Osteopenia and osteoporosis have long been recognized as complications of cirrhosis, especially in patients with chronic cholestatic liver disease.1–4 It is, however, only in the era of successful orthotopic liver transplantation (OLT) that the full clinical significance of morbidity from fracturing has been realized. Despite the high prevalence of bone disease in chronic cholestatic liver disease, the underlying etiologic mechanism of bone loss remains obscure and its management empiric and unsatisfactory.
Preexisting low bone mineral density (BMD) of the spine at the time of OLT and pretransplant fracturing have been shown to be major risk factors for posttransplant skeletal complications.5, 6 In the early months after orthotopic liver transplantation, patients suffer from rapid bone loss that leads to a high incidence of atraumatic fracturing in the early post-operative years.7–9 Since bone loss is seen early after all solid organ transplants,9, 10 it has been assumed that high doses of glucocorticoids play a major etiologic role, and this has recently been confirmed in bone histomorphometric analysis of bone biopsy specimens after OLT.11 Other factors integral to the early postoperative course may further contribute to bone loss, including immobility, and disturbances of mineral metabolism.7, 12–14 Fortunately for most patients, this first postoperative year represents the nadir in bone density and a recovery of bone mass starts to occur after the early months of bone loss.
Although some studies have assessed bone loss and bone gain after OLT, the study samples were small, and data of long-term follow-up after OLT are limited. In addition, analyses of risk factors for pretransplant hepatic osteoporosis or osteopenia and posttransplant bone loss and bone gain have not been well established, although these are important to determine preventive strategies. We aimed to determine the prevalence and predictive factors for low bone mass before OLT, posttransplant bone loss, and bone gain at the lumbar spine with long-term follow-up after OLT in a large population of 360 consecutive adult patients with end-stage primary biliary cirrhosis (PBC) and primary sclerosing cholangitis (PSC), transplanted at a single center.
OLT, orthotopic liver transplantation; BMD, bone mineral density; PBC, primary biliary cirrhosis; PSC, primary sclerosing cholangitis; MELD, model for end-stage liver disease; IBD, inflammatory bowel disease; CyA, cyclosporine A; TAC, tacrolimus; PTH, parathyroid hormone; BMD-LS, BMD at the lumbar spine; HRT, hormone replacement therapy; BMI, body mass index.
PATIENTS AND METHODS
From March 1985 to January 2001, all adult patients undergoing OLT at the Mayo Clinic Rochester with a diagnosis of either PBC or PSC were assessed clinically, biochemically, radiologically, and by measurements of BMD before OLT, at 4 and 12 months after OLT, and annually thereafter. The diagnoses of PBC and PSC were made according to well-established criteria.15–17 The study was approved by the Institutional Review Board of the Mayo Clinic. All patients were followed until January 2002, death, or retransplantation.
All patients underwent a full clinical examination at each time of evaluation: before OLT, at 4 and 12 months after OLT, and annually thereafter. Liver function status was assessed by Child-Turcotte-Pugh score and model for end-stage liver disease (MELD) score. Functional status was assessed by the Karnofsky performance scale from 0 (completely immobile) to 100 (normal activity).18 Female patients were judged to be premenopausal or postmenopausal according to clinical symptoms and biochemical testing. Nutritional status, muscle wasting, complications of liver disease, and other illnesses were noted and medications were recorded. The presence or absence of muscle wasting was a global assessment made by the transplant hepatologist. The dietary calcium intake of each patient was estimated, and general dietary instructions were given before OLT, during the early postoperative period, and annually thereafter. Oral calcium supplements (1.5 g/day) were prescribed for all patients. Vitamin D supplementation was given in response to low 25-hydroxyvitamin D levels to normalize serum levels. Inflammatory bowel disease (IBD) was diagnosed by colonoscopy and surveillance biopsies at time of activation for orthotopic liver transplantation and annually thereafter in all PSC patients.
From 1985 to 1990, all patients were treated with a standard protocol of cyclosporine A (CyA) and prednisone with or without azathioprine. From 1990 to 1993, some patients were treated with tacrolimus (TAC) and prednisone as part of the multicenter FK506 trial, the control arm of which was standard triple therapy with prednisone, CyA, and azathioprine. The TAC arm of this study received only about half the total prednisone dose of the CyA arm; patients were tapered to 5 mg per day of prednisone by 4 months, whereas the CyA patient were tapered to 10 mg per day of prednisone by 6 months. In 1994, standard immunosuppression was changed to tacrolimus, prednisone, and azathioprine, though some patients were treated as part of a multicenter trial with Neoral. In 1999, azathioprine was replaced by mycophenolate mofetil 1 g twice daily for the first 2-4 postoperative months. CyA and tacrolimus doses were adjusted according to serum levels, the desired levels depending on time from transplantation, presence of rejection, and toxicity. During the whole study period, acute cellular rejection of the allograft was treated with 1 g of intravenous Solu-Medrol on days 1, 3, and 5 after the diagnosis of the rejection. Steroid-resistant rejection was treated with monoclonal therapy with OKT3.
