Both lower and higher levels of BUN were associated with excess risks of total and ischemic stroke independent of eGFR.
Extreme BUN levels, even within the clinically normal range, may confer potential risks of stroke previously unidentified.
A U-shaped relationship was identified between BUN/Cr and total stroke and its subtypes.
Abstract
Background and aims
It remains unclear whether extreme levels of blood urea nitrogen (BUN) and BUN to creatinine ratio (BUN/Cr) can increase future risk of stroke. We conducted this study to investigate the associations of BUN and BUN/Cr with incident stroke and its subtypes.
Methods
A total of 26,835 and 26,379 participants with a mean follow-up of 7.9 years were included to investigate the associations of BUN and BUN/Cr with incident stroke, respectively. Cox proportional hazard models were used to evaluate hazard ratios (HRs) and 95% confidence intervals (CIs) for incident stroke and its subtypes.
Results
Compared with participants in the third quintile of BUN, the adjusted HRs (95% CIs) for participants in the lowest quintile were 1.21 (1.04–1.40), 1.41 (1.18–1.68) and 1.36 (0.97–1.91) for total, ischemic and hemorrhagic stroke, respectively; while for those in the highest quintile, the corresponding HRs (95% CIs) were 1.16 (1.01–1.32), 1.30 (1.11–1.53), and 1.24 (0.90–1.71). The associations remained robust when restricting the analyses to participants within clinically normal range of BUN. For BUN/Cr, compared with participants in the third quintile, participants in the lowest quintile had significant higher risks of stroke (HRs [95% CIs] were 1.19 [1.04–1.37], 1.26 [1.07–1.48], and 1.22 [0.90–1.67] for total, ischemic and hemorrhagic stroke).
Conclusions
Both high and low levels of BUN were associated with higher risks of total and ischemic stroke. Low level of BUN/Cr was associated with excess risks of total and ischemic stroke.
Introduction
Stroke is a leading cause of mortality and disability and has contributed to increased disease burden globally. Risk factor identification and early prevention are important to reduce the incidence of stroke. As indicated by reduced estimated glomerular filtration rate (eGFR) and presence of proteinuria, chronic kidney disease has been acknowledged as a risk factor for cardiovascular disease or adverse cardiovascular outcomes in both hospital- and general population-based studies. Most previous studies used eGFR as an indicator of renal function; however, the epidemiological evidence remains inconsistent regarding whether decline in eGFR is associated with incident stroke.
Blood urea nitrogen (BUN), a waste product of protein metabolism produced in the liver and excreted by the kidney, is often used in combination with creatinine (Cr) to evaluate renal function in clinical settings. However, the fact that BUN level can be influenced by multiple factors including protein intake, corticosteroids, gastrointestinal bleeding, and dehydration, suggests a role beyond the renal function indicator. A prospective study among 3355 acute ischemic stroke patients suggested that BUN rather than eGFR was significantly associated with in-hospital mortality; another prospective study showed that higher BUN levels predicted mortality among 9420 acute coronary syndrome patients independent of eGFR, serum Cr, or estimated Cr clearance. This evidence suggested that BUN may confer additional cardiovascular risks, and could be a better predictor for cardiovascular outcomes in comparison to other renal function indexes. Moreover, most previous studies were hospital-based and investigated re-hospitalization and mortality among patients with acute coronary syndrome, heart failure, or stroke, while no prospective study has examined the relation between BUN and the risk of stroke or its subtypes among a general population without cardiovascular disease history. On the other hand, BUN/Cr, an important indicator for assessing the cause of acute renal injury and clinical situations such as gastrointestinal hemorrhage, has been linked to poor prognosis among acute ischemic stroke patients, while another study with similar design and population reported non-significant relationship. To our knowledge, no study has prospectively examined the association between BUN/Cr and incident stroke in a general population.
Therefore, we conducted the present study to investigate the associations of BUN, Cr, BUN/Cr and eGFR with incident stroke and its subtypes in a large prospective cohort of the Chinese population.
Patients and methods
2.1 Study population
We conducted the present study based on the Dongfeng-Tongji cohort in Shiyan, China. A total of 27,009 retirees from Dongfeng Motor Corporation (DMC) completed baseline questionnaires, physical examinations, and provided blood samples between September 2008 and June 2010. The cohort is scheduled to resurvey every five years, and the first follow-up was completed in 2013, with a follow-up rate of 96.2%, and another 14,120 retired workers were newly recruited. Thus, the cohort currently includes a total of 41,129 participants with baseline information.
