You are on page 1of 10

Article

Association of Vascular Access Type with Mortality,
Hospitalization, and Transfer to In-Center Hemodialysis
in Patients Undergoing Home Hemodialysis
Matthew B. Rivara,* Melissa Soohoo,† Elani Streja,† Miklos Z. Molnar,‡ Connie M. Rhee,† Alfred K. Cheung,§
Ronit Katz,* Onyebuchi A. Arah,| ¶ Allen R. Nissenson,**†† Jonathan Himmelfarb,* Kamyar Kalantar-Zadeh,† and
Rajnish Mehrotra*
*Department of
Medicine, Division of
Abstract Nephrology, Kidney
Background and objectives In individuals undergoing in-center hemodialysis (HD), use of central venous Research Institute,
catheters (CVCs) is associated with worse clinical outcomes compared with use of arteriovenous access. University of
However, it is unclear whether a similar difference in risk by vascular access type is present in patients Washington, Seattle,
Washington; †Harold
undergoing home HD. Simmons Center for
Kidney Disease Research
Design, setting, participants, & measurements Our study examined the associations of vascular access type with and Epidemiology,
all-cause mortality, hospitalization, and transfer to in-center HD in patients who initiated home HD from 2007 to Division of Nephrology
2011 in 464 facilities in 43 states in the United States. Patients were followed through December 31, 2011. Data and Hypertension,
University of California
were analyzed using competing risks hazards regression, with vascular access type at the start of home HD as the Irvine Medical Center,
primary exposure in a propensity score–matched cohort (1052 patients; 526 with CVC and 526 with arteriovenous Irvine, California;

access). Division of Nephrology,
Department of
Medicine, University
Results Over a median follow-up of 312 days, 110 patients died, 604 had at least one hospitalization, and 202 of Tennessee Health
transferred to in-center hemodialysis. Compared with arteriovenous access use, CVC use was associated with Science Center,
higher risk for mortality (hazard ratio, 1.73; 95% confidence interval, 1.18 to 2.54) and hospitalization (hazard Memphis, Tennessee;
§
ratio, 1.19; 95% confidence interval, 1.02 to 1.39). CVC use was not associated with increased risk for transfer to in- Division of Nephrology
center HD. The results of analyses in the entire unmatched cohort (2481 patients), with vascular access type and Hypertension,
University of Utah,
modeled as a baseline exposure at start of home HD or a time-varying exposure, were similar. Analyses among a Salt Lake City, Utah;
propensity score–matched cohort of patients undergoing in-center HD also showed similar risks for death and |
Department of
hospitalization with use of CVCs. Epidemiology, Fielding
School of Public Health,
University of California,
Conclusions In a large cohort of patients on home HD, CVC use was associated with higher risk for mortality and Los Angeles, Los
hospitalization. Additional studies are needed to identify interventions which may reduce risk associated with Angeles, California;

use of CVCs among patients undergoing home HD. University of California,
Clin J Am Soc Nephrol 11: 298–307, 2016. doi: 10.2215/CJN.06570615 Los Angeles (UCLA),
Center for Health Policy
Research, Los Angeles,
California; **Division of
Nephrology, Department
Introduction An alternative to cannulation of AV access is the use of Medicine, David
In the United States, .110,000 individuals with ESRD of tunneled central venous catheters (CVCs) as long– Geffen School of
initiate maintenance dialysis each year (1). Although term vascular access for patients treated with home Medicine at University of
the vast majority of these patients are treated with in- HD. In patients undergoing in-center HD, use of California, Los Angeles
center hemodialysis (HD), there has been a recent re- (UCLA), Los Angeles,
CVCs is associated with greater risk of adverse clin-
California; and ††Office
surgence in use of home HD (2,3). Growth in use of ical outcomes, including higher rates of mortality and of the Chief Medical
home HD has been facilitated by data suggesting clin- hospitalization, at least in part because of greater in- Officer, DaVita, Inc., El
ical benefit with more frequent HD treatments as well cidence of infection-related complications (10–14). In Segundo, California
as the introduction of simple to operate systems for contrast, there are only limited data examining the
home HD (4,5). association of vascular access type with clinical out- Correspondence:
One of the persistent barriers to greater adoption of Dr. Matthew B. Rivara,
comes for patients undergoing home HD, individuals
Kidney Research
home HD is the need for patients or caregivers to in whom incidence of nosocomial infection may be Institute, University of
perform frequent cannulation of arteriovenous (AV) lower (15). Washington, 325
access (fistula or graft) (6–8). Data from the Frequent Using nationally representative data from a large Ninth Avenue, Box
Hemodialysis Network Trial have also shown a dialysis provider in the United States, we undertook this 359606, Seattle, WA
98104. Email: mbr@
higher incidence of interventions on AV access with study to examine the null hypothesis that, in patients uw.edu
higher frequency of use as occurs with home HD (9). undergoing home HD, use of a CVC compared with

298 Copyright © 2016 by the American Society of Nephrology www.cjasn.org Vol 11 February, 2016

