Abstract
Introduction
Data on the association between chronic kidney disease (CKD) and major hemorrhage in older adults are lacking.
Methods
We used data from a double-blind randomized controlled trial of aspirin in persons aged ≥ 70 years with prospective capture of bleeding events, including hemorrhagic stroke and clinically significant bleeding. CKD was defined as an estimated glomerular filtration rate (eGFR) < 60 ml/min per 1.73 m2 and/or urinary albumin-to-creatinine ratio (UACR) ≥ 3 mg/mmol (26.6 mg/g). We compared bleeding rates in those with and without CKD, undertook multivariable analyses, and explored effect modification with aspirin.
Results
Of 19,114 participants, 17,976 (94.0%) had CKD status recorded, of whom 4952 (27.5%) had CKD. Participants with CKD had an increased rate of major bleeding events compared with those without CKD (10.4/1000 vs. 6.3/1000 person-years [py], respectively) and increased bleeding risk (risk ratio [RR] 1.60; 95% confidence interval [CI]: 1.40, 1.90 for eGFR < 60 ml/min per 1.73 m2) and RR (2.10; 95% CI: 1.70, 2.50) for albuminuria. In adjusted analyses, CKD was associated with a 35% increased risk of bleeding (hazard ratio [HR] 1.37; 95% CI: 1.15, 1.62; P < 0.001). Other risk factors were older age, hypertension, smoking, and aspirin use. There was no differential effect of aspirin on bleeding by CKD status (test of interaction P = 0.65).
Conclusion
CKD is independently associated with an increased risk of major hemorrhage in older adults. Increased awareness of modifiable risk factors such as discontinuation of unnecessary aspirin, blood pressure control, and smoking cessation in this group is warranted.
Keywords: albuminuria, chronic kidney disease, hemorrhage, risk
Graphical abstract
CKD affects up to 15% of the global population1 and is particularly common in people older than 70 years, where prevalence is estimated at 27%.2 People with CKD are well recognized to have a higher risk of cardiovascular disease, stroke, and anemia.3 However, emerging data suggest that other sequelae including major bleeding are also important.3 Increased bleeding risk is well recognized in those with advanced CKD, for example, those receiving dialysis.4 Nevertheless, increased bleeding risk also occurs in those with milder forms of CKD, whereby the risk increases in proportion to decline in eGFR and rise in urinary albumin excretion.5 Despite the high prevalence of CKD in older populations, data on the bleeding risk in this population are lacking because current studies, including a Cochrane systematic review,6 are derived from younger populations or those with advanced kidney disease. This is an important evidence gap, given the intersection of aging populations in many Western countries, the substantial morbidity associated with major bleeding events, and because bleeding events are a major source of health care expenditure.7
Data generated from the Aspirin in Reducing Events in the Elderly (ASPREE) trial8,9 provide an unprecedented opportunity to examine bleeding risk in an older, highly characterized population with CKD. ASPREE was a randomized, community-based trial of aspirin or placebo in >19,000 people older than 70 years (or 65 years for US minorities) with baseline collection of CKD status using both eGFR and urinary albumin measurements. Because bleeding was a prespecified trial end point, data on any bleeding event were captured prospectively and adjudicated independently over a long period of follow-up (up to 7 years). As such, ASPREE trial data offer a unique opportunity to assess bleeding risk in older populations.
In this study, we assessed the incidence of bleeding in older adults and its association by CKD status. Given that ASPREE was a trial of aspirin, we further examined the association between use of aspirin and bleeding in those with and without CKD. We hypothesized that older people with CKD would have a higher rate of significant bleeding events than those without CKD and that aspirin may be an effect modifier of the relationship between CKD and bleeding.
