Early Steps to Kidney Transplantation among Persons with HIV and End-Stage Renal Disease in ESRD Network 6

Abstract

End-stage renal disease is a significant cause of morbidity and mortality in persons with HIV (PWH). Limited data exist on access to kidney transplantation for this population. A dataset inclusive of incident dialysis patients between 2012 and 2016 with follow-up through December 2017 that identifies PWH and the general dialysis population of Network 6 (Georgia, North Carolina, South Carolina), was created through merging the United States Renal Data System with the Southeastern Early Transplant Access Registry. Early steps to kidney transplantation and patient and dialysis facility-level characteristics that serve as barriers to transplantation were described. 23,414 patients were identified; 469 were PWH. Compared to non-HIV individuals, PWH were younger (49 vs 58 years, P<0.001), predominantly Black (87% vs 56% P<0.001) and male (72% vs 56% P<0.001). PWH were less likely to be referred to kidney transplant within 1 year of starting dialysis (36% vs 41% P<0.001) and waitlisted within one year of evaluation-start (14% vs 30%, P=0.05). PWH (vs. non-PWH) waited longer for referral, evaluation-start, and waitlisting and in multivariable analysis, HIV positivity was associated with a lower probability of referral (HR: 0.70; 95% CI: 0.62 – 0.80), evaluation (HR 0.66; 95% CI: 0.55 – 0.80) and waitlisting (HR 0.29; 95% CI: 0.20 – 0.41). Targeted interventions are needed to improve access to kidney transplants, particularly in waitlisting, for PWH.

Graphical Abstract

Introduction

Despite advances in HIV care and management, HIV infection continues to be an epidemic in the United States (US) with 38,739 new persons diagnosed with HIV in 2017.¹ Southern states accounted for 52% of new diagnoses with the highest rates reported in Georgia.² Unfortunately, the southern US also carries the highest burden of end-stage renal disease (ESRD). The Medicare ESRD Program is a national health insurance program for people with ESRD that consistents of 18 Network Organizations. ESRD Network 6 (GA, NC, and SC) has the second-highest rate of renal failure with 1,497 dialysis patients per million population.³ ESRD has emerged as a significant cause of morbidity and mortality,⁴ with PWH making up approximately 1.5% of the dialysis population and though having lower one- and five-year survival on dialysis.^5–7

Since a 2005 change in reporting, the true incidence and prevalence of ESRD among PWH requiring dialysis is largely unknown.^8,9 Upon initiating an ESRD service, the Centers for Medicare and Medicaid Services (CMS) Medical evidence form (CMS #2728), is completed documenting medical co-morbidities and other clinical data of the patient.¹⁰ In 2005, HIV serostatus was removed as a co-morbid condition from the CMS-2728 form, presumably due to concerns regarding disclosure of HIV status.⁸ This, coupled with the lack of national reporting before waitlisting, has limited accurate assessment of access to kidney transplant among PWH.

To improve survival on dialysis and access to kidney transplant, patients ideally should receive nephrology care prior to initiating dialysis; with the greatest benefit being at least 4 months prior to dialysis start.^11,12 The multistep process of kidney transplantation then typically starts with a referral from the dialysis facility. Once a patient is referred to kidney transplantation, they are evaluated and waitlisted (if medically eligible) by the transplant center, with hopes to eventually receive a kidney transplant. Kidney transplantation is the optimal treatment for ESRD patients, regardless of HIV status.^13–16 Despite evidence that transplanted PWH have similar rates of patient and graft survival as non-HIV patients,^17–19 PWH are less likely to be waitlisted for kidney transplant.²⁰ Reasons for this reduced access are unclear, as national surveillance data does not report information on steps before waitlisting. It is critical to investigate disparities in these early steps of kidney transplantation in order to design interventions and policy initiatives aimed to improve access to kidney transplantation in the HIV-ESRD population. In 2013, the HIV Organ Policy Equity (HOPE) Act was signed into law allowing for PWH to donate organs to other PWH, and it is estimated that 600 HIV-positive donors could save more than 1,000 lives of PWH.²¹.

Using a novel state-level HIV-ESRD dataset, this study is the first to describe referral patterns for kidney transplants among PWH. The objective of this study was to identify and describe the early steps of kidney transplantation (referral, evaluation-start, and waitlisting) for PWH and the non-HIV dialysis population in ESRD Network 6, the region with the lowest rates of kidney transplantation in the nation. Additionally, we sought to highlight patient and dialysis level characteristics that may influence access to kidney transplantation.

Materials and Methods

Study Design:

We conducted a retrospective analysis of all incident ESRD patients registered with the United States Renal Data System (USRDS) in GA, NC, and SC between 2012 and 2016, with follow-up data on outcomes through 12/31/2017. USRDS data were linked with the Southeastern Early Transplant Access Registry to obtain referral and evaluation-start dates. This study was approved by the Emory University Institutional Review Board (IRB IRB00109997).

Study Population:

All patients in ESRD Network 6 receiving dialysis between January 1, 2012, and December 31, 2016, were considered for inclusion. Patients were assigned to the first facility that provided dialysis services. Patients who were missing unique identifiers for the merging of datasets and were not adult persons between 18 and 70 years old were excluded. Those were who did not have Medicare Part D and were referred, evaluated, or waitlisted prior to dialysis start (i.e. preemptively referred, evaluated, or waitlisted) were also excluded.

