Community exposure to sodium hydroxide in a public water supply.

Introduction and Background

The literature is replete with descriptions of injuries resulting from oral ingestion of sodium hydroxide (1-6). Most of the literature describes morbidity involved with ingestion or contact with cleaning products such as liquid and granular lye products. This paper discusses the adverse health effects experienced by a community after exposure to sodium hydroxide in a public water supply.

The incident occurred in a small city located in northeastern Texas. According to the 1990 census, the population of this city is approximately 12,000 (7). The city has one hospital. The municipal water supply consists of a conventional surface water treatment plant and a distribution system that supplies treated water to residents in the city and outlying rural areas. The treatment plant includes facilities for chemical feed, rapid mixing, flocculation, clarification, filtration, disinfection, and clearwell (finished-water) storage. One of the chemicals used in the water treatment process is liquid sodium hydroxide.

On the Sunday morning of June 13, 1993, several city residents arrived at the hospital emergency department complaining of burns resulting from contact with the municipal drinking water. The hospital staff notified city officials, who immediately initiated efforts to identify and resolve the problem. The water in the clearwell storage at the treatment plant and in various parts of the distribution system was analyzed and found to have a pH of approximately 12.0. Residents were warned about the problem through local radio and television announcements and were instructed to avoid contact with the water until further notice. Acceptable water conditions were reestablished at the treatment plant within several hours. Flushing and testing of the distribution system continued for two days, however, before pH levels of less than 10.0 were obtained throughout the distribution system and residents were notified that the water was safe.

TABLE 1

Prevalence of Symptoms in 248 Patients Exposed to Sodium Hydroxide
in a Public Water Supply

Site/Symptoms             Number of Patients       Percentage

Eye                                6                  2.4

Burning                            5                  2.0
Redness                            6                  2.4
Tearing                            1                  0.4

Mouth                             50                 20.2

Blistering                         3                  1.2
Burning                           34                 13.7
Irritated throat                  11                  4.4
Swollen lips                       2                  0.8

Gastrointestinal                  62                 25.0

Abdominal pain                    28                 11.3
"Burning stomach"                  8                  3.2
Diarrhea                          10                  4.0
Nausea                            28                 11.3
Vomiting                          21                  8.5

Skin                             135                 54.4

Blisters                           4                  1.6
Burning                           78                 31.5
Irritated/itching                 38                 15.3
Rash                              15                  6.0
Redness                           52                 21.0

Other                             29                 11.7

Chest burning/pain                 3                  1.2
Dizziness                          4                  1.6
Headache                           5                  2.0
Perineal irritation                6                  2.4
Weakness                           2                  0.8

The cause of the incident was investigated over a period of several days. The investigation included a thorough inspection of the chemical feed facilities as well as individual interviews with the plant staff. The investigation revealed that the chemical feeder normally used to deliver small volumes of sodium hydroxide to the treatment process was bypassed with a gravity feed line that ran between the bulk chemical storage tank and one of the clarifier effluent troughs. The gravity bypass line was opened by the operator who was on duty Saturday night. The operator was not properly certified and had only limited experience with water treatment processes. The line was left open for several hours while the treatment plant was off line, and approximately 3,000 gallons of sodium hydroxide flowed into the effluent trough of the clarifier. Early Sunday morning, a second operator noticed the open line and closed the bypass valve, but took no further action. When the treatment plant was restarted, the sodium hydroxide flowed through the filtration units and into the four clearwell storage tanks at the plant. The volume of treated water in the storage tanks was estimated to be 3 million gallons. The sodium hydroxide increased the pH of the treated water to 12.0. The highly alkaline water in the storage tanks was then pumped to the distribution system.

During the incident, residents were instructed to seek medical attention at the hospital emergency department if they had been exposed to the affected water. Persons who first went to their private physicians were also, for the most part, sent to the emergency department for examination and treatment. A follow-up health investigation of this incident was undertaken to determine what adverse health effects were experienced by persons who were exposed to high levels of sodium hydroxide in the municipal water supply.

