Autosomal dominant polycystic kidney disease (ADPKD) is

Transcription

1 CJASN epress. Published on June 10, 2010 as doi: /CJN Renal Funtion and Healthare Costs in Patients with Polyysti Kidney Disease Krista L. Lentine,* Huiling Xiao,* Gerardo Mahniki, Adrian Gheorghian,* and Mark A. Shnitzler* *Center for Outomes Researh and Division of Nephrology, Saint Louis University Shool of Mediine, St. Louis, Missouri; and Global Health Eonomis and Outomes Researh, Novartis Argentina SA, Buenos Aires, Argentina Bakground and objetives: Charaterizing relationships of kidney funtion to healthare osts in polyysti kidney disease has appliations for eonomi evaluations of standard and emerging therapies. Design, setting, partiipants, & measurements: The administrative reords (2003 to 2006) of a private health insurer were examined to identify polyysti kidney disease patients (n 1913) from ICD9 diagnosis odes on billing laims. The first available diagnosti laim was assumed as an index date, and baseline estimated GFR () was omputed using losest serum reatinine value. The assoiations of with annualized harges were modeled by nonlinear and linear regression. Results: Medial, pharmay, and total healthare osts varied signifiantly by baseline kidney funtion, suh that mean total annualized harges (unadjusted) were approximately 5-fold higher in patients with < 15 ml/min ompared with those with > 90 ml/min. After adjustment for age and gender, total harges did not differ signifiantly among patients with > 30 ml/min, and but rose preipitously with < 30 ml/min. Eah ml/min deline <30 ml/min predited approximately $5435 higher adjusted annual harges. Results were similar after adjustment for baseline diabetes and ardiovasular disease as identified in laims, while signifiantly higher adjusted harges were deteted with 31 to 60 ml/min versus >90 ml/min in a subgroup free of diabetes and ardiovasular disease. Conlusions: Healthare harges are assoiated with advaned renal dysfuntion in polyysti kidney disease patients. Strategies that prevent loss of renal funtion below 30 ml/min have the potential to generate substantial redutions in medial harges. Clin J Am So Nephrol 5:, doi: /CJN Autosomal dominant polyysti kidney disease (ADPKD) is the most ommon inherited ysti renal disease as well as the most ommon single-gene disorder of the kidneys. ADPKD affets approximately 600,000 persons in the United States and 10 million persons worldwide (1). The natural history of ADPKD is one of inevitable pathologi and linial progression over time, haraterized by growth in the number and size of ysts with seondary destrution of funtional parenhyma. A reent prospetive study of strutural and funtional progression among 232 ADPKD patients with reatinine learane 70 ml/min at enrollment doumented average inreases in total kidney volume of ml over 3 years, and derements in GFR of ml/min per year in partiipants with baseline kidney volume 1500 ml (2). Inidene estimates for ESRD among ADPKD patients have ranged from 45% to 70% by age 65 years (3 6). As a manifestation of its progressive ourse, ADPKD ranks as the fourth most ommon ause of ESRD among adults in the United States (7). At present, no medial therapy has proven to substantially slow Reeived January 25, Aepted April 27, Published online ahead of print. Publiation date available at Correspondene: Dr. Krista L. Lentine, Saint Louis University Center for Outomes Researh, Salus Center 2nd Floor, 3545 Lafayette Avenue, Saint Louis, MO Phone: ; Fax: ; lentinek@slu.edu the growth of ADPKD-related ysts or the volumetri expansion of polyysti kidneys. However, innovative strategies to blunt renal tubular ell proliferation and fluid seretion, the proesses that drive growth of ysti bodies, omprise foi of ative researh (8,9). Considered broadly without respet to underlying ause, hroni kidney disease (CKD) is reognized as a soure of substantial morbidity and eonomi burdens. Epidemiologi studies have demonstrated assoiations between kidney dysfuntion and inreased risks of adverse outomes inluding ardiovasular disease events, hospitalizations, resoure utilization, and mortality (10,11). For example, based on data from a large Healthare Maintenane Organization (HMO), Smith et al. found that patients with CKD stages 2 to 4 had 1.9 to 2.5 times more presriptions, 1.3 to 1.9 times more outpatient visits, and 1.8 to 4.2 times more inpatient stays than age- and gendermathed ontrols (11). Notably, however, osts were similar aross these funtional stages within the CKD sample. Charaterizing relationships of kidney funtion to healthare osts among patients with CKD due to polyysti kidney disease has pratial appliations for eonomi evaluations of standard and emerging therapies in this disease. While ADPKD onfers important publi health impats based on its progressive nature within affeted patients and relative frequeny among heritable disorders, the disease is less ommon than leading auses of Copyright 2010 by the Amerian Soiety of Nephrology ISSN: /

