Insufficiency of plasmatic arginine/homoarginine during the initial postoperative phase among patients with tumors affecting the medulla oblongata heightens the likelihood of neurogenic pulmonary oedema following surgery

Abstract

Background:

This prospective clinical study aims to investigate the fluctuations of neurotransmitters in peripheral venous blood during the perioperative period and to identify independent predictors for postoperative neurogenic pulmonary oedema (NPE) in patients with medulla oblongata-involved tumours.

Materials and methods:

Peripheral venous blood samples of the enroled patients at seven perioperative time points, as well as their medical records and radiologic data were collected. High-performance liquid chromatography-tandem mass spectrometry was utilized to detect the concentrations of 39 neurotransmitters in these samples. The study applied univariate and multivariate generalized estimating equation (GEE) logistic regression analyses to explore independent predictors of postoperative NPE, and one-way repeated-measures ANOVA to compare the concentrations of the same neurotransmitter at different perioperative time points.

Results:

The study included 36 patients with medulla oblongata-involved tumours from January to December 2019, and found that 13.9% of them experienced postoperative NPE. The absence of intraoperative use of sevoflurane (P=0.008), decreased concentrations of arginine (P=0.026) and homoarginine (P=0.030), and prolonged postoperative tracheal extubation (P<0.001) were identified as independent risk factors for postoperative NPE in medulla oblongata-involved tumour patients. Pairwise comparison analysis revealed that the perioperative decreases in arginine and homoarginine concentrations mainly occurred within the postoperative 8 h.

Conclusion:

This study demonstrates that NPE is not uncommon in patients with medulla oblongata-involved tumours. The absence of intraoperative use of sevoflurane, decreased concentrations of plasmatic arginine and homoarginine, and prolonged postoperative tracheal extubation are independent predictors of postoperative NPE. These two neurotransmitters’ concentrations dropped mainly within the early postoperative hours and could serve as potential early warning indicators of postoperative NPE in clinical practice.

Introduction

Highlights

• Decreased plasmatic arginine and homoarginine within the early postoperative hours could serve as early warning indicators of postoperative neurogenic pulmonary oedema in patients with medulla oblongata-involved tumours.

• Intraoperative use of sevoflurane and prompt tracheal extubation may help reduce the risk of postoperative neurogenic pulmonary oedema in patients with medulla oblongata-involved tumours.

Neurogenic pulmonary oedema (NPE) is a grave complication that manifests acutely and dramatically following significant central nervous system (CNS) diseases, including brain tumours, ischaemic stroke, intracranial hemorrhage, status epilepticus, meningitis, closed head injury, spinal cord or brain injuries, etc.^1,2. It may appear within minutes to hours or even several days following the initial triggering event³ and has been reported to occur in up to 50% of patients with severe CNS injuries, with a mortality rate of ~9%⁴. However, the treatments of NPE have not improved significantly and the prognosis of it has remained unsatisfactory for the past few decades¹. Therefore, it is urgent to seek new methods that could effectively predict the occurrence of NPE, prevent its development and/or improve its prognosis.

A review of CNS triggers of NPE from 2012 to 2019 revealed that only one case of NPE was caused by posterior cranial fossa surgery and two cases by cerebral mass lesions (one meningioma and one medullar hemangioblastoma)⁵. No relevant systematic, prospective studies were available⁵. Partly due to its relatively unpredictable nature and a lack of aetiology-specific diagnostic markers, epidemiological data on NPE are still scarce and its morbidity and mortality are probably underestimated⁴. In our clinical practice of skull base surgery, NPE is not uncommon, particularly when lesions involve the medulla oblongata.

The medulla oblongata is recognized as one of the “NPE trigger zones”. Prior animal studies have demonstrated that neurons in the following areas of medulla oblongata are responsible for respiratory function and normally use glutamate, GABA, and/or glycine as their major neurotransmitters: (1) area A1 (located in the ventrolateral aspect of the medulla and was composed of catecholamine neurons that project into the hypothalamus)³; (2) area A5 (located in the upper portion of the medulla and project into the preganglionic centres for spinal cord sympathetic outflow)³; (3) the nuclei of the solitary tract and (4) the area postrema (related to respiratory regulation and receiving input from the carotid sinus)³.

Variations in neurotransmitter levels are a characteristic manifestation of various neurological diseases. This study postulates that these variations in neurotransmitter levels may also be detected in NPE patients triggered by medulla oblongata-involved tumour resection. High-performance liquid chromatography (HPLC)-fluorescence quantitative detection of amino acid-like neurotransmitters in plasma makes it feasible to test and validate this hypothesis. Therefore, it is essential to conduct a prospective clinical study on the incidence of postoperative NPE in patients with medulla oblongata-involved tumours to investigate the fluctuations in different neurotransmitters in peripheral blood at different perioperative time points and to explore neurotransmitters that can serve as independent predictors of postoperative NPE to provide a basis for future clinical interventions.