Methods

Medical records from the hospital in the community were reviewed. Hospital staff identified, through the review of emergency department records, all patients who reported being exposed to the affected city water and who sought medical attention at the emergency department on the day of the incident and within the following week. The majority of patients (95.6 percent) were seen either the day of the incident or the following day. The investigators abstracted several kinds of information from these medical records, including demographic data, exposure information, reported symptoms, medical findings, and treatment given. Means, frequencies, and Chi square statistical analyses were calculated with BMDP statistical software and Epi Info (8,9). The Chi square test was used to assess linear trends in rates (10). Five months after the incident, follow-up questionnaires in English and Spanish (as needed) were sent to all patients identified from the medical records to obtain information about any ongoing health problems related to the incident.

Results

The records of 248 patients were identified as belonging to persons seen at the emergency department for sodium hydroxide exposure through the public water supply of these patients, 35 lived in nearby towns, and the rest were residents of the city. A significant decrease in injury rates was noted with increasing age (p [less than] 0.001). Among city residents, the highest rate of injury occurred among children less than 10 years of age (28 per 1,000), and the lowest rate occurred among residents 60 years of age or older (6.1 per 1,000). Female residents had a slightly higher injury rate (19.4 per 1,000) than male residents (14.7 per 1,000).

Of the 248 patient records identified, 222 records contained documentation on the route of exposure. The records indicated that 43.7 percent of the patients were exposed through ingestion, 44.6 percent through dermal exposure, and 11.7 percent through both routes. The types of symptoms reported by patients can be found in Table 1. Over half of the patients reported skin problems. Another 25 percent reported gastrointestinal problems, and 20 percent reported symptoms related to the mouth and throat. The most common symptom was burning of the skin (32 percent). Reported symptoms correlated closely with reported routes of exposure. The majority of patients with mouth and gastrointestinal symptoms reported exposure to the water through ingestion, while the majority of patients with skin complaints reported dermal exposure.

Approximately one quarter of the patients were diagnosed with dermal burns, the majority of which were first-degree burns (Table 2). Seven patients were documented as having both first- and second-degree burns. Table 2 shows the incidence rates per 1,000 city residents for medical diagnoses related to the incident (35 persons who lived outside the city limits excluded). The most common diagnosis was "caustic soda exposure" (10.1 per 1,000), followed by burns to the skin (4.2 per 1,000) and gastritis/enteritis (1.3 per 1,000). Of the 135 patients with documented skin symptoms, 39 percent were diagnosed with dermal bums. Of the patients with gastrointestinal symptoms from exposure to the caustic soda water, approximately 28 percent were diagnosed with gastritis or enteritis.

For the majority of patients diagnosed with dermal burns, treatment consisted of flushing the affected skin areas with water and applying antibiotic ointment. Patients with diagnoses of gastritis or enteritis were treated with citrus juice and anti-emetic medications.

Of the 248 patients to whom follow-up questionnaires were sent, 58 (23 percent) responded. Among females, non-Hispanic whites, and older persons (60 years), significantly higher proportions of the patients returned the questionnaire than among males, Hispanic whites, and younger persons ([less than]60 years) (p [less than] .05). These differences limited the conclusions that could be drawn from the follow-up questionnaire. Seventeen (29 percent) of the respondents reported health problems that were present several months after the exposure and that they perceived as related to the sodium hydroxide in the drinking water. Most (76 percent) of those problems were skin [TABULAR DATA FOR TABLE 2 OMITTED] problems. One respondent reported continued throat pain.

Discussion

Review of morbidity related to this incident was limited to medical record review of residents seeking medical attention at the city's only emergency department. The morbidity experienced by this patient population may not have reflected the total community experience. Comparison of the 1990 Bureau of the Census population estimates for this community with the demographic characteristics of persons seeking medical treatment for this exposure showed Hispanic people and children under 10 years of age to be over-represented in the patient population. This over-representation could have occurred because these groups were more likely to be exposed to the alkaline water and/or because they were more likely to seek medical attention after exposure.

Because large numbers of residents came to the emergency department in a short time, exposure information was not always documented on the medical record. The follow-up mail questionnaire had a low response rate (23 percent), and the investigators were therefore unable to obtain more complete exposure information from this data source.

