Abstract
Background
Cardiac markers such as high-sensitivity cardiac troponin T (hs-cTnT) and N-terminal pro-B natriuretic peptide (NTproBNP) are predictors of developing acute kidney injury (AKI) during hospitalization for surgery or revascularization. However, their associations with the long-term risk of AKI in the general population are uncharacterized.
Methods
We conducted a prospective cohort study in 10 669 participants of the Atherosclerosis Risk in Communities Study (visit 4, 1996–1998, mean age, 63 years, 56% female, 22% black race) to examine the association of plasma concentrations of hs-cTnT and NTproBNP with the incident hospitalization with AKI. We used multivariable Cox regression analysis to estimate hazard ratios (HRs).
Results
During follow-up, 1907 participants had an incident hospitalization with AKI. Participants with higher concentrations of hs-cTnT had a higher risk of hospitalization with AKI in a graded fashion (adjusted HR, 1.88 [95%CI , 1.59–2.21] for ≥14 ng/L, 1.36 [1.18–1.57] for 9–13 ng/L, and 1.16 [1.03–1.30] for 5-8 ng/L compared to <5 ng/L). The graded association was also observed for NTproBNP (HR, 2.27 [1.93–2.68] for ≥272.7 pg/mL, 1.67 [1.45–1.93] for 142.4–272.6 pg/mL, and 1.31 [1.17–1.47] for 64.0-142.3 pg/mL compared to <64.0 pg/mL). The addition of hs-cTnT and NTproBNP to a model with established predictors significantly improved 10-year risk prediction for hospitalization with AKI (Δc-statistic, 0.015 [95%CI, 0.006–0.024]).
Conclusions
In middle-aged to older black and white adults in the community, higher concentrations of hs-cTnT and NTproBNP were robustly associated with an increased risk of hospitalization with AKI. These results suggest the usefulness of hs-cTnT and NT-proBNP to identify people at risk of AKI in the general population.
Keywords: Acute kidney injury, High-sensitivity cardiac troponin T, N-terminal pro-B-type natriuretic hormone, Hospitalization
Introduction
Acute kidney injury (AKI) affects nearly 4 million adults each year in the US (1). The rate of inpatient AKI has more than doubled in the last decade (2). AKI is associated with poor outcomes including prolonged hospital stay, higher mortality, and progression of kidney disease (3). The existing data on the AKI risk have been primarily focused on an early detection of AKI in critically ill patients (e.g., sepsis) (4). However, AKI often occurs in the community setting, such as from volume depletion or following the nephrotoxic exposure (e.g., nonsteroidal anti-inflammatory drugs [NSAIDs], radiocontrast agent). Thus, identifying individuals at risk for AKI in the community setting may allow for necessary preventative measures to reduce the risk of developing AKI.
Previous studies have suggested the etiological relationship between cardiovascular disease (CVD) and AKI. For example, up to 40% of patients who have undergone cardiac surgery developed AKI (5–8). AKI is also common in patients with heart failure or acute coronary syndrome (9, 10). Increased concentrations of high-sensitivity cardiac troponin T (hs-cTnT) and N-terminal pro-B natriuretic peptide (NTproBNP), biomarkers of cardiac damage and volume overload, measured at admission have been shown to predict the in-hospital development of AKI among patients undergoing major surgery or coronary revascularization (11–16).
However, the association of hs-cTnT and NTproBNP with the risk of AKI has not been characterized in the general population. If the association holds, it would expand the etiological link between cardiac abnormalities and AKI to the long-term risk in the general population, and support the clinical utilization of these markers to identify individuals at high risk for AKI. Such knowledge may also have implications on an ongoing debate on the value of cardiac markers to stratify CVD risk in the general population (17).
The objective of this study was to investigate the associations of hs-cTnT and NT-proBNP with the risk of hospitalization with AKI in a community-based cohort, the Atherosclerosis Risk in Communities (ARIC) Study.
Methods
Study Population
The ARIC Study is a prospective cohort study of 15 792 individuals aged between 45 and 64 years enrolled from 4 US communities between 1987 and 1989 (visit 1) (18). Subsequent visits occurred in 1990–1992 (visit 2), 1993–1995 (visit 3), and 1996–1998 (visit 4). For the present study, data at visit 4 were used as the baseline, because important predictors of AKI, estimated glomerular filtration rate (eGFR) and urinary albumin-to-creatinine ratio (ACR), were simultaneously measured for the first time in ARIC at this visit. Of 11 656 participants at visit 4, we excluded race other than black and white (n = 30), eGFR <15 mL/min/1.73m2 or end-stage renal disease (n = 35), missing hs-cTnT or NTproBNP (n = 78), previous history of AKI (n = 25), and missing covariates (n = 819). Thus, 10 669 participants were included in our analytic sample. The study was approved by the institutional review board at the Johns Hopkins University, and conducted in compliance with the Declaration of Helsinki. All participants provided written informed consent.
Exposures
The exposures of interest were plasma hs-cTnT and NTproBNP. Between 2010 and 2011, ARIC investigators measured plasma concentrations of hs-cTnT and NTproBNP using blood samples that were collected at visit 4 and stored at 70 °C until the assay was performed (19). The concentration of hs-cTnT was measured using a Troponin T-high sensitive STAT, Elecsys-2010 (Roche Diagnostics) on a Cobas e411 analyzer (Roche Diagnostics) with a limit of detection of 5 ng/L. The coefficient of variations was 6.0% at a mean concentration of 25 ng/L and <10% below 14 ng/L. The concentration of NTproBNP was measured using an Elecsys proBNP II immunoassay (Roche Diagnostics) on a Cobas e411 analyzer (Roche) with an assay limit of 5 pg/mL. The coefficient of variation was 5.4% at a concentration of 133 pg/mL.
