Journal Logo

Original Clinical Science—General

Excess Stroke Deaths in Kidney Transplant Recipients: A Retrospective Population-based Cohort Study Using Data Linkage

De La Mata, Nicole L. PhD1; Kelly, Patrick J. PhD1; Wyld, Melanie PhD2,3; Masson, Philip PhD4; Al-Shahi Salman, Rustam PhD5; Webster, Angela C. PhD1,6

Author Information
doi: 10.1097/TP.0000000000003091

Abstract

INTRODUCTION

Transplantation is the most effective treatment option for kidney failure and has been associated with better long-term outcomes in appropriately selected patients.1 While people with end-stage kidney disease are known to experience an excess risk of stroke and stroke deaths,2 kidney transplant recipients have a lower risk of cardiovascular disease compared with dialysis.3 This is in part due to improved kidney function and selection bias due to extensive cardiovascular risk screening in potential kidney transplant recipients.4,5

However, it is unclear if kidney transplant recipients have benefited from stroke prevention and management as much as the general population has over the past 20 years.6 In Australia and New Zealand, greater use of preventative treatment and better control of vascular risk factors have halved stroke mortality rates in the general population in the past decade.7,8 Yet, there is a lack of clinical data evaluating the effectiveness, safety, and tolerability of stroke prevention and interventions in people with kidney disease, whom are often excluded from randomized trials. Stroke interventions may also be underused or delayed due to drug interaction concerns from concurrent use of immunosuppressive agents.9

Identifying risk factors for stroke mortality in kidney transplant recipients may permit testing of targeted interventions aimed at reducing stroke risk. Observational studies to date have been limited by their small sample sizes and short follow-up.10-14 Although risk factors for stroke mortality in people with end-stage kidney disease in Australia and New Zealand have previously been described.15 Kidney transplant recipients differ from people on dialysis. Recipients typically still have chronic kidney insufficiency and experience novel exposures that may increase their risk of stroke, such as immunosuppression regimens16 and graft failure.17,18 Previous duration of dialysis has also been shown to increase the subsequent all-cause and cardiovascular mortality risk in kidney transplant recipients.19 People with polycystic kidney disease are also predisposed to hemorrhagic strokes and present different risk factors for stroke death.10,20

Our study objective was to compare stroke mortality among kidney transplant recipients to the general population and evaluate risk factors associated with stroke death in kidney transplant recipients in Australia and New Zealand. We will further explore whether kidney transplant recipients with polycystic kidney disease have different excess stroke deaths and risk factors for stroke death.

MATERIALS AND METHODS

Study Design and Setting

We conducted a population-based retrospective cohort study of all adult and children kidney transplant recipients in Australia (January 1, 1980, to December 31, 2013) and New Zealand (January 1, 1988, to December 31, 2012). Australia and New Zealand have broadly similar demographics, including life expectancies and racial background, and universal healthcare systems, where free medical care is provided in public health systems. All deaths within each country are reported under mandate to the Births, Deaths, and Marriages Register, and the cause of death is stated on a medical certificate, completed by a medical doctor. Both the primary and contributing causes of death are captured. The primary cause of death is defined as the initial disease or condition leading to a sequence of events resulting in death. Secondary or contributing causes of death are defined as other diseases or conditions that were not the underlying cause but contributed to the death.

Participants, Data Linkage, and Death Outcomes

Kidney Transplant Recipients

The Australian and New Zealand Dialysis and Transplant Registry (ANZDATA) collects prospective data from all dialysis and transplant centers in both Australia and New Zealand. Data collection captures key events as they occur in real time and annually for each patient. Core data include demographics (age, sex, country, racial background, height, weight, and smoking), comorbidities (cerebrovascular disease, diabetes, coronary artery disease, peripheral artery disease, and previous malignancy), and treatment-related information (type of dialysis or transplant, date initiated, and cause of kidney failure). The data collection methods and remit for ANZDATA have been described elsewhere.21

Data Linkage

We used data linkage between kidney transplant recipients in ANZDATA and the national death registers in Australia and New Zealand to determine the date and legal cause of death. In Australia, recipients were linked using probabilistic record linkage, matching on identifiers including date of birth, sex, and full name, to the National Death Index, which records all deaths from 1980. In New Zealand, recipients were linked using deterministic record linkage, matching on national health index number, to the Mortality Collection database, which records all deaths from 1988. Both countries reported causes of death using International Classification of Diseases (ICD) coding, Ninth version until 1996 in Australia and 1999 in New Zealand, thereafter using the 10th version, Australian Modification. We used one-to-one mapping files to convert between ICD-9 and ICD-10-AM codes. Our analysis is restricted to the years where death data were available in the national death registries at the time of data linkage in 2014; thus, our cohort consists of all kidney transplant recipients who received their transplant in Australia (1980–2013) and in New Zealand (1988–2012).

Ethics approval was granted for this study from the University of Sydney (Project No.: 2014/917), Australian Institute of Health and Welfare (Reference No.: EO2015/3/181), and the New Zealand Ministry of Health (Reference No.: 14/NTB/171). Data linkage was performed by Australian Institute of Health and Welfare for Australian recipients and the New Zealand Ministry of Health for New Zealand recipients, using best-practice privacy-preserving protocols. Once complete, only deidentified recipient data were made available to researchers for this study.

