Before respiratory patterns were linked to specific brain injuries, progress was made in the discovery of the morphology of the medullary respiratory center. Stimulation of the pneumotaxic center increases the respiratory rate. Abnormal breathing patterns had been recognized as indicative of a primary brain lesion, and most commonly known were periodic breathing patterns. Early articles on acute neurologic disease frequently used the terms “stertorous,” “gas**,” or “overventilation or hyperpnea” without further characterization. Equally common was bradypnea by which was meant low frequency breathing with little excursions.

Neurointensivists now must field questions about whether increased respiratory drive is neurological; our answer: it all depends. There has traditionally been a reluctance to attribute tachypnea to a brain lesion rather than to a compensatory response in acute illness. Rapid breathing was most commonly Kussmaul breathing in patients with a diabetic coma. Alveolar hyperventilation was typically a response to hypoxia or metabolic acidosis or associated with a specific pulmonary disorder. Hyperventilation with respiratory alkalosis had also been incidentally noted in severe hepatic stupor and in patients in the last stages of deep coma. A series of studies linking brain injury to abnormalities of the rhythm of breathing followed these descriptions.

Hofbauer was among the first to publish a book on kurzatmigkeit (shortness of breath), which included a section on breathing patterns he recorded in stroke and predominantly in tuberculous meningitis [1]. He found Cheyne-Stokes type breathing or delayed and flattened inspiration (verlängerung und verflachung), both inspiration and expiration slow in frequency (verlangsamt), never (niemals) a Biot-type breathing—this was 25 years after Biot’s original description of periodic ataxic breathing. He offered several examples of accelerated (beschleunigte) breathing in cerebral hematoma and tuberculous meningitis, although they were always accompanied with pulmonary infection. He also found breathing frequency in meningitis to be generally increased, but pus in the ventricles up to the 4th ventricle would lead to bradypnea (about 3–5 breaths) followed by apnea (about 15 seconds). He also observed persistent apnea in cerebellar abscesses. Central hyperventilation could not yet be described because there was no biochemical proof of alkalosis. In the first half of the twentieth century, hyperventilation as a direct result of an acute brain lesion was not specifically recognized in most leading medical and neurologic textbooks (Fig. 1).

Fig. 1
figure 1

Hofbauer book’s title page and examples of fast and deep breathing

Full size image

Central Neurogenic Hyperventilation

Plum and Swanson were the first to describe brain lesion–associated hyperventilation causing hypocarbia [2]. The article reports nine patients with acute lesions in the pons, mostly caused by acute embolus to the basilar artery, and compared them with 25 patients with brain lesions elsewhere who lacked hyperventilation. Plum and Swanson were struck by the “metronomically regular, moderately deep” characteristics. In their original description, the patients all had clinical signs pointing to a pontine lesion. Many patients had constricted pupils with skew deviation of the eyes or disconjugate motion, including internuclear ophthalmoplegia in four patients. They also noted decerebrate rigidity.

As expected, the patients with central hyperventilation had a marked alkalosis. The authors also noted frequent other respiratory abnormalities, with six patients having Cheyne-Stokes-type breathing and periodic breathing, usually preceding the central hyperventilation. The study also carefully eliminated hypoxemia and metabolic acidosis as plausible causes of hyperventilation. The 25 control cases involved bilateral hemispheric lesions and unilateral hemispheric lesions. Oxygen therapy or carbon dioxide inhalation could reduce or eliminate hyperventilation.

Autopsy demonstrated medial pontine destruction. The authors hypothesized that structures in the medial pontine reticular formation were inhibitory to respiration; thus, a central neurogenic hyperventilation is the result of an uninhibited stimulation of the medullary respiratory centers by the lateral pontine reticular formation. The authors found only one report with disease in the cerebral hemispheres resulting in neurogenic hyperventilation.

Up to 50 cases have since been reported [3,4,5,6,7,8,9,10,11]. A recent review synthesizes the currently available data on management of central neurogenic hyperventilation in conscious patients with central nervous system neoplasm. As suspected, addressing the underlying lesion seemed to be the most effective therapy. For instance, the majority of primary central nervous system lymphoma cases responded to chemotherapy (89%) and radiotherapy (80%), whereas subtotal surgical resection aborted hyperventilation in all appropriate cases. Opioids were effective in muting the respiration rate [12].

