Hypertension in Cameroon associated with high likelihood of obstructive sleep apnea: a pilot study

Hypertension (HTN) is a growing worldwide [1] and national public health challenge for which several health surveys in developed and developing countries have revealed a high prevalence, a large percentage of them poorly controlled [2–4]. The disease occurs as a result of an interaction between environmental, genetic, social and lifestyle factors [5, 6], requiring multifactorial and multidisciplinary management approach for favorable outcomes.

Sleep plays an important role in the regulation of body processes and as such contributes significantly to the quality of life in humans [7]. Though known to be a restorative and refreshing process, over the past 50 years habitual sleep duration has decreased by 1.5 to 2 h/day, due to the evolving lifestyle and standards of living [8]. Sleep is characterized amongst other events by complex activity of the cardiovascular autonomic mechanisms and consequent changes in arterial pressure and heart rate [8]. Poor sleep has been recognized as independent risk factor of hypertension [9, 10] as it if often characterized by recurrent episodes of intermittent hypoxia. There is growing evidence that obstructive sleep apnea (OSA) a form of Sleep Disordered Breathing (SDB) is common among patients with hypertension, ranging from 37% [11] to 56% [12] as well as evidence that managing sleep disordered breathing improves blood pressure (BP) values and response to hypertensive treatment [13].

In Cameroon, sleep studies are few [14], and to our knowledge, no study so far has been carried out to assess sleep in hypertensive participants. It was hence essential to conduct assessments of sleep patterns of hypertensive patients in Cameroon with the aid of clinician-friendly, readily-available, less costly and easy-to-use sleep questionnaires, in order to generate preliminary data on the likelihood (risk) of obstructive sleep patterns (and consequent daytime sleepiness) in patients with hypertension compared to normotensive participants from which further exploration, management and awareness-raising initiatives can be conducted.

This study sought to conduct a preliminary assessment of the likelihood (risk) of obstructive sleep apnea (OSA) and resultant daytime sleepiness in hypertensive individuals compared to normotensive participants (through an age- and sex- matched case–control study design) in the cosmopolitan city of Yaoundé, Cameroon in order to contribute to fill knowledge gaps in sub-Saharan Africa on subtle consequences of hypertension such as sleep quality which affect patient quality of life and compliance to treatment.

Our findings reported a higher likelihood (“risk”) of OSA (60% versus 20.4%, p < 0.001; OR 5.85, 95% CI 2.46–14.08; Table 4) in hypertensive compared to normotensive persons with a five-fold odds ratio even after adjusting for BMI (AOR 5.03; 95% CI, 1.90–13.33, p = 0.001). These findings are similar to what was reported in another study [18]. The suggested mechanism to explain the link between OSA and hypertension (HTN) is fluid-shift which involves nocturnal rostral fluid shift from the legs [19] leading to reduced mean upper airway cross-sectional area. Intermittent hypoxia coupled to the reduced upper airway, triggers the chemo-reflex receptors which in turn stimulates the sympathetic nervous system leading to vasoconstriction and resultant increase in BP levels contrary to expected reduction in BP levels (termed nocturnal dipping). Patients with hypertension studies have shown, are more prone to increased rostral fluid shift results in airway narrowing (which may explain why OSA is more common in patients with hypertension) which further pushes up BP worsening the hypertension [19, 20]. Further lends support to the fact that that hypertension can contribute to worsening hypertension through the OSA pathway.

Another incriminated mechanism in the pathophysiological relationship between HTN and OSA involves Aldosterone excess [21–23], a hormone which assists in the regulation of circulating blood volumes and potassium concentration via a feedback loop within the adrenal cortex. In an observational study of consecutive participants with resistant HTN, participants were evaluated for primary hyperaldosteronism using plasma rennin activity <1.0 ng/mL/h and 24 h urinary aldosterone excretion >12 μg during high urinary sodium excretion (>200 mEq/24 h), and risk of sleep apnea with the aid of the Berlin Questionnaire [21]. Patients at high risk for OSA were almost two times more likely to have primary hyperaldosteronism (36% versus 19%, p < 0.05) with low plasma renin activity (1.2 ± 1.8 ng/mL/h versus 1.9 ± 4.1 ng/mL/h) and significantly greater 24 h urinary aldosterone excretion (13.6 ± 9.6 μg versus 9.8 ± 7.6 μg, p < 0.05) compared to participants at low risk of OSA, suggesting an association between primary hyperaldosteronism and OSA [21]. In another study by the same group, increased plasma aldosterone concentration and OSA were noted in patients with resistant HTN but not in control participants, suggesting that aldosterone excess may contribute to OSA severity [23]. These authors subsequently showed that OSA prevalence was significantly higher in participants with hyperaldosteronism compared with hypertensive patients with normal aldosterone levels (84% versus 77%). These findings suggest hyperaldosteronism may be a key factor in linking resistant HTN with OSA. These studies are, however, all observational and causality cannot be directly established.

Another proposed mechanism is the involvement of the structure and function of large arteries as well as ventricular remodeling. Hypertension is tightly linked to arterial stiffness (evaluated by pulse wave velocity) [12, 24–26]. In hypertensive patients, structural, mechanical or functional remodeling (geometric alterations in the vessel wall without a change in vessel volume) due to apoptosis, low-grade inflammation and vascular fibrosis reduce small arteries and arterioles lumen diameter (determinants of vessel elasticity). These lead to decreased vascular compliance (due vascular wall structural changes) and increased peripheral resistance and impaired endothelium-dependent vasodilation (mediated in part by nitric oxide, a potent vasoactive substance produced by vascular endothelium), common in hypertensive patients [26, 27]. Vascular endothelial dysfunction (arterial stiffness) increases intra-arterial pressure (evaluated by Pulse Wave Velocity) resulting in the transudation of plasma into the interstitial space which in turn favors overnight rostral fluid shift, OSA and intermittent hypoxia. These trigger the Sympathetic Nervous System to further increase blood pressure, heart rate and arterial constriction. OSA independently is known to be associated with increased arterial stiffness which explains why the coexistence of OSA and treated hypertension is associated with additive effects on decreased vascular compliance.

