Effect of white-coat hypertension on arterial stiffness

Abstract Background: White-coat hypertension (WCH) is a debatable risk factor of cardio-cerebrovascular diseases and the current study results on the association between WCH and arterial stiffness are inconsistent. The aim was to investigate the effect of WCH on arterial stiffness using meta-analysis. Methods: Based on prespecified search strategies and inclusion criteria, Medline, Embase, Web Of Science, Cochrane Library, and BioSciences Information Service Preview databases were reviewed. A total of 20 studies involving 1538 WCH patients and 3582 normotensives (NT) were included. Literatures were screened for data extraction and quality assessment. Overall analysis and subgroup analysis were conducted in RevMan version 5.3 and Stata version 14.0 software. Results: Overall analysis showed that carotid-femoral pulse wave velocity (cf-PWV) was significantly higher in WCH group than in the NT group (P < .00001, 95% CI: 0.79–3.26). Subgroup analysis showed that in adults, cf-PWV was significantly higher in the WCH patients than in the NT subjects (P<.001, 95% CI: 0.46–0.87), while in juveniles, cf-PWV was comparable between the WCH group and the NT group (P = .25, 95% CI: −0.39 to 0.61). Conclusion: This meta-analysis showed that WCH may increase arterial stiffness in adult population.


Introduction
White-coat hypertension (WCH), also termed isolated clinic hypertension, is seen in the patients who show hypertension during the clinic visits. [1] Currently, the diagnostic criteria of hypertension has been updated and the diagnostic criteria of WCH vary by guidelines. [2,3] The widely used traditional criteria defines WCH as: Clinic systolic blood pressure ≥140 mm Hg and/or diastolic pressure ≥90 mm Hg, and mean ambulatory blood pressure <135/85 mm Hg daytime or home blood pressure <135/85 mm Hg. WCH was once considered a benign phenomenon, but several studies have established its relationship with multiple metabolic disorders such as impaired glucose tolerance, insulin resistance, and metabolic syndrome. [4,5] Ongoing studies have been directed to clarify the role of WCH in cardio-cerebrovascular impairments. [6] Arterial stiffness examination is a noninvasive tool to evaluate cardio-cerebrovascular risks. Many clinical studies and basic researches have revealed arterial stiffness as a risk factor of cardiocerebrovascular diseases. With the popularization of arterial stiffness examination, some indicators such as pulse wave velocity (PWV), ambulatory arterial stiffness index (AASI), and augmentation index have been developed. Of note, both American Heart Association scientific statement and European expert consensus have recommended PWV as the golden standard for arterial stiffness with consideration to its high accuracy and applicability. [7] To identify the target organs of WCH in cardio-cerebrovascular impairments, several clinical studies have attempted to investigate the relationship between WCH and arterial stiffness. However, their results vary due to confounding factors such as small sample size, racial difference, inconsistent methods, and discrepant inclusion criteria. [8] In light of the inconsistencies of relationship between WCH and arterial stiffness, this systematic review and meta-analysis were conducted to evaluate the relationship between WCH and arterial stiffness. of the following terms: "clinic hypertension, "office hypertension," "white-coat," "PWV," "pulse wave velocity," "arterial stiffness," "aortic stiffness," and "vascular stiffness." Publication date was limited to December 23, 2017.
Literatures of the same study population, poor research quality, and incomplete data reporting were excluded. If a paper included several independent case-control groups, they were screened and the eligible ones were included in the meta-analysis. Figure 1 shows the flowchart of study design.

Data extraction and quality assessment
Two investigators (PC and YP) independently searched literature, screened studies, and extracted data on the basis of searched strategies, and inclusion criteria. The quality of studies was assessed by population selection, comparability between cases and controls, and exposure measurement in accordance with the Newcastle-Ottawa Quality Assessment Scale (NOS). The NOS contains 8 items with a maximum score of 9 points. All studies were classified as low quality (0-3 points), medium quality (4-6 points), or high quality (7-9 points) based on NOS. [9]

Statistical analysis
The cf-PWV was compared between WCH group and NT group. All statistical analyses were conducted in RevMan software version 5.3 (The Cochrane Collaboration, Copenhagen, Denmark) and Stata version 14.0 (Stata Corp LP, College Station, TX). All the data were calculated for their 95% confidence intervals (95% CI). Statistical difference was defined as a 2-sided P value equal to or smaller than .05.
All the data were transformed into mean ± standard deviation format by either RevMan version 5.3 software or manual calculation. Publication bias analysis, sensitivity analysis, heterogeneity analysis, data synthesis, Z test, meta-regression analysis, and subgroup analysis were performed. Publication bias was analyzed with Begg and Egger tests and visually examined by funnel plot. Sensitivity analysis was performed with Cohen test and graphical methods. Twelve was used to quantitatively assess heterogeneity. When significant heterogeneity was indicated by I2>50%, the random-effects model was used to calculate effect size; otherwise, fixed-effects model was used, followed by Z test. Subgroup analysis was performed for age, blood pressure,

