Nocturnal Hypertension and Subclinical Cardiac and Carotid Damage: An Updated Review and Meta‐Analysis of Echocardiographic Studies

Abstract

Evidence on the association of nocturnal hypertension (NH) with subclinical cardiac and vascular damage is scanty. The authors performed a meta‐analysis to provide comprehensive information on this clinically relevant issue. Full articles providing data on subclinical cardiac and carotid damage as assessed by ultrasonographic methods in patients with NH as compared with patients with nocturnal normotension (NN) were considered. A total of 3657 patients (NH=2083, NN=1574) of both sexes were included in seven studies. Left ventricular mass index was higher in individuals with NH than in those with NN (112±4.7 g/m² vs 98±4.8 g/m²; standard mean difference [SMD], 0.54±0.16; confidence interval [CI], 0.23–0.85; P<.01). Similarly, common carotid intima‐media thickness was greater in patients with NH than in those with NN (751±34 μm vs 653±14 μm; SMD, 0.44±0.08; CI, 0.29–0.59; P<.01). The present meta‐analysis shows an association between NH pattern and increased likelihood of cardiac and carotid structural alterations.

A growing body of evidence indicates that nighttime blood pressure (BP) is more closely related to subclinical organ damage and cardiovascular prognosis than 24‐hour and daytime BPs.1 A mounting body of evidence also indicates that nighttime BP is a better predictor of alterations in target organs and cardiovascular events than the nondipping pattern.2, 3 The clinical value of the dipping/nondipping classification has been questioned based on pathophysiological and methodological considerations. First, dipping/nondipping categorization is strongly dependent on daytime BP values, which, in turn, are related to multiple variables including degree of physical activity, emotional state, coffee/alcohol drinking, smoking habit, and seasonal variability.4, 5 Second, both dipping and nondipping patterns have been shown to have a limited reproducibility over short‐ and long‐term periods in the setting of essential hypertension.6, 7 Third, dipping pattern is not synonymous with normal nighttime BP, as it may be associated with nocturnal hypertension (NH) according to current hypertension guideline cutoffs (ie, BP >120/70 mm Hg).8

NH has been reported to convey a higher risk of vascular and cardiac morphofunctional alterations as well as of nonfatal and fatal cardiovascular events and all‐cause mortality regardless daytime BP values.

Findings from the International Database of Ambulatory Blood Pressure involving 8711 individuals from 10 populations showed that patients with isolated NH (ie, daytime BP <135/85 mm Hg and nighttime BP ≥120/70 mm Hg) exhibited higher risks of total mortality (+29%, P=.04) and all cardiovascular events (+38%, P=.003) compared with their counterparts with nocturnal normotension (NN).9 The stronger predictive value of nocturnal systolic BP over daytime BP has been documented by a review of 24 prospective studies including 23,856 hypertensive patients and 9641 individuals from population‐based cohorts.10 Furthermore, among 859 diabetic patients followed‐up for 5 years in the Dublin Outcome Study, nighttime systolic BP was a powerful independent predictor of cardiovascular mortality, after adjustment for age, sex, body mass index (BMI), smoking, previous cardiovascular disease, and daytime BP.11 The resultant hazard ratios and 95% confidence intervals (CIs) associated with a 10‐mm Hg increase in systolic nighttime BP were 1.32 (CI, 1.12–1.69), 1.95 (CI, 1.18–3.20), and 1.24 (CI, 0.99–1.56) for total cardiovascular events, stroke, and cardiac mortality, respectively. Notably, the fully adjusted 5‐year risk of cardiovascular mortality conveyed by nighttime BP was approximately 1.5‐fold higher than that associated with daytime BP.

As for asymptomatic target organ damage, investigations performed in recent years suggest that NH (isolated or combined with elevated daytime BP) in untreated and treated hypertensive individuals is associated with more severe structural or functional alterations at the level of the heart, ascending aorta, and carotid artery.12, 13, 14

Although the link between NH and target organ damage has important implications for public health, information on this issue remains scanty. Hence, considering the relatively small number of studies conducted to date, we tried to expand present knowledge by performing a meta‐analysis to investigate whether and to what extent this BP phenotype is a risk factor for subclinical cardiac and carotid damage, as assessed by ultrasonographic methods.

