Association between Cardio-Ankle Vascular Index and Masked Uncontrolled Hypertension in Hypertensive Patients: A Cross-Sectional Study

Detection of masked uncontrolled hypertension (MUCH) that was defined for treated hypertensive individuals who had normal office blood pressure (BP) but elevated ambulatory BP remains largely challenging. Arterial stiffness is one of the leading risk markers for hypertension and can be clinically assessed by the cardio-ankle vascular index (CAVI). This study aimed to evaluate the association between CAVI and MUCH. A total of 155 hypertensive patients were included with their office BP levels and ambulatory BP monitoring measurements, which were divided into controlled hypertension (CH), MUCH, and sustained uncontrolled hypertension (SUCH) groups, respectively. There were 48 patients with CH, 56 patients with MUCH, and 51 patients with SUCH. Both MUCH and SUCH groups had a significantly higher CAVI than the CH group (9.05 (8.20–9.91) vs. 8.33 (7.75–9.15), p = 0.017, and 9.75 (8.35–10.50) vs. 8.33 (7.75–9.15), p = 0.002, respectively). There was no significant difference in CAVI values between the MUCH and SUCH groups. Multinomial logistic regression analysis exhibited that compared with the CH group, increased CAVI levels were positively associated with the presence of MUCH and SUCH (OR 2.046, 95% CI (1.239–3.381), p = 0.005; OR 2.215, 95% CI (1.310–3.747), p = 0.003) after adjusting for confounders. However, there was a similar trend of the CAVI in the MUCH and SUCH groups (OR 0.924, 95% CI (0.629–1.356), p = 0.686). In summary, our findings support, for the first time, the novel notion that CAVI as an arterial stiffness parameter is an independent risk factor for MUCH, being equally important to MUCH and SUCH. When the assessed CAVI is high in hypertensive patients with normotensive office BP levels, it is necessary to further investigate with a 24 h ambulatory BP monitoring to estimate the longstanding BP control. CAVI may be used as a noninvasive indicator to identify patients with MUCH earlier.


Introduction
High blood pressure (BP) is a modifable but poorly controlled risk factor for cardiovascular morbidity and mortality in China and worldwide. Uncontrolled BP is associated with a higher incidence of hypertension-mediated target organ damage [1]. Masked hypertension (MH), also called masked uncontrolled hypertension (MUCH) in treated patients, can be found in approximately 15% of patients with normal ofce BP [2]. Notably, the risks of cardiovascular events are substantially greater in patients with MH compared with the normotensive population and are actually close to or greater than those with sustained hypertension [3][4][5][6].
Patients with MH have a normal ofce BP albeit an elevated out-of-ofce BP while monitored using 24 h ambulatory BP monitoring (ABPM) or home BP monitoring (HBPM). ABPM measurement is more closely associated with hypertension-mediated target organ damage and the risk of cardiovascular events, is more reproducible than ofce BP measurement, and can help identify MUCH in treated hypertensive subjects [7]. However, ABPM measurement is limited often due to time restraints and intolerances when wearing [8]. Most hypertension screening and BP control assessments use ofce BP measurement, which is usually normal in patients with MH or MUCH. Te number of patients with MUCH has been underestimated due to the lack of ABPM used to assess BP control in seemingly welltreated patients. Greater recognition and identifcation of MUCH can reduce the incidence of hypertension-related complications and improve their quality of life.
Increased arterial stifness is an independent predictor of cardiovascular complications [9,10]. Cardio-ankle vascular index (CAVI) has been proposed as an index of arterial stifness based on the stifness parameter β [11]. CAVI is easy to measure and highly reproducible. It is infuenced by chronic exposure of the arterial wall to increased BP but is essentially independent of BP at the time of measurement. Some studies have reported that CAVI is high in hypertensive patients, especially in those with uncontrolled hypertension [12][13][14]. However, the relationship between MUCH and CAVI is largely unclear. Te aim of the present study is to explore whether a high CAVI value is associated with the presence of MUCH and to evaluate the independent efect of CAVI on MUCH in hypertensive patients.

Data Collection.
All participants received a standardized examination. Medical information regarding age, gender, tobacco smoking, drinking consumption, duration of hypertension, medications, history of diabetes mellitus, coronary heart disease, dyslipidemia, and atrial fbrillation were extracted from the hospital's electronic medical records database. Two authors reviewed the medical data separately.