Methods
ASPREE Study Design
The ASPREE clinical trial was a double-blind, randomized, placebo-controlled trial that evaluated the effect of enteric-coated aspirin (100 mg daily) in participants aged 70 years and older (or 65 and older for US minorities) to assess whether aspirin prolonged disability-free survival in healthy older adults.8, 9, 10 The trial enrolled 19,114 community-dwelling persons (87% Australian and 13% US adults) who were not on aspirin for secondary prevention and were free of significant health problems such as cardiovascular disease, dementia, independence-limiting physical disability, or any illness expected to limit their life expectancy to less than 5 years. Exclusion criteria included allergy to aspirin or current anemia (defined as hemoglobin of <12 g/dl for men and <11 g/dl for women). During baseline assessments, blood and urine samples were analyzed, followed by annual face-to-face interviews with annual or biennial laboratory measures and 3 monthly telephone calls for the trial duration. These telephone calls allowed for additional data collection of events (6 monthly) and helped encourage retention and check adherence to medication use (3 monthly). The trial commenced recruitment in March 2010 and ended in December 2014. Participants, their health care providers, and trial staff were blinded to treatment allocation. Trial cessation occurred in June 2017, resulting in a median of 4.7 years of follow-up. A comprehensive trial protocol has been published elsewhere.10 ASPREE was approved by multiple institutional review boards in Australia and the United States, registered with International Standard Randomized Controlled Trial Number Register (ISRCTN83772183) and www.clinicaltrials.gov (NCT01038583). All participants provided written, informed consent.
Definition of CKD
CKD was defined using Kidney Disease: Improving Global Outcomes criteria as follows: eGFR < 60 ml/min per 1.73 m2 and/or spot UACR ≥ 3 mg/mmol.11 Participants missing baseline assessment of either of the 2 kidney function markers were excluded. We calculated eGFR based on serum creatinine measurement using the Chronic Kidney Disease Epidemiology Collaboration equation.12
Definition of Major Hemorrhage
Major hemorrhage was a composite of clinically significant bleeding (intracranial and extracranial) and hemorrhagic stroke. Because there is no consensus definition for clinically significant bleeding in aspirin prevention studies, we developed a customized definition13 that incorporated commonly used criteria for major bleeding such as need for transfusion or hospitalization. To reach an end point of clinically significant bleeding, the event was required to meet 2 criteria, which were adjudicated by 2 independent physicians blinded to treatment arm (Supplementary Table S1A). First, participants were required to have overt, acute bleeding that was substantiated by their medical record or imaging or endoscopic evidence. Second, the clinically significant bleeding must have resulted in one of the following: transfusion requirement, hospitalization, surgery to control the bleeding, or death.13 Events were detected by participants’ self-report, by interviewers or primary care doctor report to trial staff, followed by ascertainment of the medical records for the bleeding end point adjudication committee. Members of the committee were trained physicians with expertise in gastroenterology, hematology, or geriatrics. Discordant adjudications were resolved by consensus involving a third opinion and discussed at regular meetings to ensure consistency in adjudications. Hemorrhagic stroke was identified by an adjudication panel of clinical experts who used the World Health Organization criteria9 and required evidence of rapidly developing clinical signs of focal (or global) disturbance of cerebral function lasting more than 24 hours (unless interrupted by surgery or death) and the hemorrhage identified on imaging as previously described.9 Further details are included in Supplementary Table S1B.
Major bleeding events were recorded as the first bleeding event that met adjudication criteria. These were subclassified into 3 categories as follows: (i) intracranial bleeding (including hemorrhagic stroke, subdural hemorrhage, extradural hemorrhage, and subarachnoid hemorrhage), (ii) gastrointestinal bleeding (inclusive of bleeds in the upper gastrointestinal tract such as peptic ulcer bleeding and those arising from the lower gastrointestinal tract such as diverticular bleeding), and (iii) all other bleeding events that met the adjudication criteria and typically included hematuria, major epistaxis, or gynecologic bleeding.
Variables for Multivariable Analyses
Trained interviewers collected a wide range of participant data at baseline, including age, sex, race or ethnicity (White, Hispanic, African American, or other), smoking status (current, former, or never), hypertension (defined as use of antihypertensive medication or a blood pressure ≥ 140/90 mm Hg at trial entry), presence of type 2 diabetes (defined as participant report, drug therapy for diabetes or a fasting glucose ≥ 126 mg/dl or 7 mmol/l), alcohol intake (self-reported as current, former, or never), waist circumference (defined as “not at risk” or “at risk” using sex-specific cutoffs of 88 cm for women and 102 cm for men), and concurrent use of nonsteroidal anti-inflammatory medication or proton pump inhibitor therapy. To evaluate the effect of CKD on bleeding risk, we selected clinically important risk factors for intracranial and gastrointestinal bleeding, supported by current trial literature and individual patient data meta-analyses. Variables included age, sex, hypertension, nonsteroidal anti-inflammatory medications, smoking, alcohol, elevated waist circumference, and diabetes.14,15
Statistical Analyses
In this post hoc secondary analysis of ASPREE data, we describe clinical and demographic characteristics among trial participants using simple frequencies or means with standard deviations. The rates of first occurrence of major bleeding in people with and without CKD were calculated per 1000 py for overall bleeding events and separately for subtypes of gastrointestinal, intracranial, and other bleeding. Cumulative incidence curves of bleeding events according to CKD status were constructed, accounting for competing risk of death. The association between major hemorrhage and eGFR or UACR as categorical or continuous variables was explored in the Cox proportional hazards regression models. To determine if the association between the continuous eGFR and UACR and log-hazard may be nonlinear, penalized splines were used. Multivariable models inclusive of clinically important risk factors for bleeding were constructed and the proportional hazards assumption was checked using Schoenfeld residuals. Statistical tests for interaction between aspirin and CKD status were undertaken in the Cox model to test for a differential effect on bleeding in those with CKD in the presence of aspirin. In all analyses, a P value < 0.05 was considered statistically significant. All analyses were performed using software R (R Core Team, Vienna, Austria, 2020).
