resurfaced for 10 minutes every hour; during games, the ice was resurfaced for 5 minutes after each 15-minute period. A Plexiglas shield surrounded the ice to protect spectators from airborne hockey pucks. Two ventilation systems were used in the arena: air vents for passive air exchange and exhaust fans.

The ice resurfacer was powered by an internal combustion engine using propane fuel. If these engines are not properly tuned and the fuel mixture in the carburetor receives too little oxygen, elevated levels of carbon monoxide may be produced; if the mixture has too much oxygen, elevated levels of nitrogen dioxide may be produced.¹⁶

Ice arenas in Minnesota are required to measure ambient levels of carbon monoxide and nitrogen dioxide on a weekly basis during the months of operation. Measurements are taken 120 cm above the ice in the center of the arena. Ambient levels of nitrogen dioxide above 0.5 ppm are considered to be elevated and are required to be reported to the Minnesota Department of Health.

Questionnaires were administered to all hockey team members who attended the two games. In addition, cheerleaders and band members, who were present during the second game, were interviewed. Information was obtained on symptoms (including cough, hemoptysis, shortness of breath, dyspnea, chest pain, headache, and weakness), onset and duration of each symptom, general health status (including history of asthma or other respiratory problems), length of time in the arena, and location in the arena during the games (in the stands or on the ice). For hockey players, information also was obtained on position played and length of time on the ice. All interviews were completed within 10 days after attending a game at the arena. A case was defined as acute onset of cough, hemoptysis, or dyspnea during a hockey game or within 48 hours of attending a hockey game at the arena. Attack rates for teams were compared using standard univariate analysis.¹⁷

Spirometry was performed within 10 days of exposure and again at 2 months after exposure for all members from two hockey teams: team C (with a single exposure) and team D (with multiple exposures). Spirometry also was performed on members of a basketball team from one of the schools, which served as an unexposed group for comparison. Pulmonary function testing was performed at the high schools that the players attended using a portable spirometer (Microloop, Medical Graphics Corp, St Paul, Minn). The best result of three attempts was recorded. Intermountain Thoracic Society predicted values (which control for age, height, and weight) were used to determine results by percent of predicted.¹⁸

We reviewed medical records for hockey players who reported seeing a physician. Information was obtained on physical examination, chest roentgenogram findings, and treatment prescribed during initial and follow-up clinic visits.

Air quality records at the ice arena were reviewed for the hockey season. No measurements had been obtained during the two games in question. Therefore, to simulate conditions during the games, the ice resurfacer was operated for 30 minutes and levels of nitrogen dioxide and carbon monoxide in the arena were measured. The use of the ventilation systems during the two games also was reviewed.

To evaluate air quality monitoring in indoor ice arenas nationally, and to obtain an estimate of the number of ice arenas located in each state, a telephone survey of all 50 state health departments was conducted. The Ice Skating Institute of America, the US Figure Skating Association, and the National Hockey Association also were contacted to obtain estimates of the number of indoor ice arenas located in the United States.

Questionnaires were completed on 92 (94%) of 98 hockey players with a single exposure (teams A, B, and C), 34 (100%) of 34 players with multiple exposures (team D), 16 (76%) of 21 cheerleaders, and 25 (96%) of 26 band members. Overall, 116 cases were identified.

Symptoms reported by at least 30% of the 69 case hockey players who had a single exposure are listed in Table 1. A typical case was characterized by acute onset of cough and dyspnea within 1 hour of playing a game at the arena. At the time of onset, the cough was frequently so severe that players had difficulty driving home after the game. The mean duration of cough was 16 days in players with acute exposure. The dyspnea was described most often as “aching lungs” or “a tightness in the chest” that made it difficult to inhale deeply. Hemoptysis was characterized by blood-tinged sputum. Similar symptoms were noted among players on team

D. However, because many of the players on team D complained of chronic cough, they were not included as acute cases. The mean duration of cough for players on team D was 41 days. Eighteen (14%) of 126 hockey players reported a history of reactive airway disease (asthma). Of these, 16 (89%) reported an exacerbation of their asthma symptoms after playing at the arena.

Attack rates for the groups are listed in Table 2. Although the attack rate for acute onset of symptoms for members of team D was only 56%, 11 (73%) of 15 players on team D who did not have acute onset of symptoms admitted to chronic respiratory symptoms (primarily cough). The attack rates for cheerleaders (who were on the ice) and band members (who sat in the stands) were similar to the attack rates for hockey players on teams A, B, and C. However, hockey players and cheerleaders were 3.2 times as likely as band members to develop hemoptysis (P=.05, Mantel-Haenszel χ² test). Length of time spent in the arena, length of time spent on the ice, and position played did not substantially increase the risk of developing hemoptysis.

Results of initial and follow-up spirometry performed on team C (single exposure), team D (multiple exposures), and the unexposed comparison group of basketball players are shown in Table 3. Overall, no differences in five lung function parameters at initial testing or at 2 months’ follow-up (when comparing percent of predicted) were noted between the two hockey teams and the basketball comparison group.

Ninety-two hockey players sought medical attention; abnormal findings and treatment prescribed at the initial clinic visit are shown in Table 4. Ten patients had follow-up physician visits; none had ongoing signs or symptoms noted.

Mechanics at the ice arena reported that the ice resurfacer had not been running properly during the preceding 6 months and that it had been emitting