Our primary objective was to assess and compare, using the UKPDS risk engine, the CHD risk of patients with DM2 who were either current or former smokers. The secondary objective was to compare these results with those obtained with the Registre Gironí del Cor (REGICOR) and Framingham scales.

A total of 156 Spanish diabetes specialists participated in this nationwide, cross-sectional, observational, epidemiological, multicenter study, which employed a stratified, multistage probability sample without replacement. The sampling frame was all health regions of the 17 Spanish autonomous communities, similar to other studies sponsored by Pfizer Spain.7 The number of clinics was proportional to the population of each region.

The study was conducted from February to December 2012. To ensure similarity in terms of sex and age range, each investigator included 8 patients with DM28 (4 current smokers and 4 ex-smokers) who visited their office for any reason. The inclusion criteria were (a) patients of either sex, (b) an age of 18 years or older and c) current smokers (daily consumers of at least 1 cigarette/day for the last month) or former smokers (at least 10 years since quitting). The exclusion criteria were (a) former smoker for fewer than 10 years, (b) a history of cardiovascular events or known nephropathy (secondary cardiovascular disease prevention) and (c) unwillingness to participate in the study. All included participants provided their informed written consent. The study was approved by the ethics committees of all the centers involved and by the Spanish Ministry of Health. The following data were recorded: (a) demographic data (province and area [rural, intermediary or urban]); (b) anthropometric variables, including height, weight and waist circumference; (c) blood pressure measured in the consultation9; (d) the last metabolic blood profile including HbA1c and lipid levels; (e) DM2 state according to the American Diabetes Association (ADA) classification; and (f) current smoking status, age when patient started smoking, duration of smoking cessation (ex-smokers) and packs/year (current smokers).

The data were collected during a single visit, and there was no follow-up. All gathered data were based on the information provided by the participants and on the data contained in the case histories. All participants continued to undergo their usual diabetic treatment indicated by their physicians.

Using the participants’ clinical data, the CHD risk at 10 years was estimated with the UKPDS equation6 using Microsoft Excel's implementation of the UKPDS Risk Engine v2.0 provided by the University of Oxford.10

This equation includes the age at diagnosis, sex, ethnicity, DM2 duration (years), HbA1c level, smoking status, systolic blood pressure (SBP) and total and high-density lipoprotein (HDL)-cholesterol. To assess the influence of the population used for the development of risk scales and their suitability for patients with diabetes, we also calculated the risks based on the Framingham equation calibrated for the Spanish population aged 35–74 years (REGICOR),11 using the online risk calculator (v2.1, April 2012), and on the updated (2008) Framingham risk equation12 using the Microsoft Excel worksheet designed by the University of Edinburgh.13 The REGICOR scale includes sex, age, presence of DM2, smoking status, total cholesterol level, SBP and diastolic blood pressure (DBP), adjusting the total result according to HDL-cholesterol levels. The Framingham scale is based on age, sex, smoking status (smoker/nonsmoker), diabetes status, SBP, total cholesterol and HDL-cholesterol. Long-term former smokers with diabetes show small differences in tobacco-related cardiovascular deaths when compared with nonsmokers. These differences are even smaller in terms of CHD,14 with a significant risk reduction observed after 10 years of smoking cessation. Accordingly, our ex-smokers were coded as nonsmokers in the REGICOR and Framingham risk scales.

The results are expressed as mean±standard deviation (SD), mean (95% confidence interval [95%CI]) or counts (percentage), unless stated otherwise. Nominal variables were analyzed with the chi-squared test or Fisher's exact test as appropriate. We performed a z transformation of the continuous variables, expressing the deviation from the mean as SD. We then used univariate one-way general linear models (GLMs) to evaluate the differences in the continuous variables between current and ex-smokers based on a single analysis. Given that there were differences between the two types of smokers using the t test, age was introduced as a covariate in the analysis of other variables. The effect of the risk scales, duration of diabetes, blood glucose control and smoking status on the estimated CHD at 10 years was analyzed using a two-way GLM, introducing age as a covariate. Significant differences among the risk scales were then analyzed by applying the Bonferroni correction. The relative risk reduction (RRR), regarded as the estimated reduction of coronary heart event risk between former and current smokers, was calculated as follows: [(current smokers estimated event rate−former smokers estimated event rate)/former smokers estimated event rate]×100. The number needed to treat (NNT) was calculated as follows: [1/(current smokers estimated event rate−former smokers estimated event rate)]. The SPSS statistical package (version 17.0)15 was employed throughout the analysis. Statistical significance was considered for p<.05.

