Introduction

There is growing evidence that obstructive sleep apnea (OSA) represents a major cardiovascular risk factor and is associated with increased cardiovascular and cerebrovascular morbidity and mortality.1, 2

Recent studies have shown that the patients with OSA and hypertension are more likely to display signs of organ involvement—thickening of the carotid wall of the intima-media thickness (cIMT), left ventricular hypertrophy (LVH) and/or microalbuminuria.1, 2, 3 A major drawback of the studies done so far is that the role of OSA in these changes could not be established unequivocally. Most of the reports evaluating the association of OSA and hypertension with early atherosclerosis included patients with high burden of cardiovascular risk factors including established cardiovascular or cerebrovascular disease, severe obesity or diabetes, or patients receiving varied pharmacological therapy.1, 2, 3

However, the relationship between alterations in cardiovascular system and OSA has not been previously studied in middle-aged patients with never-treated hypertension without overt cardiovascular diseases.1, 2, 3, 4

The aim of our study was to evaluate the prevalence of newly diagnosed OSA in middle-aged patients with never-treated essential hypertension stages 1–2 without overt cardiovascular diseases. Moreover, we addressed the question if there is a relationship between the severity of newly diagnosed OSA and alterations in cardiovascular system. Therefore, we evaluated several signs of organ involvement including markers of left ventricular (LV) structure and function, carotid wall thickening and microalbuminuria.

Methods

Study subjects

In years 2008–2010, 121 consecutive patients with never-treated essential hypertension, age 18–50 years, fulfilling inclusion criteria, attending the outpatient clinic of the Department of Hypertension, Institute of Cardiology, Warsaw, Poland, were asked to enter the study.

Before entering the study, clinical examination was performed to determine demographics, cardiac history, known duration of hypertension, cardiovascular risk factors and to exclude related comorbidities. Hypertension was diagnosed if office systolic and diastolic blood pressure (BP) levels were 140 and/or 90 mm Hg or more and if the mean daytime ambulatory systolic BP and diastolic BP levels were higher than 130 and/or 85 mm Hg.5 Thus, patients with white-coat hypertension were not included into the study.

Subjects were excluded for any of the following reasons: secondary hypertension and/or related comorbidities including positive history or clinical signs of coronary heart disease, cerebrovascular disease, diabetes mellitus, previous treatment with the hypoglycemic, hypolipemiant and/or antihypertensive agents, neoplastic, hepatic or renal disease, estimated glomerular filtration ratio <60 ml min−1 per 1.73 m2, disabling diseases or refusal to provide an informed consent.

Study design

The protocol included evaluation of: (1) office BP and ambulatory BP measurements, (2) sodium and potassium, lipids, blood count, fasting plasma glucose (FPG), creatinine and uric acid concentrations; (3) OSA: Epworth Sleepiness Scale and full night polysomnography and (4) markers of cardiovascular alterations in all patients. The study was approved by the Ethics Committee of the Institute of Cardiology, Warsaw, Poland and informed written consent was obtained from each patient.

Measurements and definitions

Anthropometric evaluations

All patients underwent standard clinical evaluation. Known duration of hypertension and neck and waist circumferences were recorded. Large neck circumference was defined as ⩾41 cm or ⩾43 cm for women and men, respectively.6 On the basis of weight and height, body mass index was calculated. Abdominal obesity was defined as a waist circumference >102 and >88 cm for men and women, respectively.5 Current smoking, history of alcoholism, adrenergic medications and oral contraceptives were also documented.

Laboratory methods

Biochemical evaluation of blood samples taken after overnight fasting were determined by routine methods, and included sodium and potassium, lipids, blood count, FPG, creatinine and uric acid concentrations. Estimated glomerular filtration ratio was calculated using the Modification of Diet in Renal Disease formula.7 Glucose and lipid metabolism abnormalities as well as metabolic syndrome (MS) were diagnosed on the basis of the 2007 European Society of Hypertension/European Society of Cardiology (ESH/ESC) guidelines—normal FPG: <5.6 mmol l−1 and impaired FPG: 5.6–6.9 mmol l−1; increased low-density lipoprotein cholesterol concentration: >3.0 mmol l−1.5 For diagnosis of MS, three of five criteria had to be met: (1) BP ⩾130/⩾85 mm Hg—this criterion was met by all patients; (2) abdominal obesity as defined above; (3) high-density lipoprotein cholesterol male <1.0 mmol l−1, female <1.2 mmol l−1; (4) triglycerides >1.7 mmol l−1; and (5) FPG ⩾5.6 mmol l−1.5

Office BP

Office BP was measured by a trained nurse with a patient in the sitting position after a 5-min rest, using an automated device (Omron 705IT, Omron, Kyoto, Japan). Three consecutive readings were performed at two separate visits. If the difference between readings was higher than 10 mm Hg, further measurements were taken so as to obtain three consecutive consistent readings at each visit. The average of these readings was recorded.