Materials and methods

Ethical approval

All procedures performed in studies involving human participants were in accordance with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards and were reviewed and approved by the ethics committee of our hospital (KY2019-107-02).

Inclusion/exclusion/elimination criteria of this prospective cohort study

The inclusion criteria comprised of the following

1. Consecutive patients who underwent surgeries at the Neurosurgical Skull Base Tumour Ward of our hospital from January 2019 to December 2019;

2. Patients aged between 14 AND 80 years;

3. Preoperative MRI indicating medulla oblongata-originated tumours or perimedulla-originated tumours directly compressing the medulla oblongata;

4. Voluntary informed consent signed by the patient and their family members.

The exclusion criteria included the following

1. Patients under the age of 14 or above the age of 80;

2. Pregnant or breastfeeding women;

3. Preoperative MRI indicating tumours not originating from or around the medulla oblongata, and perimedulla-originated tumours without medulla oblongata compression;

4. Patients with severe primary tumours in other organs or systems (e.g. liver, kidney, lung, etc.), infectious diseases (e.g. hepatitis B, syphilis, AIDS, etc.), or illnesses that affect their survival;

5. Patients with a medication history that may influence the test results;

6. Patients who were unable to cooperate during the tests due to their illness or other reasons, or with mental or legal disabilities;

7. Patients or their families unwilling to sign the informed consent form.

The elimination criteria comprised of the following

1. Patients who subjectively withdrew their informed consent form;

2. Patients with severe life-threatening complications other than NPE during the perioperative period (e.g. cerebral hernia/infarction/hemorrhage, multiple organ failure, etc.).

Data collection

The medical records and radiologic data of all enroled patients were collected: medulla oblongata involvement, types of tumours, gender, age, history of hypertension, history of hydrocephalus, intraoperative use of sevoflurane, intraoperative use of muscular relaxants, intraoperative blood loss, intraoperative blood transfusion, surgical duration, duration of postoperative tracheal intubation, postoperative 2nd tracheal intubation, duration of postoperative mechanical ventilation, postoperative tracheotomy, postoperative pulmonary infection, postoperative neurogenic pulmonary oedema (Table 1). The differentiation between tumour origins (whether they originated from the medulla oblongata or perimedulla with direct compression on the medulla oblongata) was based on intraoperative observations.

Table 1.

Baseline characteristics of total 36 medulla oblongata-involved tumour patients.