center HD. calcium. The Institutional Review Boards at time-varying exposure that was updated at the start of the Los Angeles Biomedical Research Institute and the each 91-day period of follow-up. and transfer to in-center HD. Inc. score–matched cohort to determine the associations of ed with an AV graft (n=323) in the primary cohort. the follow. cause validity of our findings. we performed competing risk regression to examine the puted using multiple imputation with five repetitions (17). Inc. we sought to juxtapose our serum albumin. assessed as distinct exposures rather than pooled together pacity. primary health insurance. and erythropoietin dose were set of analyses. time to first hospitalization. total iron–binding ca. sex. Because of the small number of patients treat. the presence of effect modification by facil- Dialysis facility experience with home HD was defined as ity experience with incident home HD overall and use of the number of 91-day patient periods in which patients on CVC for home HD was tested through assessment of the home HD were treated at that facility. bicarbon- findings for patients undergoing home HD with those ate. Data for serum albumin. alkaline phosphatase analyses. In the first potassium. culture. cause of ESRD. body mass index. year of start of matched analysis of patients only treated with conventional . blood hemoglobin. Standardized differ- DaVita. and weekly dose of erythropoietin. Finally. percentage of lymphocytes. was modeled as a static baseline exposure in the survival parathyroid hormone. analysis (Supplemental Figure 2). serum calcium. the referent group comprised patients treated patients included in the cohorts compared with those ex. vascular access type at the start of home HD missing for . body mass index. diabetes status. we performed a propensity score– of ESRD. creatinine. with an AV access. Censoring reasons included transplant. or transfer to in. FL). Additionally. potassium. time–varying serum albumin and creat- primary health insurance. dialysis facility home HD experience. February. In the second set. sex. of home HD. Similarly. This model was used to calculate the probability of each patient being treated Materials and Methods with a CVC at the time of start of home HD (propensity Data Source scores). Data were complete for age.1 SDs (18). For were used to determine demographics and comorbidities. AV fistula and AV graft were and creatinine. Three nested hierarchical University of Washington approved the study as exempt models were examined: (1) unadjusted. As an exploratory analysis. transferrin saturation. dialysis facility region. discharge to a facility operated by another dialysis pro- ed in Supplemental Tables 1 and 2. parathyroid hormone. race. and end of follow-up (Supplemental Table 4). and (3) addi- tionally adjusted for duration of dialysis treatment before Statistical Analyses start of home HD. cluded because of missing vascular access type are report. Data for race. A logistic regression outcomes through assessment of significance of the first– model was built with CVC as initial access type as the order interaction term. defined as first incidence of a positive outpatient blood Missing data for vascular access type were not imputed. from patients with an AV graft or fistula were pooled into a first hospitalization. For each single group for the primary analysis. Characteristics of analysis. count. cumulative iron dose. and diabetes status. matched cohort (n=2481). Patients who under. Unadjusted time to event competing risks survival up period was divided into 91-day periods from the date of analyses (19) were performed within the propensity first dialysis. age. association of vascular access type with first bacteremia. and cardiovascular comorbid.Clin J Am Soc Nephrol 11: 298–307. for comparison and to enhance the external sex. for at least 60 days (16). outcome and the following variables as predictors: age. each outcome. Patients qualitatively compared with the standardized differences who were only treated with conventional three times per between the groups in the unmatched cohort to confirm week in–center HD were analyzed as part of a secondary the success of the matching (Supplemental Table 3). ferritin. ences between the CVC and AV access groups in the went treatment with home HD for $45 days were included matched cohort were calculated for each variable and in the primary analysis (Supplemental Figure 1). modeling vascular access as a laboratory (Deland. discontinuation of dialysis or recovery of renal func- Data from dialysis facility electronic medical records tion. 2016 Vascular Access and Outcomes in Home Hemodialysis. blood he- for patients undergoing conventional three times per moglobin. Missing covariate data were im. and white blood cell count were missing into a single AV access group. For each patient. ferritin. ities. race/ethnicity. the data CVC use at the start of home HD with all-cause mortality. and time–varying blood hemoglobin. we assessed for evidence of effect For the primary analysis. alkaline phosphatase. a propensity score–matched modification by duration of CVC use during home HD on cohort was constructed to minimize the influence of bias the association between initial vascular access type and caused by confounding by indication. cardiovascular comorbidity. Rivara et al. 299 use of an AV access is not associated with higher risk of maintenance dialysis. All competing risk regression analyses within the entire un- laboratory values were measured in the central DaVita. Vascular access type was complete for 95% of patients. facility geographic region. (2) adjusted for from informed consent. home HD CVC experience was defined as the number of To further examine the robustness of our findings and 91-day periods in which patients on home HD were treated incorporate information on changes in vascular access and with a CVC at that facility. for 1%–7% of the cohort. vider. inine. Laboratory measurements were time-varying confounders during follow-up. we performed averaged for each patient for each 91-day period. In addition.1% of the cohort. Two sets of sensitivity analyses were performed. facility significance of the first–order interaction term. Propensity scores were used to identify one pa- The study population comprised patients $18 years of tient initiating home HD with an AV access for each pa- age who started maintenance dialysis from 2007 to 2011 tient with a CVC using a greedy matching algorithm and received care at one of the facilities operated by with a caliper width of 0. diabetes. white blood cell week HD. and bicarbonate. time from start of dialysis to start mortality. phosphorus.