Blood was taken for biochemical assessment after an overnight fast and tested by Mayo Medical Laboratories, using standard methods. Biochemical testing included parameters of liver and kidney function (serum albumin, total alkaline phosphatase, total and direct bilirubin, international normalized ratio, serum creatinine), as well as parameters of mineral metabolism (serum 25-hydroxyvitamin D, serum calcium, serum phosphorus, parathyroid hormone [PTH]). Serum 25-hydroxyvitamin D was measured by the method of Kao and Heser.19 Immunoreactive parathyroid hormone was measured by immunochemiluminescent-metric assay20 since June 1989. PTH values achieved by preceding methods were excluded from this study.
Measurements of BMD at the lumbar spine (BMD-LS)
BMD measurements at the lumbar spine were taken in all patients in a protocolized fashion; at time of activation for OLT, 4 months after OLT, at 1 year, and yearly thereafter. From March 1985 to April 1988, BMD measurements were performed on lunar machines by dual photon absorptiometry. Since April 1988, dual energy X-ray absorptiometry using Hologic machines, has been used. Phantoms were used to cross-calibrate the 2 machines, and a formula was established to convert DPA data to dual energy X-ray absorptiometry data. Bone mass was corrected for bone size to calculate BMD in g/cm2. Measurements of BMD-LS had a reproducibility of 2.2%. In patients with compression fractures, measurements were determined on only the intact vertebrae. BMD readings (all calibrated to dual energy X-ray absorptiometry measurements) were expressed as T scores (standard deviations from peak bone mass of a young, sex-matched reference population) and Z scores (standard deviations from age- and sex-adjusted reference values). Osteoporosis is defined according to World Health Organization criteria21; a T score of higher than −1.0 is considered normal, a T score between −1 and −2.4 is osteopenia, and a T score of −2.5 and less is osteoporosis.
Following liver transplantation, patients were followed till death or retransplantation. To study bone loss and bone gain, BMD values of all patients were synchronized at their OLT date (time = 0), taking the overall mean as the baseline BMD value. Posttransplant BMD values were derived by taking from every patient the change from baseline and plotting that from OLT. This method was used to limit bias caused by a changing population with time. The posttransplant BMD points were plotted and fitted using least squares regression with time as the predictor of BMD and using a restricted cubic spline with knots at 4 months and 3 years (natural spline function in Splus [Mathsoft, Seattle, WA]).
Parameters are reported as means ± standard deviation. Associations between patient characteristics and quantitative endpoints such as BMD-LS or qualitative endpoints such as osteoporosis were tested for significance using the Student t test for comparison of means or the chi-square test for comparison of proportions. Changes over time were assessed by the analysis of variance method. The Pearson correlation coefficient was used to assess the linear association between quantitative patient characteristics and BMD. The patient characteristics that were univariately significant formed a pool of potential predictors of BMD. The backwards elimination variable selection procedure was used to find the independent variables that predicted these endpoints. Such variables with a P value less than 0.05 were included in the final multivariate regression model to predict low BMD. All analyses were performed using the SAS data analysis system (SAS Institute, Cary, NC).22
Demographics of Total Patient Population
Three hundred-sixty adult patients underwent OLT for PBC (156 patients) or PSC (204 patients) and were followed after OLT; 142 were males and 218 were females, of whom 148 were postmenopausal (Table 1). The PBC patients were predominantly females (135 of 156 patients), whereas the PSC patients were predominantly males (121 of 204 patients). The mean age was 49.5 ± 10.5 years, with PSC patients being younger than PBC patients (P < 0.01), and males being younger than females (P < 0.01). The mean Child-Turcotte-Pugh score of 8.7 ± 1.8 and MELD score of 17.3 ± 8.8 indicated advanced cholestatic liver disease. Most patients (86.1%) had near-normal activity with minor symptoms and assistance at time of OLT (Karnofsky score, 50-100); 26 (7.7%) patients required considerable assistance (score, 30-40), and 21 (6.2%) patients were severely disabled or hospitalized (score, 0-20). A total of 8 (3.2%) patients smoked at time of activation for OLT, and 19 (7.6%) patients had been active drinkers. There were no differences between subpopulations (PBC/PSC and females/males) concerning Karnofsky scoring or drinking/smoking habits. The majority of patients (81.1%) had muscle wasting at time of OLT, males more often than females (P < 0.01). IBD was seen in 139 of 204 PSC patients (66.7%), with 95.6% having chronic ulcerative colitis; 46 patients were on glucocorticoid treatment within 5 years before OLT with a mean duration of treatment of 5.0 ± 5.6 years. At time of activation for OLT, 64 patients were on cholestyramine, 111 were on ursodeoxycholic acid, 5 were on anticonvulsants, 45 were on thyroid hormone replacement therapy (HRT), 18 females were on (HRT), and 4 patients used bisphosphonates.