Variable	Total	Perimedulla-originated (medulla oblongata-compressed)	Medulla oblongata-originated	P
Medulla oblongata involvement, n (%)	36 (100.0)	19 (52.8)	17 (47.2)
Types of tumours, n (%)
Schwannoma	5 (13.9)	5 (26.3)	0 (0.0)
Transitional (mixed) meningioma	8 (22.2)	8 (42.1)	0 (0.0)
Meiningothelial meningioma	2 (5.6)	2 (10.5)	0 (0.0)
Fibrous meningioma	1 (2.8)	1 (5.3)	0 (0.0)
Secretory meningioma	1 (2.8)	1 (5.3)	0 (0.0)
Atypical meningioma	1 (2.8)	1 (5.3)	0 (0.0)
Hemangioblastoma	2 (5.6)	1 (5.3)	1 (5.9)
Cavernous malformation	5 (13.9)	0 (0.0)	5 (29.4)
Astrocytoma	5 (13.9)	0 (0.0)	5 (29.4)
Ependymoma	3 (8.3)	0 (0.0)	3 (17.6)
Diffuse midline glioma	2 (5.6)	0 (0.0)	2 (11.8)
Pilomyxoidastrocytoma	1 (2.8)	0 (0.0)	1 (5.9)
Sex, n (%)				0.054b
Male	19 (52.8)	7 (36.8)	12 (70.6)
Female	17 (47.2)	12 (63.2)	5 (29.4)
Age (year)
Median (range)	44.67 (13–75)	49.33 (29–75)	38.00 (13–60)
Mean±SD	43.472±14.458	49.790±11.830	36.410±14.513	0.004 a,*
History of hypertension, n (%)				0.605b
Absent	32 (88.9)	16 (84.2)	16 (94.1)
Present	4 (11.1)	3 (15.8)	1 (5.9)
History of hydrocephalus, n (%)				0.605b
Absent	32 (88.9)	16 (84.2)	16 (94.1)
Present	4 (11.1)	3 (15.8)	1 (5.9)
Intraoperative use of Sevoflurane, n (%)				0.219b
Absent	7 (19.4)	2 (10.5)	5 (29.4)
Present	29 (80.6)	17 (89.5)	12 (70.6)
Intraoperative use of muscular relaxants, n (%)				0.749b
Absent	20 (55.6)	10 (52.6)	10 (58.8)
Present	16 (44.4)	9 (47.4)	7 (41.2)
Intraoperative blood loss (ml)
Median (range)	350.00 (50–1700)	680.00 (200–1700)	183.33 (50–500)
Mean±SD	484.444±410.078	728.420±432.836	211.760±128.123	0.000
Intraoperative blood transfusion, n (%)				0.008
Absent	26 (72.2)	10 (52.6)	16 (94.1)
Present	10 (27.8)	9 (47.4)	1 (5.9)
Surgical duration (h)
Median (range)	6.700 (4.0–16.4)	7.80 (5.3–16.4)	5.300 (4.0–8.6)
Mean±SD	7.278±2.813	8.713±3.073	5.669±1.371	0.001
Duration of postoperative tracheal intubation (day)
Median (range)	5.00 (0–27)	4.00 (0–22)	6.00 (0–27)
Mean±SD	1.889±5.074	6.000±7.008	7.650±7.280	0.494a
Postoperative 2nd tracheal intubation, n (%)				1.000b
Absent	32 (88.9)	17 (89.5)	15 (88.2)
Present	4 (11.1)	2 (10.5)	2 (11.8)
Duration of postoperative mechanical ventilation (day)
Median (range)	0.32 (0–18)	0.33 (0–12)	1.14 (0–18)
Mean±SD	1.917±4.201	1.530±3.221	2.350±5.243	0.568a
Postoperative tracheotomy, n (%)				0.434b
Absent	28 (77.8)	16 (84.2)	12 (70.6)
Present	8 (22.2)	3 (15.8)	5 (29.4)
Postoperative pulmonary infection, n (%)				0.721b
Absent	25 (69.4)	14 (73.7)	11 (64.7)
Present	11 (30.6)	5 (26.3)	6 (35.3)
Postoperative neurogenic pulmonary oedema, n (%)				0.167b
Absent	31 (86.1)	18 (94.7)	13 (76.5)
Present	5 (13.9)	1 (5.3)	4 (23.5)

^aIndependent-sample t-test.

^bFisher exact test.

^*Indicates statistical significance, and relevant P values are emphasized in bold.

For each patient, two tubes of peripheral blood samples were drawn from the median cubital vein at seven distinct perioperative time points (T1, prior to anaesthesia induction; T2, during tumour resection; T3, during dural suturing; T4, 8 h after surgery; T5, 24 h after surgery; T6, 72 h after surgery; and T7, 7 days after surgery). One tube was used for amino acid neurotransmitter analysis (collected using a vacuum negative pressure blood collection vessel containing EDTA) and the other was used for catecholamine testing (collected using an inert separation gel procoagulant tube; centrifuged at a speed of 3000 RPM and a temperature of 4°C for 3 min). The blood volume of each tube was more than 1.5 ml. All the samples were stored away from light at −80°C.

The concentrations of various neurotransmitters were determined using high-performance liquid chromatography (HPLC)-tandem mass spectrometry^6,7. The neurotransmitters analyzed included 32 amino acids (3-aminoisobutyric acid, 3-methylhistidine, 6-aminocaproic acid, β-alanine, γ-aminobutyric acid, aminobutyric acid, aminoadipic acid, phenylalanine, alanine, proline, glycine, glutamate, citrulline, sarcosine, methionine, arginine, lysine, tyrosine, leucine, ornithine, hydroxyproline, kynurenine, tryptophan, serine, threonine, aspartic acid, asparagine, homoproline, valine, isoleucine, histidine, and histidine) as well as 7 catecholamines (3-metanephrine, dopamine, homovanillic acid, 3-normetanephrine, norepinephrine, adrenaline, and vanillylmandelic acid). The quantitative analysis of all the samples was performed by Zhejiang Baichen Medical Technology Co., Ltd (www.biozonn.com).