Ambulatory BP measurements

In all patients, ambulatory BP measurements were recorded using SpaceLabs 90207 or 90217 (Redmond, WA, USA). Readings were obtained every 15 min during the day (0600–2200 hours) and every 30 min during the night (2200–0600 hours). Differentiation between sleep and activity periods was made after recoding, based on data from the patient diary. Nocturnal decrease in BP was quantified as the relative decrease in nocturnal BP for both systolic and diastolic BP: ((daytime pressure—nighttime pressure)/daytime pressure) × 100 and expressed as a percentage. Subjects were classified as dippers if the proportional decrease from awake to asleep BP was ⩾10%.

Polysomnography

The Epworth Sleepiness Scale was evaluated before the sleep study. In all patients, data regarding mean sleep duration time and snoring were also assessed. The diagnosis of OSA was made by standard attended polysomnography with an Alice 5 (Respironics, Murrysville, PA, USA) device was performed in all subjects. The polysomnography recordings were scored manually using 30-s epochs following Rechtschaffen and Kales’ criteria for sleep and wake determination and sleep staging.8 Abnormal respiratory events were evaluated according to standard criteria of the American Academy of Sleep Medicine Task Force.9

Apnea was defined as the absence of airflow for ⩾10 s. Hypopnea was scored if there was a ⩾30% reduction in nasal pressure signal from baseline with a ⩾4% desaturation for ⩾10 s. Hypopnea was also scored with ⩾50% flow reduction associated with arousal. The apnea/hypopnea index (AHI) was the sum of apneas and hypopneas per hour of sleep. OSA was diagnosed as mild OSA (AHI 5–15 events h−1) or moderate-to-severe OSA (AHI>15 events h−1).

Assessment of cardiovascular alterations

Carotid ultrasonography

The carotid ultrasound followed by a duplex color Doppler examination was done with a Phillips ATL 5000 (Philips, Eindhoven, The Netherlands) and a linear probe 7.5–12 MHz. Both left and right carotid arteries were analyzed. Multiple measurements on the distal wall from anterolateral and posterolateral longitudinal views were recorded. cIMT was measured at three points (3 mm apart) in two segments 1 cm from the flow divider caudally (carotid bulb) and 1 cm caudally from the beginning of the common carotid bulb (common carotid). The cIMT value was calculated as an arithmetical mean of 24 measurements from bulb and common carotid segments of both sides.

Echocardiography

Studies were performed with the use of a high-quality commercially available echocardiograph (Vivid 7, GE Healthcare, Little Chalfont, UK) equipped with 2.5 MHz imaging transducer. End-diastolic LV diameter (LVEDd) and end-systolic LV diameter, interventricular septum thickness (IVSTd) and posterior wall thickness (PWTd) were calculated from two-dimensionally guided M-mode tracing from the parasternal long-axis view, in accordance with the American Society of Echocardiography recommendations.10 LV mass was calculated from end-diastolic measurements using anatomically validated formula,11 and normalized by body surface area–LV mass index. LVH was defined when LV mass index was equal or exceeded 125 g m−2 in men and 110 g m−2 in women. Relative wall thickness (RWT) was calculated as the PWTd+IVSTd/LVEDd and indicated concentric LV geometry if at least 0.42. LV systolic function was determined by the ejection fraction calculated in M-mode (Teicholz) at the parasternal long-axis and midwall fractional shortening.12 LV filling was assessed by recording mitral flow by standard pulsed Doppler technique with the sample volume placed at the tips of the mitral leaflets, from the apical four-chamber view. The following parameters were considered: peak early transmitral flow velocity (E), late-diastolic flow velocity (A) and the ratio of the early to late peak (E/A ratio). The pulsed tissue Doppler imaging examination was performed from the apical four-chamber view, with the sample volume placed along the myocardial lateral wall and the left side of interventricular septum, 1 cm above the mitral annulus. Using the average of above two values, we measured myocardial early-diastolic velocity (E′), end-diastolic velocity (A′) and their ratio (E′/A′). We evaluated also E/E′ ratio. Left atrial size was determined in the parasternal long-axis view as the maximal distance between the posterior aortic root wall and the posterior left atrial wall at end-systole and normalized by body surface area. All calculations were performed during three consecutive cycles.