Methods

Search Strategy and Study Selection

The present study was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) guidelines.15

Medical literature was reviewed in order to identify all articles evaluating the association of NH with subclinical cardiac and carotid damage (ie, left ventricular [LV] mass index [LVMI], carotid intima‐media thickness [IMT]) as assessed by ultrasonographic methods.

A computerized search was performed using PubMed, OVID, Embase, and Cochrane Library databases from December 1, 1976, up to August 31, 2015. Studies were identified by crossing the following search terms: “nocturnal hypertension,” “isolated nocturnal hypertension,” “nighttime hypertension,” “nighttime BP” with “cardiac damage,” “left ventricular mass,” “left ventricular hypertrophy,” “cardiac dysfunction,” “carotid intima‐media thickness,” “carotid atherosclerosis,” “carotid damage,” “echocardiography,” and “ultrasonography.” Checks of the reference lists of selected papers and pertinent reviews complemented the electronic search. Data were extracted by two independent investigators (CC and CS) and additional data were obtained by personal contact with authors of selected papers.

Specific inclusion criteria were: (1) full articles published in English in peer‐reviewed journals; (2) studies reporting data on at least one of the following continuous variables: average LVMI, carotid IMT (and standard error [SE] or standard deviation [SD]), and/or categorical variables such as LV hypertrophy (LVH), carotid plaque as assessed by ultrasonographic methods in at least 20 or more untreated or treated individuals with NH; (3) minimum data set including age, sex, and average ambulatory BP monitoring (ABPM) values (at least one of the following: average 24‐hour, daytime, or nighttime values).

Studies including children and adolescents were excluded. Only updated or largest reports were considered when multiple publications by the same research group were found in order to avoid double counting patients.

The first literature search identified 2401 papers. After the initial screening of titles and abstracts, 2250 studies were excluded, because they did not address the relationship between NH and cardiac or carotid damage, and 151 were reviewed. Of these, seven studies fulfilled the inclusion and exclusion criteria and contained sufficient clinical and ultrasonographic data to be included in the final review16, 17, 18, 19, 20, 21, 22 (Figure 1).

Figure 1.

Schematic flowchart for the selection of studies.

Study Quality Evaluation

Two independent reviewers (CC and CS) graded each study evaluating the following seven items: (1) office BP measured in two or more sessions vs a single session; (2) use of validated ABPM devices; (3) definition of nocturnal BP patterns based on two ABPM sessions or 48‐hour BP recording; (4) blind assessment of ultrasonographic examinations; (5) sample study including a total of at least 100 cases and controls in order to detect at least a 10% difference between groups in ultrasonographic variables; (6) study based on case‐control design; (7) statistical adjustment for covariates such as age, sex, BMI, total cholesterol, and ambulatory BP.

According to this evaluation, score quality was arbitrarily classified as poor (0–2 points), fair (3–4 points), good (5–6 points), and excellent (7 points).

Statistical Analysis

The aim of this meta‐analysis was to compare cardiac and carotid alterations expressed as continuous variables (ie, LVMI, relative wall thickness, carotid IMT) and categorical variables (ie, LVH and carotid plaque prevalence) in patients with NH compared with their NN counterparts. To this purpose, a pooled analysis of cardiac parameters was performed using fixed or random effects meta‐analysis by Comprehensive Meta‐Analysis Version 2 (Biostat, Englewood, NJ). The standard mean difference (SMD) with 95% CIs was used to calculate the statistical difference of the above‐mentioned continuous variables between patients with and without NH.

Demographic and clinical data provided by the selected studies are expressed as absolute numbers, percentage, mean±SD (or ±SE). Since two of seven studies provided the association between NH and cardiac variables separately in dippers and nondippers without providing pooled data, they were analyzed as separate studies. Meta‐regression analysis was used to test the linear relationship between LVMI, carotid IMT, and continuous explanatory variables such as age, clinic or ambulatory BP, BMI, and total cholesterol. Statistical significance was set at P<.05.