Diagnostic criteria for NPE

The diagnostic criteria for NPE were established through a comprehensive evaluation of clinical presentation, imaging features, and exclusion of various differential diagnoses⁸, under the independent diagnosis of two senior physicians from the intensive care unit (both were blinded to the patients’ perioperative neurotransmitter levels). The criteria encompassed: (1) bilateral pulmonary oedema detected by X-ray, followed by rapid resolution of opacities; (2) expectoration of pink, frothy sputum or hemoptysis; (3) PaO2:PiO2 less than 200 mmHg; (4) severe acute CNS compromise leading to increased intracranial pressure; (5) absence of left atrial hypertension on echocardiography; and (6) absence of other common causes of respiratory distress or acute respiratory distress syndrome⁵. The differential diagnoses of NPE include: (1) cardiogenic pulmonary oedema, (2) aspiration pneumonia, (3) pneumonitis, (4) sepsis, (5) negative pressure pulmonary oedema, (6) post-airway obstruction oedema, (7) ventilator-associated pneumonia, (8) ventilation-induced lung injury and (9) transfusion-related lung injury.

Statistical analysis

The baseline patient characteristics are summarized as percentages for categorical variables and as the mean±standard deviation for continuous variables. Since the concentration of neurotransmitters in each included patient was detected at multiple perioperative time points, these data were not statistically independent. Therefore, univariate and multivariate GEE logistic regression analyses were used to assess the correlations between various factors and postoperative NPE⁹. P values and odds ratios (ORs) with 95% CIs were calculated with robust standard error estimated from the GEE approach with the independent working correlation structure. In consideration of rare outcome events, only the variables with P less than 0.05 in the univariate analysis were selected as covariates adjusted for the multivariate GEE logistic analysis. One-way repeated-measures ANOVA with a Bonferroni post hoc test was applied to compare the concentrations of the same neurotransmitter at different perioperative time points. Mauchly’s sphericity test was performed to determine whether a violation of sphericity occurred, and the Greenhouse–Geisser correction (when epsilon (ε)<0.75) or Huyuh-Feldt correction (when epsilon (ε)>0.75) was adopted when the sphericity was violated¹⁰. All P values are two-sided, and significance was defined using a threshold of 0.05. Statistical analyses were performed with SPSS Statistics software (version 23.0; IBM Corporation, Armonk). This work has been reported in line with the STROCSS criteria¹¹, Supplemental Digital Content 1, http://links.lww.com/JS9/B521.

Results

Patient demographics, tumour characteristics, and neurotransmitter concentrations

A total of 36 patients diagnosed with medulla oblongata-involved tumours met the aforementioned criteria, including 17 (47.2%) cases originating from the medulla oblongata and 19 (52.8%) cases originating from the perimedulla, which compressed the medulla oblongata. Apart from the two instances of hemangioblastoma (one originating from the medulla oblongata and the other originating from the cerebellar hemispheres, compressing the medulla oblongata), the tumour types did not overlap between these two groups. The statistical significance between these two groups was observed in several patient demographics, including age (P=0.004), intraoperative blood loss (P<0.001), intraoperative blood transfusion (P=0.008), and surgical duration (P=0.001). No significant differences were found in the intraoperative application ratio of sevoflurane and muscular relaxants, the duration of postoperative tracheal intubation and mechanical ventilation, or the incidence of postoperative tracheotomy and pulmonary infection between these two groups (Table 1). Of the total patients, five (13.9%) experienced NPE that commenced on the first, second, third, fifth, and fifth postoperative days. Although the incidence of postoperative NPE in the medulla oblongata-originated group (4/17, 23.5%) was higher compared to the perimedulla-originated group (1/19, 5.3%), the difference was not statistically significant (P=0.167) (Table 1). The concentrations of thirty-nine neurotransmitters at seven distinct perioperative time points are outlined in Supplementary Table 1, Supplemental Digital Content 2, http://links.lww.com/JS9/B522. The corresponding information on assay validation of HPLC-tandem mass spectrometry is uploaded in supplementary materials (Supplementary materials-Information on assay validation), Supplemental Digital Content 3, http://links.lww.com/JS9/B523.

Univariate and multivariate analyses associated with postoperative NPE of the medulla oblongata-involved tumour patients

The univariate GEE analysis demonstrated that the emergence of postoperative NPE in patients with medulla oblongata-involved tumours was significantly impacted by the use of sevoflurane during surgery (P=0.029), the duration of postoperative tracheal intubation (P=0.021), and the levels of various amino acids including phenylalanine (P=0.004), proline (P=0.048), methionine (P<0.001), arginine (P=0.002), tyrosine (P=0.004), threonine (P=0.006), aspartic acid (P=0.007), homoarginine (P=0.035), valine (P=0.018) and norepinephrine (P=0.027) (Table 2).

Table 2.

Univariate analysis of risk factors for postoperative neurogenic pulmonary oedema.

