| 2007 JASN IMPACT FACTOR 7.111 | HOME AUTHOR INFO EDITORIAL BOARD SUBSCRIBE FEEDBACK ALERTS HELP | |||
| CURRENT ISSUE | ARCHIVES | JASN Express | ONLINE SUBMISSION | |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Epidemiology and Outcomes |


* Chronic Disease Research Group, Minneapolis Medical Research Foundation; and
University of Minnesota, Minneapolis, Minnesota
Address correspondence to: Robert N. Foley, Chronic Disease Research Group, 914 South 8th Street, Suite D-253, Minneapolis, MN 55404. Phone: 612-347-5979; Fax: 612-347-5878; E-mail: rfoley{at}nephrology.org
| Abstract |
|---|
|
|
|---|
| Introduction |
|---|
|
|
|---|
| Materials and Methods |
|---|
|
|
|---|
Patients
We studied patients who had Medicare as primary payer and were receiving hemodialysis in the United States on January 1 of each of the years from 1992 to 2002. Demographic and baseline characteristics at initiation of dialysis, including age, gender, race, and primary renal diagnosis, were obtained from the Identification and Medical Evidence portions of the Renal Beneficiary Utilization System of the Centers for Medicare & Medicaid Services. Data on comorbid conditions were derived from Medicare Part A and Part B claims, between dialysis initiation and January 1 of that year; International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) and Physicians Current Procedural Terminology (CPT) codes were used for this purpose. Intravenous vitamin D use was determined from Medicare Part A outpatient files, using Healthcare Common Procedure Coding System codes J0635 and J2500 for calcitriol and paricalcitol, respectively. Parathyroidectomy occurrence was determined from Medicare Part A inpatient claims, using ICD-9-CM code 068. Data on PTH measurement were obtained from Medicare Part A outpatient and Part B physician files, using CPT code 83970.
Statistical Analyses
The 1992 to 1996 and 1997 to 2002 groups were compared using
2 analysis and multivariate logistic regression modeling. The follow-up for each point-prevalent group was from January 1 of that year to the earliest of the following in the same year: parathyroidectomy, death, or December 31. Cox regression was the primary analytical technique used, with time to parathyroidectomy as the outcome variable. The exploratory variables included the point-prevalent year, as well as age, gender, race, primary renal diagnosis, comorbid conditions, years of hemodialysis, use of intravenous vitamin D, previous parathyroidectomy, previous renal transplantation, and PTH measurement in the preceding year. We reasoned that differences in transplantation rates could confound annual parathyroidectomy rate comparisons and therefore chose to use two different analytical approaches. In model 1, transplantation was not used as a censoring event; in model 2, it was.
We used a case-control design, performed in two separate eras (1994 to 1995 and 1998 to 1999), to estimate the association between parathyroidectomy and mortality. Our findings were similar, regardless of whether censoring was performed at renal transplantation, and only findings with the latter method are presented here. In the 1994 to 1995 analysis, the study sample consisted of patients who were receiving hemodialysis on January 1, 1994, with no history of parathyroidectomy in the preceding 2 yr. Cases were defined as any patient who underwent parathyroidectomy between January 1, 1994, and December 31, 1995; control subjects were defined as those who did not undergo parathyroidectomy during the same period. Patients were characterized, and follow-up began on the day of parathyroidectomy for cases and January 1, 1995, for control subjects. Follow-up ended at death, parathyroidectomy in the follow-up period, or December 31, 2002. Poisson regression, in 3-mo intervals, was used to estimate the relative risk of death of cases, compared with control subjects, with adjustment for the following variables: age, gender, race, cause of ESRD, time elapsed since inception of renal replacement therapy, ESRD network, use of vitamin D in the preceding year, PTH measurement in the preceding year, comorbid conditions, and previous renal transplantation. The 1998 to 1999 analysis was identical to the 1994 to 1995 analysis, except that cases and control subjects were defined 4 yr later.