The respiratory centers in the pons and medulla control both respiratory rate and alveolar ventilation. Thus, neural or humoral input or direct structural injury must be considered a potential cause of hyperventilation but also hypoventilation or apnea. Central neurogenic hyperventilation has now been well recognized in patients with catastrophic brain injury. Neurogenic hyperventilation can be seen in comatose patients with anoxic-ischemic encephalopathy and in patients with upper brainstem compression and shift from a new hemispheric lesion. If neurogenic hyperventilation is present, the disorder is often a consequence of a midbrain or pontine lesion and can be associated with pontine hemorrhages, embolus to the basilar artery, and progressive signs of brainstem compression from a hemispheric lesion. Later case reports found a correlation with brainstem tumors and hemispheric neoplasms. Other signs localizing the lesion to the brainstem are often present. For unknown reasons, patients with central neurogenic hyperventilation have a high incidence of primary brainstem lymphoma or astrocytoma. However, central neurogenic hyperventilation has also been reported in multiple sclerosis and brainstem encephalitis, but again, the responsible lesion is in the pons. The pontine lesion may impair inhibitory feedback to the medullary generators of respiratory drive and rhythm, which increases minute ventilation. Infiltration of tumor, whether lymphoma or astrocytoma, presumably destroys the inhibiting descending neurons from the pons to the medullary respiratory center. Lactate production from the tumor is another possible trigger for central neurogenic hyperventilation, although several studies seem to refute it. Some patients with less acute brainstem injury remain calm during hyperventilation but cannot hold their breath.

In traumatic brain injury, periodic hyperventilation may occur with tachycardia, fever, and sweating. This is related to sympathetic hyperactivity syndrome, which may be the most commonly underrecognized and undertreated manifestation of acute brain injury. Central neurogenic tachypnea is usually continuously present, exhausting the patient. In sustained neurogenic hyperventilation, respiratory alkalosis is substantial (pH values > 7.60 mm Hg). Infusion of potent respiratory depressants, such as fentanyl, most effectively mutes this type of breathing disorder.

We do not recall having encountered many undisputed cases but we examined a 79-year-old woman with subacute progressive gait instability. Brain magnetic resonance imaging demonstrated patchy nodular enhancement within the brainstem and cerebellum, and a brain biopsy was consistent with chronic lymphocytic inflammation with pontine perivascular enhancement responsive to steroids syndrome. (CLIPPERS) Blood gas on room air revealed pH 7.59, PaCO2 16, PaO2 103, and HCO3 15. Central neurogenic hyperventilation results from released respiratory inhibition from medial pontine nuclear damage. Our patient was comfortably tachypneic, which is often seen in a pontine lesion [6].

To make the diagnosis of central neurogenic hyperventilation, clinicians must exclude pulmonary disease. Patients with acute respiratory distress syndrome or neurogenic pulmonary edema may display a high respiratory drive, and thus spontaneous breathing could lead to uncontrolled transpulmonary pressures and possibly to patient self-inflicted lung injuries. Airway occlusion pressure is a simple, noninvasive measurement method for estimating respiratory drive during mechanical ventilation [13]. It is automatically available in almost all ventilators. Airway occlusion pressure is the negative airway pressure generated during the first 100 ms of an occluded inspiration. Because of the short duration and zero flow, it is independent from respiratory muscle weakness as well as respiratory system compliance and resistance. Excessive respiratory drive may overwhelm lung-protective reflexes (e.g., Hering-Breuer inflation-inhibition reflex), which in turn leads to high tidal volumes and further lung injury and inflammation. Clinical signs, such as dyspnea and activation of accessory respiratory muscles, strongly support the presence of high respiratory drive. Poorly functioning lungs are unlikely to “blow off” CO2 resulting in hypocapnia. In Plum and Swanson’s cases, there was severe dysfunction of medially placed pontile structures, and autopsies demonstrated medial pontile destruction in five study participants, with indirect damage to this area in a sixth. Animal models showed that a stimulation of the medullary retrotrapezoid nucleus, the key component for the CO2-dependent ventilatory drive, increases minute ventilation and breathing frequency, and stimulation of the pontine lateral parabrachial nucleus increases breathing frequency and tidal volume [14].

Significantly increased respiratory drive in neurocritically ill patients is more often due to nonneurologic causes. The many COVID-19 admissions with unrelenting respiratory drives are now legendary [15]. Perhaps, in acute brain injury, the brainstem centers generating respiratory rhythms are intact but subject to increased activation from an increased CO2-dependent ventilatory drive and mediation through catecholamines with increased intracranial pressure.

Central neurogenic hyperventilation may not be a uniform disorder but a repository of related disorders of respiratory regulation. In the absence of an isolated pontine lesion, central neurogenic hyperventilation is rarely a satisfactory explanation for measured hypocarbia. It is an oddity that made its way into textbooks but is exceedingly unusual—most certainly in the neurointensive care unit. Hoffbauer noted major changes after lesions reached the medulla, essentially identifying respiratory centers. His cases, however, included pulmonary disease from aspiration, and he could not measure blood gasses. Such a confounder is important in any assessment. To quote Plum and Swanson, “As was the case in selecting cases with central hyperventilation, patients with clinical anoxemia, frank congestive failure, or extensive pneumonitis were excluded. Even with the precautions, it was almost impossible to find persons with significant cerebral vascular disease who were completely free of manifestations of cardiopulmonary disease.”