Variable	OR (95% CI)	P
Male	1.406 (0.206–9.619)	0.728
Age	1.004 (0.940–1.072)	0.903
Without history of hypertension	0.429 (0.035–5.189)	0.505
Without history of hydrocephalus	0.429 (0.035–5.189)	0.505
Medulla oblongata-originated	0.206 (0.021–2.059)	0.179
Without intraoperative use of Sevoflurane	10.125 (1.272–80.607)	0.029 *
Without intraoperative use of muscular relaxants	1.235 (0.180–8.459)	0.830
Intraoperative blood loss	0.998 (0.995–1.002)	0.326
Without intraoperative blood transfusion	1.636 (0.160–16.726)	0.678
Surgical duration	0.928 (0.633–1.360)	0.701
Postoperative tracheal intubation duration	1.201 (1.029–1.403)	0.021 *
Without postoperative 2nd tracheal intubation	0.103 (0.010–1.024)	0.052
Postoperative mechanical ventilation duration	1.005 (0.806–1.254)	0.962
Without postoperative tracheotomy	0.360 (0.049–2.657)	0.316
Without postoperative central nervous system infection	1.821 (0.266–12.473)	0.541
Without postoperative pulmonary infection	0.232 (0.033–1.649)	0.144
3-Aminoisobutyric acid	0.837 (0.536–1.306)	0.432
3-Methylhistidine	0.058 (0.001–3.687)	0.179
6- Aminocaproic acid	1.257 (0.824–1.916)	0.288
β-Alanine	0.971 (0.669–1.409)	0.877
γ-Aminobutyric acid	0.202 (0.006–7.365)	0.384
Aminobutyric acid	0.990 (0.954–1.027)	0.578
Aminoadipic acid	1.820 (0.710–4.663)	0.212
Phenylalanine	1.012 (1.004–1.020)	0.004 *
Alanine	1.001 (0.997–1.004)	0.684
Proline	1.007 (1.000–1.014)	0.048 *
Glycine	1.005 (0.999–1.011)	0.137
Glutamate	1.000 (0.994–1.005)	0.907
Citrulline	1.011 (0.970–1.053)	0.607
Sarcosine	0.569 (0.129–2.513)	0.457
Methionine	1.056 (1.025–1.088)	<0.001 *
Arginine	1.015 (1.005–1.024)	0.002 *
Lysine	1.004 (0.997–1.011)	0.256
Tyrosine	1.016 (1.005–1.027)	0.004 *
Leucine	1.005 (0.994–1.015)	0.361
Ornithine	1.009 (0.993–1.025)	0.255
Hydroxyproline	0.914 (0.729–1.145)	0.434
Kynurenine	0.814 (0.060–11.112)	0.878
Tryptophan	1.014 (0.982–1.048)	0.393
Serine	1.002 (0.995–1.010)	0.568
Threonine	1.009 (1.003–1.015)	0.006 *
Aspartic acid	1.030 (1.008–1.053)	0.007 *
Asparagine	1.014 (0.998–1.030)	0.086
Homoproline	1.040 (0.782–1.384)	0.786
Homoarginine	0.212 (0.050–0.895)	0.035 *
Valine	1.006 (1.001–1.010)	0.018 *
Isoleucine	1.008 (0.984–1.032)	0.507
Histidine	0.998 (0.981–1.017)	0.864
3-Metanephrine	0.228 (0.008–6.121)	0.379
Dopamine	0.916 (0.801–1.049)	0.205
Homovanillic acid	0.990 (0.965–1.015)	0.409
3-Normetanephrine	0.165 (0.018–1.544)	0.114
Norepinephrine	0.454 (0.225–0.915)	0.027 *
Adrenaline	0.111 (0.000–45.222)	0.474
Vanillylmandelic acid	1.005 (0.991–1.018)	0.493

OR, odds ratio.

^*Indicates statistical significance, and relevant P values are emphasized in bold.

To eliminate the statistical impact of exogenous norepinephrine, which was administered to patients with intraoperative hypotension, norepinephrine was not included in the multivariate analysis. The multivariate GEE analysis revealed that the absence of intraoperative use of sevoflurane (P=0.008), prolonged postoperative tracheal extubation (P<0.001) and decreased concentrations of arginine (P=0.026) and homoarginine (P=0.030) were independent risk factors for postoperative NPE in the enroled patients with medulla oblongata-involved tumours (Table 3).

Table 3.

Multivariate analysis of risk factors for postoperative neurogenic pulmonary oedema.

