Central neurogenic hyperventilation in conscious patients due to CNS neoplasm: a case report and review of the literature on treatment

Abstract

Background

Central neurogenic hyperventilation (CNH) is increasingly reported in conscious patients with a CNS neoplasm. We aimed to synthesize the available data on the treatment of this condition to guide clinicians in their approach.

Methods

We describe the case of a 39-year-old conscious woman with CNH secondary to glioma brainstem infiltration for whom hyperventilation was aborted with hydromorphone, dexamethasone, and brainstem radiotherapy. We then performed a review of the literature on the treatment of CNH in conscious patients due to a CNS neoplasm.

Results

A total of 31 studies reporting 33 cases fulfilled the selection criteria. The underlying neoplasm was lymphoma in 15 (45%) and glioma in 13 (39%) patients. Overall, CNH was aborted in 70% of cases. Opioids and sedatives overall seemed useful for symptom relief, but the benefit was often of short duration when the medication was administered orally or subcutaneously. Methadone and fentanyl were successful but rarely used. Chemotherapy was most effective in patients with lymphoma (89%), but not glioma (0%) or other neoplasms (0%). Patients with lymphoma (80%) and other tumors (100%) responded to radiotherapy more frequently than patients with glioma (43%). Corticosteroids were moderately effective. Subtotal surgical resection was successful in the 3 cases for which it was attempted.

Conclusion

Definitive treatment of the underlying neoplasm may be more successful in aborting hyperventilation. Variable rates of palliation have been observed with opioids and sedatives. Treatment of CNH is challenging but successful in a majority of cases.

Since its first description by Plum and Swanson in comatose patients, central neurogenic hyperventilation (CNH) is increasingly reported in conscious patients with CNS neoplasm.¹ Owing to its rarity, optimal medical management of CNH in this context is currently unknown. Here, we describe the case of a patient with CNH secondary to brainstem involvement of a high-grade astrocytoma, for whom successful palliative treatment of respiratory symptoms was achieved with corticosteroids, hydromorphone, and brainstem radiotherapy. We then synthesize the currently available data on management of this condition with a review of the literature on the treatment of CNH in conscious patients due to a CNS neoplasm to guide clinicians in their approach.

Case

A 39-year-old woman consulted the emergency department for subacute dyspnea. Two years prior to presentation, she had been diagnosed with isocitrate dehydrogenase–positive, MGMT (O6-methylguanine–DNA methyltransferase) methylated, anaplastic astrocytoma gliomatosis cerebri involving the corpus callosum, bilateral frontal lobes, and the left mesial temporal lobe (Figure 1A). At diagnosis, there was no evidence of brainstem extension. She was initially treated with 1 year of temozolomide 5 of every 28 days, aiming to spare radiotherapy. On radiological progression, she underwent subtotal resection of the dominant right frontal lesion followed by 3 cycles of experimental adjuvant intra-arterial carboplatin. She subsequently progressed again, and sought a second opinion at our center, where a Stupp protocol (combined radiotherapy and temozolomide followed by adjuvant temozolomide) was initiated 8 months before consulting for dyspnea. Her home medication included prochlorazine, ondansetron, and lorazepam. She had never smoked and did not abuse illicit drugs or alcohol.

Fig. 1.

Brain T2-weighted noncontrast MRI. A, Initial T2 axial brain MRI, 2 years before presentation for dyspnea, showing a solid mass lesion involving the rostrum and anterior body of the corpus callosum, extending into the bilateral frontal white matter with associated vasogenic edema. Surgical resection confirmed the diagnosis of isocitrate dehydrogenase–positive, MGMT (O6-methylguanine–DNA methyltransferase) methylated glioma. B (axial) and C (sagittal), Brain fluid-attenuated inversion recovery MRI at presentation showing bilateral paracentral pontine and medullary mass effect with secondary compression of the fourth ventricule and obstructive hydrocephalus. There is also progression of the right frontal cystic mass without contrast enhancement, as well as progression of T2 hyperintensities in the right insular cortex, anterior horn of the internal capsule, and temporal lobe white matter (not shown).