Outcome
The outcome of interest was incident hospitalization with AKI, which was defined as the first hospitalization after visit 4 with an International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) code relevant to AKI (584.x) or equivalent ICD-10-CM (N17.x). In ARIC, hospitalization events are identified through annual telephone calls (semi-annual since 2012) to participants or their proxies, and active surveillance collects the information of hospitalization, including ICD codes. For the primary analysis, we included AKI at any diagnostic position as a hospital diagnosis. A previous report from ARIC Study validated that this approach had a high positive (92.0%) and negative (81.8%) predictive value (20). Participants were censored if they were lost-to-follow-up, died, or had incident end-stage renal disease. The follow-up was administratively censored at December 31, 2017.
Covariates
All covariates were based on data from visit 4 (1996–1998), except for years of education, which was based on data obtained at visit 1 (1987–1989). Demographic (age, sex, and race) and smoking history were based on a self-reported questionnaire. Body mass index was calculated by dividing weight in kilograms by height in meters squared. Sitting blood pressure was measured twice using a random-zero sphygmomanometer after 5 min of rest, and their average was used. eGFR was calculated using the Chronic Kidney Disease Epidemiology equation incorporating serum creatinine and cystatin C (21). ACR was calculated as the ratio of urinary albumin to creatinine. Plasma total cholesterol and high-density lipoprotein cholesterol (HDL-C) were measured using enzymatic methods. Diabetes was defined as a fasting glucose ≥126 mg/dL; a nonfasting glucose ≥200 mg/dL, or the reported use of antidiabetic medications. Prevalent coronary heart disease, stroke, and heart failure were defined by self-reported questionnaire at visit 1, or incident cases between visit 1 and visit 4.
Statistical Analysis
Baseline characteristics were compared by 4 hs-cTnT categories of <5, 5–8, 9–13, and ≥14 ng/L (22). These cut-offs corresponded to the 47, 77, and 91 percentiles of hs-cTnT, and we categorized NTproBNP according to the same percentiles (<64.0, 64.0–142.3, 142.4–272.6, ≥272.7 pg/mL). The cumulative incidence of hospitalization with AKI was dipicted, and to avoid the overestimation of the incidence, we accounted for death as a competing risk.
We used Poisson regression models to estimate incidence rates and 95% confidence intervals (95% CIs). Hazard ratios (HRs) were estimated using multivariable Cox regression models. Model 1 adjusted for age, sex, and race. Model 2 additionally adjusted for body mass index, ever smoking status, use of anti-hypertension medication, systolic blood pressure, total cholesterol, HDL-C, eGFR, ACR, diabetes, coronary heart disease, heart failure, and stroke. Model 3 additionally adjusted for NTproBNP in the analysis of hs-cTnT, and adjusted for hs-cTnT in the analysis of NTproBNP.
Several sensitivity analyses were performed. First, we explored the association when restricting the outcome to hospitalization with AKI as the primary diagnosis, assuming that AKI was present at the time of admission. Second, we excluded participants who had prevalent CVD at baseline, because such participants might have different response of hs-cTnT and NTproBNP (23). Third, we ran a Cox model censoring interim occurrence of CVD, or adjusting for CVD as a time-varying covariate, because the risk of AKI might be mediated by the interim occurrence of CVD. Incident CVD was defined as composite outcomes of incident coronary heart disease, stroke, or heart failure, whichever occurred first. Fourth, we performed separate analyses by years of follow-up (< 10 years vs. ≥ 10 years), since several major AKI definitions were newly proposed during the first 10 years of follow-up (24), which might have influenced the utilization of ICD codes for AKI.
Subgroup analyses were conducted for age (< vs ≥65 years), sex (male vs female), race (white vs black), diabetes (no vs yes), CKD defined by eGFR <60 mL/min/1.73m2 or ACR ≥30 mg/g (no vs yes), and prevalent CVD (no vs yes).
We also ran multivariable zero-inflated negative binomial regression models to estimate event rates of hospitalization with AKI accounting for multiple events per person, since participants could experience more than one episode of hospitalization with AKI.
Finally, we assessed the ability of hs-cTnT and NTproBNP to predict the risk of hospitalization with AKI. For this analysis, we computed the Harrell’s c-statistics in multivariable Cox models with and without the inclusion of hs-cTnT and/or NTproBNP in addition to the covariates in Model 2. In addition to the entire study period, we also explored a time horizon of 10 years, since risk prediction over a shorter, prespecified time window may be more clinically relevant (rather than the full follow-up of over ∼20 years) (25). A two-sided P-value of <0.05 was considered statistically significant. All analyses were conducted using Stata version 15 (StataCorp).
Results
Study Population
Of 10 699 participants at visit 4, the mean age was 62.8 (standard deviation [SD], 5.7) years, 56.1% were female, 21.9% were black, and 6.9% had eGFR 15–59 mL/min/1.73 m2. Table 1 shows baseline characteristics by the categories of hs-cTnT. Participants with higher concentrations of hs-cTnT were more likely to be older, male, black, and have hypertension, low eGFR, high ACR, diabetes, coronary heart disease, heart failure, and stroke. Similar patterns were generally observed among participants with higher concentrations of NTproBNP (Supplemental Table 1).
Table 1.