Variable	β	OR (95% CI)	P
Constant	2.108	8.236 (0.159–4.261×102)	0.295
Measurement time points
Before anaesthesia	−31.387	2.338×10−14 (0.000–0.000)	—
During resection	−31.486	2.117×10−14 (0.000–0.000)	—
After resection	−30.234	7.403×10-14 (0.000–0.000)	—
Postoperation 8 h	−34.871	7.171×10−14 (0.000–0.000)	—
Postoperation 24 h	−6.255	0.002 (6.686×10-5–0.055)	<0.001 *
Postoperation 72 h	−4.973	0.007 (0.001–0.050)	<0.001 *
Postoperation 7 days	—	1	—
Intraoperative use of Sevoflurane
Absent	5.307	201.845 (3.909–1.042×104)	0.008 *
Present	—	1	—
Postoperative tracheal intubation duration	0.200	1.222 (1.140–1.310)	<0.001 *
Phenylalanine	−0.014	0.986 (0.925–1.052)	0.676
Proline	0.021	1.021 (0.989–1.053)	0.201
Methionine	0.125	1.133 (0.944–1.360)	0.180
Arginine	−0.055	0.946 (0.901–0.993)	0.026 *
Tyrosine	−0.055	0.946 (0.854–1.048)	0.290
Threonine	0.000	1.000 (0.971–1.030)	0.998
Aspartic acid	0.014	1.014 (0.905–1.137)	0.810
Homoarginine	−5.337	0.005 (0.000–0.595)	0.030 *
Valine	0.004	1.004 (0.991–1.017)	0.540

QIC=50.131.

OR, odds ratio.

^*Indicates statistical significance, and relevant P values are emphasized in bold.

Significant reduction of plasmatic arginine and homoarginine was observed in the early postoperative hours

Multiple measurements of the concentrations of arginine and homoarginine at 7 different perioperative time points were further applied to determine the period during which their levels significantly decreased. Mauchly’s sphericity test revealed that the assumption of sphericity was violated for the concentrations of both arginine and homoarginine in the peripheral blood (arginine: P=0.032; homoarginine: P<0.001). The Huyuh-Feldt and Greenhouse–Geisser corrections were applied to arginine (ε=0.904) and homoarginine (ε=0.428), respectively. The corrected results exhibited significant differences in the concentrations of plasmatic arginine and homoarginine at different perioperative time points [arginine: F (5.425, 189.861)=38.569, P<0.001; homoarginine: F (2.569, 89.911)=17.951, P<0.001] (Supplementary Table 2, Supplemental Digital Content 2, http://links.lww.com/JS9/B522).

Pairwise comparison analysis showed that (1) there were no significant changes in the concentrations of arginine and homoarginine in the peripheral blood during the operations (T1–T3), while (2) significant decreases in the concentrations of both occurred within 8 hours postoperatively (T3–T4). Subsequently, (3) the concentration of arginine rebounded significantly during each measurement interval within 72 hours postoperatively (T4–T6) and resulted in a significantly higher concentration compared with its preoperative measurement (T1 vs. T7), while (4) the concentration of homoarginine decreased gradually, with only the comparison between T4 and T7 remaining statistically significant (Supplementary Table 3, Supplemental Digital Content 2, http://links.lww.com/JS9/B522; Fig. 1). These results indicated that the perioperative decreases in the arginine and homoarginine concentrations in the peripheral blood mainly occurred within 8 h after surgery.

Figure 1.

Fluctuations in the arginine (A) and homoarginine (B) concentrations in the peripheral blood during the perioperative period.

Pairwise comparisons of the concentrations of arginine and homoarginine between patients, either subgrouped by the occurrence of postoperative NPE (without NPE vs. with NPE) or by tumour origin (perimedulla-originated vs. medulla-originated), at different perioperative time points showed no statistical significance (Supplementary Table 4, Supplemental Digital Content 2, http://links.lww.com/JS9/B522).

Discussion

The incidence rate of postoperative NPE in the patients with medulla oblongata-involved tumours

To the best of our knowledge, this is the first prospective clinical study that documents the incidence and risk factors for postoperative NPE in patients with medulla oblongata-involved tumours. In line with our clinical observation, the morbidity of postoperative NPE in our enroled patients (5/36, 13.9%) was relatively higher than that reported previously. This can be mainly attributed to the tumour location in all enroled patients. As previously mentioned, when tumours originate or compress the medulla oblongata, which is an anatomical region that houses several NPE trigger zones and the vagus nerve^12,13, special attention must be paid to NPE throughout the perioperative period.

Although no statistically significant difference in the incidence rate of NPE was observed between the Perimedulla-originated group and the Medulla oblongata-originated group in this study, the latter group exhibited a significantly higher NPE incidence rate (4/17, 23.5%) compared to the former group (1/19, 5.3%). This underscores the crucial role of early diagnosis, intervention, and even prevention of NPE in patients with medulla oblongata-originated tumours. Other variables that showed significant differences between the two groups, such as age, intraoperative blood loss, blood transfusion, and surgical duration, can be explained by the differing tumour characteristics and surgical complexities.

