Database-based study of the incidence of coronary heart disease (CHD) in type 2 diabetes patients

The complexity of the metabolic disease diabetes mellitus type 2 affects many organs in the body. Accordingly, a positive correlation is often found between the severity of diabetes and the occurrence of other diseases such as neuropathy, peripheral arterial disease (PAOD) and coronary artery disease (CHD) (Bauer T, 2011). CHD means that the coronary arteries that supply the heart muscle with oxygen are narrowed. This constriction is caused by arteriosclerosis, which in turn can arise from hypertension, hyperlipidemia and increased blood sugar levels in diabetes mellitus. In particular, the increased blood sugar damages the vessels and leads to micro- and macroangiopathies. As a result, diabetics are much more likely to develop CHD than non-diabetics (Kannel W, 1979). The risk of coronary heart disease is already increased before diabetes is diagnosed, as the pre-diabetes phase with elevated blood sugar levels exists for a long time before the confirmed diabetes diagnosis (Hu FB, 2002).

>> CHD, in turn, often represents the beginning of further, very serious cardiovascular diseases. Initially, CHD only manifests itself in the form of angina pectoris, ie an attack-like narrowing of the blood vessels, which is very painful due to a lack of oxygen in the heart muscle tissue behind the vessel. The fast-acting nitrates provide acute relief here. However, if a vessel blocks completely, a myocardial infarction occurs, which results in contractile tissue being destroyed. The remaining heart muscle tissue must compensate for this loss, which often leads to hypertrophy and thus heart failure. Further heart attacks, cardiac arrhythmias and sudden cardiac death are becoming increasingly likely. In order to prevent this process from starting in the first place, it is particularly important to prevent or at least delay the occurrence of CHD as much as possible. The present study examines which factors in everyday care influence the occurrence of CHD in type 2 diabetes patients.
methods
The IMS Disease Analyzer database, which contains patient-related diagnostic and treatment information from approximately 3,000 practicing physicians in various specialties, was used as the data source for this historical cohort study. The data is extracted from the practice computer via a standardized interface and checked for plausibility. The available information is then linked with relevant additional information such as the price of a medication, coding according to ATC (Anatomical-Therapeutic-Chemical Classification) and ICD (International Classification of Diseases) and updated monthly. The database contains only anonymized data in accordance with data protection guidelines (Ogdie A, 2011).
The analyzes carried out in comparison to recognized reference statistics ensured sufficient representativeness and validity of the Disease Analyzer patient database (Becher H, 2009).
The practices in which diabetes patients are treated were selected from the database. The selection included 1,100 general practitioners, practitioners and internists, including practices specializing in diabetology. Patients with the first dia-
Diagnosis of diabetes mellitus type 2 in the period 01/2000-12/2005 at the age of over 40 at the time of diagnosis. In addition, it was ensured that these patients could be continuously observed for at least 1 year before this diagnosis and 5 years after the diagnosis. Patients who already had a CHD diagnosis before and at the time of the first diabetes diagnosis were excluded.
All diagnoses examined were determined using the ICD 10 codes. The diagnoses observed in this analysis are listed in Table 1.
statistics
Data were analyzed using SAS version 9.2 statistical software (SAS Institute Inc., Cary, NC, USA). The probability of developing CHD within 5 years of the first diabetes diagnosis was presented using Kaplan-Meier curves. The Kaplan-Meier curves were dependent on the age groups (<=50, 51-60, 61-70, >70) and on the diabetes control (measured by the mean HbA1c value (<=7%, >7-8%, >8% ) based on all measurements of a patient in the observation period starting with the first diagnosis up to the CHD diagnosis or the last day of observation.
The multivariate analysis of the influencing factors was carried out in the form of a semi-parametric survival analysis (Cox proportional hazards model). The regression model was created in such a way that possible changes in patients within the co-variable structure could be taken into account. This meant that a patient's history could be divided into several segments that could be both left and right censored. Details on this methodology can be found in Ake & Carpenter (2003) (Ake CF).
The following variables were included in the model and examined for their influence on the probability of developing CHD: age group (<50, 51-60, 61-70, >70), gender (male vs. female), region (West Germany vs. East Germany), insurance status (private vs. statutory), community size (>100,000 vs. <=100,000), mean HbA1c value (<=7%, >7-8%, >8%), body mass index (<30 , >30-35, >35), co-diagnoses (hypertension, lipid metabolism disorder, cardiac arrhythmia, heart failure, PAD), stroke or cerebral transient ischemia, myocardial infarction, renal insufficiency, polyneuropathy, depression, sleep apnea). The influence of each variable was adjusted for the other variables in the model. The significance level was set at p<0.05.