2 2 Clinial Journal of the Amerian Soiety of Nephrology Clin J Am So Nephrol 5:, 2010 CKD suh as diabetes, whih by omparison affets approximately 7.2% of the adult population in the United States (15 million Amerians) (12). The hallenge of assembling large ohorts for outomes analyses in ADPKD an be addressed by study designs that integrate data aross multiple enters or providers. To advane understanding of the assoiation of kidney funtion with healthare osts in polyysti kidney disease, we performed a retrospetive study of the administrative reords of a large health insurane provider. Administrative insurane data ompile dated reords of physiian enounters, hospitalizations, proedures, and presriptions that may be examined to lassify groups with diagnoses of interest and to study assoiated resoure utilization and healthare osts. Using linked administrative billing and laboratory data, we sought to identify a ohort of polyysti kidney disease patients reeiving are in ontemporary pratie and to haraterize the relationship of kidney funtion with medial and pharmaeutial harges to the insurer aross a broad spetrum of renal funtion s. Materials and Methods Data Soure Study data were drawn from the administrative reords of a large national health insurane provider. The insurane database ontains all medial as well as outpatient pharmay billing laims submitted for benefiiaries during periods of eligibility. Individual laim reords ontain the date in whih a given laim was paid, date in whih servies were performed, diagnosti and proedure odes, plae of servie odes, revenue odes, and amount harged. The total population in the database numbers 6 million individuals from aross the United States. For the purpose of this analysis, billing laims data were augmented with extrats from outpatient labs for a subsample of benefiiaries, permitting ombined examinations of lab results along with expenditures for healthare servies and presription mediations. Data were available for laims and labs submitted in January 2003 to Deember Final samples were limited to patients with at least 6 months of insurane eligibility after the diagnosis or apture date of interest. All study partiipants were simultaneously enrolled in medial and pharmay benefits with this ompany exlusively during the study window. This study was approved by the Institutional Review Board of Saint Louis University. Identifiation of Patients with Polyysti Kidney Disease Patients with ADPKD were identified by billing laims with a orresponding International Classifiation of Diseases 9th Edition (ICD-9) ode (753.13, polyysti kidney disease, autosomal dominant ) within the available laims data. As ysts may be reognized in ADPKD patients before the diagnosis is formalized, we also onsidered laims for polyysti kidney, unspeified type (753.12). In ategorizing patients as versus , a mutually exlusive diagnosti hierarhy was used suh that individuals with any appearane of odes during observation were onsidered to have ADPKD, whereas patients without laims but any laims were onsidered to have polyysti kidney disease/unspeified. We defined the first diagnosis laim in the data set as the analysis index date. Classifiation of Renal Funtion and Baseline Charateristis Age and gender of the study samples were asertained from insurane eligibility files. Renal funtion was omputed as estimated GFR () from losest serum reatinine values reported in the laboratory results database within 365 days of index polyysti kidney disease laim. was omputed aording to the abbreviated Modifiation of Diet in Renal Disease (amdrd) equation (13) without raial adjustment as: (ml/min per 1.73 m 2 ) 186 (serum reatinine) (Age) (0.742 if female). As rae information is not traked in the administrative data, we assumed all subjets to be nonblak in aord with best pratie in laboratory mediine reommendations (14) and the approah of previous investigations with administrative data (15,16). The omputed was assumed to apply to the individual at the time of the index laim. Level of kidney funtion was lassified aording to the GFR-based shema of the National Kidney Foundation, Kidney Disease Outomes Quality Initiative (NKF/KDOQI) (17) as: 90, 60 to 89, 30 to 59, 15 to 29, or 15 ml/min. To restrit the sample to patients with pre-esrd polyysti diseases, we exluded individuals with billing laims before the index date for dialysis, transplantation, and/or assoiated ompliations based on ICD-9 diagnosis odes (V45.11, V45.12, 585.6, V56.0, V56.31, V56.32, V56.8,, E870.2, E871.2, E872.2, E874.2, E879.1, , , , , ), ICD-9 proedure odes (39.95, 54.98, 55.69, 00.91, 00.92, 00.93), Common Proedural Terminology/Healthare Common Proedure Coding System odes ( , 90945, 90947, to 90970, 90999, 99512, 99559, G0308 