From Resistant Airway to Resistant Hypertension – Sleep Apnea and Disparities in Cardiovascular Disease

Hypertension contributes importantly to heightened cardiovascular (CV) risk in African-Americans, in whom it is more common and more severe than in whites. Despite more intensive therapy, only 48% of treated African-Americans meet BP goals.¹ African-Americans manifest more resistant hypertension, accompanied by aberrant diurnal BP patterns and more pressure-related complications and target-organ injury.

A potent determinant of resistant hypertension is obstructive sleep apnea (OSA). OSA is characterized by repetitive episodes of partial or complete upper airway collapse in sleep, with consequent autonomic, ventilatory and hemodynamic disruptions that, together with intermittent hypoxemia and sleep fragmentation, lead to transient BP surges and eventually to sustained hypertension. In medicated hypertensives, OSA compromises pressure control, especially at nighttime and in spite of polypharmacy.² It follows that OSA is extremely prevalent (70–90%) in resistant hypertension, is associated with abnormal day/night BP profiles, and worsens prognosis.^2,3 Importantly, OSA often remains undiagnosed and therefore untreated among African-Americans, despite increased prevalence and greater severity of OSA, relative to whites.

This predisposition to both hypertension and OSA, and the pathophysiological relationship between these conditions, argues strongly for a major role of OSA in difficult-to-control hypertension in African-Americans. Yet there is a striking paucity of research addressing this hypothesis. In the current issue of Circulation, Johnson and colleagues⁴ present important insights that help bridge this gap.

These investigators used as a platform the Jackson Heart Sleep Study (JHSS), an ancillary of the parent Jackson Heart Study, a community-based cohort of African-Americans. Instead of relying on patient-reported diagnosis or questionnaire scores, sleep apnea was objectively quantified through ambulatory polygraphy, a valid and cost-effective alternative to “gold standard” in-laboratory polysomnography, and accepted as a substitute diagnostic tool in uncomplicated patients with high pretest probability of at least moderate sleep apnea.⁵ Those with hypertension were classified as controlled, uncontrolled, or resistant hypertension. The latter group comprised 14.5% of the sample, while 48.2% had uncontrolled BP. Approximately one-fourth of the subjects had at least moderate OSA, defined as a respiratory event index (REI) of ≥15 events/hour. As the recommended definition of monitoring time (i.e., total recording time minus probable wakefulness and artifacts)⁶ was used, the derived REI more closely approximates the polysomnographically-based apnea-hypopnea index (AHI). In multivariable analysis, odds of resistant hypertension were two-fold higher in moderate or severe OSA, and cumulative nocturnal hypoxemia (defined as percentage sleep time with oxygen saturation <90%, %Sat<90%) was similarly predictive. Stratification by OSA severity revealed that severe OSA (REI≥30 events/hour) was independently associated with 3.58-times higher odds of resistant hypertension relative to non-OSA (REI<5 events/hour). A graded relationship was evident between severity of OSA and resistance to treatment.

Several clinically relevant observations emerge from this work. First, most of those with REI≥15 (94%) were undiagnosed, reaffirming the alarmingly widespread lack of recognition of OSA in African-Americans, even in those with high risk conditions like hypertension. The second notable finding relates to hypoxemia, a key mechanism mediating OSA-induced pathology.⁷ Intermittent hypoxemia stimulates chemoreflex-mediated central sympathetic outflow, renin-angiotensin-aldosterone system activation, angiogenesis, generation of reactive oxygen species and oxidative stress, and cytokine release, promoting inflammation and endothelial dysfunction. Hypoxemia is hence a strong predictor of unfavorable outcomes, often outperforming respiratory indexes, such as the AHI, traditionally used to diagnose and rate OSA severity. From a pathophysiologic perspective, the body senses and responds to hypoxemia and not to a respiratory index, which is a convenient but artificial metric of sleep apnea severity. Indeed, Johnson and colleagues⁴ show that nocturnal hypoxemia was significantly related to risk of resistant hypertension, and should be particularly commended for exploring the predictive value of different degrees of oxygen desaturation associated with abnormal breathing.

This leads to the third observation of compelling clinical relevance. Although differences between operational definitions of hypopnea requiring ≥3% or ≥4% desaturation would seem negligible, there is considerable impact on OSA prevalence and prognostic significance. Only OSA defined by the ≥4% desaturation (and not the 3%) threshold was significantly associated with resistant hypertension, echoing findings from the Sleep Heart Health Study, where only hypopneas inducing at least 4% oxygen desaturation portended increased risk of CV disease.⁸ Also, recent results from the Multi-Ethnic Study of Atherosclerosis indicate that this 4% threshold is more strongly associated with BP.⁹ Importantly, use of the more inclusive 3% cutoff would preclude detection of such an association, hindering recognition of the significance of OSA in resistant hypertension, and plausibly in other outcomes. The actionable relation between this degree of hypoxemia and outcomes challenges the current sleep scoring rules,⁶ which recommend concomitant ≥3% desaturation or arousal to score hypopneas, leading to marked increases in OSA diagnoses. While it is understandable that we should not miss OSA, lowering diagnostic thresholds increases prevalence, obscures relevance to meaningful outcomes, and may understandably generate caregiver complacency regarding need for therapeutic intervention. If everything is OSA, then nothing is OSA.

