The usefulness of diffusion-weighted magnetic resonance imaging performed in the acute phase as an early predictor of delayed neuropsychiatric sequelae in acute carbon monoxide poisoning

Abstract

Delayed onset of neuropsychiatric symptoms after apparent recovery from acute carbon monoxide (CO) poisoning has been described as delayed neuropsychiatric sequelae (DNS). No previous study has determined whether early use of diffusion-weighted magnetic resonance imaging (DWI) can predict which patients will develop DNS in the acute CO poisoning. This retrospective observational study was performed on adult patients with acute CO poisoning consecutively treated over a 17-month period. All included patients with acute CO poisoning underwent DWI to evaluate brain injury within 72 h after CO exposure. DWI was evaluated as follows: (1) presence of pathology, (2) number of pathologies, (3) asymmetry, and (4) location of pathology. Patients were divided into two groups. The DNS group was composed of patients with delayed sequelae, while the non-DNS group included patients with no sequelae. A total of 102 patients with acute CO poisoning were finally enrolled in this study. DNS developed in 10 patients (9.8%). Between the DNS group and the non-DNS group, presence of pathology on DWI and initial Glasgow Coma Scale (GCS) showed significant difference. There was also a statistical difference between the non-DNS group and DNS group in terms of CO exposure time, troponin I, rhabdomyolysis, acute kidney injury, and pneumonia. The presence of pathology in DWI and initial GCS (cutoff: <12) at the emergency department served as an early predictors of DNS.

Keywords

Carbon monoxide poisoning complications neurotoxicity diffusion magnetic resonance imaging

Introduction

Carbon monoxide (CO) can cause various neurologic sequelae, from no deficit to permanent injury. Neuronal cell injury occurs in CO poisoning by three main mechanisms: (1) cells undergo hypoxia because of the high affinity of CO for hemoglobin;¹ (2) CO exposure can cause inflammation through independent pathways to hypoxia, such as ischemic reperfusion injury, CO effects on vascular endothelium, oxygen-radical-mediated lipid peroxidation, and nitric oxide liberated from platelets at the time of CO exposure;^2

–6 and (3) even without inflammation, CO exposure can also trigger cell apoptosis.^4,5 Among various neurologic sequelae, delayed neuropsychiatric sequelae (DNS), which is delayed onset of neuropsychiatric symptoms after apparent recovery of neurocognitive symptoms developed by acute CO poisoning, are observed in some patients. There are various symptoms and signs of DNS, including mental deterioration, cognitive dysfunction, amnesia, gait disturbance, mutism, urinary or fecal incontinence, psychosis, depression, and Parkinsonism.^2,7,8 DNS occurs in 10–30% of victims.^7,9,10 Choi⁷ reported that the lucid interval before the appearance of neurologic sequelae is generally from 2 to 40 days, although various results have been reported.^9,10 Clinical parameters, including mental status and level of carboxyhemoglobin (CO-Hb), have been found to be unhelpful in predicting neurologic severity and sequelae.^11,12

DNS is pathologically characterized by demyelination of white matter.^13

–17 Neuroimaging, including conventional computed tomography (CT) and magnetic resonance imaging (MRI), is usually used to examine white matter abnormalities in patients with DNS. There are several reports on the roles of CT and conventional MRI for predicting neurologic severity in the acute phase of CO poisoning.^13,18
–20 Diffusion-weighted MRI (DWI; Philips, Achieva 3.0T, Eindhoven, the Netherlands) is a new generation MRI sequence based on translational movement of water. DWI can detect early changes in ischemic tissues within 30 min of onset and is sensitive to cytotoxic or vasogenic edema.^21,22 The apparent diffusion coefficient (ADC) provides a quantitative measure of water diffusion and can be used to distinguish vasogenic from cytotoxic edema on DWI. In acute ischemic stroke with cytotoxic edema, decreased water diffusion in infarcted tissue causes increased DWI signal and decreased ADC signal.²³ There are only case reports on the diagnostic usefulness of DWI, which can detect early changes in ischemic tissues and is sensitive to cytotoxic or vasogenic edema,^21,22 in the acute phase of CO exposure.^12,24,25 No previous study has determined whether early use of DWI can predict which patients will develop DNS in the acute CO poisoning.

