Relationship Between Number of Multiple Risk Factors and Coronary Artery Disease Risk With and Without Diabetes Mellitus

Abstract

It is well established that a multifactorial intervention for control of glucose, blood pressure (BP), and serum lipids has beneficial effects on macrovascular complications in patients with type 2 diabetes (T2D) (1, 2). Observational studies of T2D have also shown that composite control of glucose, BP, and lipids was associated with a lower risk for coronary artery disease (CAD) in patients with T2D (3–10), although longitudinal analyses that directly and quantitatively compared the effects of such controls in patients with diabetes mellitus (DM) and in those without DM are scarce (11–13). Recently, Rawshani et al. (13) reported that, in individuals with T2D who had five risk factors [HbA1c, low-density lipoprotein-cholesterol (LDL-C), albuminuria, smoking, and BP], achieving the target range for these factors appeared to be sufficient to reduce the risk of CAD to that of the general population (11–13). However, because that study did not investigate the association between the number of targets achieved and future CAD risk in a non-DM population, a DM and a non-DM population could not be compared in terms of the effects of management of multiple risk factors for CAD. Moreover, because it is difficult to achieve control of all modifiable risk factors in real-world settings, it can be suggested that the majority of patients would have had one or more uncontrolled risk factors for CAD. Therefore, we investigated the relationships between the degree of control of multiple risk factors according to the presence or absence of DM and subsequent CAD using the non-DM group with two controlled factors as the reference using big claims data. Among the non-DM group, the largest percentage had two factors controlled. We also assessed the combinations of those risk factors that were controlled in relation to future CAD according to glucose status.

Materials and Methods

Study design and participants

The current study analyzed data from a nationwide claims database that included 296,504 people belonging to a health insurance provider for company employees and their dependents in Japan. Details of the study have been provided elsewhere (14–16). Additional material to this manuscript is available in an online repository (17). Participants aged 18 to 72 years and followed for at least 3 years beginning on 1 April 2008 and ending on 31 March 2013 were included, with the final follow-up ending on 31 August 2016. Criteria for inclusion were age 18 to 72 years and having been followed for at least 3 years. Excluded were patients with CAD at baseline, with LDL-C levels >7.6 mmol/L, and with no health examination data, including data from blood tests. Of the 296,504 individuals with available data, 220,894 people free of CAD at baseline and with health examination data, including blood test results, were analyzed. Their classification into non-DM or any type of DM has been described elsewhere (14–16). Details of the extraction of our study participants are provided in an online repository (17).

Definitions of risk factors and outcome

Study participants were considered to be at their goal level if a specific laboratory value at baseline was at or below the following cutoff points as recommended by the Japan Diabetes Society guidelines: systolic BP <130, diastolic BP <80 mm Hg; LDL-C <3.1 mmol/L (120 mg/dL); noncurrent smoking; and HbA1c <53.0 mmol/mol (7%) regardless of treatment of DM (the first three of those cutoff points applied to non-DM) (18). The presence of CAD was determined according to claims using the International Statistical Classification of Diseases and Related Health Problems, 10th revision codes for cardiac events and procedure codes for medical interventions, such as percutaneous coronary intervention and coronary artery bypass grafting after a 1-month follow-up after the health examination (16, 19).

Statistical analyses

A Cox regression model identified the association of the number of achieved targets for risk factors for CAD events with and without DM. All analyses were performed using SPSS version 19.0 software (IBM Corp., Armonk, NY). Statistical significance was set at P < 0.05. The Ethics Committee of the Niigata University approved this study.

