Acute Health Effects in a Community After a Release of Hydrofluoric Acid

JESSIE S. WING, M.D.*
LEE M. SANDERSON, Ph.D.*
Office of Surveillance and Analysis
Center for Chronic Disease Prevention
and Health Promotion
Centers for Disease Control
Atlanta, Georgia

JEAN D. BRENDER, Ph.D.
DENNIS M. PERROTTA, Ph.D.
RICHARD A. BEAUCHAMP, M.D.
Epidemiology Division
Texas Department of Health
Austin, Texas

*Drs. Wing and Sanderson were formerly with the Center for Environmental Health and Injury Control, Environmental Hazards and Health Effects Division, Centers for Disease Control, Atlanta, Georgia.

ABSTRACT. Approximately 3 000 persons were evacuated from a Texas community after 24 036 kg (53 000 lb) of caustic hydrofluoric acid (HF) were released from a nearby petrochemical plant. Emergency room and hospital records of 939 persons who were seen at two area hospitals were reviewed. Most persons who presented at the emergency rooms were female (56%) or black (60%), and their mean age was 33.9 y. The most frequently reported symptoms were eye irritation (41.5%), burning throat (21 %), headache (20.6%), and shortness of breath (19.4%). Physical examination results were normal for 49% of the cases; however, irritation of the eyes, nose, throat, skin, and lungs were noted on other exams. Decreased pulmonary function was demonstrated by pulmonary function tests (forced expiratory volume in the first second, < 80% of predicted value, 42.3%); hypoxemia (pO2 < 80 mm Hg, 17.4%) and hypocalcemia (< 8.5 mg/dI, 16.3%) were also noted. Ninety-four (10%) of the cases were hospitalized, and more than 83% of all cases were discharged with a primary diagnosis of "HF exposure." There are several reports of individuals who are acutely and chronically exposed to HF; however, we are unaware of other published reports that describe exposure of a community to HF. This incident represented a unique opportunity to study the immediate health impact on a community of residents who were exposed to a hazardous materials release. Results of this analysis suggest that (a) initial health problems should be followed up, (b) any long-term health effects of HF exposure must be assessed, and (c) the health impact on the population at risk should be determined.

HYDROFLUORIC ACID (HF) is a caustic, industrial compound used in the production of high-octane gasoline, aerosols, plastics, and refrigerants. It has been used in glass etching since the seventeenth century. (1)

In October 1987, a release of HF occurred at a petrochemical plant in Texas. Two pipes on a storage tank that contained isobutane-saturated HF were sheared during maintenance operations, and gas spewed from the tanks. Within 7 min of the release, a water curtain was established to contain and rehydrate escaping gas. Nonetheless, at approximately 5:20 P.M. on the night of the local homecoming celebration, a vapor cloud drifted across a community of 41 000 residents. Approximately 24 036 kg (53 000 lb) of anhydrous HF and 2 993 kg (6 600 lb) of isobutane were ventilated during a 48-h period. Within 20 min of the release, persons within .8 km (.5 mi) of the plant were evacuated. Eventually, 3 000 individuals were evacuated from an 8 km2 (5 mi2) area.

Methods

Representatives of the Texas Department of Health and the Centers for Disease Control visited the area 19 d after the HF release to review the medical records of 939 persons who reported being exposed to the chemical release.

The objectives of our study were to (a) describe the medical characteristics of the 939 cases and (b) to compare cases who were hospitalized with those who were treated and released from the emergency rooms (ERs). We defined a case as a person who, within 24 d following the HF release, came to hospital A or B and expressed concern about HF exposure, or whose medical records contained information regarding symptoms, signs, or diagnoses suggestive of HF exposure.

We completed an abstraction form for the 94 cases who were hospitalized at hospital A or B. Demographic information, presenting symptoms and signs, medical history, results of physical examination and laboratory tests, treatment, and discharge diagnoses were recorded. We used a modified form to abstract information for the 845 cases who were not hospitalized and who were treated and released from either of the two ERs. All data were transferred to dBase III files, which were then analyzed with the Statistical Analytical System. In these analyses, only data obtained from a person's first hospitalization or ER visit were included. (Thirty-four persons visited the ER multiple times, and 5 were hospitalized more than once.)

Results

The petrochemical plant is located in the southeastern area of the community and is .4 km (.25 mi) from residential areas. Most of the HF was ventilated during the first 2 hr of the release in a low-lying plume across the city in a southeast to a northwest direction.