to G0327, 0505F, 0507F, 4052F, 4053F, 4054F, 4055F, 50360, 50365), or the Diagnosis Related Group for kidney transplantation (302). The presene of omorbid diabetes mellitus and of ardiovasular disease at baseline sampling were defined by billing laims with orresponding ICD-9 diagnosis odes within 30 days of the index date, using oding algorithms similar to previous reports (18 22) as: Diabetes mellitus (250, 357.2, 362.0, , 648.0); ardiovasular disease omposite inluding oronary artery disease (414), myoardial infartion (410.x), angina petoris (411,413), heart failure (398.91, 422, 425, 428, 402.x1, 404.x1, 404.x3, V42.1), or stroke/transient ishemi attak (433 to 435). Asertainment of Healthare Charges Costs for all medial servies and outpatient pharmaeutial presriptions were derived from harges reported on laims paid by the insurer. Monetary figures were adjusted to the pries in the year 2008 Medial Care Component of the Consumer Prie Index (23). Annualized harges were omputed among patients with at least 6 months of insurane eligibility after the index date to avoid magnifiation of outlier effets in patients aptured with short eligibility windows. Annualization was performed by aggregating total harges per patient over 6 months after index date and multiplying by 2 to represent estimated osts for 1 year of are. Medial harges from inpatient soures were also examined separately. The harge ontributions of renal replaement servies were examined among those who required sustained dialysis (defined as at least two dialysis laims over 30 days) or kidney transplant during observation. Dialysis harges were annualized among dialysis-dependent patients, whereas transplant harges were defined as those assoiated with the transplant hospitalization. Statistial Analyses Data sets were merged and analyzed with SAS for Windows software, version 9.2 (SAS Institute In., Cary, NC). Distributions of age, gender, and baseline diabetes and ardiovasular disease diagnoses in patients with speified ADPKD/ versus unspeified/ polyysti kidney disease, as well as trait distributions by within a given group, were ompared by the two-sample t test (or ANOVA)

3 Clin J Am So Nephrol 5:, 2010 Renal Funtion and Healthare Costs 3 for age and the Chi-square test for ategorial fators. The Kruskal- Wallis test, a nonparametri equivalent to one-way ANOVA, was used to ompare insurane harges in patients with speified ADPKD versus unspeified polyysti kidney disease, as well as to ompare unadjusted harges aross s. Pairwise omparisons of unadjusted harges in a given versus 90 ml/min were performed by the Wiloxon test, with statistial signifiane for pairwise omparisons defined as P 0.01 based on a Bonferroni orretion for multiple omparisons. To assess the shape of the relationship of with annualized healthare harges, a general linear model inorporating splines was onstruted using a mid-range (60 ml/min) as a referene and inluding form break points (knots) at 15, 30, 45, and 90 ml/min. The resulting shape informed onstrution of a linear regression model inluding adjustment for gender and age. Models were onstruted for total harges as well as omponents of total medial (exluding outpatient pharmay), inpatient medial, and pharmay harges. Additional linear regressions on harges were fit adding omorbidities of diabetes, ardiovasular disease and their interation, as well as in a sample free of baseline diabetes and ardiovasular disease. Results Sample Charateristis Among 2,041,973 patients in the administrative database with at least one reported serum reatinine result within their period of observed insurane benefits eligibility and at least 6 months of benefits, 1913 (0.09%) had billing laims with diagnosis odes for polyysti kidney disease (496 with ; 1417 with alone). Among patients ategorized as ADPKD/ , 245 (49.4%) also had laims with unspeified/ polyysti kidney disease diagnosis odes in the available laims data. There were no signifiant differenes in the distributions of age, gender, or baseline diabetes or ardiovasular disease in those with diagnosis odes for ADPKD, speifially ompared with unspeified polyysti kidney disease (Table 1). The distributions in patients with diagnoses ompared with alone also did not signifiantly differ (P 0.31): 90 ml/min: 13.7% versus 16.0%; 60 to 89 ml/min: 33.7% versus 37.3%; 30 to 59 ml/min: 31.9% versus 29.9%; 15 to 29 ml/min: 12.1% versus 10.0%; 15 ml/min: 8.7% versus 6.5%. The mean age of the ombined study sample of polyysti kidney patients was 46.7 years, and 43% were men. Diabetes and ardiovasular disease were present in 5.3% and 2.2% of the full ohort, respetively; 1781 subjets were free of both these diagnoses. Compared with patients in the ombined ohort with 90 ml/min, those with lower s of renal funtion were older, more ommonly male, and had a higher baseline prevalene of ardiovasular disease. Unadjusted