Introduction

Hypertension is typically diagnosed by detecting an elevated blood pressure (BP) in the clinic. However, BP levels measured in the clinic may differ substantially when measured in the out-of-clinic setting by ambulatory BP monitoring (ABPM) or home BP monitoring (HBPM). The phenomenon of white coat hypertension describes those individuals with elevated clinic BP, but non-elevated out-of-clinic BP [1]. Data have demonstrated that white coat hypertension is common and likely not associated with an increased risk of cardiovascular disease (CVD) events when compared to individuals with sustained normotension, defined as non-elevated clinic and out-of-clinic BP [2,3,4, 5•]. Masked hypertension is the inverse phenomenon: an elevated out-of-clinic BP despite a non-elevated clinic BP. First coined by Pickering in 2002 [6], masked hypertension has gained increasing recognition in research and clinical practice and is now known to be a high-risk BP phenotype, associated with an increased risk of CVD events and target organ damage [5•, 6,7,8,9,10].

In this review, we will discuss the recent literature on masked hypertension including disease prevalence, controversies related to the use of ABPM versus HBPM to detect masked hypertension, potential approaches to screening for masked hypertension, adverse clinical outcomes associated with masked hypertension, recent perspectives on treating masked hypertension, and suggestions on future areas of research. The papers and topic areas are discussed with an emphasis on issues which may guide future research and the clinical management of masked hypertension.

Definitions

Originally, the term “masked hypertension” referred to individuals not taking antihypertensive medication who have a non-elevated clinic BP but have elevated out-of-clinic BP [6]. Several guidelines recommend using the daytime and/or the 24-h periods to define masked hypertension [10,11,12] though more recent recommendations [10] propose the use of the nighttime period as well. The term “masked hypertension” can also be applied to individuals taking antihypertensive medication. For these individuals, the term “masked uncontrolled hypertension” has been used. Herein, for readability, we use the term “masked hypertension” for both individuals not taking and taking antihypertensive medication. Sustained hypertension will refer to individuals with both elevated clinic and out-of-clinic BP; and sustained normotension will refer to individuals with both non-elevated clinic and out-of-clinic BP.

Prevalence

Systematic reviews of population-based studies have reported a prevalence of masked hypertension ranging from 15 to 30% among individuals with non-elevated clinic BP, with out-of-clinic BP measured on ABPM or HBPM [13, 14]. This wide range in part reflects the sampling of different populations as well as the use of different time periods (i.e., daytime, 24-h, nighttime periods) to define masked hypertension. A recent analysis by Melgarejo et al. demonstrated that the prevalence of masked hypertension also varies by geographic region [15•]. Examining ABPM data from 10 cohorts over three continents, the authors found that the prevalence of masked hypertension defined using the daytime, 24-h, and/or nighttime periods on ABPM ranged from 8.8% in Belgium (in the Belgian Population Study) to 30.5% in China (the JingNing Population Study) among individuals with non-elevated clinic BP. Significant variation by geographic region was also observed in an analysis of the ARTEMIS registry, an international network of clinics performing ABPM, where the prevalence of masked hypertension was observed to range from 9% in Europe to 16% in Asia and 17% in Africa with masked hypertension defined using 24-h ABPM [16]. A recent study from the USA highlights racial differences in the prevalence of masked hypertension. In an analysis by Wang et al., a community sample of employed adults from the Masked Hypertension Study was used to model daytime BP on ABPM in National Health and Nutrition Examination Survey (NHANES, 2005–2010). The authors estimated a prevalence of masked hypertension in the US population of 12.3% which ranged from 10.6% among Hispanics and 11.8% among Non-Hispanic whites, to 15.7% among African Americans [17•].

Individuals taking antihypertensive medication have also been found to have an increased prevalence of masked hypertension as compared to those not taking antihypertensive medication. Among 6432 individuals with non-elevated clinic BP in the International Database of Ambulatory Blood Pressure in relation to Cardiovascular Outcomes (IDACO), the prevalence of masked hypertension using daytime BP on ABPM was ~ 1.7 times higher for individuals taking versus not taking antihypertensive medication (31.9% treated versus 19.2% untreated) [18]. Similarly, an analysis of the Jackson Heart Study demonstrated among African Americans the prevalence of masked hypertension using daytime BP on ABPM was 1.5 times greater for individuals on treatment compared to those not taking antihypertensive medication (32.8 versus 21.5%, respectively) [19•].