Death Ascertainment and Cause of Death

The primary cause of death was used to determine the stroke deaths using ICD-10-AM codes, including subarachnoid hemorrhage (I60.0 to I60.9), intracerebral hemorrhage (I61.0 to I61.9), other intracranial hemorrhages (I62.0 to I62.9), ischemic strokes (I63.0 to I63.9), and unspecified strokes (I64.0 to I64.9). All other causes of death were considered as other deaths. We further summarized stroke deaths listed in the secondary causes of death where kidney disease was the primary cause of death, including diabetes (E10 to E14), kidney failure (N17 to N19), hypertensive disease (I10 to I15), glomerular disease (N00 to N08), other kidney or ureter disorders (N25 to N29), and renal tubule-interstitial disease (N10 to N16).

Time at risk was measured from the earliest transplant date until the recipient died or December 31, 2013 (Australia only) or December 31, 2012 (New Zealand only), whichever came first. Probabilistic record linkage may lead to a small proportion of incorrect links.22 Therefore, Australian patients were censored at the ANZDATA date of death if they were considered to have died and the national death registry had not captured any death. Australian patients were also considered to be alive and censored if ANZDATA had additional clinic visits after the registered date of death.

Statistical Analyses

We summarized patient characteristics using absolute numbers and proportions. We used the Kappa statistic to evaluate the agreement of fact of death between ANZDATA and the national death registries.23 We produced mortality rates and estimated standardized mortality ratios (SMR) using indirect standardization, matching on age, sex, and calendar year, where stroke mortality rates from the general population are presented in SDC File 1(SDC, http://links.lww.com/TP/B856). Poisson regression was used to test for differences in mortality rates and SMR.

We used the Fine and Gray competing risks regression model to evaluate risk factors associated with stroke death in kidney transplant recipients.24 In models examining risk factors for stroke deaths, all nonstroke deaths were considered as competing events. For comparison, we also examined risk factors for other deaths where stroke deaths were considered as competing events. For both the stroke and nonstroke models, we used the same covariates, which were selected a priori including age at transplant, sex, year of transplant, country, racial background, previous dialysis time, comorbidities, smoking status, and cause of kidney failure. Univariable models were fitted for each covariate, and the multivariable models were fitted for all covariates. To ensure consistent findings, we also used cause-specific Cox models to evaluate risk factors from the multivariate model for stroke and other deaths using the same risk factors. We estimated the cumulative incidence of stroke and other deaths from the final multivariate models.

We performed a subpopulation analysis on kidney transplant recipients with polycystic kidney disease as the cause of kidney failure. We used the same statistical methods as described in the main analysis to estimate SMR and evaluate risk factors for stroke mortality.

There were no missing data, except for height, weight, and smoking status which was not routinely collected until after 1995. Missing values for body mass index (BMI) and smoking status were imputed using chained equations with 5 iterations, using the same explanatory variables and outcome variable from the competing risk model as well as log-transformed time at risk. BMI values were imputed using linear regression, then converted into categories. Smoking status was imputed using multinomial logistic regression.

Data were analyzed using Stata version 15 (Stata Corporation, College Station, TX). The data used in this study are available from the corresponding author upon reasonable request and subject to ethical approvals.

RESULTS

Patient Characteristics

A total of 17 628 kidney transplant recipients were included in the analysis. There were 158 stroke deaths and 5126 other deaths over 175 084.3 person-years, with a median of 8.3-year follow-up posttransplant (interquartile range: 3.7–14.7 years). The remaining 12 344 recipients (70%) were in active follow-up. Overall, 15 579 recipients were on their first kidney transplant, 1798 recipients were on their second kidney transplant, and 251 were on their third or more kidney transplant. A total of 3733 kidney transplant recipients had experienced 1 graft failure, 518 recipients experienced 2 graft failures, and 79 experienced 3 or more graft failures. The agreement for fact of death between ANZDATA and the death register was almost perfect in Australia (Kappa statistic: 0.89 [95% confidence interval (CI), 0.88-0.90]) and New Zealand (Kappa statistic: 0.93 [95% CI, 0.91-0.95]) (Table S1, SDC, http://links.lww.com/TP/B856).

In our study population, over half (60%) were aged under 50 years at transplant, with a median age of 45 years (interquartile range: 33–56; Table 1). Most transplants were in Australia (88%) and during 2000 or later (56%). Over half were male (61%) and had a normal (36%) or overweight (25%) BMI. The BMI was not collected for 17% of recipients. The main cause of kidney failure was glomerulonephritis/IgA nephropathy (44%), followed by other causes (28%) and polycystic kidney disease (12%). Reflux nephropathy (9.6%), uncertain diagnosis (4.1%), and analgesic nephropathy (3.0%) accounted for 60% of the other causes of kidney failure.