The aim of this study was to evaluate differences in the findings of DWI performed within 72 h after an episode of CO poisoning according to the presence of DNS. This study also evaluated the usefulness of DWI as an early predictor of DNS in acute CO poisoning.

Methods

Patients

This retrospective observational study was performed on adults >18 years of age with acute CO poisoning. Patients were consecutively enrolled over a 17-month period (from January 2015 to May 2016), and DWI was measured during admission (within 72 h after CO exposure), if patients agreed, as part of the standard management strategy to patients with symptoms by CO poisoning since January 2015. The emergency department (ED) is located in a single, suburban, tertiary-care hospital (Republic of Korea), has more than 43,000 annual visits, and is staffed 24 h a day by board-certified emergency physicians.

A diagnosis of CO poisoning was based on patient medical history and CO-Hb >5% (>10% in smokers). All patients presenting with acute CO poisoning upon arrival to the ED were treated with 100% high-flow oxygen therapy through a face mask with a reservoir bag and hyperbaric oxygen therapy (HBOT) through a multiplace hyperbaric chamber (IBEX Medical Systems, Seoul, Korea), if indicated, including any interval of unconsciousness, neuropsychological symptoms and signs, cognitive dysfunction, cardiovascular dysfunction, severe acidosis, or CO-Hb ≥25%.²⁶ HBOT could not be performed in patients who were hemodynamically unstable, had uncontrolled irritability, or could not undergo middle ear equalization. Patients were educated about DNS at the time of hospital discharge and were encouraged to return to the hospital if they experienced DNS symptoms. Discharged patients were followed for at least 2 months based on Choi’s observation that the lucid interval is generally from 2 to 40 days.⁷ The presence of the following symptoms and signs, including mental deterioration, cognitive dysfunction, amnesia, gait disturbance, mutism, urinary or fecal incontinence, psychosis, depression, and Parkinsonism, were assessed.^2,7,8 The current status of patients was confirmed through telephone consultation in December 2016.

Exclusion criteria for patients in this study were as follows: (1) age <18 years; (2) not admitted or transferred to other hospital after treatment with HBOT; (3) failure to follow-up after discharge; (4) did not undergo DWI or DWI was performed more than 72 h after acute CO poisoning; (5) previous neurocognitive dysfunction including dementia, psychiatric disease, Parkinson disease, or cerebrovascular disease; and (6) failure to recover decreased mental status (i.e. permanent neurologic injury) or death.

Study variables and definitions

Two emergency physicians (HY and YSL) who were blinded to the study objectives and hypothesis collected data by retrospectively reviewing the electronic medical records of patients. The categorization of the patient’s group, which was done by other emergency physician (YSK), was blinded to the abstractors. The abstractors were also blinded to categorization of the patient groups and were trained prior to data collection to reduce possible bias from the data collection procedure. Explicit case report forms were used in this study. Chart abstractors and study coordinator (YSC) met periodically to resolve any disputes and review coding rules. Study coordinators monitored the performance of abstractors. Patient records and information were anonymized prior to analysis.

Variables that are important in acute CO poisoning were examined. The clinical parameters assessed were age; sex; source of CO; intention for self-harm; CO exposure time; time elapsed from rescue to arrival at the ED; co-ingestion of alcohol; past medical history; use of HBOT; initial Glasgow Coma Scale (GCS) at the ED; vital signs, symptoms, and signs at the ED; and complications during admission such as rhabdomyolysis,²⁷ acute kidney injury,²⁸ and pneumonia. Laboratory parameters were levels of CO-Hb, which was measured from prehospital sources or obtained at our hospital; and serum high-sensitivity troponin I (TnI), arterial blood gas, and serum lactate measured after ED arrival. All included patients with acute CO poisoning underwent DWI to evaluate for brain injury within 72 h after CO exposure. DWI (Philips) was performed by using a three-gradient protocol (4289/68 [repetition time (TR)/echo time (TE)]). Heavily diffusion-weighted (b = 1000 s/mm²) images and automatically generated ADC maps were studied. One experienced neuroradiologist (KMS) analyzed DWI for this study while blinded to clinical data. DWI was evaluated as follows: (1) presence of pathology, (2) number of pathologies, (3) asymmetry, and (4) location of pathology.