Baseline characteristics by categories of high-sensitivity cardiac troponin T.
| Characteristics | Overall (n = 10 669) | hs-cTnT (ng/L) |
|||
|---|---|---|---|---|---|
| <5 (n = 5064) | 5-8 (n = 3191) | 9-13 (n = 1458) | ≥14 (n = 956) | ||
| Age, year (SD) | 62.8 (5.7) | 61.2 (5.2) | 63.5 (5.6) | 65.1 (5.6) | 65.5 (5.6) |
| Female, no. (%) | 5981 (56.1) | 3722 (73.5) | 1537 (48.2) | 503 (34.5) | 219 (22.9) |
| Black race, no. (%) | 2341 (21.9) | 1070 (21.1) | 667 (20.9) | 341 (23.4) | 263 (27.5) |
| Ever smoke, no. (%) | 6239 (58.5) | 2891 (57.1) | 1819 (57.0) | 897 (61.5) | 632 (66.1) |
| BMI, kg/m2 (SD) | 28.8 (5.6) | 28.3 (5.5) | 29.0 (5.6) | 29.6 (5.6) | 29.8 (5.6) |
| Systolic BP, mmHg (SD) | 127.5 (19.0) | 125.0 (17.9) | 128.1 (18.9) | 131.5 (19.9) | 132.5 (21.4) |
| Diastolic BP, mmHg (SD) | 71.0 (10.3) | 70.6 (9.9) | 71.3 (10.2) | 71.6 (10.9) | 71.2 (11.8) |
| Antihypertensive drugs, no. (%) | 4608 (43.2) | 1813 (35.8) | 1389 (43.5) | 758 (52.0) | 648 (67.8) |
| Lab tests | |||||
| eGFR, mL/min/1.73 m2, no. (%) | |||||
| ≥90 | 4494 (42.1) | 2540 (50.2) | 1276 (40.0) | 461 (31.6) | 217 (22.7) |
| 60-89 | 5439 (51.0) | 2365 (46.7) | 1723 (54.0) | 829 (56.9) | 522 (54.6) |
| 45-59 | 519 (4.9) | 128 (2.5) | 154 (4.8) | 116 (8.0) | 121 (12.7) |
| 15-44 | 217 (2.0) | 31 (0.6) | 38 (1.2) | 52 (3.6) | 96 (10.0) |
| ACR, mg/g, no. (%) | |||||
| <10 | 8525 (79.9) | 4304 (85.0) | 2601 (81.5) | 1059 (72.6) | 561 (58.7) |
| 10-29 | 1271 (11.9) | 534 (10.5) | 359 (11.3) | 224 (15.4) | 154 (16.1) |
| 30-299 | 696 (6.5) | 204 (4.0) | 196 (6.1) | 134 (9.2) | 162 (16.9) |
| ≥300 | 177 (1.7) | 22 (0.4) | 35 (1.1) | 41 (2.8) | 79 (8.3) |
| T-Chol, mg/dL (SD) | 201 (39) | 205 (35) | 201 (35) | 197 (39) | 193 (43) |
| mmol/L (SD) | 5.2 (1.0) | 5.3 (0.9) | 5.2 (0.9) | 5.1 (1.0) | 5.0 (1.1) |
| HDL-C, mg/dL (SD) | 50 (15) | 54 (15) | 50 (15) | 46 (15) | 44 (15) |
| mmol/L (SD) | 1.3 (0.4) | 1.4 (0.4) | 1.3 (0.4) | 1.2 (0.4) | 1.1 (0.4) |
| NT-proBNP, pg/mL (SD) | 138.4 (391.7) | 89.0 (104.1) | 113.0 (201.4) | 168.8 (339.3) | 439.1 (1112.9) |
| Comorbidities, no. (%) | |||||
| Diabetes | 1765 (16.5) | 534 (10.5) | 525 (16.5) | 325 (22.3) | 381 (39.9) |
| Prevalent CHD | 888 (8.3) | 218 (4.3) | 253 (7.9) | 191 (13.1) | 226 (23.6) |
| Prevalent HF | 565 (5.3) | 167 (3.3) | 146 (4.6) | 106 (7.3) | 146 (15.3) |
| Prevalent stroke | 231 (2.2) | 67 (1.3) | 67 (2.1) | 38 (2.6) | 59 (6.2) |
Abbreviations: BMI, body mass index; BP, blood pressure; eGFR, estimated glomerular filtration rate; ACR, urinary albumin-to-creatinine ratio; T-Chol, total cholesterol; HDL-C, high-density lipoprotein cholesterol; hs-cTnT, high-sensitivity cardiac troponin T; NTproBNP, N-terminal pro-brain natriuretic peptide; CHD, coronary heart disease; HF, heart failure.
hs-cTnT, NTproBNP, and Incident AKI
During a median follow-up of 19.2 years, 1907 participants had incident hospitalization with AKI (crude incidence rate, 11.0 per 1000 person-years [95% CI, 10.5–11.5]). Figure 1 shows the cumulative incidence of hospitalization with AKI considering death as a competing risk and stratified by the 4 categories of hc-TnT and NTproBNP. For both hs-cTnT and NTproBNP, the risk of hospitalization with AKI was higher in categories with greater concentrations of hs-cTnT or NTproBNP, with an approximately 4-fold risk gradient across categories.
Fig. 1.
Cumulative incidence of hospitalization with AKI accounting for death as a competing risk according to the categories of (A) hs-cTnT and (B) NTproBNP.
In age-, sex-, and race-adjusted Cox model, participants with hs-cTnT ≥14 ng/L had a significantly higher risk of hospitalization with AKI compared to those with hs-cTnT <5 ng/L (HR, 3.51 [95% CI, 3.02–4.09]) (Model 1 in Table 2). The HRs were also significant for those with hs-cTnT 9-13 ng/L (HR, 1.87 [95% CI, 1.63–2.15]) and 5–8 ng/L (1.33 [1.18–1.49]). In age-, sex-, and race-adjusted Cox model, participants with NTproBNP ≥272.7 pg/mL had a 3.3-fold higher risk of AKI compared to participants with NTproBNP <64.0 pg/mL (HR, 3.33 [95% CI, 2.88–3.86]). The association was also significant for NTproBNP 142.4–272.6 pg/mL (HR, 1.87 [95% CI, 1.63–2.15]) and 64.0–142.3 pg/mL (1.37 [1.23–1.54]).