| Results |
|---|
|
|
|---|
|
|
|
|
| Discussion |
|---|
|
|
|---|
Parathyroidectomy rates are often used as a barometer of management success or failure in large populations of dialysis patients. One national-level study found that the proportion of prevalent ESRD patients who underwent parathyroidectomy in the United States declined significantly from 1988 to 1998 (4). Another recent national-level study found an overall decline in parathyroidectomy rates between 1990 and 1999; as found in our study, rates were lowest in 1998 (5). Although our study showed similar trends, it differed in several ways from other studies: We used 2000, 2001, and 2002 as prevalent years, excluded peritoneal dialysis patients, included patients with previous parathyroidectomy and renal transplantation, assessed a wide range of comorbidity, and restricted follow-up to a maximum of 1 yr in each prevalent cohort. Our approach to follow-up was intended to equalize the potential duration of follow-up in each annual cohort, given our belief that parathyroidectomy rates would likely increase with longer follow-up.
Outside the United States, the Okinawa Dialysis Study reported that parathyroidectomy rates were lower in the 1980s than in the 1970s (6). The European Dialysis and Transplantation Association Registry reported several years ago that the proportion of patients with a history of parathyroidectomy increased between 1982 and 1988; incidence rates of new parathyroidectomy were similar throughout this period (7). Another large-scale study was that of Malberti et al. (8), who reported findings on 14,180 patients who received renal replacement therapy between 1983 and 1996 in Lombardy. Incidence rates were relatively constant over time, unlike those in our study. Higher parathyroidectomy rates were associated with female gender, longer dialysis vintage, and use of peritoneal dialysis; younger age, diabetic nephropathy, and renal transplantation were associated with lower rates. The underlying basis for these associations, which mirror closely those found in our study, remains speculative.
It is widely known that primary hyperparathyroidism occurs more frequently in women (9), and it is possible that women with uremia are more susceptible to secondary hyperparathyroidism and/or renal bone disease. It has been reported by others that diabetic ESRD patients may have lower PTH levels than their nondiabetic counterparts and may be more susceptible to low-turnover bone disease (10,11). High-dose parenteral vitamin D is often attempted as a "final" nonsurgical strategy when parathyroidectomy is being considered (12). The association between the use of parenteral vitamin D and subsequent parathyroidectomy may reflect hyperparathyroidism that was more advanced at the beginning of the outcome assessment period. We are unsure why parathyroidectomy rates decreased so dramatically with advancing age; patient selection factors and death as a competing risk may have been involved.
Parathyroidectomy rates were heterogeneous across large geographic areas. To our knowledge, few if any studies have examined this issue. A PubMed search on October 6, 2004, using the terms parathyroidectomy, geographic, and dialysis yielded no citations. The causes of such heterogeneity could not be determined from this observational study. Differences in prevalent case mix could account for some of the heterogeneity, as could approaches to management of bone metabolism. It was interesting that the lowest parathyroidectomy rates were seen in a region of relatively high population density (the southern counties of California), whereas the highest rates were seen in a region of relatively low population density (Alaska, Idaho, Montana, Oregon, and Washington). It is tempting to speculate that access to specialist nephrology care may partly account for some of the heterogeneity in parathyroidectomy rates.
We found biphasic associations between parathyroidectomy and mortality. Compared with patients who did not undergo the procedure, higher mortality rates were seen immediately after surgery, followed, in the longer term, by lower mortality rates. Mortality patterns were similar in the 1994 to 1995 and the 1998 to 1999 case-control analyses. It is difficult to unravel the relative contributions of unmeasured comorbid conditions and patient selection to these findings. One of many possible interpretations, however, includes the following hypothesis: Eradication of refractory hyperparathyroidism improves cardiovascular risk (as has been speculated for decades) but at a short-term cost (the risk associated with parathyroid surgery).