Annualized Charges by Level There were no signifiant differenes in annualized harges among patients with diagnosis odes for ADPKD versus unspeified polyysti kidney disease (P values for harge omparisons: Total medial, 0.69; inpatient, 0.54; outpatient pharmay, 0.31; total harges, 0.76). A ombined study sample of patients with and diagnosis odes was thus onsidered in subsequent analyses. Annualized medial and pharmaeutial harges differed signifiantly aross s of kidney funtion in the study sample (Table 2). Speifially, mean annual medial harges (unadjusted) inreased from $24,427 in polyysti kidney disease patients with baseline 90 ml/min to $134,784 in those with baseline 15 ml/min (P ). Considering inpatient medial harges separately, mean annual inpatient expenses were $10,721 in polyysti kidney disease patients with baseline 90 ml/min but were $37,304 in those patients with 15 ml/min (P ). Mean annual outpatient pharmay harges inreased from $2093 to $5355 aross baseline s of 90 ml/min to 15 ml/min (P ), and mean total healthare harges per year rose from $26,521 to $140,139 aross these s of renal funtion (P ). There were 187 patients who progressed to dialysis-dependene during observation (proportions per eah baseline : 90 ml/min: 2.0%; 60 to 89 ml/min: 1.4%; 30 to 59 ml/min: 5.4%; 15 to 29 ml/min: 22.4%; 15 ml/min: 68.8%). The mean annualized harges for dialysis servies were $131,890. Thirty-five patients reeived kidney transplants during observation (proportions per baseline : 60 ml/ min: 0%; 30 to 59 ml/min: 0.9%; 15 to 29 ml/min: 5.9%; 15 ml/min: 16.3%). The mean harges per transplant hospitalization were $119,931. Among patients with 15ml/min at baseline, 71% of total medial harges were related to dialysis or transplant servies. Nonlinear and Linear Regression Exploration of the relationship of with total annualized healthare harges by spline regression in the study sample is shown in Figure 1. This method ompares additional harges assoiated with eah ml/min differene ompared with a mid-range referene of 60 ml/min using a flexible funtional form. The resulting plot demonstrates mild flutuation in harges around a baseline of $0 (negligible differene from referene) until falls below 30 ml/min, at whih point annual osts rise preipitously with further derements in. The exploration in Figure 1 was used to inform onstrution of a multivariate linear regression model. Here annual harges assoiated with kidney funtion 30 ml/min were onsidered aording to KDOQI s rather than by ml/min hanges. Patients with 90 ml/min were hosen as the referene on linial grounds, and models were adjusted for age and gender. While s 31 to 60 ml/min were assoiated with a trend toward higher annual harges ompared with 90 ml/min (P 0.21), the effet estimates for funtion s 30 ml/min were nonsignifiant. Given the sharp rise in harges with 30 ml/min seen using the spline form, we modeled the harges assoiated with eah ml/min deline below 30 ml/min. By the resulting linear fit, eah ml/min deline in 30 was assoiated with an inrease in adjusted annual harges of approximately $5435 ompared with 90 ml/min (P ). Patterns were similar for total medial (exluding pharmay) harges, suh that eah ml/min derement in was assoiated with $5332 higher adjusted annual medial harges. Outpatient pharmay harges, whih omprised a smaller fration of total harges, showed an inrease at earlier stages of kidney dysfuntion. Adjusted annual

4 4 Clinial Journal of the Amerian Soiety of Nephrology Clin J Am So Nephrol 5:, 2010 Table 1. Demographi traits of the study sample by polyysti kidney disease diagnosis and kidney funtion Combined Cohort, ADPKD, ICD , n (%) a n (%) a Polyysti Kidney Unspeified, ICD , n (%) a Full Cohort n 1913 n 496 n 1417 Age in years, mean SD b b b Male gender 828 (43.3) b 221 (44.6) b 607 (42.8) b Diabetes 102 (5.3) 31 (6.3) 71 (5.0) Cardiovasular disease 43 (2.2) b 13 (2.6) b 30 (2.1) 90 n 295 n 68 n 227 Age in years, mean SD Male gender 94 (31.9) 23 (33.8) 71 (31.3) Diabetes 14 (4.8) 3 (4.4) 11 (4.9) Cardiovasular disease 3 (1.0) 0 3 (1.3) 60 to 89 n 696 n 167 n 529 Age in years, mean SD Male gender 272 (39.1) 67 (40.1) 205 (38.8) Diabetes 31 (4.5) 8 (4.8) 23 (4.4) Cardiovasular disease 5 (0.7) 3 (1.8) 2 (0.4) 30 to 59 n 582 n 158 n 424 Age in years, mean SD Male gender 299 (51.4) 79 (50.0) 220 (51.9) Diabetes 33 (5.7) 15 (9.5) 18 (4.3) Cardiovasular disease 18 (3.1) 7 (4.4) 11 (2.6) 15 to 29 n 205 n 60 n 145 Age in years, mean SD Male gender 98 (47.8) 34 (56.7) 64 (44.1) Diabetes 13 (6.3) 3 (5.0) 10 (6.9) Cardiovasular disease 8 (3.9) 1 (1.7) 7 (4.8) 15 n 135 n 43 n 92 Age in years, mean SD Male gender 65 (48.2) 18 (41.9) 47 (51.1) Diabetes 11 (8.2) 2 (4.6) 9 (9.8) Cardiovasular disease 9 (6.7) 2 (4.6) 7 (7.6) a Perentages reflet proportions within the polyysti kidney diagnosis ode ategory with the indiated demographi or linial traits ( olumn