Ofce BP and 24 h ABPM Measurements.
Ofce BP measurement was taken by trained nurses. Patients sat quietly for 10 min before measurements, and the appropriate cuf size was used with the bladder encircling at least 80% of the same arm (Omron HEM-7051, Tokyo, Japan). Te patient's arm was placed on the desk at the heart level. Tree readings with a 1-minute interval between measurements were performed, and the last two readings were averaged as ofce BP [7].
24 h ABPM measurement, belonging to out-of-ofce BP measurements, was performed based on the 2020 International Society of Hypertension Global Hypertension Practice Guidelines [7]. An ambulatory BP monitor (Vasomedical BIOX Ambulatory Blood Pressure Monitor) was installed over the passive arm of each patient to automatically measure and record BP. In the daytime, it was programmed to perform measurements every 20 minutes between 08 : 00 and 22 : 00. In the nighttime, it was programmed to perform measurements every 60 minutes after 22 : 00. If a record contained 70% of the programmed readings, the coverage time was greater than 20 hours, and there were at least 20 readings during the day and at least 7 readings during the night, the record was considered efective. Average BP values were usually provided for daytime, nighttime, and 24 h.

CAVI Measurement. CAVI was measured using a
VaSera VS-1000 vascular screening system (Fukuda Denshi Co, Tokyo, Japan). Te patient rested on a bed in a supine position for 10 minutes before the measurements. Te CAVI examination was carried out strictly in line with the operational procedure: in a quiet environment, both ankles and brachium were secured with cufs. Electrodes for electrocardiography were placed on both wrists, and a microphone was placed on the sternum for phonocardiography.
Te CAVI value was based on the stifness parameter calculated using the following formula: CAVI � a{(2ρ/ΔP) × ln (Ps/Pd) × PWV 2 } + b, where ρ is blood density, Ps refers to the systolic blood pressure, Pd refers to the diastolic blood pressure, ΔP � Ps − Pd, PWV is pulse wave velocity between the aortic and ankle values, and a and b are constants. Te CAVI value was calculated as the average of the right and left CAVI measurements. Te normal value of CAVI (low CAVI) is <9, and a CAVI ≥9 is defned as the abnormal value (high CAVI) [16].

Statistical Analysis.
Te type of distribution of quantitative variables was analyzed using the Shapiro-Wilks or Kolmogorov-Smirnov tests. Continuous variables with normal distribution or with non-normal distribution were expressed as the mean ± SD or median with interquartile ranges (25th and 75th percentiles), respectively. Among the three groups, comparisons of continuous variables with normal distribution were carried out by ANOVA. Te Kruskal-Wallis test was performed to compare groups with non-normal distribution. Categorical variables were expressed as the number and percentage of cases, and between-group diferences were analyzed by the chi-square test or Fisher's exact test as appropriate. Te Bonferroni correction was used for all pairwise comparisons. Te Spearman correlation was used to identify the relations between the CAVI and BP parameters. Multinomial logistic regression analysis was performed to determine whether CAVI was independently associated with MUCH. Tree models were constructed step-by-step. A p-value of less than 0.05 was considered statistically signifcant. Statistical analysis was conducted in SPSS 27.0 (SPSS Inc., Chicago, USA). Graphs were plotted using the GraphPad Prism software version 9.0 (California, USA).

Clinical Characteristics and Biochemical Measurements of Diferent BP Phenotypes.
After excluding 42 participants following our exclusion criteria, a total of 155 participants comprising 48 controlled hypertensive cases, 56 patients with MUCH, and 51 patients with SUCH were included in the subsequent analysis. Te average age of patients was 64 years, and 110 (71%) of the patients were male. Table 1reveals that there were no signifcant diferences in terms of age, BMI, smoking history, drinking history, course of hypertension, comorbidities (including diabetes mellitus, coronary heart disease, dyslipidemia, and atrial fbrillation), the use of β blocker, antiplatelet agents, and statins. In addition, the values of FBG, HbA1c, creatinine, eGFR, UA, LP (a), TC, and LDL-C were comparable among the groups. But patients with MUCH were more likely to be male than the patients with SUCH. At the same time, there was a higher level of TG and a lower level of HDL-C in the MUCH group than in the CH group. Patients with MUCH and SUCH had a higher rate of the use of angiotensinconverting enzyme inhibitor (ACEI) or angiotensin II receptor blocker (ARB), while most of the patients in the SUCH group used calcium channel blocker agents (CCB). Both TG and hs-CRP values were higher in the SUCH group than those in the CH group.

