Ball AM, Gillen DL, Sherrard D, Weiss NS, Emerson SS, Seliger SL, Kestenbaum BR, Stehman-Breen C. Risk of Hip Fracture Among Dialysis and Renal Transplant Recipients. JAMA. 2002;288(23):3014-3018. doi:10.1001/jama.288.23.3014
Author Affiliations: Division of Nephrology, School of Medicine (Drs Ball, Seliger, and Kestenbaum), and Departments of Biostatistics (Mr Gillen and Dr Emerson) and Epidemiology (Drs Weiss and Stehman-Breen), School of Public Health and Community Medicine, University of Washington; Division of Nephrology, Seattle VA Puget Sound Health Care System (Drs Sherrard and Stehman-Breen); and Fred Hutchinson Cancer Research Center (Dr Weiss), Seattle, Wash.
Context Renal failure places people at particularly high risk of hip fracture.
However, the possible differential impact of dialysis and renal transplantation
on this risk is not well understood.
Objective To determine if patients who receive kidney transplants are at greater
risk of hip fracture compared with those who continue to undergo dialysis.
Design, Setting, and Participants Cohort study of 101 039 patients with end-stage renal disease placed
on the renal transplant waiting list in the United States between January
1, 1990, and December 31, 1999.
Main Outcome Measures Hip fractures, identified from Medicare claims data.
Results Among the patients included in this analysis, 971 hip fractures were
observed during the follow-up period of 314 767 person-years. The incidence
rate of hip fracture in patients receiving dialysis was 2.9 per 1000 patients
per year compared with 3.3 hip fractures per 1000 patients per year in those
who had previously received a renal transplant. Initially, the relative risk
(RR) of hip fracture associated with transplantation was 1.34-fold greater
when compared with dialysis (adjusted RR, 1.34; 95% confidence interval [CI],
1.12-1.61) but then decreased by 1% per month (adjusted RR, 0.99; 95% CI,
0.98-0.99) until the estimated risk became equal for dialysis and transplant
recipients approximately 630 days after transplantation (adjusted RR, 1.00;
95% CI, 0.87-1.15). Among transplant recipients, risk of fracture was relatively
higher in persons with a prolonged period of dialysis before transplantation.
Conclusion The high risk of hip fracture among dialysis patients is exceeded by
that among renal transplant patients during the first 1 to 3 years after transplantation.
Hip fractures are a common cause of morbidity and mortality.1,2 Relative to persons without kidney
disease, those with impaired renal function, including persons undergoing
dialysis or those who have received a renal transplant, have been observed
to be at increased risk of hip fracture.3- 6 Although
the results of one study7 suggest that the
risk among dialysis and transplant recipients is similar, the follow-up after
transplantation was relatively short. Furthermore, that study was not able
to separate the possible adverse effects of type of treatment modality from
patient characteristics associated with preferential receipt of a particular
treatment modality. Using data from the US Renal Data System (USRDS), we conducted
a cohort study of patients placed on the renal transplant waiting list to
determine if persons who have undergone renal transplantation have a different
risk of hip fracture relative to those who continue to undergo dialysis.
The USRDS database contains demographic and clinical information on
all US patients with end-stage renal disease (ESRD) who qualify for Medicare
and who have survived more than 90 days with renal replacement therapy. Briefly,
data from the USRDS are generated from various sources, including Medicare
billing records, United Network for Organ Services transplant records, and
ESRD medical evidence reports (Centers for Medicare and Medicaid Services
form 2728), network census reports, and death notification reports. The structure
of the USRDS is described in greater detail elsewhere.8
To obtain roughly comparable health status between patients receiving
a renal transplant and those undergoing dialysis, only patients listed on
the transplant waiting list were included in the dialysis group.9 Thus,
data were extracted for those patients who had not undergone transplantation
before January 1, 1990, and who were placed on the transplant waiting list
between January 1, 1990, and December 31, 1999. Patients older than 70 years
were excluded, since less than 1% of these patients had received a renal transplant.
After implementing these inclusion and exclusion criteria, data on 101 039
patients were available for the current analysis. Dialysis modality included
both peritoneal dialysis and hemodialysis, and transplant modality included
both kidney and kidney-pancreas transplantation.
Covariates that we identified a priori as potential risk factors for
hip fracture included age at time of placement on the transplant waiting list,
sex, race (African American, white, other), diabetes (as listed on the Centers
for Medicare and Medicaid Services medical evidence form), and length of time
undergoing dialysis before being placed on the transplant waiting list. Date
of transplantation was abstracted from the USRDS treatment history file, which
contains information regarding the history of modalities prescribed for each
patient in the database.
The number of hip fractures occurring during follow-up was determined
using information contained in the USRDS hospitalization file. Data contained
in the hospitalization file are derived from the Health Care Financing Administration's
standard analytical files, which contain data on inpatient hospital stays.
For the current analysis, hospitalization data were considered for January
1, 1990, through December 31, 1999. Hip fractures were identified using International Classification of Diseases, Ninth Revision codes
indicating cervical, intertrochanteric, or subtrochanteric hip fractures.