In the present study, we excluded participants with a history of coronary heart disease, stroke, cancer, nephritis, end-stage renal disease, kidney transplant, or with eGFR <15 ml/min/1.73 m2 (potential severe kidney failure) at baseline and those who were lost to follow-up, leaving 30,936 participants. We further excluded participants with missing data on serum Cr or BUN. Hence, a total of 26,835, 27,436, 26,379 and 27,436 participants were included in the final association analyses of BUN, Cr, BUN/Cr and eGFR, respectively. The detailed criteria on participant selection are depicted in Fig. 1.
Fig. 1Flow chart for the selection of study participants. DFTJ, Dongfeng-Tongji; CHD, coronary heart disease; BUN, blood urea nitrogen; Cr, creatinine; eGFR, estimated glomerular filtration rate.
All participants gave their written informed consents, the study followed the guidelines of the Helsinki Declaration and was approved by the Ethics and Human Subject Committee of Tongji Medical College, Huazhong University of Science and Technology.
2.2 Assessment of BUN, Cr, BUN/Cr, and eGFR
We collected the peripheral venous blood samples after overnight fast. Serum BUN and Cr concentrations were determined by an Architect Ci8200 automatic analyzer (Abbott Laboratories. Abbott Park, Illinois, USA) with Abbott diagnostic reagents following standard experiment procedures provided by the manufacturer. BUN/Cr was obtained by dividing serum BUN concentration (mg/L) by serum Cr concentration (mg/L). We calculated eGFR according to Modification of Diet in Renal Disease equation based on the Chinese patients with chronic kidney disease, i.e. eGFR = 175 × Scr−1.234 × age−0.179 × 0.79 (if female).
2.3 Assessment of covariates
Baseline data on sociodemographic characteristics (including age, sex, ethnicity, educational and marital status), lifestyle factors such as smoking status (current, past, never), drinking status (current, past, never), regular exercise (yes/no), medical history of disease and medication usage were obtained by trained interviewers with semi-structured questionnaires through face-to-face interviews. Participants were identified as current smokers if they smoked at least one cigarette per day for more than half a year. Participants were identified as current drinkers if they drank at least one time per week over the last six months. Physical activity was classified as regular exercise for at least 5 times per week and at least 30 min per time over the last six months.
Physical examinations were performed by trained staff at the medical center of the affiliated hospital of DMC. Hypertension was defined as resting blood pressure ≥140/90 mmHg, or self-reported physician-diagnosed hypertension, or current usage of antihypertensive medication. Hyperlipidemia was defined as total cholesterol ≥6.22 mmol/L or triglycerides ≥2.26 mmol/L or high-density lipoprotein cholesterol <1.04 mmol/L or low-density lipoprotein cholesterol ≥4.14 mmol/L at medical examination, or self-reported physician-diagnosed hyperlipidemia, or current usage of lipid-lowering medication. Diabetes was defined as fasting glucose ≥7.0 mmol/L according to the American Diabetes Association, or self-reported physician-diagnosed diabetes, or taking oral hypoglycemic medication or insulin. Body mass index (BMI) was calculated as weight in kilogram divided by squared height in meter. Serum uric acid (SUA) concentration was determined by an Architect Ci8200 automatic analyzer (Abbott Laboratories. Abbott Park, Illinois, USA).
2.4 Ascertainment of incident stroke
All participants were covered by DMC’s health care system, which enabled us to track morbidity and mortality. The outcome of interest in the present study was incident stroke (ICD-10: I60–I61, I63–I64, I69.0-I69.1, and I69.3-I69.4), which was defined as sudden or rapid onset of a typical neurological deficit of vascular origin that persisted more than 24 h or till death that first occurred after enrollment but before December 31, 2018. Stroke subtypes were classified by physicians into ischemic stroke and hemorrhagic stroke according to computed tomography and/or magnetic resonance imaging results.
For all suspected stroke events first occurred after enrollment, a panel of physicians reviewed all available medical records and confirmed stroke diagnosis following diagnosis criteria. In the present study, stroke cases with available medical records, neuroimaging and confirmed by physicians were regarded as definite stroke cases; those whose confirmatory information was obtained by telephone or from medical insurance records but with no medical records and neuroimaging were regarded as probable cases. Both definite and probable stroke cases were used for the analyses of total stroke, whereas only definite stroke cases were included for the analyses of stroke subtypes because we cannot ascertain stroke subtypes for probable cases.