Patients were divided into two groups. The DNS group was composed of patients with delayed sequelae of mental deterioration, cognitive dysfunction, gait disturbance, mutism, urinary or fecal incontinence, psychosis, depression, or Parkinsonism, which were confirmed in follow-up. The non-DNS group included patients with none of these sequelae. This study received institutional review board approval (approval number: CR316115).

Study end points

The primary goals of this study were to compare DWI findings associated with the presence of DNS and to investigate the usefulness of DWI as an early predictor of DNS in acute CO poisoning.

Statistical analysis

Categorical variables are presented as frequencies and percentages, while continuous variables are presented as means and standard deviations or as medians and interquartile ranges. χ² or Fisher’s exact tests were used to compare categorical variables, while two-sample t- or Mann–Whitney U-tests were used to compare continuous variables. Normality was first assessed using the Shapiro–Wilk test. Multivariate logistic regression analysis was used to identify the factors to predict the DNS, and the results are expressed as odds ratio (OR) with 95% confidence interval (CI). The area under the curve (AUC) for the predictive ability for the presence of DNS was determined using receiver operating characteristic (ROC) curves. Optimal cutoff points of predictors were evaluated by ROC curves and the Youden index. p Values less than 0.05 were considered statistically significant, and all statistical analyses were conducted using SPSS Statistics for Windows, version 23.0 (IBM, Armonk, New York, USA) and MedCalc Statistical Software version 17.5.3 (MedCalc Software, Ostend, Belgium).

Results

Patient characteristics

A total of 152 consecutive adult patients with acute CO poisoning were identified during the study period. Fifty patients were excluded based on the following criteria: younger than 18 years (4 patients), not admitted or transferred to other hospital after treatment with HBOT (15 patients), failed to follow-up (4 patients), DWI not performed because patients or caregiver did not agree (13 patients), previous neurocognitive dysfunction including cerebral infarct or dementia (3 patients), and failure to recover from neurologic complications including decreased mental function (11 patients). After exclusion, a total of 102 patients with acute CO poisoning were enrolled in this study (Figure 1).

Figure 1.

Flow diagram of patient selection. CO: carbon monoxide; DWI: diffusion-weighted magnetic resonance imaging.

Baseline characteristics and laboratory findings are shown in Table 1. Of the 102 analyzed patients, the median age was 55.5 years (18–89 years), and there were 59 male patients (57.8%). Charcoal was the most common source of CO (74 patients, 72.5%), and intention for self-harm was present in 35 patients (34.3%). The median CO exposure was 4 h, and HBOT was performed in 97 patients (95.1%). The most common symptoms and signs in the early phase of acute CO poisoning included loss of consciousness (58 patients, 63.7%) and cognitive dysfunction developed in 29 patients (39.7%). Neurologic symptoms including leg weakness, sensory change in extremities, facial sensory change, dysarthria, or diplopia occurred in nine patients (8.8%). The median CO-Hb, TnI, and lactate of all patients were 25.4%, 0.090 ng/mL, and 2.03 mmol/L, respectively. DNS developed in 10 patients (9.8%), and symptoms of DNS were mental deterioration, cognitive dysfunction, and gait disturbance. The presence of pathology on DWI was shown in 25 patients (24.5%). Of complications during hospitalization, rhabdomyolysis (30 patients, 29.4%) was the most common (Table 2).

Table 1.

Comparison of baseline characteristics according to delayed neurologic sequelae.