960. mg/mo 0 [0.3 11.548] 5500 [1894. % Diabetes 38 34 32 38 38 Hypertension 17 24 23 18 17 GN 19 18 18 19 20 Other 26 24 27 5 5 Comorbid conditions. 527] 343 [223. 558] 387 [200.261. 534] 342 [209.4 5. 549] 333 [204. 300] 0 [0.060. g/dl 3.3 Serum ferritin.3 11. 300] White blood cell count.4 11. 573] 327 [181. 12. 583] 252 [95. % White 75 67 73 74 73 Black 17 22 20 17 19 Hispanic 6 6 2 6 5 Asian 1 3 3 1 1 Other 1 2 2 2 2 Clinical Journal of the American Society of Nephrology Cause of ESRD. units/wk 5842 [2199.960.6 9. 369] 308 [124.570] 5819 [2097. 382] 197 [71.2 Blood hemoglobin. kg/m2 2968 3067 31610 2968 2967 Time from start of dialysis 157 [45.5 11.2 5.960.5 Serum calcium. 0] 0 [0.9 .161. 13.2 5. 621] 417 [197.656] 4878 [1599. Characteristics of unmatched (n=2481) and propensity score–matched (n=1052) study cohorts stratified by initial vascular access type at the time of initiation of home hemodialysis Unmatched Cohort Propensity Score–Matched Cohort Characteristics CVC (n=579) AV Access (n=1794) Missing (n=108) CVC (n=526) AV Access (n=526) Age.5 3.361. mg/dl 5. mg/dl 241655 252646 244652 24565 243646 Median erythropoietin dose. % Atherosclerotic heart disease 28 27 19 27 26 Congestive heart failure 53 49 34 52 52 Other cardiovascular diseases 26 22 14 24 24 Primary insurance Medicare 39 39 46 39 37 Medicaid 3 3 4 3 5 Other 58 57 50 58 58 Dialysis facility region Northeast 15 17 17 15 15 West 18 21 20 19 16 Midwest 34 24 19 32 36 South 34 39 45 35 33 Laboratory data Serum albumin. % men 61 67 67 62 66 Diabetes mellitus.5 [0.860. 407] to HHD.6 8. % 61 60 57 61 60 Body mass index.462.960.062.462.6 8. yr 54615 52614 52614 53615 52614 Sex. d Race/ethnicity.006] Iron dose. 300] 0 [0. pg/ml 320 [179.5 3. 11.860.5 4. 13.461.488] 4400 [1468. 598] Serum total iron binding capacity. g/dl 11.161. 13.8 7.960. mg/dl 8.2 5.960. 300 Table 1.362. 530] Serum phosphorus. 606] 166 [52. 725] 323 [184. 597] 380 [205. ng/ml 328 [185. 300] 12.161.4 3.5 7.161.562.161.6 Serum parathyroid hormone.3 7.6 7.060.161.161.7 8. 3103/ml 7.

