The Blue Death

by Dr. Robert D. Morris

  Johnson took a radical new approach. Rather than filtering the water, he would disinfect it with chlorine. Chlorine had been tested in a few small plants and had been applied to sewage as a disinfectant, but it had never been used to treat water on this scale. As Johnson supervised the construction of the plant, he got a call from Jersey City. The beleaguered contractor for the Jersey City Reservoir had heard about his project. If Johnson’s system was good enough for cows and pigs in Chicago, he wondered, why wouldn’t if work for the people of New Jersey?

  By the summer of 1908, Johnson had built a treatment plant near the outlet of the Jersey City Reservoir unlike any other plant in the country. Inside three tanks, each the size of a small swimming pool, contained 10,500 gallons of a substance then referred to as chloride of lime, the same white substance that the General Board of Health had encouraged communities to spread along the streets of London fifty years earlier to slow the spread of cholera. The contractor turned on the plant and returned to Mayor Fagan, demanding that he be paid.

  Jersey City sued, insisting that the contractor was obligated to provide water free from the upstream sewage. The contractor insisted he only needed to supply safe water and had done so. In the end the courts agreed. In December 1908 the New York Times acknowledged what would prove to be one of the most important public health interventions of the twentieth century with four short paragraphs on its back page. Sandwiched in between articles about a threatened boycott of automobiles built in New Jersey and the appointment of a fourth member to the New Jersey Fish and Game Commission, was the paper’s report on the first routine use of chlorine for a municipal water supply.

  “So successful has been this experiment,” the article stated, quoting an unspecified source, “that any municipal water plant, no matter how large, can be made as pure as mountain spring water.” The article went on to note that New York had decided to add chlorine to the waters of the immense Ashokan Reservoir which was still under construction in the Catskills. Indeed, water suppliers all over the country had been looking for just such a technology. Those that did not filter leaped on a treatment system that required a small fraction of the capital and space needed to construct a filtration plant. At the same time, operators of filtration plants needed a method to protect against the pathogens that inevitably find their way through a filter bed.

  Engineers around the country and in Europe had experimented with other ways to disinfect water including ozone and ultraviolet light, but nothing provided the low cost and ease of use associated with chlorine. Within six years half the water treatment plants in the United States were using chlorine to disinfect some 2 billion gallons of water each day. By 1924 three thousand cities had turned to chlorine (mostly in the form of chlorine gas) to treat almost 4 billion gallons of water.

  Employing chlorine together with rural reservoirs, filtration plants, or both, municipal water suppliers around the country soon had systems that routinely provided water free from disease-causing bacteria. Serious waterborne diseases like cholera, typhoid, and amebic dysentery dropped from routine to rare. In 1900 an average American had a 5 percent chance of dying of a gastrointestinal infection before the age of seventy. By 1940 that rate had dropped 0.03 percent and by 1990 it had fallen to about 0.00005 percent.

  By 1960 drinking water had disappeared from the national consciousness as an issue of concern. Civil engineering had been the high technology of the nineteenth and early twentieth centuries, but the luster faded. With the opportunity to reengineer watersheds all but gone and the technology of treating water mature, providing drinking water became mundane. The primary goal for those maintaining public water supplies became not innovation, but invisibility. If you did your job right, no one would notice. If you were noticed, it meant the water smelled bad or tasted funny, or, heaven forbid, had caused a detectable disease outbreak. Success bred complacency. Soon complacency drifted into somnolence. Then, as the century wore on, problems began to disturb this slumber.

  9

  THE TWO-EDGED SWORD

  The cold nights of autumn had already begun to paint the leaves of Boston as I walked from Vanderbilt Hall past the classic quadrangle of Harvard Medical School to the uninspired concrete monolith that houses the university’s School of Public Health. Once inside I asked the security guard for directions to the offices of the Technology Assessment Group. To my left was a large bank of elevators. The guard, a man whose accent I could place only as from somewhere on the Indian subcontinent, gestured in the other direction past the kitchen toward a single small elevator next to a cleaning closet. That elevator went in only one direction, down. As the elevator descended past the basement and into the bowels of the Kresge Building, I felt sure the guard had misheard me.

  I had arranged to spend two months during my senior year of medical school working with Tom Chalmers, a leader in clinical research and the former dean of Mount Sinai Medical School. Dr. Chalmers had pioneered the use of a new and important tool in medical research known as meta-analysis. After leaving Mount Sinai, he had paired up with Fred Mosteller, a chaired professor of mathematical statistics at Harvard, to create the Technology Assessment Group (TAG), which was devoted to the methods and application of meta-analysis. I had expected to find these two giants of medicine and statistics in a suite of offices overlooking the Boston skyline, but as I made my way past the copy center to an oddly angled cluster of offices in the second basement of the Kresge Building, I understood that Harvard was not like other places.

  When I arrived, TAG’s offices were empty except for a short man with a graying shock of black hair, a day or two of stubble on his face, and black-rimmed Coke bottle glasses. “Hi, I’m Bruce,” he said with a New York accent and a wry smile, “Bruce Kupelnick. You must be Tom’s new person.” I would soon learn that Tom had many “persons.”