perentages ). There were no statistially signifiant differenes in distributions of listed linial traits in and subgroups at the P b P 0.05 for variation in the distribution of linial traits aording to within a polyysti kidney diagnosis ode ategory. pharmay harges were approximately $1500 higher in patients with 31 to 60 ml/min ompared with 90 ml/min (P 0.003), and then rose an estimated $199 per ml/min deline 30 (P 0.001). By ontrast, inpatient medial harges were not signifiantly assoiated with renal funtion after ovariate adjustment. Although not ommon in the study sample of polyysti kidney patients, both omorbid diabetes mellitus and ardiovasular disease were assoiated with inreased healthare harges (Table 3). In the absene of ardiovasular disease, diabetes at baseline predited approximately $27,847 higher annual harges independent of age, gender, and. In the absene of diabetes, ardiovasular disease was assoiated with approximately $188,220 higher annual harges independent of age, gender, and. There was a signifiant interation between these onditions in relation to osts, suh that omorbid diabetes and ardiovasular disease had similar ost impliations as diabetes alone in this sample. The age and genderadjusted assoiation of with annual harges in the full sample was similar after additional adjustment for diabetes and ardiovasular disease. Speifially, eah ml/min deline in 30 was assoiated with $5316 higher adjusted annual harges. Considering the sample of 1781 polyysti kidney disease patients who were free of diabetes and ardiovasular disease diagnoses at baseline, 31 to 60 ml/min was assoiated with approximately $10,788 higher total annual harges ompared with 90 ml/min that now reahed statistial signifiane (P 0.045) (Table 4). Subsequent derements in 30 min/ml were assoiated with estimated inreases in

5 Clin J Am So Nephrol 5:, 2010 Renal Funtion and Healthare Costs 5 Table 2. Annualized medial, pharmay and total harges aording to kidney funtion in the sample of polyysti kidney disease patients a Medial Charges ($) Level n Total Medial Inpatient Medial Outpatient Pharmay Charges ($) Total Charges ($) Mean Median Mean Median Mean Median Mean Median , , , to , , to , , , to , , ,806 10, ,784 42,910 37, ,139 46,779 P b a Charges were adjusted to the $2008 medial omponent of the Consumer Prie Index. b Variation in annualized osts by was ompared by the Kruskal-Wallis test, a nonparametri equivalent to oneway ANOVA. Pairwise omparisons of unadjusted harges in a given versus 90 ml/min were performed by the Wiloxon test, with statistial signifiane for pairwise omparisons defined as P 0.01 ( ) based on a Bonferroni orretion for multiple omparisons. Figure 1. Assoiations of (referened to 60 ml/min) with annualized total healthare harges in polyysti kidney disease patients by spline regression. Plot illustrates annual harge differenes ompared with referene of 60 ml/min. Charges were adjusted to the $2008 medial omponent of the Consumer Prie Index. total harges of $4914 per ml/min. Assoiations of with total medial only, inpatient and outpatient pharmay harges in the sample without diabetes and ardiovasular disease were similar to results in the full sample. Disussion ADPKD is a progressive geneti disorder that leads to advaned kidney failure in approximately half of affeted patients by mid-adulthood. We examined administrative reords of a large health insurane provider to assemble a ohort of patients with polyysti kidney disease and investigate the relationship of renal funtion with healthare harges to the insurer. The study yielded several key findings: (1) medial, pharmay, and total healthare osts varied signifiantly by baseline kidney funtion, suh that mean total annual harges (unadjusted) were approximately fivefold higher in polyysti kidney patients with 15 ml/min ompared with those with 90 ml/min; (2) after adjustment for age and gender, total harges did not differ signifiantly in the full sample among those with 30 ml/min, but rose preipitously in patients with 30 ml/min ompared with higher s of funtion; (3) omorbid diabetes and ardiovasular disease were signifiant orrelates of harges in patients with polyysti kidney disease, but the assoiation of and harges persisted after adjustment for these omorbidities; and (4) in a subgroup free of diabetes and/or ardiovasular disease at baseline, a signifiant rise in adjusted harges was deteted at 31 to 60 ompared with 90 ml/min. Quantifiation of the osts assoiated with health states offer valuable inputs for eonomi evaluations inluding ost-benefit analyses of standard and emerging disease management strategies. The dramati inrease in healthare harges for polyysti kidney disease patients with 30 ml/min is important but not surprising. While the study sample in this analysis was seleted to be free of indiations of dialysis or transplantation at baseline, GFR 30 ml/min