Table 2.
Risk of incident hospitalization with AKI across the categories of hc-cTnT and NTproBNP.
| hs-cTnT (ng/L) | ||||
|---|---|---|---|---|
| Category | <5 | 5-8 | 9-13 | ≥14 |
| Incidence ratea (95% CI) | 7.8 (7.2-8.4) | 11.3 (10.5-12.3) | 16.1 (14.5-17.9) | 25.6 (22.8-28.9) |
| Adjusted HRb (95% CI) | ||||
| Model 1 | 1 [Reference] | 1.33 (1.18-1.49) | 1.87 (1.63-2.15) | 3.51 (3.02-4.09) |
| Model 2 | 1 [Reference] | 1.16 (1.03-1.30) | 1.36 (1.18-1.57) | 1.88 (1.59-2.21) |
| Model 3 | 1 [Reference] | 1.13 (1.01-1.27) | 1.28 (1.11-1.48) | 1.68 (1.42-1.98) |
| NTproBNP (pg/mL) | ||||
| Category | <64.0 | 64.0-142.3 | 142.4-272.6 | ≥272.7 |
| Incidence ratea (95% CI) | 8.7 (8.1-9.3) | 10.9 (10.0-11.8) | 14.2 (12.7-15.9) | 23.3 (20.6-26.3) |
| HR (95% CI) | ||||
| Model 1 | 1 [Reference] | 1.37 (1.23-1.54) | 1.87 (1.63-2.15) | 3.33 (2.88-3.86) |
| Model 2 | 1 [Reference] | 1.31 (1.17-1.47) | 1.67 (1.45-1.93) | 2.27 (1.93-2.68) |
| Model 3 | 1 [Reference] | 1.28 (1.14-1.44) | 1.61 (1.40-1.86) | 2.10 (1.78-2.47) |
Crude incidence rate per 1,000 person-years.
Model 1 adjusted for age, sex, and race. Model 2 was additionally adjusted for body mass index, ever smoke, use of anti-hypertension medication, systolic blood pressure, diabetes, coronary heart disease, heart failure, stroke, total cholesterol, HDL-C, eGFR, and ACR. Model 3 additionally adjusted for NTproBNP for analysis of hs-cTnT and hs-cTnT for analysis of NTproBNP.
The additional adjustment for other confounders including eGFR and ACR somewhat attenuated the association, but the association remained strong for both hs-cTnT and NTproBNP (Model 2 in Table 2). For hs-cTnT, the HRs were 1.88 (95% CI, 1.59–2.21) for ≥14 ng/L, 1.36 (1.18–1.57) for 9–13 ng/L, and 1.16 (1.03–1.30) for 5–8 ng/L compared to <5 ng/L. For NTproBNP, the HRs were 2.27 (95% CI, 1.93–2.68) for ≥272.7 pg/mL, 1.67 (1.45–1.93) for 142.4–272.6 pg/mL, and 1.31 (1.17–1.47) for 64.0–142.3 pg/mL compared to <64.0 pg/mL. The HRs were mostly unchanged in the model including both hs-cTnT and NTproBNP categories in addition to the covariates in Model 2 (Model 3 in Table 2).
hs-cTnT and NTproBNP were independently associated with risk of hospitalization with AKI in cross-categories of hs-cTnT and NTproBNP (Fig. 2). Participants with hs-cTnT ≥14 ng/L and NTproBNP ≥272.7 pg/mL had a nearly 4-fold greater risk of hospitalizations with AKI compared to those with hs-cTnT <5 ng/L and NTproBNP <64.0 pg/mL after accounting for potential confounders (HR, 3.90 [95% CI, 2.97–5.10] in Model 2).
Fig. 2.
Hazard ratios of hospitalization with AKI in cross-category of hs-cTnT and NTproBNP. The risk of hospitalization with AKI was higher across strata of hs-cTnT and NTproBNP. The risk of AKI was highest for participants with hs-cTnT ≥14 ng/L and NTproBNP ≥272.7 pg/mL, with a nearly 4-fold greater risk as compared to those with hs-cTnT <5 ng/L and NTproBNP <64.0 pg/mL. The model was adjusted for age, sex, race, body mass index, ever smoke, use of antihypertension medication, systolic blood pressure, diabetes, coronary heart disease, heart failure, stroke, total cholesterol, HDL-C, eGFR, and ACR. Colors reflect the magnitude ofrisk of incident AKI: green, low risk; yellow, moderately increased risk; orange, high risk; red, very high risk.
In sensitivity analyses, the associations were consistent for both hs-cTnT and NTproBNP when assessing the 422 cases of hospitalization with AKI at primary position (Supplemental Table 2), excluding participants who had prevalent CVD (Supplemental Table 3), censoring the interim incident CVD cases (Supplemental Table 4), or adjusting for incident CVD as a time-varying covariate (Supplemental Table 5). The associations were statistically significant in both <10 years and ≥10 years of follow-up, with more evident results in the former (Supplemental Table 6).
Subgroup Analyses
The associations were significant regardless of the subgroups of age, sex, race, diabetes, CKD, or prevalent CVD (Fig. 3). For hs-cTnT, there was no significant interaction for any of the analyzed subgroups (P-interaction, all >0.05). For NTproBNP, the interaction was significant for age and diabetes, although the HRs were qualitatively consistent between those aged <65 and ≥65 years (HR, 1.22 [95% CI, 1.151.29] vs 1.35 [1.26–1.45]; P-interaction, 0.028) or those with and without diabetes (HR, 1.31 [1.24–1.38] vs 1.19 [1.09–1.29]; P-interaction, 0.009).