Results
Participant Characteristics
Of 19,114 participants in the ASPREE trial enrolled at baseline, eGFR calculation was available for 18,650 (97.6%) participants, and urinary albumin was available for 18,131 participants (94.9%). In total, 17,976 (94.0%) participants had CKD status available using both or either definition (Figure 1). Of these, 27.5% (4952) had CKD, with 18.4% (3427 of 18,650) with eGFR < 60 ml/min per 1.73 m2, and 11.5% (2085 of 18,131) had UACR ≥ 3 mg/mmol (Table 1). Trial participants with CKD were older than those without CKD (mean age 77 ± 5.2 years vs. 75 ± 4.2 years) and more likely to have hypertension (83% of those with CKD vs. 71% of those without CKD) and type 2 diabetes (15% of those with CKD vs. 9% of those without CKD). There were no differences in smoking, alcohol intake, or nonsteroidal anti-inflammatory drug use.
Figure 1.
Flow diagram. CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate.
Table 1.
Baseline characteristics according to CKDa status
| Characteristic | No CKDa |
CKDa |
Total |
|---|---|---|---|
| (n = 13,024) | (n = 4952) | (N = 17,976) | |
| Age (yr) | |||
| Mean (SD) | 75 (±4.2) | 77 (±5.2) | 75 (±4.6) |
| Age category | |||
| 65–74 yr | 8190 (63%) | 2268 (46%) | 10,458 (58%) |
| 75–79 yr | 3332 (26%) | 1406 (28%) | 4738 (26%) |
| 80+ | 1502 (12%) | 1278 (26%) | 2780 (15%) |
| Sex | |||
| Male | 5728 (44%) | 2094 (42%) | 7822 (44%) |
| Female | 7296 (56%) | 2858 (58%) | 10,154 (56%) |
| CKD characteristics | |||
| eGFR (ml/min), median (IQR) | 77.9 (69.6–85.9) | 56.6 (50.4–65.8) | 74.0 (63.4–84.4) |
| eGFR 45–59 ml/min (no. of participants) | NA | 2821 | 2821 |
| eGFR 30–<45 ml/min (no. of participants) | NA | 559 | 559 |
| eGFR <30 ml/min (no. of participants) | NA | 47 | 47 |
| Urinary ACR, median (IQR) | 0.7 (0.4–1.1) | 1.8 (0.7–4.9) | 0.8 (0.5–1.5) |
| Race/ethnicity | |||
| White AUS/US | 11,871 (91%) | 4506 (91%) | 16,377 (91%) |
| Hispanic | 340 (3%) | 126 (3%) | 466 (3%) |
| Black | 627 (5%) | 251 (5%) | 878 (5%) |
| Other | 186 (1%) | 69 (1%) | 255 (1%) |
| Hypertensionb | |||
| No | 3775 (29%) | 822 (17%) | 4597 (26%) |
| Yes | 9249 (71%) | 4130 (83%) | 13,379 (74%) |
| SBP (mm Hg) | |||
| Mean (SD) | 140 (±16) | 140 (±17) | 140 (±17) |
| Diabetes mellitusc | |||
| No | 11,822 (91%) | 4208 (85%) | 16,030 (89%) |
| Yes | 1202 (9%) | 744 (15%) | 1946 (11%) |
| BMI (kg/m2) | |||
| Mean (SD) | 28 (±4.6) | 29 (±5.0) | 28 (±4.7) |
| Missing | 55 (0.4%) | 24 (0.5%) | 79 (0.4%) |
| Waist circumference (cm) | |||
| Mean (SD) | 97 (±13) | 98 (±13) | 97 (±13) |
| Missing | 132 (1.0%) | 65 (1.3%) | 197 (1.1%) |
| WHO waist circumference classificationd | |||
| Not at risk | 6062 (47%) | 1948 (39%) | 8010 (45%) |
| At risk | 6830 (52%) | 2939 (59%) | 9769 (54%) |
| Missing | 132 (1.0%) | 65 (1.3%) | 197 (1.1%) |
| Smoking | |||
| Never | 7191 (55%) | 2731 (55%) | 9922 (55%) |
| Former | 5332 (41%) | 2027 (41%) | 7359 (41%) |
| Current | 501 (4%) | 194 (4%) | 695 (4%) |
