Peripheral artery disease (PAD) is an important global health problem and associated with considerably high morbidity and mortality [1]. It is a disease process resulting from obstruction of large peripheral arteries, exclusive of the coronary and intracranial cerebrovascular system, commonly due to atherosclerosis [2]. This chronic slowly progressive disease is usually characterized by occlusion of lower limb arteries ultimately causing acute or chronic limb ischemia. Although the association of PAD with higher risk of ischemic events has been identified, this particular manifestation of systemic atherosclerosis is largely under diagnosed and undertreated [2, 3]. The main systemic atherosclerotic vascular diseases, namely coronary artery disease (CAD), cerebrovascular disease (CVD) and PAD are leading causes of morbidity and mortality and all these diseases share the common pathophysiological process of atherothrombosis [4].

Advanced age, family history, smoking, diabetes mellitus, hypertension and dyslipidemia are commonly identified traditional cardiovascular risk factors of PAD [5–7]. A number of “nontraditional” risk factors for PAD have also been recognized including race and ethnicity, elevated inflammatory markers such as C-reactive protein, fibrinogen, leukocytes and interleukin-6, genetics, hypercoagulable states of altered blood levels of D-dimer, homocysteine, lipoprotein, and an abnormal waist-to-hip ratio [8]. The risk-factor identification is important because PAD is associated with reduction in functional capacity and quality of life as well as increased cardiovascular morbidity and mortality from myocardial infarction and CVA [9]. It is also associated with personal, social, and economic burden [3]. The risk factor modification plays an important role in managing patients with PAD in primary care setting and prevention of its complications [10]. Early diagnosis of PAD is essential to improve quality of life, to prevent further functional impairment, and to reduce mortality and morbidity from CAD and CVD. For sustainable preventive strategies in a country, it is mandatory to identify the prevalence of the disease and identifying country-specific modifiable risk factors. A recent study found the age and sex adjusted prevalence of PAD in Sri Lanka to be as 3.6% [11]. There are no studies on risk factors of PAD in Sri Lanka. This study was intended to identify the country specific risk factors of PAD which will help to address preventive measures of PAD in Sri Lanka.

All the selected cases (79) and controls (158) participated in the study making the response rate 100%. The mean age of cases was 64.3 years (SD 7.6) and median age was 65 years (IQR 60–70 years). The mean and median ages of controls were 64.3 years (SD 7.7) and 66 years (IQR 60–71 years) respectively.

This is the first case control study on PAD in Sri Lanka. This study was able to identify country specific risk factors of PAD which will help to address preventive measures of PAD in Sri Lanka. Use of validated ABPI cut off values in selecting cases and controls minimized any misclassification in our study. Selecting non diseased individuals as cases as well as selecting disease individuals as control leads to under estimation of the risk [28]. The cutoff levels of ABPI for selection of cases and controls were based on the finding of a validation study of ABPI [12]. All the cases of our study were incident cases detected during the community based prevalence survey [11]. As latent period of PAD is long, modification of risk factors following diagnosis becomes a major drawback if prevalent cases were selected for a case control study. Hence, we were able to minimize differential recall of possible exposure to risk factors between cases and controls in our study by only enrolling incident cases.

The final MLR model was able to predict 68–78% of the variance of PAD, indicating that over two-thirds of the variables determining PAD have been identified in our study. Thus, findings of this study are useful for healthcare providers to take preventive action to reduce the burden of disease in the country by early identification of risk factors. In the multivariate analysis, presence of history of diabetes mellitus, hypertension or dyslipidemia for 10 or more years, smoking, elevated CRP (>3 mg/dl), hyperhomocysteinemia (>15 mg/dl) were found to be independent risk factors for PAD among Sri Lankan adults. There are number of supportive research findings similar to our study. American Diabetes Association reported diabetes mellitus as the strongest risk factor of PAD [29]. National Health and Nutrition Examination Survey found Odds ratio (OR) of 2.08 for presence of DM (95% CI 1.01–4.28) [30]. In the San Diego population study the OR of diabetes mellitus for PAD was reported as 6.9 (p < 0.001) [19]. Dyslipidemia also has been identified as a significant risk factor for PAD in many studies. Ridker et al. reported dyslipidemia as an independent risk factor (RR 3.0, 95% CI 1.5–6.1) [31]. National Health and Nutrition Examination Survey also found the OR for dyslipidemia as high as 1.7 (95% CI 1.01–2.74) [30]. A cross sectional study conducted in Finland by Korhonen et al.(OR 2.3, 95% CI 1.56–6.58), Rotterdam follow up study (OR 1.32, 95% CI 1.07–1.64) and San Diego Population Study (RR 1.85, p = 0.01) have reported hypertension as a significant risk factor of PAD [19, 32, 33].