Variables and Outcomes:

This study included two primary analytic approaches: (i) descriptive analysis comparing kidney transplant referral, evaluation, and waitlisting between PWH and HIV-negative persons on dialysis in Network 6 and (ii) Cox regression and time-to-event analysis, describing patient-level and dialisys facility factors associated with probability of (a) referral, (b) evaluation, and (c) waitlisting.

For the descriptive analysis, referral was examined in two ways: overall referral (transplant center receipt of the referral form at any time during the study period) and referral within one year (referral within one year of dialysis start). This time frame was selected as dialysis facilities are required to educate ESRD patients about transplant within 60 days of dialysis start.²²

Among those who were referred, evaluation-start was similarly examined in two ways: overall evaluated (having an evaluation-start date at any point during the study period) and evaluation within 6 months of referral date. Evaluation-start was defined as the date when a patient physically initiated a required component of the transplant evaluation which included either the first visit to the transplant center, visit to a satellite clinic, or attendance at a required transplant education course among those referred.²² Evaluation within 6 months of a patient’s first referral date was chosen because the median time from referral to evaluation-start has been identified in prior studies as ~3 months.²³

Waitlisting, among those who were evaluated, was also examined as overall waitlisted (being added to the national deceased donor waitlist at any point during the study period) and waitlisted within one year of the evaluation-start date. Waitlisting within one year of evaluation-start date was chosen because based on prior works the median time to waitlisting was 282 days.²⁴ Although patients can be referred, evaluated, and waitlisted more than once (to the same center or a different center), we restricted the analyses to the first referral, evaluation, and waitlisting event in the study period. There were 369 patients who were defined as being waitlisted but lacked a referral date and less than 11 patients who were defined as being evaluated though lacked a referral date. These patients were treated as being referred given referral has to occur before evaluation and waitlisting and were given a referral date of the day prior to waitlisting.

Patient characteristics included demographic and clinical data reported by clinicians from the CMS #2728 form at the time of dialysis start. Clinical data comprised of age, sex, race/ethnicity, attributed cause of ESRD (based on ICD9 and ICD10 codes), and characteristics potentially affecting the eligibility of transplant (body mass index dichotomized at >/≤35 kg/m2, tobacco use, cancer, chronic obstructive pulmonary disease, and other co-morbid conditions). Facility-level characteristics examined profit status, hospital-based vs free-standing facility, number of patients within a facility, and the ratio of patients to full-time social workers within a facility. Variables that were not associated with the exposure (HIV status) and outcome (referral, evaluation-start, or waitlisting) were not included in the Cox regression model.

Data Sources:

The creation process for this dataset has previously been described.^22,25 Briefly, the Early Transplant Access registry was created through collecting patient-level referral data from all nine adult transplantation centers in Georgia: Augusta University Medical Center (Augusta, GA), Emory Hospital (Atlanta, GA), Piedmont Hospital (Atlanta, GA), Carolinas Medical Center (Charlotte, NC), Duke Transplant Center (Durham, NC), University of North Carolina Hospital (Chapel Hill, NC), Wake Forest Baptist Hospital (Winston Salem, NC), Vidant Medical Center (Greenville, NC), and Medical University of South Carolina (Charleston, SC). Referral data were securely sent to ESRD Network 6, which served as the data coordinating rcenter. The merging of these two datasets allows for identification of those who were referred, evaluated, and waitlisted to kidney transplantation and the subpopulation that were not.

Data are extracted from the medical evidence form, deidentified, and pre-linked with United Network for Organ Sharing information on waitlisting and transplantation events.²⁶ The CMS-2728 form provides evidence of an ESRD condition, registers patients to USRDS, and documents clinical data related to the patient’s comorbidities.¹⁰ Facility characteristics were obtained from the Dialysis Facility Report (DFR). The DFR is publicly available data that includes treatment patterns, hospitalization, mortality, and transplantation patterns of Medicare-certified dialysis facilities in the U.S.²⁷ Characteristics of the patients’ residential neighborhoods, as defined by patient 5-digit zip code tabulation area, were obtained from the 2010–2014 American Community Survey²⁸ and linked by patient zip code at the start of dialysis to USRDS. To identify PWH, we used Medicare Part D Prescription Claims data, a sub-database within USRDS, for persons prescribed antiretroviral medication (Supplemental Table S1). For concerns that PWH who did not use Medicare to fill prescriptions would be misclassified as HIV negative, the cohort was limited to only those who had Medicare Part D. HIV positivity was defined as either having a prescription for an antiretroviral medication during the study period or having a diagnosis of HIV Associated Nephropathy (HIVAN) as the primary cause of ESRD.

Statistical Methods:

Descriptive statistics for patient demographic and clinical factors were stratified by HIV status and examined using χ² for categorical variables and Student’s t-tests or Wilcoxon rank-sum test for continuous variables. We examined the time to each event and cumulative incidence of referral, evaluation-start, and waitlisting (censoring for death or end of the study period) by calculating the cause-specific hazard ratios (HR) and respective 95% confidence intervals (CIs) using the Aalen-Johnsen estimator.²⁹ The cumulative incidence function was compared between PWH and non-PWH at 1 year, 6 months after referral, and 1 year after evaluation-start for referral, evaluation-start, and waitlisting, respectively.^22,23 In order to assess the effect of HIV status and other patient and dialysis facility factors on time to referral, evaluation-start, and waitlisting, multivariable Cox proportional hazards models were used to evaluate the sub-distributional hazards for both outcomes. The competing risk of death was accounted for by using conditional time-dependent weights. For multivariable patient-level analyses, covariates that were either significant in bivariable analyses, clinically relevant and associated with both the exposure (HIV status) and outcomes (referral, evaluation, and waitlisting) were included in multivariable-adjusted models. The proportional hazards assumption was tested using time-dependent variables and was met for the association of HIV status with each outcome. Lastly, in sensitivity analyses, we compared those with and without medicare prescription claims, those who were and were not preemptively referred, evaluated, and waitlisted, conducted a propensity score weighted analysis, and a propensity matched analysis to account for confounding variables.