Over 4 weeks, the patient had progressively developed hyperventilation, with spontaneous tachypnea causing dyspnea at rest and permanent tingling in the peribucal area and extremities. Hyperventilation persisted during sleep and was episodically relieved by breathing into a paper bag and taking lorazepam. The patient denied cough, chest pain, sputum production, hemoptysis, orthopnea, fever, or chills. She did not have any focal neurological symptoms, positional headache, nausea, or visual obscurations. The complete review of systems was otherwise unremarkable.

On examination, the patient had a normal mental status, but was asthenic and tachypneic. Her vital signs were blood pressure 108 over 75 mm Hg, heart rate 53 beats per minute, respiratory rate 30 to 36 breaths per minute, peripheral oxygen saturation 100% on room air, and temperature 36.4 °C. Her cardiopulmonary exam was normal. She did not exhibit digital clubbing or peripheral edema. A survey neurological examination, which did not include fundoscopy, was also unremarkable.

The patient’s complete blood count, kidney and liver biochemistry, and cardiac enzymes were normal. Arterial blood gas analysis demonstrated profound hypocapnea and chronic respiratory alkalosis (pH 7.59, PaCO₂ 6.6 mm Hg, HCO₃ 14.1 mmol/L, PaO₂ 160 mm Hg). Her electrocardiogram showed sinus bradycardia (45 bpm) and a long corrected QT interval (523 msec). Pulmonary function tests and transthoracic cardiac ultrasound, performed 2 weeks after the onset of symptoms, had not disclosed any anomalies. Her chest radiograph was normal. There was no evidence of pulmonary embolism on thoracic CT angiography. Brain MRI demonstrated T2 hypersignal and edema in the paracentral pons and medulla, associated with mass effect and compression of the fourth ventricule causing obstructive hydrocephalus (Figure 1B and C).

On admission, a diagnosis of CNH secondary to glioma brainstem infiltration was made. A treatment with dexamethasone (2 mg every 6 hours) was initiated to address brainstem vasogenic edema. Owing to good mental status and no intracranial hypertension symptoms, the patient did not require cerebrospinal fluid (CSF) shunting despite radiological obstructive hydrocephalus. She received 0.25 mg of subcutaneous hydromorphone every 3 hours in an attempt to decrease her respiratory drive. After 2 days, the patient’s respiratory rate normalized (15 breaths/min) and the dyspnea subjectively improved. She declined control arterial gas analysis, content with achieved palliation. The patient was discharged home the following day with oral dexamethasone (2 mg every 6 hours) and hydromorphone (0.5 mg every 3-4 hours as needed). At 1 week follow-up, the patient no longer reported hyperventilation, nor did she make use of the opioids. The steroids were weaned and the patient underwent brainstem radiotherapy (34 Gy in 5 fractions) because the brainstem was not included in the prior radiotherapy field. A repeat MRI performed 1 month later showed decreased T2 hypersignal in the brainstem, but progression in the supratentorial matter. At the 2-month follow-up, the patient remained neurologically stable without signs of CNH relapse.

Methods

We searched Ovid MEDLINE from 1946 to January 2020 using the strategy prepared by an information specialist and provided in the supplemental material. Two authors (J.N.B. and M.C.B.) independently reviewed the study abstracts to identify eligible studies. To be included in the review, studies had to provide a detailed case report of a conscious patient with CNH secondary to CNS neoplasm. Case series and cohort studies (observational and interventional) were also eligible. There was no restriction on patient age, sex, or underlying neoplasm pathology. We excluded studies that did not report attempted medical or surgical treatment of CNH or the results of these treatments, did not provide sufficient evidence supporting the diagnosis of CNH, or were published in a language other than English or French. We searched the citation list of all included studies to further identify published cases.