is the funtional at whih many patients begin preparing for dialysis or transplant and fae inreased burdens of kidney failure-related ompliations suh as anemia, hypertension, and bone and mineral disorders (24). Renal replaement therapy often begins after GFR falls below 10 to 15 ml/min. Notably, outpatient medial harges appeared to be the main driver of the assoiation of with total harges in this analysis, with a lesser but onsistent ontribution from outpatient pharmay harges, whereas inpatient harges were not signifiantly assoiated with renal funtion in this sample. Other investigators have shown positive assoiations between healthare osts and proximity to dialysis in CKD patients. For example, St. Peter et al. found that osts in a Mediare/Mediaid population inreased from $993 per-pa-

6 6 Clinial Journal of the Amerian Soiety of Nephrology Clin J Am So Nephrol 5:, 2010 Table 3. Relationships of with annualized healthare harges by multivariate linear regression in the full study sample (n 1913) a Variables Parameter Estimate ($ per year SEM ) b P Parameter Estimate ($ per year SEM ) P Total Charges Interept 5886 (8537) (8266) 0.70 Age (per year) 579 (174) (170) 0.08 Male gender 107 (4134) (3988) ml/min referene referene 61 to 90 ml/min 1866 (5656) (5456) to 60 ml/min 7590 (6087) (5869) 0.19 Per ml/min (449) (433) Diabetes mellitus 27,847 (9292) Cardiovasular disease 188,220 (15847) Diabetes ardiovasular disease interation 185,935 (29767) Total Medial Charges Interept 6228 (8551) (8281) 0.74 Age (per year) 547 (174) (170) 0.11 Male gender 865 (4162) (4015) ml/min referene referene 61 to 90 ml/min 2067 (5676) (5476) to 60 ml/min 6292 (6096) (5878) 0.28 Per ml/min (448) (433) Diabetes mellitus 25,104 (9265) Cardiovasular disease 186,664 (15797) Diabetes ardiovasular disease interation 184,495 (29667) Inpatient Charges Interept 3465 (27799) ,240 (26967) 0.52 Age (per year) 1437 (520) (517) 0.05 Male gender 4782 (14202) ,690 (13628) ml/min referene referene 61 to 90 ml/min 1262 (20653) (19776) to 60 ml/min 5328 (20169) ,895 (19331) 0.54 Per ml/min (1164) (1114) 0.61 Diabetes mellitus 18,965 (24284) 0.44 Cardiovasular disease 176,913 (30285) Diabetes ardiovasular disease interation 193,981 (53821) Outpatient Pharmay Charges Interept 542 (712) (714) 0.35 Age (per year) 41 (15) (15) 0.02 Male gender 581 (343) (343) ml/min referene referene 61 to 90 ml/min 18 (472) (471) to 60 ml/min 1491 (502) (501) Per ml/min (37) (36) Diabetes mellitus 2484 (789) Cardiovasular disease 1066 (1341) 0.43 Diabetes ardiovasular disease interation 1514 (2464) 0.54 a Charges were adjusted to the $2008 medial omponent of the Consumer Prie Index. b Regression model among the full sample, exluding parameters for diabetes and ardiovasular disease. Regression model among the full sample, inluding parameters for diabetes and ardiovasular disease.

7 Clin J Am So Nephrol 5:, 2010 Renal Funtion and Healthare Costs 7 Table 4. Relationships of with annualized healthare harges by multivariate linear regression in the polyysti kidney disease sample free of baseline diabetes and ardiovasular disease (n 1781) a Variable Parameter Estimate ($ per year SEM ) Total Charges Interept 3403 (7540) 0.65 Age (per year) 318 (156) 0.04 Male gender 4664 (3641) ml/min referene 61 to 90 ml/min 3344 (4966) to 60 ml/min 10,788 (5386) Per ml/min (403) Total Medial Charges Interept 3042 (7541) 0.69 Age (per year) 291 (156) 0.06 Male gender 5581 (3660) ml/min referene 61 to 90 ml/min 3460 (4975) to 60 ml/min 9604 (5384) 0.07 Per ml/min (402) Inpatient Charges Interept 5199 (26,584) 0.85 Age (per year) 1339 (520) 0.01 Male gender 27,079 (13,632) ml/min referene 61 to 90 ml/min 5722 (19,262) to 60 ml/min 10,428 (19,194) 0.59 Per ml/min (1106) 0.12 Pharmay Charges Interept 595 (735) 0.42 Age (per year) 35 (15) 0.02 Male gender 681 (353) ml/min referene 61 to 90 ml/min 108 (484) to 60 ml/min 1428 (518) Per ml/min (38) a Charges were adjusted to the $2008 medial omponent of the Consumer Prie Index. tient per-month 2 years before dialysis to $6300 per-patient per-month 6 months after dialysis initiation, with a positive gradient in between these periods (25). Similar observations have been made in managed are populations (26). Consistent with our observation of nonsignifiant hanges in adjusted harges in polyysti kidney disease patients with 90 to P 30 ml/min, Smith et al. found that while are of CKD patients was more expensive than that of ontrols, provider-side osts were stable aross pre-esrd CKD stages 2 to 4 in an HMO setting (11). This study extends understanding of relationships of renal funtion and healthare osts to patients with polyysti kidney disease as the ause of CKD, demonstrating sharp arual of eonomi impat at advaned stages of renal insuffiieny. Assuming novel interventions to slow progression of ADPKD prove linially effetive in the future, the eonomi