha 1, low-density lipoprotein receptor binding protein 5, and estrogen receptor, but polymorphisms in these genes have not yet been linked to an increased risk of fracture in patients without liver disease.1 Studies in PBC indicate that polymorphisms in the vitamin D receptor and collagen 1 alpha 1 gene do not appear to contribute to the risk of osteoporosis.9
Subsequent bone loss after the peak bone mass is reached occurs because of the uncoupling of bone resorption and bone formation. Increased osteoclast activity is mediated through osteoclastogenic proinflammatory cytokines such as interleukin 1 and tumor necrosis factor, both of which have been implicated in hepatic inflammation and fibrosis. More recently, the receptor activator of NF kappa beta (RANK) and the receptor activator of NF kappa beta ligand (RANKL), in addition to osteoprotegerin, have been shown to be involved in osteoclastic bone resorption.10 RANK is found on osteoblasts, and through its interaction with RANKL on osteoclasts, there is an increase in mature osteoclast survival and the differentiation of immature osteoclasts. Osteoprotegerin secreted by osteoblasts into the bone microenvironment blocks the interaction between RANK and RANKL, regulating bone turnover.1 The exact role of RANK/RANKL in the pathogenesis of abnormal bone turnover in patients with chronic liver disease remains unclear, with conflicting findings on the levels of these proteins in serum. In one study of patients with cirrhosis, the osteoprotegerin levels were raised compared with those of controls, but the levels of RANKL were similar, and the osteoprotegerin/RANKL ratio was higher in patients with osteoporosis; this suggested that osteoprotegerin could be compensating for the negative bone turnover.11
Reduced bone formation in chronic liver disease has also been linked to low levels of insulin-like growth factor 1 (IGF-1) in serum.12 In osteopenic rats, a treatment with IGF-1 results in increased BMD. IGF-1 levels in patients with cirrhosis are also related to the severity of liver disease as measured by the Child-Pugh score.13 However, a direct causal link between IGF-1 levels and osteoporosis in chronic liver disease in humans has not been established.
Although unconjugated bilirubin inhibits osteoblast activity and function in vitro and in animal models, studies of patients undergoing liver transplantation have found no correlation between unconjugated, conjugated, and total bilirubin levels and BMD. A further recent study of hyperbilirubinemic Gunn rats also failed to show a difference between BMD and osteocalcin levels in the rats and wild-type rats, suggesting that elevated serum bilirubin alone is not a major contributory factor to hepatic osteodystrophy.14
Vitamin K is also involved in bone metabolism, mediating the carboxylation of glutamyl residues on bone proteins such as osteocalcin. In a population without liver disease, suboptimal vitamin K status has been associated with an increased risk of fractures, and supplementation has been associated with improvements in BMD.15
Prevalence of Osteoporosis in Chronic Liver Disease
Cirrhosis increases the risk of fracture by about 2-fold. Various studies over the last 2 decades have shown that the prevalence of osteoporosis in patients with cirrhosis is between 12% and 55%12, 16–24 (Table 2). The difference between studies is probably a reflection of differences in age, liver disease etiology, nutritional state, hypogonadism, and liver disease severity. In one study of 58 patients with viral cirrhosis, the risk of osteoporosis was shown to be associated with the severity of cirrhosis, with Child-Pugh A patients having a higher BMD than those with Child-Pugh C disease.18 In a study of 243 patients with mixed end-stage liver disease requiring transplantation, the only independent risk factors for osteoporosis were a lower body mass index in women and increasing age.20 The presence or absence of cholestasis did not help in ascertaining the risk of low bone mass density.
|n||Type of Cirrhosis||Osteoporosis (%)||End-Stage Liver Disease*||Fractures (%)||Reference|
|60||Mixed||47||Diamond et al.16 (1980)|
|74||Viral||20||No||6.7||Chen et al.17 (1996)|
|58||Mixed||43||Yes||Monegal et al.18 (1997)|
|32||Viral||55||No||Gallejo-Rojo et al.12 (1998)|
|81||PSC||16||No||3||Angulo et al.19 (1998)|
|243||Mixed||37||Yes||Ninkovic et al.20 (2000)|
|207||HCV plus alcohol||20||Yes||Carey et al.21 (2003)|
|104||Mixed||12||Yes||Sokhi et al.22 (2004)|
|156||PBC||43.7||Yes||22||Guichelaar et al.23, 24 (2006/2007)|
|204||PSC||32.5||Yes||16||Guichelaar et al.23, 24 (2006/2007)|
Noncirrhotic Biliary Disease
Noncirrhotic biliary disease has also been linked with increased risk of fracture (Table 2).