Variables	Total (n = 102)	Non-DNS (n = 92, 90.2%)	DNS (n = 10, 9.8%)	p Value
Age	55.5 (36.8–69)^a	55 (36–68)^a	69.5 (50.5–77.8)^a	0.065
Male gender (%)	59 (57.8)	54 (58.7)	5 (50.0)	0.739
Source of CO				1.000
Charcoal (%)	74 (72.5)	66 (71.7)	8 (80)
Firewood (%)	22 (21.6)	20 (21.7)	2 (20)
LPG (%)	4 (3.9)	4 (4.3)	0 (0)
Intention for self-harm (%)	35 (34.3)	32 (34.8)	3 (30)	1.000
CO exposure time (h)	4 (2–9.5)^a	3.5 (2–9)^a	12 (7.7–16)^a	<0.001
Time elapsed from rescue to arrival at the ED (h)	2.6 (1.3–4.2)^a	2.6 (1.5–4.2)^a	0.6 (0.6–7.2)^a	0.132
Co-ingestion of alcohol (%)	27 (26.7)	25 (27.5)	2 (20)	1.000
Past history
DM (%)	11 (10.8)	11 (12)	0 (0)	0.595
HTN (%)	26 (25.5)	22 (23.9)	4 (40)	0.272
Hyperlipidemia (%)	6 (5.9)	6 (6.5)	0 (0)	1.000
HBOT (%)	97 (95.1)	89 (96.7)	8 (80)	0.074
Initial GCS at the ED	15 (12–15)^a	15 (13–15)^a	10 (7.5–12.8)^a	0.002
SBP	133 (117–149)^a	132 (117–149)^a	134 (111–145)^a	0.685
Symptoms
Headache (%)^b	34 (40.5)	33 (41.8)	1 (20)	0.644
Dizziness (%)^b	42 (50)	40 (50.6)	2 (40)	1.000
Dyspnea (%)^b	14 (16.7)	14 (17.7)	0 (0)	0.584
Chest pain (%)^b	4 (4.8)	4 (5.1)	0 (0)	1.000
General weakness (%)	31 (36.9)	29 (36.7)	2 (40)	1.000
Loss of consciousness (%)^c	58 (63.7)	55 (64)	3 (60)	1.000
Cognitive dysfunction (%)^d	29 (39.7)	27 (38)	2 (100)	0.154
Neurologic symptoms (%)	9 (8.8)	7 (7.6)	2 (20)	0.214
Motor symptoms (%)	2 (2)	1 (1.1)	1 (10)	0.187
Sensory symptoms (%)	3 (2.9)	3 (3.3)	0 (0)	1.000
Dysarthria (%)	4 (3.9)	3 (3.3)	1 (10)	0.342
Diplopia (%)	1 (1)	1 (1.1)	0 (0)	1.000
CO-Hb (%)	25.4 (13.7–36.6)^a	23.8 (13.5–36.7)^a	27.4 (13.2–35.0)^a	0.731
Troponin I (ng/mL)	0.090 (0.015–0.672)^a	0.055 (0.015–0.633)^a	0.814 (0.375–1.45)^a	0.003
Bicarbonate (HCO₃) (mmol/L)	20.5 (18.0–22.4)^a	20.6 (18.6–22.3)^a	18.7 (16.1–23.3)^a	0.456
Lactate (mmol/L)	2.03 (1.25–3.86)^a	2.01 (1.17–3.27)^a	3.78 (1.72–4.28)^a	0.106

DNS: delayed neuropsychiatric sequelae; CO: carbon monoxide; LPG: liquefied petroleum gas; ED: emergency department; DM: diabetes mellitus; HTN: hypertension; HBOT: hyperbaric oxygen therapy; GCS: Glasgow Coma Scale; SBP: systolic blood pressure; CO-Hb: carboxyhemoglobin.

^aMedian (interquartile range).

^bEighty-four patients were investigated except for patients who were unconscious at the ED.

^cNinty-one patients were investigated except for patients who were unconscious at the ED.

^dSeventy-three patients were investigated except for patients who were unconscious at the ED.

Table 2.

Presence of pathology in diffusion-weighted magnetic resonance imaging and complications during admission.

Variables	Total (n = 102)	Non-DNS (n = 92, 90.2%)	DNS (n = 10, 9.8%)	p Value
Presence of pathology in DWI (%)	25 (24.5)	18 (19.6)	7 (70)	0.002
Complications
Rhabdomyolysis (%)	30 (29.4)	23 (25)	7 (70)	0.006
Acute kidney injury (%)	9 (8.8)	5 (5.4)	4 (40)	0.005
Pneumonia (%)	6 (5.9)	2 (2.2)	4 (40)	0.001

DNS: delayed neuropsychiatric sequelae; DWI: diffusion-weighted magnetic resonance imaging.

Comparisons of general characteristics and laboratory findings based on the presence of DNS

Comparisons of baseline characteristics and laboratory findings are shown in Table 1. CO exposure time and initial GCS significantly differed between patients in the non-DNS and DNS groups (3.5 h vs. 12 h, p < 0.001 and 15 vs. 10, p = 0.002, respectively). The median TnI level was also significantly higher in the DNS group than in the non-DNS group (0.814 ng/mL vs. 0.055 ng/mL, p = 0.003). Initial CO-Hb level, use of HBOT, and neurologic symptoms did not significantly differ between groups. Comparisons of DWI analysis and complications are shown in Table 2. The presence of pathology on DWI was significantly higher in the DNS group than in the non-DNS group (70% vs. 19.6%, p = 0.002). Of complications during admission, rhabdomyolysis, acute kidney injury, and pneumonia were significantly more common in the DNS group than in the non-DNS group.

