by Dan Fagin
The meetings with Eric Green would continue, intermittently, for three more months. After each long session, Green and his technical adviser, Doug Allen, would meet privately with each side and offer their confidential opinions about the strengths and weaknesses of the day’s arguments. Exactly what Green said to each side would be forever secret, under the terms of his consulting contract. But by the end of the thirteen expert presentations, in early 2001, he was apparently telling both sides that a financial settlement of the case was both possible and desirable—and the companies were listening. More than three years after Schlichtmann began his unlikely quest for a lawsuit-free resolution of the case, attorneys for Union Carbide, Ciba, and United Water signaled that they were finally ready to start talking about a settlement. They might not be convinced that the families were in the right, but the prospect of a high-profile lawsuit and maybe even a huge jury verdict was more frightening than ever. There was also the wild card of the long-awaited case-control studies. If their results were clear, one way or the other, there would be little point in further settlement talks. At that point, the companies either could be facing hundreds of millions of dollars in liabilities, or none at all.
Jerry Fagliano tried not to think about how much was riding on the outcome of the studies that had been the center of his working life for more than four years. He cut himself off from any news about what the lawyers were doing but could not help being aware that the outcome of his investigation could add or subtract many millions of dollars from any out-of-court settlement. The studies could even help determine which families would benefit, if they ended up linking certain types of cancer to local pollution while excluding others. “We were certainly aware of that, and it put a tremendous amount of pressure on us to get the study right,” Fagliano remembered. Linda Gillick made it impossible to forget what was at stake. She almost never spoke in public about the legal case—that was a private matter for the families and their lawyers, as far as Gillick was concerned—but she never missed a chance to remind Fagliano that much more than money was riding on the outcome of his studies. The families were counting on them to explain what had so far been unexplainable: the cause of the cluster. Fagliano still attended monthly meetings of Gillick’s advisory committee to explain, over and over, why the studies were taking so long. Gillick usually greeted him with a sharply worded reminder that the families were out of patience. By now, she and Fagliano had attended nearly one hundred public meetings together and dozens more in private. They had developed a mutual respect. But in public, Gillick was as relentless as ever, even when the audience was sparse, as was usually the case now that most residents and reporters had lost interest. Easing up was not her style.
By 2000, Fagliano was waiting for just one thing: the completion of the computer model of the town’s water system, which would tell him which wells had supplied which homes. The lawyers had used a much simpler model in the maps they devised for the expert presentations in front of Eric Green, but their version was a mere stick figure compared to the masterpiece Morris Maslia was devising for Fagliano’s studies, at a cost of more than $5 million. Maslia made no apologies for the slow pace. Creating 420 water distribution simulations, one for each month of the thirty-five-year study period, and then testing their reliability, was extremely complicated and difficult.3
Finally, in the fall of 2000, Maslia was ready. In November, at the same time that Barry Finette was analyzing the first blood samples from Toms River and the lawyers were holding their first session with Eric Green, Maslia showed his model to a panel of expert advisers, who suggested only minor changes. In January, he sent his simulation results to Fagliano, who already had the results of the air pollution dispersion model created by the team at Rutgers. As with Maslia’s water model, the air model did not try to determine which particular pollutants were in specific parts of town at various times; the historical records were far too vague for that. Instead, the models were comparative tools that showed how pollutants in general were dispersed around Toms River and thus could be used to estimate the relative burden borne by each neighborhood, not the actual burden.
Jerry Fagliano did not need to know what was actually in the air and water all those years ago because his case-control study, too, was based solely on comparisons. What he needed to know was which neighborhoods, based on the models, got more pollution and which got less during the three critical periods: 1962 to 1996 for Ciba air emissions, 1962 to 1975 for Holly Street well water, and 1982 to 1996 for Parkway water.4 Ever since the shocking release of Michael Berry’s cluster study in 1996, there had been no compelling new information about cancer and pollution in Toms River, despite the millions of dollars spent on investigations. With the completion of the computer models, Fagliano and his collaborators were finally in a position to say something new.
There was just one more preliminary step: The researchers needed to calculate a numeric score, an “exposure index,” for every address at which a case or control child lived. They would need index scores not only for historical exposure to air emissions from the Ciba factory and the nuclear plant, and to water from each public well field, but also to water drawn from the private backyard wells that had been so common in town until the 1970s.5 And they would need to calculate these scores in two different ways because there were two case-control studies: a birth record study and an interview study.
Fagliano knew almost nothing about the 528 children in the birth record study. Forty-eight had been diagnosed with cancer, and the rest were healthy controls matched to case children by age and sex. He did not even know whether they still lived in Toms River.6 With so little information about residential history, it would be impossible even to guess at the relative pollution burden each child had faced before being stricken. But because he knew the dates and home addresses on their birth certificates, it would be possible to use the air and water models to estimate the relative exposure each child’s mother faced during pregnancy.7 That seemed especially useful now that the interview results were suggesting that prenatal exposures might be very important.