Two-sided P-values were used for all analyses and P-value ≤ 0.05 was considered statistically significant. SAS version 9.4 (SAS Institute Inc) was used for data management and statistical analysis.

Results

Study Population and Baseline Character:

A total of 50,441 patients initiated an ESRD service during the study time period (January 1, 2012, through December 31, 2016); of these 38,689 patients had Medicare Part D claims data. 28,475 were adults between 18 to 70 years old and those who were preemptively referred, evaluated, or waitlisted (n=5,061) were excluded. Upon joining the Southeastern Early Transplant Access Registry with USRDS Medicare Claims data, 309 observations were duplicated and excluded. 23,414 patients were included in the primary analysis, of which 469 were HIV positive (Figure 1).

Figure 1.

Study cohort flowchart for incident patients with ESRD in ESRD Network 6: 2012–2016, with follow-up through 12/31/2017.

Demographic and clinical data on incident dialysis patients stratified by HIV status are listed in Table 1. The median age in the study cohort was 57 (IQR 48 – 64), with PWH being significantly younger (49 vs 58, P<0.001). PWH had a higher proportion of males (69% vs 56%, P<0.001) and black race (87% vs 57%, P<0.001). HIVAN was the leading primary cause of ESRD in PWH compared with diabetic nephropathy for HIV negative persons (P<0.001). HIV negative persons had a higher percentage of BMI and all recorded co-morbid conditions (P<0.001) except for tobacco use. Regarding markers of socioeconomic status, PWH were significantly less likely to receive pre-ESRD care (47% vs 61%, P<0.001), less likely to be informed of transplant options (84% vs 88%, P=0.003), more likely to reside in an urban zip code (27% vs 16%, P<0.001), more likely to have Medicaid insurance (45% vs 26%, P<0.001) or no insurance upon starting dialysis (20% vs 14%, P<0.001) and more likely to live in high poverty neighborhoods (44% vs 32%, P<0.001). There were no significant differences in the type of dialysis facility in which patients were treated, though the general dialysis population was more likely to be treated in a for-profit vs. non-profit facility (86% vs 80%, P <0.001), less likely to be in a large dialysis facility (dialysis facilities with > 75 people, 57% vs 60%, P<0.016) and overall less access to a full-time dialysis social worker.

Table 1.

Baseline Demographics of PWH and HIV Negative Persons with ESRD in ESRD Network 6: 2012 – 2016 at the time of ESRD start.