TABLE 1. - Characteristics of kidney transplant recipients by death status
Characteristics All-cause stroke deaths Other deaths Alive Total
n (%) a n (%) a n (%) a n (%) a
Total, (%) b 158 (1) 5126 (29) 12 344 (70) 17 628 (100)
Age at transplant (y)
 ≤29 14 (9) 571 (11) 2862 (23) 3447 (20)
 30–49 71 (45) 1868 (36) 5132 (42) 7071 (40)
 50–64 64 (41) 2303 (45) 3672 (30) 6039 (34)
 65–74 9 (6) 375 (7) 663 (5) 1047 (6)
 ≥75 0 (–) 9 (<0.2) 15 (<0.2) 24 (<0.2)
 Median [IQR] 49 [41–55] 50 [40–58] 43 [31–54] 45 [33–56]
Sex
 Female 78 (49) 2043 (40) 4793 (39) 6914 (39)
 Male 80 (51) 3083 (60) 7551 (61) 10 714 (61)
BMI category (kg/m2)
 Underweight (<18.5) 5 (8) 209 (7) 1021 (9) 1235 (7)
 Normal (≥18.5–<25.0) 39 (59) 1357 (45) 5117 (44) 6513 (37)
 Overweight (≥25.0 and <30.0) 17 (26) 960 (32) 3510 (30) 4487 (25)
 Obese (≥30.0) 5 (8) 507 (17) 1878 (16) 2390 (14)
 Not available 92 (–) 2093 (–) 818 (–) 3003 (17)
Previous dialysis time (mo)
 0–<6 30 (19) 704 (14) 2925 (24) 3659 (21)
 ≥6–<12 36 (23) 927 (18) 1756 (14) 2719 (15)
 ≥12–<18 19 (12) 766 (15) 1468 (12) 2253 (13)
 ≥18–<36 40 (25) 1428 (28) 2739 (22) 4207 (24)
 ≥36 33 (21) 1301 (25) 3456 (28) 4790 (27)
 Median [IQR] 16 [8–30] 19 [9–36] 18 [6–40] 19 [7–39]
Y of transplant
 ≤1990 85 (54) 2156 (42) 1138 (9) 3379 (19)
 1991–1999 58 (37) 1865 (36) 2416 (20) 4339 (25)
 2000–2009 15 (9) 993 (19) 5558 (45) 6566 (37)
 2010–2013 0 (–) 112 (2) 3232 (26) 3344 (19)
Country
 Australia 147 (93) 4587 (89) 10 742 (87) 15 476 (88)
 New Zealand 11 (7) 539 (11) 1602 (13) 2152 (12)
Racial background
 White 144 (91) 4429 (86) 10 145 (82) 14 718 (83)
 Nonwhite 14 (9) 697 (14) 2199 (18) 2910 (17)
Comorbidities
 Cerebrovascular disease 10 (6) 178 (3) 292 (2) 480 (3)
 Diabetes 23 (15) 868 (17) 1587 (13) 2478 (14)
 Coronary artery disease 9 (6) 555 (11) 910 (7) 1474 (8)
 Peripheral artery disease 7 (4) 342 (7) 516 (4) 865 (5)
 Previous malignancy 66 (42) 2495 (49) 3509 (28) 6070 (34)
Smoking status
 Current/former 31 (20) 1611 (31) 4473 (36) 6115 (35)
 Never/unknown 127 (80) 3515 (69) 7871 (64) 11 513 (65)
Cause of kidney failure
 Diabetes 17 (11) 711 (14) 1239 (10) 1967 (11)
 Hypertension/renal artery disease 7 (4) 247 (5) 489 (4) 743 (4)
 Glomerulonephritis/IgA nephropathy 60 (38) 2045 (40) 5680 (46) 7785 (44)
 Polycystic kidney disease 28 (18) 608 (12) 1560 (13) 2196 (12)
 Other 46 (29) 1515 (30) 3376 (27) 4937 (28)
aColumn percentages.
bRow percentages.
BMI, body mass index; IQR, interquartile range.

Deaths

Of the 158 stroke deaths, there were 57 (36%) intracerebral hemorrhages, 11 (7%) intracranial hemorrhages, 20 (13%) ischemic strokes, 19 (12%) subarachnoid hemorrhages, and 51 (32%) unspecified strokes. A further 76 stroke deaths were listed in the secondary causes of death where kidney disease was the primary cause of death. These included 19 (25%) intracerebral hemorrhages, 16 (21%) ischemic strokes, 4 (5%) intracranial hemorrhages, 2 (3%) subarachnoid hemorrhages, and 35 (46%) unspecified strokes. Of the 5126 other deaths, the leading underlying cause of death was coronary heart disease (n = 787), followed by diabetes (n = 440) and kidney failure (n = 427).

The cumulative incidence for stroke mortality was significantly higher in those with preexisting cerebrovascular disease compared with those without (P < 0.001). At 2 years posttransplant, it was 0.14% in those with and 0.04% in those without cerebrovascular disease (Figure 1). This increased at 5 years posttransplant to 0.49% for those with and 0.15% for those without cerebrovascular disease. The cumulative incidence was not higher in those with cerebrovascular disease for other deaths (P = 0.631).

FIGURE 1.
FIGURE 1.:
Cumulative incidence in transplant recipients by preexisting cerebrovascular disease. Estimates for (A) stroke deaths and (B) other deaths.

Mortality Rates

The overall mortality rate for stroke was 90.2 (95% CI, 77.2-105.5)/100 000 person-years, for intracerebral hemorrhage was 32.6 (95% CI, 25.1-42.2)/100 000 person-years, and for ischemic stroke was 11.4 (95% CI, 7.4-17.7)/100 000 person-years. The stroke mortality rates over time since transplant showed no distinguishable pattern during the first year (P > 0.9), unlike rates for other deaths which was highest in the first 3 months posttransplant (P < 0.001) (Figure 2). Thereafter, stroke mortality rates steadily increased over time since transplant (P < 0.001), from 47.7 (95% CI, 23.9-95.5)/100 000 person-years in the first year to 140.0 (95% CI, 77.5-252.8)/100 000 person-years in the tenth year posttransplant. Rates for other deaths decreased after the first year (P < 0.001) and steadily increased thereafter.

FIGURE 2.
FIGURE 2.:
Crude mortality rates per 100,000 person-years This is given for the first year (top) and up to 10 y (bottom) for (A) stroke deaths (left) and (B) other deaths, by time since transplant and time since graft failure.