2.5 Statistics analysis
To explore the complex relationships between the indicators and incident stroke, we categorized BUN, Cr, and BUN/Cr into quintiles to examine their associations with stroke risk. We also categorized these indicators into tertiles and quartiles to explore their associations with stroke risk. We categorized eGFR into three groups including G1 (eGFR≥90 ml/min/1.73 m2) as normal eGFR level, G2 (60–90 ml/min/1.73 m2) as mildly decreased eGFR level, and G3 (15–60 ml/min/1.73 m2) as mildly-to-severely decreased eGFR level according to the KDIGO 2012 Clinical Practice Guideline for evaluation and management of chronic kidney disease. We presented baseline characteristics of participants as mean ± SD or median (interquartile range) for continuous variables, and frequency (percentage) for categorical variables. Differences across groups were compared using one-way analysis of variance tests or Kruskal-Wallis tests for continuous variables according to the data distribution, and chi-square tests for categorical variables. Person-years for each participant were calculated from the date of enrollment to the date of first stroke event, death, or December 31, 2018, whichever came first.
Cox proportional hazard models were used to evaluate the associations of BUN, Cr, BUN/Cr, and eGFR with incident stroke. The results were presented as hazard ratios (HRs) (95% confident intervals [CIs]), taking the third quintile of BUN, Cr, and BUN/Cr, and the G1 of eGFR (eGFR≥90 ml/min/1.73 m2) as reference groups, respectively. All models were stratified by age at risk (in 5-year intervals), sex, and adjusted for year of recruitment (2008–2010, 2013), BMI, smoking status, drinking status, education, regular exercise, hypertension, diabetes mellitus, hyperlipidemia, diuretics usage, and family history of stroke. When analyzing the associations of BUN with incident stroke, eGFR was further adjusted. For covariates with missing values, an additional level was added to indicate the missingness for categorical variables, whereas the missing values of continuous variables were replaced with the median value and a dichotomous variable indicating the missingness was added. We used restricted cubic spline to assess the potential nonlinear associations of BUN, Cr, BUN/Cr, and eGFR with incident stroke and its subtypes, with 3 knots placed at 5th, 50th, 95th percentiles. We set the 10th percentiles as references for BUN, Cr and BUN/Cr, and excluded BUN, Cr and BUN/Cr < 0.5th percentile or >99.5th percentile to avoid influence from extreme values. We set 90 ml/min/1.73 m2 as reference for eGFR, and excluded eGFR >150 ml/min/1.73 m2, given that a GFR of 150 ml/min/1.73 m2 is almost the upper limit of most normal GFR measurement in this middle-aged and elderly group (2 SD above the mean). Moreover, we performed stratified analyses by baseline characteristics, including age (<65 or ≥65 years), sex (male or female), BMI (<24 or ≥24 kg/m2), current smokers (yes or no), current drinkers (yes or no), hypertension (yes or no), diabetes (yes or no), and hyperlipidemia (yes or no). Potential interactions were tested by adding multiplicative interaction terms into the models. In addition, because SUA, a potential indicator of decreased renal function, has been inversely associated with vascular stiffness, we also investigated the association between SUA and incident stroke.
Several sensitivity analyses were conducted. We repeated the analyses according to tertiles and quartiles of BUN, Cr, and BUN/Cr. We evaluated the associations of BUN, Cr, BUN/Cr, and eGFR with risk of total stroke after further excluding probable stroke cases. Besides, because dietary protein intake has a great impact on blood level of BUN, we performed sensitivity analyses by additionally adjusting for dietary consumption of grain, soy foods, meats, fish, eggs, and diary, which are the main sources of protein intake for the Chinese, and each food group was dichotomized with a cut-off frequency of ≥5 times per week. In addition, we examined whether the observed relations between BUN, Cr, BUN/Cr and incident stroke persisted when restricted to participants with eGFR ≥60 ml/min/1.73 m2. Given that severe liver disease is an important cause of lower BUN level and may increase the risk of stroke, we performed a sensitivity analysis excluding individuals with liver disease or those with twice the upper limit of normal liver enzymes (alanine aminotransferase ≥80 U/L, aspartate aminotransferase ≥80 U/L, alkaline phosphatase ≥300 U/L) to control the confounding by potential liver dysfunction. We also performed the analysis of BUN and incident stroke among individuals with normal BUN levels (3.1–9.5 mmol/L).