Table 1. Demographic Data of Patients With End-stage PBC and PSC Before OLT
NOTE: Values of BMD/, and T/ and Z scores are expressed as means ± SE; the remaining variables are expressed as mean ± SD.
Years since first diagnosis until OLT.
Poor nutritional status, muscle wasting, Karnofsky scoring, smoking, and drinking history were recorded for 248 of the total patients.
Differences between PBC vs. PSC, females vs. males:
To determine any temporal changes, the 16-year study period (1985-2000) was divided into 3 transplantation periods: 1985-1989 (n = 93), 1990-1995 (n = 153), and 1996-2000 (n = 115) (Table 2). The percentage of patients with PBC and PSC and the gender distribution remained the same over these 3 periods, but PBC patients became older and more postmenopausal with time. Age and menopausal status did not change in the PSC population. Increases in mean body mass index (BMI) and decreases in muscle wasting were seen over time, but no temporal differences were seen in Child-Turcotte-Pugh, MELD, or Karnofsky scores.
Table 2. Temporal Changes in Clinical, Biochemical and Bone Density Parameters Before OLT
Assessment of Baseline BMD-LS and Biochemical Variables
At time of activation for OLT, 38% of patients had osteoporosis, 39% osteopenia, and only 23% of patients had normal bone mass (Table 1). PBC patients had a higher prevalence of osteoporosis and lower BMD than PSC patients (P < 0.001) and, although female patients had lower BMD values than male patients (P < 0.0001), no significant differences concerning T and Z scores were observed between the subpopulations. The T scores of the 64 cholestyramine-treated patients did not differ significantly from T scores in the untreated patients (T scores, −2.1 ± 1.4 vs. −2.0 ± 1.4), nor did the T scores of the 46 glucocorticoid-treated patients differ from untreated patients (T scores, −2.2 ± 1.6 vs. −2.0 ± 1.4). In addition, PSC patients with and without IBD before OLT had similar T scores at baseline (T scores, −2.0 ± 1.2 vs. −1.8 ± 1.4).
In the early period of 1985-1989, 57% patients had pretransplant osteoporosis with a mean T score of −2.5 ± 1.6 and Z score of −2.0 ± 1.6; in the latest period of 1996-2000, 26% had osteoporosis with a mean T score of −1.7 ± 1.2 and Z score of −1.0 ± 1.2. The increase in pretransplant bone mass with time was seen in the female, male, and PSC subpopulations, but bone mass remained stable in the PBC patients (Table 2).
Biochemical indices at time of OLT reflected end-stage cholestatic liver disease with low albumin (3.0 ± 0.5 g/dL), high total and direct bilirubin (11.2 ± 11.0 mg/dL and 7.0 ± 7.1 mg/dL, respectively), and high alkaline phosphatase (1,107.8 ± 855.1 U/L). Levels of serum calcium (8.8 ± 0.6 mg/dL), 25-hydroxyvitamin D (20.2 ± 17.9 ng/mL), phosphorus (3.5 ± 0.8 mg/dL), creatinine (1.2 ± 0.9 mg/dL), and PTH (2.3 ± 2.0 pmol/L) were in the normal range. Male patients had higher alkaline phosphatase levels than female patients (1,177.5 ± 934.4 vs. 1,000.7 ± 706.5, respectively; P < 0.01); no other differences were observed between subpopulations (PBC/PSC, females/males). Increases in vitamin D levels and decreases in serum bilirubin and alkaline phosphatase levels were seen with time over the study period (Table 2).
