"On March 19, we had infants with no other exposure than formula reconstituted with tap water," said Ned Colange, chief medical officer of the Colorado Department of Public Health and the Environment (CDPHE). "That's when we pulled the trigger."

[ILLUSTRATION OMITTED]

Salmonella

outbreaks from drinking water are extremely rare. Also, CDPHE had not yet received results from tests on water samples. But although restaurant closures and widespread economic hardship were bound to ensue--although a population of 10,000 would be unable to wash dishes, take showers, or drink water from the tap--the agency issued a bottled-water alert.

[ILLUSTRATION OMITTED]

"We gave a bottled-water order without knowing definitely that there was Salmonella in the water," confirmed Steve Gunderson, director of CDPHE's Water Quality Control Division. "That involves some risk. Just imagine if it wasn't [the water]." He also pointed out that in most waterborne outbreaks, an initial event--perhaps a flood, a pipe rupture, or a fire at a water plant--points the investigation in a given direction. In this case, no indication of a problem had presented itself before people started getting sick. CDPHE made the call on the basis of an epidemiological investigation conducted over just a few of days.

Alamosa sits in the San Luis Valley, about 230 miles southeast of Denver. The Rio Grande runs past the city, and the landscape is scenic, comprising desert, state and national parks, and farmland surrounded by the forests and jagged peaks of the Rocky Mountains (see photos on page 55). At an altitude of over 7,000 feet, the climate can be demanding. Average annual snowfall is 41.7 inches. Generally only the months of July and August are completely frost free.

The drinking water comes from an aquifer deep enough to be considered sterile. At the time of the outbreak, Alamosa had a "closed" system--meaning it pumped directly from the aquifer into closed tanks. Under these circumstances it theoretically should not be possible for microbes to contaminate the water. So chlorination was not required.

And yet, more than 400 cases were reported over the course of the outbreak. Given the spectrum of symptoms associated with Salmonella infection--everything from bloody diarrhea and potentially fatal infection of the blood to minor digestive symptoms that people might not even notice--Colange believes that the number of actual cases was much higher. For every case that presents for care, he told JEH, another 30 cases are likely to have occurred, according to a multiplier used by the Centers for Disease Control and Prevention (CDC). By those numbers, he figures that "almost everyone in Alamosa probably had Salmonella."

The first case was reported on March 6. This report went to the Alamosa Public Health Nursing Service, which began a preliminary investigation. On Friday, March 14, an increase in cases led the nursing service to turn to the state health department. By that point, about 19 cases had been reported. "One [case] gets you worried, and two gets you even more worried," Colange said. "At 19 you think you have a real problem."

CDPHE's epi investigation started over the weekend. Since Salmonella is usually foodborne, staff surveyed outbreak victims about restaurants, picnics, and foods they had eaten. The nursing service had particular concerns about one restaurant. "More people seemed to have eaten there than not," Colange observed. So that restaurant was specifically named in the survey form. A couple of cases had also occurred in people who did not live in Alamosa. Those people had visited the city, though. And had drunk the water. As the epi investigation of the first 19 reported cases was winding up late on Tuesday, March 18, a staff member from CDPHE's Infectious Disease and Environmental Epidemiology Division called Colange and said he though the source might be the water supply.

"And I said, 'I have never heard of such a thing,'" Colange told JEH.

In fact, as the staff member told Colange, such outbreaks have occasionally occurred, including one in 1965 in Riverside, California, that made an estimated 16,000 people sick. Riverside had the same kind of water system as Alamosa.

Tests for the presence of Salmonella in the water would take some time to produce results. In the meantime, CDPHE conducted a quick screening test for total coliform bacteria. The results were positive. Routine monitoring, conducted just two weeks earlier, on March 5, had not found coliform bacteria in the water.

"Clearly, something had changed," Gunderson observed. "We had years of records from the system, and it had not been a problem."

As for the results of the epidemiological survey: No conclusive foodborne patterns emerged. Ill people had in various proportions eaten various foods in various establishments. The only thing everyone seemed to have in common was ingestion of water from the Alamosa system. And then there were those five infants who had eaten no food at all--just formula reconstituted with tap water--all of whom were ill.