The role of nitric oxide (NO) in NPE’s generation

Despite numerous theories proposed to explain the generation of NPE, its precise pathophysiological mechanisms remain incompletely understood⁵. The most commonly endorsed theory suggests that sympathetic overstimulation (catecholamine storm) leads to widespread vasoconstriction⁵. Surges of sympathetic activity, particularly through increased alpha-adrenergic activity induced by hypothalamic or medullary lesions, are believed to result in elevation of systemic hypertension, an increase in left atrial and pulmonary hydrostatic pressure, and culminate in pulmonary oedema.

The lungs are recognized as the primary site of NO production in the circulatory system¹². Nitric oxide synthases (NOS), a group of enzymes, produce NO by converting L-arginine into NO and L-citruline¹⁴. Three NOS isoforms have been identified, including two constitutive forms: neuronal NOS (nNOS) and endothelial NOS (eNOS), and an inducible form (inducible NOS, iNOS)¹⁵. NO is a potent vasodilator in the bronchial circulation and may regulate airway blood flow. It also modulates vascular tone via its vasodilatory properties. Excessive NO levels can cause hypotension associated with sepsis, whereas reduced NO levels in the lungs may contribute to pathologic states linked to pulmonary hypertension. Previous studies have shown that vagotomy in rats promoted pulmonary oedema, with significantly higher oedema observed in experimental animals than in controls^12,16, and iNOS may play a direct role in vagotomy-induced pulmonary oedema.

Risk factors for postoperative NPE in patients with medulla oblongata-involved tumours

As demonstrated by the present investigation, the absence of intraoperative use of sevoflurane, protracted postoperative tracheal extubation, and reduced levels of arginine and homoarginine in the perioperative period are independent risk factors for the occurrence of postoperative NPE.

This study has unearthed an unexpected correlation between the inhalational anaesthetic sevoflurane and the incidence of NPE. Previous research has established the neuroprotective and neurotoxic properties of the inhaled anaesthetic sevoflurane¹⁷. Recent investigations have suggested that low-dose inhalation of sevoflurane can engender a potent neuroprotective effect in various acute CNS injuries, such as cerebral ischaemia–reperfusion injury, traumatic brain injury, cerebral hypoxia injury, and subarachnoid hemorrhage, primarily through antiapoptotic and anti-inflammatory effects^18–20. Sevoflurane preconditioning has also been shown to suppress microglia activation in the early stages of CNS injury by lowering the expression of inflammatory mediators, such as iNOS. In addition to mitigating neuroinflammation, experiments indicate that sevoflurane anaesthesia can confer lung-protective effects, such as reducing morphological damage, oxidative stress, and inflammatory response in a mouse model of ventilator-induced lung injury²¹.

The current clinical study has confirmed that the intraoperative use of sevoflurane serves as a protective measure against the onset of postoperative NPE in patients suffering from medulla oblongata-involved tumours. Although the exact mechanism behind this phenomenon remains elusive, this study highlights the need for further research to establish a optimal protocol for intraoperative use of inhaled sevoflurane in patients with medulla oblongata-involved tumours to adequately exert its lung-protective effect while minimizing its impact on the nerve-protective effect of intraoperative electrophysiological monitoring.

Arginine, a conditionally essential amino acid, is metabolized by two key enzymes, arginase and NOS, which compete for the same substrate^22,23, resulting in dynamic equilibrium at physiological concentrations. It has been reported that circulating arginine levels decline under conditions of stress, such as surgery or trauma, and are proportional to the severity of the injury²⁴.

Two mechanisms related to these enzymes may contribute to the decrease in circulating arginine. (1) Competition of arginase for arginine: The release of anti-inflammatory cytokines early after stress²⁵ activates arginase, which, by competing for arginine, further promotes the degradation of arginine and inhibits NO production in the arginine/NO pathway. (2) Accelerated consumption of arginine by iNOS: Elevated levels of proinflammatory cytokines induced by stress²⁶ enhance the transmembrane transport of L-arginine into endothelial cells²⁷ and increase metabolism via NOS. Excessively expressed iNOS, in turn, augments inflammatory responses¹⁴. Therefore, a positive feedback loop between proinflammatory cytokines and iNOS is formed, accelerating the aforementioned transmembrane transport of arginine to synthesize NO and resulting in a subsequent decrease in endogenous arginine levels.