In PSC, the risk of fracture is related to increasing age, the duration of coexisting inflammatory bowel disease, and more advanced biliary disease.19 The prevalence of osteoporosis at the lumbar spine as measured by BMD varies from 8.6%-32% in end-stage PSC. In an important study reported in this journal23 of 204 patients with PSC awaiting transplantation, 15% already had fractures. However, bone mass in end-stage PSC does appear to have improved over the last 2 decades.24 This is not necessarily related to the use of ursodeoxycholic acid because in a small prospective study,25 there was no improvement in BMD in those patients given the drug over a 4-year period.
In contrast to PSC, a direct relationship between PBC and osteoporosis has been a subject of controversy for many years as patients with PBC are classically postmenopausal women who are at risk of osteoporosis.2, 3 However, 2 recent studies have confirmed a 4-fold increased risk of osteoporosis and a 2-fold increased risk of fractures in this group of patients in comparison with age-matched controls. In a Spanish study of 142 women with PBC with a mean age of 53, the prevalence of osteoporosis was 38% versus 10% in the control group, giving a relative risk of 3.83. It is not known how this translated into increased fracture risk. A higher Mayo risk score, older age, advanced histological stage, and lower body mass index are risk factors for osteoporosis.26 In a larger cohort study from the United Kingdom in which 930 patients with PBC were age-matched and sex-matched with 9202 subjects, the absolute excess of fractures was 12.5 per 1000 person years.27
There is little doubt that the prevalence of osteoporosis and fractures in patients with end-stage liver disease has fallen over the last 2 decades. In patients with biliary disease listed for transplantation, the Mayo group showed that between 1985 and 1989, 57% were osteoporotic versus 34% between 1990 and 1995 and 26% between 1996 and 2000. This may have been due to better nutrition and lower corticosteroid use.24
Chronic Liver Disease
Osteoporosis is unusual in the absence of cirrhosis, cholestasis, and risk factors outlined in Table 1, but there have been reports of increased prevalence in patients with noncirrhotic chronic liver disease.
Osteoporosis, as expected, is more common in patients with hemochromatosis and hypogonadism. A small study of hemochromatosis in the late 1990s suggested that high iron levels, rather than cirrhosis, are associated with osteoporosis in this group of patients.28 This has been emphasized by a further recent study of 38 men with hemochromatosis gene (HFE)-related iron overload, of whom only 13% were hypogonadal, whereas 34% were osteoporotic. Osteoporosis was not related to the presence of cirrhosis, and there was a fall in BMD with rising hepatic iron levels.29 However, the mechanism by which iron affects bone turnover has yet to be explained.
An adverse affect of ribavirin and interferon on BMD during the treatment of chronic hepatitis C, reported in one study, has not been confirmed by subsequent investigators.30
However, in a recent study, BMD was shown to be low in patients without cirrhosis with predominately chronic hepatitis C infection. Further studies are needed to verify this finding as there was no control group and no details were given of concurrent alcohol intake.31
BMD falls in the first 3 months following liver transplantation, rising again to pretransplant levels within 2 years. Fracture rates of 15%-27% have been reported (Table 3), with most fractures occurring within the first 2 years of transplantation.23 Studies for which lumbar spine x-rays were taken have reported higher levels of posttransplant fracture than those that relied on clinical symptoms. However, unlike pretransplant fractures, most posttransplant fractures appear to be symptomatic and are commonly vertebral fractures followed by rib fractures.2, 23
|n||Liver Disease||Mean Length of Follow-Up (Months)||Osteoporosis (%)||Fractures (%)||Fracture Detection||Reference|
|58||Mixed||12||43||Radiological||Monegal et al.18 (1997)|
|130||Mixed||3||39||27||Radiological||Ninkovic et al.20 (2000)|
|130||Mixed||36||21*†||Radiological and clinical||Leidig-Bruckner48 (2001)|
|207||Mixed||12||17||Radiological||Carey et al.21 (2003)|
|360||Biliary disease||63||44||34*||Radiological||Guichelaar et al.23 (2007)|
The role of calcinurin inhibitors in bone turnover following transplantation remains controversial. Posttransplant bone biopsies have shown that tacrolimus-treated patents have an earlier recovery of bone metabolism and trabecular bone structure than those taking cyclosporine, but their effect on early bone loss remains unclear.7 The cumulative steroid dose has been implicated in impairing bone formation in these first few months, and a change in practice with lower steroid usage after transplantation probably explains the fall in fracture frequency observed anecdotally over the last decade and supported by the publication of Guichelaar et al.23 in this month's issue of HEPATOLOGY.