Other

Urinary albumin excretion (UAE) was assessed using two 24-h urine collections (the average of two collections was recorded). Increased cIMT, LVH and microalbuminuria were defined according to ESH/ESC criteria.5

Statistical analysis/analysis

Data analysis was carried out using statistical software PASW Statistics 18 (SPSS, Chicago, IL, USA). The results are presented as mean±1 s.d. or median and interquartile range. The values of variables between groups were compared—continuous and discrete variables: Mann–Whitney test or univariate analysis of variance with post-hoc Dunnett’s t-test or T3 Dunnett’s t-test depending on Levene’s test of homogeneity of variances; categorical variables: χ2 test or Fisher’s exact test. Parameters identified as statistically significant based on univariate analysis (P<0.05) were included in the multivariate logistic or linear regression models in order to determine the combined effect of several variables on the incidence or value of the characteristic. P<0.05 was considered statistically significant.

Results

We evaluated 121 consecutive patients with never-treated essential hypertension (mean age 35.9±10.1 years; 97 men and 24 women). Mild OSA was diagnosed in 36 patients (29.8%) and moderate-to-severe OSA in 24 patients (19.8%). Polysomnography parameters are shown in the Supplementary Table 1. Patients with moderate-to-severe OSA as compared with patients without OSA were characterized by higher frequency of MS and its components (Table 1).

Table 1 Clinical characteristics of the studied patients
Full size table

Table 2 presents office BP and ambulatory BP measurement levels across the evaluated groups. Although generally patients with OSA were characterized by higher ambulatory systolic BP as compared with patients without OSA, the difference reached significance only for comparison between patients with mild OSA and without OSA (Table 2). No significant differences in BP profile between the groups were noted (Table 2).

Table 2 Clinic and ambulatory blood pressure levels in relation to the presence of OSA and its severity
Full size table

Analysis was performed to determine the predictive value of the clinical variables associated with the moderate-to-severe OSA. Increasing age (odds ratio (OR) 2.00; P=0.015 for 10-year increase), large neck circumference (OR 3.34; P=0.025) and increased triglycerides levels (OR 3.07; P=0.032) but no other parameters including ambulatory BP measurement levels, nighttime BP decrease and Epworth Sleepiness Scale score predicted independently the presence of moderate-to-severe OSA (Supplementary Table 2). When in this model, components of MS were replaced by the presence of MS—it was the only variable significantly associated with the presence of moderate-to-severe OSA (OR 8.9; 95% CI 3.0–26.2; P<0.001).

The overall frequency of established target organ damage was low—LVH, increased cIMT or carotid plaques and microalbuminuria were observed in 4%, 10% and 5% of patients, respectively. Patients with moderate-to-severe OSA but not with mild OSA were characterized by higher cIMT as compared with patients without OSA (Figure 1). There were significant correlations between AHI, mean saturation and minimal saturation and cIMT (r=0.30; P=0.002; r=−0.23; P=0.014; r=−0.31; P<0.001, respectively). However, in a multivariate model only age and decreased high-density lipoprotein were significantly and independently correlated with cIMT (Supplementary Table 3).

Figure 1
figure 1

Carotid intima-media thickness according to the presence of obstructive sleep apnea (OSA) and its severity. The data are shown as mean value ±1 s.d.

Full size image

Patients with moderate-to-severe OSA but not patients with mild OSA were characterized by higher 24-h UAE as compared with patients without OSA (14.5±6.9 vs. 10.0±8.0 mg 24 h−1; P=0.014 and 9.8±4.0 vs. 10.0±8.0 mg 24 h−1; P=0.99, respectively). In a multivariate model, UAE was independently related with FPG and AHI (Supplementary Table 4).

Patients with moderate-to-severe OSA as compared with patients without OSA were characterized by higher indices of LV concentric remodeling as well as by worse indices of diastolic function (Table 3). There were no differences in parameters describing neither LV systolic function nor in LVM and LV mass index. In separate multivariate analysis, AHI correlated independently of other factors both with RWT and E′ (Table 4).

Table 3 Echocardiographic parameters in relation to the presence of OSA and its severity
Full size table
Table 4 Variables related with relative wall thickness (a) and E′-wave velocity (b)—univariate and multivariate linear regression analysis
Full size table

Discussion

First, the novel finding of our study is to document high prevalence of newly diagnosed OSA in middle-aged patients with never-treated hypertension without overt cardiovascular diseases. However, it is difficult to compare our results with those from other studies as in most of them the prevalence of OSA was determined in patients with well-controlled hypertension or drug-resistant hypertension.3

The non-dipping pattern of BP during the night has been reported as a feature of patients with OSA.3 However, in two recent studies, which included patients with resistant hypertension, no associations between higher nighttime BP levels nor nocturnal dipping pattern and the presence of OSA were reported.13, 14 Our results also showed no differences in nighttime systolic BP and diastolic BP levels nor in BP nighttime decline between patients with OSA as compared with those without OSA. It is also of interest, that in contrast to some previous studies, blunted nocturnal dip was not correlated with the markers of cardiovascular alterations. Also, in a recent study involving a large group of middle-aged untreated hypertensive patients, Cuspidi et al.15, 16 found that dippers have a similar subclinical cardiac and extra-cardiac organ damage as their nondipper counterparts.