| Alcohol | |||
| Never | 2128 (16%) | 1011 (20%) | 3139 (17%) |
| Former | 749 (6%) | 332 (7%) | 1081 (6%) |
| Current | 10,147 (78%) | 3609 (73%) | 13,756 (77%) |
| Treatment | |||
| Placebo | 6573 (50%) | 2470 (50%) | 9043 (50%) |
| Aspirin | 6451 (50%) | 2482 (50%) | 8933 (50%) |
ACR, albumin-to-creatinine ratio;AUS, Australia; BMI, body mass index; CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate; IQR, interquartile range; NA, not applicable; SBP, systolic blood pressure; SD, standard deviation; US, United States; WHO, World Health Organization.
CKD defined as either eGFR < 60 ml/min per 1.73 m2 and/or urine albumin-to-creatinine ratio ≥ 3 mg/mmol.
Hypertension defined as use of antihypertensive medication or a blood pressure ≥ 140/90 mm Hg at trial entry.
Type 2 diabetes defined as participant report, drug therapy for diabetes, or a fasting glucose ≥ 126 mg/dl or 7 mmol/l.
WHO waist circumference classification defined as “not at risk” or “at risk” using sex-specific cutoffs of 88 cm for women and 102 cm for men.
Rates of Bleeding Events According to CKD Status and Site of Bleeding
A total of 595 participants experienced a major hemorrhage event during trial follow-up, with 68 participants experiencing more than 1 event. Overall, participants experienced a total of 663 bleeding events, including 244 gastrointestinal bleeding events, 169 intracranial bleeding events, and 250 “other” bleeding events including hematuria, trauma-related bleeding, epistaxis, and gynecological bleeding. Of the 68 participants who had more than 1 bleeding event, 3 participants had 3 bleeding events and 65 participants had 2 bleeding events (details on types of events are provided in Supplementary Table S2).
Bleeding event rates according to CKD status and site of bleeding are shown in Figure 2. Participants with CKD had an increased rate of bleeding compared with those without CKD, estimated at 10.4 (95% CI: 9.2, 11.9) per 1000 py compared with 6.3 (95% CI: 5.7, 7.0) per 1000 py (RR 1.6; 95% CI: 1.4, 1.9; P < 0.0001). Rates of both gastrointestinal bleeding and other bleeding types in participants with CKD were also increased compared with those without CKD (4.6 vs. 2.5 and 4.2 vs. 2.7 per 1000 py, respectively), whereas the rates of intracranial bleeding were similar between the 2 groups (2.4 vs. 2.0 per 1000 py). Albuminuria was associated with a larger unadjusted risk for overall bleeding (RR 2.10; 95% CI: 1.70, 2.50) (Figure 2) compared with reduced eGFR (RR 1.40; 95% CI: 1.10, 1.60). This difference was also more pronounced for gastrointestinal bleeding, where rates in those with and without albuminuria were 6.3 per 1000 py (95% CI: 4.9, 8.2) versus 2.6 per 1000 py (95% CI: 2.2, 3.0), and for reduced eGFR < 60 ml/min per 1.73 m2, the rate was 4.0 per 1000 py (95% CI: 3.1, 5.2) compared with 2.8 per 1000 py (95% CI: 2.4, 3.2) for those with eGFR ≥ 60 ml/min per 1.73 m2 (Figure 2).
Figure 2.