Evidence suggests that OSA is associated with left ventricular (LV) hypertrophy [12, 24, 26]. Heart remodeling associated with OSA can be explained by several mechanisms. Acutely, recurrent hypoxia may have a direct effect on the aorta, promoting reduction in the amount of glycosaminoglycans and collagen as well as at the level of the heart [25]. Obstructive apneas is also characterized by negative intrathoracic pressure against an occluded upper airway, leading to LV transmural pressure increase (an important determinant of LV afterload). Cyclic increases in the BP due to apnea also contribute to increase LV afterload. These limit the aorta and proximal large vessels from serving as an elastic buffering chamber (the Windkessel function) which can store as much as 50% of the LV stroke volume during systole, leading to increased LV mass and LAD [25].

Hypertension on its part is not only associated with increased arterial stiffness but also with early signs of heart remodeling manifesting as increased Left Atrial Diameter (LAD), Interventricular Septal Wall Thickness (IVST), Left Ventricular Posterior Wall Thickness (LVPWT), LV mass index and percentage of LV hypertrophy as well as heart remodeling.

Anthropometric parameters such as BMI and neck circumference have equally been shown to affect sleep patterns in people including hypertensive patients. In our study, hypertensive participants were significantly more obese than normotensive participants. Despite adjusting for BMI with the aid of multivariate logistic regression analyses, hypertensive participants were 5 times more likely to have having a greater “risk” of OSA than normotensive participants [AOR: 5.03; 95% CI: 1.90–13.33, p = 0.001]. This was similar to what was reported in another study that studied the association between hypertension and risk of OSA assessed with the aid of the Berlin Questionnaire [28].

Though the prevalence of snoring (assessed with the aid of the Berlin Questionnaire was higher in participants with hypertension compared to normotensives (58.0% versus 44.0% respectively), the difference was not significant, p = 0.167); Table 4. This was inconsistent with the results of a similar study which revealed a significantly higher prevalence of snoring in participants with resistant hypertension [29]. This discrepancy could be explained by our relatively small sample size compared to the latter study.

We also did not observe a trend increasing risk of OSA with respect to the BP severity (assessed as per JNC classification of hypertension severity). This absence of a trend may be due to our study sample size. Larger scale studies are needed to further investigate the relationship between OSA and hypertension severity.

Resultant daytime sleepiness (measured by Epworth Sleepiness Score) was not significantly different between cases and normotensive participants. This finding is similar to that reported by other studies [27, 29–31]. The absence of a difference may be as a result of repeated sympathetic nervous system firing during the day which maintains an alert and awake state.

All the hypertensive participants in our sample were on anti-hypertensive drugs (a mean of 2 drugs per person). Amongst these participants, 12 (24%) had well controlled BP (<140/90 mmHg) as recommended by the European Society of Hypertension (ESH) and the Joint National Committee on prevention, detection, evaluation and treatment of high blood pressure [9, 10] as opposed to 38 (76%) who were not controlled (>140/90 mmHg). This proportion of controlled hypertension though higher than that of previous blood pressure control studies in Cameroon [32, 33], is lower than that obtained by Katte and collaborators in a similar setting and on a larger sample [34] as well as rates reported in the US as part of the National Health and Nutrition Examination Survey (NHANES) [8].

The tendency for persons with hypertension to seek treatment is the result of the efforts of several screening and health education campaigns. Significantly greater mean BMI, mean neck circumference and mean waist circumference among the uncontrolled HBP cases compared to those with controlled BP could possibly link account for the poor BP control.

This study has some limitations that should be considered in designing future sleep studies. First, its cross-sectional study design gives only a snapshot and its results may be different or more conclusive if a prospective design had been chosen. The relatively small sample size might explain the unintentional absence of smokers in the study sample, which affects the study’s generalizability to the general Cameroonian population where 1 in every 9 persons smokes [34]. This weakness is however a strength and suggests that the observed high likelihood of obstructive sleep apnea in patients with hypertension compared to controls, is likely due to the pathological changes explained above, observed in hypertension and not (directly) smoking, a known risk factor for sleep apnea. Sleep questionnaires though user-friendly, easy to use and cheaper, can be subjective tools in assessing sleep compared to other methods like polysomnography and actigraphy. This was minimized by coaching participants on filling the sleep questionnaires. Studies are needed to determine the sensitivity, specificity and validity of these sleep questionnaires in our settings, as quick and reliable tools to detect disordered sleep, by comparing them with gold standards. Finally this pilot study involved a small hospital-based sample. The weakness of hospital data can best be dealt with by conducting large and community-based studies which time and resources could not permit. However, the study’s hospital-based participants were age-and-sex matched with community-based controls to improve quality.

This study provides preliminary data on sleep disorders in a sample of hypertensive participants in a sub-Saharan situation. The higher prevalence of the risk of OSA in hypertensive compared to normotensive participants is worth noting. Further larger scale and prospective studies however need to be carried out using detailed investigations (actigraphy, polysomnography) and biological markers to shed more light on the pattern of sleep in participants with hypertension in our setting as opposed normotensives and how it impacts blood pressure values. More emphasis on subtle consequences of hypertension like sleep perturbation should be given during follow-up of patients with hypertension so as to limit its impact on treatment adherence, disease progression and quality of life. Sleep questionnaires validated for use in our setting against gold standards could be used in the clinical settings as simple and quick tools to determine those at risk and in need for further investigation.