The prevalence of masked hypertension also varies when different ABPM periods (daytime, 24-h, and nighttime) are used to define out-of-clinic hypertensions status. Examining data from the IDACO cohort, Asayama et al. determined that the prevalence of masked hypertension among individuals with non-elevated clinic BP ranged from 13.8% when out-of-clinic hypertension was defined only using the 24-h BP to 27.9% when using the daytime, 24-h, and/or nighttime BP [5•]. In the Jackson Heart Study, the prevalence of masked hypertension defined separately using the daytime, 24-h, and nighttime periods were 22.1, 26.6, and 41.7%, respectively [20•]. When all three periods were used—having elevated daytime, 24-h, and/or nighttime BP—the overall prevalence was 44.1%. Therefore, it is possible that, in some populations, using only daytime BP or 24-h period to define masked hypertension would considerably under-estimate the prevalence of masked hypertension. Which period should be used to define masked hypertension, and if the same criteria are appropriate for all populations, is an important area of ongoing and future research.

Out-of-Clinic Blood Pressure Measurement Strategies to Define Masked Hypertension

Multiple studies have demonstrated that out-of-clinic BP, as measured by ABPM or HBPM, is a better predictor of target organ damage [21, 22] and fatal and nonfatal cardiovascular events [23,24,25] than clinic BP. ABPM are wearable, portable devices which are worn continuously and measure BP automatically at predetermined intervals (typically every 15–30 min over a 24-h period) usually with the oscillometric method. In contrast, HBPM typically includes the use of a patient-initiated oscillometric BP measurement device. As compared to ABPM, almost all home BP devices record measurements only during the awake period. However, HBPM may be more practical than ABPM, as it is less cumbersome for the patient, less expensive, and more widely available [13, 19•, 26••].

The question of which method should be used to detect masked hypertension is complicated by the question of which method is superior for measuring out-of-clinic BP. Multiple studies have examined whether ABPM or HBPM better predicts CVD events and mortality [27•]. A systematic review by Shimbo et al. found nine articles including seven unique cohorts where both ABPM and HBPM were used, and CVD events and/or mortality outcomes were reported [27•]. The authors found that both ABPM and HBPM were independently associated with CVD events and mortality; however, there was insufficient evidence to consider either modality superior for predicting CVD events and/or mortality.

There also is substantial disagreement in the diagnosis of masked hypertension when using ABPM versus HBPM. In a study by Stergiou et al. [28] among individuals who underwent both ABPM and HBPM to make the diagnosis of masked hypertension, only 44% were diagnosed as having masked hypertension on both modalities; 34% were diagnosed on only ABPM and 22% diagnosed on only HBPM. A separate study by Viera et al. found daytime ABPM and HBPM to agree 72.1% of the time (kappa 0.36) for diagnosing masked hypertension [29•]. As different diagnoses can be reached when using alternate out-of-clinic BP measurement modalities, the preferred method for clinical practice remains uncertain.

Additionally, the optimal methodology to measure clinic BP is also an ongoing area of investigation. In most studies, clinic BP has been measured using either a mercury sphygmomanometer or an automated oscillometric device by trained medical staff. However, there is recent interest in the use of a fully automated oscillometric device that is able to obtain multiple clinic measurements without an observer present (i.e., unattended clinic BP) [30,31,32,33]. Compared to attended clinic BP, unattended measurements obtained with an automated device are lower [33, 34•] and correlate more strongly with daytime BP on ABPM (r = 0.145 versus 0.571) [31]. It has been suggested that routine use of an unattended automated device could decrease the prevalence of white coat hypertension [11, 33]. However, measuring lower BP levels in the clinic may have the unintended effect of increasing the prevalence of masked hypertension as individuals who would have previously been identified as having sustained hypertension will now have masked hypertension, because clinic BP is no longer elevated. The consequences of using the unattended clinic BP method on the diagnosis of masked hypertension should be considered if such strategies for clinic BP measurement become standard practice.