Following graft failure, stroke rates showed no pattern during the first year (P > 0.9), while other death rate was highest in the first 3 months (P < 0.001) (Figure 2). There was some indication that stroke death rates were highest in the first year after graft failure, but this was not statistically significant (P = 0.408). Other death rates after graft failure was highest in the first year, then decreased thereafter (P < 0.001).

The stroke mortality rates showed a clear decreasing trend over time when stratified by calendar year (Figure 3A), from 202 (95% CI, 76-541)/100 000 person-years in 1988 to 67 (95% CI, 32-140)/100 000 person-years in 2013 (P = 0.001). The stroke mortality rates increased with older age in both sexes (P < 0.001), being slightly higher in females (P = 0.031) (Figure 3B). The stroke mortality rate per 100 000 person-years increased from 39 (95% CI, 10-156) in females and 52 (95% CI, 19-137) in males aged 30 years to 193 (95% CI, 87-430) in females and 257 (95% CI, 138-478) in males aged 70 years.

FIGURE 3.
FIGURE 3.:
Crude stroke mortality rates per 100,000 person-years (left) and standardized mortality ratios (SMRs; right). This is given by calendar year in (A, B) and age and sex in (C, D).

Stroke SMR

The overall SMR for stroke was 4.2 (95% CI, 3.6-4.9), for intracerebral hemorrhage was 6.3 (95% CI, 4.9-8.2), and for ischemic stroke was 4.2 (95% CI, 2.7-6.5). The stroke SMR decreased over time (P < 0.001) (Figure 3C). The SMR were highest before 2000, where the kidney transplant recipients had at least 5 times the stroke deaths than expected. The SMR reduced from 11.6 (95% CI, 4.4-30.9) in 1988 to 3.1 (95% CI, 1.5-6.4) in 2013.

The stroke SMR were highest in those who were younger (P < 0.001) and were slightly greater in females (Figure 3D). In those aged 30–49 years, females had over 19 times (SMR: 19.7, 95% CI, 12.9-30.3) and males had 9 times (SMR: 9.1, 95% CI, 5.6-14.6) the stroke deaths than expected (Table 2). The excess stroke deaths decreased with older age, where those aged ≥75 years had comparable stroke mortality rates to the general population. Similar trends in SMR were seen in ischemic stroke and intracerebral hemorrhage deaths, with a larger SMR in those who were younger and of female sex.

TABLE 2. - Standardized mortality ratio estimates for all-cause stroke, intracerebral hemorrhages, and ischemic hemorrhages, and in those with polycystic kidney disease as the cause of kidney failure, by age group
O E SMR (95% CI) O E SMR (95% CI)
Age at death (y) 30–49 50–64
 Male
  Any stroke death 17 1.9 9.1 (5.6-14.6) 37 7.9 4.7 (3.4-6.5)
  Intracerebral hemorrhage 7 0.6 12.6 (6.0-26.3) 13 2.3 5.6 (3.3-9.7)
  Ischemic stroke 1 0.2 5.6 (0.8-39.9) 3 1.1 2.8 (0.9-8.6)
  Unspecified stroke 3 0.1 27.1 (8.7-84.0) 13 1.4 9.4 (5.5-16.3)
 Female
  Any stroke death 21 1.1 19.7 (12.9-30.3) 35 4.0 8.8 (6.3-12.3)
  Intracerebral hemorrhage 10 0.2 43.6 (23.5-81.0) 15 1.1 14.3 (8.6-23.6)
  Ischemic stroke 5 0.1 59.8 (24.9-143.7) 6 0.3 18.5 (8.3-41.1)
  Unspecified stroke 2 0.1 38.4 (9.6-153.4) 10 0.5 21.0 (11.3-39.1)
65–74 75+
 Male
  Any stroke death 19 8.7 2.2 (1.4-3.4) 7 5.0 1.4 (0.7-3.0)
  Intracerebral hemorrhage 3 2.1 1.4 (0.5-4.4) 2 0.9 2.3 (0.6-9.2)
  Ischemic stroke 1 1.3 0.8 (0.1-5.6) 1 0.7 1.5 (0.2-10.6)
  Unspecified stroke 11 3.1 3.5 (1.9-6.3) 3 2.8 1.1 (0.3-3.3)
 Female
  Any stroke death 15 5.1 2.9 (1.8-4.8) 6 4.1 1.5 (0.7-3.2)
  Intracerebral hemorrhage 5 1.2 4.2 (1.8-10.2) 1 0.7 1.5 (0.2-10.6)
  Ischemic stroke 1 0.6 1.6 (0.2-11.6) 2 0.5 4.0 (1.0-15.9)
  Unspecified stroke 6 1.7 3.5 (1.6-7.7) 3 2.5 1.2 (0.4-3.8)
CI, confidence interval; E, expected; O, observed; SMR, standardized mortality ratios.

Risk Factors for Stroke Mortality

Higher risk of stroke death was associated with older age at transplant, having ever had a kidney transplant fail, earlier era of transplant, preexisting cerebrovascular disease, and not previously having had cancer in the multivariate model (Table 3). These risk factors were also associated with a higher risk of other deaths, except for preexisting cerebrovascular disease. Additional risk factors for other deaths included BMI, previous duration of dialysis, being from a White racial background, preexisting coronary or peripheral artery disease, smoking status, and cause of kidney failure. These findings were consistent with estimates from the cause-specific Cox model (Table S2, SDC, http://links.lww.com/TP/B856).