	Study Population (%)	HIV + (%)	HIV − (%)	P-valuea
Total Number of Patients	23,414	469 (2)	22,945 (98)
Age (IQR)	57 (48 – 64)	49 (40 – 56)	58 (48 – 64)	<0.001
Sex
Male	13,055 (56)	340 (72)	16,610 (56)	<0.001
Female	9,890 (44)	129 (27)	13,040 (44)
Race/Ethnicity
White	9,604 (41)	60 (13)	9,544 (42)	<0.001
Black	13,356 (57)	408 (87)	12,948 (56)
Other	454 (2)	<11*	453 (2)
Hispanic	716 (3)	<11*	709 (3)	0.058
Primary cause of ESRD
Diabetic Nephropathy	10,745 (46)	63 (13)	10,682 (47)	<0.001
Glomerulopathy	1,686 (7)	33 (7)	1,653 (7)
Hypertensive Nephrosclerosis	1,990 (9)	18 (4)	1,972 (9)
HIV Associated Nephropathy	243 (1)	243 (52)	-
Other	8,750 (37)	112 (24)	8,638 (38)
Facility % of Incident Patient Clinical and Laboratory Measures
BMI ≥ 35 (kg/m2)(IQR)	6,541 (28)	54 (12)	6,487(28)	<0.001
Received nephrology care prior to dialysis b
Yes	14,214 (61)	221 (47)	13,993 (61)	<0.001
No	6,443 (28)	188(40)	6,255 (27)
Unknown	2,755 (12)	60 (13)	2,695 (12)
Mode of dialysis b
Hemodialysis	21,096 (90)	437 (93)	20,659 (90)	0.196
Peritoneal Dialysis	2,306 (10)	32 (7)	2,274 (10)
Co-Morbid Conditions b
Hypertension	20,879 (89)	385 (82)	20, 494 (89)	<0.001
Congestive Heart Failure	5,992 (26)	57 (12)	5,935 (27)	<0.001
Atherosclerotic Heart Disease	1,946 (8)	15 (3)	1,931 (8)	<0.001
Cerebrovascular Disease	2,039 (9)	20 (4)	2,019 (9)	0.006
Peripheral Vascular Disease	1,884 (8)	19 (4)	1,865 (8)	0.001
Chronic obstructive pulmonary disease	1,846 (8)	22 (5)	1,824 (8)	0.010
Cancer	1,135 (5)	<11*	1,129 (5)	0.003
Tobacco use	2,444 (10)	81 (17)	2,363 (10)	<0.001
Diabetes
On Insulin	10,402 (44)	78 (17)	10,324 (45)	<0.001
On oral medications	2,255 (10)	19 (4)	2,236 (10)	<0.001
Without medications	1,493 (6)	22 (5)	1,471 (6)	0.131
Other Cardiovascular Disease	3,625 (15)	43 (9)	3,582 (16)	0.001
Patient incident year, No. of incident patients, median (IQR)
Year of Incident ESRD
2012	4,948 (21)	129 (28)	4,819 (21)	0.010
2013	4,871 (21)	102 (22)	4,769 (21)
2014	4,566 (20)	97 (21)	4,469 (19)
2015	4,595 (20)	73 (16)	4,522 (20)
2016	4,434 (20)	68 (17)	4,366 (19)
Facility % of Incident Patient Socioeconomic Characteristics
Primary Health Insurance Provider
Medicare	12,050 (51)	220 (47)	11,830 (52)	0.046
Medicaid	6,129 (26)	209 (45)	5,920 (26)	<0.001
Private	5,098 (22)	59 (13)	5,039 (22)	<0.001
Other Insurance	2,575 (11)	16 (3)	2,559 (11)	<0.001
No insurance	3,384 (14)	94 (20)	3,290 (14)	0.005
Neighborhood poverty (% of zip code residents below poverty)
0%–19.9% below poverty	15,856 (68)	264 (56)	15,592 (68)	<0.001
≥20% below poverty	7,558 (32)	205 (44)	7,353 (32)
Median % Black (IQR)	32 (16 – 50)	41 (23 – 63)	32 (16 – 50)	<0.001
Median % High School Graduates (IQR)	83 (78 – 88)	83 (78 – 89)	82 (78 – 88)	0.169
Zip Code Type c
Urban	3,745 (16)	123 (27)	3,622 (16)	<0.001
Suburban	4,272 (19)	106 (23)	4,166 (18)
Rural	14,865 (64)	233 (50)	14,632 (65)
Town	202 (1)	<11*	201 (1)
Dialysis Facility Characteristics, No. (%)
Profit status d
For-Profit	20,024 (86)	373 (80)	19,651 (86)	0.009
Non-profit	3,342 (14)	95 (20)	3,247 (14)
Type of Facility d
Free-standing	22,577 (97)	449 (96)	22,900 (97)	0.415
Hospital-based	791 (3)	19 (4)	72 (3)
Facility size (No. of patients) e
≤ 25	928 (4)	13 (3)	915 (4)	0.016
25–54	4,219 (18)	62 (13)	4,157 (18)
55–75	4,970 (21)	112 (24)	4,858 (21)
> 75	13,297 (57)	282 (60)	13,015 (57)
Social work on full time staff d
0	6,879 (29)	118 (25)	6,761 (29)	<0.001
1	14,422 (62)	277 (59)	14,145 (62)
2	1,782 (8)	53 (11)	1,729 (8)
3	124 (0.5)	<11*	120 (0.5)
4	154 (1)	16 (3)	138 (1)

Abbreviations: BMI: Body Mass Index; ESRD: End-Stage Renal Disease; IQR: Interquartile range; PWH: Persons with HIV

^aP-Values compare PWH and HIV negative persons. This is an unadjusted comparison.

^b2 patients (0.001%) were missing information on Pre-ESRD care; 4 (0.02%) patients were missing information on Mode of dialysis; 3 patients (0.001%) were missing information on Co-morbid conditions.

^c330 patients (1.4%) were missing information on Zip code type.

^d46 patients (0.2%) were missing information on Profit status; 46 patients (0.2%) were missing information on Type of facility; 52 patients (0.2%) were missing information on Social work on full time staff.

^eTotal number of patients in a facility was obtained from year 2016.

^*In compliance with CMS privacy policies, all aggregate counts of 10 or less must be suppressed, thus indicated as <11.

Referral, Evaluation, and Waitlisting:

Among the 23,414 patients, 12,145 (52%) were referred to kidney transplantation, including 9,629 (41%) within the first year of dialysis start (Table 2). There were no significant differences in referral over the entire study period among PWH and HIV negative counterparts (52% vs 52%, P=0.274), though PWH were significantly less likely to be referred within 1 year of dialysis start (36% vs 41%, P<0.001). Out of those referred, there were no significant differences between PWH and the non-HIV population for evaluation-start (50% vs 56% P=0.369, representing 27% and 25% of the cohort respectively). There were also no differences noted in evaluation-start within 6 months of referral between the two populations (45% vs 50%, P=0.372, representing 25% and 26% of the cohort respectively). In terms of waitlisting, out of those who were evaluated 4,061 patients (representing 17% of the cohort or 34% of those evaluated) were placed on the deceased donor waiting list, with a higher percentage of non-HIV persons waitlisted compared with PWH (34% vs 20%, P=0.002). Furthermore, PWH were less likely to be waitlisted within 1 year of evaluation-start (35% vs 49%, P=0.001).

Table 2.

Proportions of Referral, Evaluation, Waitlisting and Death in ESRD Network 6: 2012 – 2016 with follow up through 12/31/2017.