Subgroup analysis.
The studies were stratified by the history of antihypertensive drug use, age, instrument for  [28] Greece 387 (52.7) 54.3 ± 0. inspecting PWV and study quality, and Table 2 shows all subgroup analysis results. For adults, PWV was significantly higher in WCH group than in NT group (P<.001, 95% CI: 0.46-0.87), but PWV was not different between WCH group and NT group in juveniles (P = .253, 95% CI: À0.39 to 0.61). In the subgroup analysis of 9 studies excluding patients with diabetes mellitus or cardiovascular diseases, heterogeneity was significantly reduced (I2 = 45%; Fig. 5), and PWV differed between

Discussion
Meta-analysis evaluated the relationship between WCH and arterial stiffness. It was found that adult WCH patients had significantly higher cf-PWV than normal population, indicating higher risks of cardio-cerebrovascular diseases in these patients. However, juveniles did not show the phenomenon, probably attributable to short duration of WCH and a low degree of arterial stiffness. Moreover, only 2 studies containing 164 juveniles were analysed, which might not have fully represented the real situations of juveniles. More studies are needed to reveal the relationship between WCH and arterial stiffness in the juvenile population. During literature screening, some studies were identified which used AASI and augmentation index to quantify arterial stiffness. [29,30] These accessory examinations have been accepted by clinical practitioners. In particular, AASI calculated from ambulatory blood pressure monitoring is easy to use. Nevertheless, PWV, as the golden standard of arterial stiffness, has a markedly higher diagnostic accuracy than other indexes. Metaanalysis included clinical studies which had employed PWV as an examination method to best show the relationship between WCH and arterial stiffness. Recently, Upala et al [31] published another meta-analysis about the relationship between WCH and arterial stiffness, but they reported no significant association between WCH and arterial hypertension on the basis of 4 eligible observational studies containing persistent hypertension group, WCH group and normal control group. In our opinion, due to their inclusion methods, they might have excluded many casecontrol studies which only contained WCH group and normal control group thus the study did not sufficiently reveal the relationship between WCH and arterial stiffness. Based on prespecified search strategies and inclusion criteria, Medline, Embase, Web Of Science, Cochrane Library, and BIOSIS Preview databases were reviewed. A total of 20 studies involving 1538 WCH patients and 3582 normotensives were included in our study, which would better reflect the effect of WCH on arterial stiffness.  A limitation of this meta-analysis is that we had no individual patient data, only the literature data can be combined and analyzed. A further limitation was that the heterogeneity for overall analysis was noticeable (I2 = 82%), so the meta-regression analysis and subgroup analysis were performed. Subgroup analysis is the highlight of meta-analysis, especially that of the patients without diabetes mellitus or cardiovascular diseases. Maine-Syracuse case-control study has demonstrated the significant relationship between type-2 diabetes mellitus (especially uncontrolled type-2 diabetes mellitus) and arterial stiffness. Previous studies have proven the close relationship between cardiovascular diseases (e.g., coronary artery disease) and arterial stiffness. [32,33] Therefore, subgroup analysis for the patients without diabetes mellitus or cardiovascular diseases was conducted. The results showed significantly reduced heterogeneity in the eligible studies, which was further reduced by the secondary subgroup analysis stratified by history of antihypertensive drug use. In this way, subgroup analyses identified the relationship between WCH and arterial stiffness. By stepwise subgroup analyses, the eligible criteria was gradually narrowed to reduce the heterogeneity and to enhance the reliability of study results. Meta-regression analysis also identified diabetes mellitus and cardiovascular diseases as important sources of overall heterogeneity.
This study showed that WCH may cause arterial stiffness in adult population. This kind of mechanisms may help uncover the multiple target organ damages in the future. WCH is common in clinical practice, but its pathophysiological mechanisms and target organ damages remain unclear. As a result, many clinicians are confused about its diagnosis and treatments. Based on these study findings, more attention is to be given to the role of WCH in cardio-cerebrovascular target organ damages, and reasonable diagnostic and therapeutic standards of WCH should be further explored.

Author contributions
Formal analysis: Yan Wang.