Clinical Characteristics After OLT
After OLT, patients spent, on average, 25.6 ± 21.0 days in the hospital during the first 4 months. Fifty-one patients (14.2%) were retransplanted after OLT at a mean of 1.3 ± 2.6 years; 78 patients (21.7%) died after transplantation at a mean of 5.1 ± 4.3 years; 160 patients (44.4%) had rejection episodes after OLT; and 43 patients (11.9%) sustained nonanastomotic biliary strictures. The average daily dose of prednisone during the first month was 140.0 ± 49.3 mg; this was less in patients treated with TAC compared to those receiving CyA (P < 0.01; 124.8 ± 42.0 vs. 145.8 ± 51.5 mg/dL, respectively). Similarly, the average daily prednisone dose from month 1 to month 4 was higher in patients treated with CyA (P < 0.01; 54.4 ± 17.9 mg) compared to those treated with TAC (39.6 ±13.4 mg). Mean CyA serum levels were 242.2 ± 89.0 ng/mL (215 patients) and 222.5 ± 67.6 ng/dL (201 patients) at 1 and 4 months, respectively; mean TAC levels were 9.4 ± 3.6 ng/dL (119 patients) and 9.0 ± 2.8 ng/dL (117 patients) at 1 and 4 months. There are no differences in immunosuppression between males and females or between PBC and PSC. Temporal changes in immunosuppression were seen with the more recent transplanted patients having less rejection episodes, less hospitalization days, less treatment with CyA, more treatment with TAC, lower cumulative prednisone doses and lower CyA serum levels when compared to the earlier 2 transplantation periods (Table 3).
Table 3. Temporal Differences in Clinical, Immunosuppressive, and Biochemical Parameters After OLT
Hospitalization days during first 4 mo posttransplant.
Values are listed as average daily doses of prednisone and average daily serum levels of TAC and CyA.
Laboratory differences between the 3 transplantation periods were assessed at 4 mo, 1 y, and 2 y posttransplant. There were no significant differences concerning total and direct bilirubin, alkaline phosphatase, calcium, and vitamin D at all time points.
No patients were treated with TAC from 1985 to 1989. PTH was measured by an earlier, less reliable assay in this era.
Hormone replacement therapy (HRT) was used in 57 of 148 postmenopausal females after OLT, starting at 939.4 ± 587.3 days after OLT; only 27 patients started during the first 2 years after OLT. Sixteen patients received bisphosphonates after OLT: etidronate in 1 patient (in first post-OLT year), pamidronate in 2 patients (in first and second year after OLT,) and alendronate in 13 patients (all starting after the first 2 posttransplant years). Sixty-three patients were enrolled in a randomized trial of calcitonin therapy or no treatment after OLT; this trial showed no effect of calcitonin therapy on bone mass after OLT.23
Biochemical Characteristics After OLT
After OLT, liver function improved (P < 0.001), and at 4 months, levels of total bilirubin (2.1 ± 5.2 mg/dL) and alkaline phosphatase (299.9 ± 499.0 U/L) had decreased and serum albumin (3.7 ± 0.6 g/dL) increased. Albumin further increased by 1 year posttransplant to 4.0 ± 0.4 g/dL after which it remained stable. Alkaline phosphatase showed a small but significant increase from 4 to 8 years after OLT, leading to alkaline phosphatase levels of 207.1 ± 291.1 U/L at year 8. Kidney function deteriorated with increased serum creatinine at 4 months (1.3 ± 0.6 mg/dL, P < 0.0001) and with a further increase by year 4 posttransplant (1.4 ± 0.5 mg/dL, P < 0.05), after which creatinine remained stable. Serum calcium (9.1 ± 0.8 mg/dL), 25-hydroxyvitamin D (31.6 ± 13.8 ng/mL), phosphorus (3.8 ± 0.7 mg/dL), and PTH (4.3 ± 4.5 pmol/L) all increased after OLT (P < 0.05). PTH further increased to 4.8 ± 3.4 pmol/L at 1 year posttransplant, whereas vitamin D continued to increase during the first 4 posttransplant years to 40.9 ± 14.9 ng/mL (P < 0.001). Phosphorus decreased to 3.5 ± 0.6 mg/dL (P < 0.05) at year 4 posttransplant, after which it remained stable. Recently transplanted patients had higher albumin and PTH levels and lower creatinine and phosphorus levels compared to earlier patients (Table 3).