Fig. 3.
Subgroups analyses by demographic and clinical categories according to the categories of (A) hs-cTnT and (B) NTproBNP. The association of hs-cTnT and NTproBNP with the risk of hospitalization with AKI was consistently observed across the subgroups of age, sex, race, diabetes, CKD, and prevalent of CVD. Interaction was not significant except for age and diabetes, where the association of NTproBNP was stronger for age ≥65 than <65 years; and those with diabetes than without diabetes. The model was adjusted for age, sex, race, BMI, ever smoke, use of anti-hypertension medication, systolic BP, diabetes, coronary heart disease, heart failure, stroke, total cholesterol, HDL-C, eGFR, and ACR Color.
Event Rate Ratio Accounting for Multiple Events per Person
When multiple events per person were taken into account, participants in the highest quartile of hs-cTnT and NTproBNP had a 84% and 73% higher event rate of hospitalization with AKI compared to those in the lowest quartiles in the multivariable zero-inflated negative binomial regression model with the covariates in Model 2 (event rate ratio, 1.84 [95% CI, 1.51–2.23] and 1.73 [1.42–2.11], respectively) (Supplemental Table 7).
Discriminative Ability of hs-cTnT and NTproBNP to Predict the Risk of Hospitalization with AKI
In a base model with the covariates in Model 2, the Harrell c-statistic to predict the risk of hospitalization with AKI was 0.7213. The c-statistic was significantly improved when log-transformed hs-cTnT or NTproBNP was each added to the model (Δ-c-statistics, 0.0046 [95% CI, 0.0018–0.0074] or 0.0081 [0.00400.0123], respectively), and highest when both log-transformed hs-cTnT and NTproBNP were entered into the model (0.0108 [0.0063–0.0152]) (Table 3).
Table 3.
c-statistic improvement by adding hs-cTnT and NTproBNP.
| Variables | (A) During the entire follow-up |
(B) During the first 10 years of follow-up |
||||
|---|---|---|---|---|---|---|
| Harrell c-statistic | c-statistic difference from the base model | P-value for difference | Harrell c-statistic | c-statistic difference from the base model | P-value for difference | |
| Base model | 0.7213 | Reference | 0.7803 | Reference | ||
| + log hs-cTnT | 0.7259 | 0.0046 (0.0018-0.0074) | 0.001 | 0.7899 | 0.0096 (0.0021-0.0172) | 0.013 |
| + log NTproBNP | 0.7294 | 0.0081 (0.0040-0.0123) | < 0.001 | 0.7886 | 0.0083 (0.0004-0.0161) | 0.039 |
| + log hs-cTnT and NTproBNP | 0.7321 | 0.0108 (0.0063-0.0152) | < 0.001 | 0.7949 | 0.0145 (0.0056-0.0235) | 0.001 |
The base model was adjusted for age, sex, race, BMI, ever smoke, use of anti-hypertension medication, systolic BP, diabetes, coronary heart disease, heart failure, stroke, total cholesterol, HDL-C, eGFR, and ACR.
The base Harrell c-statistic was higher when restricting the predicting outcome to hospitalization with AKI that occurred within 10 years of follow-up (c-statistic, 0.7803) (Table 3). We confirmed that the addition of each or both log-transformed hs-cTnT and/or NTproBNP to the model significantly improved the prediction (e.g., Δ-c-statistics, 0.0145 [95% CI, 0.0056–0.0235] by adding both hs-cTnT and NTproBNP). The degree of c-statistic improvement for adding both hs-cTnT and NTproBNP was smaller than eGFR (Δ-c-statistics, 0.0206 [95% CI, 0.0095–0.0317]), but greater than several established risk factors for AKI such as age (0.0123 [0.0045–0.0200]), ACR (0.0077 [0.0001–0.0153]), or diabetes (0.0067 [0.0008–0.0126]).
Discussion
In this cohort, participants with higher concentrations of hs-cTnT and NTproBNP had a significantly higher risk of incident hospitalization with AKI in a graded fashion. The association was robust to several sensitivity analyses and consistent across demographic and clinical subgroups including the presence of CKD and history of CVD. The associations of hs-cTnT and NTproBNP with AKI were independent of each other, and participants with hs-cTnT ≥14 ng/L and NTproBNP ≥272.7 pg/mL had a nearly 4-fold increased risk of incident hospitalization with AKI compared to those with the reference category, even after accounting for various potential confounders. Finally, the addition of hs-cTnT and NTproBNP significantly improved the risk prediction for hospitalization with AKI by a degree greater than that for age, ACR, and diabetes.
To our knowledge, this is the first study to demonstrate the association of hs-cTnT and NTproBNP with the long-term risk of AKI in the general population. Previous studies examined the concentration of cardiac markers at admission and the short-term risk of AKI during the hospital stay (11–16). For example, the Translational Research Investigating Biomarker Endpoints in AKI (TRIBE-AKI) (26) is a large multi-center consortium exploring the utility of biomarkers in predicting risk of AKI following cardiac surgery, and has shown that higher preoperative concentrations of hs-cTnT and BNP were associated with a higher risk of postoperative AKI (13, 16).
Our findings extend the literature and demonstrate an association of cardiac biomarkers with long-term risk of AKI over 10 to 20 years in a well-documented cohort of the general population. We also rigorously adjusted for potential confounders and accounted for competing risk of death, and the potential impact of prevalent and interim incident CVD. Furthermore, we demonstrated that hs-cTnT and NTproBNP improved the prediction for developing AKI in the general population. Although the delta c-statistics were relatively small, the degree of improvement was greater than that for age, ACR, and diabetes.