Heterogeneity was estimated using the I ² test; random effect models were applied when heterogeneity across studies was high (I ²>75). Publication bias was assessed using the funnel plot method (trim and fill test).

Results

Characteristics of the Studies

Table 1 shows the main characteristics of analyzed studies, including, sample size, clinical setting, mean age, sex distribution, BMI, office systolic and diastolic BP, and mean 24‐hour, daytime, and nighttime BP values in NH patients.

Table 1.

Summary of Studies Reporting Data on Echocardiographic Left Ventricular Hypertrophy in Patients With Nocturnal Hypertension

AuthorReference	Setting	Sample Size, No.	Age, y	Men, %	BMI, kg/m2	Office SBP/DBP, mm Hg	Mean 24‐Hour SBP/DBP, mm Hg	Mean Daytime SBP/DBP, mm Hg	Mean Nighttime SBP/DBP, mm Hg
Wijlkman16	DM	30	62±3	57	29.9±6.0	122±5/72±6	NA	132±11/79±7	126±9/72±6
Cuspidi17	HC	477	46±11	63	25.4±3.6	147±13/97±8	139±11/90±8	144±11/94±8	128±11/79±8
Chatsistamatiou18	HC	105	51±12	69	28.4±3.8	147±14/93±8	130±9/80±9	134±13/82±9	123±9/74±7
Ogedegbe19	PBS	81	55±12	48	30.2±6.6	124±15/76±9	126±6/75±6	127±6/77±6	124±8/71±7
Koroboki dippers 20	HC	195	55±10	60	28.0±4.0	159±18/97±11	NA	150±9/92±10	129±8/76±8
Koroboki nondippers 20	HC	243	58±11	56	28.0±5.0	154±17/94±11	NA	141±12/86±10	135±12/78±12
Androulakis21	HC	161	52±1a	53	28.9±0.6a	150±1/94±0.9a	133±0.7/81±0.8a	NA	NA
Wang dipper 22	CKD	129	44±15	76	23.8±3.3	157±22/94±14	147±11/88±7	152±12/90±7	132±10/79±7
Wang nondipper 22	CKD	661	49±16	59	23.2±3.5	153±23/89±14	145±14/84±9	145±14/85±9	144±16/83±11

Abbreviations: BMI, body mass index; CKD, chronic kidney disease; DBP, diastolic blood pressure; DM, type 2 diabetes mellitus; HC, hypertensive cohort; NA, not available; PBS, population‐based study; SBP, systolic blood pressure. ^aData are presented as absolute numbers, percentages, means±standard deviation, or standard error.

Overall, 3657 patients (NH=2083, NN=1574) of both sexes were included in seven studies (sample size range, 30–661 NH and 62–406 NN participants) performed in three different geographical areas (Europe=5, Asia=1, North America=1). Of note, the majority of the pooled NH population (58.1%) was examined in Europe.

Ultrasonographic assessment of cardiac and carotid damage associated with NH pattern was the primary aim of all studies. Participants had been recruited from hypertensive cohort,17, 18, 20, 21 population‐based sample,19 type 2 diabetic16 and chronic kidney disease settings.22

Data on Study Quality Evaluation

The score of each article ranged from three (fair quality) to five (good quality). One study had a score of three,16 three studies had a score of four,18, 19, 22 and three studies had a score of five points.17, 20, 21 Interobserver reproducibility in evaluating the quality of studies was 95%.