Follow-up of Posttransplant BMD-LS After OLT
T and Z scores at the lumbar spine fell at 4 months after OLT (Table 4) (Fig. 1), with a high incidence of osteoporosis (51%) in the total patient population. Thereafter, BMD increased up to 4 years after OLT and remained above pretransplant levels in the total patient population. Z scores continued to improve from 4 to 8 years, whereas T scores remained stable (Fig. 1). The same pattern of changes in posttransplant BMD, with bone loss followed by bone gain, was seen when comparing PBC and PSC patients (Fig. 2) and male patients and premenopausal and postmenopausal female patients (Fig. 3). However, although all groups of patients lost bone mass early after OLT, recovery of bone mass differed depending on the initial severity of osteopenia or osteoporosis. Patients with the lowest baseline BMD experienced the greatest gain in bone mass after OLT, with BMD exceeding baseline levels by 1-2 years after OLT; on the other hand, patients with the highest baseline BMD failed to recover to baseline BMD values at any time after OLT (Fig. 4).
Table 4. BMD-LS in 360 OLT Recipients Over 8 years
The rates (adjusted to annual rates) of early bone loss and bone gain were analyzed in the total patient population and in the subpopulations using changes in BMD from baseline. A high rate of bone loss in the first 4 months was seen (15.9 ± 18.9 %/y), after which bone mass started to increase from 4 to 12 months at an annual rate of 6.4 ± 14.3% and from 12 to 24 months at an annual rate of 6.7 ± 20.3%. Three hundred seventeen patients (82.2%) lost bone mass during the first 4 months after OLT, of which 100 patients (41.3%) lost 0-5%, 91 patients (37.6%) lost 10-15%, and the remaining 8 patients (2.5%) lost 15-30% of baseline BMD. Eleven patients (4.6%) had stable BMD after OLT, and 32 (13.2%) patients gained 1-5% bone mass during the first 4 months. The rate of bone loss in the first 4 postoperative months was significantly more in patients with PSC compared to those with PBC (18.0 ± 20.0 %/y vs. 13.3 ± 17.8 %/y, respectively). No other significant differences in rates of bone loss (0-4 months) and bone gain (4-24 months) were seen between PBC and PSC subpopulations, or between females and males. Postmenopausal women treated with HRT, started on average at 3 years posttransplant, have better T scores from 4 to 8 years after OLT than postmenopausal women without HRT (mean T scores at 2, 4, and 8 years after OLT in patients with HRT are −2.15 ± 0.17, −1.88 ± 0.22, and −1.43 ± 0.48, respectively compared to the scores of those without HRT: −2.28 ± 0.22, −2.10 ± 0.37, and −2.55 ± 0.88, respectively).
After OLT, the rate of early posttransplant bone loss during the first 4 months after OLT was not significantly different in the 3 transplantation periods (Fig. 5). Bone gain during the first 2 years, however, was greater in the more recently transplanted patients, with 11.2% ± 19.3 gain in patients transplanted during 1996-2000 compared to 4.4% ± 13.8 during the 1985-1989 period and 3.8% ± 10.2 during 1990-1995.
Predictors for Pretransplant Bone Density
Univariate analysis (Table 5) showed that pretransplant T scores of the lumbar spine correlated positively with male gender, BMI, and albumin, and negatively with age, PBC disease, duration of disease (y), OLT number, postmenopausal status, muscle wasting, and alkaline phosphatase. No correlations were seen with Child-Turcotte-Pugh score, MELD score, Karnofsky scoring, smoking and alcohol use, IBD status in PSC patients, and other biochemical indices, including calcium and vitamin D levels. Multivariate analysis indicated that independent risk factors for low pretransplant BMD were decreased BMI, female gender, an older age, increased alkaline phosphatase, decreased albumin, and higher OLT number.
Table 5. Univariate Correlations for Pretransplant and Posttransplant BMD in Patients With End-stage Cholestatic Liver Disease
NOTE: No correlations before or after OLT were seen with Child-Turcotte-Pugh and MELD score, pretransplant nutritional status, Karnofsky scoring, or alcohol use. In addition, no correlations were seen with most pretransplant labs, including calcium and vitamin D markers. Following OLT, bone loss and bone gain rates did not correlate with posttransplant rejection, hospitalization days, and serum levels of CyA and TAC.
Age and duration of disease are reported in decades. Alkaline phosphatase, cyclosporine levels, and OLT numbers are reported as hundreds of units.