Most of the previous studies have assessed the risk for PAD only in relation to the presence or absence of diabetes mellitus, dyslipidemia and hypertension and not related to the duration of the illness. Although the exact time of onset of these conditions is difficult to estimate, at least the known duration of disease after detection provides important information about development of PAD. Therefore, we assessed the risk of PAD according to the history of duration of diabetes mellitus, dyslipidemia and hypertension. Having diabetes mellitus, dyslipidemia or hypertension for 10 years or more duration was found as significant risk factors of PAD. Korhonen et al. [32], reported that newly diagnosed pre-diabetes or diabetes per se is not associated with PAD, whereas long-lasting diabetes mellitus remains a well-established risk factor. In a prospective cohort study in USA, Joosten et al. [34] highlight that incidence of PAD significantly increases with the duration of diabetes mellitus, dyslipidemia and hypertension.

We found that ever smoking as significant risk factor for PAD in the MLR analysis. Many studies have identified smoking as a potent risk factor of PAD with consistent dose response relationship. A study carried out in Beijing, found both current smoking and former smoking (OR 1.5, 95% CI 1.1–2.1) as a significant risk factor of PAD [35]. The San Diego Population Study has identified pack year smoking more than 20 (OR 1.6, p < 0.01) and Framingham offspring study-pack year smoking of 10 years (OR 1.3, CI 1.2–1.4) as significant risk factors of PAD [19, 36]. A systematic review by Willigendael et al. found that the prevalence of PAD had increased by 2–3-fold among current smokers compare to nonsmokers and a clear dose–response relationship [37]. In our study pack year smoking was not found as a significant risk factor in the final MLR model. No significant association was found between use of alcohol and PAD in our study. Similar results have been reported in previous studies as well [33–36]. However, Strong heart study found that current alcohol consumption is negatively associated with PAD (OR 0.26, p < 0.03) [38]. However, it was based on a minority population in USA. Therefore results of Fabsitz et al. study cannot be considered conclusive.

We assessed the association between blood levels of HsCRP and homocysteine with PAD. In the present study we found elevated HsCRP (>3 mg/L) and Hyperhomocysteinemia (>15 mg/dl) to as significant risk factors of PAD. Many previous case control and follow up studies [31, 39, 40], have identified higher level of HsCRP as a significant risk factor of PAD. Garofolo et al. in a cross sectional study in Brazil [41], found that the risk of homocysteinemia for PAD was significant (OR 1.5, 95% CI 1.02–2.25). Aronow and Ahnalso reported homocysteine as a significant risk of for PAD (OR 1.12, p < 0.001) [42]. Systematic reviews carried out by Boushey et al., [43] and Khandanpour et al., [44] also found homocysteine as a significant risk factor of PAD (OR 4.3, 95% CI 1.7–6.9) supporting our results.

Literature shows inconsistent results for sex as a risk factor of PAD. Bennett et al. reported that males are at significant higher risk for PAD among migrant South Asians in UK (OR 2.81, 95% CI 1.31–6.0). In the same study they found sex was not a significant risk factor among black ethnicity the (OR 0.65, 95% CI 0.21–2.01) [45]. In contrary, Sigvant et al. found a higher prevalence of PAD among females [46]. Similar to our findings, many studies has reported non significant association between sex and PAD [19, 30]. Similarly, way inconsistent results have been reported by many studies for obesity as a risk factor of PAD. In our study BMI, HR and WHR were not found as significant risk factors of PAD. The Beijing study, reported that risk for PAD was 1.05 (95% CI 1.02–1.08) for each 1 kg/m²- increment in BMI [35]. Lu et al. has found that WHR increases the risk of PAD in men (OR 4.68, 95% CI 2.1–10.3) and higher WC increases the risk of PAD in women (OR 2.94, 95% CI 1.01–8.8) [47]. Some studies have reported BMI as a protective factor for PAD [19, 48]. However, many others reported the association between obesity and PAD was not significant [30, 33, 36, 49–53]. It should also be noted that findings in studies conducted in different populations, particularly in minority groups or in diverse settings has shown slight differences. Thus, association of obesity and sex with PAD was inconclusive.

These results demonstrate diabetes mellitus hypertension or dyslipidemia as well as smoking as country specific risk of PAD. In addition, there is a high risk of PAD among persons with high level of homocystein and HsCRP. Longer the duration or higher level exposure to these risk factors has increased the risk of PAD. These findings emphasis the need of routine screening of PAD among individuals with the identified risk factors.