The rate, cost and outcomes of parathyroidectomy in the united states dialysis population from 2016–2018

Patients with end-stage kidney disease (ESKD) often have secondary hyperparathyroidism (SHPT) which is characterized by elevated parathyroid hormone (PTH) in response to kidney failure. SHPT is commonly associated with mineral and bone disorder (MBD) which includes associated effects on serum levels of calcium (Ca) and phosphate (P) [1].Typically, SHPT and MBD are managed through the use of vitamin D, phosphate binders, and/or calcimimetics [2]. However, patients who cannot be managed with currently available options may undergo a parathyroidectomy to reduce PTH and facilitate the management of Ca and P [1, 3].

Parathyroidectomy is relatively uncommon. The rate of parathyroidectomy was estimated at 5.4 per 1,000 person-years using United States (US) inpatient data from 2002 through 2011 and was relatively stable through the latter part of the time period [4]. Rates were highest in younger patients (age 20–44 years) and were higher in women than men, and inpatient mortality decreased from 1.7 to 0.8% from 2002 to 2011. A separate study using 2013 national readmission data showed that approximately 17.2% of patients with SHPT had an unplanned hospital re-admission within 30 days of a parathyroidectomy [5]. National parathyroidectomy rates or readmission rates for the US have not been published since these investigations. A study of 181 patients who underwent parathyroidectomy between January 1, 2008 and December 31, 2010 showed that the monthly physician charges increased from $1,531 in the 6 months before parathyroidectomy to $1,931 in the 6 months afterward [6]. The authors also showed that physician encounters increased after parathyroidectomy, and that 31% of patients had a diagnosis code for hypocalcemia after parathyroidectomy. However, the study was limited by its small sample size and lack of inpatient data.

Therefore, the goals of this study were to provide more current estimates of the parathyroidectomy rate in adult, US hemodialysis (HD) and peritoneal dialysis (PD) patients, and to quantify changes in costs as well as key outcome measures after parathyroidectomy, including subsequent parathyroidectomy, mortality, surgical side effects, dialysis visits, office visits, hospitalizations, and selected laboratory values.

Please see Supplementary Materials for more details on the methods used in this study.

This study presents the most current parathyroidectomy rates to date and is the first to provide detailed cost, utilization and laboratory information for the year before and after parathyroidectomy. Parathyroidectomy remains a relatively uncommon event for dialysis patients. The rate of parathyroidectomy per 1,000 person years declined from 6.5 to 5.3 between 2016 and 2018; however, it is difficult to extrapolate long-term trends from 3 years of data. These rates are slightly higher than the overall rate of 5.4 per 1,000 estimated by Kim et al. using annual inpatient data from 2002 to 2011 [4]. Kim et al. used inpatient data that included transplant patients (who we excluded) and could not exclude multiple parathyroidectomy hospitalizations from the same person. In contrast, we required Medicare Part A, B, and D coverage and Kim et al. estimated rates across all payers. Despite these differences, our finding that the parathyroidectomy rate is higher in younger patients is consistent with Kim, et al., and we extend these findings to show that parathyroidectomy is extremely rare in the oldest patients. We also show that there is a 1.7-fold higher parathyroidectomy rate in PD patients compared to HD patients after accounting for age, race, and dialysis vintage.

Patients undergoing parathyroidectomy cost the healthcare system an incremental $25,000, the majority of which is from inpatient facility payments for the parathyroidectomy itself but which also includes the higher rate of hospitalization and hospitalization days during the 12 months afterward. The incremental cost of parathyroidectomy was even higher in PD patients ($39,000). In terms of other consequences of parathyroidectomy, both subsequent parathyroidectomy and all-cause mortality were at their highest levels during the 30-day period after the parathyroidectomy. The higher mortality after surgery is consistent with the findings from Kestenbaum, et al., although the absolute rate in that study (386 per 1,000 person-years) is more than twice the rate in our study [18].  Since the Kestenbaum study was based on patients undergoing parathyroidectomy prior to 2000, the documented decline in the inpatient mortality rate from 2002 to 2011 [4] suggests that the difference in rates between our study and that by Kestenbaum et al. are related to improvements in post-operative survival for patients over time.

