Study Population

This prospective cohort study was conducted at the First Affiliated Hospital of Shihezi University. A total of 2,531 patients with HF were consecutively recruited from February 2013 to July 2017. The diagnosis of HF was confirmed by physical examination, laboratory tests, and echocardiography, according to the American College of Cardiology/American Heart Association guidelines.¹³ The exclusion criteria were as follows: (1) severe valvular heart disease as the primary etiology of HF; (2) life-threatening complications, such as severe liver dysfunction, kidney dysfunction (defined as serum creatinine levels >265 µmol/L), or a history of malignancy and life expectancy less than 1 year; (3) second-degree or third-degree atrioventricular block, unless a pacemaker was present; (4) acute myocardial infarction or unstable angina within 1 month before admission; and (5) refusal to participate in the follow-up process. Finally, 2,122 patients were included in the association analysis. The flowchart of the selection process is provided in Figure 1. This study was approved by Review Board of the First Affiliated Hospital of Shihezi University, and written informed consent was obtained from all participants. The investigation conformed to the principles of the Declaration of Helsinki.

View Figure

• Download as Powerpoint
• Download Figure

Clinical Follow-Up and Data Collection

In this study, follow-up was conducted every 3 months through outpatient interview or by telephone. The primary end points were defined as cardiovascular deaths or heart transplantation. Baseline demographic data included patient age; sex; and history of hypertension, diabetes, hyperlipidemia, and smoking, which were obtained from the patients’ records at the time of enrollment. All patients underwent fasting blood sampling and laboratory tests in the early morning after admission to the study. Hypertension was characterized by a systolic blood pressure higher than 140 mm Hg, a diastolic blood pressure higher than 90 mm Hg, or current treatment with an antihypertensive drug. Diabetes was identified based on a fasting plasma glucose (FPG) level greater than 7.8 mmol/L or a glucose level greater than 11.1 mmol/L at 2 hours after an oral glucose tolerance test. Hyperlipidemia was defined as a total plasma cholesterol level greater than 5.72 mmol/L or plasma triglyceride levels greater than 1.70 mmol/L. A history of smoking of more than 2 pack-years or smoking within the preceding year was defined as smoking. The use of β-blockers as an eligibility criterion was at the discretion of the treating physician and was defined as being used for more than 6 months.

Statistical Analysis

Statistical analyses were conducted using SPSS, version 15 (IBM Corp) for Windows (Microsoft Corporation). Normally distributed quantitative data were presented as mean (SD), while abnormally distributed quantitative data were presented as median (IQR). Categorical variables were presented as numbers (percentages). The SUA/Cr ratio was calculated by dividing the SUA levels by the serum creatinine levels. Participants were allocated to the following 3 groups according to SUA/Cr ratio tertile: tertile 1, SUA/Cr ratio less than 3.60; tertile 2, SUA/Cr ratio between 3.60 and 5.24; and tertile 3, SUA/Cr ratio greater than 5.24. Differences among the groups were examined using χ² tests, 1-way analysis of variance, 2‐tailed t tests, and Kruskal-Wallis tests, where appropriate. Logistic regression analysis was performed to identify risk factors for a high SUA/Cr ratio. Variables with significant effect (P < .1) in the univariate model were entered into the multivariate model.

The sample size was calculated using SPSS, version 15, software. The results indicated that this study could achieve a power of 90% at a significance level of .05 when the populations in tertiles 1, 2, and 3 were 502 individuals each. In reality, the sample sizes were 707 in tertile 1, 708 in tertile 2, and 707 in tertile 3, so the sample sizes were deemed sufficient to attain the desired power of the test. Before Cox proportional hazards regression, the Schoenfeld test was conducted, demonstrating that the Pearson correlation coefficient was 0.588 between Schoenfeld residuals and failure time rank. This finding indicates that the Cox proportional hazards regression model satisfied the proportional hazard assumption.

Cox proportional hazards regression models were used to calculate the hazard ratios (HRs) and 95% CIs for the associations between SUA/Cr ratio and the prognosis of HF with and without adjustment for covariates. Covariates with P < .1 in univariate models or with clinical importance were selected for multivariate Cox proportional hazard regression modeling. Uric acid and creatinine were not included in the multivariate Cox proportional hazard regression model to avoid collinearity and potential spurious results because the SUA/Cr ratio was calculated from these variables. Forest plots were constructed to identify whether the HRs of increased SUA/Cr ratio differed among the subgroups based on baseline characteristics, FPG levels, UA, creatinine, or β-blocker use. Statistical significance was defined as a P < .05 using 2-tailed tests.

Patient Characteristics

A total of 2,122 patients with HF were included in the final analysis (Fig. 1). Among these patients, 66% were male, and the mean (SD) age was 59.3 (14.7) years. A total of 88.9% of patients had hypertension, 32.1% had diabetes, 24.9% had hyperlipidemia, 38.4% were currently smoking, and 84.1% used β-blockers. The mean (SD) SUA/Cr ratio was 4.71 (2.09). The median (IQR) SUA level was 370 (306-439) µmol/L, and the median (IQR) creatinine level was 83 (67-103) mg/dL. Table I shows the baseline demographic and clinical characteristics of patients according to SUA/Cr ratio tertiles. Patients with higher SUA/Cr ratios were older; more likely to have a history of diabetes; and had higher levels of low-density lipoprotein cholesterol and UA but a lower percentage of β-blocker use, lower heart rates, and lower levels of creatinine (P < .05). No differences were observed among the groups in terms of sex (P < .22); the presence of hypertension (P < .94) and hyperlipidemia (P < .60); smoking status (P < .35); systolic blood pressure (P < .22) and diastolic blood pressure (P < .74); and levels of total cholesterol (P < .67), triglycerides (P < .93), high-density lipoprotein cholesterol (P < .05), and FPG (P > .29). Multivariate logistic analysis showed that older age (odds ratio, 1.02 [95% CI, 1.01-1.03]; P < .001), no β-blocker use (odds ratio, 0.68 [95% CI, 0.53-0.89]; P = .004), and a lower heart rate (odds ratio, 0.99 [95% CI, 0.99-1.00] P = .008) were associated with a higher SUA/Cr ratio (Supplemental Table I).