BP Measurements and Definition of NH

Brachial clinic BP was measured according to recommendations of contemporary international guidelines. Mercury sphygmomanometers were used in three studies17, 19, 22 and an oscillometric device in one20 The remaining studies did not provide any information on the type of device used. Of note, clinic BP was determined by taking multiple measurements during at least two visits in three studies.17, 18, 21

Ambulatory BP was measured on the nondominant arm using validated devices in all studies. A Spacelabs 90207 device (Spacelabs, Redmond, WA) was used in six of seven studies, and a Takeda device (TM 2430, Tokyo, Japan) was used in the remaining study.22 Instruments were set to take BP readings at different time intervals (15–20 minutes during daytime and 15–30 minutes during nighttime).

Nocturnal hypertension was defined as a mean nighttime systolic and/or diastolic BP ≥120/70 mm Hg in five studies, and as a mean nighttime systolic BP ≥120 mm Hg in two studies.20, 22

Nighttime period was defined according to different criteria. Day‐active and night‐rest periods were defined according to individual diaries in three studies.16, 18, 19 A nighttime interval from 11 pm to 7 am was used in two studies17, 21 and the interval from 1 am to 6 am 20 and from 10 pm to 7 am 22 was adopted in the remaining two studies.

Clinical Characteristics of NH Patients

All NH patients included in the selected studies were free of clinical or laboratory evidence of heart failure, coronary heart disease, and cardiac valve defects. Previous or current antihypertensive treatment was an exclusion criterion in all studies but one.16 The mean age range was 44 to 62 years16, 22 (pooled mean 52.5 years) and 60% of participants were men (n=1255). The average BMI ranged from 23.2±3.5 kg/m² 22 to 30.2±6.8 kg/m² 19 (pooled mean 27.2 kg/m²). The average nighttime systolic BP value varied from 123±9 mm Hg18 to 144±16 mm Hg22 and diastolic BP from 71±7 mm Hg19 to 83±11 mm Hg22; pooled mean nighttime systolic and diastolic BP was 130 mm Hg and 77 mm Hg, respectively.

Clinical Characteristics of NN Patients

The prevalence of male NN patients was 50% (n=785). The mean age ranged from 3722 to 61 years16 and BMI from 22.0±3.2 kg/m² 22 to 30.2±7.5 kg/m² 19 (Table 2). The pooled average age was 50 years and BMI 26.7 kg/m². Mean nighttime systolic BP ranged from 105±9 mm Hg22 to 113±6 mm Hg20 and nighttime diastolic BP from 61±5 mm Hg16 to 67±6 mm Hg20 The pooled mean nighttime systolic and diastolic BP was 108 mm Hg and 64 mm Hg, respectively.

Table 2.

Summary of Studies Reporting Data on Echocardiographic Left Ventricular Hypertrophy in Patients With Nocturnal Normotension

AuthorReference	Setting	Sample Size, No.	Age, y	Men, %	BMI, kg/m2	Office SBP/DBP, mm Hg	Mean 24‐Hour SBP/DBP, mm Hg	Mean Daytime SBP/DBP, mm Hg	Mean Nighttime SBP/DBP, mm Hg
Wijlkam16	DM	70	61±3	60	28.4±5.2	119±7/71±6	NA	125±9/75±6	106±7/61±5
Cuspidi17	HC	62	47±11	53	24.3±3.6	142±14/92±6	125±7/78±5	132±9/84±6	109±6/64±4
Chatsistamatiou18	HC	77	50±13	60	27.4±3.6	144±13/92±9	122±9/73±8	126±10/77±8	108±7/62±5
Ogedegbe19	PBS	176	52±11	34	30.2±7.5	117±16/74±9	116±7/70±5	120±7/74±6	108±7/62±5
Koroboki dippers 20	HC	406	54±10	49	28.0±4.0	150±17/94±10	NA	132±10/83±9	108±8/65±7
Koroboki nondippers 20	HC	93	58±10	34	28.0±4.0	143±14/90±9	NA	120±7/75±7	113±6/67±6
Androulakis21	HC	158	53±1a	48	29.2±0.6a	147±1/92±0.9a	121±0.8/74±0.6a	NA	NA
Wang dippers 22	CKD	274	38±15	62	23.0±3.7	132±19/82±12	119±10/73±7	123±10/76±7	105±9/64±7
Wang nondippers 22	CKD	258	37±14	50	22.0±3.2	126±17/81±12	114±8/70±6	115±8/72±6	109±7/67±7

Abbreviations: BMI, body mass index; CKD, chronic kidney disease; DBP, diastolic blood pressure; DM, type 2 diabetes mellitus; HC, hypertensive cohort; NA, not available; PBS, population‐based study; SBP, systolic blood pressure. ^aData are presented as absolute numbers, percentages, means±standard deviation, or standard error.