home hemodialysis. 2543 4.0 1 and remained constant over the remainder of the follow- 2268 2462 up period (Figure 1). % per 100 person-years for patients using a CVC compared with 46. and 386 patients transferred to in-center HD over 2549 person-years of follow-up.6 7. Me- dian treatment time per session and mean number of treat- ments per week were 168 (IQR=150–191) minutes and 4. . February. 174 patients with known initial vascular access type died. In contrast to the un- matched cohort. CVC. were more likely to be Missing (n=108) black. arteriovenous.5 to 62.360. NC). Statistical analyses were performed using SAS. mg/dl 56. Overall. respectively. 2007 and December 31. mEq/L hospitalization and transfer to in–center HD rates were Potassium.0 start of in-center HD with all-cause mortality and first 2369 2463 hospitalization. patients with a CVC were older. irrespective of initial vascular access type (Supplemental Table 4).5 7. 2016 Vascular Access and Outcomes in Home Hemodialysis.1 (95% confidence interval [95% CI].1 to 9. version 9.3 (SAS Institute Inc. were more likely have diabetes or cardiovascular comorbidity. 9. Of 2268 2462 these. (Continued) Creatinine.7 to 52. The propensity score–matched cohort comprised 526 pairs of individuals with CVC or AV ac- cess at time of start of home HD. In contrast. AV. 114] Between January 1. 2481 patients had at least one 91-day period in which there was available information on vascular access type. AV Access (n=526) Time to event competing risk survival analyses were performed to determine the associations of CVC use at the Propensity Score–Matched Cohort 82 [65. 1199 were hospitalized at least one time. 13. Within the propensity score–matched cohort.563.431 pairs of patients).0) Lymphocyte. 301 three times per week in–center HD (22. 82 [65. and had lower baseline serum albumin and 76 [63.963. the CVC (n=579) prevalence of AV access increased to 86% by end of year 83 [65. mEq/L Table 1. Results Home HD Study Population CVC (n=526) Data are presented as means6SDs. .4 treatments for patients with a CVC and 161 (IQR=146–180) minutes and 4. compared with patients initiating home HD with an AV access. 41. median [interquartile range]. 102] 4.4 to 62. 5. The Bicarbonate.2 2468 2463 [IQR] =45–369) for patients with a CVC and 308 days (IQR=124–583) for patients with AV access (Table 1). central venous catheter.161. 116] 4. 112] 4. Among all study pa- tients. U/L Characteristics person-years compared with 6. Rivara et al.3 treatments for pa- tients with AV access.6) and 16.0 2269 2363 creatinine (Table 1). or percentage.863. 72% started home HD with an AV access. 116] 4.6 7.460.9) per 100 Alkaline phosphatase. . 14.2) per 100 person-years for patients with AV access (Table 2). there were no meaningful differences between the groups with CVC or AV access in any of the measured baseline demographic characteristics or labora- tory variables.261.9) for patients using AV access. AV Access (n=1794) Unmatched Cohort The median duration of maintenance dialysis treatment before start of home HD was 157 days (interquartile range 78 [61. HHD.460..9 (95% CI.3 (95% CI.9) and 18.062.8 (95% CI.9 to 22.6 6.10% of patients who initiated home HD with an AV access switched to use of a CVC (Figure 2).3 (95% CI. regained kidney function.50% after 1 year had switched to use of an AV access (Figure 2). for patients initiating home HD with CVC. 50.6 7.9 patients started home HD in 464 facilities in 43 states. Cary. For the unmatched cohort (2481 patients).Clin J Am Soc Nephrol 11: 298–307. had a shorter interval from date of first-ever dialysis to home HD. or transferred to a nonaffiliated dialysis facility were similar. The proportions of pa- tients who underwent kidney transplantation.460.1 (95% CI. 2011.460.263. the crude mortality rate for patients with CVC was 12.2 to 19.

regardless of level of ad. 2.02 to 1.18 to 2. with a trend toward with higher risk for all-cause mortality (HR.99 to 1. the relationship between CVC use unmatched cohort were similar to analyses within the and risk for transfer to in-center HD was significantly propensity score–matched cohort (Figure 3). However. Association of Vascular Access Type with All-Cause effect modification by either dialysis facility home HD ex- Mortality in Home HD perience (P value for interaction =0. pensity score–matched cohort. cohort after full adjustment for potential confounders The results of secondary analyses within the entire (Figure 3). In-Center HD in Home HD ure 3).28).01).81) or vascular access type and risk for transfer to in-center HD in facility home HD CVC experience (P value for interac. There were no meaningful differences between the justment.39. 1.28).302 Clinical Journal of the American Society of Nephrology Figure 1. Analyses in dergoing in-center HD comprised 22. patients HD CVC experience (P value for interaction =0.19. There was no evidence of effect substantially attenuated (HR. 95% CI. there was no evidence for groups with CVC or AV access in any of the measured . the association with all-cause mortality was 3.95) modification by facility home HD experience (P value for (Figure 3).36. 1. 1. central venous catheter. arteriovenous. 1. 1. the propensity score–matched cohort or the unmatched tion =0. Supplemental Table 5). CVC.431 pairs of individ- the unmatched cohort were similar to those in the pro.73. 95% CI.54) (Fig.10). 1. interaction =0. AV. Association of Vascular Access Type with Hospitalization in Association of Vascular Access Type with All-Cause Home HD Mortality and Hospitalization in Conventional In–Center Compared with patients undergoing home HD with an HD AV access. There was no effect modification by dialysis facility There was no statistically significant association between home HD experience (P value for interaction =0. undergoing home HD with CVCs had greater risk for all-cause mortality compared with patients using an AV Association of Vascular Access Type with Transfer to access (hazard ratio [HR]. | Vascular access over study follow-up among patients undergoing home hemodialysis (HD. 95% stronger risk for patients undergoing dialysis in facilities CI. uals with CVC or AV access at time of start of dialysis.39). modified by the cumulative dialysis facility CVC experi- adjusted model.0. n=2481).73 to 3. 0. Using an un.22). As for mortality. patients using a CVC had a higher risk for The propensity score–matched cohort of patients un- hospitalization (HR. After adjustment for potential con. time–varying CVC use was associated ence (P value for interaction .61) or facility home Within the propensity score–matched cohort. 95% CI. with lower home HD CVC experience (Supplemental Figure founders.