Whom to Screen

ABPM and HBPM are recommended by many guidelines, scientific statements, and position papers for excluding white coat hypertension among those with elevated clinic BP [10,11,12, 35]. The optimal approach for the detection of masked hypertension is unknown [36]. Screening all individuals with non-elevated clinic BP for masked hypertension is impractical. For example, Booth et al. determined that such an approach in the USA would require 118.6 million adults undergo ABPM [37]. Consequently, several studies have now examined the utility of using clinic BP thresholds to identify populations at increased risk for having masked hypertension with varying success. A higher clinic BP including BP levels in the prehypertensive range (120–139/80–89 mmHg) is associated with a higher prevalence of masked hypertension. However, in a US population, screening individuals with clinic BP in the prehypertension range was found to have a sensitivity for detecting masked hypertension of 82.5% [37] and would still require 59.3 million adults to be referred for ABPM. Among individuals with clinic BP in the prehypertension range, Viera et al. tried to identify a clinic BP threshold above which masked hypertension may be more likely and individuals should be referred for ABPM [38•]. Although, in this cohort, a clinic BP cutoff of 120/82 mmHg would have the best operating characteristics for detecting masked hypertension, they determined that it would result in high false positive rates as approximately 40% of individuals meeting this BP cutoff would not have masked hypertension [38•]. The authors concluded that clinic BP alone may not be a sufficient screening tool to guide decisions on whom to screen for masked hypertension.

A CVD risk-based approach, limiting the use of out-of-clinic BP measurement to those individuals with multiple risk factors for masked hypertension, has similarly been examined [39, 40]. Prior studies have identified risk factors such as male sex, smoking, diabetes, and higher clinic BP to be associated with masked hypertension [13]. One theorized approach for identifying masked hypertension is therefore to screen individuals with metabolic syndrome as components of the metabolic syndrome have been associated with masked hypertension and higher out-of-clinic BP [41]. However, studying the association between metabolic syndrome and masked hypertension, Colantonio et al. found that using metabolic syndrome, which includes clinic BP, to identify individuals at risk for having masked hypertension would not provide additional predictive information beyond clinic BP alone [39]. A strategy with more promise may be to focus masked hypertension screening on those individuals at increased baseline CVD risk. We previously found that higher 10-year predicted CVD risk, using the pooled cohort risk equations, was associated with a higher prevalence of masked hypertension [42•]. Although 10-year predicted risk was not superior to clinic BP for predicting masked hypertension, the majority of individuals with masked hypertension had a 10-year predicted CVD risk ≥ 10%. Risk prediction equations may therefore help identify individuals with masked hypertension who would derive the most benefit from antihypertensive treatment. Whether this finding can be applied to all populations, and how such a strategy may impact CVD outcomes, is unknown.

Adverse Outcomes

Masked hypertension has also been associated with target organ damage. Prior studies have found masked hypertension to be associated with renal dysfunction (reduced estimated glomerular filtration ratio: eGFR, proteinuria) and vascular dysfunction (increased pulse wave velocity) [43, 44•, 45•]. Masked hypertension has also been associated with increased left ventricular mass index. In a recent meta-analysis by Cuspidi et al., as compared to individuals with sustained normotension, individuals with masked hypertension had an increased left ventricular mass index (79.2 ± 0.35 g/m2 versus 91.6 ± 4.0 g/m2, respectively) and an increased prevalence of left ventricular hypertrophy (3.7% versus 14.1%, respectively) [46••]. A separate meta-analysis found masked hypertension to be associated with increased carotid intima-media thickness, a presumed measure of early carotid atherosclerosis [47••].

At present, multiple studies have found masked hypertension to be associated with an increased risk for CVD events including stroke or myocardial infarction. Compared to sustained normotension, the hazard ratio (HR) for CVD events for masked hypertension approached that of sustained hypertension (2.09, 95% confidence interval (CI) 1.55–2.81 and 2.59, 95% CI 2.0–3.35, respectively) [8]. Studies included in this meta-analysis used ABPM, and masked hypertension was defined based on elevated daytime BP in all but one study which examined mean 24-h BP; none of the studies defined masked hypertension by nighttime BP measurements. More recently, masked hypertension based on nighttime readings is also associated with an increased risk of CVD events. In a study by Booth et al., masked hypertension based on nighttime readings was associated with a greater than twofold increased risk of CVD events among African Americans [20•]. There is also evidence that masked hypertension identified using HBPM is associated with increased CVD risk. Using data from the International Database of HOme blood pressure in relation to Cardiovascular Outcome (IDHOCO), Stergiou et al. found, among 6458 participants who had undergone HBPM, risk for CVD events was higher for those with masked hypertension (HR 1.55, 95% CI 1.12–2.14) compared to sustained normotension [48•].