  Bruce showed me to my “desk,” actually half of a small desk in the middle of Tom’s office. The other half belonged to another one of Tom’s people. Bruce took great pleasure in showing me around the office, introducing people as they arrived. He had left the University of Chicago in the midst of his doctoral dissertation in history with plans to write a book. The book had proved to be far more elusive than he had anticipated. He had found an ideal spot as Tom’s assistant, embedding himself in the unique intellectual buzz of Boston.

  As people arrived the conversation turned to the success of the Red Sox, who were in the process of winning the American League East. Amid the gathering wonder as to how the Sox would manage to blow it this year, a tall angular figure swept into the office. His thinning gray hair had just a hint of its original red and appeared windswept, as if the sheer pace of his life were blowing it back. A matching mustache bounced on his upper lip. Tom Chalmers smiled, stuck out a long arm, and shook my hand as he greeted me with unaffected enthusiasm. He invited me into his office, but not before adding his thoughts on Boston’s shot at the pennant.

  Tom’s office had a narrow desk facing the wall at a height that allowed him to stand and work. He rarely sat. From it he held court on the rare days when he was not on the road. (I counted eleven days in the office during the two months I was there.) Tom’s presence brought a steady stream of visitors. One by one Tom’s people came by to discuss their projects.

  For the moment I had an audience with the master. From his desk Tom produced a thick folder crammed with photocopied papers. I leafed through the dog-eared articles, as he explained that each of the papers examined the relationship between chlorinated drinking water and cancer. A student of his had assembled them for a course he taught on meta-analysis. She had not been able to come up with a system for combining the results of the papers. Tom suggested that my first project should be to see if I could figure out a way to make the study work.

  To a student the “suggestion” of a research mentor has the weight of an edict from on high. I sat down at the fraction of a desk that had been carved out for me and set to work. I dug into the stack of papers in search of the pat
tern that would allow me to piece together the puzzle. I soon realized that I would need to burrow back to the birth of the idea that spawned them if I wanted to understand the papers. One of the best places in the world to do that was next door at Harvard’s Countway Medical Library.

  Every minute of every day, the world’s researchers are adding to the ever-expanding inverted pyramids of medical knowledge. In the stacks of Countway, I excavated down to the well-worn bricks buried deep within the pyramid. As I did so, a larger story took shape.

  On April 22, 1915, with dawn breaking outside the Belgian town of Ypres, German seventeen-inch howitzers suddenly opened fire. For thirty minutes French and Algerian forces huddled in their trenches, and then, abruptly, the guns fell silent. Carefully the men rose up and peered out toward the rising sun. They scanned the horizon expecting to see German soldiers running through no-man’s-land with fixed bayonets. At first they saw only the sun rising over a barren landscape with a light breeze blowing from the enemy lines. Then something began to grow on the horizon. As they watched, a vast greenish yellow cloud began to slither toward them.

  They could not know the nature of the demon that approached. The initial bombardment had provided the Germans cover while they unleashed 168 tons of deadly gas. The east wind carried the poison over four miles of trench lines where it tore at the lungs of some ten thousand soldiers, killing half of them. Some had survived by running from the cloud, but when the Germans advanced behind the gas, they encountered thousands of Allied soldiers coughing violently, temporarily blinded, and stumbling among the corpses of their countrymen.

  The chemical that the Germans had chosen to launch the era of gas warfare was chlorine. The extreme reactivity of chlorine that made it a potent disinfectant in drinking water made it a deadly weapon in chemical warfare. At the time of its introduction for water treatment, the concentrations used in drinking water did not appear to have any detectable toxic effects in humans. Jersey City could not even find a scientist to argue that chlorination of drinking water might have deleterious effects on human health. The tremendous benefits associated with inactivation of harmful pathogens made arguments against its use seem heretical at best.

  The level of chlorine used to disinfect drinking water is far below that needed to cause acute toxic effects like those inflicted by the German attack. In the decades following its introduction in Boonton, any ill effects of chlorine seemed hardly worth consideration when compared to the obvious benefits of disinfection. But its ability to react with and kill microbes would create an unanticipated problem. I would find the story in the papers that Tom Chalmers handed me on that October morning in Cambridge.

  The first hint of a problem came in 1970, when Johannes Rook, a chemist for the Rotterdam Waterworks, filled a bottle with the city’s treated drinking water. With massive waterborne outbreaks of cholera and typhoid a distant memory, water suppliers had moved on to other concerns. Among them was improving the taste of the drinking water, a major source of customer complaints, particularly when chlorine doses were high. Charged with this task, Rook had a problem; standard methods for finding chemical contaminants in drinking water were useless to him.

  From his experience as a chemist for a Dutch brewery, Rook knew that much of the flavor in a glass of beer floats in the air above it, filling our noses before we even have the first sip. Once these volatile organic compounds (VOCs) have evaporated, the flat beer that remains offers little taste. In 1970 the standard methods for testing water failed to capture the VOCs. Rook needed to find a new way to test water. Undaunted he adapted a method he had used for testing beer. What he found when he used that method to test water shook the complacent world of drinking water treatment.