Independent predictors for baseline BMD (model r-square, 17%), posttransplant bone loss during the first 4 months (model r-square, 6%), and posttransplant bone gain from 4 to 24 months (r-square, 5%)
Several pretransplant factors (Table 5) correlated with posttransplant loss at the lumbar spine bone during the first 4 months; more bone loss at 4 months was seen with PSC as the underlying liver disease, younger age at time of OLT, no presence of IBD disease before OLT, smoking at time of OLT, higher pretransplant BMD, and shorter duration of disease before OLT. The posttransplant factor that correlated positively with bone loss was an increased serum direct bilirubin level at 4 months; a trend was seen with total bilirubin and posttransplant nonanastomotic biliary strictures. Bone loss was not affected by gender, postmenopausal status, pretransplant Child-Turcotte-Pugh and MELD scores, pretransplant nutritional status, BMI and muscle wasting, Karnofsky score, alcohol intake, or any pretransplant biochemical index other than bilirubin. In addition, no correlations were seen with posttransplant rejection, hospitalization days, glucocorticoid doses or serum levels of TAC, although a trend was seen with 4-month serum CyA levels. Multivariate analysis of the pretransplant and posttransplant risk factors indicated that independent risk factors for posttransplant bone loss were PSC disease, shorter duration of disease, and higher posttransplant serum direct bilirubin.
Predictive Factors for Posttransplant Bone Gain
Univariate analysis indicated that bone gain at the lumbar spine during the first 2 posttransplant years (Table 5) was increased in those who were premenopausal at time of OLT, and in those with low pretransplant BMD, low 4-month BMD, and higher OLT number. Bone gain was increased with higher posttransplant levels of vitamin D and PTH. Less bone gain occurred in the presence of increased posttransplant levels of creatinine, bilirubin, alkaline phosphatase, and phosphorus, by the development of nonanastomotic biliary strictures, and by higher average daily doses of glucocorticoids (trend). Bone gain after OLT was not affected by posttransplant rejection, longer hospitalization stay or serum levels of calcineurin inhibitors. Multivariate analysis of the pretransplant and posttransplant factors indicated that the only independent predictive factor for posttransplant bone gain was OLT number.
Osteoporosis and its milder form, osteopenia, are important complications of advanced chronic liver disease and are found with a high incidence in patients awaiting liver transplantation, especially for chronic cholestatic liver disease.1–4 Despite its frequency, hepatic bone disease is generally overshadowed by the more urgent complications of chronic liver disease and may remain unrecognized unless the diagnosis is specifically sought. Following OLT, however, this situation changes, as early aggressive bone loss occurs in almost all patients.5–10 In patients who are already osteoporotic or osteopenic, this further bone loss results in an increase in fracture rates.7–9 Studies have indicated that this early period of bone loss is followed by recovery of bone metabolism11–14 and a subsequent gain of BMD, but long-term follow-up data are limited. Although immunosuppression is assumed to play a role in posttransplant bone loss, risk factors for bone loss and bone gain after OLT have not been well established, and larger studies assessing bone disease before and after OLT are lacking. Whether temporal changes have led to a reduction in pretransplant and posttransplant bone disease is unclear. We therefore studied prospectively a large cholestatic population before and after OLT with long-term follow-up after OLT, to assess predictive factors for low bone mass before OLT and for posttransplant bone loss and subsequent bone gain.
Previous studies have shown a correlation between bone disease in PBC and PSC and advanced histologic disease,4, 24, 25 and this present study confirms the high prevalence of osteoporosis and osteopenia in patients with advanced PBC and PSC. Gender did not influence the degree of osteopenia or osteoporosis seen in our patient population. Just as females in general have lower BMD than males, cholestatic females had lower BMD than cholestatic males, but when values were adjusted for age and sex, similar Z scores were found in PBC and PSC patients, supportive of a common etiologic role in these 2 cholestatic diseases. The important etiologic role of cholestasis on pretransplant BMD has been suggested in previous studies,24, 26–28 but the actual connection between low BMD and cholestasis remains obscure. There was a direct correlation in our study between alkaline phosphatase levels and pretransplant BMD, but the relative importance of liver and bone isoenzymes to this correlation is unknown, and no correlation was seen with bilirubin levels. Low serum albumin correlated with low BMD, but it is difficult to know if this represents an effect of nutritional status or of hepatic synthetic function on bone metabolism. As in previous studies,29–31 no link between pretransplant osteopenia and any abnormality of calcium or mineral metabolism was found. All patients received calcium supplements and vitamin D therapy to correct low serum levels, but despite this, pretransplant osteopenia was very common, suggesting no important role for abnormalities in calcium or vitamin D metabolism in the etiology of cholestatic osteopenia. Whether allelic variants of the vitamin D receptor have an effect on hepatic bone loss remains to be determined.32–34
Over the last 2 decades, changes have occurred in the management of advanced liver disease, in immunosuppressive regimens, and in the allocation and waiting time for liver transplantation. Over this time, an improvement is seen in pretransplant BMD—spinal T scores increased from −2.5 before 1990 to −1.7 after 1996—and this may give more insight into etiology. The severity of the liver disease, as reflected by MELD and Child-Turcotte-Pugh scores, has not changed, the duration of disease before OLT has increased, and patients have become older. On the other hand, BMI has increased, muscle wasting and nutritional status (including vitamin D levels) have improved and bilirubin is lower; these factors may all have contributed to increased BMD before OLT.