	Entire Cohort (%)	HIV + (%)	HIV − (%)	P-valuea
	23,414	469 (2)	22,945 (98)
Patient informed of transplant options b
Yes	20,743 (89)	396 (84)	20,347 (88)	0.003
Overall referral rate	12,145 (52)	255 (52)	11,890 (52)	0.274
Referred within one year of dialysis start	9,629 (41)	171 (36)	9,458 (41)	<0.001
Overall evaluation rate c	6,824 (56)	128 (50)	6,696 (56)	0.372
Evaluated within 6 months of referral	6,114 (50)	115 (45)	5,999 (50)	0.670
Overall waitlisting rate c	4,061 (34)	51 (20)	4,010 (34)	0.002
Waitlisted within 1 year of evaluation-start	1,994 (29)	18 (14)	1,976 (30)	0.001
Median time to referral (in months, IQR)	4.4 (2.1 – 10.0)	5.7 (2.9 – 16)	4.4 (2.1 – 9.9)	<0.001
Median time to evaluation (in months, IQR)	7.5 (4.7 – 13.7)	9.0 (5.2 – 22.0)	7.5 (4.7 – 13.6)	0.002
Median time to waitlisting (in months, IQR)	12.8 (8.6 – 20.3)	18.7 (11.0 – 30.8)	12.8 (8.6 – 20.2)	0.016
Total died during study period	9,095 (39)	184 (39)	8,911 (39)	0.862
Died without a referral	6,000 (26)	116 (25)	5,884 (26)	0.654
Median time to death (in months, IQR)	22.0 (9.9 – 37.3)	22.2 (11.3 – 37.3)	22.0 (9.8 – 37.3)	0.659

Abbreviations: ESRD: End-Stage Renal Disease; IQR: Interquartile range

^aP-Values compare PWH and HIV negative persons.

^b40 patients (0.2%) were missing information on Patient informed of transplant options.

^CDenominator used for evaluation was of those referred and the denominator for waitlisting was of those evaluated. Percentage listed use the respective denominators.

The median time to referral in the study cohort was 4.4 months (IQR: 2.1 – 10.0 months). Compared with non-HIV individuals, PWH waited significantly longer for a referral to kidney transplantation (5.7 months vs. 4.4 months, P<0.001). An equivalent trend was observed for time to evaluation-start and waitlisting in PWH vs. non-HIV (9.0 months vs 7.5 months for evaluation-start, P=0.002, and 18.7 months vs 12.8 months for waitlisting, P=0.016). During the study period, 39% of the study population died prior to receiving a kidney transplant, 26% of whom died prior to referral for kidney transplantation. The median time to death was 22 months and no differences were observed between PWH and HIV negative individuals for percetange of death and median time to death.

Figure 2a–c demonstrates the cumulative incidence of time to referral, evaluation-start, and waitlisting comparing PWH and HIV negative persons censored by death or end of the study, whichever occurred first. For the cumulative incidence of time to referral, referral within one year from dialysis start was considered to be an appropriately timed referral. There was not significant between PWH and HIV negative persons at 1 year, (cumulative incidence 0.36; 95% CI: 0.32 – 0.40 vs cumulative incidence 0.41; 95% CI: 0.40 – 0.41, P=0.07 ), and no difference by the end of the study period (P=0.196). Of the individuals who were referred, initiating a transplant evaluation within 6 months of referral was considered to be an appropriately timed evaluation-start. There were no differences noted in the cumulative incidence of time to evaluation-start at 6 months after referral (cumulative incidence 0.50; 95% CI: 0.43 – 0.56 vs cumulative incidence 0.55; 95% CI: 0.54 – 0.56. P=0.05), or throughout the study period (P=0.10). Regarding waitlisting among those who were evaluated, waitlisting within one year of evaluation-start was considered to be appropriate timing of waitlisting. There was an early separation between PWH and HIV negative counterparts that was maintained at 1 year and throughout the study (cumulative incidence 0.14; 95% CI: 0.08 – 0.22 vs cumulative incidence 0.34; 95% CI: 0.33 – 0.35, P<0.001).

Figure 2a-c.

Cumulative incidence of referral, evaluation, and waitlisting comparing PWH to HIV negative persons between 2012 and 2016 with follow up through 12/31/2017. A. Cumulative incidence of referral at 1 year was 41% (95% CI 0.40 – 0.42) for HIV negative persons and 36% (95% CI 0.32 – 0.40) for PWH. B. Cumulative incidence of evaluation at 6 months was 55% (95% CI 0.54 – 0.56) for HIV negative persons and 50% (95% CI 0.43 – 0.56) for PWH. C. Cumulative incidence of waitlisting at 1 year was 34% (95% CI 0.33 – 0.35) for HIV negative persons and 14% (95% CI 0.08 – 0.22) for PWH.

Multivariable-Adjusted Analyses

After adjusting for comorbidities and transplant factors (Table 3), HIV status was negatively associated with transplant referral (HR 0.70; 95% CI: 0.60 – 0.80), evaluation (HR 0.67; 95% CI: 0.55 – 0.81) and in particular waitlisting (HR 0.29; 95% CI: 0.21 – 0.42). Black race and Hispanic ethnicities were positively associated with referral and evaluation. Common comorbidities such as congestive heart failure, peripheral vascular disease, obesity, chronic obstructive pulmonary disease, and tobacco use were progressively, negatively associated with proceeding through each transplant step. Patient socioeconomic characteristics such as public insurance (Medicare and Medicaid) was largely associated with lower access to transplant, while suburban and rural residence were mostly associated with increased access to transplantation. The profit status of the dialysis facility was the only dialysis facility factor associated with a lower probability of referral and evaluation-start, however hospital based facilities were associated with increased referral and waitlisting.