For all eligible studies, we extracted and tabulated the following data: patient age and sex, neoplasm pathology (primary CNS lymphoma [PCNSL], glioma, other neoplasm), attempted medical or surgical treatments (with specified dose and route of administration, when specified), clinical and paraclinical results regarding respiratory and general status, radiological results, duration of follow-up, and reported side effects. Reported treatment outcomes specific to CNH were stratified according to neoplasm pathology as success (relief of hyperventilation with minimal or no side effects), failure (persistence of hyperventilation), mitigated outcome (unsustained benefit or mixture of positive and negative effects), or uncertain outcome. Treatments received for the causative neoplasm before development of CNH were not collected.

Results

The search strategy provided 211 unique citations, of which 30 studies describing a total of 32 patients satisfied the selection criteria. Other studies were excluded because 175 did not report a case of CNH in a conscious patient with CNS neoplasm, 4 did not report attempted treatment or its result, and the full text was unavailable for 5 studies.

Data extracted from all eligible studies and this case are reported in Table 1. Out of a total of 33 patients, 22 (67%) were male and the mean age was 39 years (range, 3-87 years; interquartile range 22-55). The underlying CNS neoplasm was PCNSL in 15 cases (45%), glioma in 13 cases (39%), and other neoplasm in 5 cases (15%). Overall, 4 patients (12%) received oxygen supplementation and 5 (15%) underwent rebreathing in a paper bag. Opioids were employed in 19 cases (58%): Eleven were treated with morphine, 1 with hydromorphone, 3 with fentanyl, 1 with methadone, and 3 with an unspecified opioid. Sedatives were tested in 14 cases (42%): Six received midazolam, 4 received diazepam, 2 received another benzodiazepine, and 2 received propofol. Finally, tumor progression believed to cause CNH was addressed with corticosteroids in 19 patients (58%), with radiotherapy in 13 patients (39%), with chemotherapy in 11 patients (33%), and with subtotal resection in 3 patients (9%; cases included pilocytic astrocytoma, medulloblastoma, and bilateral acoustic schwannomas). Four patients (12%) underwent mechanical ventilation while receiving treatment.

Table 1.

