Prognostic Significance of Abnormal Ankle–Brachial Index Among Long-term Hemodialysis Patients in Kinshasa, the Democratic Republic of the Congo

Objective Early identification of atherosclerosis using a non-invasive tool like ankle–brachial index (ABI) could help reduce the risk for cardiovascular disease among long-term hemodialysis patients. The study objective was to assess the frequency and impact of abnormal ABI as a marker of subclinical peripheral artery disease (PAD) in chronic hemodialysis patients. Methods This was a historic cohort study of kidney failure patients on long-term hemodialysis for at least 6 months. The ABI, measured with two oscillometric blood pressure devices simultaneously, was used to assess subclinical atherosclerosis of low limb extremities. Abnormal ABI was defined as ABI <0.9 or >1.3 (PAD present). Survival was defined as time to death. Independent factors associated with abnormal ABI were assessed using multiple logistic regression analysis. Kaplan–Meier method (log-rank test) was used to compare cumulative survival between the two groups; a P value <0.05 was statistically significant. Results Abnormal ABI was noted in 50.6% (n=43) of the 85 kidney failure patients included in the study; 42.4% (n=36) had a low ABI, and 8.2% (n=7) had a high ABI. Factors associated with PAD present were cholesterol (adjusted odds ratio [AOR], 1.02; 95% confidence interval [CI], 1.01–1.04; P=0.019), inflammation (AOR, 9.44; 95% CI, 2.30–18.77; P=0.002), phosphocalcic product (AOR, 6.25; 95% CI, 1.19–12.87; P=0.031), and cardiac arrhythmias (AOR, 3.78; 95% CI, 1.55–7.81, P=0.009). Cumulative survival was worse among patients with PAD present (log-rank; P=0.032). Conclusion The presence of PAD was a common finding in the present study, and associated with both traditional and emerging cardiovascular risk factors as well as a worse survival rate than patients without PAD.


INTRODUCTION
Peripheral artery disease (PAD) of the lower extremities, an important manifestation of systemic atherosclerosis, 1 is commonly seen in patients with kidney failure undergoing long-term hemodialysis (LTHD). In fact, the prevalence of PAD in LTHD patients has been reported to be high, ranging from 17% to 48%, and associated with increased cardiovascular morbidity and mortality. 2,3 Peripheral artery disease shares similar risk factors with coronary artery disease and cerebrovascular disease. 3 Therefore, its early diagnosis and management can help improve the prognosis of LTHD patients 4 by avoiding or at least delaying adverse events, such as amputations, cardiovascular events, and death. 5 In this regard, the ankle-brachial index (ABI) and pulse wave velocity are common non-invasive tools used to assess arterial health quantitatively with regard to blocked arteries and arterial stiffness, respectively. 6 A low ABI has been reported to predict the future risk of cardiovascular disease and influence outcomes among LTHD patients. 7 In the Democratic Republic of the Congo, very few studies have been carried out on the prevalence and prognostic significance of cerebral and cardiac diseases among LTHD patients, 8,9 and data on the prevalence and prognostic significance of PAD in LTHD patients are not yet available. Therefore, the present study aimed to assess the burden and the prognostic significance of PAD among LTHD patients in Kinshasa.

Study Population and Design
We conducted a historic cohort study that included all patients who attended four hemodialysis centers (University Hospital of Kinshasa, Medical Center of Kinshasa, Afia Medical Care, Ngaliema Medical Center) in Kinshasa, from March to December 2016. The patients had undergone hemodialysis two or three times a week with high-flux dialyzers, at a blood flow rate of 250-300 mL/min and dialysate flow rate of 500 mL/min, during each four-hour dialysis session. Patients aged 18-75 years who underwent long-term hemodialysis (LTHD) for at least 6 months were recruited. Exclusion criteria were atrial fibrillation, bilateral below the knee amputations, and recent hospitalization (less than 4 weeks prior to study enrollment). The study protocol was approved by the Ethics Committee of Kinshasa School of Public Health/University of Kinshasa (ESP/CE/013/2016), and written informed consent was obtained from all patients. All clinical investigations were conducted according to the principles expressed in the Declaration of Helsinki.