TABLE 3. - Summary of the multivariable results for the competing risks model, with stroke (n = 158) and nonstroke (n = 5216) deaths as the events of interest
Stroke death Other death
SHR (95% CI) P SHR (95% CI) P
Age at transplant (y) <0.001 <0.001
 ≤29 0.33 (0.17-0.63) 0.50 (0.45-0.56)
 30–44 ref ref
 45–54 2.08 (1.39-3.11) 1.96 (1.81-2.13)
 ≥55 1.80 (1.14-2.82) 3.66 (3.35-4.01)
Sex
 Female ref ref
 Male 0.78 (0.55-1.10) 0.153 1.04 (0.97-1.11) 0.288
Body mass index (kg/m2) 0.425 0.017
 Underweight (≤18.4) 0.96 (0.53-1.72) 1.05 (0.92-1.19)
 Normal (18.5–24.9) ref ref
 Overweight (25.0–29.9) 0.69 (0.47-0.99) 1.08 (1.00-1.16)
 Obese (≥30.0) 0.49 (0.28-0.85) 1.27 (1.16-1.38)
Ever graft failure
 No ref ref
 Yes 1.88 (1.32-2.67) <0.001 3.28 (3.06-3.52) <0.001
Previous dialysis time (mo) 0.737 <0.001
 0–<6 ref ref
 ≥6–<12 1.05 (0.64-1.72) 1.21 (1.09-1.34)
 ≥12–<18 0.71 (0.39-1.28) 1.35 (1.21-1.50)
 ≥18–<36 0.92 (0.56-1.50) 1.49 (1.36-1.64)
 ≥36 0.98 (0.59-1.64) 1.69 (1.53-1.86)
Y of transplant <0.001 <0.001
 ≤1995 ref ref
 1996–2005 0.40 (0.26-0.63) 0.64 (0.59-0.68)
 2006–2013 0.18 (0.07-0.47) 0.40 (0.36-0.46)
Racial background
 Nonwhite ref ref
 White 1.74 (0.98-3.09) 0.057 1.11 (1.01-1.22) 0.038
Comorbidities
 Cerebrovascular disease 2.97 (1.58-5.56) 0.001 1.05 (0.86-1.28) 0.631
 Coronary artery disease 0.66 (0.34-1.27) 0.209 1.14 (1.03-1.26) 0.014
 Peripheral artery disease 0.83 (0.37-1.82) 0.635 1.23 (1.08-1.40) 0.002
 Previous malignancy 0.65 (0.47-0.89) 0.008 0.93 (0.87-0.99) 0.024
Smoking status 0.760 <0.001
 Never ref ref
 Current 0.84 (0.51-1.38) 1.36 (1.24-1.50)
 Former 0.78 (0.53-1.16) 1.17 (1.09-1.26)
Cause of kidney failure 0.683 <0.001
 Diabetes 1.39 (0.79-2.44) 2.08 (1.88-2.31)
 Hypertension/renal artery disease 1.16 (0.53-2.56) 1.21 (1.03-1.42)
 Glomerulonephritis/IgA nephropathy ref ref
 Polycystic kidney disease 1.26 (0.79-2.00) 0.94 (0.86-1.04)
 Other 0.97 (0.66-1.44) 1.16 (1.07-1.25)
The multivariate model was also adjusted for country. Global P Value was Wald tests for heterogeneity.
Bold indicates significance at P < 0.05.
CI, confidence interval; SHR, subhazard ratio.

Recipients who were older at transplant had a higher risk of stroke death (P < 0.001). Those aged ≥55 years at transplant were 80% (SHR: 1.80, 95% CI, 1.14-2.82) more likely to die of stroke compared with 30–44 years. Recipients who experienced graft failure were 88% (SHR: 1.88, 95% CI, 1.32-2.67) more likely to die of stroke (P < 0.001). However, previous duration of dialysis was not associated with their risk of stroke death (P = 0.737), but longer previous duration of dialysis was associated with an increased risk of death from other causes. Recipients transplanted in more recent years had a lower risk of stroke death. Those transplanted in 2006–2013 were 82% (SHR: 0.17, 95% CI, 0.07-0.47) less at risk of stroke death and 60% (SHR: 0.40, 95% CI, 0.36-0.46) less at risk of death from other causes compared with recipients transplanted in ≤1995 (P < 0.001). Recipients with cerebrovascular disease were nearly 3 times (SHR: 2.97, 95% CI, 1.58-5.56) more likely of stroke death (P = 0.001), yet cerebrovascular disease was not associated with other deaths. The unadjusted estimates for stroke and other death are presented in Table S3 (SDC, http://links.lww.com/TP/B856).

Recipients With Polycystic Kidney Disease

There were 2196 kidney transplant recipients with polycystic kidney disease followed over 19 491.2 person-years, where 28 stroke deaths and 608 other deaths occurred. Most strokes were hemorrhagic (9 intracerebral, 5 subarachnoid, and 1 intracranial hemorrhages), similarly distributed to the main analysis (55% versus 53%). Of the 608 other deaths, the leading underlying cause of death was cystic kidney disease (n = 109), followed by coronary heart disease (n = 94) and other kidney and ureter disorders (n = 39). The overall stroke mortality rate was 143.7 (95% CI, 99.2-208.1)/100 000 person-years, and the overall stroke SMR was 4.2 (95% CI, 2.9-6.0). Due to the low number of stroke deaths, we did not evaluate associated risk factors.

DISCUSSION

In this large cohort of 17 628 kidney transplant recipients, we found a high excess of stroke deaths in recipients with over 4 times the stroke deaths compared with people of the same age, sex, and calendar year from the general population. The greatest excess was for hemorrhagic strokes, where kidney transplant recipients had over 6 times the expected stroke deaths. While excess stroke deaths did reduce over time, it remained 3 times higher than expected in the most recent year and was highest among young recipients. Further, we described risk factors for stroke death in kidney transplant recipients where those with preexisting cerebrovascular disease had nearly 3 times the risk of stroke death. Previous duration of dialysis was not associated with the risk of stroke death but was associated with a greater risk of death from other causes.