Case Reports and Series With Respective Treatments and Outcomes

Case	Patient age/sex	Underlying pathology	Treatment	Clinical result	Clinical follow-up	Radiological follow-up
Lange and Laszlo, 19652	51 M	Primary malignant cerebral reticulosis	Oxygen supplementation (FiO2 100%). Morphine (10-15 mg, routine and frequency unspecified)	Oxygen: no effect on RR. Morphine: partial decrease in RR	Neurological deterioration, died at 43 d	Not reported
Rodriguez et al, 19823	53 F	Brainstem (pontine and medullary) astrocytoma	Oxygen supplementation Morphine (4-10 mg every 4-6 hours IV)	Oxygen: No response. Morphine: decrease in RR and pH	Rapid progression of brainstem dysfunction. Died 7 d after admission	Not reported
Sunderrajan and Passamonte, 19844	41 M	CNS lymphomatoid granulomatosis	Morphine sulfate (dose and route unspecified). Prednisone (1 mg/kg/d, route and duration unspecified). Cyclophosphamide (2 mg/kg/d, route and duration unspecified)	Morphine: failure to control hyperventilation. Corticosteroids and cyclophosphamide under mechanical ventilation: improvement in hyperventilation at d 18	5 wks, no relapse	Decreased densities in left frontal lobe and disappearance of diffuse, bilateral contrast enhancing lesions on CT scan.
Gottlieb et al, 19875	23 F	Medulloblastoma	Surgical resection, neuraxis radiotherapy	Complete remission of hyperventilation	2 y, no relapse	Not reported
Nakasu et al, 19886	7 F	Pilocytic astrocytoma	Methylprednisolone (125-250 mg/d for 4 d IV). Right suboccipital craniectomy and surgical resection	Methylprednisolone: hyperventilation persisted. Surgical resection: clinical improvement	At 3 mo, still < 0.5 mg/d betamethasone, respiration slightly shallow at 24 breaths/min	Not reported
Pauzner et al, 19897	61 F	Primary cerebral lymphoma	Rebreathing in paper bag. Diazepam (IV, dose unspecified). Morphine (IV, dose unspecified). Corticosteroid therapy (dose and route unspecified). Radiotherapy (40 Gy during 2 wks) while patient was under general anesthesia and mechanical ventilation	Paper bag and diazepam: failure to alter respiratory pattern. Morphine: favorable but transient response. Corticosteroids and radiotherapy under general anesthesia: no recurrence of hyperventilation after sedation was stopped	Despite respiratory improvement, lapsed into a coma and died on d 99.	Initial radiological response to radiotherapy on CT scan, with subsequent relapse corresponding to clinical neurological decline
Jaeckle et al, 19908	32 F	Brainstem (pontine and medullary) anaplastic astrocytoma	Oxygen supplementation (FiO2 100%). Midazolam (5 mg IV). Morphine sulfate (5 mg IV). Methadone (5 or 10 mg orally every 8 h). Dexamethasone (10 mg every 6 h, route unspecified). Chemotherapy (procarbazine, lomustine, vincristine)	Oxygen: no response. Midazolam: somnolence, no reduction in RR. Morphine: decrease in RR without mental status depression. Methadone: decrease in RR with both doses, but confusion occurred with 10 mg dose. Weaned following dexamethasone administration. On relapse, 5 mg orally every 8 h was restarted, with eventual decrease to oral 2.5 mg at bedtime. Dexamethasone: weaned to 1 mg orally daily prior to relapse. At relapse, stopped after resolution of dyspnea	Neurologically stable at 27 mo on oral methadone 2.5 mg at bedtime	MRI at 4 and 11 mo showed reduction in enhancement and abnormal T2 signal.
Tobias and Heideman, 19919	11 M	Brainstem (pontine) glioma	Hyperfractionated radiation therapy. Dexamethasone (4 mg every 12 h). Morphine (unspecified dose and route). Fentanyl infusion (initial dose: 0.5 µg/kg/h IV, titrated by 0.5 µg/kg/h for RR 20-30) followed by sufentanil infusion (IV, same target)	Morphine: Excessive somnolence with decrease in respiratory rate for brief period (30-60 min). Fentanyl: reversal of dyspnea, patient discharged home with subsequent increases in fentanyl dose (up to 37 µg/kg/h). Fentanyl changed for sufentanil (up to 30 µg/kg/h) when IV volumes became too important	Steady increase of opioid infusions up until 2 mo, after which patient died	Not reported