benefits of suh interventions are likely to be greatest when funtion loss below approximately 30 ml/min an be prevented. Cardiovasular disease has been regarded as the most ostly disease in the United States, as the ombination of high prevalene and ase management that may entail expensive proedures was estimated to aount for $400 billion diret and indiret osts for the nation in 2006 (27). Diabetes mellitus was assoiated with approximately $174 billion in total osts in the United States in 2007 (28), although some of the osts of diabetes are mediated by end organ ompliations suh as ardiovasular disease and kidney failure. We therefore estimated the assoiations of baseline ardiovasular disease and diabetes with healthare harges in our sample of polyysti kidney disease patients. While unommon in this sample, both omorbidities were signifiantly assoiated with annual total harges. Importantly, however, the relationship of with osts was not signifiantly affeted by adjustment for these onditions. In a sample seleted to be free of baseline diabetes and ardiovasular disease, the signifiant assoiation between and total harges expanded to the patients with 31 to 60 ml/min. Our study has limitations. As with any observational study, our findings represent assoiations and do not prove ausal links between the studied variables. A ross-setional design was used based on the nature of the available data. Although prospetive longitudinal studies may demonstrate a somewhat different form of the relationship of GFR with healthare osts, based on the nature of the variables of interest, it is unlikely that reverse ausation (i.e., that osts aused hanges in ) explain the findings. Billing laims have been demonstrated to provide sensitive measures of diagnoses suh as diabetes and ardiovasular disease (18,22), but administrative oding algorithms for identifying polyysti kidney disease have not been validated with external data. Analogous to the approah of prior studies of ADPKD in administrative data (29,30), we ombined patients with and laims in the main analysis based on the similarity of these groups. Lak of information on partiipant rae may have produed some mislassifiation of renal funtion. As the amdrd equation inludes a saling fator of 1.21 for blak rae, was underestimated for Afrian Amerian partiipants by approximately 21%. However, as Afrian Amerians aount for only 12.5% of the general United States population and are underrepresented among the privately insured in this ounty (31), potential mislassifiation of CKD based on rae likely affeted only a small number of partiipants. We used harges as a surrogate measure for osts based on the billing information available in the soure

8 8 Clinial Journal of the Amerian Soiety of Nephrology Clin J Am So Nephrol 5:, 2010 data. While harges are frequently examined as proxy measures in studies of healthare eonomis, harge data may vary based on inlusion of different ost items and modifiation of harges over time and between enters. Moreover, harges an be approximately 1.6 to 1.8 times greater than osts (32). However, the relative relationship between ost and harges is unlikely to vary aross s of renal funtion and, therefore, should not invalidate the key findings of this investigation. Findings may differ under other reimbursement systems suh as publi insurane. We were also unable to study indiret osts to the patient. Despite these limitations, the data soure used enabled study of insurane data in nearly 2000 patients with polyysti kidney disease, a sample muh larger than ould be assembled at a single-enter enter or even multi-institutional. In onlusion, we observed a strong assoiation of healthare harges with advaned renal dysfuntion in privately insured polyysti kidney disease patients, independent of age and omorbid diseases suh as diabetes and ardiovasular disease. This evidene is relevant to eonomi evaluations of new interventions that ould prevent or delay progression of polyysti kidney disease. Strategies that prevent loss of renal funtion below 30 ml/min have the potential to generate substantial redutions in medial harges. Aknowledgments Support for data use fees and researh staff time was provided by Novartis Pharmaeutials, Corp. The analyses, interpretation, medial writing, and reporting of these data are the sole responsibility of the authors. Institution at whih work was performed: Saint Louis University, St. Louis, Missouri. Dislosures None. Referenes 1. Grantham J, Cowley BJ, Torres VE: Progression of autosomal dominant polyysti kidney disease (ADPKD) to renal failure. In: The Kidney: Physiology and Pathophysiology, edited by Seldin DW, Geibish G, Philadelphia, Lippinott Williams and Wilkins, 2000, pp Grantham JJ, Torres VE, Chapman AB, Guay-Woodford LM, Bae KT, King BF, Jr., Wetzel LH, Baumgarten DA, Kenney PJ, Harris PC, Klahr S, Bennett WM, Hirshman GN, Meyers