Correlation between CAVI and Each BP Parameter in
Patients with CH, MUCH, and SUCH. Spearman correlation analysis was used to examine the correlations between the CAVI and BP parameters including ofce BP levels and ABPM measurement (Figures 1(a)-1(d)). Te level of CAVI was positively related to 24 h SBP, daytime SBP, and nighttime SBP (all p < 0.001). However, there were no signifcant correlations between ofce BP, 24 h DBP, daytime DBP, nighttime DBP, and the CAVI values.

Comparation of the CAVI for Patients with CH, MUCH, and SUCH.
Both MUCH and SUCH groups had signifcantly higher CAVI values than that in the CH group (9.05 (8.20-9.91) vs. 8.33 (7.75-9.15), p � 0.017, and 9.75 (8.35-10.50) vs. 8.33 (7.75-9.15), p � 0.002, respectively). However, there was no signifcant diference in CAVI values between the MUCH and SUCH groups (Figure 2(a)). Of the 48 patients with CH, 15 (31.3%) had a high CAVI. Additionally, 31 cases (55.4%) had a high CAVI in the MUCH group of 56 subjects and 30 cases (58.8%) in the SUCH group, exhibiting that the trend to the high CAVI in the MUCH group was signifcantly higher than that in the CH group (χ 2 � 6.089, p � 0.014). However, there was no signifcant diference in comparison with the SUCH group (χ 2 � 0.131, p � 0.718), as is illustrated in Figure 2(b).

Te Association between CAVI and Diferent BP
Phenotypes. Figure 3 demonstrates the results of the multinominal logistic regression analysis of the CAVI value for the diferent BP phenotypes. Univariate logistic regression analysis showed that the increase of CAVI was signifcantly Journal of Healthcare Engineering and positively associated with the MUCH(OR 1.501, 95% CI (1.106-2.036), p � 0.009, Figure 3(a)). After adjustment for age, sex, duration of hypertension, smoking history, the use of ACEI/ARB and CCB, higher level of CAVI remained signifcantly associated with the presence of MUCH (OR 2.046, 95% CI (1.239-3.381), p � 0.005, Figure 3(a)). Compare with CH, increased CAVI levels were positively associated with the presence of SUCH after adjusting for confounders (OR 2.215, 95% CI (1.310-3.747), p � 0.003, Figure 3(b)). However, there was a similar trend of the CAVI in the MUCH and SUCH groups (OR 0.924, 95% CI (0.629-1.356), p � 0.686, Figure 3(c)).