Early predictors of DNS in acute CO poisoning

Clinically important variables and statistically different variables in this study (age, CO exposure time, initial GCS, TnI, and presence of pathology on DWI) were analyzed by multiple logistic regressions to identify early predictors related to the development of DNS. The presence of pathology on DWI (OR, 6.438; 95% CI, 1.201–34.499, p = 0.030) and initial GCS (OR, 0.772; 95% CI, 0.609–0.979, p = 0.033) were significant early predictor for DNS (Table 3). The AUC of presence of pathology on DWI and initial GCS (cutoff: <12) for early prediction of DNS were 0.752 (95% CI, 0.580–0.925) and 0.774 (95% CI, 0.680–0.851), respectively (Table 4).

Table 3.

Predictors of delayed neurologic sequelae analyzed by multiple logistic regression.

Variables	Univariate analysis			Multivariate analysis
Variables	OR	95% CI	p Value	OR	95% CI	p Value
Age	1.037	0.994–1.081	0.090	1.031	0.981–1.083	0.227
CO exposure time (h)	1.236	1.090–1.401	0.001	1.143	0.984–1.328	0.080
Initial GCS	0.755	0.629–0.907	0.003	0.772	0.609–0.979	0.033
Troponin I	1.110	0.948–1.299	0.196	1.032	0.834–1.276	0.774
Presence of pathology in DWI	9.593	2.256–40.783	0.002	6.438	1.201–34.499	0.030

OR: odds ratio; CI: confidence interval; CO: carbon monoxide; GCS: Glasgow Coma Scale; DWI: diffusion-weighted magnetic resonance imaging.

Table 4.

Diagnostic accuracy of predictors for the development of delayed neurologic sequelae.^a

Variables	AUC	Sensitivity (%)	Specificity (%)	PPV (%)	NPV (%)
Initial GCS <12	0.774 (0.680–0.851)	80 (44.4–97.5)	77.2 (67.2–85.3)	28.1 (13.1–47.7)	97.2 (90.3–99.7)
Presence of pathology in DWI	0.752 (0.580–0.925)	70 (34.8–93.3)	80.4 (70.9–88)	28.4 (12.4–49.8)	96 (88.9–99.2)

AUC: area under curve; PPV: positive predictive value; NPV: negative predictive value; GCS: Glasgow coma scale; DWI: diffusion-weighted magnetic resonance imaging.

^aValue (95% confidence interval).

Characteristics of DWI findings based on the presence of DNS

A comparison of DWI findings is shown in Table 5. There was no significant difference between two groups in terms of checked days of DWI after ED arrival. The number of pathologies (≥3) on DWI was higher in the DNS group than in the non-DNS group, although there was no significant difference between the two groups (42.9% vs. 27.8%, p = 0.640). The globus pallidus was the most common pathologic site (76%). Although not significantly different, globus pallidus was more common in the non-DNS group than in the DNS group. However, periventricular white matter and centrum semiovale were only found in the DNS group (Figure 2).

Figure 2.

Representative diffusion-weighted magnetic resonance imaging in the patients with delayed neurologic sequelae after carbon monoxide poisoning. (a) Centrum semiovale (diffusion-weighted magnetic resonance imaging). (b) Centrum semiovale (apparent diffusion coefficient). (c) Periventricular white matter (diffusion-weighted magnetic resonance imaging). (d) Periventricular white matter (apparent diffusion coefficient).

Table 5.

Comparison of diffusion-weighted magnetic resonance imaging findings according to delayed neurologic sequelae in patients with pathology on diffusion-weighted magnetic resonance imaging.