Data Collection and Procedure
Information on demographic and medical data including sex, age, smoking history (ever versus never), kidney failure complications (encephalopathy, pericarditis, hypervolemia, acidosis, anemia, hypocalcemia), comorbidities (stroke and transient ischemic attack history, arrhythmias, heart failure, coronary artery disease, viral hepatitis C, human immunodeficiency virus, diabetes, hypertension), and dialysis parameters (number of sessions, vascular access, urea clearance, interdialytic weight gain) were obtained from interviews and the patients' medical records. Body mass index was calculated as weight divided by height squared in kg/m 2 . Hypertension was defined as blood pressure (BP) ≥140/90 mmHg or taking antihypertensive drugs, and diabetes was defined based on a fasting blood glucose level of ≥126 mg/dL or taking antidiabetic drugs. Patients with a history of cerebrovascular accidents, including cerebral bleeding and infarction, were defined as having a cerebrovascular disease, and those with a history of angina or myocardial infarction, or ischemic changes in electrocardiography, were defined as having coronary artery disease. Laboratory parameters (blood urea nitrogen, serum creatinine, serum potassium, uric acid, bicarbonate, calcium, phosphorus, vitamin D, parathormone, hemoglobin and hematocrit, C-reactive protein, albumin and total protein, alkaline reserve, troponin, ProBNP) obtained one month or less following enrollment were retrieved from patients' medical records.

Evaluation of Cardiac Structure and Function
Echocardiographic examination was performed by an experienced cardiologist using a VIVID 7 system (General Electric Medical Systems, Milwaukee, WI, USA), with the patient breathing quietly while lying in the left lateral decubitus position. The cardiologist was blinded to other data. Two-dimensional Mmode images were recorded from the standardized views. Echocardiographic measurements included left ventricular internal diameter at diastole, the left ventricular posterior wall thickness at diastole, interventricular septal wall thickness at diastole, Ewave deceleration, and peak early and late diastolic transmittal filling velocity.

Measurement of ABI
Since ABI may be influenced by hemodialysis, 10 all ABI values were obtained by a trained and experienced physician 10-30 minutes before hemodialysis and after 5 minutes' rest in the supine position. The ABI values were measured once in each patient using the Bidop ES-100 V3 arterial doppler device (Hadeco, Kawasaki, Japan), which automatically and simultaneously acquires oscillometric BP measurements in both arms and ankles. 11 Occlusion and monitoring cuffs were placed tightly around the upper arms and both sides of the lower extremities in the supine position. Measurements were obtained from the posterior tibial arteries in the lower extremities, since the pedis dorsal artery is congenitally absent in 4% to 12% of the population. 12 Systolic BP (SBP) was measured twice at each site, in rapid succession and alternating, to obtain an average value. 10 Using the SBP ankle value the ABI was calculated as the ratio of ankle SBP/arm SBP. The ABI values were defined as follows: PAD present, abnormal ABI (<0.9 or >1.3) and PAD absent, normal ABI (0.9-1.3).

Statistical Analysis
Statistical analysis was performed using SPSS, version 21. Continuous variables were expressed as mean ± standard deviation (SD) (normal distribution) or median and range (skewed distribution). Categorical variables were expressed as absolute (n) and relative (in percent) frequencies. Comparison of the means of two groups or more was done using Student's t test or one-way analysis of variance (ANOVA) with Scheffé's multiple test, respectively. Logistic regression analysis was used to assess independent factors associated with PAD. The endpoint was survival (time-to-death): the data of surviving patients at the end of the study (December 2019), patients lost to follow-up, or patients shifted to transplantation or peritoneal dialysis were also analyzed in the study. Kaplan-Meier analysis was used to describe survival, and the comparison of different survival rates was done using the log-rank test; statistical significance was defined as P<0.05.
The echocardiographic and ABI parameters of the study population are summarized in Table 4.

Frequency and Clinical Profile of Abnormal ABI
Abnormal ABI was observed in 43 (50.6%) patients, of which 36 (42.4%) had a low ABI and 7 (8.2%) had a high ABI (Table 1). Compared to patients with a normal ABI, patients with an abnormal ABI were on average older (56.4±14.4 versus 49.1±16.6 years; P=0.034), and a significantly higher proportion were 60 years old or more (51.2% versus 31%; P=0.047). These patients also had, on average, significantly decreased DBP (80.4 mmHg versus 88.7 mmHg; P=0.013) and a significantly higher proportion of inflammation (79.1% versus 52.4%; P=0.009). Differences observed in other parameters of interest did not reach the level of statistical significance (Table  1). With reference to kidney failure and hemodialysis parameters, patients with an abnormal ABI tended to have a higher proportion with temporary catheter access at initiation of hemodialysis (93.0% versus 81.0%; P=0.056); however, the difference observed was not statistically significant. The two subgroups were similar for the other variables of interest (Table 2).