Results
The database contained 30,402 type 2 diabetes patients who were first diagnosed between 2000 and 2005. 20,218 of these patients could be followed continuously over the next 5 years. However, only 15,096 did not have a CHD diagnosis at the time of diabetes diagnosis (see Figure 1). This means that 25.3% of diabetes patients were diagnosed with CHD before or at the same time as diabetes.
The selected patients were on average 64.5 years old and 46.3% were male. 6.2% were privately insured. 24.6% lived in East Germany. The HbA1c value calculated for the entire observation period was 6.7 and the body mass index was 30.2. Most patients already had one or more severe cardiovascular diseases or diabetic complications such as hypertension (79.1%), lipid metabolism disorders (55.1%), heart failure (18.1%), cardiac arrhythmias (22.8%) (see Table 2). Based on the patients who did not have CHD at the time of the first diabetes diagnosis, 3,370 (22.3%) of the patients were diagnosed with CHD within the next 5 years. The proportion of patients with a diagnosis of CHD was highest among patients aged 70 years and over (27.3%) and lowest among patients aged 50 and under (7.8%) (Figure 2). After adjustment for the remaining variables, increasing age also appears to be a significant influencing factor. The older the age, the greater the hazard ratio. In the 50-59 age group the value was 1.39, in 60-69 year olds it was 1.95 and in over 70 year olds it was 2.96 (Fig. 4).
Male diabetics were more frequently affected by CHD than women (24.9% vs. 20%) (Figure 3). The hazard ratio for males was 1.41.
HbA1c levels have been shown to be another influential factor. In patients with the mean HbA1c value above 8, the hazard ratio was 1.62. A higher proportion of CHD patients was found among diabetics in East Germany (HR=1.24). Other co-diagnoses have also been found to be important influencing factors. A history of myocardial infarction increased the probability of CHD almost fivefold (HR 4.67), followed by heart failure (HR 1.60), PAD (HR 1.57), diabetic polyneuropathy (HR 1.29), and cardiac arrhythmias (HR 1.22) and stroke or transient ischemic attack (TIA) (1.20); a lipid metabolism disorder was also relevant (HR 1.08). In addition, the influence of depression is not negligible (HR 1.10) (Figure 5). We did not find any significant effects of sleep apnea and hypertension.
discussion
What was striking in the analysis was that 25.3% of all patients who were diagnosed with type 2 diabetes mellitus for the first time already suffered from CHD or at least this diagnosis was made in parallel with diabetes. This result underlines the risk of developing CHD in the pre-diabetes stage as already reported in the literature (Rieder A, 2004).
Age and gender have been shown to be crucial influencing factors for the occurrence of CHD, ie the diagnosis of CHD, during the observation period. In particular, old age (>70 years) represents a high risk, both unadjusted and adjusted. While age has been investigated as a risk factor for various diseases and particularly often for cardiovascular diseases (Jurado J, 2009) (Zellweger MJ, 2009) , the connection between diabetes, age and the occurrence of CHD has not yet been clearly demonstrated in such a large cohort. The incidence rate is highest in patients >70 years old; Men are more affected. The result that men are more likely to develop CHD is confirmed by various studies, with different hormonal makeup being cited as the cause (Khaw KT, 2007) (Yang C, 2010). When it comes to CHD prevalence, the influencing factors age and gender also dominate, as a study from everyday German healthcare shows (Wittchen HU, 2007).
It is pathophysiologically obvious that co-morbidities such as cardiac arrhythmias, heart failure and TIA or events that have already occurred such as myocardial infarction and stroke are predictors of the diagnosis of CHD; Perhaps there was already a de facto CHD that was only diagnosed and thus documented during the observation period. The negative influence of depression and other mental disorders on cardiovascular diseases and in particular on CHD is known (Brown JM, 2011) (Taylor V, 2010) and was confirmed in this analysis.
The meaning and significance of the surrogate parameter HbA1c is controversial (Richter B, 2007). However, the present study showed that an HbA1c value > 8 massively increases the risk of developing CHD (HR = 1.62). This result can be explained pathophysiologically by the fact that the elevated blood sugar levels promote angiopathies. The connection between diabetes and CHD has already been impressively demonstrated in the Framingham study as well as in more recent cohort studies (Kannel W, 1979) (Harris MI, 1998).
What is astonishing is that even after adjusting for the remaining variables, more patients in East than in West Germany developed CHD. The supply situation in Germany is such that more old and sick people live in the East than in the West. But this fact should not play a role after adjusting for age, gender, co-morbidities and - as this study shows - the reliable predictor HbA1c value, which is a measure of the successful control of diabetes. Obviously, other factors that have not been measured and are particularly present in East Germany must have a negative influence on health.
While there is also a connection between sleep apnea (Martinez D, 2011) and hypertension (Soler EP, 2010) and CHD in the literature, there was no evidence of this in this retrospective study with a good 15,000 patients. <<