Variables	Total (n = 102)	Non-DNS (n = 92, 90.2%)	DNS (n = 10, 9.8%)	p Value
Presence of pathology (%)	25 (24.5)	18 (19.6)	7 (70)	0.002
Number of pathology				0.640
≤2 (%)	17 (68)	13 (72.2)	4 (57.1)
≥3 (%)	8 (32)	5 (27.8)	3 (42.9)
Checked days of DWI after ED arrival	2 (2–2)	2 (2–2)	2 (2–2)	0.844
Asymmetry (%)	10 (40)	6 (33.3)	4 (57.1)	0.378
Location of pathology^a
Globus pallidus (%)	19 (76)	15 (83.3)	4 (57.1)	0.298
Putamen (%)	1 (4)	1 (5.6)	0 (0)	1.000
Periventricular white matter (%)	1 (4)	0 (0)	1 (14.3)	0.280
Centrum semiovale (%)	1 (4)	0 (0)	1 (14.3)	0.280
Cerebellum (%)	5 (20)	2 (11.1)	3 (42.9)	0.113
Cortex (%)	8 (32)	6 (33.3)	2 (28.6)	1.000
Frontal lobe (%)^b	1 (12.5)	0 (0)	1 (50)	0.250
Temporal lobe (%)^b	6 (72.5)	4 (66.7)	2 (100)	1.000
Parietal lobe (%)^b	4 (50)	2 (33.3)	2 (100)	0.429
Occipital lobe (%)^b	3 (37.5)	1 (16.7)	2 (100)	0.107

DNS: delayed neuropsychiatric sequelae; DWI: diffusion-weighted magnetic resonance imaging; ED: emergency department.

^aTwenty-five patients were investigated.

^bEight patients were investigated.

Discussion

In this study, the presence of pathology on DWI performed in the acute phase, regardless of location, was statistically different according to the presence of DNS. The presence of pathology on DWI was an early predictor of DNS in acute CO poisoning. DWI performed in the very early period (within 72 h after CO exposure) can sensitively identify acute cytotoxic edema caused by hypoxia from CO exposure, which cannot be detected in any other imaging systems. We suggest that, with more CO-induced hypoxic effects in the acute phase, there is more demyelination of cerebral white matter, which is a pathological characteristic of DNS.^13

–17 Therefore, the presence of pathology on DWI may be an early predictor of DNS. Until now, there were only case reports about DWI in the acute phase of CO poisoning,^12,25 and there are few case reports related to DWI and DNS.²⁵ Teksam et al. reported two cases that indicate that DWI abnormalities in CO poisoning may prevent misdiagnosis and prediction of prognosis in acute CO poisoning.²⁵

In this study, the DNS group had more patients with more than three pathologic sites than the non-DNS group, although there was no statistically significant difference. O’Donnell et al. reported on 19 patients who underwent magnetic resonance (MR) without DWI within 126 h after CO exposure and all patients with pathologic lesions in more than two areas on MRI showed subsequent impaired functional status.¹⁸ They suggested that the extent of the abnormality on conventional brain MRI can be of use in determining degree of severity and prognosis, especially when clinical, biochemical, and hematological parameters are equivocal.¹⁸

DNS after CO poisoning results from progressive demyelination in the deep cerebral white matter.^29,30 The most common demyelination areas are the periventricular white matter and centrum semiovale.^14,16,31 This study included patients with pathology in the periventricular white matter and centrum semiovale on DWI performed during the acute phase, and they were all in the DNS group. We think that white matter pathology including periventricular white matter and centrum semiovale in the acute phase after CO intoxication may represent severe brain damage that can lead to DNS. However, Teksam et al. reported that, although two cases contained lesions of the periventricular white matter, one patient did not develop DNS and another experienced DNS 2 months after CO exposure.²⁵ Parkinson et al. reported that, although CO poisoning can result in brain injury manifested by white matter hyperintensities in conventional MR (not included DWI) performed in the early phase, it did not seem that white matter hyperintensities are related to CO poisoning severity.¹⁹ They suggested that centrum semiovale hyperintensities were only related to worse cognitive performance.¹⁹ Although cerebellar abnormalities in CO poisoned patients are rare, one study that examined variables predictive of subsequent adverse outcomes showed that cerebellar abnormalities on presentation were associated with an increased risk of cognitive sequelae at 6 weeks.³² In this study, there were more cerebellar lesions in the DNS group than in the non-DNS group, although there was no statistically significant difference (42.9% vs. 11.1%, p = 0.113). After inhalation, CO directly binds to heme iron in the globus pallidus, which is rich in iron.¹⁴ Ischemia in the globus pallidus, caused by hypoxic effects of CO exposure, is a specific characteristic of severe CO poisoning.^18,31,33 Several studies have suggested that the globus pallidus is the most common site of involvement in CO poisoning, and damage usually occurs immediately.^16,18 Although the globus pallidus was the most common site on DWI in this study, there was no significant difference according to the presence of DNS.