Dubaybo et al, 199110	55 M	Medullar compression by invasive laryngeal carcinoma	Dexamethasone (IV, unspecified dose). Radiation therapy to base of brain (4400 rads over 3 wks)	Normalization of RR, reduction in pH both after dexamethasone and radiation therapy	Not reported	Not reported
Krendel et al, 199111	43 M	CNS lymphoma	Chemotherapy (systemic and intraventricular, no details specified). Cranial radiation therapy	Gradual decrease in RR and pH with increase in pCO2 over 2-3 wks	Not reported	Not reported
Karp and Nahum, 199212	57 M	Lymphomatous meningitis	Methotrexate (biweekly, intrathecal)	Complete resolution of hyperventilation at 2 wks	Developed infectious meningitis and died 2 mo later	Not reported
Siderowf et al, 199613	57 M	Brainstem (pontine) glioma	Morphine (unspecified dose and route). Steroids (unspecified high-dose, IV). Palliative radiotherapy	No response.	Died 37 d after admission	Not reported
Chang et al, 200014	40 F	CNS lymphoma	Flunitrazepam (1 mg orally 2×/d). Alprazolam (1 mg orally 3×/d). Morphine (5 mg IV every 6 h). Midazolam (3 µg/kg/min). Propofol (15 µg/kg/min). Chemotherapy (carmustine, dexamethasone, methotrexate, cytosine arabinoside)	Flunitrazepam and alprazolam: No effect. Morphine, midazolam, and propofol: decrease in RR while patient remained arousable. Hyperventilation gradually resolved within 1 wk.	Not reported.	Not reported.
Sakamoto et al, 200115	72 M	Malignant brainstem (pontine and medullary) lymphoma	Rebreathing from paper bag. Diazepam (unspecified dose, IV). Oxygen inhalation	Failure to alter the respiratory pattern	Not reported	Not reported
Shahar et al, 200416	7 M	Glioblastoma multiforme with suspected brainstem infiltration	Methylprednisolone (dose unspecified, IV). Midazolam (dose and route unspecified)	No apparent effect on RR and dyspnea.	Lapsed into coma after 2 mo and died after 3 mo	Not reported
Toyooka et al, 200417	40 M	Brainstem dissemination of pineal glioblastoma	Rebreathing in paper bag. Diazepam (dose unspecified, IV)	Failure to alter respiratory pattern	Placed under mechanical ventilation and died of pneumonia 1 mo later	Not reported
Gaviani et al, 200518	25 F	Brainstem (pontine and medullary) glioma	Morphine sulfate (1 mg IV then 15 mg orally 2×/d)	Normalization of RR.	Four months, no relapse.	MRI stability at three months.
Laigle-Donadey et al, 200519	54 M	B-cell brain (including medullary) lymphoma	Morphine derivative (Moscontin 40 mg orally/d). Corticosteroids (molecule unspecified, 1 mg/kg orally/d). Chemotherapy (IV and intrathecal: methotrexate, lomustine, procarbazine)	Morphine and corticosteroids: no effect. Chemotherapy: normalization of RR, pH, and pCO2 at 6 wks	8 y, no relapse	Complete MRI remission 6 wks after chemotherapy induction
Tarulli et al, 200520	87 M	B-cell brain (including pontine) lymphoma	Morphine (dose unspecified, IV 2×/d). Methylprednisolone (1 g/d IV for 5 d)	Morphine: failure to reduce RR. Methylprednisolone: decrease in RR, increase in pCO2	Transferred to hospice care after diagnosis	Not reported
Schmid et al, 200521	75 M	CNS infiltration of transformed chronic lymphocytic leukemia	Midazolam (5 mg IV). Chemotherapy (cytosine-arabinoside, methotrexate, dexamethasone, intrathecal). Corticosteroids (prednisone, IV)	Midazolam: immediate reduction in RR but with transient effect and repeated administration necessary. Chemotherapy and corticosteroids: normalization of RR at 4 d.	Despite respiratory improvement, rapid neurological decline leading to death within several weeks.	Not reported