  Using equipment of his own design, Rook collected the chemicals that evaporated from the water. He then separated these VOCs by injecting them into a gas chromatograph, a device used to help identify unknown chemicals. On one side of the chromatograph, ink flowed from a thin finger of metal onto a long strip of paper as it rolled past. As each different VOC passed through the chromatograph, the pen jerked, leaving a series of spikes, like a range of narrow mountains on the paper’s red grid. The size of the spike corresponded to the amount of the chemical that was present. When he examined the graph, he was shocked. The largest spike corresponded to chloroform, a chemical he had never expected to find.

  Chloroform was no longer seen as the benign chemical that John Snow had offered with such confidence to the queen of England. In 1945 scientists had shown that high concentrations caused liver cancer in mice and could also destroy their kidneys. Inadequate regulations had allowed its continued use in a number of products, but its presence in drinking water raised the stakes dramatically.

  Where could it have come from? No industry in Holland could account for the amount of chloroform present. To find out, Rook began to test samples of the water coming into the treatment plant. The results were even more alarming than the initial discovery of the spike. The river water flowing into the plant contained no chloroform at all. Rook could draw only one conclusion. The process of chlorination was causing the formation of toxic chemicals including chloroform in the drinking water.

  In the sixty years since its introduction, chlorination had become central to drinking water treatment. The possibility that, as it made the water safe from pathogens, it was introducing a new set of poisons had staggering implications. Rook checked with the plant’s health officer who assured him that many cough medicines contained chloroform and its presence in drinking water should not be a source of concern. Nonetheless, Rook recognized that the news could be explosive. For the next four years, he kept his findings quiet as he conducted more tests.

  In 1972, as Rook refined his study, a report from the International Agency for Research on Cancer upped the ante. The authors of that study pointed to the 1945 study, which, while small, clearly suggested that chloroform could cause cancer. More research was urgently needed.

  Four years would pass before Gordon Robeck picked up his phone on October 15, 1974. As he did he wondered why a reporter from the Miami Herald would be calling him. Robeck was head of the drinking water treatment research group at the Environmental Protection Agency (EPA) and prepared himself for routine questions about pesticides or industrial chemicals in the water supply. But this reporter had a question he never expected. Somehow this reporter had heard that chlorination caused the formation of chloroform in drinking water. Was it true?

  As it turned out, Tom Bellar, a chemist at EPA’s drinking water labs in Cincinnati, had been investigating chemical contaminants in the city’s water at the same time Rook had been analyzing the flavor of Rotterdam’s water. Bellar had studied air pollution before shifting to water and, like Rook, brought a unique perspective to the chemical analysis of water. Faced with the challenge of developing a better method to measure volatile chemicals, Bellar had come up with a method very similar to Rook’s. When he applied the method, he discovered, just as Rook had, that the treatment of drinking water produced an array of chlorine-containing compounds, including chloroform. For the four years that followed, the EPA and Rook had sat on the same explosive news.

  To avoid release of the information, the EPA simply chose not to publish its findings. Rook followed a strategy that was almost equally effective. He published his findings in Dutch. Even though he included the graph from his gas chromatograph he did not mention that the large peak corresponded to chloroform. Only an astute reader of the Dutch literature would have understood the implications of Rook’s work.

  One such reader found his way onto the staff of Miami’s pollution control program. In 1974 he mentioned Rook’s work to a friend, Mike Toner, a young science reporter at the Miami Herald. Rook’s and Bellar’s discovery was about to detonate.

  As Toner dug into the story, he placed the call to Robeck. Robeck and the EPA had been able to stay silent as long as nobody knew the right questions to ask. They had hoped to contain the story until they could find a so
lution to the problem. When Toner asked him pointblank if the chlorination of drinking water resulted in the formation of chloroform he had no choice but to answer “Yes.”

  The October 17, 1974, edition of the Miami Herald carried Toner’s story. The national media picked up the story and by the end of the month NBC had begun a three-part series on chlorinated chemicals in drinking water. This was not the managed release of information EPA had in mind. But the story was just beginning.

  Those called to epidemiology find themselves in the peculiar business of picking through recent history in search of public health mistakes. From cigarettes to fried foods, epidemiologists have shown us the fallacy in our assumptions about the safety of the things we eat, drink, and inhale. In 1975, when the EPA released a report showing that water supplies all over the United States were contaminated with a variety of chemicals including chloroform, environmental epidemiologists pricked up their ears.

  The story turned ominous when, in 1976, the National Cancer Institute reported that long-term exposure to chloroform caused cancer at concentrations far lower than those studied in 1945.

  Well-intended efforts to protect us from pathogens in drinking water had resulted in a natural experiment on a massive scale. Tens of millions of people had been subjected to a lifetime of exposure to a known carcinogen. Epidemiologists around the United States set about assembling and analyzing the results of that experiment. The thousands of hours put in by more than seventy epidemiologists ultimately produced the stack of papers that Tom Chalmers handed me in the fall of 1990. In just thirteen years, they had generated more than twenty-two studies, all designed to answer a seemingly simple question: Does exposure to chlorination by-products in drinking water cause cancer?