35) and hospitalization (HR. 1. access (Figure 4).37. Rivara et al. . 2016 Vascular Access and Outcomes in Home Hemodialysis. 1. n=579).Clin J Am Soc Nephrol 11: 298–307. (B) Patients with initial vascular access as arteriovenous (AV) access (n=1794). baseline demographic characteristics or laboratory vari. February. 95% ables (Supplemental Tables 6 and 7).34 to 1. 1. | Vascular access over study follow-up among patients undergoing home hemodialysis (HD) stratified by initial vascular access type. Individuals with a CI.25 to 1.30. 303 Figure 2.40) compared with individuals with an AV CVC had a significantly higher risk for death (HR. 95% CI. 1. (A) Patients with initial vascular access as central venous catheter (CVC.

9) 17. 2. there was no evidence of effect modifica- Crude Rate tion by duration of CVC use on the association of initial (95% CI)a vascular access type with outcomes.4 to 62. patients dialyzing with a graft did not have a higher risk for all-cause mortality. HD in the home setting may lead Central venous catheter Central venous catheter to higher risk for rare but potentially catastrophic AV Rate per 100 person-years.20–22).3) 5.9 (5. Additionally. Patients undergoing in-center HD have three times per week expo- Patient-Yr of sure to dialysis facilities.3) 12.0 (4.3) 14. Results of an exploratory analysis examining the Table 2.304 Clinical Journal of the American Society of Nephrology Sensitivity and Exploratory Analyses In sensitivity analyses modeling vascular access type at Transfer to In-Center Hemodialysis 18.7 (14.1 to 9. These findings were observed in an analysis of a propensity score–matched cohort as well as after adjustment for po- tential confounders in a larger unmatched cohort using 12. in which they come into contact Follow-Up with other chronically ill patients as well as facility staff 1920 563 613 629 members.9 to 15.6 (40.3 (12. there are important ways in 68 42 78 96 which home HD differs from in-center HD that may sub- stantially alter the risks associated with CVC use. Access Although these differences provide a rationale for why Unmatched cohort CVC risk may be different for in-center HD and home HD. 95% CI. results were similar to those from the primary analyses (Supplemental Table 8). thrombolytic . we found that treatment with a CVC was associated with a Events higher risk for death and hospitalization but not transfer to 317 287 361 838 in-center HD compared with treatment with an AV access. Crude Rate (95% CI)a the magnitude of higher risk for death and hospitalization observed with use of a CVC in home HD was similar to Mortality that observed in a parallel analysis of patients undergoing conventional in–center HD. and event rate by initial vascular access type in propensity score–matched (n=1052) and unmatched (n=2481) cohorts association of initial vascular access type with first bacter- Events emia events showed higher risk for bacteremia with use of 102 100 111 275 CVC within the propensity score–matched cohort (HR.1 (14.9 to 22.7 to 52. Beyond potential differences in risk 95% CI. 3.1 to 6. Arteriovenous access Arteriovenous access access–related adverse events (26.4 (9. In a second set of sensitivity analyses assessing AV fistula or AV graft as distinct exposures.8 to 63. Finally. the results of our study show that higher risks for mortality and hospitalization associated with CVC use also exist for patients undergoing home HD and are similar in magni- tude (22). However.8 (41.1) baseline or time–varying vascular access. irrespective of the level of statistical adjustment (HR. including multidrug-resistant organisms (23–25). Follow-up. 95% CI.5 to 15.1) the start of home HD as the primary exposure within the Crude Rate (95% CI)a entire unmatched cohort. number of events.4 (51.3 (13.6) 57.59. Numerous studies over the past two decades have shown associations between CVC use and higher risk for adverse clinical outcomes in patients undergoing in-center Events HD (10–14. hospitaliza- tion.3 (50.20) as well as within the unmatched cohort.8 to 46. Although rates of colonization with multidrug- resistant organisms are unknown in patients undergoing home HD. These results support the findings of previous observational studies that have shown similar rates a of complications.21 to 4. 1. 95% confidence interval. regardless of the level of covariate adjust- ment.68 for the fully adjusted 56. it is plausible that less frequent contact with Propensity score–matched cohort health care facilities would result in less frequent coloni- zation and infection. including infection.7 to 16.5) 6.1 (9. for nosocomial infection. Hospitalization Discussion In this large study of patients on incident home HD.7 to 21.6) 43.0) 16. 2.60 to 4.27).2 to 19.22.9) 46. or transfer to in-center HD compared with those using a fistula. presenting opportunities for exposure to exoge- nous organisms.7) model).