Samples taken downwind 1 h after the release contained 10 ppm HF; and minimal traces of HF were found in samples obtained 2 h after the release. The following day, vegetation and exterior surfaces located within 3.2 km (2 mi) northwest of the plant were damaged; i.e., foliage was brown, car windows were etched, and car paint was streaked. (2)

Hospital A, the smaller of the two hospitals in our study, is located 2.4 km (1.5 mi) from the plant. It was visited by 67.4% of all the cases. This hospital (a 120-bed facility) instituted a disaster plan on the night of the release to reinforce its medical staff and supplies. "Disaster tags" were provided to the first 231 cases (after which supplies were exhausted) who presented to Hospital A. Location and activity of a case at the presumed time of exposure to HF were obtained from information recorded on the disaster tags.

Cases who were hospitalized were generally older than nonhospitalized cases (Table 1, mean age 41.5 y versus 33.1 y, respectively), We compared the cases with the 1980 U.S. Census results for the city, and we found that more blacks (2-fold increase), females (1.3-fold increase), and older persons (median age 33.9 y versus 28.9 y) were represented in our study.

The range of time during which cases presented to the ERs after HF exposure is presented in Figure 1. On the day of the release (day 1), 44% of cases who presented to the ERs were hospitalized; thereafter, the proportion of cases who were hospitalized decreased. However, most cases who were not hospitalized visited ERs on day 3 (27%) or day 4 (21 %).

The majority (82.7%) of cases who were given disaster tags were not occupationally exposed at the time of the release; they were either at their homes or in their neighborhoods. There were 39 occupationally exposed cases (i.e., police personnel, plant employees, evacuation personnel) who were identified by information on their disaster tag; however, most (72%) were not hospitalized during this visit.

Symptoms. Cases who were not hospitalized and who had no medical complaints on presentation to the ER numbered 120 (14.2%), whereas only 1 (1.1 %) case who was hospitalized had no complaints (p < .001). Most of these nonhospitalized cases who voiced no complaints presented to the ERs either on day 1 (n = 60, 7%) or day 4 (n = 22, 2.6%) of the episode.

The symptoms reported most frequently among all cases were eye irritation, throat burning, headache, shortness of breath, and throat soreness (Table 2). Hospitalized cases reported more headache (p < .05), shortness of breath, cough, nausea, vomiting, dizziness, and skin burning than did nonhospitalized cases (p < .001).

Physical examination. Physical examination results were normal for 49% of the cases; however, conjunctival and oropharyngeal irritation, pulmonary findings (i.e., rhonchi, rales, wheezing, or stridor), and dermal irritation (i.e., skin blisters, rash) were noted during other exams (Table 3). Erythema was relatively frequent (including erythematous tonsils, n = 7). There were also other cases with papules, tracheal ulcers, a hypertensive episode (240/120 mm Hg), hyperventilation, and interuterine pregnancy.

Pulmonary function tests. Forced expiratory volume in 1 s (FEV1.0) was less than 80% of the predicted values for 30.2% and 53.7% of the nonhospitalized and hospitalized cases, respectively (p = .01) (Table 4). The ratio of FEV1.0 to forced vital capacity (FVQ was low for 3.2% of the nonhospitalized cases compared with 29.9% of hospitalized cases (p < .001).

When stratified by exposure type, although only 10 of the occupationally exposed cases performed pulmonary function tests, results for 70% (n = 7) were in the < 80% predicted range compared with 40% (n = 48) who were not occupationally exposed.

Other laboratory results. The criteria employed for obtaining and analyzing laboratory specimens were unclear, especially for cases who were not hospitalized. Arterial blood gas (ABG) determinations, which were made for 155 cases, revealed that 17.4% were hypoxemic (pO2 < 80 mm Hg). More hospitalized cases (24%) experienced alkalotic pH levels (pH 7.46-7.62) than did nonhospitalized cases (10%, p < .001).

Discharge diagnoses. Eighty-three percent of nonhospitalized cases had a primary ER diagnosis that was related to HF exposure, compared with 94% of the hospitalized cases. Other diagnoses among the hospitalized cases included asthma/bronchospasm, bronchitis, respiratory distress, congestive heart failure/pulmonary edema, pneumonia, tracheobronchitis, atelectasis, and esophagogastritis. More than half of these cases were hospitalized for 1 d or less (range < 1-12 d).