Echocardiographic Findings in NN and NH Patients

LV diameters and wall thickness were measured by M‐mode technique under two‐dimensional control, and LV mass was calculated using necropsy‐validated equations23 in all studies providing this kind of information.17, 18, 19, 20, 21, 22

LV mass was normalized to body surface area (BSA) in five of the selected studies and to height^2.7 in one.18 Nonindexed LV mass values were provided by Ogedegbe and colleagues19 In the pooled study population, mean LV mass indexed to BSA ranged from 88 g/m² 21 to 123 g/m² 20 in NH patients and from 83 g/m² 21 to 119 g/m² 16 in their NN counterparts.

As shown in Figure 2A, mean LVMI was 111.8±4.7 g/m² in NH patients and 98.2±4.8 g/m² in NN patients.16, 17, 20, 21, 22

Figure 2.

(A) Left ventricular mass index in patients with nocturnal hypertension (NH) and nocturnal normotension (NN). (B) Common carotid intima‐media thickness (IMT) in patients with nocturnal hypertension (NH) and nocturnal normotension (NN). Means±standard errors; number of patients in each group is reported in the histograms.

Figure 3 reports the SMD of LV mass indexed to BSA in 1896 NH and 1322 NN individuals: the value was positive in favor of NH patients (0.54±0.16; CI, 0.23–0.85; P<.001).

Figure 3.

Forest plot for unadjusted standard mean difference (SMD) of left ventricular mass index (LVMI) in patients with nocturnal hypertension (NH) (n=1896) and those with nocturnal normotension (NN) (n=1322) (random model, P<.001). CI indicates confidence interval. The relative weight of each study is reported on the right side.

A funnel plot excluded the presence of relevant publication bias of studies comparing LVMI in NH and NN patients. Adjustment for publication bias did not change the difference in LVMI between groups.

Sensitivity analysis showed that the final result was not substantially affected by a single study effect.

Only two studies provided data on LVH prevalence.17, 19 In the report by Cuspidi and colleagues,17 35% of NH and 13% of NN patients were found to have LVH; the corresponding figures reported by Ogedegbe and colleagues19 were 9% and 3%, respectively.

Ultrasonographic Carotid Findings in NH and NN Patients

Four of seven studies including a total of 1533 NH and 829 NN patients provided findings on carotid IMT.17, 18, 21, 22

IMT was measured at the common carotid artery level in all selected studies and calculated on two‐dimensional longitudinal section as the distance from the leading edge of the first echogenic line to the leading edge of the second echogenic line, according to Pignoli and colleagues24 and Salonen and colleagues25 In selected studies, mean carotid IMT ranged from 669±140 μm17 to 810±330 μm22 in NH and from 620±190 μm20 to 693±22 μm21 in NN patients.

As shown in Figure 2B, mean common carotid IMT was 751±34 μm in NH and 653±14 μm in NN patients.

Figure 4 reports the results of the meta‐analysis from the selected studies: SMD of common carotid IMT was positive in favor of NH individuals (0.44±0.08; CI, 0.29–0.59; P<.01). A funnel plot excluded the presence of relevant publication bias of studies comparing carotid IMT in NH and NN individuals. The adjustment for publication bias did not abolish the difference in carotid IMT.

Figure 4.

Forest plot for unadjusted standard mean difference (SMD) of common carotid intima‐media thickness (IMT) in patients with nocturnal hypertension (NH) (n=1533) and nocturnal normotension (NN) (n=829) (random model, P<.001). CI indicates confidence interval. The relative weight of each study is reported on the right side.