This flammation. n=22. (28. February. We found that that over one half of patients who for colonization and subsequent catheter–related bacter. | Association of central venous catheter (CVC) use with risk for all-cause mortality and hospitalization among a propensity score– matched cohort of patients undergoing conventional three times per week in–center hemodialysis. AV access. associ. Figure 4. even after dialysis (13. home HD with a CVC switch to use of AV access. body mass index. has been shown to predict finding emphasizes the importance of continued engagement cardiovascular events in patients undergoing maintenance with patients regarding optimal vascular access. Hazard ratios are subdistribution hazard ratios from competing risks regression. 305 Figure 3. There are multiple potential mechanisms by which The results of our study support current guidelines that CVCs may lead to adverse clinical outcomes. catheter tips are associated with cardiac arrhythmias as ated with CVCs among patients undergoing daily or noc. or access-related hospitalization. . establish and maintain AV access.431). The propensity score–matched cohort included 1052 patients (CVC. AV access. CVCs may also predispose to chronic systemic in. well as such catastrophic complications as cardiac perfora- turnal HD compared with those undergoing in-center HD tion and tamponade (31). presenting a risk at home. dialysis facility home hemodialysis experience. n=526. A CVC repre. The reference group was arteriovenous (AV) access. | Association of central venous catheter (CVC) use with risk for all-cause mortality. and transfer to in-center hemodialysis in patients undergoing home hemodialysis. diabetes. serum creatinine. serum albumin. and blood hemoglobin.30). Rivara et al. Hazard ratios are sub- distribution hazard ratios from competing risks regression. remained on home HD for at least 1 year and initially started emia. which in turn.29).862 patients (CVC. sex. malpositioned the initial transition to maintenance dialysis. Results are presented for the propensity score–matched cohort of 44. The fully adjusted model was further adjusted for duration of dialysis treatment before the start of home hemodialysis. Finally. regardless encourage patients and their providers to work diligently to of whether used for in-center or home HD. although rare. The minimally adjusted model was adjusted for age. n=526).Clin J Am Soc Nephrol 11: 298–307. hospitalization. including when dialyzing sents a foreign body in the vascular space. and race and/or ethnicity. n=22. The reference group was arteriovenous (AV) access. 2016 Vascular Access and Outcomes in Home Hemodialysis. administration.431. the unmatched cohort included 2481 patients.