Our study showed that the prevalence of MS was significantly higher in subjects with mild and moderate-to-severe OSA than in patients without OSA. We also found that the MS independently of other factors predicts the presence of moderate-to-severe OSA in hypertensive subjects. Our results support the concept that OSA is tightly linked to MS and represents an independent atherosclerotic burden in this population.17

Second, what is the novelty of our study, we examined whether there are echocardiographic differences between never-treated hypertensive patients with and without newly diagnosed OSA. Our results confirm this assumption, being in line with the findings of other studies reporting high prevalence of concentric hypertrophy or remodeling in population with hypertension and OSA.18, 19, 20, 21 In fact, we documented only increased IVSTd and PWTd in patients with moderate-to-severe OSA as compared with patients without and with mild OSA. Our analyses also showed positive and independent association between RWT (index of LV geometry) and severity of OSA. Thus, in spite of normal LVM, these patients had a propensity to remodel ventricle in concentric geometry. The association between OSA and LV geometry have been previously investigated in studies conducted in patients with obesity and hypertension.21, 22 Both overweight and increased BP determine a chronic LV overload and are two of the main factors leading to LVH. However, in our study abdominal obesity and BP levels were not related to RWT in the multivariate analysis, suggesting that severity of OSA had a more relevant effect on LV structure. This is in line with the findings of the study by Cioffi et al.,23 which included 157 patients with suspected OSA undergoing polysomnography. The patients were older as compared with our study and comprised normotensive and both treated and untreated hypertensive subjects. It was also shown that moderate-to-severe OSA was related with increased RWT and concentric geometry independently of BP levels and obesity. The authors proposed that this phenomenon may be related to some consequences of severe OSA including significant hypoxemia, increased sympathetic drive and/or transmural pressure because of the respiratory effort during apneic/hypoapneic event.

In our study, systolic function assessed by ejection fraction was normal in all three groups. What more, midwall fractional shortening did not differ between study groups. It is in contrary to previous studies in which in hypertensive patients systolic shortening is often depressed.12, 24 It may be explain by relatively normal thickness of LV wall and LV mass index in all three groups. However, diastolic function differed between study groups. In patients, without evidence of systolic left heart failure, we found a tendency toward diastolic dysfunction. In addition to this, myocardial diastolic function decreased with increasing OSA severity even so all the values were within what is thought to be the ‘normal range’.

The previous studies on LV diastolic function in patients with hypertension and/or OSA have yielded conflicting results,25, 26, 27, 28, 29, 30, 31, 32 depending on methodological approaches and selection of study population. Most studies have reported impaired diastolic function in patients with OSA. However, Niroumand et al.29 showed that OSA does not impair LV diastolic function in a large population. In this study, diastolic function was assessed only by conventional Doppler echocardiography (E/A ratio), which have limited value for making diagnosis of LV diastolic function.29

Tissue Doppler imaging-derived parameters seem to be more sensitive in detecting subtle changes in diastolic function in patients with OSA.27, 28 The findings obtained in this study are in line with previous data obtained by tissue Doppler imaging in patients with essential hypertension26, 30 or in general population.32 In an outpatient clinic, population conventional Doppler and tissue Doppler imaging demonstrated that patients with hypertension had an impaired LV relaxation and elevated LV end-diastolic pressure, expressed by the E/E′ ratio.26 Wachter et al.32 in a cross-sectional cohort demonstrated that moderate-to-severe OSA was independently associated with impaired LV filling as assessed by comprehensive echocardiography including tissue Doppler. In our study, E/E′ ratio was significantly higher in group with moderate-to-severe OSA comparing with group without OSA. Our results may bring hypothesis that the diastolic disturbances in these patients could likely be the cumulative burden of hypertension and sleep disturbance.