Our results also extend our knowledge about the association of hs-cTnT and NTproBNP with other kidney outcomes such as end-stage renal disease and CKD progression (27, 28). Potential mechanisms underlying these associations include low cardiac output and impaired renal perfusion due to higher rates of heart failure and myocardial infarction (29, 30), neurohormonal activation (e.g., sympathetic nervous system, inflammation, oxidative stress) (31), and higher prevalence of treatment for cardiac diseases (e.g., diuretics) among individuals with increased concentrations of hs-cTnT or NTproBNP. Nonetheless, we need to acknowledge that the association of hs-cTnT or NTproBNP with the risk of AKI remained consistent even after taking account for interim CVD events during follow-up. Therefore, future studies are needed to explore other mechanisms linking these cardiac markers to AKI.
The present study has several clinical implications. Our findings suggest that hs-cTnT and NTproBNP can identify those at risk for AKI. Thus, when available, such information may be utilized in the life-style counsel (e.g., avoid volume depletion) and risk-benefit assessment prior to nephrotoxin exposure (e.g., NSAIDs, radiocontrast agents) to reduce the risk of AKI. Of note, although hs-cTnT and NTproBNP are not routinely measured in the clinical practice, some experts suggest the value of cardiac markers to stratify CVD risk in the general population (17). In addition, previous studies have demonstrated the utility of hs-cTnT and NTproBNP in risk stratification for a wide range of outcomes, such as peripheral artery disease (32), infections (33), and bleeding (34), making both hs-cTnT and NTproBNP efficient biomarkers with broad applications. Finally, studies have suggested that hs-cTnT and NTproBNP concentrations were modifiable in response to interventions such as statin therapy (35), lifestyle modification (36), and physical activity (37). Thus, whether interventions to reduce the concentration of hs-cTnT and NTproBNP can reduce the risk of AKI in the general population should be investigated in future studies.
Limitations
The present study has several limitations. First, our approach of using ICD codes to identify cases of AKI may have low sensitivity, resulting in missing some cases of AKI, particularly mild AKI not requiring hospitalization. However, ICD codes have been reliably used to identify AKI cases including in the national estimate for AKI incidence (2). Additionally, severe AKI is clinically relevant, given the positive relationship between the severity of AKI and the subsequent risk of adverse outcomes (3, 38).
Second, the nature of observational studies may not allow for excluding the possibility of residual confounding. However, we rigorously accounted for potential confounders, including major risk factors for AKI, such as eGFR, ACR, and incident CVD. Third, our study population consisted of middle to older age black and white individuals in the community, and therefore, the generalizability of our findings to other populations may require caution. Finally, whether our findings can be extrapolated to other cardiac markers, such as cardiac troponin I (39, 40), should be investigated in future studies.
Conclusions
In conclusion, in this cohort of middle-aged to older black and white adults in the community, increased concentrations of hs-cTnT and NTproBNP were independently and robustly associated with incident hospitalization with AKI. Our results suggest the usefulness of hs-cTNT and NT-proBNP as AKI relevant biomarkers to identify at risk of AKI risk in the general population.
Supplemental Material
Supplemental material is available at Clinical Chemistry online.
Supplementary Material
Nonstandard Abbreviations:
- AKI
acute kidney injury
- ARIC
Atherosclerosis Risk in Communities Study
- hs-cTnT
high-sensitivity cardiac troponin T
- NTproBNP
N-terminal pro-B natriuretic peptide.
Author Contributions
All authors confirmed they have contributed to the intellectual content of this paper and have met the following 4 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; (c) final approval of the published article; and (d) agreement to be accountable for all aspects of the article thus ensuring that questions related to the accuracy or integrity of any part of the article are appropriately investigated and resolved.
Y. Sang, statistical analysis; J. Coresh, financial support, administrative support, provision of study material or patients.
Authors’ Disclosures or Potential Conflicts of Interest
Upon manuscript submission, all authors completed the author disclosure form. Disclosures and/or potential conflicts of interest:
Employment or Leadership
None declared.
Consultant or Advisory Role
A. Shah, Philips Ultrasound; S. D. Solomon, Akros, Alnylam, Amgen, Arena, AstraZeneca, Bayer, BMS, Cardior, Cardurion, Corvia, Cytokinetics, Daiichi-Sankyo, Gilead, GSK, Ironwood, Merck, Myokardia, Novartis, Roche, Takeda, Theracos, Quantum Genetics, Cardurion, AoBiome, Janssen, Cardiac Dimensions, Sanofi-Pasteur, Tenaya, Dinaqor, Tremeau, CellProThera, Moderna; C. M. Ballantyne, Abbott, Denka Seiken, Roche; K. Matsushita, Kyowa Kirin, Akebia; R.C. Hoogeveen, Denka Seiken.
Stock Ownership
None declared.
Honoraria
None declared.
Research Funding
E. Selvin was supported by NIH grants K24DK106414 and R01DK089174 from the National Institute of Diabetes and Digestive and Kidney Diseases. This research was also supported by NIH grant R01HL134320 awarded to C. M. Ballantyne and E. Selvin by the National Heart, Lung, and Blood Institute. The Atherosclerosis Risk in Communities study has been funded in whole or in part with Federal funds from the National Heart, Lung, and Blood Institute, National Institutes of Health, Department of Health and Human Services, under Contract nos. (HHSN268201700001I, HHSN268201700002I, HHSN268201700003I, HHSN268201700005I, HHSN268201700004I). Reagents for the high sensitivity cardiac troponin T and NTproBNP assays were donated by the Roche Diagnostics Corporation. C. M. Ballantyne, funding from Abbott and Roche to institution.
