CARBON MONOXIDE INTOXICATION. THREE YEARS' EXPERIENCE
OF THE FIRST IN-HOSPITAL MULTIPLACE HYPERBARIC CHAMBER IN BELGIUM

P. Germonpré, M.D., C. Jeuneau, A. Van Renterghem, M.D.
Center for Hyperbaric Oxygen Therapy, Military Hospital "Queen Astrid"

Abstract

Operational since April 2nd, 1991, the Center for Hyperbaric Oxygen Therapy at the Military Hospital in Brussels has been the first hospital-based multiplace hyperbaric chamber in Belgium. Its activity has been increasing constantly, and at this moment some 1100 treatments are performed annually. Referrals are primarily from the Brussels region, although patients have been referred from as far as Luxemburg. Data will be presented on the 272 carbon monoxyde intoxicated patients, treated between April 2nd 1991 and September 30th, 1994. Epidemiological data, treatment policy and referral procedure will be discussed, as will be the role of the Center in the achievement of secondary prevention of carbon monoxyde intoxication.

Introduction

Carbon monoxyde (CO) intoxication is still frequent in Belgium. Annually, about 10.000 cases are detected, of whom 10% are hospitalized for prolonged normobaric oxygenotherapy or hyperbaric oxygen therapy. The prevalence of CO intoxication is geographically determined: at least four major intoxication areas can be distinguished: at the south - the Hainaut region; in central Belgium - the Brussels City center; at the north - around Antwerp and Liege. Not surprisingly, these areas are caracterized by the presence of large areas of old, often pauperized housings, often inhabited by immigrant populations. Before 1991, each of these "at risk" areas disposed of one or two monoplace hyperbaric treatment facilities. Annual treatment rates varied between 80 and 250 CO intoxication cases per HBO center.

In 1991, the monoplace hyperbaric chamber at one of the Brussels university hospitals broke down and was not replaced for various reasons. Around the same time, the Military Hospital, located in roughly the same area of the Brussels Region, took in use a multiplace hyperbaric chamber, and has been asked to take over the HBO treatment of CO intoxicated patients from that university hospital. Owing to the obvious advantages that a multiplace hyperbaric chamber offers, several other hospitals started referring patients to our hyperbaric facility. Today, annually some 100 CO intoxicated patients are treated, which represents almost 20% of the "per annum" hyperbaric treated CO intoxications in Belgium. The Military Hospital does not serve as a primary civil healthcare facility. Its Emergency Department is destined for intake medical examinations for military personnel. The Military Hospital does however operate a EMS team that takes part in the advanced life support service in a specified zone of the Brussels region. Furthermore, the Military Hospital Burn Center is open for civil patients, and is one of the most important burn care facilities in Belgium.

General demographic data:

There has been a progressive increase in referrals for CO intoxication: 29 patients in 1991 (9 months), 52 (1992), 110 (1993) and 119 in 1994, of which 82 (9 months) are discussed here (Total 272 cases). The monthly distribution figures show a "classical" peak incidence during the months of october/november and a second rise during march/april. This is an indicator of the importance of weather conditions in the pathogenesis of CO intoxication (cfr. infra). These figures also show a peak incidence in november, due to an unusually high number of cases in november 1993, during the second week of which an "epidemic" of CO intoxication (more than 50 cases in three days) was noted in the south and central part of Belgium, owing - again - to particularly defavorable weather conditions. Mean age of the patients was 27 years, with a proportion of children (< 15 years) of 27.9 %. Sex ratio was almost 1 (132 women for 139 men: 0.95). A considerable number of treated patients were immigrants: 46%. In this group were also the most collective intoxications.

Referrals:

A majority of patients have been referred from other hospitals (60.3%). Of these inter-hospital transfers, 41% were from Brussels hospitals, 40% from the Walloon Region, 19% from the Flanders Region, where less CO intoxications occur, and hyperbaric facilities are geographically more evenly distributed.

24.3% were directly transferred to our hospital by EMS teams; 6.3% were referred by our own EMS team - in these cases more often secondary to fire exposure (35%) - and 8.8% presented themselves (with own transportation) at our Emergency Department. Causes (table 3)

Causes:

The cause of intoxication could be determined in all cases. Gas water heaters (36.8%) are the most frequent cause, followed closely by coal stoves (27.6%). Overall, there is an almost 50% distribution between water and room heating appliances. 10.7% were caused by fire. This high figure is explained by the presence of a large burn center in our hospital, and a systematic blood sample for HbCO in symptomatic fire victims, even if not burned. Other causes (6.6%) included car or explosion motor exhaust fumes and more rare causes such as charcoal burning in an inside (!) barbeque and polluted heating air in a large retail store.