As previously mentioned, there is often disagreement between ABPM and HBPM when diagnosing masked hypertension. It is unknown if individuals with masked hypertension on ABPM but not HBPM have the same risk of CVD events as individuals with masked hypertension on HBPM but not ABPM. Studying Japanese adults in the Ohasama cohort, Satoh et al. looked at the association of stroke with masked hypertension diagnosed on ABPM only (using daytime, 24-h, and/or nighttime periods), HBPM only, or on both out-of-clinic BP measurement modalities [49••]. They found that masked hypertension detected on both ABPM and HBPM was associated with an increased risk of stroke (HR 2.05, 95% CI 1.23–3.41) compared to sustained normotension defined as having non-elevated clinic BP, non-elevated BP on ABPM, and non-elevated BP on HBPM. Further, this risk was also elevated among individuals with masked hypertension diagnosed only on ABPM but not on HBPM (HR 1.93, 95% CI 1.15–3.24), and when diagnosed on HBPM but not ABPM (HR 2.26, 95% CI 1.32–3.89). There were no direct statistical comparisons made between individuals with masked hypertension on only ABPM versus those with masked hypertension on only HBPM [49••]. Future studies aimed at contrasting how masked hypertension diagnosed by ABPM versus HBPM relates to outcomes are an important area for research with substantial implications for guidelines.

Treatment

Few studies have examined masked hypertension treatment, and none have examined how treatment may affect CVD events. The European Society of Cardiology/European Society of Hypertension is the only major society to recommend pharmacologic treatment or lifestyle measures for the treatment of masked hypertension and acknowledges that there is minimal evidence (Level of Evidence C) to support this Class IIa recommendation [10]. When considering management strategies for masked hypertension and related phenotypes, important questions remain including: [1] what out-of-clinic BP target should clinicians use to monitor treatment response, [2] if out-of-clinic BP is used, should it be monitored via ABPM (using daytime, 24-h, or nighttime BP values) or HBPM, and [3] what may be the effect of treatment of masked hypertension on CVD events.

Lifestyle modifications may be an appropriate and effective intervention for the treatment of masked hypertension. A study by Bromfield et al. categorized participants in the Jackson Heart Study as having poor, intermediate, or ideal factors of cardiovascular health as defined by the American Heart Association’s “Life’s Simple 7”—a composite measure that includes body mass index, physical activity, diet, cigarette smoking, blood pressure, cholesterol, and glucose [50•]. In multivariable adjusted analysis, masked hypertension based on daytime BP was less common among participants with a healthier lifestyle. Specifically, masked hypertension was less common among those individuals who had ideal versus poor scores for physical activity and cigarette smoking, ideal versus poor scores for diet, and ideal versus intermediate scores for blood pressure. The finding that better cardiovascular health is associated with a lower prevalence of masked hypertension suggests that lifestyle modifications may reduce the risk of masked hypertension. Further studies are required to support this hypothesis.

To our knowledge, there are currently three clinical trials examining treatment of masked hypertension (www.clinicaltrials.gov; Table 1). These studies are investigating important unknown areas of masked hypertension management including how clinical strategies targeting clinic versus out-of-clinic BP and treatment with antihypertensive medications may affect masked hypertension, target organ damage, and cardiovascular events. The findings of these studies may have an important impact regarding the treatment of masked hypertension.

Table 1 A summary of current clinical trials examining the impact of treatment of masked hypertension (www.clinicaltrials.gov)
Full size table

Conclusions

A recent study among US adults found that the majority of incident CVD events occur in individuals with non-elevated clinic BP [51]. Given the prevalence of masked hypertension and its association with cardiovascular outcomes, many of these individuals may have undetected and untreated masked hypertension. Improving outcomes for this prevalent BP phenotype should be a primary focus of future hypertension research. Despite recent advances in our understanding of masked hypertension, challenges remain which have implications for academic research and clinical care (Table 2). As we strive to improve public health outcomes and decrease cardiovascular disease burden, it is imperative that we continue to address these gaps in knowledge and work toward improving the management of individuals with masked hypertension.

Table 2 Future areas of masked hypertension research
Full size table