Our main finding is that transient ICP elevations are highly associated with sleep apnea in all patients. High amplitude ramp-type transient ICP elevations were related to sleep apnea during REM sleep and sinusoidal-type with NREM sleep. All patients have transient ICP elevations without sleep apnea, especially those with IIH and pediatric-onset hydrocephalus, with similar ICP changes during REM sleep periods. Furthermore, CPAP treatment reduced the number of transient ICP elevations and changed the mean ICP during sleep.
Transient ICP elevations and sleep apnea
Intracranial pressure is influenced by various factors, including the volume of intracranial components (e.g., brain tissue, blood, CSF), the compliance of the intracranial cavity, and the resistance to cerebrospinal fluid outflow. Sleep apnea has been linked to various physiological changes affecting ICP, including heart rate, blood pressure, and respiratory function. Here we found that sleep apnea induced transient ICP elevations in all patients, like in patients without intracranial pressure disturbances4, suggesting this is a general physiological consequence of sleep apnea. The relationship between sleep apnea and patients with ICP changes has been known for many years.1, 5–7 Patients with iNPH frequently have severe sleep apnea.11–13 Accordingly, the patients with iNPH have the highest number of sleep apnea-associated transient ICP elevations of the four groups of patients (Table 1).
CO2 and transient ICP elevations
Traditionally, CO2 has been considered essential in generating transient ICP elevations, but several studies have shown mixed results.1,4, 17–19 Recently, we have shown that transient ICP elevations (B-waves) can be generated by sleep apnea and abnormal respiratory movements in the chest and not by elevated levels of PaCO2 when ICP is below 20 mmHg.4 Cerebral blood flow (CBF) regulation, through vasodilation or constriction of the cerebral blood vessels, is critical for the cerebrovasculature to respond to changes in O2 and CO2, resulting in changes in cerebral blood volume (CBV) and ICP.20–22
Furthermore, ventilation and PaCO2 are closely linked. Thus, CO2 is critical for CBF regulation and stable breathing, and compromised regulation of CBF can lead to irregular breathing. Accordingly, CO2 is most likely a modulator of transient ICP elevations when ICP is within normal levels and may primarily affect transient ICP elevations by influencing the rhythm of breathing and secondary in changing the arterial CBV. Accordingly, our results show sleep apnea-associated transient ICP elevations without changes in CO2 or O2 (Fig. 3 & Supplementary Fig. 4). However, increased CO2 is probably an important factor in the transient ICP elevations with plateau morphology seen in the three patients with REM sleep prolonged desaturation.
Physiology transient ICP elevations
The total CBV consists of 70% capillary and venous blood and only 30% arterial blood,23 and changes in CBV during hypercapnia and hypocapnia only causes changes in arterial blood volume without changes in venous and capillary blood volume.24 Thus, rapid changes in the outflow of venous blood, not affected by CO2, have significant potential to alter ICP.
Respiratory inhalation drives deoxygenated venous blood outflow from the brain and induces a counterbalancing of CSF inflow, following the Monro–Kellie doctrine of relative compartment changes.25,26 A temporary cessation of breathing, as during an apnea, will generate negative intrathoracic pleural pressure increasing venous return and initially decreasing ICP. However, as the apnea precedes, an excessively venous return to the heart will raise the central venous pressure (CVP)5 and ultimately decrease cerebral venous blood's outflow, increasing CVP and ICP. When the apnea stops and breathing is restored, the intrapleural pressure will change to positive instantly, thus further increasing CVP and ICP. Our results are consistent with this physiological association between sleep apnea and transient ICP elevations, as there is a close temporal delay between the onset of the apnea and the subsequent increase in ICP. On average, in a total apnea count of 3270 in all patients, the apnea starts 3.6 seconds before the lowest ICP level and proceeds throughout the ICP elevation ending close to the peak of the ICP increase. The ICP peak coincides with a ventilatory overshoot when respiration resumes. This is supported by a highly significant phase correlation of transient ICP elevations with the peak of respiratory movements, CVP, and arterial blood pressure, with or without changes in CO2.5,19,27
The magnitude of ICP and the manifestation of transient ICP elevations might concurrently reflect the level of compliance of the intracranial cavity. In situations where the ICP is elevated or the CSF compartment is less able to compensate for acute changes, e.g., changes in blood flow or irregular breathing (especially during REM sleep or sleep apnea), the transfer of these changes to the intracranial compartment as transient ICP elevations may be accentuated. Thus, resulting in increased and more pronounced transient ICP elevations. Furthermore, the timing of the apnea in the respiratory cycle, the level of negative intrathoracic pleural pressure, or the apnea duration could explain the heterogeneity in the amplitude of sleep apnea generated transient ICP elevations we observed since short apneas and irregular breathing only caused transient ICP elevation with small amplitudes.