A total of 1062 patients with DM2 were initially included, 172 of whom were subsequently excluded (Fig. 1), resulting in 890 patients ultimately analyzed. The patients were distributed to 2 separate groups according to smoking habit: 444 current smokers and 446 ex-smokers. Ninety-five percent (95%) and 96% of the active and ex-smokers were 35 years of age or older, respectively. The smokers had a lower body mass index (BMI) than ex-smokers despite having similar waist circumferences. Obesity (BMI ≥30kg/m2) was also lower in the current smokers (current vs. ex-smokers: 165 [37%] and 190 [43%]; chi-squared, 7.9, p=.019). The mean HbA1c and triglyceride levels were higher in the current smokers (Table 1Table 1).

Figure 1.

Flow of patients included in the study.

(0.2MB).

The likelihood of CHD at 10 years according to the UKPDS risk engine was significantly greater for the smokers (25.4%; 95% CI 23.6–27.2) than for the ex-smokers (20.8%; 95% CI 19.3–22.4), after adjusting for age (F, 850.2; p<.001) (Fig. 2A). The RRR of the former smokers versus the current smokers was 22% at 10 years. The estimated NNT to avoid a coronary event at 10 years was 22 individuals. Smokers also had a greater mean predicted risk of a fatal coronary event at 10 years (17.9%; 95% CI 16.3–19.5) than the ex-smokers (15.4%; 95% CI 14.0–16.9) (F, 875.2; p<.001, adjusted for age) (Fig. 2A). The RRR of the ex-smokers for a fatal coronary event at 10 years was 15% compared with the current smokers. The estimated NNT was 43 individuals. The risk of CHD at 10 years (UKPDS risk ≥20%) was lower for the former smokers (178, 40%) than for the current smokers (223, 50%) (odds ratio [OR], 1.5; 95% CI 1.2–2]; chi-squared, 9.6; p=.002), as was the risk for a fatal coronary event (114 [26%] vs. 140 [32%], respectively; OR, 1.3; 95% CI 1.0–1.8; chi-squared, 3.9; p=.049). The RRR was similar for those patients diagnosed with diabetes 10 or more years ago (36.7% [95% CI 33.6–39.8] vs. 28.7% [95% CI 26.2–31.3] for the current smokers vs. ex-smokers, respectively; RRR, 28%) and for those participants who had been diagnosed with diabetes within the past 10 years (17.2% [95% CI 15.6–18.8] vs. 13.4% [95% CI 12.0–14.8] for the current smokers vs. former smokers, respectively; RRR, 29%). In addition, the patients with poor blood glucose control (HbA1c >7%) benefited more from smoking cessation (RRR of ex-smokers vs. current smokers, 24%) than those individuals with HbA1c ≤7% (RRR of ex-smokers vs. smokers, 10%) (Fig. 2B). As for the influence of age and sex (Fig. 2C), we detected a higher RRR of predicted CHD risk in the former smokers versus current smokers, regardless of the age in men and in women under 60 years of age (Fig. 2B).

Figure 2.

(A) Estimated risk of coronary heart disease at 10 years according to the UKPDS risk engine as a function of smoking status. The data were analyzed with a one-way general linear model, introducing age as a covariate. Current smokers are represented by black circles and ex-smokers by white circles. *p<.05 for smoking status. (B) Influence of blood glucose control upon the estimated risk of coronary heart disease as a function of smoking status. The data were analyzed with a two-way general linear model, introducing age as a covariate. Current smokers are represented by black circles and ex-smokers by white circles. *p<.05 for smoking status. †p<.05 for blood glucose control status. ‡p<.05 for the interaction between smoking status and blood glucose control. (C) Influence of age and gender upon estimated relative risk reduction of coronary heart disease among ex-smokers versus smokers. Data are shown as means. Estimated risk of coronary heart disease in current smokers is represented by black circles. Arrowheads represent estimated relative risk reduction of ex-smokers vs. current smokers.

(0.22MB).

The study included patients between 35 and 75 years of age (Table 2). There were statistically significant differences among the 3 risk scales in predicting CHD risk at 10 years. The patients showed the highest estimated risk when assessed by the UKPDS risk engine, while the lowest values were observed with the REGICOR scale. Nonetheless, all equations showed a reduced estimated CHD risk at 10 years in former smokers compared with current smokers. According to the UKPDS risk engine, however, this higher risk only applies to men. The number of patients presenting a very high CHD risk (UKPDS and Framingham score ≥20% or REGICOR score ≥10%), considering all patients as a whole (current smokers vs. ex-smokers) was 207 (51%) versus 142 (37%) for the UKPDS risk engine (OR, 1.8; 95% CI 1.4–2.4]; chi-squared, 17.7; p<.001); 235 (58%) versus 101 (26%) for the Framingham scale (OR, 4.0; 95% CI 2.9–5.4]; chi-squared, 84.3; p<.001) and 198 (49%) versus 91 (23%) for the REGICOR scale (OR, 3.2; 95% CI 2.3–4.3]; chi-squared, 56.6; p<.001).