For these birth record study children, calculating an exposure index would be straightforward. For drinking water exposure, Morris Maslia would identify the location in the pipeline network closest to the birth address of each child and then run the model to see which combination of wells had supplied the home’s water for each month of the thirty-five-year study period.8 Fagliano’s team would then take the model results for the nine months that preceded each child’s birth and calculate a prenatal exposure score. For example, a child whose prenatal water supply was 60 percent Parkway water, 5 percent Holly water, and 35 percent other wells would be classified as having high exposure to Parkway water but low Holly exposure.9 The Rutgers team devised similar scales for air emissions from Ciba and the nuclear plant, and the health department used them to generate prenatal air exposure scores for each child and each type of exposure.
The scores were a simple and sound way to compare exposure but were vulnerable to criticism because they did not take into account where the children lived after they were born and which type of water (tap or bottled) their mothers consumed while pregnant. Depending on the study’s outcome, critics on one side or the other would have legitimate reasons to complain. But there was a smaller group of Toms River children that Fagliano knew much more about. These were the 199 children whose families had been interviewed in depth by state health investigators. They were the subjects of his second study, the interview study. Forty of the interview study children had been diagnosed with leukemia or brain or nervous system cancers. One hundred and fifty-nine were healthy controls matched to cases based on sex and age.10
Thanks to the interviews, Fagliano knew where each of these children lived all the way from one year before birth until the month of diagnosis. He knew how many glasses of water per day each mother and child had drunk and whether there was a filter on their tap. True, his information depended on memories and thus was subject to re
call bias, but it was far better than having no information at all. So for the interview study, Fagliano’s team was able to devise a more elaborate exposure formula that took into account changing addresses and varying personal behaviors. A habit of drinking a lot of tap water, for example, would raise a child’s water exposure score.11 These adjustments made the interview study, despite its smaller size, less vulnerable to criticism than the birth record study.
By the beginning of 2001, Jerry Fagliano and his collaborators at the state health department had a full set of exposure scores for all 702 children in the two studies.12 After five years, the preparation was over. There were no more decisions to be made about which hypotheses to investigate, which children to include, or which formulae to apply. There were no more interviews to conduct or computer models to test. There was nothing left to do but run the numbers and interpret the results.
The biomarker Barry Finette had decided to look for in the Toms River blood samples was a gene with a name only a biochemist could love: hypoxanthine-guanine phosphoribosyl transferase, or HPRT. He wanted to find out whether the white blood cells of children who drank Toms River water had more HPRT mutations than the cells of unexposed children who lived out of town. If genetic mutations were the key to triggering cancer—as Theodor Boveri’s century-old speculations, Hermann Muller’s fruit fly experiments, and Alfred Knudson’s retinoblastoma studies all strongly suggested—then mutation frequency was a logical indicator of cancer risk.
Biomarkers like HPRT were at the heart of the young field of molecular epidemiology. The term had been coined in 1982 by a junior researcher at Columbia University, Frederica Perera, in a paper she wrote with her mentor, I. Bernard Weinstein.13 An admirer of Knudson’s ideas about multi-hit carcinogenesis, Weinstein studied interactions between pollutants and DNA at the most intimate level: the chemistry of individual molecules. Perera and Weinstein thought it might be possible to use the pattern-recognition tools of traditional epidemiology to identify which of those molecular couplings were associated with disease. Again, the evidence would be indirect—correlation, not causation—but it would be powerful because these biomarkers could be measured directly in human blood and tissue without requiring extrapolation from animal tests.
If the molecular epidemiologists were successful, they would be pioneering a new way to assess the health risks of chemicals—and also to win legal cases like the one in Toms River. Even more importantly, their work might lead to the development of early warning medical tests to save the lives of undiagnosed cancer victims. As the people of Toms River knew all too well, conventional cancer epidemiology was like a fire engine that arrived long after your house had burned down. By the time answers were available, many people would already be sick or dead, and it would be too late to do anything but guess which decades-old exposures might be responsible. Molecular epidemiology held out the hope of working much faster—but only if people like Perera could find biomarkers that could be reliably measured and were true surrogates for disease risk.
Perera’s first biomarker candidate was benzo(a)pyrene, the same compound Ernest Kennaway had identified back in 1932 as the crucial carcinogen in coal tar, the fountainhead of the chemical industry. A potent carcinogen, B(a)P had a distinctive way of making mischief in human cells: Its molecules would bind tightly to DNA strands by sharing pairs of electrons in tight, covalent bonds. When a cell replicated, these adducts could trigger genetic mutations. Hunting for these “B(a)P-DNA adducts” in human cells, Perera thought, could be a very useful way of assessing an exposed person’s risk of getting cancer. Starting in 1980, Perera and her collaborators at the National Cancer Institute looked for adducts in B(a)P-exposed mice, rabbits, and dogs and also in the white blood cells and lung tissue of humans, some of whom had lung cancer. What they found would establish the tantalizing pattern of molecular epidemiology: hints of significance cloaked in ambiguity. Animals injected with B(a)P did indeed have more DNA adducts, but the correlation was murkier in humans with cancer who smoked.14 Adducts were not a surefire “dosimeter” of exposure and risk, as Perera had hoped. Carcinogenesis was too complex, involving too many steps and too many possible pathways, for any one biomarker to be an accurate predictor of risk. Still, the results were interesting enough to keep going, she thought.