Table 3.

Multivariable-Adjusted Cox regression model for the relationship between demographic and clinical variables with referral, evaluation, and waitlisting: 2012–2016^a with follow up through 12/31/2017

	Time to Referral		Time to Evaluation		Time to Waitlisting
Variable	Adjusted HR (95% CI)	P-Value	Adjusted HR (95% CI)	P-Value	Adjusted HR (95% CI)	P-Value
Patient Related Characteristics
HIV status
HIV −	Ref		Ref		Ref
HIV +	0.70 (0.62 – 0.80)	<0.001	0.66 (0.55 – 0.80)	<0.001	0.29 (0.20 – 0.41)	<0.001
Age	0.98 (0.98 – 0.98)	<0.001	0.98 (0.97 – 0.98)	<0.001	0.96 (0.96 – 0.96)	<0.001
Sex
Female	Ref		Ref		Ref
Male	1.09 (1.04 – 1.13)	<0.001	1.06 (1.00 – 1.12)	0.035	1.05 (0.97 – 1.13)	0.234
Race
White	Ref		Ref		Ref
Black	1.23 (1.18 – 1.29)	<0.001	1.12 (1.05 – 1.19)	0.001	0.97 (0.89 – 1.05)	0.452
Other	1.22 (1.07 – 1.40)	0.003	1.20 (1.00 – 1.44)	0.043	1.29 (1.04 – 1.60)	0.023
Hispanic	1.14 (1.03 – 1.28)	0.016	1.22 (1.06 – 1.40)	0.007	1.24 (1.04 – 1.47)	0.017
BMI ≥ 35 (kg/m2)	0.90 (0.86 – 0.94)	<0.001	0.78 (0.73 – 0.83)	<0.001	0.52 (0.47 – 0.57)	<0.001
Received Pre-ESRD care
No	Ref		Ref		Ref
Yes	1.17 (1.12 – 1.23)	<0.001	1.18 (1.11 – 1.25)	<0.001	1.27 (1.17 – 1.38)	<0.001
Comorbid Conditions
Hypertension	1.19 (1.11 – 1.27)	<0.001	1.16 (1.06 – 1.27)	0.001	1.22 (1.08 – 1.38)	0.002
Congestive Heart Failure	0.89 (0.85 – 0.93)	<0.001	0.84 (0.79 – 0.90)	<0.001	0.60 (0.54 – 0.67)	<0.001
Atherosclerotic Heart Disease	0.94 (0.87 – 1.02)	0.143	0.99 (0.89 – 1.12)	0.919	0.92 (0.77 – 1.10)	0.375
Cerebrovascular Disease	0.77 (0.72 – 0.84)	<0.001	0.70 (0.62 – 0.79)	<0.001	0.64 (0.53 – 0.77)	<0.001
Peripheral Vascular Disease	0.82 (0.76 – 0.90)	<0.001	0.72 (0.63 – 0.81)	<0.001	0.63 (0.51 – 0.77)	<0.001
COPD	0.77 (0.71 – 0.85)	<0.001	0.63 (0.55 – 0.74)	<0.001	0.27 (0.19 – 0.38)	<0.001
Cancer	0.54 (0.48 – 0.61)	<0.001	0.61 (0.52 – 0.73)	<0.001	0.52 (0.41 – 0.66)	<0.001
Tobacco use	0.86 (0.81 – 0.93)	<0.001	0.75 (0.68 – 0.83)	<0.001	0.45 (0.38 – 0.53)	<0.001
Diabetes
On Insulin	0.96 (0.92 – 0.997)	0.037	0.98 (0.92 – 1.03)	0.404	0.80 (0.74 – 0.87)	<0.001
On oral medications	1.04 (0.97 – 1.11)	0.311	1.00 (0.91 – 1.09)	0.954	0.91 (0.79 – 1.03)	0.138
Without medications	0.98 (0.90 – 1.06)	0.609	0.93 (0.83 – 1.04)	0.198	0.88 (0.75 – 1.03)	0.113
Patient Socioeconomic Characteristics
Primary Health Insurance Provider
Medicaid	0.89 (0.83 – 0.94)	<0.001	0.83 (0.76 – 0.90)	<0.001	0.67 (0.59 – 0.77)	<0.001
Medicare	0.86 (0.82 – 0.91)	<0.001	0.85 (0.79 – 0.91)	<0.001	0.74 (0.67 – 0.82)	<0.001
Employer	1.34 (1.26 – 1.43)	<0.001	1.43 (1.31 – 1.56)	<0.001	1.92 (1.69 – 2.18)	<0.001
Other Insurance	1.12 (1.04 – 1.21)	0.004	1.20 (1.09 – 1.33)	0.001	1.52 (1.31 – 1.76)	<0.001
No insurance	0.99 (0.92 – 1.07)	0.852	1.02 (0.92 – 1.13)	0.716	0.97 (0.84 – 1.13)	0.703
Zip code Type
Urban	Ref		Ref		Ref
Town	1.05 (0.84 – 1.32)	0.676	1.04 (0.77 – 1.42)	0.793	1.61 (1.08 – 2.38)	0.019
Rural	1.13 (1.07 – 1.19)	<0.001	1.15 (1.06 – 1.24)	0.001	1.15 (1.03 – 1.28)	0.015
Suburban	1.27 (1.19 – 1.35)	<0.001	1.40 (1.28 – 1.52)	<0.001	1.34 (1.19 – 1.51)	<0.001
Neighborhood poverty (% of zip code residents below poverty)
0%–19.9% below poverty	Ref		Ref		Ref
≥20% below poverty	0.99 (0.94 – 1.04)	0.608	0.98 (0.91 – 1.06)	0.669	0.89 (0.80 – 0.99)	0.027
Median % Black (IQR)	1.001 (1.00 – 1.002)	0.048	1.001 (1.00 – 1.003)	0.082	1.00 (0.998 – 1.001)	0.602
Median % High School Graduates (IQR)	1.006 (1.002 – 1.01)	0.004	1.005 (1.00 – 1.01)	0.043	1.01 (1.01 – 1.02)	<0.001
Dialysis Facility Characteristics
Profit status
Non-profit	Ref		Ref		Ref
For-profit	0.89 (0.84 – 0.94)	<0.001	0.81 (0.75 – 0.88)	<0.001	0.88 (0.78 – 0.98)	0.022
Type of Facility
Free-standing	Ref		Ref		Ref
Hospital-Based	2.05 (1.78 – 2.36)	<0.001	1.80 (1.50 – 2.16)	<0.001	0.98 (0.80 – 1.24)	0.881

