Headache and drowsiness in a 22 year old studentBMJ 2008; 337 doi: https://doi.org/10.1136/bmj.a1481 (Published 24 September 2008) Cite this as: BMJ 2008;337:a1481
- Hon Man Jor, fixed term specialty training appointment in core medical training1,
- Harry R Dalton, consultant gastroenterologist and honorary senior lecturer12
- 1Cornwall Gastrointestinal Unit, Royal Cornwall Hospital, Truro TR1 3LJ
- 2Peninsula College of Medicine and Dentistry, Royal Cornwall Hospital
- Correspondence to: H R Dalton
A 22 year old female medical student presented as an emergency with headache and drowsiness. She shared a flat with three other students, all of whom were well, and who found her in bed after she failed to turn up for lectures. She had no relevant medical history and was well the day before. The only drug she takes is the oral contraceptive pill.
On examination she was flushed and drowsy, afebrile, and had no skin rash. Her pulse was 100 (sinus rhythm), blood pressure was 120/80 mm Hg, the Glasgow coma score was 11/15 (localising pain, inappropriate speech, eye opening in response to speech), and pulse oximetry showed an oxygen saturation on air of 98%. Her blood gases were as follows: arterial blood gas on air pH 7.35 (normal range 7.35-7.45), partial pressure of oxygen (pO2) 8.2 kPa (>10.6 kPa), partial pressure of carbon dioxide (pCO2) 5.8 kPa (4.7-6 kPa), bicarbonate ions 24 mmol/l (24-28 mmol/l). The rest of the examination, including a detailed neurological examination and fundoscopy, was normal.
1. What is the diagnosis?
2. How is the diagnosis confirmed?
3. What is the treatment?
1. Carbon monoxide poisoning. This is suggested by the clinical picture and the mismatch between the normal oxygen saturation of 98% on air (measured by pulse oximetry) and the hypoxia evident in the blood gases (pO2 of 8.2 kPa on air).
2. By measuring carboxyhaemoglobin.
3. 100% oxygen. Hyperbaric oxygen is used for selected severe cases.
Around 50 people in the UK die each year from carbon monoxide poisoning, and about 10% of these deaths are accidental.1 The colourless and odourless carbon monoxide gas forms when carbonaceous fuels burn in limited oxygen. Most cases of accidental carbon monoxide poisoning are caused by inadequate ventilation or poor maintenance of gas appliances. Malfunctioning gas appliances (suggested by a yellow or orange tip of the flame) can emit up to five times more carbon monoxide than properly maintained equipment (blue flame).2 In the United States, 46% of fatal cases of carbon monoxide poisoning are non-accidental (mainly from car exhaust fumes), and 28% involve smoke inhalation in fires.3
Carbon monoxide has 250 times the affinity of oxygen for the iron in haemoglobin and when it binds it makes haemoglobin unavailable for oxygen carriage. It distorts the structure of the haemoglobin, which makes it bind more tightly to oxygen; this shifts the oxygen dissociation curve to the left and reduces the release of oxygen molecules to the tissues.4 Furthermore, carbon monoxide binds to cytochrome c oxidase in the mitochondrial membrane, thereby inhibiting oxidative metabolism and causing “chemical asphyxiation” at the cellular level.5
The key to the diagnosis in this case is the mismatch between the normal oxygen saturation of 98% on air and the hypoxia evident in the blood gases (pO2 of 8.2 kPa on air). Oxygen saturation is commonly measured by pulse oximetry. The pulse oximeter uses two light emitting diodes that emit pulses of red (wavelength of 660 nm) and infrared (wavelength of 940 nm) light every five to 10 microseconds.6 Deoxyhaemoglobin absorbs more infrared light, whereas oxyhaemoglobin absorbs more red light. The measurement of oxygen saturation relies on the difference in the optical absorbance of oxyhaemoglobin and deoxyhaemoglobin. Carboxyhaemoglobin has a similar absorption coefficient to oxyhaemoglobin, so oxygen saturation as assessed by pulse oximetry is the sum of the carboxyhaemoglobin and oxyhaemoglobin readings. Thus, in patients with carbon monoxide poisoning pulse oximetry gives falsely high readings for oxygen saturation.7
2. Confirming the diagnosis
The definitive diagnostic test for carbon monoxide poisoning is to measure carboxyhaemoglobin with a blood gas analyser that incorporates a co-oximeter. Unlike the pulse oximeter, which uses only two wavelengths, the co-oximeter measures light absorbance at six or more discrete wavelengths and it directly measures the various forms of haemoglobin.8 Normal carbon monoxide values are 1-3% in non-smokers and up to 15% in heavy smokers.9 Carboxyhaemoglobin values decay with time. The use of the Clark nomogram allows back calculation of the likely estimated peak carboxyhaemoglobin value.10
Symptoms of carbon monoxide poisoning (box) usually occur when carbon monoxide reaches 10-30%. However, the severity of the clinical symptoms does not correlate with the carbon monoxide value.11 Recovery time should therefore be based on the return of normal neurological function rather than normal carbon monoxide values.
Symptoms of carbon monoxide poisoning
Mild to moderate symptoms
Abnormal neurological signs
Loss of consciousness
Carbon monoxide poisoning is treated by giving the patient 100% oxygen as soon as possible. This can be delivered via a tight fitting non-rebreather mask,12 continuous positive airway pressure mask, or endotracheal tube.13 The high flow oxygen increases the oxygen content of the blood by maximising the fraction dissolved in the plasma. This provides oxygen directly to the underoxygenated tissues, and the rate of eliminating carbon monoxide is increased as well. The half life of carboxyhaemoglobin is four to six hours when patients breath normal air. This is reduced to 40-80 minutes when 100% oxygen is given and even further to 15-30 minutes when hyperbaric oxygen (100% oxygen at 2.5 atmospheric pressure) is given.14 At least six hours of 100% oxygen is recommended for mild to moderate carbon monoxide poisoning.15
Hyperbaric oxygen may be used (if a hyperbaric chamber is available) within six hours if carboxyhaemoglobin is greater than 25%, or if the patient is pregnant or has serious neurological or cardiovascular symptoms. However, little evidence is available to show that hyperbaric oxygen is better than normobaric oxygen in terms of clinical outcome and prognosis.16 12
Cite this as: BMJ 2008;337: a1481
Competing interests: None declared.
Patient consent not required (patient anonymised, dead, or hypothetical).