Transient ICP elevations without sleep-disordered breathing
ICP elevations in patients with IIH and pediatric-onset hydrocephalus during the REM sleep phase were also observed without changes in respiration (Fig. 4) and oxygen saturation or CO2 (Supplementary Fig. 4). This suggests that other mechanisms generate transient ICP elevations and that the mechanism may differ depending on the underlying clinical condition. Especially patients with IIH and pediatric-onset hydrocephalus display similar ICP morphology compared to patients with iNPH and adult-onset hydrocephalus.
Surges in cardiac sympathetic and parasympathetic activity are an essential REM sleep feature, increasing cardiovascular variability. In REM sleep, a rapid and rhythmic increase in CBF occurs,28–30 and there is a causal influence of slow waves in CBF velocity on slow waves in ICP with a frequency between 0.095 and 0.155 Hz magnified by increasing ICP.31 Increased CBV is the critical factor in increasing ICP, whether due to decreased venous outflow, increased arterial inflow, or a combination of the two. CSF only has limited buffering capacity for acute and sudden changes in ICP. Thus changes in CSF volume are less likely to be involved in the sudden ICP increase we observed with REM onset due to the speed of changes, at least in patients with normal CSF circulation. However, the slowly declining ICP level through the REM sleep phase could be caused by a slowly adapting CSF circulation, where CSF is removed from the ventricles into the spinal canal. Accordingly, in patients with pediatric-onset hydrocephalus and IIH, the rapid increase in the ICP during the REM phase of sleep suggests that changes in CBF and CBV mainly cause the changes in ICP. We speculate if different underlying mechanisms cause transient ICP elevations during REM sleep, but most often is caused by sleep apnea or changes in respiration. However, in some patients, especially during high ICP or reduced venous buffer capacity due to venous hypertension, transient ICP elevations may be generated by oscillating changes in the cardiovascular system dictated by the autonomic nervous system (Fig. 5). Future studies should try to pinpoint the mechanism generating these slow oscillations.
Clinical implications
Idiopathic normal pressure hydrocephalus and sleep apnea
Patients with iNPH often suffer from severe sleep apnea,11–13 which is also linked to cerebral microbleeds and brain ischemias.35,36 It is possible that the transient increases in intracranial pressure caused by sleep apnea, as shown here and previously,4,5,32 could damage the brain's periventricular tissue and capillaries, leading to decreased compliance and the progression of chronic hydrocephalus, particularly in patients with iNPH. Furthermore, pressure peaks resulting from sleep apnea and the retrograde flow of venous blood through incompetent jugular valves 33,34 may enlarge the ventricles through transient venous hypertension hindering CSF absorption, particularly in the presence of brain atrophy. These factors may play a role in the pathogenesis of iNPH and warrant further investigation.
Idiopathic intracranial hypertension and pediatric-onset hydrocephalus
The exact pathophysiology of IIH is still unclear. However, several studies have shown that patients with IIH often have increased venous pressure, intracranial or systemic, and compromised venous outflow.37–39 Some have suggested that elevated intracranial venous pressure may be a universal mechanism in patients of various etiologies.37 This study does not address the metabolic and hormonal factors involved in the pathophysiology of IIH, which could also alter central venous pressure.