Discussion

Hydrofluoric acid, when present as an acid mist, can extensively irritate the eyes and respiratory tract, which can result in intense lacrimation, sore throat, cough, lower airway inflammation, and possible pulmonary edema-an effect that can be manifested several hours later. The compound can cause topical and systemic toxicity, the signs of which may only be apparent 24 h after initial contact. (4) Hydrofluoric acid can also penetrate the skin and hydrolyze; subcutaneous tissue can be damaged and deep, painful ulcerations may occur. (5-7) Acute toxicity may affect the gastrointestinal tract (diarrhea), liver (direct hepatotoxicity), and kidneys (acute renal failure). Metabolic changes may occur, which are usually secondary to hypocalcernia, hypomagnesemia, and their sequelae. (8) Systemic fluorosis, myocardial injury, and death have been reported after severe HF burns. (9-11)

The differences in demographic characteristics of our cases may have resulted from variations in geographic distribution of housing in the city and proximity to industrial areas, differential use of medical facilities, demographic changes since the 1980 Census, (12,13) social and economic variations, true exposure and health effect differences, or other artifacts. The local homecoming celebration, which occurred on the same day as did the release, may have also influenced community activity and exposure.

Most cases who came to the hospitals soon after the release probably experienced airborne exposure and inhalation, both of which would result from the HF "fog." The spectrum of symptoms reported included the,expected complaints of irritation to the eyes, nose, throat, and lungs.' Pulmonary findings were prominent, especially among cases who were hospitalized. Complaints of headache, dizziness, nausea, and vomiting were also expressed frequently. (5)

The initial arterial blood gas determinations revealed that 17% of all cases tested were hypoxemic. Perhaps this represents intrapulmonary shunting of blood after HF inhalation, a finding that has been reported previously for persons who inhale other acid fumes. (4) Intrapulmonary shunting and hypoxemia may have contributed to symptoms of headache and dizziness reported by some of our cases. An impressive number of hospitalized cases also had alkalosis (pH to 7.62), but there was no strong association between alkalosis, serum calcium levels, or tachypnea.

For many, FEV1.0 values were low. Perhaps this is a health effect or a dose response for cases who were occupationally exposed to the HF release. The FEV1.0 values, which were worse for hospitalized cases, may be a predictor of hospitalization. Or, they may have resulted from personnel technique, possible malfunctioning spirometry equipment, testing of patients en masse in emergency settings, or another artifact. (15,16)

One reported side effect of HF exposure is hypocalcemia. Most patients tested were eucalcemic; however, 16.3% were mildly hypocalcemic (7.5-8.4 mg/di). There was very limited information available on fluoride levels among our cases. A urine fluoride level (1.7 mg/1) reported for one hospitalized case was not elevated. Urinary excretion of fluoride, which is determined easily and probably reflects the amount of chemical absorbed, is used as an index of fluoride exposure and has been related to osteofluorosis.(8, 14)

Historically, medical treatment for HF exposure has varied.(3, 17-20). In our study, medical personnel at hospitals A and B prescribed calcium gluconate or calcium chloride preparations for ocular, dermal, respiratory, and intravenous therapy; supplemental oxygen inhalation; and bronchodilator or steroid medications.

Fourteen percent of all nonhospitalized cases had no complaints upon presentation to the ERs. The time of presentation for the nonhospitalized cases was bimodal. These individuals may represent those who were concerned about HF exposure immediately after the release (day 1), those who presented on Monday (day 4, after the evacuation period ended), or those "worried well" who may have presented to the hospitals for other reasons.

There were anecdotal accounts of cases hospitalized soon after the incident for purposes of observation (and hospitalization periods were usually brief). At the time of the HF release, most physicians' offices and clinics were closed. This record review probably captured the most critically ill or the most extensively exposed cases immediately after the release; a few cases, however, may have been "lost" to medical centers in other communities.

This study also illustrates the pattern of health care utilization by a community after a toxic exposure. The disparity in time of presentation (Fig. 1) and presenting symptoms and signs suggest that early findings, particulary among hospitalized cases, may reflect airborne exposure to acid mist that causes mucous membrane and respiratory irritation (i.e., conjunctival and oropharyngeal irritation, dyspnea). Cases who presented later in the post-impact period may reflect continued HF exposure and/or nontreatment and symptoms and signs (i.e., burning, itchy skin, rashes) that may have occurred during the clean-up phase of the episode (especially among nonhospitalized cases). Many cases who were not hospitalized and who complained of skin rashes presented on day 4 (n - 6, 29%) or day 6 (n - 5, 24%) of the episode.

The data have several limitations, including the validity of self-reported historical information, incompleteness or under-reporting of some variables, and treating patients en masse in an emergency setting. The medical records from the hospitalized cases may be more complete than those of nonhospitalized cases, and the records may not necessarily reflect more health problems in the hospitalized group. However, after a toxic chemical release, more medical problems in a patient may also be a predictor of hospital admission.