All statistical analyses were conducted using SAS version 9.4 (SAS institute Inc., Cary, NC) and R software (version 3.5.0; R Core Team). A two-sided p < 0.05 was considered statistically significant.
Results
3.1 Characteristics of study population
Supplementary Table 1 presents the baseline characteristics of study participants according to quintiles of BUN and BUN/Cr, respectively. Among 26,835 participants eligible for BUN analyses, individuals with higher BUN level were older, more likely to be male, smokers and drinkers, with higher BMI, and had a higher proportion of hypertension, hyperlipidemia, diabetes and diuretics usage.
Among 26,379 participants eligible for BUN/Cr analyses, participants with higher BUN/Cr level were more likely to be female, non-current smokers, non-current drinkers, and had a lower proportion of hypertension, hyperlipidemia, but a higher proportion of diabetes.
The baseline characteristics among the 27,436 participants eligible for Cr and eGFR analyses are shown in Supplementary Table 2.
For analyses of BUN, Cr, and BUN/Cr, 2004 (1777 definite cases, including 1418 ischemic stroke and 359 hemorrhagic stroke, and 227 probable cases), 2083 (1853 definite cases, including 1476 ischemic stroke and 377 hemorrhagic stroke, and 230 probable cases) and 1989 (1767 definite cases, including 1411 ischemic stroke and 356 hemorrhagic stroke, and 222 probable cases) incident stroke cases occurred during a mean follow-up of 7.9, 8.0 and 7.9 years (Fig. 1), respectively. Analyses of Cr and eGFR shared the same population.
3.2 Associations of BUN, Cr, BUN/Cr, and eGFR with incident stroke
As shown in Table 1, compared with the third quintile of BUN, the multivariable-adjusted HRs (95% CIs) of total stroke were 1.21 (1.04–1.40) for the participants in the lowest quintile and 1.16 (1.01–1.32) for those in the highest quintile. When stratified by stroke subtypes, the participants in the lowest, second and the highest quintile of BUN were associated with 41% (18%–68%), 23% (3%–46%), and 30% (11%–53%) higher risks of ischemic stroke, while the HRs (95% CIs) of hemorrhagic stroke were 1.36 (0.97–1.91) for the lowest quintile, and 1.24 (0.90–1.71) for the highest quintile, respectively. Compared with subjects in the third quintile of BUN/Cr, the HRs (95% CIs) of participants in the lowest quintile were 1.19 (1.04–1.37) for total stroke, 1.26 (1.07–1.48) for ischemic stroke and 1.22 (0.90–1.67) for hemorrhagic stroke. Compared with individuals with a normal eGFR level, significantly higher risk of total stroke was observed in individuals with a mildly-to-severely decreased eGFR level (HR 1.18, 95% CI 1.01–1.39) after adjustment for potential confounders. However, compared with participants in the middle quintile of Cr, we did not observe any statistically significant association between low or high Cr level and total stroke as well as its subtypes.
Table 1Adjusted HRs (95% CIs) for incident stroke according to groups of BUN, Cr, BUN/Cr and eGFR.
a Models were stratified by age at risk (in 5-year intervals), sex, and adjusted for year of recruitment (2008–2010, 2013), body mass index, smoking status, drinking status, education level, regular exercise, hypertension, diabetes, hyperlipidemia, family history of stroke, diuretics usage, and eGFR (only for analyses of BUN).
In accordance with the categorical results, restricted cubic spline regression analyses showed U-shaped relationships of BUN, BUN/Cr and eGFR with incident total stroke (all poverall<0.05, pnonlinearity<0.05; Fig. 2). For stroke subtypes, we observed significant nonlinear associations of BUN with ischemic stroke, eGFR with hemorrhagic stroke, and BUN/Cr with both ischemic stroke and hemorrhagic stroke (Supplementary Fig. 1).
Fig. 2The restricted cubic spline curves for associations of BUN, Cr, BUN/Cr and eGFR with total stroke. The solid lines represent adjusted hazard ratios according to the change of BUN, Cr, BUN/Cr or eGFR, while the dotted lines represent the 95% confidence intervals in each panel. Knots were placed at the 5th, 50th and 95th percentiles, with the 10th percentiles set as reference for BUN, Cr and BUN/Cr, and 90 ml/min/1.73 m2 set as reference for eGFR, respectively. Covariates included in the analyses were age, sex, year of recruitment (2008–2010, 2013), body mass index, smoking status, drinking status, education level, regular exercise, hypertension, hyperlipidemia, diabetes, family history of stroke, diuretics usage, and eGFR (only for the analysis of BUN). BUN, blood urea nitrogen; Cr, creatinine; eGFR, estimated glomerular filtration rate.