However, of particular interest is the relationship between increased intraabdominal pressure associated with central obesity and IIH. We note that gastric bypass is reported to achieve a much higher success in relieving symptoms than CSF VP-shunting.39 Furthermore, increased intrathoracic pressure results in increased ICP in animal models,40,41 in severely obese patients with IIH,38 and head trauma patients.42 Increased intrathoracic or abdominal pressure possibly increases superior vena cava pressure and secondarily cerebral venous pressure resulting in the decreased cerebral venous outflow.41 Interestingly, one of the patients with pediatric-onset hydrocephalus had similar REM sleep-associated high amplitude transient ICP elevations without sleep apnea, and a subsequent invasive venous pressure measurement revealed a moderately elevated superior sagittal venous pressure (12 mmHg). Furthermore, high-grade venous stenoses and cerebral hyperemia are common in childhood hydrocephalus.43 Thus, we speculate if cardiovascular changes and increased CBF during REM sleep are causing high amplitude ramp-type transient ICP elevations in patients with elevated venous pressure (and thereby lower compensatory reserve).
Venous system and intracranial pressure disturbances
The jugular veins are open when lying down, connecting the central venous system directly to the brain. Thus allowing changes affecting the CVP, e.g., during apnea or increased abdominal pressure, to be transmitted to the brain and affect ICP. The brain is protected during a seated or standing position by collapsing jugular veins, uncoupling the brain from the central venous system. Could abnormal positional changes in ICP or high ICP during REM sleep be diagnostic markers for increased CVP in patients with IIH? The role of the venous system in patients with ICP disturbances could be underestimated, and studies of the venous system a highly needed in future studies of these patients.
Continuous positive airway pressure
Continuous positive airway pressure reduced the number of transient ICP elevations with waveform abnormalities in all three patients. The increase in mean ICP in the two patients with iNPH could result from increased resistance to venous return caused by CPAP but could also be a general consequence of CPAP treatment. A case report has described a patient with iNPH who experienced worsening symptoms after starting CPAP treatment but saw improvement after a VP-shunt was inserted. Patients with iNPH have a higher frequency of retrograde jugular venous flow, possibly because of incompetent jugular valves, which may also contribute to the observed increase in ICP during CPAP treatment.33 To fully understand the impact of CPAP on ICP and venous pressure in patients, future studies should measure ICP and pressure in the venous system during CPAP treatment. Interestingly CPAP treatment in the patient with pediatric-onset hydrocephalus and a VP-shunt reduced the mean ICP (Supplementary Table 2), but the generalizability of this finding to other types of hydrocephalus and iNPH requires further study. If CPAP treatment increases ICP in patients with iNPH, it could potentially be harmful, making it essential to understand the effects of CPAP on ICP in this patient population. This study showed an increase in mean ICP in two patients with iNPH after CPAP treatment, raising concerns about the treatment's safety.
Transient ICP elevations
Transient ICP increases do not occur with regular intervals or duration. They have very different appearances and durations between patients and within patients, which speaks against a universal underlying and generating mechanism. This favors a role for sleep apnea in generating transient ICP elevations in a significant proportion of the observed ICP elevations, as apneas and hypopneas vary in duration, degree, and timing. However, several mechanisms are probably responsible for generating transient ICP elevations. When ICP is pathologically raised, physiological changes during REM sleep can play an essential role in generating transient ICP elevations (Fig. 5).
Furthermore, changes in the venous system, whether intermittent or chronic, could have a previously underestimated role in patients with ICP disturbances, which should be investigated to clarify the role of possibly elevated venous pressure in different clinical conditions. It could also be relevant to question whether transient ICP elevations generated by sleep apnea could be involved in the well-known association of sleep apnea with stroke and cardiovascular disease.44 Additionally, it will be helpful to establish an animal model of sleep apnea with simultaneous ICP measurement to elucidate the mechanism of sleep apnea in generating transient ICP elevations.