Further studies could be conducted to (1) follow up cases for whom positive findings were noted (symptoms, physical findings, or laboratory results), (2) obtain additional information on occupationally exposed cases; (3) review pulmonary function test data (including original spirometry plots), (4) map cases for whom locations during exposure were known (home address may not be an adequate proxy because many were exposed outside their homes), and (5) consider an evaluation of the population at risk. Epidemiologic and medical evaluations of individuals who were potentially exposed but who did not present at these hospitals could be conducted to determine the full extent of the impact and possibly to justify the establishment of a registry for exposure and health effects. The emergency response to this incident should be evaluated and coordinated among industry, community, civil defense, hospitals, and medical personnel, and methods to improve disaster preparedness and treatment protocols should be addressed.

* * * * *

The authors thank Drs. Ralph Morris and Ed lbert, Ms. Sue Broussard, Ms. Toni Harris, and staff at hospitals A and B for their cooperation and support.

Submitted for publication March 21, 1990; revised; accepted for publication December 27, 1990.

Requests for reprints should be sent to: J. S. Wing, M.D., Office of Surveillance and Analysis, Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control, 1600 Clifton Road, MS K-30, Atlanta, GA 30333.

* * * * *

References

1 . Halton DM, Dranitsaris P, Baynes CJ. Toxicity levels to humans during acute exposure to hydrogen fluoride. Ottawa, Canada: Atomic Energy Control Board, 1984; INFO-0143.

2. lbert ER. Anhydrous hydrofluoricle acid vapor release (Report). La Marque, TX: Galveston County Health District, 1987.

3. Shewmake SW, Anderson BG. Hydrofluoride acid burns: a report of a case and review of the literature. Arch Dermatol 1979;115: 593-96,

4. Speizer F. Harrison's principles of internal medicine, 11th ed. New York: McGraw-Hill, 1987, pp. 1073-75.

5. Waldbott GL, Lee JR. Toxicity from repeated low-grade exposure to hydrogen fluoride-case report. Clin Toxicol 1978;13: 391-402.

6. Burke WJ, Hoegg UR, Phillips RE. Systemic fluoride poisoning resulting from a fluoride skin burn. J Occup Med 1973;15(l): 39-41.

7. Rom WN, Barkman H. Respiratory irritants. Environmental Occup Dis 1983;273-83.

8. Braun J, Stob H, Zober A. Intoxication following the inhalation of hydrogen fluoride. Arch Toxicol 1984;56:50-54.

9. Mayer TG, Gross PL. Fatal systemic fluoride due to hydrofluoric acid burns. Ann Emer Med 1985;14:149-53.

10. Tepperman PB. Fatality due to acute systemic fluoride poisoning following a hydrofluoric acid skin burn. J Occup Med 1980;22 (10):691-92.

11. Menchel SM, Dunn WA. Hydrofluoric acid poisoning. Am J Forensic Med Pathol 1984;5(3):245-48.

12. U.S. Department of Commerce Bureau of the Census. Statistical abstract of U.S. 0988): national data book and guide to sources, 108th ed. Washington, DC: U.S. Government Printing Office, 1988.

13. U.S. Department of Commerce Bureau of the Census: State & metropolitan area data book: 1986. Washington, DC: U.S. Government Printing Office, 1987.

14. Hogstedt C. Fluorides. In: Aitio A, Riihimaki V, Vainio H, Eds, Biological monitoring and surveillance of workers exposed to chemicals. Washington, DC: Hemisphere, 1984; pp. 177 86.

15. Knudson Rj, Slatin RC, Lebowitz MD, Burrows B. The maximal expiratory flow-volume curve: normal standards, variability, and effects of age. Am Rev Respir Dis 1976; 113:587-600.

16. Gardner RM. Standardization of spirometry: a summary of recommendations from the American Thoracic Society. Ann Intern Med 1988;108:217-20.

17. Trevino MA, Herrmann GH, Sprout WL. Treatment of severe hydrofluoric acid exposures. J Occup Med 1983;25:861-63.

18. Vance MV, Curry SC, Kunkel DB, Ryan Pj, Ruggeri SB. Digital hydrofluoric acid burns: treatment with intraarterial calcium infusion. Ann Emer Med 1986; 15(8):59-65.

19, Haggerty MA, Soto-Greene M, Reichman LB. Caring for victims of toxic gas inhalation. J Crit Illness 1987;2(7):77-87.

20. Carney SA, Hall M, Lawrence JC, Ricketts CR. Rationale of the treatment for hydrofluoric acid burns. Br j Ind Med 1974;31: 317-21.

To learn more about fluoride air pollution, see www.fluoridealert.org/f-pollution.htm