An independent determinant of pretransplant low BMD identified in this patient population was low BMI, as seen in other studies.2, 24, 25 BMI reflects both lean and fat tissue mass and influences BMD by several mechanisms: lean tissue stimulating the skeleton by providing mechanical stress,35, 36 and fat tissue effecting the production of leptin37, 38 and estradiol in female patients.39, 40 The importance of lean tissue on BMD is further emphasized by the independent correlation seen in our study with muscle wasting. Muscle wasting may simply reflect less physical activity in these patients, although most study patients here were ambulatory before OLT despite their advanced liver disease. These findings further stress the importance of adequate nutritional status and physical activity to prevent pretransplant bone loss. Before OLT, male patients had significantly more muscle wasting than female patients, a difference that could not be explained on the basis of their Karnofsky performance status alone. In males, however, muscle mass is also dependent on the anabolic effect of testosterone.41, 42 Testosterone levels have been shown to be reduced in males with advanced cholestatic liver disease, with a 70% incidence of hypogonadism (free testosterone <9 pg/dL).11 Hypogonadism may therefore be another contributing factor to muscle wasting (and osteopenia) in our male population, as has been suggested by histomorphometric analysis in male cholestatic patients.11 Unfortunately, free testosterone levels were not available for most of our patients. As expected, age and menopausal status also correlated with increased pretransplant bone loss here, both well-recognized risk factors for osteoporosis and osteopenia in the general population.
Between OLT and 4 months posttransplant, the incidence of osteoporosis and osteopenia increased abruptly in all patient groups due an average of 5% bone loss over only 4 months, a very high rate of bone loss rarely seen in other clinical situations. Males and females were equally affected. The rate of bone loss did not change with time, despite changes in immunosuppressive regimens. The rate of bone loss was highly variable from patient to patient, but the vast majority (82%) of patients lost bone mass during the first 4 months. Since bone loss after transplantation probably occurs early in this 4-month period, perhaps by 1-2 months,43 it is possible that some of the 18% whose bone mass at 4 months was not lower than their pretransplant level had actually lost and then gained back bone mass by 4 months. Several demographic factors were identified in this study as risk factors for this early bone loss: younger age, PSC rather than PBC, no IBD disease, shorter duration of disease, current smoking history, and higher baseline BMD values. Surprisingly, the only posttransplant factors negatively associated with bone loss were bilirubin levels at 4 months and nonanastomotic strictures, and no correlations were seen with posttransplant glucocorticoids, serum levels of calcineurin inhibitors, hospitalization days, and rejection episodes. Our finding that liver recipients with higher BMD at OLT (who are also the younger patients with PSC) lose more bone after OLT than those with lower BMD is in agreement with previous studies,44, 45 but the reason for this is unknown. Nor is it obvious why PSC patients without IBD but the same baseline BMD or with shorter disease duration should lose more bone. Some of the patients with IBD were on glucocorticoids but had the same baseline BMD as those without glucocorticoid therapy; in these patients, glucocorticoids were already exerting an influence on bone mass, and it could be speculated that this may have reduced the effect of posttransplant glucocorticoid use. The fact that cholestasis, as reflected by high bilirubin levels at 4 months or nonanastomotic strictures, worsened bone loss emphasizes again that important effect of cholestasis on bone metabolism.