Typical textbook indications for parathyroid surgery include "the unequivocal evidence of secondary HPT [hyperparathyroidism] (very high levels of serum PTH and/or the presence of osteitis fibrosa on bone biopsy), together with any of the following: (1) persistent hypercalcemia not attributable to other causes; (2) persistent hyperphosphatemia despite proper use of phosphate binders; (3) serum Ca x P [calcium times phosphorus] product that consistently exceeds 70 to 75 mg2/dl2; (4) progressive skeletal and articular pain, fractures, or deformities as a result of secondary HPT; and (5) calciphylaxis" (13). For a given severity of hyperparathyroidism, the decision to recommend parathyroidectomy often involves a careful balance of several factors. These include the likelihood of long-term survival on dialysis and transplantation, anesthetic risk, and the availability of appropriate surgical and radiologic services. The management of renal bone disease changed dramatically during the years included in this study. For example, assays that measure the intact PTH molecule became the norm (14). Several active forms of vitamin D sterol became available, including some with a lower tendency to raise serum calcium levels for a given decrease in PTH levels; aluminum use became uncommon, largely supplanted by calcium-based phosphate binders and non-calcium-containing phosphate binders, such as sevelamer (1520). It is likely that the declining rates of parathyroidectomy between 1992 and 1998 partly reflect these advances. The relatively sudden rise in parathyroidectomy rates after 1998 was unexpected.
Several obvious questions present themselves, not least of which is whether the trends observed are artifactual, a consequence of the methods used. Inclusion in this study was defined on the basis of dialysis use on January 1 of each calendar year, and follow-up was restricted to the ensuing 12 mo. We chose this model, as opposed to an incident model (in which all newcomers to hemodialysis in a calendar year are studied), because parathyroidectomy rates were expected to be very low initially. We attempted to characterize the patients extensively in each year, but it is possible that unmeasured factors could have changed from year to year. The study was based on administrative claims, and calcium, phosphorus, and PTH levels were not available. It thus is impossible to disprove the hypothesis that the trends observed in this study reflect changes in the threshold levels of hyperparathyroidism that triggered referral for surgery. The decision to recommend parathyroidectomy is heavily dependent on PTH levels; it is possible that changes in the assays used may have led to changes in PTH levels, leading to parathyroidectomy, without any real biologic changes in successive patient cohorts. Unfortunately, this hypothesis is not one that we can address directly from the data available to us. In practice, a single code is available for "parathormone," irrespective of the type of assay used. PTH was measured more frequently in later cohorts. It is conceivable, therefore, that the overall burden of diagnosed and undiagnosed severe hyperparathyroidism may have remained constant over time; in this scenario, the rise in parathyroidectomy rates in later years might reflect a rising proportion of diagnosed disease, offset by a declining proportion of undiagnosed disease. This being said, the same analytical methods were used for cohorts before and after 1998, and it is difficult to believe that methodologic flaws fully account for the biphasic trend in parathyroidectomy rates before and after 1998. It is also possible that heterogeneity of therapeutic approach accounts for some of our findings. If the recently introduced Kidney Disease Outcomes Quality Initiative guidelines (21) are embraced, then it is possible that harmonization of treatment practices could lead to lower parathyroidectomy rates. Finally, with large data sets, differences that are trivial from a clinical perspective can be associated with a high degree of statistical significance.
The trends seen in this study may reflect changing enthusiasm for the aggressive use of calcium-augmenting therapies between 1992 and 2002. This enthusiasm may have been tempered by the risk of low-turnover bone disease and the realization that extraskeletal calcium deposition may place patients at risk (22,23). If one argues that the risk of untreatable or autonomous hyperparathyroidism increases with higher PTH levels and longer exposure periods, then higher parathyroidectomy rates might be expected with less aggressive treatment practices. This speculation cannot be addressed directly using the findings of this study. Similarly, other possible explanations, such as the use of different PTH assays and the emergence of newer vitamin D sterols and phosphate binders, cannot be addressed. Large observational studies are relatively good for discerning trends but relatively poor for elucidating causal mechanisms. We believe, however, that the study suggests associations of clinical importance and establishes a reasonable foundation for monitoring future trends and initiating hypothesis-driven prospective studies to better understand observed trends. Despite its flaws, this study suggests a disturbing pattern: An unexplained recent rise in parathyroidectomy rates at a time when the therapeutic armamentarium for preventing severe hyperparathyroidism was expanding considerably.
| Acknowledgments |
|---|
We thank Dana D. Knopic and James Kaufmann, PhD, for assistance with manuscript preparation and editing, respectively.