Bleeding event rates and unadjusted rate ratios for overall bleeding and intracranial, gastrointestinal, and other bleeding events, according to CKD status, eGFR level, and albuminuria levels. CI, confidence interval; CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate; py, person-years; UACR, urinary albumin-to-creatinine ratio.
Risk of Major Hemorrhage According to CKD Status
The adjusted risk of major hemorrhage according to CKD status and the components is shown in Table 2. All estimates are adjusted for age, sex, smoking, hypertension, and aspirin use. CKD was independently associated with an increased risk of bleeding (HR 1.37; 95% CI: 1.15, 1.62; P < 0.001). Albuminuria (UACR ≥ 3 mg/mmol) was associated with an increased risk of bleeding in adjusted models (HR 1.74; 95% CI: 1.42, 2.13; P < 0.001) (Table 2), and for each doubling of the UACR, the risk of bleeding increased by 18% (95% CI: 13, 24; P < 0.001) (Supplementary Tables S3 and S4). For eGFR < 60 ml/min per 1.73 m2, the risk of bleeding was not increased (HR 1.10; 95% CI: 0.90, 1.33; P = 0.35) (Table 2 and Supplementary Table S3). However, further examination using splines suggested that the relationship between eGFR and bleeding risk was nonlinear (P = 0.009 vs. linear model, Supplementary Table S4), such that the bleeding risk increased when the eGFR fell below 40 ml/min per 1.73 m2 (Figure 3).
Table 2.
Adjusted risk of major hemorrhage according to CKD status, albuminuria, and eGFR levels
| Adjusted risk of major hemorrhage according to CKD status | |||
|---|---|---|---|
| Variablea | HR (95% CI) | P value | |
| CKD | No (Reference) | 1.00 | |
| Yes | 1.37 (1.15, 1.62) | <0.001 | |
| eGFR | ≥60 ml/min/1.73 m2 (Reference) | 1.00 | |
| <60 ml/min/1.73 m2 | 1.10 (0.90, 1.33) | 0.353 | |
| Albuminuria | <3 mg/mmol (Reference) | 1.00 | |
| ≥3 mg/mmol | 1.74 (1.42, 2.13) | <0.001 | |
CI, confidence interval; CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate; HR, hazard ratio.
Variables adjusted for clinically important risk factors for hemorrhage including age, sex, hypertension, smoking status, and aspirin use.
Figure 3.
Adjusted risk of major hemorrhage according to eGFR (solid line represents estimated hazard ratio, dotted lines represent 95% CI). CI, confidence interval; eGFR, estimated glomerular filtration rate; HR, hazard ratio.
Cumulative Incidence Curves for Bleeding According to CKD Status and Risk Factors
Incidence curves for bleeding according to CKD status are shown in Figure 4 according to overall CKD status, eGFR level, and albuminuria, inclusive of unadjusted HRs. These indicate that albuminuria is particularly associated with increased bleeding risk. In addition, cumulative incidence curves for bleeding according to other clinically important risk factors are shown in Figure 5, illustrating the increased bleeding risk over time with age, male sex, current smoking, and hypertension.
Figure 4.
Cumulative incidence curves for major hemorrhage according to overall CKD status, albuminuria, and eGFR (hazard ratios are unadjusted for other risk factors). Shading represents 95% CI around the curve. CI, confidence interval; CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate; HR, hazard ratio.
Figure 5.
Cumulative incidence curves for major hemorrhage stratified for age, sex, smoking, and HTN. Shading represents 95% CI around the curve. CI, confidence interval; HTN, hypertension.
Effect Modification of Major Hemorrhage With Aspirin Use
There was no statistically significant interaction between CKD status and aspirin use based on bleeding risk (P value for interaction = 0.66). The adjusted risk for major bleeding in the aspirin arm was 1.32 (95% CI: 1.06, 1.65) in those with CKD versus those without CKD, which was similar to the risk of bleeding in the placebo arm of 1.43 (95% CI: 1.10, 1.84) in those with CKD versus without CKD. A forest plot illustrating this is shown in Supplementary Figure S1.