Abbreviations: BMI: Body Mass Index; ESRD: End-Stage Renal Disease; IQR: Interquartile range; PWH: Persons with HIV

^a386 observations (1.7%) were deleted from the analysis due to missing information.

Sensitivity Analysis:

Given the final cohort consisted of only those who had a Medicare prescription claim, we compared those with Medicare prescription claims to those who did not to evaluate the difference between the two groups (Supplemental Table S2). Additionally, since we excluded those who were preemptively referred, evaluated, or waitlisted, we assessed the differences between those patients who proceeded through the transplant steps preemptively and those who did not (Supplemental Table S3). We performed a propensity score weighted analysis in attempt to account for confounding variables. age, sex, race, prior nephrology care, BMI, poverty level, Medicare, and Medicaid status. The estimated probabilities were then included in the final model (Supplemental Table S4). The results of propensity weighted analysis were consistent with findings for the primary outcomes of HIV status. HIV status remained negatively associated with referral, evaluation, and waitlisting: HR 0.72; 95% CI: 0.63 – 0.81; HR 0.67; 95% CI: 0.55 – 0.81; HR 0.30; 95% CI: 0.21 – 0.42, respectively. Lastly, we performed a propensity matched analysis using one-by-k nearest neighbor propensity matched algorithm. The most power that could be achieved while still maintaining balance of covariates was at 1:10 matching, without replacement, caliper at 0.20. This achieved a cohort of 3,010 persons. Balance of covariates was assessed graphically by evaluating the common support, pre-and post-match distribution of propensity scores and the standard means difference for variables that were dichotomous. The results of the propensity matched analysis were also consistent with the findings for the primary outcomes of HIV status (data not included).

Discussion

Although there has been increasing attention on access to kidney transplantation, little is known about the role early steps to transplantation plays in access to transplantation, as national data before waitlisting are limited. This study, which utilized novel data on early transplant access steps in GA, NC, and SC, demonstrated that PWH had a 71% lower relative risk in proceeding through the early steps of kidney transplantation compared with the general dialysis population of ESRD Network 6. Additionally, PWH waited 1.2 months longer to be referred, 1.5 months longer to be evaluated, and 4.9 months longer to be waitlisted for kidney transplantation. Factors such as older age, BMI ≥ 35 (kg/m2), comorbid conditions, public insurance, residing in a high poverty neighborhood, and for-profit dialysis facilities, were identified as barriers to referral, evaluation-start, and/or waitlisting for kidney transplantation.

In July 2019, the Advancing American Kidney Health initiative was launched, with the aim of decreasing the number of American developing kidney failureand needing dialysis, while increasing availability of kidney transplantation. While dialysis incidence decreased during the study period³⁰, one step towards advancing this goal is improving when PWH receive nephrology care. In this study, 47% of PWH compared with 61% of HIV negative persons received nephrology care prior to dialysis start (P<0.001). Lack of access to nephrology care has been associated with a lower likelihood of receiving a kidney transplant³¹ and possibility limits preemptive referrals to kidney transplantation.³² In one study, out of 42 kidney transplant eligible PWH, only two (5%) received a transplant evaluation prior to dialysis start.³³ A similar trend was observed in this study with only 8% of PWH compared with 18% of HIV negative individuals being preemptively referred to transplant (Supplementary Table S3).