Adachi et al, 200722	51 F	Malignant brain lymphoma	Rebreathing from paper bag. Diazepam (unspecified dose, IV). Fentanyl (50 µg/h IV then 50 µg/h transdermic [Durotep patch 5 mg = 50 µg/h])	Paper bag and diazepam: no increase in pCO2. Fentanyl: reversal of tachypnea, increase in pCO2, improved level of consciousness with both IV and transdermic fentanyl; patient discharged home	Arterial blood gases remained normal at 1 mo. Lymphoma then progressed and patient died on d 58.	Multiple lymphoma infiltrations became apparent on MRI on d 28.
Blüher et al, 200723	3 M	Malignant meningeal melanoma	Chemotherapy (ifosfamide, carboplatin, etoposide)	Death	Died within the year from cardiovascular collapse	Not reported
Enam and Ali, 201124	47 M	Primary brainstem B-cell lymphoma	Midazolam (dose and route unspecified). Chemotherapy (methotrexate, vincristine, dexamethasone)	Immediate but transient reduction in RR.	Developed septic shock and died 2 mo later	Repeat MRI after 4 cycles of chemotherapy showed 50% reduction in size of lymphoma
Van Wamelen et al, 201125	29 M	Brainstem (pontine and medullary) glioma	Dexamethasone (4 mg every 6 h, route unspecified). Cranial radiotherapy	Normalization of RR, pH, and pCO2 after radiotherapy	Symptom-free at 3 mo	Repeat brain MRI showed reduction of pontine tumor mass without significant gadolinium uptake
Carvalho et al, 201126	22 M	Bilateral acoustic nerve schwannomas	Surgical resection of left tumor	Dramatic clinical and laboratory improvement	Not reported	Not reported.
Pantelyat et al, 201427	69 M	Diffuse large B-cell lymphoma	Methylprednisolone (dose unspecified, IV)	No clinical improvement	Transferred to hospice and died 5 d later	Not reported
Ledet et al, 201428	5 F 10 M 16 F	Diffuse intrinsic pontine glioma	Corticosteroids (dose and route unspecified). Opioids (dose and route unspecified). Radiotherapy	Improvement with corticosteroids for 1 patient and during radiotherapy for another. No response to opioids	The patient who failed to respond died of tumor progression within 1 wk. No follow-up reported for other 2 patients	Not reported
Karastaneva et al, 201729	4 M	Acute lymphoblastic leukemia	Chemotherapy (Berlin-Frankfurt-Munster 2000 treatment protocol). Craniospinal irradiation	Complete resolution of symptoms after remission	Not reported	Not reported
Sweidan et al, 201730	46 M	Diffuse large B-cell lymphoma	Fentanyl, propofol, and midazolam (doses and routes not specified). High-dose steroids (dose and route unspecified). Whole-brain radiation. Rituximab (8 weekly doses). Mechanical ventilation	Fentanyl, propofol, midazolam: failure to decrease hyperventilation. Steroids, radiotherapy, and chemotherapy: resolution of hyperventilation after weaning from ventilator. Neurological improvement 2 wks later.	Clinically stable 3 wks after extubation	Not reported
Murata et al, 201931	60 M	Diffuse large B-cell brain (including pontine) lymphoma	Methylprednisolone (1 g IV × 3 d)	Partially effective.	Not reported	Progression of left frontal lobe tumor-like lesion on MRI on d 31
Neves Briard 2020 (present case)	39 F	Bifrontal glioma with probable brainstem (pontine and medullary) invasion	Rebreathing in paper bag. Lorazepam (1 mg orally before bed). Dexamethasone (2 mg every 6 h IV then orally). Hydromorphone (0.25 mg every 4 h SC then 0.5 mg every 3-4 h orally). Brainstem radiotherapy	Paper bag and lorazepam: very brief reversal of hyperventilation. Dexamethasone and hydromorphone: reversal of dyspnea and decrease in RR without mental status depression; patient discharged home.	Clinically stable at 2-mo follow-up without dexamethasone or hydromorphone	Repeat MRI 1 mo after brainstem radiotherapy showed decreased hypersignal and edema in pontine and medullary hypersignal but progression in supratentorial matter