Pretransplant vertebral fracture is more predictive of posttransplant fracture than BMD, and this emphasizes the need to optimize bone health before transplantation. However, BMD is clearly important as a reduced BMD before and after transplantation has been associated with higher rates of fracture. There is no consensus on the effects of the type of pretransplant liver disease, gender, age, and menopausal status on the risk of posttransplant fracture. Although most patients before transplantation have normal serum parathyroid hormone (PTH) levels, the PTH, vitamin D, and normal free testosterone levels rise in the first year following transplantation.24, 32
Assessment and Indications for BMD
There has been some debate, along with some differences in recommendations, about when to assess bone density in patients with chronic liver disease,2, 3, 33 and this reflects the difficulty of predicting the risk of osteoporosis and fracture on an individual basis (Table 4). The path of least resistance is to assess BMD in all patients with chronic liver disease. There is a clear consensus that BMD should be undertaken in anyone who has had a fragility fracture or who is on long-term corticosteroids and that treatment should be initiated if osteoporosis is confirmed. BMD should also be assessed prior to liver transplantation and probably in all patients with cirrhosis. The indications for BMD are less clear in patients with cholestatic liver disease. The American Gastroenterological Association guidelines2 suggest that BMD should be considered in all patients with PBC at diagnosis, whereas other recommendations limit BMD to patients with bilirubin greater than 3 times the upper limit of normal.3 The finding from the Mayo Clinic that 26% of PSC patients requiring transplantation are osteoporotic suggests that this group should also be considered for BMD screening, particularly if they have concurrent inflammatory bowel disease.24
|Indication||BSG3||AGA2||Pares and Guanabens33|
|PBC at diagnosis||+|
|Liver transplantation assessment||+||+||+|
|Postmenopausal and other risk factors for osteoporosis||+||+|
|Premature menopause (<45 years) and secondary amenorrhea||+|
The frequency of follow-up BMD measurements remains debatable. In the absence of osteoporosis, there is probably no need to repeat BMD more frequently than 3 times a year.2, 3 There is currently no role, as discussed later, for the routine use of serum and urinary markers of bone turnover in stratifying risk for fracture and the need for BMD or in assessing worsening bone health during follow-up.
Current recommendations are that the treatment for osteoporosis should be initiated in patients older than 65 years with chronic liver disease who are likely to need the equivalent of 7.5 mg daily or more of corticosteroids for more than 3 months, although there is probably an increased risk at a lower dose. In younger patients with chronic liver disease, treatment is indicated if BMD shows a T score of less than −2.5.1–3, 33
Although serum markers of bone turnover, such as procollagen peptides of type 1 collagen, are being used in postmenopausal osteoporosis to assess the response to treatment, they have been poorly studied in patients with chronic liver disease and currently cannot be used in patients with chronic liver disease.
The evidence for the treatment of osteoporosis in patients with liver disease is based on trials of patients with postmenopausal osteoporosis. There have been only a few small studies of patients with chronic liver disease, mainly PBC, and improvements in BMD rather than more clinically important fracture rates have been used as outcome measurements (reviewed in Pares and Guanabens33).
Increasing weight-bearing exercises that reduce bone loss and increase core strength to prevent falls should be used in addition to pharmacological therapy
Calcium and vitamin D supplementation has been shown to reduce fractures in elderly patients living in sheltered accommodations;1 consequently, patients in trials of other treatment modalities for osteoporosis are routinely given vitamin D and calcium supplementation, usually 800 IU of vitamin D and 1 g of calcium daily. There have, however, been few clinical trials on the effects of vitamin D supplementation alone, the optimal dose, and the formulation in patients with cirrhosis. In a nonrandomized study of 25 patients with alcoholic cirrhosis and low serum vitamin D levels, hydroxyvitamin D3 supplementation did increase BMD over baseline values.2 In a further study of 76 patients with cirrhosis, 1-alpha-25-hydroxyvitamin D3 resulted in a small increase in BMD at the lumbar spine, but no data were given on the effect on fractures.34 Given the low serum levels of vitamin D in patients with cirrhosis, rising PTH level, and falling urinary calcium loss after transplantation, further studies of vitamin D and calcium supplementation are probably warranted in patients with chronic liver disease.
Alendronate, etidronate, ibandronate, and risidronate, all of which prevent bone resorption, are used for treating postmenopausal osteoporosis. These drugs are usually given with calcium and vitamin D. They have been studied only in a small number of patients with chronic liver disease, and so data on their efficacy in this setting are lacking (reviewed in Pares and Guanabens33). Most studies have been conducted with PBC patients rather than patients with cirrhosis. In PBC, alendronate increases bone mass. In one study comparing daily alendronate with etidronate given over a 2-year period, the improvement in lumbar BMD was more marked in the alendronate group. Markers of bone turnover also fell with alendronate, although no lumbar spine fractures were seen in either group.35 There were no esophageal complications from daily alendronate, although another study of PBC has suggested that alendronate is better tolerated once weekly than once daily. In a further 12-month placebo-controlled study of osteopenic PBC patients, alendronate was associated with improvements in both spinal and femoral BMD, but the study was not powered to show an effect on fracture rates.36 There are no studies of the safety or efficacy of bisphosphonates in cirrhosis, but anecdotally bisphosphonates appear to be well tolerated once weekly, although it would be reasonable to exercise caution in using the drug in patients with recent esophageal banding/sclerotherapy.