[ILLUSTRATION OMITTED]

City officials did not want it to be the water. "There was a very uncomfortable period when the water people were having a hard time believing it was their water," Colange said.

Was the decision to issue the bottled-water alert a hard one?

Gunderson told JEH that there was a lot of discussion among CDPHE staff, but that there was a strong feeling about the water, especially in light of the total-coliform results from the water samples. And of course, if it was the water, the hazard was imminent.

Ultimately, Colange was the one responsible for making the call to stop 10,000 people from using the tap water. "Yeah, I would say my job probably depended on my level of certainty," he remarked. Nevertheless: "I wasn't too worried about it. I didn't lose any sleep over this."

There was, he acknowledged, some skepticism in Alamosa. For one thing, new cases seemed to be occurring after the bottled-water order was issued. On March 19, the day of the order, 79 cases had been reported. The number increased to 138 on March 21, two days later, and to 216 on March 24. "As late as Easter Sunday [March 23], there were people in the city who questioned whether it was the drinking water," Gunderson told JEH. To complicate matters, 16 states, including Colorado, were experiencing a Salmonella outbreak associated with cantaloupe from Honduras. Here was a tempting thought: Maybe the city's drinking water wasn't the source after all? Maybe it was cantaloupe from Honduras?

On the latter point, results from stool sampling vindicated the gumshoe epidemiology: The cantaloupe-associated Salmonella was a different strain from the one occurring in Alamosa. And as far as the case count went, appearances were deceiving. Colange explained:

There's the incubation period, then there's the I-think-I'll-get-over-this-on-my-own period, and there's the I'm-not-getting-any-better-I-want-to-see-the-doctor. Then you perturb it with an announcement from the local public health department that there's this problem and you might want to see your doctor. Then you perturb it the other way because we're over-whelming the health care system, so we say, "Don't go to the doctor unless. ..." So cases can go up and down.

With date of onset taken into account, it became clear that right after the bottled-water order went into effect, the number of new cases had started to go down.

[ILLUSTRATION OMITTED]

On Monday, March 24, all results from polymerase chain reaction (PCR) tests on the water came back positive, indicating the presence of Salmonella DNA in the water. "I will admit," Colange said, "that in an odd way--and it was odd--I was pleased." And two days later, results from pulsed-gel electrophoresis tests showed that the Salmonella subtype present in the water matched that present in stool samples from patients. "So you have the complete story," he added. "You have all the epi evidence hanging together, you have the Salmonella growing out of the water, and it's the same Salmonella that was making people sick."

As of this writing, one question remains: How did the pathogen get into a closed water system? Testing showed that on emerging from the deep wells, the water was fine. CDPHE considered a number of possibilities: cross-connections, a small leak in a shallowly set pipe somewhere in the distribution system, or--and this was the scenario that both Colange and Gunderson considered most likely--a problem with one of the city's water storage tanks.

Alamosa had three tanks: two towers, one of which dated to the turn of the last century, and an underground cement-lined tank with a cement cap and a pitched roof. The older of the two towers had some small holes, but the underground tank became a prime suspect when it was found to have cracks and to have visibly leaked. One theory was that runoff from a heavy snowfall that had occurred shortly before the outbreak might have carried contaminants, perhaps from bird droppings, into the tank.

More testing showed that water going into the underground storage tank was coliform-negative and that water coming out was coliform positive. The tank was drained and repaired, and no visible sources of contamination were found inside it. In other words, there was no smoking gun. Gunderson told JEH that they may never know for sure what the source of the outbreak was.

From now on, the Alamosa system will be chlorinated. Monitoring of the water will continue at increased frequency to ensure that contamination is not ongoing.

The incident may also have ramifications for other unchlorinated systems. "It got our attention," Colange said, "and it has actually gotten national attention, because there are a lot of these systems. And the concern is just that they're old." CDPHE intends to review chlorine waivers across the state and work with municipalities "to ensure that there is not a repeat of this incident."

Editor's Note: In April, following the extraordinary waterborne Salmonella outbreak in Alamosa, Colorado, our staff reporter spoke at length with officials from the Colorado Department of Public Health and the Environment. This month's Briefings, based on those interviews, comprises three articles, covering the investigation, some special challenges posed by the environment in which the outbreak occurred, and the emergency response.

-RB

Rebecca Berg, Ph.D.