Is Brain Injury in Obstructive Sleep Apnea Reversible?

Obstructive sleep apnea (OSA) is accompanied by structural brain changes, as shown by neuroimaging findings from several groups. These neural changes likely contribute to central nervous system dysfunction in OSA, including psychological and physiologic comorbidities. However, the nature of how the brain is altered in OSA remains unclear, and in particular the question of whether irreversible brain injury predominates over acute but reversible pathology is unresolved. The study by O'Donoghue and colleagues¹ in this issue of SLEEP helps address that question, and the findings suggest that the structural changes in OSA consist of a mixture of irreversible atrophy and cell death, and non-lethal pathology such as inflammation.

The hippocampal metabolite changes are consistent with cells in this structure being in an inflammatory state. Intermittent hypoxia in animal models triggers inflammatory responses,² so presumably each OSA patient experiences a similar phenomenon due to nightly exposure to apneic events. The findings by O'Donoghue et al.¹ of differences in hippocampal metabolite levels associated with inflammation before but not after CPAP treatment are consistent with this theory. The question of whether hippocampal function can be restored with therapy was not addressed in this study, but other groups have shown improvement with CPAP therapy, albeit without resolution of cognitive function associated with hippocampal roles,^3,4 which suggests partial normalization of function. O'Donoghue et al. also found some improved cognition, although their sample was less impaired than the typical clinic population, thereby limiting the possible effect size that could be detected.

The most notable neuropsychological improvement with CPAP was in emotional functioning, with a significant improvement in depressive symptoms and quality of life. The hippocampus is a critical structure for mood regulation,⁵ and the changes in metabolite levels may reflect a normalization of such function. An interesting question is whether the depressive symptoms declined because of normalization of hippocampal function following removal of the hypoxic insults, or whether the reduction in depressive symptoms led to correction of the hippocampal regulation. An intervention that addresses the elevated depressive symptoms in OSA would help address that question. Furthermore, specifically targeting depression in OSA may help with CPAP compliance and other behavior modification, in addition to improving mood and brain function.

The white matter changes reported by O'Donoghue et al. indicate that axonal or glial pathology is present in OSA, consistent with earlier findings.^6–9 Furthermore, the pathology remains after high-compliance CPAP treatment, suggesting non-reversible changes. While white matter changes are typically interpreted as reflecting axonal and myelin damage, they very likely also reflect injury to glia, which are at least as numerous as neurons in the brain.¹⁰ Alterations in neural structures of OSA patients are seen as gray matter loss, changes in metabolite levels, and impaired water diffusion,^8,9,11–18 and the latter two may correspond to neuronal loss as well as inflammation. However, yet another unanswered question is to what degree changes in brain structures affect brain functions, a question underscored by the lack of correlation between metabolite and neuropsychological measures in the study.¹ The partial recovery of neuropsychological functions with CPAP demonstrates improvement of function, but the lack of full recovery even with these OSA patients who were highly compliant to CPAP therapy also suggests a degree of potentially permanent functional deficits.

Perhaps the main limitation of the study by O'Donoghue et al.¹ is the limited generalizability to the typical OSA patient. The researchers recruited a homogenous sample and supported the patients for their CPAP use, resulting in an impressive compliance level and retention rate. While these OSA subjects were impaired on various neuropsychological measures, their cognitive performance and emotional status were above those of standard sleep clinic populations.¹⁹ Furthermore, although essential for the design of the small studies mandated by neuroimaging logistics, the strict exclusion criteria leave open the question as to the interaction between OSA and comorbidities that may lead to brain injury, including hypertension and diabetes,^20–22 which are strongly associated with OSA.^23,24 Other methodological limitations of the study by O'Donoghue et al.¹ include the poor resolution and signal-to-noise of MRS in the brainstem, so that localized pathology in key brainstem structures could not be ruled out. Finally, individual MRI techniques cannot resolve the nature of the structural brain changes. Ultimately a combination of measurements will be required to distinguish between cell death, neuronal shrinkage, inflammation, and axon, myelin, and glial damage. For example, the data by O'Donoghue et al. that suggest an inflammatory state in the hippocampus must be reconciled with other findings of reduced hippocampal volume.^25,26 Presumably a combination of acute and chronic brain pathology is present in untreated OSA patients, which would require a combination of volumetric, tissue water content and diffusivity, and perhaps metabolite measurements, to verify.

In summary, O'Donoghue et al.¹ have provided further evidence that OSA impacts brain functioning and structure, with changes indicative of injury to white matter axons and glia, and of an inflammatory state in the hippocampus. The inflammatory changes appear to resolve with CPAP treatment, which also is associated with partial recovery of cognitive function, and strong improvement in mood and quality of life. However, white matter changes likely reflect injury that is not reversible with CPAP.

DISCLOSURE STATEMENT

The author has indicated no financial conflicts of interest.

CITATION

Macey PM. Is brain injury in obstructive sleep apnea reversible? SLEEP 2012;35(1):9-10.