| Footnotes |
|---|
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
G. A. Block, S. Zeig, J. Sugihara, G. M. Chertow, E. M. Chi, S. A. Turner, and D. A. Bushinsky Reply Nephrol. Dial. Transplant., October 1, 2008; 23(10): 3371 - 3372. [Full Text] [PDF] |
||||
![]() |
S. Mazzaferro, M. Pasquali, A. Farcomeni, A. R. Vestri, A. Filippini, A. M. Romani, G. Barresi, and F. Pugliese Parathyroidectomy as a therapeutic tool for targeting the recommended NKF-K/DOQITM ranges for serum calcium, phosphate and parathyroid hormone in dialysis patients Nephrol. Dial. Transplant., July 1, 2008; 23(7): 2319 - 2323. [Abstract] [Full Text] [PDF] |
||||
![]() |
C. Lomonte, L. Vernaglione, D. Chimienti, A. Bruno, S. Cocola, A. Teutonico, F. Cazzato, and C. Basile Does Vitamin D Receptor and Calcium Receptor Activation Therapy Play a Role in the Histopathologic Alterations of Parathyroid Glands in Refractory Uremic Hyperparathyroidism? Clin. J. Am. Soc. Nephrol., May 1, 2008; 3(3): 794 - 799. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. Cunningham, J. Floege, G. London, M. Rodriguez, and C. M. Shanahan Clinical Outcomes in Secondary Hyperparathyroidism and the Potential Role of Calcimimetics NDT Plus, January 1, 2008; 1(suppl_1): i29 - i35. [Abstract] [Full Text] [PDF] |
||||
![]() |
T. Drueke, D. Martin, and M. Rodriguez Can calcimimetics inhibit parathyroid hyperplasia? Evidence from preclinical studies Nephrol. Dial. Transplant., July 1, 2007; 22(7): 1828 - 1839. [Full Text] [PDF] |
||||
![]() |
M. E. Rodriguez, Y. Almaden, S. Canadillas, A. Canalejo, E. Siendones, I. Lopez, E. Aguilera-Tejero, D. Martin, and M. Rodriguez The calcimimetic R-568 increases vitamin D receptor expression in rat parathyroid glands Am J Physiol Renal Physiol, May 1, 2007; 292(5): F1390 - F1395. [Abstract] [Full Text] [PDF] |
||||
![]() |
G. Rashid, J. Bernheim, J. Green, and S. Benchetrit Cardiovascular Events and Parathyroid Hormone--Suggestion of a Further Link J. Am. Soc. Nephrol., April 1, 2007; 18(4): 1023 - 1026. [Abstract] [Full Text] [PDF] |
||||
![]() |
T. Cantor, Z. Yang, N. Caraiani, and E. Ilamathi Lack of Comparability of Intact Parathyroid Hormone Measurements among Commercial Assays for End-Stage Renal Disease Patients: Implication for Treatment Decisions Clin. Chem., September 1, 2006; 52(9): 1771 - 1776. [Abstract] [Full Text] [PDF] |
||||
![]() |
N. Velasco, M. S. MacGregor, A. Innes, and I. G. MacKay Successful treatment of calciphylaxis with cinacalcet--an alternative to parathyroidectomy? Nephrol. Dial. Transplant., July 1, 2006; 21(7): 1999 - 2004. [Full Text] [PDF] |
||||
![]() |
C. Basile, C. Lomonte, L. Vernaglione, F. Casucci, D. Chimienti, A. Bruno, S. Cocola, E. A. Verrelli, and F. Cazzato A high body mass index and female gender are associated with an increased risk of nodular hyperplasia of parathyroid glands in chronic uraemia Nephrol. Dial. Transplant., April 1, 2006; 21(4): 968 - 974. [Abstract] [Full Text] [PDF] |
||||
![]() |
C. P. Schmitt, T. Odenwald, and E. Ritz Calcium, Calcium Regulatory Hormones, and Calcimimetics: Impact on Cardiovascular Mortality J. Am. Soc. Nephrol., April 1, 2006; 17(4_suppl_2): S78 - S80. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
HOME
CURRENT ISSUE
ARCHIVES
JASN Express
ONLINE SUBMISSION
AUTHOR INFO
EDITORIAL BOARD SUBSCRIBE FEEDBACK ALERTS HELP |