The duration of exposition was estimated as SHORT (< 1 hr): n=135 (49.6%), MEDIUM (1-6 hrs): n=75 (27.6%) or LONG (> 6 hrs): 60 (22.1%) (2 missing cases).

Symptoms and signs:

133 patients (49%) had loss of consciousness as the most important symptom, of which 27 (20.3%) had still impairment at arrival at our hospital. 77 cases (28.3%) presented with headache as the main complaint, accompanied with nausea or vomiting (n=13, 15%), marked muscular weakness (inability to walk) (n=19, 22%) or vertigo (n=30, 35%). There were 20 cases (7.3%) in which cardiological symptoms were present, of which 6 occurred without loss of consciousness nor neurologic impairment.

Of the SHORT and MEDIUM exposures, 119 (56.6%) presented only minor symptoms (headache, fatigue) upon admission at our hospital. The mean initial HbCO level in these patients was 24.7% (SD 12.3%), taken 1.18 hrs (SD 0.9 hrs) after discovery of the intoxication. The patients that still presented 'major' symptoms (n=91) had a mean HbCO level of 27.8% (SD 14.8%), taken 1.32 hrs (SD 2.08 hrs) after discovery. Of the LONG exposures, 31 (51.6%) had still obvious neurologic impairment when arriving at our hospital. The initial HbCO level in these patients was lower (26.5% (SD 9.8%), taken 2.1 hrs (SD 3.3 hrs) after discovery) than the same group of patients with SHORT or MEDIUM exposures. The mean time before arrival at the Military Hospital was equal (3.1 vs 3.6 hrs) for both groups.

6 pregnant women (2.2%) were treated, of whom 3 had serious symptoms. In one of these the foetus (8 months) was dead upon discovery of the (comatose) patient, and the patient died the next day. The duration of exposition was estimated to have been 8 hours, and the initial HbCO level was 70%. The other patients (mean HbCO level: 21.7%) did well and no gross neurologic anomalies were detected at the birth (one patient lost to follow-up).

16 patients had clinically evident smoke inhalation (5.8%), as judged by the presence of soot on tongue and nostrils, moderate to severe respiratory discomfort, and blackish expectorations; of those, 12 (75%) were discharged 24hrs after HBOT. Of the remaining patients, only one had to be intubated for respiratory distress syndrome.

In 19 patients (7%), a minor intoxication was present with no serious symptoms or signs. Of those, 10 were treated with HBOT because of high (>20%) HbCO levels; the others were treated with normobaric oxygen.

Treatment and outcome:

248 patients (91.2%) were treated with HBOT, according to a standard protocol of 2.5 ATA oxygen, 90 minutes. 23 patients (8.5%) were treated with 100% normobaric oxygen for 6 hours or more, or with 35% oxygen (1 patient with COPD and arterial pCO2 of 70mmHg).

Of these 22 patients, 9 were not treated because of minor intoxication, 5 because of intolerance of HBOT: 2 middle ear barotrauma (0.7%), 3 because of claustrophobia or agitation (1.1%), 7 because of known contra-indications for HBOT (tubal dysfunction, active bronchitis, sinusitis) or refusal (4). 8 of these normobaric oxygen cases were children.

Upon completion of HBOT, 247 patients were re-transferred to the referring hospital or sent home after a short observation period (<6 hours), with instructions for secondary prevention (90.8%). 22 patients (8.1%) were hospitalized at the ICU, and 2 died (0.7%). Final outcome is not known in the majority of cases. We do not have reliable data on the occurence of delayed neurologic sequelae, since most of the patients are lost to follow-up after transfer to the referring hospital.

Discussion

1. Number of treatments and referrals:

There has been a logical increase of CO intoxicated patients during the first three years of operation, that reflects the overall increase in HBO treatments performed at our Center. The actual rate represents approximately 2/3 of the HBO treatments for CO intoxication in the Brussels Region (the remaining being treated at two hospitals with monoplace chamber), and 20% of the annual HBO treatments for CO intoxication in Belgium. Moreover, there has been a geographical spreading of referring institutions (hospitals/EMS). During 1991, 82.8% (24 cases) of the referrals were from the Brussels Region, whereas this represents only 45.7% in 1994 (37 cases). Although no exact data on this is available, the number of comatose and intubated/ventilated CO intoxicated patients has also increased: presumably these were treated locally with normobaric oxygen before.