Soon, Perera and dozens of other researchers were setting up larger case-control studies and counting DNA adducts in highly exposed people all over the world. The field soon expanded far beyond adducts to embrace other biomarkers, too. Investigators tested populations for mutations in oncogenes (including the “Philadelphia chromosome” that had inspired Knudson) and also tumor-suppressor genes—especially the TP53 gene on the short arm of Chromosome 17, the most frequently mutated gene in human cancer.15 There were other genetic biomarkers, too, including indicator genes that did not play a direct role in causing or preventing cancer but seemed to be useful surrogates for assessing overall risk. Today, researchers can quickly and inexpensively scan the entire human genome for mutations, so indicator genes have become less important. But in 2000, when Barry Finette was launching his Toms River studies, they were vital.
HPRT, the biomarker Finette was using in his initial Toms River study, was one of the most popular indicator genes—though not necessarily for the right reasons. No one was sure if HPRT was an important gene for carcinogenesis; the evidence from factory studies was mixed, though there was good evidence that radiation and chemotherapy treatment triggered HPRT mutations in cancer patients.16 But the gene had other qualities that made it a splendid biomarker, including the fact that it was easy to propagate, or clone, inexpensively in the laboratory.17 That was an essential trait for human studies, in which every blood sample was precious and had to be carefully conserved. Even so, focusing a cancer epidemiology study on HPRT was a little like looking for lost keys under the streetlamp because the light was better there. It was a very convenient biomarker gene, but hardly the most promising of the thousands of possible candidates.
Finette was handicapped not only by a questionable biomarker but also by an experimental design that relied on some very speculative assumptions. He planned to compare HPRT mutation rates in the siblings of TEACH children to rates in children of the same sex and age who lived out of town. The plan assumed that all of the siblings, no matter where they actually lived in Toms River, had been highly exposed to contaminated water and that none of the out-of-town children had been similarly exposed. But water contamination was a fact of life throughout South Jersey, and within Toms River varied significantly in different parts of town. Indeed, Jerry Fagliano’s case-control study was premised on this variability, since it compared various neighborhoods within the township. Just as questionably, Finette’s study had to assume that siblings were apt surrogates, since the DNA of the TEACH children was already too damaged by chemotherapy to analyze directly. Finally, the blood samples in Finette’s study were not collected until 2000, many years after local water and air pollution were at their worst. Finette thus had to assume that the children’s DNA still showed damage from many years earlier and had not been masked by subsequent mutations that accumulate naturally as the cells in a child’s body keep dividing with age. For all those reasons, Finette worried that he would not find the answers the Toms River families sought.
He was right to be pessimistic. It took Finette’s research team five months to count HPRT mutations in white blood cells from the forty-nine Toms River case siblings and the forty-three control children from out of town. In early 2001, he reported his results to the lawyers: There was no appreciable difference in mutation frequency between the two groups.18 The experiment, Finette thought, had been crippled by the cascade of assumptions built into it, especially the long gap between the years of peak pollution and the collection of the blood samples in 2000. If there had ever been a spike in mutation frequency in the Toms River children—and Finette still believed there had been—the study was too tardy to detect it. As usual, science had arrived too late
to make a difference in Toms River.
Finette had other ideas for Toms River research that he would pursue for years to come, returning over and over to the plastic tubules in his lab’s freezer, like a pilgrim in search of enlightenment. He would go beyond merely counting mutations and instead look to see if specific genetic changes were present in cells of local children but not out-of-town controls. His ideas would take many years to test—far too long to influence the outcome of the legal case—and there was no good reason to think they would ultimately bear fruit. Molecular epidemiology, in its simplest form, was premised on the idea that diseases could be predictably associated with single, specific genetic variations, but most cancers did not play by those rules. Still, the families and their lawyers wanted Finette to keep going. They were interested in more than just leverage in the upcoming settlement negotiations. If there was any chance to someday learn something new about the cause of the cluster, they wanted to pursue it. So Finette and a dwindling team of assistants would keep working, quietly, as the Toms River drama built to a climax. And then they would keep working for years afterward, searching for faint clues in a dark sea of genetic data.
The families had put their faith in three kinds of case-control studies, and now it was clear that two of them—Finette’s blood work and the National Toxicology Program’s rat study of SAN trimer—would take many more years to complete and were so weighed down by scientific complexity that they were unlikely to end in a meaningful result. Realistically, there was just one hope left.