M.400 facilities across . Mylonakis E: 2004: Global overview of patient numbers. Chan C. Nissenson A. Kim SJ. Quinn RR. This work was supported by National Institutes of Health/ Multivariate Behav Res 46: 399–424. J Am Soc Institute of Diabetes and Digestive and Kidney Diseases. access and hospitalization risk in long-term hemodialysis tions between vascular access type and outcomes. 16: 1449–1455. Fine JP. provides Shaffer R. 17. Sands JJ. 1. Nephrol Dial Transplant 20: 2587–2593. Ravani P. James SH. El Segundo. Outcomes in Caring for ESRD (CHOICE) Study. Lockridge R. 2010 access-related hospitalizations. Eustace JA. James MT: Associations between hemodialysis States. Rivara MB. Hemodial Int 14: chronic hemodialysis. J Am that we observed vary in patient subgroups. Saito A. Taub K: Why patients cohort. Blagg C. Pisoni RL. in a large. and R01DK095668 (to K. JAMA Pediatr 167: 656–661. Ing TS: Solute kinetics with short-daily home resistant gram-negative bacteria among patients who require hemodialysis using slow dialysate flow rate. a statistical technique shown to reduce bias from with ESRD do not select self-care dialysis as a treatment option. J Am Soc care program. Himmelfarb J. Rayner HC. J Am Stat Assoc 94: 496–509. patients: A meta-analysis of prevalence. Saran R. and R. 2013 modeled vascular access type as a time-varying exposure 10. 2012 formed the primary analysis in a propensity score–matched 8. Powe NR: Vascular account for whether patients used high–frequency. ASN Dialysis Advisory Group: How to substantial power to detect differences in outcomes among overcome barriers and establish a successful home HD program. Kohn OF. J Am Soc Nephrol Despite its strengths.). Kimata N. Finkelstein F. Ravel V. National access type and clinical outcomes: A systematic review. Ornt DB. Coresh J. Ting GO. Rocco MV. Klag MJ. Curran SP: Intensive home haemodialysis: patients in . Kliger the primary outcomes through application of competing risk AS. Pastan S. J Am Soc Nephrol 24: 498–505. inflammatory markers. Eggers P. Powe NR. Centers for Medicare & Medicaid Services (CMS). Kerr PG: Vascular events limited power to examine whether the associations access and all-cause mortality: A propensity score analysis. end-stage renal disease prospective payment sys. Arrington CJ.B. Kalantar-Zadeh K. Molnar MZ. Curr Opin Nephrol Hypertens 23: 586–591.. Lacson E Jr. D’Agata EMC: Multidrug- 5. Port tions. Astor BC. Hakim RM: Change in vascular hospitalizations. Soc Nephrol 15: 477–486. Moeller S. Furthermore. Mehrotra R: Predictors of treatment with dialysis modalities in observational studies for comparative associated with use of CVCs among patients undergoing effectiveness research. Second. risk factors. Atkins RC. countries participating in DOPPS: An instrumental variable ally. 2009 12. Fed Regist 75: 49029–49214. 2014 tem. 1999 K. Chan CT. Nephrol 25: 2131–2141. McClellan WM: Vascular access and increased risk of death among hemodialysis patients. Palmer SC. Tennankore KK. 21. Cummings P: Missing data and multiple imputation. 2008 39–46. de Sa E. Nephrol Dial Transplant 30: 1208–1217. Mortis G. 2014 with an AV access was associated with higher risk for death 16. Moist LM: Association between vascular access type and patient mortality among Disclosures elderly patients on hemodialysis in Canada. Hayes WN. Zacharioudakis IM. and significance.. Am J Kidney Dis 41: 380–385. MacRae JM. Clin J Am Soc Nephrol 7: 2023–2032. Tai DJ.-Z. Austin PC: An introduction to propensity score methods for Acknowledgments reducing the effects of confounding in observational studies. Arah OA. Hemmelgarn BR. 2002 References 22. Mehrotra R: The changing landscape of home dialysis in 24. there FK: Facility hemodialysis vascular access use and mortality in remains the possibility of residual confounding.). Lazarus JM. We were also not able to 13. 2013 18. use of a CVC compared Hemodial Int 18: 481–487. 2003 nonrandom exposure assignment (32–34).K. Am J Kidney chines or whether treatments were administered during Dis 64: 954–961. Wolfe RA.K. 2014 Meta-analysis of methicillin-resistant Staphylococcus aureus 4. treatment modalities Vancomycin-resistant enterococci colonization among dialysis and associated trends.40 states. An Pannu NI.Z. Tennankore K. Clin J Am Soc Nephrol 3: 752–758. 2005 Am J Kidney Dis 65: 88–97. Pop-Vicas A. 2005 11.. Bethesda. Nat Rev Nephrol 8: 515–522. we observed similar results when we access complications with frequent hemodialysis. Stanley K. Zhang JC. Clin J Am Soc Nephrol 5: 1996–2003. we used data from 6. 