TABLE I. Patient Demographics^a

View Fullscreen

SUA/Cr Ratio and the Prognosis of HF

The median (IQR) follow-up time in this study was 15 (11-26) months. During the follow-up period, patients were excluded for a variety of reasons: 390 (18.4%) patients with HF died; 32 were younger than 18 years of age; 38 were diagnosed with malignant disease; and 42 patients had severe liver dysfunction, while 35 patients had severe kidney dysfunction. Patients with severe kidney dysfunction may require hemodialysis, which can affect the SUA/Cr ratio. These patients often experience numerous complications that can influence the prognosis of HF and potentially affect study findings. A total of 31 patients declined to participate in follow-up procedures, and 185 patients were lost to follow-up. The number of deaths in the lowest tertile was 100 (14.1%), in the middle tertile was 135 (19.1%), and in the highest tertile was 155 (21.9%).

Univariate Cox proportional hazard regression model analysis was conducted to evaluate the predictive value of all variables for the prognosis of HF. Table II shows that several variables were associated with the prognosis of HF, including age (HR, 1.03 [95% CI, 1.03-1.04]; P < .001), diabetes (HR, 0.62 [95% CI, 0.51-0.76]; P < .001), smoking (HR, 0.83 [95% CI, 0.70-0.99]; P = .043), β-blocker use (HR, 7.18 [95% CI, 5.46-9.44]; P < .001), total cholesterol (HR, 0.98 [95% CI, 0.97-0.98]; P < .001), triglycerides (HR, 0.68 [95% CI, 0.59-0.78]; P < .001), high-density lipoprotein cholesterol (HR, 0.61 [95% CI, 0.45-0.82]; P = .001), and low-density lipoprotein cholesterol (HR, 0.82 [95% CI, 0.73-0.92]; P = .001). Next, variables with P < .1 and variables of clinical importance (hypertension and hyperlipidemia) in the univariate model were selected for multivariate Cox proportional hazard regression modeling. To evaluate the effects of different co-variables on the association between SUA/Cr ratio and teh prognosis of HF, we created 4 adjusted models. As demonstrated in Table III, SUA/Cr ratio was associated with HF prognosis in the univariate Cox proportional hazard regression model. The highest SUA/Cr ratio was identified as an independent predictor of mortality risk in HF (HR, 1.62 [95% CI, 1.26-2.09]; P < .001). The statistical significance remained (HR, 1.71 [95% CI, 1.23-2.37]; P = .001), even after adjusting for various covariates (model 4; see Table III).

TABLE II. Univariate Analyses of the Prognostic Factors of Patients With HF

View Fullscreen

TABLE III. Association of Serum Uric Acid to Creatinine Ratio With Mortality Risk of HF in the Study Cohort

View Fullscreen

The cumulative survival rates were analyzed using the Kaplan-Meier survival curve. At the end of 1, 2, 3, 4, and 5 years, these rates were 90.7%, 84.9%, 81.4%, 75.1%, and 63.7% in tertile 1; 86%, 79.1%, 74.7%, 69.9%, and 63.8% in tertile 2; and 86%, 76.9%, 70.6%, 61.7%, and 52.3% in tertile 3 (Fig. 2). The survival rate was lower in those individuals with the highest SUA/Cr ratio than in those individuals with the lower ratio (log-rank, 14.7; P < .001).

View Figure

• Download as Powerpoint
• Download Figure

Subgroup analyses were conducted to further assess the associations between SUA/Cr ratio and the prognosis of HF in different clinically relevant subgroups. The mortality risk of HF increased with higher SUA/Cr ratios in female patients, patients with a history of hypertension and β-blocker use, and patients with SUA levels less than 428 µmol/L and creatinine levels less than 97 mg/dL (Fig. 3). Stratification by age; by history of diabetes, hyperlipidemia, and smoking; and by FPG level, however, had no obvious effect on the association between SUA/Cr ratio and the prognosis of HF.

View Figure

• Download as Powerpoint
• Download Figure

Correlation Between SUA/Cr Ratio and Clinical Characteristics

Stratified analysis was also conducted to detect the association between SUA/Cr ratio and clinical characteristics. As illustrated in Figure 4A and Figure 4B, patients in tertiles 2 and 3 exhibited reduced left ventricular ejection fraction and increased left ventricular end-diastolic diameter compared with patients in tertile 1 (P < .001). No significant differences were observed, though, in interventricular septal thickness at diastole or in left ventricular posterior wall diastolic thickness among the different groups (Fig. 4C and Fig. 4D).

View Figure

• Download as Powerpoint
• Download Figure

Study Limitations

The study had some limitations. First, the sample size was limited, so a larger cohort study should be conducted in the future. Second, this was a single-center study, and the conclusions deserve verification in other centers. Third, some unmeasured or residual confounding effects may have been present in this observational study. Fourth, this study only focused on baseline SUA and creatinine levels, disregarding any dynamic changes in SUA/Cr ratio that could have provided more valuable information for understanding the underlying mechanism. Fifth, caution should be exercised when extrapolating the findings of this study to other populations as all patients included were from China. To confirm and expand these findings, further prospective studies involving diverse populations will be required. Finally, the exact mechanism by which SUA/Cr ratio affects the prognosis of HF remains unknown.