Derivation and Validation of a Score for Predicting Poor Neurocognitive Outcomes in Acute Carbon Monoxide Poisoning

Key Points Question Can a novel clinical scoring system predict poor neurocognitive outcomes after acute carbon monoxide poisoning? Findings This prognostic study developed and externally validated a prediction model including 5 risk factors associated with poor neurocognitive outcome at 1 month, creatine kinase level, hyperbaric oxygen therapy, Glasgow Coma Scale score, age, and shock (COGAS score), among patients with carbon monoxide poisoning. COGAS score showed excellent discrimination performance. Meaning These findings suggest that use of a reliable prediction model during the early phase of carbon monoxide poisoning could help identify patients at risk of poor neurocognitive sequelae.


eAppendix 1. Definition of the Study Variables
Carbon monoxide (CO) exposure duration, as reported by the patient or patients' guardians, was the expected maximum duration of CO exposure, measured from the time of normal state of consciousness to patient rescue. Any state of loss of consciousness was defined as a case of such, regardless of the length of loss of consciousness. Shock was diagnosed when a vasopressor was required to resuscitate the patient and if lactate levels exceeded 2.0 mmol/L.

eAppendix 2. Global Deterioration Scale Explanation
The Global Deterioration Scale (GDS) is a validated, reliable instrument for describing the clinical progression of dementia. 1 It is also used to determine the prognosis of patients with carbon monoxide (CO) poisoning, 2-4 severe chronic obstructive pulmonary disease, Alzheimer's disease, and vasculopathy-related dementia. 1,[5][6][7] Although the GDS score is not as diverse as a CO battery, it has the advantage of being able to identify neurocognitive functions, such as memory and concentration, as well as activities of daily living, through interviews. Moreover, many neurocognitive function tests may be difficult for patients with sequelae. The Short-Form General Health Survey-36, a commonly used testing tool, has a set of self-reported questions; however, it is limited in evaluating patients with severe neurological impairment as it requires an individual's ability to understand and address the questions. Digit span, trail making, and clock drawing are good evaluation tools but require short-term memory and visuospatial functions. Therefore, the GDS score can be used for all patients with CO poisoning regardless of poisoning severity. The scale consists of seven stages, with higher scores indicating greater severity.

Stage Cognitive dysfunction
Clinical characteristics 1 No cognitive decline Patients appear clinically normal.
No complaints of memory deficits.
No evident memory deficit on clinical interview.

Very mild cognitive decline
Patients complain of memory deficits.
Most frequently, patients: (a) forget where they have placed familiar objects (b) forget the name of someone they formerly knew well.
No objective evidence of memory deficit on clinical interview.
No objective deficits in employment or social situations.
Patients display appropriate concern about their symptoms.

Mild cognitive decline Earliest clear-cut deficits.
Objective evidence of memory deficit was obtained only through an intensive interview conducted by a trained geriatric psychiatrist. Concentration deficit may be evident on clinical testing.
Patients may demonstrate a reduced ability to: The subtlety of the clinical symptoms may be exacerbated by denial that is often manifested by these patients. Mild-tomoderate anxiety also accompanies the symptoms, typically when the patients are forced to cope with challenging employment and social demands that render them unable to negotiate. Denial is often the dominant defense mechanism. The evident decline in patients' intellectual and cognitive capacities is too overwhelming with a loss of full conscious acceptance and recognition. Flattening of effect and withdrawal from previously challenging situations is observed.

Moderately severe cognitive decline
Patients can no longer survive without some assistance.
During interviews, patients are unable to recall a major relevant aspect of their current lives. Examples include the following: (a) difficulty recalling their address or telephone number, names of close family members, such as grandchildren, or the name of the high school or university from which they graduated (b) some disorientation with time (date, day of the week, season) or location (c) well-educated patients may have difficulty counting backwards from 40 by fours or from 20 by twos.
Patients retain the knowledge of many major facts regarding themselves and others. They invariably know their own names and generally know their spouse and children's names. They require no assistance with toileting and eating but may have some difficulty choosing the proper clothing to wear and may occasionally clothe themselves improperly (e.g., put their shoes on the wrong feet). Patients are largely unaware of all recent events and experiences in their lives.
They retain some knowledge of their past but is very uncertain. They are generally unaware of their surroundings, the year, or the season and may have difficulty counting backward, and sometimes forward, from 10. Patients require substantial assistance with activities of daily living. These symptoms are quite variable and include the following: (a) delusional behavior (e.g., patients may accuse their spouse of being an impostor, may talk to imaginary figures in the environment, or their own reflection in the mirror) (b) obsessive symptoms (e.g., continual repetition of simple cleaning activities) (c) anxiety symptoms, agitation, and previously nonexistent violent behavior (d) cognitive abulia (i.e., loss of willpower because they cannot dwell on a thought long enough to determine a purposeful course of action).

7
Very severe cognitive decline All verbal abilities are lost.
Frequently, there is no speech ability at all; only grunting remains.
Patients have urinary incontinence and require assistance with toileting and eating. They lose psychomotor skills (e.g., the ability to walk). The brain may find it difficult to tell the body what to do. Generalized cortical and focal neurologic signs and symptoms are frequently present. Variables included in this model were as follows: older age (>50 years), low GCS (≤12), shock, no use of hyperbaric oxygen therapy, creatine kinase (>320 U/L), hypertension, and serum bicarbonate (≤19.6 mmol/L). ROC = receiver operating characteristic; AUC = area under the ROC curve; GDS = Global Deterioration Scale