Results

Baseline characteristics of our study participants with and without DM are shown in Table 1. The median ± SD age was 44.0 ± 8.5 years in subjects without DM and 50.0 ± 8.2 years in subjects with DM. Over a mean follow-up duration of 4.8 years, 479 and 231 CAD events occurred in subjects without DM and subjects with DM, respectively. Of 207,286 persons without DM, 26.3%, 39.6%, and 29.0% were at target for one, two, and three factors that applied to non-DM participants, respectively, whereas of 13,608 persons with DM, 26.7%, 36.4%, 23.8%, and 6.1% were at target for one, two, three, and four factors that applied to DM participants, respectively. In both subjects with and without DM, the highest percentages of achievement of targets were those for two risk factors. Baseline characteristics according to status of achievement of individual and composite risk factor targets with and without DM are summarized in an online repository (17). Among non-DM participants, compared with those achieving each BP and LDL-C target, those not achieving each targets were older, male, and more likely to be a current smoker and to have higher BMI, systolic BP, and LDL-C levels and lower high-density lipoprotein cholesterol levels. Noncurrent smokers were younger and had almost the same systolic BP and LDL-C levels as current smokers (17). Among 13,608 individuals with DM, 36.7%, 41.1%, 55.3%, and 37.6% were at target for BP, LDL-C, HbA1c, and smoking, respectively (Table 1). Compared with the non-DM group, differences in age, male sex, and lipid or BP variables in the DM group were smaller between groups that did and did not achieve BP, LDL-C, and HbA1c targets (17). Tables 2 and 3 show the hazard ratios and 95% CIs for CAD events. After adjustment for covariates, groups that had achieved BP targets had 46% and 42% lower risks of CAD compared with those not achieving BP targets in the non-DM and DM groups, respectively. Those in the non-DM group and in the DM group who achieved the LDL-C target had 46% and 49%, respectively, lower risks for CAD events than those who did not achieve the LDL-C target. Among the DM group, achieving the HbA1c target was associated with a 35% lower risk of CAD events compared with the DM group not achieving the HbA1c target. Noncurrent smokers had 47% and 38% lower risks compared with current smokers in the non-DM and DM groups, respectively. Among the non-DM population, control of either BP, LDL-C, or current smoking was associated with an ∼45% lower risk for CAD compared with lack of control of any of those risk factors. Among the DM population, achievement of the lipid target was more closely associated with a lower risk for CAD than control of other factors, such as glucose, BP, or current smoking.

Compared with participants who had achieved two targets, the risks of CAD in those who had any one and no target achieved were two and four times greater, respectively, regardless of the presence of DM (Table 2). Reduced CAD risk was observed with more than two risk factor targets reached in both subjects with and without DM (Table 2). The effect of composite control (i.e., control of all four modifiable risk factors) was sufficient to bring CAD risk in those with DM below that for the non-DM group who had met two targets among BP, LDL-C, and current smoking, whereas the risk of CAD in the DM group with all four risk factors out of target range was 9.4 times greater than in the non-DM group who had two targets achieved (Table 2). Hazard ratios for CAD events according to type of combination of risk factor targets are available in an online repository (17). Individuals with diabetes who achieved control of one or any combination of two risk factors had a significantly increased risk for CAD compared with those without DM with all three targets achieved (17). The risks of CAD in those with DM who had any combination of two out of four risk factors (HbA1c, BP, LDL-C, smoking) were 1.7 to 2.1 times higher compared with participants in the DM group who had achieved all four targets (17), suggesting that there was little difference in the risk of CAD among combinations of any two out of four risk factors (HbA1c, BP, LDL-C, smoking).

Discussion

We present a historical prospective study to determine the degree of control of multiple risk factors according to the presence or absence of DM and to compare the effects of the number and combination of targets reached for BP, LDL-C, HbA1c, and current smoking on future CAD in real-world conditions using big claims data. Compared with those who had two targets achieved, the risks of CAD in those who had any one and no target achieved were two and four times greater, respectively, regardless of the presence or absence of DM. Reduced CAD risk was observed with increased numbers of risk factor targets reached in both DM and non-DM compared with any two targets achieved. The effect of composite control of four modifiable risk factors was sufficient to bring CAD risk in DM below subjects in the non-DM group who had achieved two targets.

Among patients without diabetes, <30% achieved control targets for BP, lipids, and smoking, suggesting that stricter control of risk factors is essential to reduce future CAD in patients without DM. Likewise, in previous observational studies in Western countries, <10% of patients with DM were at goal levels for all three factors (BP, lipids, and glucose) (8). Moreover, our study found that only 6% reached targets when four factors, which included current smoking, were considered. Despite the finding that that CAD events have been greatly reduced around the world in recent years (20–22), a substantial proportion of patients with DM have one or more of these uncontrolled risk factors (23, 24). These findings suggest that unfavorable risk factor control in DM might contribute to the development of atherosclerosis, resulting in CAD. In our study, CAD risk reduction was observed according to the number of achieved target levels of modifiable risk factors regardless of the presence or absence of DM (17). Wong et al. (8) showed the relationships between the effects of the number of achieved targets for risk factors and CAD in DM. In that study, patients with one, two, or all three risk factors for BP, lipids, and glucose at target levels (vs. none) had lower risks of CAD of 41%, 56%, and 60%, respectively (8). Recently, Rawshani et al. (13) found that patients with diabetes who had five risk factors (HbA1c, LDL-C, albuminuria, smoking, and BP) within target range appeared to have at most a marginal excess risk of death, myocardial infarction, or stroke compared with subjects without DM who had four risk factors at target levels (13). Although our study participants were younger than those in the Rawshani et al. (13) study, our study and theirs found that patients with diabetes who had risk-factor variables within the target ranges appeared to have little risk of CAD compared with non-DM groups. Composite control of modifiable risk factors contributes to reducing the risk of CAD in a relatively wide range of age groups. Moreover, to reflect more accurately what is observed in real life regarding the relationship between the number of risk factors and CAD risk with and without DM, we set as the reference the non-DM group that had two targets achieved, which represented the largest percentage of achievement of targets in that group, instead of the non-DM group having no risk factor targets achieved. As a result, the effect of composite control of four modifiable risk factors was sufficient to bring CAD risk in those with DM below that for the non-DM reference group. Unfortunately, we do not have data on albuminuria. Albuminuria had almost a twofold risk for CAD events in patients with T2D (25, 26). Further study is needed to investigate the impact of albuminuria on CAD compared with other risk factors, such as BP, lipids, blood glucose, and smoking. In our study, being at the target level for BP, lipids, or smoking reduced the risk of CAD by 45% in people without DM (Table 3). The risk of CAD in DM with any combination of two out of four risk factors (HbA1c, BP, LDL-C, smoking) was approximately two times higher compared with the DM group with all four targets achieved (17), suggesting that there was little difference in the risk of CAD with the achievement of targets for any two out of four risk factors (HbA1c, BP, LDL-C, smoking). Further studies are needed to assess differences in the level of risk according to the nature of the risk factors not at target with an adequate number of patients.