Expert Testimony
None declared.
Patents
None declared.
Other Remuneration
M. Grams, payment from DCI for travel to speak at annual director meeting.
Role of Sponsor
The funding organizations played no role in the design of study, choice of enrolled patients, review and interpretation of data, preparation of manuscript, or final approval of manuscript.
Acknowledgments
The authors thank the staff and participants of the ARIC study for their important contributions.
Disclaimer: The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
References
- 1. Hoste EAJ, Kellum JA, Selby NM, Zarbock A, Palevsky PM, Bagshaw SM, et al. Global epidemiology and outcomes of acute kidney injury. Nat Rev Nephrol 2018;14:607–25. [DOI] [PubMed] [Google Scholar]
- 2. Pavkov ME, Harding JL, Burrows NR.. Trends in hospitalizations for acute kidney injury - United States, 2000-2014. MMWR Morb Mortal Wkly Rep 2018;67:289–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Coca SG, Singanamala S, Parikh CR.. Chronic kidney disease after acute kidney injury: a systematic review and meta-analysis [eng] [Meta-Analysis Research Support, N.I.H., Extramural Research Support, U.S. Gov't, Non-P.H.S. Review]. Kidney Int 2012;81:442–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Alge JL, Arthur JM.. Biomarkers of AKI: a review of mechanistic relevance and potential therapeutic implications. Clin J Am Soc Nepherol 2015;10:147–55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Mariscalco G, Lorusso R, Dominici C, Renzulli A, Sala A.. Acute kidney injury: a relevant complication after cardiac surgery. Ann Thorac Surg 2011;92:1539–47. [DOI] [PubMed] [Google Scholar]
- 6. Karkouti K, Wijeysundera DN, Yau TM, Callum JL, Cheng DC, Crowther M, et al. Acute kidney injury after cardiac surgery: focus on modifiable risk factors. Circulation 2009;119:495–502. [DOI] [PubMed] [Google Scholar]
- 7. Rosner MH, Okusa MD.. Acute kidney injury associated with cardiac surgery. Clin J Am Soc Nepherol 2006;1:19–32. [DOI] [PubMed] [Google Scholar]
- 8. Park M, Coca SG, Nigwekar SU, Garg AX, Garwood S, Parikh CR.. Prevention and treatment of acute kidney injury in patients undergoing cardiac surgery: a systematic review. Am J Nephrol 2010;31:408–18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Cruz DN. Cardiorenal syndrome in critical care: the acute cardiorenal and renocardiac syndromes. Adv Chronic Kidney Dis 2013;20:56–66. [DOI] [PubMed] [Google Scholar]
- 10. Ronco C, Haapio M, House AA, Anavekar N, Bellomo R.. Cardiorenal syndrome. J Am Coll Cardiol 2008;52:1527–39. [DOI] [PubMed] [Google Scholar]
- 11. Siniscalchi A, Gamberini L, Mordenti A, Bernardi E, Cimatti M, Riganello I, et al. Postoperative troponin T elevation as a predictor of early acute kidney injury after orthotopic liver transplantation: a preliminary retrospective study. Transplant Proc 2012;44: 1999–2001. [DOI] [PubMed] [Google Scholar]
- 12. Akgul O, Uyarel H, Pusuroglu H, Isiksacan N, Turen S, Erturk M, et al. High BNP level as risk factor for acute kidney injury and predictor of all-cause mortality in STEMI patients. Herz 2014;39:507–14. [DOI] [PubMed] [Google Scholar]
- 13. Patel UD, Garg AX, Krumholz HM, Shlipak MG, Coca SG, Sint K, et al. Preoperative serum brain natriuretic peptide and risk of acute kidney injury after cardiac surgery. Circulation 2012;125:1347–55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Zelt JGE, Mielniczuk LM, Liu PP, Dupuis JY, Chih S, Akbari A, Sun LY.. Utility of novel cardiorenal biomarkers in the prediction and early detection of congestive kidney injury following cardiac surgery. J Clin Med 2018;7: 540. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Omar AS, Mahmoud K, Hanoura S, Osman H, Sivadasan P, Sudarsanan S, et al. Acute kidney injury induces high-sensitivity troponin measurement changes after cardiac surgery. BMC Anesthesiol 2017;17:15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Bucholz EM, Whitlock RP, Zappitelli M, Devarajan P, Eikelboom J, Garg AX, et al. Cardiac biomarkers and acute kidney injury after cardiac surgery. Pediatrics 2015;135:e945. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Farmakis D, Mueller C, Apple FS.. High-sensitivity cardiac troponin assays for cardiovascular risk stratification in the general population. Eur Heart J 2020; doi: 10.1093/eurheartj/ehaa083. [DOI] [PubMed] [Google Scholar]
- 18.The Atherosclerosis Risk in Communities (ARIC) Study: design and objectives. The ARIC investigators. Am J Epidemiol 1989;129:687–702. [PubMed] [Google Scholar]
- 19. Agarwal SK, Avery CL, Ballantyne CM, Catellier D, Nambi V, Saunders J, et al. Sources of variability in measurements of cardiac troponin T in a community-based sample: the atherosclerosis risk in communities study. Clin Chem 2011;57:891–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Grams ME, Waikar SS, Macmahon B, Whelton S, Ballew SH, Coresh J.. Performance and limitations of administrative data in the identification of AKI [Eng]. Clin J Am Soc Nepherol 2014;9:682–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Inker LA, Schmid CH, Tighiouart H, Eckfeldt JH, Feldman HI, Greene T, et al. Estimating glomerular filtration rate from serum creatinine and cystatin C [eng] [Research Support, N.I.H., Extramural Validation Studies]. N Engl J Med 2012;367:20–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Filion KB, Agarwal SK, Ballantyne CM, Eberg M, Hoogeveen RC, Huxley RR, et al. High-sensitivity cardiac troponin T and the risk of incident atrial fibrillation: the Atherosclerosis Risk in Communities (ARIC) study. Am Heart J 2015;169:31–8 e3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Thygesen K, Mair J, Giannitsis E, Mueller C, Lindahl B, Blankenberg S, et al. How to use high-sensitivity cardiac troponins in acute cardiac care. Eur Heart J 2012;33:2252–7. [DOI] [PubMed] [Google Scholar]