Homoarginine is an endogenous homologue of arginine that presents with an additional methylene group in the carbon chain²⁸. Although normally present in low concentrations in most bodily fluids^29,30, as confirmed in the present study (Fig. 1B), prior research has indicated that the decline in circulating homoarginine is correlated with the severity of left ventricular dysfunction^31–34 and pulmonary arterial hypertension (PAH)³⁵.

The results of this clinical research suggested that plasmatic arginine and homoarginine of the whole series decreased in the early period postoperatively, and the trend had no correlation with the way tumours affected the medulla oblongata. Moreover, it proved that decreased concentrations of arginine and homoarginine are independent risk factors for postoperative NPE.

Stress-induced reduction in plasmatic arginine and homoarginine was limited to early postoperative hours

The stress-induced reduction in circulating arginine³⁶ possesses two characteristics: proportional to the severity of injury²⁴ and time-bound³⁶. Previous research has shown that the increase in arginase activity after severe trauma is more than five times greater than after minor elective surgery²⁴. The activity of arginase peaks at 2–4 h after surgery or trauma, and its release is limited to a few hours³⁶. Therefore, it is logical to assume that systemic arginine depletion after a major surgery will be limited both in extent and in duration. The concentration of arginine after a major surgery is expected to start with a sharp decline followed by a bounce-back. In this prospective study of patients with medulla oblongata-involved tumours, the significant reduction in arginine concentration during the early postoperative period and its subsequent recovery support these previous hypotheses.

The potential clinical implications of the study’s findings are twofold. First, plasmatic arginine and homoarginine could serve as potential early warning indicators of postoperative NPE if their concentrations significantly decrease within 8 hours postoperatively compared to baseline levels. This could lead to early intervention measures and improved patient outcomes. However, further clinical trials are necessary to confirm this hypothesis. Second, active supplementation with arginine and homoarginine may be an effective strategy to prevent postoperative NPE. Arginine is a conditionally essential amino acid, and its replacement has been recommended. Intravenous infusion of arginine can lead to a maximum plasma concentration within 20-30 minutes, while oral administration takes around 60 min to reach peak levels³⁷. Animal tests have confirmed that high doses of L-arginine can reduce pulmonary oedema and improve survival rates³⁸. Future studies on medulla oblongata-involved tumours patients with postoperative NPE may focus on comparing patients with and without L-arginine supplementation and determining the optimal dosage and timing of supplementation.

Limitations

Potential limitations of this study should be taken into consideration. First, selection bias is inevitable due to the single-centre-based design and the enroled patients who were predominantly Han Chinese population. Second, the present rigorous criteria restricted the sample size and the statistical power, and further restricted the exploration of the difference of the incidence of postoperative NPE between the medulla oblongata-originated group and the perimedulla-originated group.

Conclusion

The findings of this study suggest that reductions of plasmatic arginine and homoarginine could potentially serve as early warning indicators of postoperative NPE in patients with medulla oblongata-involved tumours. Therefore, detections of the concentrations of these two neurotransmitters in peripheral blood are suggested both before anaesthesia and within 8 h after surgery. Active supplementation with arginine and homoarginine may be effective in preventing postoperative NPE. Future studies should investigate the optimal dosage and timing of L-arginine supplementation and further compare patients with and without supplementation. Additionally, intraoperative use of sevoflurane and prompt tracheal extubation may help reduce the risk of NPE in these patients. Overall, these results provide valuable insights into the prevention and management of NPE in patients with medulla oblongata-involved tumours undergoing surgery.

Ethical approval

All procedures performed in studies involving human participants were in accordance with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards and were reviewed and approved by the ethics committee of Beijing Tiantan Hospital, Capital Medical University (KY2019-107-02). The patients/participants provided written informed consent to participate in this study.

Consent

Written informed consents were obtained from the patients for publication of this article. Copies of the written consents are available for review by the Editor-in-Chief of this journal on request.

Sources of funding

This work was supported by the Open Projects of Clinical College (Department) of Capital Medical University (Clinical study of the application of arginine and sevoflurane for the prevention of postoperative neuronal pulmonary oedema in patients with skull base/brainstem tumours).

Author contribution

L.W. designed this prospective clinical study. Y.Z. and G.H.Z. acquired the data. Q.Z. and K.W. analyzed and interpreted the data. L.W. and Q.Z. drafted the article. All authors critically revised the article. L.W., Z.W., J.T.Z., W.J. and G.J.Z. provided administrative/technical/material support. All authors read and approved the final manuscript.

Conflicts of interest disclosure

There are no conflicts of interest.

Guarantor

Guo-Jun Zhang.

Data availability statement

All data generated or analysed during this study are included in this published article and its supplementary information files.

Provenance and peer review

This paper is not commissioned.