In this study, initial GCS at the ED was significantly lower in the DNS group and early predictor of DNS in acute CO poisoning. Hypoxia and inflammation caused by CO can cause neuronal injury in the brain and result in decreased consciousness.² This may mean that CO poisoning, which affects consciousness in the acute phase, can more easily develop into DNS in the sub-acute or chronic phase. In a report by Choi, most cases of DNS were associated with loss of consciousness in the acute phase of CO poisoning.⁷

In this study, CO exposure time was significantly longer in the DNS group than in the non-DNS group. This may be because of increased neuronal damage with longer CO exposure time. There is a study indicating that CO exposure lasting more than 24 h is a risk factor of DNS in patients with acute CO poisoning.³² We also observed that TnI was significantly elevated in the DNS group compared to the non-DNS group. Because the DNS group had a longer CO exposure time, more myocardial damage may have occurred, resulting in elevated TnI caused by CO-induced hypoxia and ischemic reperfusion injury. This is in line with the observation of greater incidence of rhabdomyolysis and acute kidney injury in the DNS group than in the non-DNS group. In addition, patients with DNS experienced more pneumonia complications than the non-DNS group. We think that aspiration pneumonia caused by decreased mental status might be more developed in the DNS group than in the non-DNS group because initial GCS at the ED was significantly lower in the DNS group.

This study had several limitations. First, it was limited by its retrospective design. Missing data during collection were also a limitation. Second, this study was conducted at a single hospital, resulting in a small sample size. Because of low number of patients with DNS, it was not possible to determine which specific DWI lesions have prognostic value. However, all patients with acute CO poisoning and DWI since January 2015 were investigated to reduce this possible bias. Third, this study may be susceptible to ascertainment bias. The patients or families presumably were aware of their DWI results, and this may have influenced their self-reported outcomes. Fourth, the follow-up period was not the same in all patients. However, all patients were followed for at least 2 months in this study. Choi reported that the lucid interval before the appearance of neurologic sequelae is generally from 2 to 40 days.⁷ Additional follow-up was also conducted via telephone. Although follow-up via telephone may be not an objective tool to identify DNS, we considered DNS that cannot be recognized by patients as not clinically important. Fifth, patients excluded due to various reasons may lead to selection bias. Sixth, serial DWI was not measured. Therefore, we could not investigate serial changes in patients with pathology in the DWI. A well-designed prospective study is necessary to clarify these limitations.

Conclusions

The presence of pathologic lesions on DWI and initial GCS less than 12 at the ED served as an early predictor of development of DNS.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Ernst

Zibrak

. Carbon monoxide poisoning. N Engl J Med 1998; 339: 1603–1608.

Weaver

. Carbon monoxide poisoning. N Engl J Med 2009; 360: 1217–1225.

Zhang

Piantadosi

. Mitochondrial oxidative stress after carbon monoxide hypoxia in the rat brain. J Clin Invest 1992; 90: 1193.

Thom

. Dehydrogenase conversion to oxidase and lipid peroxidation in brain after carbon monoxide poisoning. J Appl Physiol 1992; 73: 1584–1589.

Thom

. Leukocytes in carbon monoxide-mediated brain oxidative injury. Toxicol Appl Pharmacol 1993; 123: 234–247.

Ischiropoulos

Beers

Ohnishi

. Nitric oxide production and perivascular nitration in brain after carbon monoxide poisoning in the rat. J Clin Invest 1996; 97: 2260.

Choi

. Delayed neurologic sequelae in carbon monoxide intoxication. Arch Neurol 1983; 40: 433–435.

Pan

Wan

. Factors affecting the prognosis of patients with delayed encephalopathy after acute carbon monoxide poisoning. Am J Emerg Med 2011; 29: 261–264.

Hart

Kennedy

Adams

. Neurological manifestation of carbon monoxide poisoning. Postgrad Med J 1988; 64: 213–216.

10.

Sawa

Watson

Terbrugge

. Delayed encephalopathy following carbon monoxide intoxication. Can J Neurol Sci 1981; 8: 77–79.