CM, Zhang X, Zhu F, Miller JP: Volume progression in polyysti kidney disease. N Engl J Med 354: , Mitheson H, Williams G, Castro JE: Clinial aspets of polyysti disease of the kidneys. BMJ 1: , Bear JC, MManamon P, Morgan J, Payne RH, Lewis H, Gault MH, Churhill DN: Age at linial onset and at ultrasonographi detetion of adult polyysti kidney disease: Data for geneti ounselling. Am J Med Genet 18: 45 53, Churhill DN, Bear JC, Morgan J, Payne RH, MManamon PJ, Gault MH: Prognosis of adult onset polyysti kidney disease re-evaluated. Kidney Int 26: , Fik GM, Gabow PA: Natural history of autosomal dominant polyysti kidney disease. Annu Rev Med 45: 23 29, Inidene and prevalene of ESRD. United States Renal Data System. Am J Kidney Dis 32: S38 S49, Calvert JP: Strategies to inhibit yst formation in ADPKD. Clin J Am So Nephrol 3: , Leuenroth SJ, Crews CM: Targeting yst initiation in ADPKD. J Am So Nephrol 20: 1 3, Go AS, Chertow GM, Fan D, MCulloh CE, Hsu CY: Chroni kidney disease and the risks of death, ardiovasular events, and hospitalization. N Engl J Med 351: , Smith DH, Gullion CM, Nihols G, Keith DS, Brown JB: Cost of medial are for hroni kidney disease and omorbidity among enrollees in a large HMO population. JAm So Nephrol 15: , Cheung BM, Ong KL, Cherny SS, Sham PC, Tso AW, Lam KS: Diabetes prevalene and therapeuti target ahievement in the United States, 1999 to Am J Med 122: , Levey AS, Greene T, Kusek JW, Bek G: A simplified equation to predition glomerular filtration rate from serum reatinine. J Am So Nephrol 11: 155A, Smellie WS, Shaw N, Bowley R, Stewart MF, Kelly AM, Twomey PJ, Chadwik PR, Houghton JB, Ng JP, MCulloh AJ: Best pratie in primary are pathology: Review 10. J Clin Pathol 60: , Middleton RJ, Foley RN, Hegarty J, Cheung CM, MElduff P, Gibson JM, Kalra PA, O Donoghue DJ, New JP: The unreognized prevalene of hroni kidney disease in diabetes. Nephrol Dial Transplant 21: 88 92, Philipneri MD, Roa Rey LA, Shnitzler MA, Abbott KC, Brennan DC, Takemoto SK, Buhanan PM, Burroughs TE, Willoughby LM, Lentine KL: Delivery patterns of reommended hroni kidney disease are in linial pratie: Administrative laims-based analysis and systemati literature review. Clin Exp Nephrol 12: 41 52, National Kidney Foundation: K/DOQI linial pratie guidelines for hroni kidney disease: Evaluation, lassifiation, and stratifiation. Am J Kidney Dis 39: S1 S266, Hebert PL, Geiss LS, Tierney EF, Engelgau MM, Yawn BP, MBean AM: Identifying persons with diabetes using Mediare laims data. Am J Med Qual 14: , Kasiske BL, Snyder JJ, Gilbertson D, Matas AJ: Diabetes mellitus after kidney transplantation in the United States. Am J Transplant 3: , Stevens LA, Fares G, Fleming J, Martin D, Murthy K, Qiu J, Stark PC, Uhlig K, Van Lente F, Levey AS: Low rates of testing and diagnosti odes usage in a ommerial linial laboratory: Evidene for lak of physiian awareness of hroni kidney disease. J Am So Nephrol 16: , U.S. Renal Data System: USRDS 2008 Annual Data Report, Methods Appendix. Bethesda, National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases, Available at Aessed July 6, Lentine KL, Shnitzler MA, Abbott KC, Bramesfeld K, Buhanan PM, Brennan DC: Sensitivity of billing laims for

9 Clin J Am So Nephrol 5:, 2010 Renal Funtion and Healthare Costs 9 ardiovasular disease events among kidney transplant reipients. Clin J Am So Nephrol 4: , Bureau of Labor Statistis. Consumer Prie Index, U.S. Average Medial Care. Available at servlet/surveyoutputservlet. Aessed August 7, London R, Solis A, Goldberg GA, Wade S, Ryu S: Health are resoure utilization and the impat of anemia management in patients with hroni kidney disease. Am J Kidney Dis 40: , St Peter WL, Khan SS, Ebben JP, Pereira BJ, Collins AJ: Chroni kidney disease: The distribution of health are dollars. Kidney Int 66: , Robbins JD, Kim JJ, Zdon G, Chan WW, Jones J: Resoure use and patient are assoiated with hroni kidney disease in a managed are setting. J Manag Care Pharm 9: , Mensah GA, Brown DW: An overview of ardiovasular disease burden in the United States. Health Aff (Millwood) 26: 38 48, National Diabetes Statistis, Estimated Diabetes Costs in the United States in Available at diabetes.niddk.nih.gov/dm/pubs/statistis/#osts. Aessed August 7, Freedman BI, Souie JM, Chapman A, Krisher J, MClellan WM: Raial variation in autosomal dominant polyysti kidney disease. Am J Kidney Dis 35: 35 39, Perrone RD, Kausz TA, Chavers BM, Weinhandl ED, Collins AJ: Inidene of autosomal dominant polyysti kidney disease (ADPKD) among ESRD patients in the United States. J Am So Nephrol 16 (Suppl): 358A, DeNavas-Walt C, Protor BD, Smith JC: Inome, poverty, and health insurane overage in the United States: Current Population Reports, US Census Bureau Issued Health Care Finaning Administration. Prospetive Payment System: Standardizing File. Washington DC, Department of Health and Human Servies, 1998