Finally, carotid plaque prevalence was reported in only one of the selected studies (22% in NH and 21% in NN individuals).17

Correlation Analyses

A meta‐regression analysis of data from the pooled NH and NN population (n=3238) showed that LVMI was significantly correlated with age (slope, 1.26; P<.001), BMI (slope, 3.34; P=.03), and average nighttime systolic BP (slope, 0.62; P=.04). No significant correlation was found with daytime systolic BP (slope, 0.55; P=.14) or diastolic BP either in the diurnal (slope, 0.53; P=.29) and nocturnal period (slope, 0.87; P=.15).

Meta‐regression data from a total of 2301 participants documented that common carotid IMT was significantly associated with average nighttime systolic BP (slope, 4.13; P<.01), nighttime diastolic BP (slope, 6.37; P=.01), and daytime systolic BP (slope, 4.0; P=.02). No significant correlations were found between carotid IMT, diastolic daytime BP (P=.22), age (P=.15), BMI (P=.32), and total cholesterol (P=.19).

Discussion

The present meta‐analysis performed on seven studies published since 2009 provides comprehensive and updated information on subclinical cardiac and vascular damage, as assessed by cardiac and carotid ultrasonography, in a large pooled population from different clinical settings (ie, untreated essential hypertensive patients, untreated patients with chronic kidney disease, an untreated population‐based sample of African Americans, and type 2 diabetes patients) with NH compared with NN. The main findings of our analysis were the following: (1) LV mass indexed to BSA and common carotid IMT were higher in NH than in NN patients; (2) such differences in cardiac and carotid structure were unaffected by the presence of a publication bias or single study effect; and (3) in meta‐regression analyses conducted in the pooled population of NH and NN participants, nighttime systolic BP emerged as a significant correlate of both LVMI and IMT; this was not the case for systolic daytime BP. Many aspects of our results merit further discussion.

A consistent body of evidence supports the view that subclinical modifications at the cardiac, vascular, and renal level induced by high BP have an adverse prognostic significance independently of BP levels and traditional risk factors. Ultrasonographic markers of subclinical cardiovascular damage, in particular increased LV mass and carotid IMT, have been shown to predict the risk of incident cardiovascular disease in essential hypertensive patients, as well as in population‐based cohorts.26, 27 The relationship of LVMI and carotid IMT with incident cardiovascular outcomes is continuous, and ample evidence indicates that even high‐normal LV mass or IMT values are associated with increased likelihood of cardiovascular events.28, 29, 30 Among the 1716 participants in the Pressioni Monitorate e Loro Associazioni (PAMELA) study, we were able to show that patients stratified in the highest quintile of LV mass indexed to BSA exhibited a 2.69 (CI, 1.05–6.96) fully adjusted relative risk of incident cardiovascular disease as compared with patients in the first quintile.29 Eikendal and colleagues30 assessed the relationship between carotid IMT and the first cardiovascular event (myocardial infarction or stroke) among 3067 adults 45 years and younger free from symptomatic cardiovascular disease at baseline over a mean follow‐up of 16 years; the authors found a relationship between IMT and myocardial infarction or stroke, independently of other risk factors, with the hazard ratio being 1.40 per SD increase in IMT.30

Our meta‐analysis underlines that NH patients have a greater burden of subclinical organ damage of established prognostic value, such as increased LVMI and thicker carotid wall. Of note, both markers of organ damage were significantly more altered in NH patients as compared with their counterparts after adjustment for potential publication bias and single‐study effect.

