Coronary artery disease (CAD) is a leading global cause of mortality not only in the general population but also in patients with chronic kidney disease (CKD) (1–4). CKD has resulted in a heavy health burden in Taiwan, where a high prevalence and incidence of CKD have long been reported (5, 6). CKD is associated with arterial calcification and stiffness caused by uremic effects and electrolyte imbalance (2, 7). CKD is also accompanied by hypertension, dyslipidemia, oxidative stress, and chronic inflammation, all of which may accelerate the progression of atherosclerosis (8, 9). However, traditional risk factors cannot fully explain the shared pathogenesis between CAD and CKD (10). Therefore, the identification of new circulating biomarkers is warranted to evaluate and predict mortality in patients with CAD and CKD.

Brain-derived neurotrophic factor (BDNF) is a neurotrophic factor that protects the growth of neurons and synaptic plasticity (11–14). In addition to its neuroprotective effects (15, 16), BDNF is also important for cardiac development (17–20). A reduction in circulating BDNF levels is associated with chronic inflammation and oxidative stress (21–24). Tropomyosin-related kinase receptor B (TrkB), a receptor of BDNF, is expressed on endothelial cells and vascular smooth muscles (17, 25), and BDNF has been reported to be associated with a reduction in pulse pressure (26, 27). BDNF may be protective against mortality, most likely via TrkB signaling (28–30). An animal study conducted in rodents revealed that BDNF could prevent damage to glomerular podocytes via TrkB signaling (31), which is also thought to be essential for early renal cell differentiation, kidney structure formation, and renal function maintenance (32, 33). A low circulating BDNF concentration was reported to significantly predict the incidence of CKD in the Hyogo Sleep Cardio-Autonomic Atherosclerosis Study (34).

Although BDNF has been reported to be protective against adverse cardiovascular events and mortality in longitudinal studies (29, 35), the composite role of BDNF and CKD in the prediction of all-cause mortality in patients with established CAD remains unclear. Since BDNF has benefits in regard to the common risk factors for CAD and CKD, including arterial stiffness (26, 27), chronic inflammation (22, 23), and oxidative stress (21, 24), we hypothesized that BDNF acts as a mediator in the relationship between the heart and kidney. Therefore, in this study, we examined the composite effects of BDNF and CKD on all-cause mortality in patients with established CAD.

A total of 348 patients with CAD were enrolled in the present study. There was a positive correlation between serum BDNF levels and eGFR (correlation coefficient = 0.108, p = 0.045; Figure 1). A total of 45 primary endpoint events occurred during the median follow-up time of 6.0 years. Based on their serum BDNF levels and CKD status, all the patients were divided into four groups (Figure 2): 132 patients were in the high BDNF without CKD group (13 primary endpoints, 9.8%), 129 patients were in the low BDNF without CKD group (9 primary endpoints, 7.0%), 42 patients were in the high BDNF with CKD group (6 primary endpoints, 14.3%), and 45 patients were in the low BDNF with CKD group (17 primary endpoints, 37.8%). The baseline characteristics of these four groups are presented in Table 1. There were significant differences among these four groups according to age, UACR, eGFR, and BDNF (all p-values < 0.001), while none of the other assessed characteristics reached statistical significance, as determined by the Kruskal–Wallis test.

Table 2 presents the results of the univariable analysis of the association between risk factors and all-cause mortality. The risk of mortality was significantly different among the four groups categorized by serum BDNF levels and CKD status (p < 0.001). The patients in the low BDNF with CKD group had a higher mortality risk (37.8%) than those in the high BDNF without CKD group (9.8%), the low BDNF without CKD group (7.0%), and the high BDNF with CKD group (14.3%). In addition, the proportion of patients aged ≥60 years was higher in the mortality group than in the survival group (80.0 vs. 54.1%, p = 0.002). The prevalence of UACR ≥30 mg/g is also significantly higher in the mortality group than in the survival group (35.6 vs. 17.8%, p = 0.010). None of the other assessed risk factors exhibited statistically significant differences between the mortality and the survival groups.

Figure 3 shows that the survival rate was the lowest in the low BDNF with CKD group according to the Kaplan-Meier survival analysis (log rank test: p < 0.001). We conducted multivariable Cox regression analyses, as presented in Table 3. The risk of reaching the primary endpoint was significantly higher in the low BDNF with CKD group than in the high BDNF without CKD group (hazard ratio = 3.186, 95% CI: 1.482–6.846, p = 0.003) after adjusting for age and UACR which were the significant predictors for mortality in Table 2. We also examined the interaction between BDNF and CKD status on the primary endpoint, and the results are shown in Table 4. The variable BDNF × CKD status was found to contribute to the multivariable logistic regression model (odds ratio = 6.413, 95% CI: 1.497–27.47) and Cox regression model (hazard ratio = 3.640, 95% CI: 1.006–13.173) after adjusting for age and UACR.

Our main finding is that CKD accompanied by a lower serum BDNF concentration predicted a higher risk of all-cause mortality during a median follow-up time of 6.0 years among the 348 patients with CAD. To the best of our knowledge, this is the first study to ascertain the composite effect of low serum BDNF levels and CKD on long-term mortality. In line with our findings, Kaess et al. (35) reported that a lower serum BDNF level is predictive of a greater cardiovascular risk based on the data from the population in Framingham. Zhou et al. (39) reported that polymorphisms in the BDNF gene are associated with ischemic stroke in a genetic study of Chinese Han patients with large-arterial atherosclerosis. Jiang et al. (29) also reported that a lower circulating BDNF level significantly predicted a higher mortality rate in patients with angina pectoris. However, in these above studies, probably due to the nature of the research purposes, neither the eGFR values nor the CKD status of the study participants were characterized. Given that both the prevalence and incidence of CKD are high in Taiwan (5, 6), it might be clinically meaningful and of public health importance to investigate whether the combination of low BDNF and CKD is associated with higher mortality in patients with CAD.