Third, this study showed that in middle-aged, never-treated hypertensive patients with newly diagnosed moderate-to-severe but not mild OSA, cIMT was significantly greater as compared with those without OSA. However, the range of cIMT values observed in most of our patients was below the currently accepted arbitrary cutoff point for increased cIMT. In contrast to our, in some other studies there was a significant difference in cIMT between those with mildly elevated AHI scores and the control groups. In most of the studies, cIMT was evaluated in normotensive subjects with OSA or in patients with OSA and hypertension who were older, had longer known duration of hypertension and were characterized by higher prevalence of target organ damage than those included in our study.

The impact of OSA and hypertension alone compared with the association of both conditions on markers of carotid atherosclerosis is not fully understood. Drager et al.33 showed that the association of OSA and hypertension has additive effects on markers of carotid atherosclerosis. The study showed that cIMT is positively related to systolic BP and AHI indicating that association between OSA and atherosclerosis could be mediated by the presence of hypertension.33

Silvestrini et al.34 were the first to demonstrate that male patients with severe OSA and without any symptoms of cardiovascular and cerebrovascular disease have an increased predisposition to carotid atherosclerosis, independent of the coexistence of the classic vascular risk factors.

It should be noted that our study differs in some ways from the previous reports. In only one study, Baguet et al.35 characterized cIMT and plaque occurrence in patients with newly diagnosed OSA and never-treated hypertension without known cardiovascular disease. In this group of 83 patients with moderate-to-severe OSA, mean cIMT was 0.67 mm, 30% of subjects had carotid wall hypertrophy and the significant relationship was found between cIMT and plaque formation and nocturnal oxygen desaturation. However, it is rather difficult to compare our study with those of Baguet et al.35 as in contrast to our group, the prevalence of untreated hypertension was 65% and the investigated patients were older.

In our study, a positive correlation was also found between AHI, mean saturation and minimal saturation and cIMT. However, in a multivariate model only age and decreased high-density lipoprotein cholesterol concentration were significantly and independently correlated with cIMT. Most of the studies have shown that cIMT is independently related to the severity of OSA, as expressed by AHI or nocturnal hypoxemia. It has been reported that not only severe OSA, but even mildly elevated AHI scores predispose to early atherosclerosis.

On the contrary, as in our study Kaynak et al.36 showed, that although cIMT values and the prevalence of plaque increased with respiratory disturbance index, cIMT was not significantly affected by the severity of OSA.

Fourth, hypertensive patients with moderate-to-severe newly diagnosed OSA but not patients with mild OSA were characterized by higher UAE as compared with patients without OSA. There was a significant correlation between UAE and AHI, which remained significant in a multivariate model. This study showed that AHI correlated with UAE within the normoalbuminuric range.

Recent data indicate that the association of UAE with major cardiac events extends to urine albumin levels below the currently accepted arbitrary cutoff point for microalbuminuria. Our data are in agreement with the results of Ursavas et al.,37 who reported that low-grade UAE (0–30 mg g−1) was associated with OSA in normotensive/non-diabetic patients, being independent of age and body mass index. In contrast to our study, subjects with hypertension were excluded from the study.

In another study, Tsioufis et al.38 performed in 62 untreated hypertensive patients with OSA, it was documented that UAE was increased within the normal range in hypertensive individuals with OSA as compared with those without OSA proportionally to OSA severity independently of confounders. Supporting our study, in a multivariate model AHI was an independent predictor of UAE. Our study and those of Tsioufis et al.38 may indicate that the association of upper-airway dysfunction with albumin excretion may provide an explanatory mechanism for the OSA-related risk in hypertension.

Limitations

We included only middle-aged or younger subjects with never-treated hypertension without a history of diabetes or cardiovascular disease, thus mitigating the potential confounding effects of age and comorbidities on our results. However, our results may not be applicable to the broader population of patients with hypertension.

The absence of a control group may be considered as a study limitation. However, it does not preclude the demonstration of an independent relationship between moderate-to-severe OSA and markers of alteration in cardiovascular system.

Finally, multivariate analysis can only describe the association, but not prove cause and effect. However, data from trials using continuous positive airway pressure in patients with OSA documented that waning of OSA can result in the improvement of markers of cardiovascular alterations.

Summary

In summary, our study documented relatively high prevalence of newly diagnosed OSA in middle-aged never-treated hypertensive subjects. High suspicion of OSA should be particularly raised in patients with MS. Our study showed a relatively low incidence of established subclinical organ damage. However, patients manifesting moderate-to-severe OSA were characterized by more pronounced markers of alterations in cardiovascular system. This relationship reflects predisposition of middle-aged hypertensive subjects with OSA to develop established target organ damage. It may be postulated that early recognition of OSA in patients with hypertension may prevent the development of organ damage, which reflects an intermediate stage in the continuum of cardiovascular disease and is a determinant of overall cardiovascular risk.