Arterial stiffness in hypertensive and type 2 diabetes patients in Ghana: comparison of the cardio-ankle vascular index and central aortic techniques

Epidemiological transition has been underway for many years in most lower/middle income countries, as in Ghana. Chronic adult diseases including hypertension, obesity, resulting metabolic syndrome and diabetes are now common [1, 2]. Consequently, premature macro- and micro-vascular conditions, including hypertensive heart failure (if not yet much coronary disease), stroke, renal failure, and arterial complications of Type 2 diabetes (T2D) now burden patients and health services [3, 4]. Arterial stiffness is a powerful predictor of such future cardiovascular (CV) complications and all-cause mortality in all patient groups studied, even after adjusting for age, blood pressure (BP) or other risk factors [5, 6]. In addition, arterial stiffness is one of the earliest detectable manifestations of adverse structural and functional changes within the vessel wall [7]. The gold standard for measuring arterial (aortic) stiffness is carotid-femoral pulse wave velocity [8]. Cardio-ankle vascular index (CAVI), an index of arterial stiffness developed in Japan from measuring heart-ankle PWV (haPWV) and brachial BP, is reported to be independent from BP at the time of measurement; hence CAVI may indicate organic stiffness of the arterial wall, minimising the influence of BP fluctuations [9].

There is little arterial function research in sub-Saharan Africa population, and none to our knowledge, on CAVI and haPWV in patients with T2D or hypertension [10]. We used CAVI and another calibrated cuff-based device, the Arteriograph, to measure aortic PWV to test the hypothesis that arterial stiffness would progressively increase with increasing burden of CV risk, mainly as having hypertension or Type 2 diabetes in Ghanaians.

Our findings suggest that these indices of arterial stiffness, CAVI, haPWV from which CAVI is derived, and PWVao are significantly increased in Ghanaian patients with diabetes and hypertension more than in those with either alone, or without both conditions. Our data here are the first from a sub-Saharan African patient population where generally, atheromatous disease is still uncommon; haemorrhagic stroke, hypertensive heart and renal failure remain the major causes of vascular events [15]. There was no difference in CAVI and PWVao values between diabetes without hypertension, and hypertension patients without diabetes, suggesting that both conditions are similarly related to arterial pathology, even when the CAVI computation adjusts for BP differences. That conclusion is supported by the similar heart-ankle pathway PWV data, which has to be extracted from the software (based on the time delay between the phonocardiogram’s 2^nd sound to arrival of the pressure wave at the ankle). While the PWV of the so-called brachial-ankle arterial pathway, in general use in Japan, is prognostic [16], its cardiac-brachial component is not.

The key importance of arterial stiffness is that its prognostic value persists when other known risk factors including any BP (systolic, pulse or mean) have been taken into consideration [17]. Cruickshank et al. [18] showed how arterial stiffness, measured by aortic PWV, predicted mortality in T2D patients, in those with ‘impaired’ glycaemia and in glucose-challenged controls and hence that arterial PWV could be considered an integrated index of vascular health. That sample included African-Caribbeans, who still have lower coronary heart disease, but not stroke, risk than the British population average; similar to but not as extreme a difference as in Ghana. CAVI, a novel index, is thought to reflect collective overall compliance in the aorta, femoral and popliteal arteries, independent of BP, based on the formula quoted above (Methods) [9]. CAVI has also been shown to be superior to ankle-brachial PWV in predicting cardiovascular events. This superiority of CAVI had been attributed to its independence, theoretically, from BP at the time of measurement, so that CAVI may indicate the intrinsic stiffness of the arterial wall [9, 14, 19]. CAVI may also be a useful screening tool for moderate to advanced levels of arterio-, rather than athero-sclerosis [20], the major difference between traditionally high ‘BP’- and plaque- driven pathology. We found that patients with co-existent diabetes and hypertension have dramatic reduction arterial compliance, or increase stiffness, than patients with single condition. The lack of difference in CAVI/haPWV but a clearer difference in the Arteriograph’s PWVao between the T2D without hypertension and hypertension group without T2D suggest less arteriosclerosis in the more muscular peripheral arteries in the legs here. In indigenous black South Africans, where smoking rate and atherosclerosis remain (currently) low, arterial stiffness as carotid-femoral PWV is still pressure-dependent [21]. Such data illustrate how PWV generally and CAVI may add value to a routine BP ‘check’. Without sub-Saharan intervention trials, their routine use will not yet be judged ‘cost-effective’.