Symptoms reported among community residents who sought medical attention after exposure to sodium hydroxide in the public water supply were similar to symptoms reported in the literature (14). Generally, the symptoms reported by the residents were of a milder nature than those associated with ingestion of cleaning products that contain lye, most likely because of the dilution of the sodium hydroxide in the water system. No serious health sequelae were reported on the follow-up questionnaires. The highest pH detected in the public water supply was 12.0. Previous research has indicated that a pH of 12.5 or more can cause esophageal ulceration (11).

Conclusion

The results of this study show the adverse health effects in a population exposed to a high pH (12.0) in a municipal water supply Because of the prompt recognition of the common exposure by hospital staff, the rapid reporting of the situation to the appropriate city officials, and the notification of the public through radio and television announcements, the numbers of persons exposed to the affected water and the duration of exposure were minimized.

Review of this incident underscores the importance of properly trained water treatment plant operators who are vigilant in monitoring and controlling the quality of drinking water delivered to the public. Following the incident, several steps were taken by city officials to prevent another such problem from occurring at the treatment plant. Plant staffing was upgraded to ensure that a properly trained operator was on duty at all times. New operations procedures were developed to improve monitoring of the treatment process and to establish better communication among the plant staff. The bypass line between the bulk chemical tank and the clarifier effluent trough was physically disconnected. New chemical feed facilities, including small, intermediate feed tanks, were installed to eliminate all direct-feed capability between the bulk chemical tanks and the treatment process. Automated monitoring equipment was installed to provide early warning and automatic plant shutdown in the event of a water quality problem.

Action was also taken by state environmental officials to prevent similar incidents from occurring at other surface water treatment plants in Texas. To this end, design standards for chemical feed facilities were changed to require the use of small, intermediate feed tanks or equivalent protection between bulk chemical storage tanks and the treatment process. The effect of this regulatory change has been to minimize the possibility of chemical overfeed incidents and to ensure a higher level of public health protection.

REFERENCES

1. Kikendall, J.W. (1991), "Caustic Ingestion Injuries," Gastroenterology Clinics of North America, 20(4):847-857.

2. Moore, W.R. (1986), "Caustic Ingestions," Clinical Pediatrics, 25(4):192-196.

3. Gorman, R.L., B. Benson, M.T. Khin-Maung-Gyi, W. Klein-Schwartz, E. Krenzelok, T. Litovitz, M. McCormick, N. McElwee, G.M. Oderda, and H. Spiller (1992), "Initial Symptoms as Predictors of Esophageal Injury in Alkaline Corrosive Ingestions," American Journal of Emergency Medicine, 10(3):189-194.

4. Gumaste, V.V., and P.B. Dave (1992), "Ingestion of Corrosive Substances by Adults," American Journal of Gastroenterology, 87(1):1-5.

5. Moazam, F., D. Miller, D.L. Mollitt, and J.L. Talbert (1987), "Caustic Ingestion and Its Sequelae in Children," Southern Medical Journal, 80(2): 187-190.

6. Howell, J.M. (1986), "Alkaline Ingestions," Annals of Emergency Medicine, 15(7):820-825.

7. U.S. Bureau of the Census (1990), 1990 Census of Population and Housing, Texas, Summary Tape File 1A.

8. Dixon, W.S., ed. (1992), BMDP Statistical Software Manual, Berkeley: University of California Press, pp. 129-153.

9. Dean, A.G., A.H. Burton, K.A. Brendel, D. Coulombier, J.A. Dean, R.C. Dicker, R.F. Fagan, D.C. Smith, and K.M. Sullivan (1994), Epi Info, Version 6: A Word Processing, Database, and Statistics Program for Epidemiology on Microcomputers, Atlanta: Centers for Disease Control and Prevention, pp. 373-413.

10. Gahlinger, P.M., and J.H. Abramson (1993), Computer Programs for Epidemiologic Analysis, Honolulu: Makapuu Medical Press, pp. 164-167.

11. Vancura, E.M., J.E. Clinton, E.P Krenzelok, and E. Ruiz (1980), "Toxicity of Alkaline Solutions," Annals of Emergency Medicine, 9(3):118-122.

Corresponding Author: Jean D. Brender, Ph.D., Infectious Disease Epidemiology and Surveillance Division, Texas Department of Health, 1100 West 49th St., Austin, TX 78756.