3.3 Sensitivity analyses and stratified analyses
We also examined the relationships between SUA levels and incident stroke but found no significant association (Supplementary Table 3). In the sensitivity analyses, categorizing BUN, Cr, and BUN/Cr into tertiles and quartiles, excluding probable stroke cases, additionally adjusted for consumption of major food groups, excluding participants with eGFR <60 ml/min/1.73 m2, excluding individuals with potential liver dysfunction, or excluding participants with clinically abnormal level of BUN did not materially change the results (Supplementary Tables 4–10).
Results of stratified analyses showed that the association of extreme levels of BUN with total stroke risk was more evident among males or current drinkers (Fig. 3). Likewise, the association of total stroke with extreme levels of BUN/Cr appeared to be more evident among participants who were male, current drinkers, or with hypertension, or without hyperlipidemia (Fig. 3). The association of eGFR with incident stroke was more robust among males, and participants without a history of hypertension or hyperlipidemia, while no significant interaction was observed except for hyperlipidemia (p for interaction = 0.04; Supplementary Fig. 2).
Fig. 3Association of BUN and BUN/Cr with total stroke in subgroups. The HRs (95% CIs) for the extreme quintiles of BUN and BUN/Cr were calculated with the third quintile of BUN and BUN/Cr as reference groups. All models were stratified by age at risk (in 5-year intervals), sex and adjusted for year of recruitment (2008–2010, 2013), body mass index, smoking status, drinking status, education levels, regular exercise, hypertension, hyperlipidemia, diabetes, family history of stroke, diuretics usage, and eGFR (only for the analyses of BUN), except for the corresponding stratified factor. BUN, blood urea nitrogen; Cr, creatinine; eGFR, estimated glomerular filtration rate; HR, hazard ratio; CI, confidence interval.
Discussion
In this prospective cohort study, we found that both lower and higher levels of BUN were associated with greater risks of total stroke and ischemic stroke independent of eGFR levels. Moreover, BUN/Cr was associated with increased risk of total stroke and ischemic stroke in U-shaped relationships. We also found that mildly-to-severely decreased eGFR was associated with higher risk of incident total stroke.
In this study, we unexpectedly found that BUN was associated with incident stroke in a nonlinear manner, and both lower and higher levels of BUN were associated with excess risks of total stroke and ischemic stroke. In a case-control study that included 180 patients who suffered from stroke within 10 days of post cardiac surgical procedures and 180 matched controls, Martinson et al. reported that postoperative BUN was associated with higher risk of stroke. However, the study population of post-cardiac surgery patients limited the generalizability of their findings. Furthermore, previous studies have indicated that the associations between BUN level and coronary heart disease or other cardiovascular outcomes were independent of eGFR, and BUN may be a better prognostic biomarker for cardiovascular outcomes, among other renal function indicators. We previously found that higher BUN level was associated with 17% higher risk of coronary heart disease after adjusted for eGFR levels. A prospective study among 3355 patients with acute ischemic stroke found that those with higher BUN at admission had 3.75-fold higher risk of in-hospital mortality after adjusted for eGFR, while eGFR was associated with neither mortality nor adverse outcomes at discharge. Among 9420 acute coronary syndrome patients, Kirtane et al. also observed that higher BUN level was associated with greater risk of mortality, and demonstrated that the prognostic role of BUN was independent of serum Cr, estimated Cr clearance and eGFR. In line with these studies, we identified significant associations of higher BUN with risk of incident stroke after adjustment for eGFR, and the findings remained robust after excluding participants with eGFR <60 ml/min/1.73 m2. We also identified that lower BUN was associated with increased risk of stroke. All these associations remained when restricted to participants within a clinical normal range of BUN. However, only a few studies have reported the association between lower BUN level and increased risk of stroke or cardiovascular outcomes. Our results suggested that maintaining a moderate level of BUN may yield additional benefits in stroke prevention. Future studies with larger sample size are required to replicate our findings and to elucidate the proper normal range of BUN for stroke prevention.