Discussion
Data from this large community-based trial of older adults indicate that those with CKD, defined as either eGFR < 60 ml/min per 1.73 m2, UACR ≥ 3 mg/mmol, or both, have a 35% increased risk of major hemorrhage compared with individuals without CKD. Although both reduced eGFR and albuminuria were associated with increased risk of bleeding in unadjusted analyses, the results in the adjusted analyses were more nuanced. For albuminuria, the results remained consistent after multivariate adjustment and demonstrated increased risk of bleeding whether modeled categorically or continuously, such that for a doubling in UACR, the risk increased by 18% (95% CI: 13%, 24%). For eGFR, when assessed as a continuous variable, the risk of bleeding was nonlinear and increased once eGFR fell below 40 ml/min per 1.73 m2. These relationships may be explained by the fact that albuminuria reflects structural damage to the kidneys, whereas a slightly reduced eGFR of 55 ml/min per 1.73 m2 may be normal kidney function for an 85-year-old person. In addition, these data suggest that albuminuria, already recognized as a stronger predictor of cardiovascular risk, may be a better marker for other sequelae of CKD such as bleeding.
In addition to CKD, other important associations for increased bleeding risk included age ≥ 80 years, male sex, smoking, hypertension, and aspirin use, as shown by the cumulative incidence curves. Although a single risk factor may not increase risk to a clinically significant degree, the accumulation of risk factors multiplies the risk of bleeding, and their presence is likely to be clinically important when CKD is also present. We explored whether aspirin might be an effect modifier of the relationship between bleeding and CKD but did not find evidence that this occurred. This result is consistent with data from the Hypertension Optimal Treatment study16 and the International Polycap Study-3,17 both of which did not demonstrate any evidence of effect modification by CKD severity. However, given that both CKD and aspirin are independently associated with increasing bleeding events, clinicians should be aware of the potential additive effect. Of note, current protein pump inhibitor use was not associated with any significant reduction in bleeding risk in our cohort.
These results are generalizable to primary care and hospital outpatient clinic populations of older people because our study population of older people with milder CKD is largely reflective of those patient populations. These findings also complement and provide novel data to the current literature on bleeding risk in those with CKD, which is primarily derived from hospitalized populations and administrative databases, and are retrospective in design. For example, in a study from Taiwan18 using International Classification of Diseases codes to define CKD and upper gastrointestinal bleeding, patients with either CKD or end-stage renal disease were associated with 2-fold and 5-fold increased risk of bleeding, respectively (HR 1.95; 95% CI: 1.62, 2.35 and HR 5.24; 95% CI: 4.67, 5.86) compared with matched controls without CKD. A study of a CKD cohort from a health care database in Canada5 found a nearly 4-fold relative risk associated with albuminuria > 300 mg/g (RR 3.7; 95% CI: 3.0, 4.5) but it included younger population. In other studies, the presence of CKD increased the need for transfusion in colonic diverticular bleeding19 and was independently associated with increased bleeding following cardiac stents.20 Patients undergoing dialysis also had rates of bleeding gastroduodenal ulcers that were more than double than those of healthy people with normal renal function,21 whereas people with CKD also experienced greater rates of re-bleeding after a significant gastrointestinal bleed than the non-CKD group.22 However, none of these studies were derived from exclusively older populations or large clinical trials with adjudicated events and long-term follow-up, as occurred in ASPREE.
When albuminuria was used to define CKD, the association with bleeding risk was stronger than when eGFR was used, which is consistent with previous studies5 and data indicating that albuminuria is a stronger marker of risk of cardiovascular disease and end-stage kidney failure. This finding may hold relevance for risk prediction scores for bleeding. Current scoring systems such as Hypertension, Abnormal liver/renal function, Stroke history, Bleeding history or predisposition, Labile INR, Elderly, Drug/alcohol usage (HAS-BLED)23 use reduced eGFR to assign CKD status; however, our data suggest that albuminuria may be a more sensitive predictor. Future studies on risk prediction systems incorporating albuminuria would be of interest.
There are limitations to our study that should be mentioned. We did not capture minor bleeding events that did not result in admission to hospital or transfusion, such as minor per rectal bleeding, hematuria, or epistaxis. In addition, both eGFR and UACR were determined by a single spot collection, which may result in misclassifying participants in either direction. We considered adjusting our risk of bleeding models for severity of CKD, but we did not have enough participants within the stage-4 CKD group to accurately assess this. The eligibility criteria focused on community-dwelling, otherwise healthy older people and may limit generalizability to a CKD population in tertiary care with a high prevalence of comorbidities. Finally, the numbers of participants with non-White ethnicity are low. Strengths of this study include its large sample size, prospective event capture, and physician review of records and adjudication of bleeding events, ensuring minimal misclassification bias and missed events.