With the HIV epidemic in the southeast predominately affecting young males who identify as being black, it is not surprising that the HIV population was almost 10 years younger than the general dialysis population, majority male and mostly black persons. Interestingly, Black race was associated with an increased probability of referral and evaluation-start. Patzer et al also noted in their study evaluating dialysis facility referral and evaluation-start in the Southeast, that those who identify as black were associated with a higher percentage of referral.³⁴ This finding in part could be secondary the selection bias of the black race generally being ³⁰less likely to be preemptively referred. Given that preemptive referral and waistlitng were excluded in this study, the remaining population in this study was predominately Black, making it appear as if person who identify as being Black were associated with an increased probability of referral and evaluation-start.³⁵

In Network 6, 52% of patients with a Medicare Part D claim were referred to kidney transplantation between 2012 and 2016 and overall, PWH and HIV negative persons were referred and evaluated for transplantation at similar proportions. However, PWH were less likely to be referred to kidney transplantation within one year of dialysis start, suggesting that this population may experience additional delays in the referral process. Reasons why PWH experience delays to referral are unclear, though factors identified in prior studies include lack of pre-ESRD nephrology care, not having permanent access at dialysis initiation, and not being informed of transplant options.³⁶. With no specified target or metric, it is unknown though whether all candidates suitable for kidney transplantation are being referred and evaluated. Additionally, referral itself may not be the main driver of disparities in access to kidney transplant, but rather the difficulty in navigating the complex and challenging multistep process of kidney transplantation.²⁵

Significant dropouts were observed as individuals proceeded through the transplant steps for both PWH and the HIV negative population, though PWH were less likely to be waitlisted despite being younger with fewer comorbidities. Furthermore, HIV serostatus was negatively associated with all three transplant steps. Single-center experiences have suggested uncontrolled HIV, non-compliance, substance use, insurance-related issues, inability to complete transplant requirements, incomplete records and co-infection with hepatitis C, prevent PWH from waitlisting.^37,38Transplant candidacy requirements specific to PWH include having a CD4 count >200 cells/mm³ and being virally suppressed for at least 6 months, potentionally extending the evaluation process.³⁹ Lastly, 22% of adult kidney transplant centers still consider HIV seropositivity an absolute contraindication to transplantation.⁴⁰ Unfortunately, our study lacks the specific data to elucidate how these factors affected access to kidney transplant among our HIV population. These limitations underscores the need for national reporting on early steps to transplantation and a better understanding of the potential barriers that prevent individuals from achieving the ultimate goal of transplantation.

Determination of transplant candidacy is a complex and difficult process involving patient characteristics, dialysis facility characteristics, and transplant center factors. It is important to distinguish that barriers may be different for each transplant step since referral generally occurs at the dialysis facility and evaluation-start and waitlisting at the transplant center. For example, the profit status of a dialysis facility was associated with a lower probability of referral and evaluation-start, though not associated with waitlisting, a finding that has been previously identified.⁴¹ However, markers of socioeconomic status such as having non-private insurance or coming from a high poverty neighborhood either globally decreased access to transplantation or were associated with lower waitlisting; factors that would disproportionately affect PWH.^42,43 Successfully proceeding through the transplant steps often requires knowledge of how to navigative the medical system, self-advocavy having a solid understanding of the listing process, and being aware of pre-emptive referres.⁴⁴ Understanding barriers for each transplant step will not only require national reporting on referral and evaluation, but also teasing apart the patient-level, dialysis facility-level, and transplant center-level factors at play.

This study offers the first region-level characterization of PWH with ESRD proceeding through the early steps of kidney transplantation. Prior studies have reported solely on waitlisting, but due to the unique Southeastern Early Transplant Referral Dataset, we captured referral and evaluation-start, in addition to waitlisting. We identified that waitlisting is a crucial step in access to kidney transplants among incident HIV-ESRD persons in Network 6.

The conclusions of our study is limited by a few considerations. First, prescription data only captured those who had a Medicare part D claim for antiretrovirals, possibly underestimating the HIV-ESRD population and misclassifying PWH who have private insurance or did not fill any antiretrovirals. In 2016, approximately 81% of hemodialysis patients were enrolled in Medicare Part D so we believe the majority of the HIV-ESRD population was captured.⁴⁵ Also, this study focused on incident ESRD patients and not prevalent ESRD patients, thus the entire HIV-ESRD population in Network 6 was not encapsulated.

Data from this study does not allow the assessment of the impact that psychiactric illness, social support, incomplete or missing records, and HIV viral control had on kidney transplant waitlisting among PWH. Our study lacks data on pre-transplanation HCV serostatus. With 25% of PWH being coinfected with HCV⁴⁶, some providers many have considered coinfection a relative contraindication to kidney transplant during the study period. Medications used for the treatment of HIV, such as lamivudine, tenofovir and emtricitabine, are also used for purposes such as pre-exposure prophylaxis (PREP) for HIV or treatment of hepatitis B. Patients taking medications for these indications may have been misclassified, particularly the use of lamivudine for HBV, though the use of tenofovir derivatives is unlikely to have been misclassified given low uptake of PrEP in the southeast⁴⁷ and with tenofovir requiring renal adjustments, it is rarely used for hepatitis patients with ESRD. Lastly, given that this study focused on Network 6, the findings of this study may not be generalizable to other regions.

Conclusions

In ESRD Network 6, PWH and HIV negative persons had similar rates of referral and evaluation-start for kidney transplant, though PWH were less likely to achieve waitlisting. Factors identified as being associated with a lower probability of referral, evaluation-start, and/or waitlisting included HIV positiveity, increasing age, morbid obesity, comorbid conditions, non-private insurance, and living in a high poverty neighborhood. Barriers specific to the HIV population need to be further elucidated with HIV-related clinical and laboratory data. Interventions in improving access to kidney transplantation in the Southeast should target minimizing patient dropout between transplant evaluation and waitlisting.