Because patients younger than 65 years who have maintained a successful renal
transplant for at least 3 years are no longer considered eligible for Medicare,
the USRDS hospitalization file does not identify fractures that occur in these
individuals. Thus, in our analyses, patients younger than 65 years were censored
if they had maintained a successful renal transplant for more than 3 years.
Hip fracture incidence rates were calculated as the observed number
of hip fractures per total patient time at risk, where patients were considered
at risk from the time of initial placement on the transplant waiting list
until the occurrence of a hip fracture, death, termination of Medicare coverage,
loss to follow-up, or December 31, 1999. In the event that no Medicare payments
for renal replacement therapy were recorded for 1 year, patients were considered
lost to follow-up from the date of the last known payment.
Survival analysis was used to model time from initial placement on the
waiting list until hip fracture. To account for changes in ESRD treatment
modality between the start and end of the study, treatment modality was modeled
as a time-dependent covariate. Transplantation was analyzed according to intention
to treat, ie, once patients underwent transplantation they were entered into
and remained in the transplant group regardless of the status of the graft.
Preliminary diagnostics revealed the presence of nonproportional hazards with
respect to the association between transplantation and the risk of hip fracture,
suggesting that the relative risk (RR) of hip fracture associated with transplant
varied with time. To model this association, a time-dependent covariate that
indicated the total time each patient had been a member of the transplant
group was also entered into the regression model. Adjustment covariates were
included in the multivariate analysis if they were considered a priori to
be potential confounders or independent risk factors for fracture. To further
investigate the possible burden of dialysis on the risk of fracture before
undergoing renal transplantation, a subanalysis was conducted within patients
who actually received a transplant before December 31, 1999. A Cox proportional
hazards model was used to model the relationship between time undergoing dialysis
before transplantation and the risk of hip fracture after adjustment for potential
confounders and independent predictors of fracture. For this subanalysis,
patients were considered at risk from the day of transplantation until the
time of fracture, death, termination of Medicare coverage, loss to follow-up,
or December 31, 1999. All statistical analyses were performed using SAS v8.2
(SAS Institute Inc, Cary, NC) and S-Plus v6.1 (Insightful Inc, Seattle, Wash).
Of the 101 039 patients on the renal transplant waiting list, 41 095
(40.7%) never received a transplant during follow-up, whereas 59 944
(59.3%) underwent transplantation at some point before fracture, death, termination
of Medicare coverage, loss to follow-up, or December 31, 1999. Patients who
received a renal transplant during follow-up tended to be younger, male, white,
and free of diabetes, and spent a shorter time undergoing dialysis before
being placed on the transplant waiting list (Table 1). Median time undergoing dialysis before placement on the
transplant waiting list was 196 days (range, 0 days-9.10 years), with 22 766
patients being listed before initiating dialysis. The median follow-up time
after being placed on the transplant waiting list was 2.98 years (range, 1
day-10 years). With respect to patients who received a transplant, median
time from being placed on the waiting list to the time of transplantation
was 200 days (range, 0 days-8.56 years), with 9349 patients undergoing transplantation
on the day of their first recorded dialysis session.
Hip fracture rates observed in the study are displayed in Table 2. A total of 971 patients experienced an incident hip fracture
during the total follow-up time of 314 767 person-years. Increasing age,
female sex, white race, diabetic nephropathy, increasing length of time undergoing
dialysis before being listed for transplantation, and receiving a renal transplant
were all associated with increased estimates of the incidence of hip fracture.
The observed incidence rate of hip fracture in patients who did not undergo
transplantation was 2.9 fractures per 1000 person-years compared with 3.3
fractures per 1000 person-years for patients who underwent transplantation.
Cox regression estimates are presented in Table 3. After adjustment for other factors in the table, age was
a strong risk factor for fracture. For each decade of life, the risk of hip
fracture was estimated to be 55% higher (95% confidence interval [CI], 1.47-1.64; P<.001). Patients aged 56 to 70 years experienced a
3.27-fold greater risk of hip fracture compared with patients younger than
40 years (95% CI, 2.77-3.87; P<.001). Women were
estimated to have a 64% greater risk of hip fracture compared with men (95%
CI, 1.44-1.86; P<.001). After adjustment for other
covariates of interest, black patients were estimated to have experienced
a 62% lower risk of hip fracture compared with white patients (95% CI, 0.32-0.45; P<.001). Patients with diabetic nephropathy were estimated
to have a nearly 3-fold greater risk of hip fracture compared with those without
diabetic nephropathy (adjusted RR, 2.96; 95% CI, 2.61-3.36; P<.001). For each month of dialysis before being placed on the transplant
waiting list, the risk of hip fracture was estimated to be 2% higher (95%
CI, 1.02-1.03; P<.001). Patients who underwent
dialysis for more than 12 months before placement on the transplant waiting
list were estimated to have a risk of hip fracture 91% higher than patients
undergoing dialysis treatment only 0 to 3 months before being listed (95%
CI, 1.60-2.26; P<.001).