Although posttransplant glucocorticoids are assumed to be a major factor in early bone loss, data correlating glucocorticoids with posttransplant BMD are few.45–47 In this study, posttransplant bone loss did not correlate with either the average daily dose of glucocorticoids or episodes of rejection that required bolus Solu-Medrol therapy. Recent histomorphometric findings11 in cholestatic patients at time of OLT have shown a direct correlation between posttransplant bone volume losses and glucocorticoid doses. In addition, this histomorphometric study indicated that the main insult leading to early posttransplant bone loss, occurs very early after OLT, and is probably mainly related to further decreases in bone formation, a well-recognized effect of glucocorticoid therapy. All patients in this BMD study received high doses of glucocorticoids during and early after OLT, and this may have been sufficient to maximize their skeletal effect on bone metabolism and thus obscure any effect of dose tapering on preventing bone loss. Studies using glucocorticoid-free regimens will be needed to fully appreciate their effect on bone loss in the early posttransplant period. The contribution of calcineurin inhibitors to posttransplant bone loss has not been well established and may be overwhelmed by the profound effects of glucocorticoids on bone metabolism. Analysis of this study was not able to demonstrate any definite effect of serum levels of TAC or CyA on bone mass nor any difference between the 2 drugs. In this study, however, there were too few patients who received neither CyA nor TAC to allow any assessment of the effect of calcineurin inhibitor-free immunosuppression.
Data on the long-term effect of OLT on BMD are very limited but have suggested some improvement in BMD with time.48, 49 Our study with long-term follow-up of posttransplant BMD in a large cholestatic population indicated an increase in BMD during the first 2 years; after this, T scores remained stable throughout the 8 years of the study, whereas Z scores, which are age adjusted, continued to increase. These studies indicate that our cholestatic patients do not lose bone mass with age at the usual rate of 1-2% per year, presumably secondary to the concurrent ongoing beneficial effects of OLT on BMD recovery.
Although no early correlations after OLT were shown in our study, glucocorticoid doses would seem to be important for influencing the rate of bone gain. Bone gain after OLT was not significantly influenced by sex or disease, but it was reduced in postmenopausal patients and in patients with increased cholestasis or elevated serum creatinine. In addition, not only do patients with high baseline BMD lose more bone after OLT, but they also have less bone gain over the subsequent posttransplant years. Conversely, it would appear that patients with the most severe bone loss before OLT—that is, those with the most compromised bone metabolism—have the most to gain by the improved metabolic milieu after OLT. Although similar bone losses were seen after OLT throughout the study period, bone gain during the first 2 years after OLT is greater in the more recently transplanted patients (11% per year for patients after 1996 compared to 3.8% and 4.4% in the earlier 2 transplantation periods). During this latter period of 1996-2000, there was less rejection, less biliary stricturing, less prednisone dosing, and TAC rather than CyA as the primary immunosuppressant. In addition, the patients from the last transplantation period had lower posttransplant creatinine and phosphorus levels and higher PTH levels when compared to the earlier transplantation periods.
Despite no apparent connection between vitamin D and pretransplant osteopenia, vitamin D correlated with bone gain after OLT. Vitamin D is important in osteoblastogenesis,50 and previous studies have shown that it correlates with increased number of osteoblasts in cholestatic osteoporosis;51 this may explain the greater bone gain with higher levels of vitamin D in the posttransplant period. The importance of adequate mineral metabolism to bone gain was further illustrated by positive correlations with posttransplant PTH and negative correlations with posttransplant phosphorus.
In summary, most patients (77%) with advanced PBC and PSC have osteopenic bone disease and only 23% of patients have normal bone mass. After adjusting for differences in age and sex, a similar severity of osteopenia was found in patients with PBC and PSC. At time of OLT, risk factors for hepatic osteopenia are low BMI, older age, postmenopausal status, the presence of muscle wasting, high alkaline phosphatase, and low serum albumin. An improvement in bone mass occurred with time over the 15-year study period (in PSC, females and males), perhaps due at least in part to better nutritional status, increased BMI, increased vitamin D, and less cholestasis. After OLT, aggressive bone loss occurs during the first 4 months, and this did not change over time despite changes in immunosuppressive regimens with less posttransplant glucocorticoid doses, less rejection episodes and less nonanastomotic biliary stricturing. Risk factors for bone loss were younger age, PSC, higher pretransplant BMD, no IBD, shorter duration of disease, current smoking and ongoing cholestasis at 4 months. After the first 4 postoperative months, bone gain then occurs during the first 2 years and was increased in the more recently transplanted patients. Other factors favoring improvement in bone mass are lower baseline and/or 4-month BMD, premenopausal status for females, lesser glucocorticoids, no ongoing cholestasis, and higher levels of vitamin D and parathyroid function. Bone mass therefore improves most in patients with the lowest BMD who undergo successful transplantation and have normal hepatic allograft function and improved gonadal and nutritional status. In addition, patients with osteoporosis or osteopenia can be expected to gain bone mass for at least 8 years, despite getting older.