Belozeroff et al. showed that hypocalcemia was the most common surgical complication, followed by bleeding, and that wound infection and seroma were rare [6]. Our study provides more precise estimates of these events, correcting for background rates that might lead to an overestimate in a dialysis population. Also, we used serum corrected calcium instead of diagnosis codes to identify hypocalcemia and showed that hypocalcemia is almost twice as common than shown by Belozeroff et al. (31%), with 58% of patients having cCa < 8.5 mg/dL and 31% having cCa < 7.5 mg/dL in month 2 after parathyroidectomy. These results are consistent with the overall trends in cCa and P which were stable for the year before parathyroidectomy, declined after surgery, and then rose gradually over the course of the following year to a mean level that was still lower than the pre-surgery values. These findings suggest that clinicians should be aware of the signs and symptoms of hypocalcemia in these patients, some of whom may have hypocalcemia for a year after surgery.

We also showed that inpatient hospitalizations increased by 1.4 per person per year and hospital days increased by 12.1 days per person per year. Given that outpatient HD occurs on 3 out of 7 days (43%), one would expect exactly 5.2 fewer dialysis visits with 12.1 more days in the hospital. Hence, the increase in hospitalizations likely explains the lower rate of dialysis visits.

Patients who undergo parathyroidectomy are a select subgroup of patients with SHPT. At the time of parathyroidectomy in our 2016–2017 parathyroidectomy cohort, 47% of patients were receiving phosphate binders, 47% were receiving IV vitamin D, and 43% were receiving calcimimetics. In comparison, in the overall ESKD population in quarter 1 of 2017, 63% of patients were receiving phosphate binders, 42% were receiving IV vitamin D, and 30% were receiving calcimimetics [7]. Hence, the differences in the proportions were such that the parathyroidectomy population had 13% more patients receiving calcimimetics, a 16% fewer patients receiving phosphate binders, and 5% more patients receiving IV vitamin D. This suggests that patients undergoing parathyroidectomy may be more likely to be using calcimimetics and IV vitamin D, and less likely to be using phosphate binders. However, given that phosphate binder use may be under-reported compared to other sources, these results should be interpreted with caution [19].

This study is strengthened by its broad coverage of US dialysis patients, the availability of detailed laboratory information, and the availability of comprehensive Medicare claims data that include direct health care expenditures across all settings of care (inpatient, outpatient, pharmacy, and home care). As with any study using observational data, there are limitations and other factors to be considered when interpreting these findings. Our focus was on HD and PD patients who had Medicare as their primary payer. Our rates were estimated for 2016–2018, and our parathyroidectomy cohort was limited to patients who had a parathyroidectomy in 2016–2017. When estimating the annual parathyroidectomy rates, it is likely that some procedures were not the “first-ever” parathyroidectomy for some patients. However, because we excluded patients who had a parathyroidectomy within the 3 months prior to each year, we minimized the potential for double-counting parathyroidectomy episodes. Also, we were not able to evaluate whether changes in the parathyroidectomy rate were associated with changes in vitamin D or calcimimetics use because oral medication data were only available in patients with Medicare Part D coverage. Finally, we note that these results apply to the US and that in other regions of the world parathyroidectomy utilization and practice patterns may be different.

The cohorts created in this study only included patients with Medicare fee-for-service coverage. However, since commercially insured patients initiating dialysis are covered by their primary private insurer during a 36-month coordination of care period and are not included in these data, these results may not apply to the full dialysis population. As a result, in both the parathyroidectomy rate and cost estimation analyses, there is a risk that younger patients in our study may be different from the total under-65 years of age population initiating dialysis. Also, because different kinds of costs including medications are reimbursed as part of the bundled dialysis payment, we could not separate all medication costs from medical costs. Similarly, costs by type may vary depending on the specific details of the bundled payments for a particular year. As a result, cost increases for medications used to manage patients after parathyroidectomy (e.g., for hungry bone syndrome) cannot be enumerated in this study. Also, to the extent that parathyroidectomy might reduce the risk and associated cost of long-term outcomes beyond one year, these analyses could not capture the associated cost reductions. Finally, we did not have parathyroid hormone levels in the data, which would have been useful for characterizing the short-term and long-term effectiveness of parathyroidectomy.

In conclusion, parathyroidectomy is a relatively uncommon event that continues to occur at a rate of about 5.3 to 6.5 per 1,000 person-years. The incremental 12-month cost of parathyroidectomy is approximately $25,000, and the most common complication of surgery is the risk of hypocalcemia which occurs in more than half of patients. Both mean cCa and P decline after surgery and, although they rise during the following year, they remain lower than the presurgical values. Because of the implications of parathyroidectomy for clinicians, payers and patients, the USRDS should consider reporting parathyroidectomy-related metrics in its annual report. This would facilitate clear communication of the clinical and economic outcomes associated with parathyroidectomy.