11.

Hurley

Hopkins

Bigler

. Applications of functional imaging to carbon monoxide poisoning. J Neuropsychiatry Clin Neurosci 2001; 13: 157–160.

12.

Stephen

Donal

Siobhain

. Carbon monoxide poisoning: novel magnetic resonance imaging pattern in the acute setting. Int J Emerg Med 2012; 5: 30.

13.

Miura

Mitomo

Kawai

. CT of the brain in acute carbon monoxide intoxication: characteristic features and prognosis. Am J Neuroradiol 1985; 6: 739–742.

14.

Chang

Han

Kim

. Delayed encephalopathy after acute carbon monoxide intoxication: MR imaging features and distribution of cerebral white matter lesions. Radiology 1992; 184: 117–122.

15.

Kwon

Chung

. Delayed postanoxic encephalopathy after carbon monoxide poisoning. Emerg Med J 2004; 21: 250–251.

16.

Chu

Jung

Kim

. Diffusion-weighted MRI and 99mTc-HMPAO SPECT in delayed relapsing type of carbon monoxide poisoning: evidence of delayed cytotoxic edema. Eur Neurol 2004; 51: 98–103.

17.

Kim

Chang

Song

. Delayed encephalopathy of acute carbon monoxide intoxication: diffusivity of cerebral white matter lesions. Am J Neuroradiol 2003; 24: 1592–1597.

18.

O’Donnell

Buxton

Pitkin

. The magnetic resonance imaging appearances of the brain in acute carbon monoxide poisoning. Clin Radiol 2000; 55: 273–280.

19.

Parkinson

Hopkins

Cleavinger

. White matter hyperintensities and neuropsychological outcome following carbon monoxide poisoning. Neurology 2002; 58: 1525–1532.

20.

Silver

Cross

Fox

. Computed tomography of the brain in acute carbon monoxide poisoning. Clin Radiol 1996; 51: 480–483.

21.

Le Bihan

Breton

Lallemand

. MR imaging of intravoxel incoherent motions: application to diffusion and perfusion in neurologic disorders. Radiology 1986; 161: 401–407.

22.

Schäbitz

Fisher

. Diffusion weighted imaging for acute cerebral infarction. Neurol Res 1995; 17: 270–274.

23.

Schlaug

Siewert

Benfield

. Time course of the apparent diffusion coefficient (ADC) abnormality in human stroke. Neurology 1997; 49: 113–119.

24.

Ozcan

Ozcam

Kosar

. Correlation of computed tomography, magnetic resonance imaging and clinical outcome in acute carbon monoxide poisoning. Braz J Anesthesiol 2016; 66: 529–532.

25.

Teksam

Casey

Michel

. Diffusion-weighted MR imaging findings in carbon monoxide poisoning. Neuroradiology 2002; 44: 109–113.

26.

Gesell

. Hyperbaric oxygen 2009: indications and results: the hyperbaric oxygen therapy committee report. Durham, NC: Undersea and Hyperbaric Medical Society, 2008.

27.

Melli

Chaudhry

Cornblath

. Rhabdomyolysis: an evaluation of 475 hospitalized patients. Medicine 2005; 84: 377–385.

28.

Levey

Levin

Kellum

. Definition and classification of kidney diseases. Am J Kidney Dis 2013; 61: 686–688.

29.

Thom

. Carbon monoxide-mediated brain lipid peroxidation in the rat. J Appl Physiol 1990; 68: 997–1003.

30.

Thom

Bhopale

Fisher

. Delayed neuropathology after carbon monoxide poisoning is immune-mediated. Proc Natl Acad Sci USA 2004; 101: 13660–13665.

31.

Porter

Hopkins

Weaver

. Corpus callosum atrophy and neuropsychological outcome following carbon monoxide poisoning. Arch Clin Neuropsychol 2002; 17: 195–204.

32.

Weaver

Valentine

Hopkins

. Carbon monoxide poisoning: risk factors for cognitive sequelae and the role of hyperbaric oxygen. Am J Respir Crit Care Med 2007; 176: 491–497.

33.

Thom

Fisher

Manevich

. Roles for platelet-activating factor and· NO-derived oxidants causing neutrophil adherence after CO poisoning. Am J Physiol Heart Circ Physiol 2001; 281: H923–H930.