Abbreviations: F, female; IV, intravenous; M, male, RR, respiratory rate; pCO₂, partial pressure of carbon dioxide; SC, subcutaneous.

Reported treatment and clinical outcomes are respectively supplied in Tables 2 and 3, stratified for neoplasm pathology (PCNSL, glioma, and other neoplasm). Overall, 70% of patients were relieved of CNH through their respective treatment regimens. Oxygen supplementation (0/4), rebreathing from a paper bag (0/5), and diazepam (0/4) were never effective. Morphine was frequently tested, with variable outcomes (5/11). Methadone (1/1) and fentanyl (2/3) were successful, but rarely used. Propofol (1/2) and midazolam (3/6) were moderately successful, despite frequent association with mechanical ventilation. Chemotherapy was most effective in patients with PCNSL (8/9), but not glioma or other neoplasms (0/2). Patients with PCNSL (4/5) and other tumors (2/2) also responded to radiotherapy more frequently than patients with glioma (3/7). Corticosteroids were moderately effective, regardless of pathology type (4/8 in PCNSL, 3/10 in glioma, 1/1 in other). CNH resolved after all cases (3/3) for which subtotal surgical resection was performed.

Table 2.

Treatment Outcomes

Treatment	Patients, No.	Success, n	Failure, n	Uncertain outcome, n	Mitigated outcome, n	Success rate, n/No. (%)
Primary CNS lymphoma (15 cases)
Oxygen supplementation	1	0	1	0	0	0/1 (0)
Rebreathing in paper bag	3	0	3	0	0	0/3 (0)
Morphine	6	2	3	0	1	2/6 (33)
Fentanyl	2	1	1	0	0	1/2 (50)
Midazolam	4	3	1	0	0	3/4 (75)
Diazepam	3	0	3	0	0	0/3 (0)
Other benzodiazepine	1	0	1	0	0	0/1 (0)
Propofol	2	1	1	0	0	1/2 (50)
Corticosteroids	8	4	2	0	2	4/8 (50)
Radiotherapy	5	4	0	0	1	4/5 (80)
Chemotherapy	9	8	0	1	0	8/9 (89)
Glioma (13 cases)
Oxygen supplementation	2	0	2	0	0	0/2 (0)
Rebreathing in paper bag	2	0	1	0	1	0/2 (0)
Morphine	5	3	1	0	1	3/5 (60)
Hydromorphone	1	1	0	0	0	1/1 (100)
Fentanyl	1	1	0	0	0	1/1 (100)
Methadone	1	1	0	0	0	1/1 (100)
Midazolam	2	0	2	0	0	0/2 (0)
Diazepam	1	0	1	0	0	0/1 (0)
Other benzodiazepine	1	0	0	0	1	0/1 (0)
Corticosteroids	10	3	4	2	0	3/10 (33)
Radiotherapy	7	3	3	1	0	3/7 (43)
Chemotherapy	1	0	0	1	0	0/1 (0)
Surgical resection	1	0	0	0	0	1/1 (100)
Other neoplasm (5 cases)
Oxygen supplementation	1	0	1	0	0	0/1 (0)
Corticosteroids	1	1	0	0	0	1/1 (100)
Radiotherapy	2	2	0	0	0	2/2 (100)
Chemotherapy	1	0	1	0	0	0/1 (0)
Surgical resection	2	2	0	0	0	2/2 (100)

Table 3.

Clinical Outcomes

Neoplasm	Cases, No.	CNH abortion (n, %)	Clinical outcome
			Alive (n, %)	Follow-up duration in wks (mean, range)	Dead (n, %)	Death due to neoplasm, n	Death due to other cause, n	Timing of death in wks (mean, range)	Unknown clinical outcome (n, %)
PCNSL	15	11 (73)	3 (20)	141 (5-416)	7 (47)	5	2	7 (1-14)	5 (33)
Glioma	13	9 (69)	5 (38)	31 (8-108)	6 (46)	5	1	5 (1-12)	2 (15)
Other	5	3 (60)	1 (20)	104 (104-104)	2 (40)	2	0	6 (6-6)	2 (40)

Abbreviations: CNH, central neurogenic hyperventilation; PCNSL, primary CNS lymphoma.

Reported long-term outcomes did not disclose any cases in which patients died as a direct consequence of CNH. The majority of reported deaths were secondary to progression of the underlying neoplasm (Table 3). Overall, the reports did not provide radiological outcomes in 70% or clinical long-term outcomes in 27% of cases.

Discussion

First described in 1959 by Plum and Swanson, CNH is a rare cause of dyspnea, hypocapnea, and respiratory alkalosis, particularly in conscious patients.¹ By definition, the patient must demonstrate paraclinical evidence of hyperventilation (low PaCO₂) without evidence of an underlying cardiopulmonary or toxicometabolic cause. Hyperventilation must occur both in awake and asleep states, thereby excluding psychogenic causes. Respiratory alkalosis induces metabolic disturbances and can lead to cerebral vasoconstriction and hypoperfusion, tetany, electrocardiographic changes, and psychomotor symptoms such as agitation, decreased concentration, and irritability.³²

