TOXIC SHOCK
Unexpected Deaths from an Improved Product
It was Monday, September 12, 2011. I was on a plane to Washington, DC, to speak at the tenth anniversary of the Measles Initiative. Many groups were involved in a global effort to interrupt measles transmission, an effort coordinated by the American Red Cross. Deaths had declined from more than 3 million per year to several hundred thousand. Amazing progress. The meeting would celebrate 1 billion measles immunizations administered to children around the world in the previous decade. The event was bound to be exciting.
On the plane I began reading the New York Times and suddenly gasped as I saw a picture of Bruce Dan. The article explained that he had died, at age 64, of complications following a bone marrow transplant for leukemia. The article recounted his life of adventure. He had received a bachelor’s degree in aeronautics from MIT and a master’s in biomedical engineering from Purdue. He then decided to go into medicine, getting his degree from Vanderbilt University. His wife described him as “a Renaissance man, always reinventing himself.” He later spent time at the Journal of the American Medical Association and then became a medical news anchor.
Bruce’s death took me by surprise, and I began to reminisce. Bruce had been part of the EIS while I was the CDC director. He, with two other physicians in that program, Kathy Shands and Art Reingold, plus a host of other workers at the CDC, had labored in 1980 to solve the mystery of women suddenly dying of a shock syndrome. The women would go from healthy to toxic to death in a matter of days. It was absolutely scary.
As soon as word got out that this condition existed, reports came in to the CDC by the dozens. The deaths were distributed throughout the country, and each had seemed to be a rarity at the place of report. When the reports were combined, however, it became clear that this was a national problem. Doctors were put on full alert, and the CDC investigators scrambled to make sense of it. The June 27, 1980, edition of the CDC’s weekly newsletter reported on fifty-five cases of what was to become known as toxic shock syndrome (TSS) (1). Most patients were women, and most were between teenage years and age 50.
Senator Ted Kennedy held a hearing, asking the CDC to report on what it knew. Scientists frequently want some separation from politics, but it is not possible. Public health funding depends on appropriations from politicians. Therefore, when the CDC is called on to testify, the agency always has to agree. The problem is that answers are often unavailable early in an investigation; yet, the questioner will continue to probe. My general approach was to do the best preparation possible, bring a person with more knowledge on the subject, and then agree to get more information to provide for the record when we could not answer a question. I also learned over time to embrace the interest of politicians, to anticipate their knowledge needs before they asked, and to enlist them in the great adventure of public health. Every public health decision is ultimately based on a political decision.
For the TSS Senate hearing, I was accompanied by Dr. Shands and by Mrs. Robert Haley. Mrs. Haley had recovered from the syndrome and was asked to appear to report on what it was like to be a patient with this condition. Her husband, Robert Haley, worked at the CDC in hospital infections.
At one point, Senator Kennedy asked when we thought we would have an answer. Trying to combine optimism and realism and based on past investigations, I said, “I hope we have an answer by this fall.” Dr. Shands told me later that she almost fell off her chair. They were still in the dark, and she did not see how they could unravel the mystery that soon.
But they did. The CDC, the Wisconsin State Department of Health and Social Services, and the Utah State Health Department undertook studies on the practices and products that might be in common use by the patients. While the reports included a wide spectrum of ages and included both male and female cases, the preponderance of female patients between the teen years and age 50 made it possible to narrow the focus to the menstrual cycle and then to tampons as a major risk factor, especially one tampon.
Late one Friday afternoon the investigation team reported on the risk inherent in using Rely tampons, manufactured by Procter & Gamble. Women did not make up all cases of toxic shock. When women were affected, not all cases involved the menstrual cycle. And if cases were related to the menstrual cycle, not all women affected had used Rely tampons. So it was a matter of figuring out the relative risk of various activities or products. If all cases had been in women who used Rely, then the answer would have been apparent and noncontroversial. Aware that the CDC had been criticized after the swine flu epidemic in 1976, for an appearance of basing decisions on scant evidence (see the “Swine Flu” section of chapter 9), three different and independent statistical groups at the CDC were convened to work over the weekend to analyze the statistics and render an opinion on the validity of the conclusion reached by the toxic shock investigative team.
On Monday morning, we gathered to hear each of the statistical groups report on its conclusions. I anticipated that all three would find the relationship to Rely tampons was firm. The three teams described their approaches, and two supported the conclusions with no qualifications. The third team expressed concern that calculations could be biased and therefore in error.
It was time to make a judgment. Based on all of the analysis, and aware of the concern of the third statistical team, we reached a tentative conclusion that Rely tampons were a risk factor. Certainty might be elusive but the health of the public required that we share our opinion. The CDC released its findings and on September 22, 1980, Procter & Gamble recalled Rely tampons voluntarily, notified consumers, and retrieved the product from the market.
The impact of the report was dramatic. In 1980, a total of 890 cases of TSS were reported, 812 (91 percent) were associated with menstruation, and 38 women with menstrual TSS died. In 1989, 61 cases of TSS were reported, and 45 (74 percent) were menstrual. There were no menstrual TSS deaths in 1988 and 1989.
Some scientists were concerned that the marked reduction was an artifact of the CDC’s passive surveillance system, wherein cases are reported voluntarily. Therefore, a coalition of states developed a multistate active surveillance system, with investigators searching for cases. It found the same pattern. The rate of menstrual TSS declined from 6.2 cases per 100,000 women between 12 and 49 years of age in 1980 to 1 per 100,000 women later in the decade.
The precise mechanism by which Rely tampons increased the risk of TSS is unknown but may well have involved its heightened absorbency, which would require women to replace the tampons less often. This might have allowed time for bacterial growth and toxin production. But the result of taking them off the market was clear. The reporting of nonmenstrual toxic shock has remained fairly constant over the years.
Several lessons stand out. First, this was a life-and-death situation, and once again an EIS team, in a relatively short time period, took a chaotic situation and made sense of it. Second, the manufacturers had released a product that they felt was superior to other products on the market because of its high absorbency. An adverse outcome of low incidence is difficult to find in the relatively small numbers of subjects used in premarketing tests; therefore, postmarketing surveillance is essential to identify adverse effects of low incidence. A third lesson is the value of erring on the side of public health. The makers of Rely did not hesitate. Once the problem was identified, they quickly took the product off the market.
Fourth, once again public health was presented with a problem somewhat different than that found in clinical medicine. In the latter, new findings may generate a great deal of excitement, but they are generally not accepted until repeated by another investigator. In public health, a second study is sometimes not possible. In the toxic shock incident, removing the product from the market was prudent but prevented a confirmatory study. The decision on reporting always has to be on the side of minimizing the chance for suffering and death—but with an awareness that an inaccurate report based on insufficient evidence could devastate an industry. Finally, the additional step of having three statistical groups reanalyze the findings was prudent.