Although the recommended target values for BP and LDL-C in DM and non-DM differ among guidelines, we chose the following target values in accordance with Japanese guidelines regardless of the presence or absence of DM (18): BP <130/80 mm Hg and LDL-C <3.1 mmol/L. These results (17) did not change using target levels of BP <140/90 mm Hg and LDLC <4.1 mmol/L. Patients having any one, any two, any three, or all four factors controlled among BP, LDL-C, HbA1c, and current smoking had a decreased risk of CAD compared with those having no target achieved in both the DM and non-DM groups (17). Our definition of LDL-C control (<3.1 mmol/L) was based on Japan Diabetes Society guidelines (18). The results of trials with statins argue for more aggressive lowering of LDL-C levels for the primary prevention of CAD. The current ADA guidelines recommend an LDL-C target of <1.8 mmol/L for patients with diabetes (27). However, because the incidence of CAD in Japan is lower than in the United States and other Western countries, there was not an adequate number of patients with CAD in categories of LDL-C <1.8 mmol/L to investigate the risk for CAD. Further studies are needed to clarify the various target levels in people with and without DM.

Strengths of this study include its large sample size and accurate definition of CAD based on data from medical practices, which allowed precise identification of almost all patients with incident CAD during follow-up. However, several limitations should also be considered. (1) We showed cardiovascular risk factors as percentages, not as continuous variables, in multivariate analysis. (2) The possibility of introduction of selection bias should be considered. This study was a historical cohort and included only people who had undergone physical examinations with blood tests. In addition, determination of being at a target value for a given factor was based on a single measurement at baseline. (3) Although we evaluated the risk factors that might affect CAD, residual confounding factors may exist, such as undetected comorbidities in patients or varying degrees of T2D severity and/or duration. (4) There was no information on participants’ education levels, lifestyle choices, treatment adherence, health insurance status, socio-economic limitations, and comorbidities, all of which would affect outcomes. (5) We had no information on the symptomatic nature of CAD.

It was also not possible to identify those whose treatments had changed during follow-up; nor could the duration of diabetes and distinction between type 1 (T1D) and type 2 diabetes patients be ascertained. However, type 2 diabetes is more common than type 1 diabetes and accounts for 95% of diabetes cases in Japan.

In summary, these findings show that composite control of modifiable risk factors has a large effect in patients with and without DM. The effect was sufficient to bring CAD risk in patients with DM below that for the non-DM group who had two targets achieved.

Acknowledgments

The authors thank Mami Haga, Niigata University Faculty of Medicine, for excellent secretarial assistance and Mitsuru Hashiramoto, Mitsuru Clinic, for warm encouragement. We also thank JMDC Inc. for technical support.

Financial Support: This work was supported by the Japan Society for the Promotion of Science (18K17897 to K.F.).

Author Contributions: M.Y.-H. developed the study design, researched the data, contributed to discussions, wrote the manuscript, and reviewed and edited the manuscript. K.F. planned and supervised this research, researched the data, contributed to discussions, wrote the manuscript, and reviewed and edited the manuscript. H.S. developed the study design, contributed to discussions, and reviewed and edited the manuscript.

Disclosure Summary: The authors have nothing to disclose.

Abbreviations:

Abbreviations:
 
• BP

  blood pressure

 
• CAD

  coronary artery disease

 
• DM

  diabetes mellitus

 
• LDL-C

  low-denisty lipoprotein-cholesterol

 
• T2D

  type 2 diabetes

References and Notes