Bisphosphonates have been used in 7 studies following liver transplantation in an attempt to reduce fractures. Five of these studies assessed the use of intravenous pamidronate. In 1 of these, which used historical controls only, a reduction in fractures was shown. Three other studies showed an improvement in BMD, but this was mainly limited to trabecular bone (lumbar spine) and not cortical bone. In the fifth randomized study of 99 patients, there was no effect on BMD and fractures, but this was probably a reflection of the absence of lumbar spine bone loss and a low incidence of fractures in the controls (8%).37 Oral alendronate has been studied only in an uncontrolled trial following transplantation. Zoledronic acid, a more potent bisphosphonate, given within 7 days of transplantation and then at 1, 3, 6, and 9 months appears to reduce acute bone loss in the first 3 months following transplantation.38 Complications include temporary secondary hyperparathyroidism and postinfusional hypocalcaemia in some patients. The lack of an effect on the fracture rate in all these studies probably reflects the size of the studies, given the overall fall in the rate of posttransplant fractures over the last decades as described by Guichelaar et al.23 in this journal.
A recently recognized rare complication of bisphosphonates is osteonecrosis of the jaw.39 The majority of reports have concerned cancer patients receiving intravenous zoledronic acid or pamidronate, but there have been reports of patients without cancer receiving lower oral doses. About two-thirds of the cases have occurred following tooth extraction.
Hormone Replacement Therapy
This has become second-line therapy in postmenopausal osteoporosis because of the concern about the risks of thromboembolic disease and gynecological malignancy. It can be used safely in patients with chronic liver disease and is ideally given transdermally. There have been only 2 small randomized controlled trials of patients with PBC and patients without cirrhosis. In one study, transdermal hormone replacement therapy was shown to improve BMD after 2 years, but the study was not powered to show any effect on fractures.40 In the second study of 31 patients, hormone replacement therapy reduced bone loss at the femoral neck. However, the difficulty in recruiting women to such a trial was evidenced by the high dropout on treatment, and thus it was not possible to recruit enough patients to show an effect on fracture rates.41 In a further open labeled study of 18 patients with PBC, only 1 patient had to stop the transdermal hormone replacement therapy because of rising aminotransferases.42
This is a selective estrogen receptor modulator that has less effect on markers of bone turnover than bisphosphonates but has been shown to reduce the rate of vertebral fractures in postmenopausal osteoporosis, although it has no effect on femoral fracture rates.1 There are few data on its efficacy in chronic liver disease, with one pilot study suggesting that it may prevent bone loss in PBC when given for a year.43
Hypogonadal patients with chronic liver disease can be treated with transdermal testosterone as it leads to stable hormone levels, preventing exposure of the liver to surges in levels seen with oral or depot preparations. One concern about restoring testosterone levels to normal in patients with cirrhosis has been the theoretical risk of increasing the risk of hepatocellular carcinoma.
Strontium reduces vertebral fracture and nonvertebral fracture in postmenopausal osteoporosis with the main side effect of diarrhea. Its mechanism of action remains unclear, and to date, there have been no studies of its efficacy in chronic liver disease. Its use may be considered in patients intolerant of bisphosphonates.
Recombinant PTH has been used for the secondary prevention of fragility fractures in postmenopausal women intolerant of bisphosphonates. It acts by stimulating new bone formation.
Osteoporosis remains the most clinically important form of hepatic osteodystrophy and is easy to miss clinically. Despite a fall in incidence over the last 2 decades, up to a quarter of patients requiring transplantation for end-stage biliary disease will be osteoporotic, and as many as 12% will already have fractures. It is important to identify such individuals before they develop fractures as this increases the risk of posttransplant fractures and significant morbidity can be associated with lumbar spine fractures even in nontransplant patients with cirrhosis. Identifying patients with chronic liver disease at risk of osteoporosis still relies on BMD, but clinical risk factors such as corticosteroid use and hypogonadism should be taken into account when risk is assessed. The evidence for the treatment of osteoporosis in chronic liver disease is still based on large studies of postmenopausal women because of the few studies of chronic liver disease. Bisphosphonates do improve BMD in cholestatic patients and following transplantation and are well tolerated, but the studies to date have been too small to show a beneficial effect on fracture rates.1, 2