To our knowledge, this is the first study to examine excess stroke deaths in kidney transplant recipients. While previous studies have found dialysis patients have 3-fold increase in stroke deaths compared with the general population,2 the larger 4-fold increase in stroke deaths we observed in kidney transplant recipients is likely due to a combination of factors. Despite cardiovascular screening in potential recipients, there is still a high prevalence of traditional cardiovascular risk factors among kidney transplant recipients, including hypertension, diabetes, and obesity.25 In addition, predisposing factors, which led to the development and progression of kidney failure, do not remit following transplantation and contribute to the acquisition of new or progression of existing cardiovascular diseases.17 Certain immunosuppression drugs may also increase the risk of developing hypertension, high cholesterol, and diabetes posttransplantation, such as calcineurin inhibitors and corticosteroids.26-28 Further, the higher excess of stroke deaths in transplant recipients compared with dialysis patients is likely due to the younger age distribution where there is a greater excess of stroke deaths (transplant recipients: 45 years median age; dialysis patients: 59 years median age).

Improvements in graft survival, immunosuppression, and cardiovascular management have reduced the risk of stroke death in kidney transplant recipients over time. Ever experiencing graft failure was associated with nearly double the risk of stroke death. This is somewhat expected given the underlying mechanisms where decreased kidney function increases the risk of stroke, including platelet dysfunction, inflammation, accelerated atherosclerosis, and coagulation abnormalities.29-31 In addition, previous duration of dialysis was not found to increase the risk of stroke death in our study but did increase the risk of death from other causes. These findings align with other studies in kidney transplant recipients1,13,32 and suggest that longer dialysis vintage does not have longstanding effects on the risk of stroke death. The probability of graft failure has reduced over time, likely due to reduced transplant waiting times and changes in immunosuppression regimens.33-35 Initial regimens in Australia have moved away from cyclosporine (65% in 2000 to 2% in 2013) to tacrolimus (26% in 2000 to 87% in 2013).36,37 Similarly, New Zealand has slowly increased the use of tacrolimus from 5% in 2000 to 17% in 2013. Tacrolimus is associated with fewer acute rejections, better graft survival, and lower systolic blood pressure and triglycerides.26,33,34 Further, cyclosporine is associated with increased risk of hypertension and hyperlipidemia, more so than tacrolimus.26,27 While it is also likely that kidney transplant recipients have benefited from improved cardiovascular management over time, their effectiveness in kidney transplant recipients is mainly extrapolated from studies in people without kidney impairment. The Assessment of LEscol in Renal Transplantation (ALERT) clinical trial found no evidence that lipid-lowering therapy reduced the risk of stroke death38 and an observational study found statin use did not reduce stroke risk.12 Hence, it is possible that cardiovascular management is less effective at reducing stroke risk in kidney transplant recipients, resulting in the over 3 times excess stroke deaths still occurring in the most recent era. Young transplant recipients also had the greatest excess of stroke deaths. This may be a result of underrecognition of their cardiovascular risk and less access to cardiovascular management.

Preexisting cerebrovascular disease nearly tripled the risk of stroke death in kidney transplant recipients, indicating that kidney transplant recipients may be receiving ineffective or are not accessing secondary stroke prevention. This aligns with the ALERT clinical trial study which found previous cerebrovascular event is associated with a 3.5-fold increase in the risk of ischemic stroke.39 There are no clinical trials assessing stroke prevention or specific guidelines for managing stroke risk in kidney transplant recipients. Findings from an observational study indicated that warfarin was not associated with reduced risk of stroke in kidney transplant recipients with atrial fibrillation.40 While meta-analyses in people with kidney impairment found that antiplatelet agents did not reduce the risk of stroke, but novel oral anticoagulants did by 30%.41,42 Further, an Australian study found that more often general practitioners do not prescribe antiplatelet therapy in people with kidney impairment, despite their increased stroke risk, compared with other groups.43 Therefore, more studies are needed to evaluate whether kidney transplant recipients are receiving ineffective stroke prevention or are not accessing appropriate stroke prevention.

The main strength in our study is the large cohort size that includes the entire kidney transplant recipient population in Australia and New Zealand since 1988. However, there were limitations to our study. First, we were limited to clinical data captured in ANZDATA that does not collect concurrent medications or other relevant stroke risk factors, such as atrial fibrillation, blood pressure, or family history of stroke. We therefore were unable to examine these in our SMR estimates or associations with stroke death. Addressing modifiable risk factors may also lead to improved outcomes in kidney transplant recipients. While further studies with more granular cardiovascular data would be beneficial, our study is the first to report population-based estimates of excess stroke deaths and risk factors among kidney transplant recipients. Second, deaths occurring overseas would not have been captured in data linkage. However, it is likely there were few overseas deaths given the high agreement between ANZDATA and death registers. Further, probabilistic data linkage used in the Australian kidney transplant recipients may result in incorrect links. Though this has been previously reported to be relatively low at <5 per 1000 records.44 Third, we used the underlying cause of death from the national death registers to determine stroke deaths in our cohort. While misclassification may occur, the accuracy of stroke deaths given in the Australian death register has been evaluated with a high specificity at 99% and sensitivity at 59%.45 Hence, stroke deaths are more likely to be underreported, and our estimates are conservative of the true excess of stroke deaths. In addition, inclusion of contributing causes of death would have resulted in capturing and additional 76 stroke deaths. However, it is difficult to disentangle the degree to which the stroke caused the death when relying on contributing causes of death. Thus, using the primary cause of death is more reliable in identifying stroke deaths that are unrelated to postsurgery or trauma. Fourth, we used the general population as the reference population for our estimates. Other matching methods, such as based on propensity scores, may led to different findings, but these methods may be subject to selection bias46 or overmatching47 which can bias results. Using the general population is the least biased comparison group and is a robust epidemiological method.