- 24. Waikar SS, Liu KD, Chertow GM.. Diagnosis, epidemiology and outcomes of acute kidney injury. Clin J Am Soc Nepherol 2008;3:844–61. [DOI] [PubMed] [Google Scholar]
- 25. Goff DC Jr, Lloyd-Jones DM, Bennett G, Coady S, D'Agostino RB, Gibbons R, et al. 2013 ACC/AHA guideline on the assessment of cardiovascular risk: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. Circulation 2014;129:S49–73. [DOI] [PubMed] [Google Scholar]
- 26. Parikh CR, Coca SG, Thiessen-Philbrook H, Shlipak MG, Koyner JL, Wang Z,. et al. Postoperative biomarkers predict acute kidney injury and poor outcomes after adult cardiac surgery. J Am Soc Nepherol 2011;22:1748–57. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Bansal N, Katz R, Dalrymple L, de Boer I, DeFilippi C, Kestenbaum B, et al. NT-proBNP and troponin T and risk of rapid kidney function decline and incident CKD in elderly adults. Clin J Am Soc Nepherol 2015;10:205–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Kim Y, Matsushita K, Sang Y, Grams ME, Skali H, Shah AM, et al. Association of high-sensitivity cardiac troponin T and natriuretic peptide with incident ESRD: the Atherosclerosis Risk in Communities (ARIC) study. Am J Kidney Dis 2015;65:550–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Ishigami J, Grams ME, Naik RP, Caughey MC, Loehr LR, Uchida S, et al. Hemoglobin, Albuminuria, and Kidney Function in Cardiovascular Risk: The ARIC (Atherosclerosis Risk in Communities) Study. J Am Heart Assoc 2018;7:e007209. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Saunders JT, Nambi V, de Lemos JA, Chambless LE, Virani SS, Boerwinkle E, et al. Cardiac troponin T measured by a highly sensitive assay predicts coronary heart disease, heart failure, and mortality in the Atherosclerosis Risk in Communities Study. Circulation 2011;123:1367–76. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Remuzzi G, Perico N, Macia M, Ruggenenti P.. The role of renin-angiotensin-aldosterone system in the progression of chronic kidney disease. Kidney Int Suppl 2005;68:S57–65. [DOI] [PubMed] [Google Scholar]
- 32. Matsushita K, Kwak L, Yang C, Pang Y, Ballew SH, Sang Y, et al. High-sensitivity cardiac troponin and natriuretic peptide with risk of lower-extremity peripheral artery disease: the Atherosclerosis Risk in Communities (ARIC) Study. Eur Heart J 2018;39:2412–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Ishigami J, Hoogeveen RC, Ballantyne CM, Folsom AR, Coresh J, Selvin E, Matsushita K.. Associations of high-sensitivity cardiac troponin and natriuretic peptide with subsequent risk of infection in persons without cardiovascular disease: the Atherosclerosis Risk in Communities study. Am J Epidemiol 2019;188:2146–55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Mathews L, Ishigami J, Ding N, Hoogeveen RC, Kucharska-Newton A, Ballantyne CM, et al. Cardiac biomarkers and subsequent risk of hospitalization with bleeding in the community: atherosclerosis risk in communities study. J Am Heart Assoc 2020;9:e013560. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Everett BM, Zeller T, Glynn RJ, Ridker PM, Blankenberg S.. High-sensitivity cardiac troponin I and B-type natriuretic peptide as predictors of vascular events in primary prevention: impact of statin therapy. Circulation 2015;131:1851–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Chu P, Gotink RA, Yeh GY, Goldie SJ, Hunink MG.. The effectiveness of yoga in modifying risk factors for cardiovascular disease and metabolic syndrome: a systematic review and meta-analysis of randomized controlled trials. Eur J Prev Cardiolog 2016;23:291–307. [DOI] [PubMed] [Google Scholar]
- 37. deFilippi CR, de Lemos JA, Newman AB, Guralnik JM, Christenson RH, Pahor M, et al. Impact of moderate physical activity on the longitudinal trajectory of a cardiac specific biomarker of injury: results from a randomized pilot study of exercise intervention. Am Heart J 2016;179:151–6. [DOI] [PubMed] [Google Scholar]
- 38. Chertow GM, Burdick E, Honour M, Bonventre JV, Bates DW.. Acute kidney injury, mortality, length of stay, and costs in hospitalized patients [eng]. J Am Soc Nepherol 2005;16:3365–70. [DOI] [PubMed] [Google Scholar]
- 39. Omland T, Pfeffer MA, Solomon SD, de Lemos JA, Rosjo H, Saltyte Benth J, et al. Prognostic value of cardiac troponin I measured with a highly sensitive assay in patients with stable coronary artery disease. J Am Coll Cardiol 2013;61:1240–9. [DOI] [PubMed] [Google Scholar]
- 40. Hijazi Z, Siegbahn A, Andersson U, Lindahl B, Granger CB, Alexander JH, et al. Comparison of cardiac troponins I and T measured with high-sensitivity methods for evaluation of prognosis in atrial fibrillation: an ARISTOTLE substudy. Clin Chem 2015;61:368–78. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.