Discussion
Our fndings suggested that the CAVI value is positively related to ambulatory SBP profles but not associated with ofce BPs and ambulatory DBPs. In patients with MUCH, there is a higher CAVI value compared with patients with CH. However, there is no signifcant diference in CAVI values between the MUCH and SUCH groups. After adjusting for confounders, increased CAVI levels are positively and independently associated with the presence of MUCH, although there is a similar trend of the CAVI in the MUCH and SUCH groups.   MUCH in hypertensive patients is defned as normal ofce BP but elevated out-of-ofce BP (including HBPM or ABPM), among which the risk of cardiovascular events is close to or even greater than that of SUCH. Based on the use of ofce BP to monitor BP control, physicians will substantially overestimate the number of patients who have well-controlled BPs, leaving many higher-risk patients at an excess risk [17]. It is of great importance for MUCH to be accurately detected so that medication can be adjusted accordingly.
Te clinical characteristics of patients with MH and MUCH are poorly defned [18]. Available 24 h ABPM-based studies have identifed that high-normal ofce BP, age, male, smoking, obesity, diabetes, proteinuria, and high cardiovascular disease risks were associated with MH [19][20][21][22]. In our present study, we identifed the clinical profles of MUCH patients as more likely to be male, low HDL-C, and high TG. Notably, we observed that the ofce BP in the CH and MUCH groups were similar. However, those with MUCH had signifcantly higher levels of 24 h SBP, 24 h DBP, daytime SBP, daytime DBP, nighttime SBP, and nighttime DBP than patients with CH, which is consistent with the recently published research [21]. Te fndings indicated that MUCH has a negative impact on blood vessels and results in subsequent damage to multiple organs.
Arterial stifness is becoming increasingly important in clinical applications as an early marker of hypertensionmediated target organ damage [2]. Some studies have reported that CAVI is high in hypertensive patients, especially in those with uncontrolled hypertension [12]. However, the magnitude of the relationship between arterial stifness and MUCH is poorly reported. Previous studies found that the BP level was associated with arterial stifness, and showed prognostic information, based only on ofce BP measurements [14], or arterial stifness was mostly measured by carotid-femoral PWV [23,24]. One study reported that CAVI was almost linear in relation to SBP, DBP, and mean BP [14] based on the ofce BP. In our study, we found the CAVI value was positively related to ambulatory SBPs (including 24 h SBP, daytime SBP, and nighttime SBP) but not associated with ofce BPs and ambulatory DBPs (including 24 h DBP, daytime DBP, nighttime DBP) in hypertensive patients. In a recent meta-analysis, the association between SBP and the progression of arterial stifness assessed by PWV was found, while the correlation between mean arterial pressure or DBP and PWV was weak because of limited data [25].
CAVI is regarded as an arterial stifness parameter independent of the time of its measurement, with advantages including being noninvasive, easier to obtain, and representing a global index of both central and peripheral stifness. It thus appears to be a suitable candidate for the routine evaluation of vascular organ damage and the estimation of cardiovascular risk, especially in hypertensive subjects [26]. 24 h ABPM is an accurate assessment of BP levels, which provides plenty of information on the systolic and diastolic day and night BP profles and ambulatory average BP levels [27], and is of great use in the identifcation of MH and MUCH. Due to the limited availability of ABPM, it is necessary to look for simple and noninvasive approaches to identify MUCH early in clinical practice. In this study, we found that the CAVI value was signifcantly higher in the MUCH group than that in the CH group, but it was similar between the MUCH and the SUCH groups. Te trend of the high CAVI in the MUCH group was also higher than that in the CH group. Furthermore, adjusting for covariables of age, sex, duration of hypertension, smoking history, and the use of ACEI/ARB and CCB, the increase of CAVI is still positively associated with the MUCH compared with the CH. Until now, it is unclear whether arterial stifness is a cause or a consequence of hypertension or whether arterial stifness and hypertension are mutually reinforcing [28]. Findings from our current study support the notion that CAVI, as an arterial stifness parameter, is an independent risk factor for MUCH and it is equally important to MUCH and SUCH. We recommend that even in treated hypertensive patients with normotensive ofce BPs when the assessed CAVI is high, patients should have a further 24 h ABPM to estimate the longstanding BP control.

Limitations
Te current study has several limitations that must be taken into account. Firstly, the recruitment of the subjects was based on hospitalized patients, which may have unpredictable selection bias. Secondly, the participants might have some activities during ABPM, which to some extent may afect the accuracy of the measurement results. Tirdly, cross-sectional studies could not evaluate the causal relationship directly. Finally, the sample size of this work was relatively small.

Conclusions
Our fndings support, for the frst time, the novel notion that CAVI as an arterial stifness parameter is an independent risk factor for the presence of MUCH, and that it is equally important to MUCH and SUCH. In future clinical work, when the assessed CAVI is high and above the normal range in hypertensive patients with normotensive ofce BP, it is necessary to further investigate with a 24 h ABPM to estimate the BP control. In hypertensive patients with normotensive ofce BPs, CAVI might be a potential noninvasive measurement to help identify MUCH earlier. In the future, larger prospective cohort studies are needed to further confrm the causal relationship between CAVI and MUCH among the hypertensive population. Te detection of MUCH using deep learning can be the new direction [29].

Data Availability
Te data that support the fndings of this study are available from the corresponding author upon reasonable request. Te data are not publicly available due to the information that could compromise the privacy of research participants.