In conclusion, kidney transplant recipients have a high excess of stroke deaths, particularly young recipients, and risk factors for stroke death include preexisting cerebrovascular disease and graft failure. While there were improvements over time, it is unclear whether kidney transplant recipients have access to stroke prevention or are receiving effective stroke prevention. Preexisting cerebrovascular disease being a risk factor presents an opportunity for secondary prevention. Further studies are needed to assess the benefits and harms of current stroke management in kidney transplant recipients or determine whether specialized stroke prevention and intervention need to be developed through novel clinical trials.

ACKNOWLEDGMENTS

The data reported here have been supplied by the Australia and New Zealand Dialysis and Transplant Registry (ANZDATA), the Australian Institute of Health and Welfare (AIHW), and the New Zealand Ministry of Health. We would like to acknowledge the assistance provided by the AIHW in the data linkage process. The interpretation and reporting of these data are the responsibility of the Editors and in no way should be seen as an official policy of interpretation of ANZDATA, AIHW, or the New Zealand Ministry of Health.

REFERENCES

1. Wolfe RA, Ashby VB, Milford EL, et al. Comparison of mortality in all patients on dialysis, patients on dialysis awaiting transplantation, and recipients of a first cadaveric transplant. N Engl J Med. 1999; 341:1725–1730
2. De La Mata NL, Masson P, Al-Shahi Salman R, et al. Death from stroke in end-stage kidney disease. Stroke. 2019; 50:487–490
3. Meier-Kriesche HU, Schold JD, Srinivas TR, et al. Kidney transplantation halts cardiovascular disease progression in patients with end-stage renal disease. Am J Transplant. 2004; 4:1662–1668
4. Campbell S, Pilmore H, Gracey D, et al. KHA-CARI guideline: recipient assessment for transplantation. Nephrology (Carlton). 2013; 18:455–462
5. Gansevoort RT, Correa-Rotter R, Hemmelgarn BR, et al. Chronic kidney disease and cardiovascular risk: epidemiology, mechanisms, and prevention. Lancet. 2013; 382:339–352
6. Feigin VL, Lawes CM, Bennett DA, et al. Worldwide stroke incidence and early case fatality reported in 56 population-based studies: a systematic review. Lancet Neurol. 2009; 8:355–369
7. Australian Bureau of Statistics. Causes of death, Australia.2015:Canberra, Australia: Australian Bureau of Statistics; Available at http://www.abs.gov.au/AUSSTATS/abs@.nsf/Lookup/3303.0Main+Features100012015?OpenDocument. Accessed July 18, 2018
8. The New Zealand Ministry of Health. Mortality and demographic data 2012.2015:Wellington, New Zealand: The New Zealand Ministry of Health; Available at http://www.health.govt.nz/system/files/documents/publications/mortality-and-demographic-data-2012-nov15.pdf. Accessed March 24, 2017
9. Palacio S, Gonzales NR, Sangha NS, et al. Thrombolysis for acute stroke in hemodialysis: international survey of expert opinion. Clin J Am Soc Nephrol. 2011; 6:1089–1093
10. Aakhus S, Dahl K, Widerøe TE. Cardiovascular disease in stable renal transplant patients in Norway: morbidity and mortality during a 5-yr follow-up. Clin Transplant. 2004; 18:596–604
11. Aull-Watschinger S, Konstantin H, Demetriou D, et al. Pre-transplant predictors of cerebrovascular events after kidney transplantation. Nephrol Dial Transplant. 2008; 23:1429–1435
12. Findlay MD, Thomson PC, MacIsaac R, et al. Risk factors and outcome of stroke in renal transplant recipients. Clin Transplant. 2016; 30:918–924
13. Oliveras A, Roquer J, Puig JM, et al. Stroke in renal transplant recipients: epidemiology, predictive risk factors and outcome. Clin Transplant. 2003; 17:1–8
14. Aalten J, Hoogeveen EK, Roodnat JI, et al. Associations between pre-kidney-transplant risk factors and post-transplant cardiovascular events and death. Transpl Int. 2008; 21:985–991
15. De la Mata NL, Alfaro-Ramirez M, Masson P, et al. Absolute risk and risk factors for stroke mortality in patients with end stage kidney disease (ESKD): retrospective population-based cohort using data linkage. Nephrology. 2017; 22Suppl 318–50
16. Miller LW. Cardiovascular toxicities of immunosuppressive agents. Am J Transplant. 2002; 2:807–818
17. Fellström B, Jardine AG, Soveri I, et al.; ALERT Study Group. Renal dysfunction is a strong and independent risk factor for mortality and cardiovascular complications in renal transplantation. Am J Transplant. 2005; 5:1986–1991
18. Soveri I, Holdaas H, Jardine A, et al. Renal transplant dysfunction–importance quantified in comparison with traditional risk factors for cardiovascular disease and mortality. Nephrol Dial Transplant. 2006; 21:2282–2289
19. Meier-Kriesche HU, Kaplan B. Waiting time on dialysis as the strongest modifiable risk factor for renal transplant outcomes: a paired donor kidney analysis. Transplantation. 2002; 74:1377–1381