The case here reported shares similar features with those previously published in the literature. Gliomas involving the pons and the medulla have consistently been described²⁰; however, a significant and disproportionate quantity of reported cases involve primary CNS lymphoma.³¹ The physiopathology of CNH is debated; all reports describe slowly infiltrative lesions that primarily involve the brainstem. The original hypothesis by Plum and Swanson was that CNH resulted from a disconnection between pontine and medullary respiratory centers, a hypothesis that animal experiments failed to confirm.^1,33 Other authors suggested that brainstem chemoreceptors could be activated following acidification of the local CSF or brain tissue by tumor-secreted lactate.²⁰ These changes would affect only the tumor microenvironment, resulting in an unremarkable CSF analysis. However, elevated CSF lactate has also been observed in healthy dogs hyperventilated for several hours, suggesting that hyperventilation could in fact cause cerebral lactate production by hypocapnea-induced vasoconstriction and ischemia. This alternative hypothesis, which may apply only for cases of acute hyperventilation, has not been further studied.³⁴ Finally, in animals, tachypnea was induced by stimulating the lateral parabrachial nucleus, a finding that is consistent with the confirmed pathological localization of the brainstem lesion in multiple reported cases.^2,3,20 CNH has also been reported in conscious patients with supratentorial subcortical infarcts,³⁵ hemorrhage,³⁶ and demyelinating lesions.³⁷ Considering the wide range of lesion localizations, the heterogeneity in response to therapy, and the complexity of the central respiratory network, there may be multiple, distinct mechanisms involved in the pathophysiology of CNH.³³ In our patient, exact pathological localization could not be confirmed, but glioma infiltration of the pons and medulla was supported by radiological evidence of a bilateral brainstem infiltration and progression of the disease in supratentorial matter, as well as a clinical remission following treatment with corticosteroids and brainstem radiotherapy.

To our knowledge, this review is the first to review and synthesize the currently available data on management of CNH in conscious patients with CNS neoplasm. As suspected, addressing the underlying lesion seemed the most effective therapy. For instance, the majority of PCNLS cases responded to chemotherapy (89%) and radiotherapy (80%), whereas subtotal surgical resection aborted hyperventilation in all appropriate cases. However, many cases are not amenable to aggressive treatment because of the nature or localization of the culprit lesion, or because of the general status or wishes of the patient, stressing the importance of reliable palliative treatment. In this review, 45% of the patients ultimately died, the vast majority because of progression of the underlying neoplasm. Despite this, CNH was successfully aborted in 70% cases, relieving a strong majority of patients from their discomfort without significant secondary effects, such as a decreased level of consciousness. In our review, opioids and sedatives (eg, morphine, hydromorphone, fentanyl, midazolam) overall seemed useful for symptom relief, but the benefit was often of short duration when the medication was administered orally or subcutaneously. Transdermal application of fentanyl overcame this limitation in the only case in which its use was reported.²² Oral methadone also appeared effective and durable.⁸ Our findings suggest that certain therapeutic avenues are less likely to be useful, such as oxygen supplementation, rebreathing into a paper bag, and benzodiazepines.

Owing to the rarity of CNH in conscious patients with CNS neoplasm, this review included only observational case reports and one case series. Large variability in the quality and quantity of information contained in these reports may limit the validity of our findings. Most studies had information on the initial response to treatment, but not all reported time to improvement or duration of effect. It is particularly important to stress that several treatments were often administered concurrently, with clinicians judging the success or failure of respective agents according to the time at which patients improved or declined. A publication bias may also be present, overestimating the benefit of reported treatments. Moreover, restriction of case selection by language restrains the scope of our review. These factors limit our capacity of inference. However, because CNH is rare in the studied population, it is unlikely that high-quality interventional studies will be possible. Therefore, despite these shortcomings, our review provides the most comprehensive data currently available to help clinicians in their approach to CNH in conscious patients with CNS neoplasm.

Conclusion

Treatment of CNH is challenging but successful in a majority of cases. Definitive treatment of the underlying neoplasm may be more successful in aborting hyperventilation. Variable rates of palliation have been observed with opioids and sedatives. Further research is warranted to evaluate the potential role of promising palliative avenues, such as transdermal fentanyl and oral methadone.

Conflict of interest statement.

None declared.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.