In conclusion, CKD in otherwise healthy older people is associated with elevated risk of major bleeding as albuminuria increases or when eGFR is reduced to <40 ml/min per 1.73 m2. Given the association of CKD and bleeding in older persons, preemptive strategies to reduce bleeding risk such as discontinuation of unnecessary aspirin use, smoking cessation, and blood pressure control are warranted.
Disclosure
All the authors declared no competing interests.
Acknowledgments
The ASPREE trial was supported by grants from the National Institute on Aging and the National Cancer Institute at the National Institutes of Health (grant numbers U01AG029824 and U19AG062682), the National Health and Medical Research Council of Australia (grant numbers 334047 and 1127060), the Victorian Cancer Agency (Australia), and Monash University (Australia).
Footnotes
Figure S1. CKD status, aspirin use, and bleeding risk.
Table S1A. Criteria used for adjudication of clinically significant bleeding events in the ASPREE trial. A bleeding event had to fulfill criteria in both categories A and B to be adjudicated as a confirmed event.
Table S1B. Criteria used for adjudication of hemorrhagic stroke in the ASPREE trial.
Table S2. Number of bleeding events and types in persons with more than one clinically significant bleed.
Table S3. Model estimates for risk of major hemorrhage, inclusive of clinically significant risk factors for CKD status (model A), reduced eGFR (model B), and albuminuria (model C).
Table S4. Model estimates for risk of bleeding with urinary albumin (log2 transformed) (model A) and eGFR (model B) modeled as continuous variables.
Strobe Statement.
Supplementary Material
Figure S1. CKD status, aspirin use, and bleeding risk.
Table S1A. Criteria used for adjudication of clinically significant bleeding events in the ASPREE trial. A bleeding event had to fulfill criteria in both categories A and B to be adjudicated as a confirmed event.
Table S1B. Criteria used for adjudication of hemorrhagic stroke in the ASPREE trial.
Table S2. Number of bleeding events and types in persons with more than one clinically significant bleed.
Table S3. Model estimates for risk of major hemorrhage, inclusive of clinically significant risk factors for CKD status (model A), reduced eGFR (model B), and albuminuria (model C).
Table S4. Model estimates for risk of bleeding with urinary albumin (log2 transformed) (model A) and eGFR (model B) modeled as continuous variables.
Strobe Statement.
References
- 1.GBD Chronic Kidney Disease Collaboration Global, regional, and national burden of chronic kidney disease, 1990-2017: a systematic analysis for the Global Burden of Disease Study. Lancet. 2020;395:709–733. doi: 10.1016/S0140-6736(20)30045-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Polkinghorne K.R., Wolfe R., Jachno K.M., et al. Prevalence of chronic kidney disease in the elderly using the Aspirin in Reducing Events in the Elderly study cohort. Nephrology (Carlton) 2019;24:1248–1256. doi: 10.1111/nep.13565. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Hallan S.I., Matsushita K., Sang Y., et al. Age and association of kidney measures with mortality and end-stage renal disease. JAMA. 2012;308:2349–2360. doi: 10.1001/jama.2012.16817. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Jalal D.I., Chonchol M., Targher G. Disorders of hemostasis associated with chronic kidney disease. Semin Thromb Hemost. 2010;36:34–40. doi: 10.1055/s-0030-1248722. [DOI] [PubMed] [Google Scholar]
- 5.Molnar A.O., Bota S.E., Garg A.X., et al. The risk of major hemorrhage with CKD. J Am Soc Nephrol. 2016;27:2825–2832. doi: 10.1681/ASN.2015050535. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Palmer S.C., Di Micco L., Razavian M., et al. Antiplatelet agents for chronic kidney disease. Cochrane Database Syst Rev. 2013;(2):CD008834. doi: 10.1002/14651858.CD008834.pub2. [DOI] [PubMed] [Google Scholar]