2. Treatment and referral selection:

Referrals are generally accepted after telephone contact between the referring doctor and the EMS doctor or anaesthetist on call at the Military Hospital. Decision for referral is taken on the basis of clinical presentation and HbCO level. In cases where no loss of consciousness has occurred, only mild symptoms are present and transfer time is estimated to be more than 1 hour, initial normobaric oxygen treatment at the local hospital is encouraged. HBOT is then reserved for those cases that have not fully recovered after 1 hour of 100% oxygen via non-rebreather mask. HBO treatment criteria currently applied are listed in table 6. The last three criteria (important symptoms after >1hr NBOT, symptomatic children <12yrs of age, and HbCO level > 20%) are "relative", i.e. the individual case is considered with regard to symptoms, duration of exposure, estimated transfer time. Referral for HBOT is accepted even after a > 6 hours treatment delay in case of marked symptoms.

Because the Military Hospital's Hyperbaric Center serves only as a referral institution and not as a primary health care facility, our proportion of HBO treated cases (vs. normobaric oxygen) is high (12/1). For comparison, at a university hospital in Central Brussels, this proportion would be approximately 1/1. The population that is being treated at our Hyperbaric Center is not representative for the CO intoxicated population in Belgium. Referrals are - whenever possible - selected via a telephone contact before transfer of the patient. Only 9 patients were not HBO treated because of minor indication (3.3%).

Patients are retransferred to the referring hospital. They are given a standard letter of discharge with recommendations for further examinations and therapy if needed, and for a clinical neurological control after 1 month. We currently do not ask to be informed about the results.

3. Symptoms and severity of intoxication:

Although not statistically exploitable for the moment, we find that patients with SHORT or MEDIUM exposures to high environmental CO concentrations (e.g. water heating appliances) tend to be in a better general condition when arriving at our hospital, despite more frequent initial loss of consciousness (42% vs 38.3% for the LONG exposures) and higher initial HbCO levels (24.7% vs 22.1% for the LONG exposures). LONG exposures on the contrary, result more frequently in prolonged impairment of consciousness or other major symptoms (cardiac, muscular weakness, severe headache, vomiting), even in cases where no initial loss of consciousness could be documented (8/31). It is likely that patients of the first group are suffering from hypoxic hypoxemic symptoms mainly, caused by high HbCO levels, and recover rapidly upon administration of normobaric oxygen, whereas patients from the second group present an intoxication that is more of a "tissular" nature, cause by a slowly progressive impregnation of intracellular haem groups and electron tranfer chain enzymes by CO, more or less independent from the HbCO levels. Although we were not able to show a statistically significant correlation between duration of exposure and short term or long term outcome (mainly due to the impossibility to obtain reliable follow-up data), we feel that the estimated duration of intoxication exposure has a major importance in the determination of the severity of the CO intoxication, and should be systematically noted in the patients' medical record.

4. Secondary prevention of CO intoxication:

For each patient, a detailed register is kept, including data on cause and duration of exposition. These data are communicated to the National Register of CO Intoxication. This register is kept by the National Poison Center in Brussels, and serves mainly as an epidemiological data base. Analysis of these data helped to detect a "new" cause of CO production during the epidemic in november 1993. It appeared that over 50% of the intoxicated patients during that period were rather youg people with a good standard of living, that occupied rather new housings. Moreover, these patients were almost all intoxicated by modern, coal burning, room heating devices. A detailed analysis, performed by the Poison center, included a meteorological review and a visit to these houses. It was found that these modern coal stoves were too powerful for the volume of the heated room, and were operated at their lowest during that week. Average temperature was 15°C, and a so-called "temperature inversion" was present, causing combustion gases to remain stagnant in the chimneys. As a result of this investigations, a nationwide prevention campaign has been organised, with frequent "CO-warnings" during radio and television weather forecasts, in case of defavorable weather conditions.

In approximately 30% of the cases, the patient declared experiencing difficulties in resolving the problem of CO production, either by lack of expertise or by a weak juridicial position in case of a rental housing. Recently we obtained the cooperation of the City Laboratory of Hygiene in performing, when needed, a free investigation at the site and delivering a legally valid report to the CO victim and to the owner of the house.

References:

• Mostin, M., Tissot, B.: National Poison Center Survey on CO Intoxication 1986 and 1992; PC Brussels, 1992
• Tissot, B., Perissino, A., Dechamps, P., Germonpré, P: CO intoxication: risk factor analysis and National Register. Presented at the "Symposium on CO Intoxication", Brussels, Dec 10, 1994.