The results of our analyses showed that both age and UACR were significantly different among the four groups and were also significantly associated with mortality. As age and albuminuria are related to CKD, these factors may be confounders in our analyses (39–42); this may explain the higher proportion of patients with UACR ≥30 mg/g in the mortality group than in the survival group. The above observation is in line with a previous study reporting a high UACR associated with CKD (43).

Patients in the low BDNF and CKD group have an approximately three-fold mortality risk than those in a high BDNF without CKD group after adjusting for the other important risk factors. However, the risks of long-term mortality were not significantly between the high BDNF without CKD group and the high BDNF with CKD group or the low BDNF without CKD group. The finding indicates that there may be a synergistic effect of low BDNF and CKD on long-term mortality. This contention is further supported by the multivariable logistic regression models which revealed that the significant interaction variable BDNF × CKD contributes to the regression model after the other associated risk factors are adjusted.

While several studies have ascertained the protective effect of BDNF on cardiovascular disease and mortality (29, 35, 39), relatively few studies have explored the relationship between BDNF and CKD, especially in patients with CAD. As TrkB is expressed in human kidneys (44), the signaling through this molecule probably contributes to the link between BDNF and renal pathophysiology. TrkB signaling has been reported to be essential in early kidney cell differentiation (32). Garcia-Suarez et al. (33) also suggested that TrkB is a key factor involved in renal structure and function because a significant reduction in glomerular areas, an absence of the macula densa, and an increasing number of extraglomerular mesangial cells were observed in TrkB-deficient mice (33).

Brain-derived neurotrophic factor was found to protect the kidneys from apoptosis induced by endoplasmic reticulum stress in a mouse model (45). Li et al. (31) reported that BDNF repairs podocytes by increasing the length and number of podocyte cell processes using an in vitro model of focal segmental glomerulosclerosis. The protective benefits were reflected by the in vivo improvement of proteinuria and glomerular lesions after BDNF administration in mice with adriamycin nephropathy. In the human kidney, TrkB is exclusively expressed in podocytes (31), and BDNF probably protects podocytes from injury under different conditions. BDNF is also localized in the processes and cell bodies of podocytes in patients with diabetic nephropathy (46). A lower circulating BDNF level significantly predicted the development of CKD in Japanese individuals in the Hyogo Sleep Cardio-Autonomic Atherosclerosis Study (34). Given the results of these previous studies, BDNF may be protective against kidney injury. Our observational findings on long-term survival can be partly explained by the protective effects of BDNF on the renal and cardiovascular systems. However, we were unable to collect laboratory or dialysis data during the follow-up period in the present study. Notably, decreases in both renal function and circulating BDNF concentration synergistically exacerbated mortality risk. This might be a compensatory mechanism between renal function and the effect of BDNF on long-term protection in patients with CAD. Considering that CKD is highly prevalent in patients with CAD and increases mortality (47), our study suggests that serum BDNF levels can be used as a new clinical indicator to categorize the risk of long-term mortality in patients with CAD and CKD.

Brain-derived neurotrophic factor mRNA is expressed in several human tissues, and platelets are a major source of BDNF in the peripheral blood (44, 48, 49). Activated platelets increase the BDNF release, and the short-term administration of antiplatelet drugs can decrease BDNF release from platelets into the serum (49, 50). In contrast, serum BDNF concentrations were not related to stable antiplatelet use in patients with CAD (51). As we enrolled patients with established CAD in this study, most (97.4%) of the participants had taken antiplatelet drugs at baseline assessment. Therefore, the proportion of patients using antiplatelet therapy was not significantly different across the four groups categorized by CKD and serum BDNF levels in this study. Moreover, several limitations of the present study have not been completely addressed. First, we did not directly assess the possible underlying compensatory mechanisms between renal and BDNF functions in patients with CAD. Second, DM might be a confounder in this study because it is associated with nephropathy and mortality. We excluded patients with known DM. Therefore, the results cannot be applied to the population with DM, even though some patients were found to have high glucose levels at baseline in this study. Third, glomerulonephritis associated with immune disease was excluded from the present study. However, renal biopsy was not performed, and the underlying causes of CKD were unknown. Fourth, the use of angiotensin-converting enzyme inhibitors, angiotensin II receptor antagonists, or statins at baseline did not decrease the mortality risk. These non-significant findings might have resulted from changes in medications that were not assessed during the follow-up period. Finally, age and albuminuria are related to CKD, and they might be confounders in our analyses because they were significantly different among the four study groups; these parameters were also significantly associated with mortality. Although we adjusted for these confounders in the multivariable regression analysis, we did not further group the patients according to their age and UACR level because of the limited number of cases. Future larger study cohorts are necessary to ascertain the effects of BDNF and renal function in different populations.

Chronic kidney disease accompanied by a low serum BDNF concentration is predictive of a significantly high all-cause mortality rate during a median follow-up of 6.0 years among patients with CAD. Further studies are warranted to determine the mechanism and causal relationship between BDNF and CKD on long-term mortality in patients with CAD.

The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.

The studies involving human participants were reviewed and approved by the Institutional Review Board of Taichung Veterans General Hospital. The patients/participants provided their written informed consent to participate in this study.

C-YH, WS, and I-TL contributed to the conception and design of the study. WS and I-TL collected the clinical data and revised the manuscript. C-YH organized the database, performed the statistical analyses, and drafted the manuscript. All authors have contributed to the manuscript and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.