Most studies where arterial stiffness was evaluated using CAVI have been performed in non-African, South-East Asian populations [9, 14, 19, 20]. In agreement with our findings, Wang et al [22] showed that CAVI is elevated in a Chinese population with coexistent hypertension and diabetes. The dynamics of CAVI, as an index of arterial stiffness in sub-Saharan Africans, may differ from that in Asians or Caucasians [23], as found by Uurtuya et al in comparable healthy young subjects [24] and hypertensive T2D patients in Japan and Mongolia [25]. Arterial stiffness measured by PWV here should reflect arterial function in these patient samples, but due to the relatively small number of controls, estimating general stiffness in sub-Saharan African populations at large will need community-based population studies.

The major determinants of PWVao in this study were heart rate, systolic BP and age. PWV, as a measure of arterial stiffness, is affected by BP variation, independent of intrinsic stiffness. Changes in heart rate greatly affect variation in pulse waveform amplification from aortic segments to peripheral arteries, increasing the brachial systolic BP [26]. Heart rate in the multivariate model may mask variation in PWVao due to diabetes and hypertension status. When heart rate was removed from the model (data not shown), diabetes (β = 0.122, p = 0.027) and hypertension status (β = 0.207, p < 0.001) became independently associated with PWVao. CAVI, unlike PWVao, was not related to heart rate nor to BP in multivariate regression, even when systolic BP was substituted with diastolic, mean or pulse pressures (data not shown), as found elsewhere [27]. The capture of standard major CVD risk factors by CAVI is similar to other reports [20, 24, 28, 29]. Contrary to other studies [28‐30], none of the biochemical parameters predicted variation in CAVI values here. This might be the result of a robust model; all the independent variables were simultaneously forced into the regression. The model attempts to mimic the complex interaction among various cardiovascular risk factors in vivo.

CAVI and haPWV had independent positive association with age but negative with BMI in all participants. The effect of age on CAVI was stronger in healthy participants, similar to Japanese studies, where CAVI increased at a rate of 0.5units/10 years in both men and women [31]. We speculate that the U-shape relationship between most anthropometric measurements and vascular health [32] may explain the negative association of anthropometric indices with CAVI and haPWV in our findings. The relationship between CAVI and body composition defined as true lean body or fat mass have been reported in a population-based study. Nagayama et al. [28] reported that CAVI was associated with baseline anthropometric indices and decreased after reduction in body weight by caloric restriction. The association between obesity and arterial stiffening might be attributed to insulin resistance, reduction in nitric oxide synthesis and pro-inflammatory state associated with obesity [33, 34].

Blood vessel damage in T2D is also characterised by resistance vessel hypertrophy and endothelial leakage (as in ‘micro’-albuminuria) [35]. Hypertension without diabetes is characterised by ‘eutrophic’ remodelling of resistance vessels without hypertrophy, at least, until decompensation occurs, leading to increased endothelial dysfunction and peripheral vascular resistance [36]. These pathophysiological trajectories are interlinked in several ways; T2D and hypertension induce changes in arterial structure through similar yet independent pathways [37]. The mechanism by which T2D per se contributes to increased arterial stiffness and hypertension is not yet certain [36]. However, decreased nitric oxide activity, activation of the renin-angiotensin system, mitogen activated kinase pathways, advanced glycated end-product generation and increased oxidative stress all seem to contribute in T2D [17, 38]. CAVI has recently been shown to correlate with oxidized lipoprotein(a) in older [30], but not younger [24], patients, and is associated with changes in oxidative stress status in T2D patients with abnormal levels of LDL [39]. However, the physiological mechanisms underlying elevated arterial stiffness in T2D and hypertensive patients in these Ghanaian patients will need biopsy and post-mortem studies to clarify tissue pathology.

This study has shown indices of arterial stiffness, particularly, CAVI and haPWV, are useful tools for screening and discrimination of cardiovascular risk status in Ghanaian patients with T2D and hypertension. These indices may offer more precise targets for treatment than (higher) BP alone, particularly in low resource sub-Saharan African setting.