DISCUSSION
The main findings of the present study are as follows. First, abnormal ABI (PAD present), mostly with a low ABI, was observed in half of the chronic hemodialysis patients. Second, PAD was associated Data are expressed as mean±standard deviation, absolute (n) and relative (in percent) frequencies.
with advanced age. Third, factors independently and significantly associated with PAD were P×Ca, arrhythmia, inflammation, and cholesterol. Fourth, cumulative survival was worse in patients with an abnormal ABI compared to those with a normal ABI.
Abnormal ABI was found in half of the chronic hemodialysis patients in this study, with a higher frequency than those reported by Tian (28%), 13 Ašćerić (35%), 14 and Ozgur (44%). 15 Differences in sample size, study population characteristics, and criteria used to define abnormal ABI could explain the difference between their studies of abnormal ABI frequencies. Of note, due to financial constraints, most patients in our study had less than three dialysis sessions per week, resulting in accumulation of some uremic toxins, such as asymmetric dimethylarginine (ADMA), a well-known nitric oxide synthase inhibitor responsible for endothelial dysfunction and subsequent accelerated atherosclerosis. 16 This study found that older age, especially >60 years, was associated with abnormal ABI, which is in agreement with previous reports by Ašćerić 14 and Ozgur. 15 The aging process could lead to accelerated atherosclerosis 17  Data are expressed as mean±standard deviation, median (interquartile range), absolute (n) and relative (in percent) frequencies.
resistance, with a subsequent constellation of multiple cardiovascular risk factors, all of which have obesity in common as the main underlying factor. 18,19 Hence, advanced age is clearly a risk factor for atherosclerosis, particularly if obesity is present.
Phosphocalcic product, inflammation, cholesterol, and arrhythmia emerged as the main independent factors positively associated with abnormal ABI in multivariate analysis. Our finding is consistent with that of previous reports of an association of Data are expressed as mean±standard deviation, median (interquartile range), absolute (n) and relative (in percent) frequencies.
ABI    traditional and emerging risk factors with abnormal ABI as a marker of atherosclerosis in the general population as well as in chronic hemodialysis patients. 10,20,21 Increased calcium levels and phosphocalcic product, a marker of mediacalcosis, have already been reported in chronic hemodialysis patients with abnormal ABI or incompressible ankle by Miguel 10 and Van Jaarsveld. 22 Patients with abnormal ABI had increased levels of C-reactive protein as an inflammation biomarker in the present study. Our finding agrees with that of previous reports of an association of inflammation with accelerated atherosclerosis in chronic hemodialysis patients. 20,23 There is a mutually triggering vicious cycle between inflammation and free radical production leading to oxidative stress and subsequent endothelial dysfunction and atherosclerosis. Indeed, proinflammatory cytokines can lead to excessive endothelial production of free radicals, and the latter can increase via activation of the nuclear factor kappa beta, the transcription of proinflammatory genes. 23 Our finding of higher levels of total cholesterol and low-density lipoprotein cholesterol (LDLc) in chronic hemodialysis patients with abnormal ABI is consistent with that reported by Jabbari. 7 Increased cholesterol levels in chronic hemodialysis patients could be due to inflammation and uremic toxin-induced insulin resistance with subsequent lipid and glucose homeostasis disorders. 24,25 In addition, since chronic hemodialysis is frequently associated with malnutrition, the latter could induce increased hepatic production of lipids due to the fall in oncotic pressure; such cholesterol levels are seen in nephrotic syndrome. 26 In chronic hemodialysis, high levels of total cholesterol and LDL-c are associated with a high likelihood of carotid atheroma plaque formation. 27,28 In addition to multiple traditional and emerging cardiovascular risk factors, the association of abnormal ABI with arrhythmia in chronic hemodialysis patients in the present study could be explained by the presence of valvular calcifications as reported by Ureña-Torres. 29 Although vascular access failure has been reported to be frequently associated with abnormal ABI in chronic hemodialysis patients, 30 the lack of association observed in the present study, where most patients used either temporary or permanent catheters, could be due to the small study sample size.
In the present study, survival was lowest among patients with an abnormal ABI, consistent with the findings of Miguel 10 and Adragao 31 who reported higher mortality rates among chronic hemodialysis patients with abnormal ABI. 32 This could be explained by the fact that abnormal ABI is not only a marker of local endothelial dysfunction but also of extended endothelial dysfunction involving microcirculation of vital organs, such as heart, brain, kidneys, and lungs, which can lead to multiple organ failure. 3 Therefore, abnormal ABI could be used for the prediction of global cardiovascular risk in chronic hemodialysis patients.
The interpretation of the results of the present study should take into account some limitations. First, the retrospective design of the present study precludes the establishment of any temporal relationship between exposure and outcomes. Second, the sample size did not allow sufficient power for statistical tests to identify potential relationships between variables of interest.

CONCLUSION
Peripheral artery disease as assessed by ABI was a common finding in the present study and associated with both traditional and emerging cardiovascular risk factors as well as a low survival rate compared to patients without PAD. The validation of the present findings in a prospective study with a representative sample of chronic hemodialysis patients is planned.