The association of BUN/Cr with incident stroke has been less studied. In this study, we, for the first time, found that lower BUN/Cr was associated with higher risks of incident total stroke as well as ischemic stroke. A prospective study reported that BUN/Cr was significantly associated with increased risk of all-cause mortality and coronary heart disease, but not incident stroke. However, the study was conducted among 3401 hemodialysis patients, which may limit the power and generalizability given that a great number of uremic toxins, including indoxyl sulfate, p-cresyl sulfate, are increased in end-stage renal disease, and may be involved in the development of cardiovascular disease. In contrast, our findings were based on a large cohort of generally healthy individuals and investigated both stroke and its subtypes.
The association of eGFR levels with stroke remained inconsistent. The Rotterdam Study among 4937 participants reported that decreased GFR was associated with higher risk of hemorrhagic stroke in a dose-response relationship but not total or ischemic stroke, whereas a previous meta-analysis among 284,672 participants indicated that an eGFR <60 ml/min/1.73 m2 was associated with 43% higher risk of incident stroke and 98% higher risk of ischemic stroke. A prospective study among 539,287 Swedish subjects reported that declined eGFR was associated with increased risks of ischemic stroke in both men and women, while hemorrhagic stroke was only related to decreased eGFR among women. However, other prospective cohort studies reported null findings. The disparity could be explained by demographic differences in the study population, sample size, case numbers, and the numbers of participants with low eGFR. In this study, we found that low eGFR was significantly associated with incident total stroke, but not stroke subtypes. The relatively small sample size of low eGFR group may limit the statistical power.
The mechanisms underlying the association of BUN with incident stroke are largely unknown. We herein propose possible mechanisms based on currently available evidence in Supplementary Fig. 3. One potential biological pathway linking elevated BUN and stroke is through disturbances of glucose homeostasis and consequent diabetes. Among factors that can concurrently be attributed to BUN levels and stroke risk, elevated BUN level may reflect the decrease of water volume in the body in addition to impaired renal function, leading to higher risk of stroke; on the other hand, reduced BUN level may result from low protein intake or severe liver disease, both of which may increase the risk of stroke. However, in this study, low protein intake and liver dysfunction were unlikely potential confounders, because the results were robust when we adjusted for consumption of major protein sources, or excluded participants with a history of liver disease and those with two times the upper limit of normal liver enzymes.
For the association between low BUN/Cr levels and incident stroke, inflammation, renin–angiotensin–aldosterone system activation, renal sympathetic nervous system activation, oxidative stress, and endothelial dysfunction after acute renal injury may serve as the possible underlying mechanisms. Because BUN/Cr is a composite index of BUN and Cr, the added value in addition to BUN in stroke pathology remains unclear and should be further explored in future studies. Multiple explanations exist regarding the association between decreased eGFR and increased risk of stroke. One potential principal pathway is endothelial cell dysfunction, other conditions including reduced nitric oxide production, enhanced nitric oxide inactivation, inflammation, and oxidative stress in renal insufficiency may also contribute. Moreover, dysregulated toxic solutes excretion and mineral metabolism could be other potential mechanisms. Further studies are warranted to explore the underlying mechanisms of the associations of BUN, BUN/Cr and eGFR with incident stroke.
The strengths of the present study are worth noting. To our knowledge, this is the first prospective study evaluating the associations of levels of BUN and BUN/Cr with risk of stroke in a large cohort of a generally healthy population. Notably, extreme levels of BUN were significantly associated with incident total stroke and ischemic stroke even among individuals within normal ranges of BUN. Nevertheless, our study also has several limitations. This is a single-center study, and our findings were based on middle-aged and older Chinese, which may limit the generalization of the results. In addition, due to the decline in muscle mass among middle-aged and elderly people, the eGFR level derived from serum Cr concentration may be overestimated, and a diagnosis of acute renal injury cannot be made because the Cr concentration was measured only once. Meanwhile, although we additionally adjusted for consumption frequencies of main sources of protein intake in the Chinese population in the sensitivity analyses, the amount of protein intake, which has major impact on BUN levels, was not available in our study and could not completely rule out the possible effect of low protein diet on stroke risk. Moreover, although we have accounted for several known confounding factors, we cannot fully exclude residual confounding.
In conclusion, we found that lower levels of BUN and BUN/Cr could be independent risk factors of incident total and ischemic stroke. A higher level of BUN also increased the risk of total stroke and ischemic stroke. More importantly, the associations of BUN with total stroke and ischemic stroke were independent of eGFR and remained significant among participants with normal or mildly decreased eGFR levels. Monitoring and maintaining a moderate BUN level may yield benefits in the prevention of stroke, especially ischemic stroke. Further studies with large sample size are warranted to validate our findings and elucidate the underlying mechanisms.
Leave A Comment