After adjustment for age, sex, race, diabetic nephropathy, and time
undergoing dialysis before being placed on the transplant waiting list, patients
who underwent transplantation were estimated to have an initial 34% higher
risk of hip fracture compared with patients continuing with dialysis (95%
CI, 1.12-1.61; P = .002). However, this relative
risk was estimated to decrease 1% each month following transplantation (95%
CI, 0.98-0.99; P<.001). Figure 1 depicts the risk of hip fracture associated with transplantation
as a function of time. It was estimated that the risk of fracture among patients
who underwent transplantation became equal to that of dialysis patients approximately
630 days after transplantation, although pointwise CIs suggest that the risks
may become equal anywhere between the first and third years after transplantation.
The burden of dialysis before the time of transplantation on the risk
of hip fracture was further investigated among a subgroup of patients on the
transplant waiting list who eventually underwent transplantation during follow-up.
After adjustment for age, sex, race, and diabetic nephropathy, patients who
received dialysis treatment 3 to 12 months before transplantation were estimated
to have a 67% greater risk of hip fracture than patients whose duration of
dialysis had been fewer than 3 months (95% CI, 1.22-2.29). Patients on the
transplant waiting list who received dialysis for more than 12 months before
transplantation were estimated to have a 2.2-fold greater risk of hip fracture
than those who had undergone dialysis for fewer than 3 months (95% CI, 1.68-2.95).
We observed that, relative to patients continuing dialysis, renal transplantation
was associated with a 34% greater risk of hip fracture soon after transplant.
However, the magnitude of the increased risk waned with increasing time since
transplantation. It was estimated that after approximately 630 days the risk
of hip fracture was greater among patients who continued with dialysis. In
addition, among persons who underwent renal transplantation, the risk of hip
fracture was greater among those with longer length of time spent undergoing
dialysis before transplantation.
In contrast to our previous report,7 we
found a higher risk of hip fracture associated with renal transplantation
compared with dialysis during the first 630 days after transplantation. However,
our previous study may have underestimated the relative risk of hip fracture
because we included all ESRD patients, including less-healthy dialysis patients
who were not eligible for transplantation. Although we adjusted for age and
sources of comorbidities in our previous study, it is likely that we were
unable to fully account for differences in health status between dialysis
patients both eligible and ineligible for transplantation. The current study
included only those patients placed on the waiting list for renal transplantation,
resulting in a cohort of relatively more uniform health status. In addition,
our previous study used data from a smaller number of patients included in
a USRDS special study, resulting in less power to detect differences.
The particularly high RR of hip fracture that we observed soon after
renal transplantation is likely attributable to rapid bone loss during the
first 6 months after renal transplantation.10- 14 Potential
mechanisms that might explain this rapid bone loss include high doses of steroids
used during induction therapy and initially after transplantation, and postoperative
immobility followed by increased physical activity due to improved health
Our observation that the excess RR of hip fracture associated with renal
transplantation gradually diminishes during the first 1 to 3 years following
transplantation, after which the risk is less than that for persons who continue
with dialysis, is consistent with reports20 demonstrating
increases in bone density with time since renal transplantation. Resolution
of pretransplantation conditions deleterious to bone architecture and decreases
in immunosuppressive therapy are likely explanations for this improvement.
In addition, a variety of factors specific to dialysis could increase the
incidence of hip fracture among those patients with longer exposure to dialysis,
including low bone mineral density due to poor nutritional status, debilitation,
hypogonadism, multiple comorbidities, chronic acidosis, inactivity, heparin
use, pretransplantation use of steroids, aluminum intoxication, hyperparathyroidism,
and abnormal calcium metabolism.3,11,21- 25 This
hypothesis is supported by our data, which suggest that a longer time undergoing
dialysis before renal transplantation is associated with a greater risk of
hip fracture after transplantation. Therefore, the diminishing RR of hip fracture
over time associated with renal transplantation likely results from a combination
of decreasing risks in transplant patients and increasing risks in dialysis
This study has limitations. If patients were not hospitalized for their
fractures, we would have underestimated the incidence of hip fracture. However,
prior studies have shown that most hip fractures require hospitalization for
treatment.26 Also, given the observational
nature of the data, it is possible that there were unmeasured characteristics
associated with receipt of a particular type of treatment modality that may
have also been relevant to the risk of hip fracture. We tried to minimize
bias from this source by identifying only patients eligible for transplantation
and by adjusting for factors that were associated both with type of treatment
modality and the incidence of hip fracture.
The data suggest that, in the short term, recipients of kidney transplants
are at greater risk of hip fracture compared with those with renal failure
who continue with dialysis. However, after 1 to 3 years, the risk among transplant
recipients appears to be lower. Additionally, we found that among transplant
recipients, the longer the time undergoing dialysis before transplantation,
the greater the risk of hip fracture after transplantation.