- 7.Campbell H.E., Stokes E.A., Bargo D., et al. Costs and quality of life associated with acute upper gastrointestinal bleeding in the UK: cohort analysis of patients in a cluster randomised trial. BMJ Open. 2015;5 doi: 10.1136/bmjopen-2014-007230. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.McNeil J.J., Nelson M.R., Woods R.L., et al. Effect of aspirin on all-cause mortality in the healthy elderly. N Engl J Med. 2018;379:1519–1528. doi: 10.1056/NEJMoa1803955. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.McNeil J.J., Wolfe R., Woods R.L., et al. Effect of aspirin on cardiovascular events and bleeding in the healthy elderly. N Engl J Med. 2018;379:1509–1518. doi: 10.1056/NEJMoa1805819. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.ASPREE Investigator Group Study design of ASPirin in Reducing Events in the Elderly (ASPREE): a randomized, controlled trial. Contemp Clin Trials. 2013;36:555–564. doi: 10.1016/j.cct.2013.09.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Kidney Disease: Improving Global Outcomes (KDIGO) CKD Work Group KDIGO 2012 clinical practice guideline for the evaluation and management of chronic kidney disease. Kidney Int Suppl. 2013;3:1–150. doi: 10.1038/kisup.2012.73. [DOI] [Google Scholar]
- 12.Levey A.S., Stevens L.A., Schmid C.H., et al. A new equation to estimate glomerular filtration rate. Ann Intern Med. 2009;150:604–612. doi: 10.7326/0003-4819-150-9-200905050-00006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Margolis K.L., Mahady S.E., Nelson M.R., et al. Development of a standardized definition for clinically significant bleeding in the ASPirin in Reducing Events in the Elderly (ASPREE) trial. Contemp Clin Trials Commun. 2018;11:30–36. doi: 10.1016/j.conctc.2018.05.015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Kelly D.M., Rothwell P.M. Prevention and treatment of stroke in patients with chronic kidney disease: an overview of evidence and current guidelines. Kidney Int. 2020;97:266–278. doi: 10.1016/j.kint.2019.09.024. [DOI] [PubMed] [Google Scholar]
- 15.Mahady S.E., Margolis K.L., Chan A., et al. Major GI bleeding in older persons using aspirin: incidence and risk factors in the ASPREE randomised controlled trial. Gut. 2021;70:717–724. doi: 10.1136/gutjnl-2020-321585. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Jardine M.J., Ninomiya T., Perkovic V., et al. Aspirin is beneficial in hypertensive patients with chronic kidney disease: a post-hoc subgroup analysis of a randomized controlled trial. J Am Coll Cardiol. 2010;56:956–965. doi: 10.1016/j.jacc.2010.02.068. [DOI] [PubMed] [Google Scholar]
- 17.Mann J.F.E., Joseph P., Gao P., et al. TIPS-3 investigators. Effects of aspirin on cardiovascular outcomes in patients with chronic kidney disease. Kidney Int. 2022;103:403–410. doi: 10.1016/j.kint.2022.09.023. [DOI] [PubMed] [Google Scholar]
- 18.Luo P.J., Lin X.H., Lin C.C., et al. Risk factors for upper gastrointestinal bleeding among aspirin users: an old issue with new findings from a population-based cohort study. J Formos Med Assoc. 2019;118:939–944. doi: 10.1016/j.jfma.2018.10.007. [DOI] [PubMed] [Google Scholar]
- 19.Nagata N., Niikura R., Aoki T., et al. Risk factors for adverse in-hospital outcomes in acute colonic diverticular hemorrhage. World J Gastroenterol. 2015;21:10697–10703. doi: 10.3748/wjg.v21.i37.10697. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Manzano-Fernandez S., Marin F., Pastor-Perez F.J., et al. Impact of chronic kidney disease on major bleeding complications and mortality in patients with indication for oral anticoagulation undergoing coronary stenting. Chest. 2009;135:983–990. doi: 10.1378/chest.08-1425. [DOI] [PubMed] [Google Scholar]
- 21.Sugimoto M., Sakai K., Kita M., Imanishi J., Yamaoka Y. Prevalence of Helicobacter pylori infection in long-term hemodialysis patients. Kidney Int. 2009;75:96–103. doi: 10.1038/ki.2008.508. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Liang C.M., Hsu C.N., Tai W.C., et al. Risk factors influencing the outcome of peptic ulcer bleeding in chronic kidney disease after initial endoscopic hemostasis: a nationwide cohort study. Medicine. 2016;95:e4795. doi: 10.1097/MD.0000000000004795. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Pisters R., Lane D.A., Nieuwlaat R., de Vos C.B., Crijns H.J., Lip G.Y. A novel user-friendly score (HAS-BLED) to assess 1-year risk of major bleeding in patients with atrial fibrillation: the Euro Heart Survey. Chest. 2010;138:1093–1100. doi: 10.1378/chest.10-0134. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.