2004 15.). Coresh J. Hemodial Int 18: 616–624. This approach Benefits and barriers. Ziakas PD. United States Renal Data System: USRDS Annual Data Report. Gioberge S. Na Y. our study 9. McLaughlin K. Zervou FN. Larive B. Cheung AK. Zervou FN. CA. 2014 A. subdistribution of a competing risk. Golper T.M. Future studies should investigate Brunelli S. Katz R. Frequent Hemodialysis Network Trial Group: Risk of vascular regression (35). and R. 20. Sherer S. such as patients. Soucie JM. an CHOICE Study: Type of vascular access and survival among important consideration to ensure accuracy in attribution of incident hemodialysis patients: The Choices for Healthy risk to vascular access type. Chan CT: Vascular In conclusion. Fink NE.N. 2010 25. given the observational nature of our study.R. First. nationally representative cohort access-related infection in nocturnal home hemodialysis.-Z. McDonald SP. Wang W. is an employee of DaVita. enhances external validity of our findings and also. MD. Krishnan M. Fourth. patients with differing vascular access. we per. 2014 Nephrol 24: 465–473. 2014 the day or at night. Albert JM. Strippoli GFM. or a baseline exposure at the time of start of home HD. Inc. and mortality in incident hemodialysis patients: The Choices for Healthy Outcomes in dialysate volume equipment or conventional HD ma- Caring for End-Stage Renal Disease (CHOICE) Study. Kuttykrishnan S. home HD or whether there are key subgroups among 2015 whom this higher risk does not exist. Mylonakis E: the United States. Strom J. Kidney Int 62: 620–626. 2015 3. low– access type. 2013 2. the overall number of 14. Brown G: ESRD patients in 23. whether interventions exist that may reduce the risk Streja E. our study has several limitations. Rice LB. Manns B. Mehrotra R. 2012 7. Astor B.306 Clinical Journal of the American Society of Nephrology Our study has multiple strengths. Ziakas PD. Suri RS. R21AG047306 (to M. Gillespie B. Zacharioudakis IM. Coe FL. Addition. Am J Kidney Dis 53: 475–491. Young BA. Gray RJ: A proportional hazards model for the Grants T32DK007467 (supporting M. Lockridge RS. Although we identified a number of important associa. limiting our ability to determine associa. 2011 National Institute of Diabetes and Digestive and Kidney Diseases 19. of patients on incident home HD. Gassman J. Third. Al-Jaishi AA. colonization and risk of infection in dialysis patients. 2010 . Tonelli Overview of the Epidemiology of Kidney Disease in the United M. Grassmann A. and hospitalization. HHS: Medi. Craig JC. Manns B. Polkinghorne KR. Shafi T. National Institutes of Health. Banerjee T. Oliver MJ. data were not available on cause-specific mortality or analysis. Final rule. accounts for important inter–related competing risks among Lindsay RM.. Battistella M.M.R. Hons R. Ethier J. Heagerty PJ. Thomas C.

2015 KJ: When do we need competing risks methods for survival analysis 28. “The Burden of Harm—What Is the Ideal controversy. 2008 31. Davies-Cole JO. d’Gama C. 30. 2012 versus peritoneal dialysis: A propensity score approach. 2016 Vascular Access and Outcomes in Home Hemodialysis.06570615/-/ pensity analysis of late versus early nephrologist referral DCSupplemental. van Stralen KJ. Semin Dial 21: 401–404.. Ikizler TA: Hemodialysis central venous catheters as a source of inflammation and its implications. Lok CE. Jager 65: 116–121. Winkelmayer WC. Chai SJ. Kurkjian KM. and mortality on dialysis. Mittleman MA. Avorn J: A pro. Tennankore KK. Levin R. Rubin DB: Estimating causal effects from large data sets using propensity scores. Perl J. Chan CT: Central venous catheter outcomes in in nephrology? Nephrol Dial Transplant 28: 2670–2677. Leffondré K. Schlossberg DS. Hung AM. Noordzij M. Kortas C. . asnjournals. Published online ahead of print. Rivara et al. Vincenti DM. and mortality results. 2003 cjasn. February. 33. 2015 Nocturnal Hemodialysis Study: The London Daily/Nocturnal Hemodialysis Study– design. Am J Kidney Dis 42[1 Suppl]: 5–12. Lucero CA. Fink JC. 2003 Vascular Access for Home Hemodialysis?. Owen WF Jr. Faratro R. London Daily/ Received: June 17. 32.Clin J Am Soc Nephrol 11: 298–307. J Am Soc Nephrol 14: 486–492. 307 26.” on pages 205–206. Ann Intern Med 127: 757–763. 2013 nocturnal hemodialysis. Zoccali C. J Am Soc 27. 2002 Adverse technical events in home hemodialysis. Lindsay RM. 1997 This article contains supplemental material online at http://cjasn. Patel PR: Vascular access hemorrhages 34. Vesely TM: Central venous catheter tip position: A continuing See related editorial.org. Leitch R.org/lookup/suppl/doi:10. 2015 Accepted: October 14. Chan CT: Nephrol 13: 2353–2362. Heidenheim AP. J Vasc Interv Radiol 14: 527–534. Dekker FW. contribute to deaths among hemodialysis patients. Kidney Int 70: 1348–1354. Am J Kidney Dis 35. Pliskin JS. Wong E. Ellingson KD. morbidity. Arduino MJ. Glynn RJ.2215/CJN. Palekar RS. 2006 29. Fung S. Publication date available at www. Levin R. Kidney Int 82: Avorn J: Comparing mortality of elderly patients on hemodialysis 686–692. Magri 2003 JM. Winkelmayer WC.