These results are in keeping with available evidence that absolute nighttime BP values are superior to daytime values in predicting target organ damage and, more importantly, cardiovascular events. For instance, in the PAMELA cohort, a 10‐mm Hg increase in nighttime systolic BP was accompanied by a much greater increase in cardiovascular mortality than a 10‐mm Hg increase in daytime systolic BP, independently of baseline BP levels.31

Almost all patients (98.5%) included in the selected studies were not taking antihypertensive treatment at the time of cardiac and carotid ultrasonographic assessment. This represents a qualifying feature of our analysis, as antihypertensive drugs may modify the relationship between BP levels and markers of target organ damage such as LV mass and carotid IMT. It has been suggested that inhibitors of the renin‐angiotensin‐aldosterone system (and calcium antagonists) are more effective than other classes of drugs in reducing LVMI, in the presence of similar BP reductions.32 Furthermore, calcium channel blockers have been shown to exert a more pronounced blunting effect on carotid atherosclerosis progression than other antihypertensive drugs, including diuretics, β‐blockers, and angiotensin‐converting enzyme inhibitors.33

NH was defined according to diagnostic criteria of European guidelines (ie, nighttime BP of ≥120 mm Hg systolic or ≥70 mm Hg diastolic) in five of seven studies; only nighttime systolic BP ≥120 mm Hg was used in the remaining two reports.20, 22 Of note, in all studies but one19 NH was defined irrespectively from daytime BP levels. Patients with isolated NH, characterized by elevated nighttime BP and normal daytime BP, represented less than 5% of the pooled population.

Literature data on the reproducibility of this BP pattern over time are scanty, as NH was defined by repeated ABPM sessions only in a few studies. In a pioneering paper, White and colleagues34 documented that NH reproducibility (nighttime BP ≥125/80 mm Hg) over a period of 4 to 8 weeks was much higher (κ >0.46) than that of nocturnal nondipping (κ <38). Investigating the short‐term variability of NH in 658 essential hypertensive patients who underwent two ambulatory BP recordings within a 2‐ to 4‐week interval,17 we found that this pattern was reproducible in approximately 85% of the cases. Recently, Abdalla and colleagues35 reported good short‐term reproducibility of the NH pattern, defined as nighttime and daytime hypertension (κ=0.65), in a community‐based sampled of adults (n=282) living in upper Manhattan. In contrast, the trait of isolated NH was poorly reproducible (κ=0.21).

Study Strengths and Limitations

Some strengths and limitations of the present report deserve discussion. A major strength is the inclusion in the sample population of untreated individuals with NH, defined according to BP thresholds recommended by current hypertension guidelines.36, 37 In addition, NH associated with daytime hypertension, a more reproducible trait than isolated NH, was the prevalent ambulatory BP pattern of selected cases (95%). Notably, between NH and NN patients, the significant SMD in LVMI and common carotid IMT (0.54 and 0.46, respectively) and the lack of publication bias or single study effect were strong arguments in favor of clinically relevant cardiovascular organ damage associated with NH. Furthermore, among risk factors correlated with subclinical organ damage, BMI did not differ between groups. Mean age was slightly higher in NH patients (P=.02), whereas total cholesterol was higher in NN patients (P=.03).

Our study has several limitations. First, due the cross‐sectional design of all studies included in our meta‐analysis, a causal relationship between NH pattern and subclinical cardiovascular damage remains unproven. Second, methodological differences adopted to define nighttime and daytime periods may have impacted the definition of NH phenotype and consequently influenced our findings. Third, our analysis of organ damage was restricted to LVMI and common carotid IMT and information on other important markers such as LV geometry, left atrium dimensions, LV systolic/diastolic function, and carotid plaque prevalence were lacking or insufficient to be analyzed. Fourth, the restriction to papers published in English may have partly affected our findings.

Conclusions

Available knowledge on the relationship between NH and subclinical cardiac or carotid damage is based on results provided by only a few single studies. The present meta‐analysis expands information on this issue by showing that alterations of cardiac and carotid structure, as assessed by ultrasonography, are more advanced in patients with NH compared with their NN counterparts. In a practical perspective, the present meta‐analysis reinforces the view that effective BP control during the nighttime period may play a pivotal role in preventing the progression of cardiac and vascular damage.

Disclosure

The authors report no conflicts of interest.

J Clin Hypertens (Greenwich). 2016;18:913–920. DOI: 10.1111/jch.12790. © 2016 Wiley Periodicals, Inc.