Finally, we need to consider why OSA and resistant hypertension interact so strongly. Although sympathetic overdrive, volume overload and excess aldosterone are thought to be primary culprits, the nexus between OSA and resistant hypertension is likely not restricted to direct biological mechanisms raising BP and reducing responsiveness to therapy, but also via other pathways, which are conceivably potentiated in African-Americans. Adherence to antihypertensive drugs is low amongst African-Americans, and hypertensive patients at high risk of OSA exhibit reduced compliance with medications.¹⁰ Moreover, daytime sleepiness, a sentinel symptom of OSA, is more frequent in African-Americans, and accounts for a large proportion of nonadherence to antihypertensive regimens.¹¹ Recognizing these behavioral and other factors will more quickly lead to developing remediation strategies.

A potential weakness of the study, noted by the authors, is the lack of ambulatory BP (ABP). Out-of-office BP, preferably ABP, is recommended to rule out white coat resistant hypertension.³ Ambulatory BP, particularly related to nighttime measures, outperforms conventional office BP, and its superior prognostic value is especially evident among African-Americans. Moreover, there is an urgent need for comprehensive and longitudinal BP phenotyping in African-Americans with resistant hypertension and concurrent OSA, to better discern temporal patterns and track progression.

Their data nevertheless make a compelling argument for improved detection and management of OSA in African-Americans. Barriers to OSA screening and treatment acceptance, such as access to healthcare and awareness of health implications of OSA, can be overcome by implementation of culturally-tailored education programs for both providers and communities, interactions with racially-concordant healthcare staff, and increased accessibility of sleep testing. Preliminary data from similar initiatives to enhance OSA screening are encouraging.¹² Recent findings from the Jackson Heart Study, showing that heart failure is more common in African-Americans and that overall adiposity, as measured by body mass index (BMI), is independently associated with an increased risk of incident heart failure,¹³ speak directly to these objectives. Given the high prevalence of undiagnosed OSA in African-Americans, and the direct relationship between BMI and OSA, it is likely that occult OSA may have contributed to incident heart failure, in part by increasing blood pressure. A pivotal question however, is whether treating OSA will lower BP and/or risk of outcomes such as heart failure.

Although current guidelines³ recommend identification and correction of comorbid OSA in resistant hypertension, whether treating OSA lowers BP significantly remains unclear. Antihypertensive effects of continuous positive airway pressure (CPAP), the therapy of choice for OSA, are generally modest, but patients with drug-resistant hypertension may experience greater changes, with pooled reductions of about −6 mmHg estimated from randomized controlled trials (RCT),¹⁴ along with improved nocturnal BP. Nevertheless, achievement of target BP with CPAP is unproven, and, due to short follow-up of most of these studies, evaluation of long-term benefits is elusive. Furthermore, there is substantial variability in responsiveness as the magnitude of BP decline is largely a function of adherence to CPAP. This is especially problematic because CPAP compliance is notoriously suboptimal, and tolerance is further diminished in African-Americans. Moreover, in spite of improvements in sleep, daytime symptoms, and intermediate outcomes, recent large-scale RCTs such as the Sleep Apnea Cardiovascular Endpoints (SAVE) trial¹⁵ failed to demonstrate survival advantages of CPAP. Notably, since underrepresentation of African-Americans in clinical trials persists (less than 1% of participants in SAVE were black), it is unclear to what extent such findings can be extrapolated to this population. Similar considerations on generalizability can be made regarding preferred antihypertensive drug class in OSA, or combination therapy of CPAP and antihypertensive agents, which also lack robust efficacy data. Rigorous RCTs to inform on optimal pharmacological management of high BP in OSA, with emphasis on targeted BP control strategies in high-risk groups, are urgently warranted.

Given the devastating, disproportionate toll of high BP in African-Americans, identifying and eradicating sources of inequality in hypertension risk and consequences is a national priority. As advocated also by the American Heart Association,¹ it is time to bring to center stage the role of sleep in these disparities. Johnson and colleagues⁴ take an important step towards answering this call to action. Understanding the contribution of sleep disorders and sleep habits in the excessive CV risk experienced by African-Americans, and the potential benefits of normalizing sleep, may go a long way toward mitigating such disparities.