20. Adams HP Jr, Dawson G, Coffman TJ, et al. Stroke in renal transplant recipients. Arch Neurol. 1986; 43:113–115
21. McDonald SP, Russ GR. Australian registries-ANZDATA and ANZOD. Transplant Rev (Orlando). 2013; 27:46–49
22. Blakely T, Salmond C. Probabilistic record linkage and a method to calculate the positive predictive value. Int J Epidemiol. 2002; 31:1246–1252
23. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics. 1977; 33:159–174
24. Fine JP, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc. 1999; 94:496–509
25. Ojo AO. Cardiovascular complications after renal transplantation and their prevention. Transplantation. 2006; 82:603–611
26. Fellström B. Risk factors for and management of post-transplantation cardiovascular disease. Biodrugs. 2001; 15:261–278
27. Mangray M, Vella JP. Hypertension after kidney transplant. Am J Kidney Dis. 2011; 57:331–341
28. Pascual M, Curtis J, Delmonico FL, et al. A prospective, randomized clinical trial of cyclosporine reduction in stable patients greater than 12 months after renal transplantation. Transplantation. 2003; 75:1501–1505
29. Arnold J, Sims D, Ferro CJ. Modulation of stroke risk in chronic kidney disease. Clin Kidney J. 2016; 9:29–38
30. El Husseini N, Kaskar O, Goldstein LB. Chronic kidney disease and stroke. Adv Chronic Kidney Dis. 2014; 21:500–508
31. Lee M, Saver JL, Chang KH, et al. Low glomerular filtration rate and risk of stroke: meta-analysis. BMJ. 2010; 341:c4249
32. Weiner DE, Carpenter MA, Levey AS, et al. Kidney function and risk of cardiovascular disease and mortality in kidney transplant recipients: the FAVORIT trial. Am J Transplant. 2012; 12:2437–2445
33. Australian and New Zealand Dialysis and Transplant Registry. 37th Annual ANZDATA Report, Chapter 7: Transplant Waiting list. Adelaide, Australia: 2014. Available at https://www.anzdata.org.au/wp-content/uploads/2016/12/c07_waitinglist_print_20150929.pdf. Accessed March 21, 2019
34. Gallagher M, Jardine M, Perkovic V, et al. Cyclosporine withdrawal improves long-term graft survival in renal transplantation. Transplantation. 2009; 87:1877–1883
35. Gallagher MP, Hall B, Craig J, et al.; Australian Multicenter Trial of Cyclosporine Withdrawal Study Group and the ANZ Dialysis and Transplantation Registry. A randomized controlled trial of cyclosporine withdrawal in renal-transplant recipients: 15-year results. Transplantation. 2004; 78:1653–1660
36. Australian and New Zealand Dialysis and Transplant Registry. 27th Annual ANZDATA Report, Chapter 8: Transplant. Adelaide, Australia2004[Cited March 21, 2019.] Available at https://www.anzdata.org.au/wp-content/uploads/2016/12/Ch08Transplantation-3.pdf. Accessed March 21, 2019
37. Australian and New Zealand Dialysis and Transplant Registry. 41st Annual ANZDATA Report, Chapter 7: Kidney Transplantation. Adelaide, Australia2018Available at https://www.anzdata.org.au/wp-content/uploads/2018/11/c07_transplant_2017_v1.0_20181220.pdf. Accessed March 1, 2019
38. Holdaas H, Fellström B, Cole E, et al.; Assessment of LEscol in Renal Transplantation (ALERT) Study Investigators. Long-term cardiac outcomes in renal transplant recipients receiving fluvastatin: the ALERT extension study. Am J Transplant. 2005; 5:2929–2936
39. Abedini S, Holme I, Fellström B, et al.; ALERT Study Group. Cerebrovascular events in renal transplant recipients. Transplantation. 2009; 87:112–117
40. Lenihan CR, Montez-Rath ME, Shen JI, et al. Correlates and outcomes of warfarin initiation in kidney transplant recipients newly diagnosed with atrial fibrillation. Nephrol Dial Transplant. 2015; 30:321–329
41. Palmer SC, Di Micco L, Razavian M, et al. Effects of antiplatelet therapy on mortality and cardiovascular and bleeding outcomes in persons with chronic kidney disease: a systematic review and meta-analysis. Ann Intern Med. 2012; 156:445–459
42. Sardar P, Chatterjee S, Herzog E, et al. Novel oral anticoagulants in patients with renal insufficiency: a meta-analysis of randomized trials. Can J Cardiol. 2014; 30:888–897
43. Razavian M, Heeley EL, Perkovic V, et al. Cardiovascular risk management in chronic kidney disease in general practice (the AusHEART study). Nephrol Dial Transplant. 2012; 27:1396–1402
44. Centre for Health Record Linkage. Master Linkage Key Quality Assurance.2012:Sydney, AustraliaAvailable at http://www.cherel.org.au/media/24160/qa_report_2012-a.pdf. Accessed March 13, 2018
45. Harriss LR, Ajani AE, Hunt D, et al. Accuracy of national mortality codes in identifying adjudicated cardiovascular deaths. Aust N Z J Public Health. 2011; 35:466–476
46. Grimes DA, Schulz KF. Compared to what? Finding controls for case-control studies. Lancet. 2005; 365:1429–1433
47. Miettinen OS. Matching and design efficiency in retrospective studies. Am J Epidemiol. 1970; 91:111–118

Supplemental Digital Content

Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.