REYE SYNDROME
Following the Evidence
Sometimes a disease problem descends abruptly, as with the first case of Lassa fever in the United States in 1969. Sometimes the problem is abrupt but the explanation delayed, as with Legionnaires’ disease in 1976. But often the problem is understood only over time, when enough evidence is finally available to allow for insights.
This chapter highlights the agonizingly slow unraveling of a problem and the equally slow steps in responding adequately to it. Part of the problem was working in a field at the edge of the science. At the time of this outbreak, we in the public health world did not even know how to do meta-analysis (a mathematical approach of putting studies together to give the aggregate findings more power than the individual observations), and, therefore, we could not assemble experiences into a convincing whole. One enemy was a powerful aspirin industry, which did not want to lose profits from giving aspirin to children, even if it could be harmful.
This account appears repetitious at times but that only points out the careful line we had to walk in uncovering the truth. This account may also read more like a scientific report than a story of solving a mystery, but that is also the point. It shows science at its best, finally leading to a policy change and then a behavioral change.
Reye syndrome, or Reye’s syndrome, took time to understand. The condition may have been recognized in the 1920s. However, it was a paper by Reye, Morgan, and Baral, in the Lancet in 1963 (1), that gave it the status of a syndrome, and suddenly many pediatricians realized that they might have seen the condition. This syndrome is characterized by fever, lethargy, confusion, and vomiting and may include a rash on the hands and the feet. The injury is most pronounced in the brain and the liver. Adults are rarely affected. Children may go into coma or have seizures; children die in one-third of cases.
Shortly after the Lancet paper, EIS officer George Johnson reported on an outbreak affecting sixteen children with influenza B who developed neurologic symptoms and signs (2). It was the beginning of a frustrating journey to investigate a mystery while fighting off snipers from industry and government.
Arizona Study
In 1979, Karen Starko, another EIS officer, conducted a study in Phoenix, Arizona, that finally provided traction for public health action. She reported, along with her colleagues, on seven pediatric cases of Reye syndrome during an influenza A outbreak. These scientists compared the course of these children’s illness and treatment with those for sixteen control classmates who had also been sick but had not developed Reye syndrome (3).
All the patients had taken medications for their influenza symptoms. Whereas all seven case-patients had been given salicylates (that is, nonsteroidal anti-inflammatory drugs such as aspirin), only half of the sixteen controls had received these drugs. While the numbers were small, this was a turning point in understanding the risk of aspirin for this condition.
A July 1980 MMWR review (4) reported on more than 300 cases of Reye syndrome in a four-month period. It spoke of the association of the syndrome with certain viruses and other “factors, such as medications or toxins” that might contribute to the disease. Among the toxins were “isopropyl alcohol … warfarin, and aflatoxins.” It then mentioned the Starko study before concluding that “further investigations are needed to more clearly define the possible role of salicylates use and toxins in the pathogenesis of Reye Syndrome.”
On November 7, 1980, the MMWR reported on studies in Ohio and Michigan that added to the 7 cases reported by Starko. Michigan reported on 25 patients, and Ohio collected 159 cases from pediatric centers. These studies included control groups. The editorial comment stated: “The results of these studies suggest that during certain viral illnesses the use of salicylates—even before the onset of vomiting—may be a factor in the pathogenesis of Reye syndrome. In view of these data, parents should be advised to use caution when administering salicylates to treat children with viral illnesses, particularly chickenpox and influenza-like illnesses [emphasis added]” (5).
Looking back, the reader may be bewildered by the CDC’s reluctance to provide stronger warnings. The problem was a desire to show “statistical significance” before making a recommendation.
The following month, in December 1980, Starko and her colleagues published the results of their 1979 study in the journal Pediatrics (3). Now their findings were in a peer-reviewed journal, not just a government publication. As the evidence mounted, there were great efforts to test the validity of these findings. While control groups increase confidence in findings, it is always possible that parents’ recall is different because their children were severely ill, whereas the control children were less so. It is also possible that the more severe illness of Reye syndrome led to the increased use of aspirin rather than the other way around. And could the investigators have “led the witness” by a preconceived idea that aspirin was a hazard? For all of these reasons, some suggested using caution about recommending against the use of aspirin in these cases.
February 1982 CDC Statement
By late 1981, the CDC concluded the evidence was sufficiently strong that “caution when administering salicylates” was no longer sufficient. The association was strong, but causality was more difficult to prove. After consulting various groups, we decided to be totally transparent in saying that causality could not be proved but that doctors and parents were entitled to have all of the information that we possessed. Plans were made to publish the accumulated information plus the recommendation to withhold aspirin from children having chickenpox or flu in the MMWR on February 12, 1982.
The pressure from aspirin manufacturers took me by surprise. One would think that if a problem existed with aspirin use in children with influenza or chickenpox, the manufacturers would want to know that to protect both children and themselves. That was not the reaction at all. First, representatives from the aspirin manufacturers came to the CDC with boxes of “new” data, which we were obliged to examine. We found nothing new, but it took weeks of study to determine that. Next, the aspirin manufacturers began sending me messages, calling me at home and even at my parents’ home when I went for a visit. The messages and pleas were varied, but the bottom line was always that the CDC’s reputation was at stake and did I want to be the one that ruined its record for scientific integrity? All it would take is to make an announcement that aspirin should not be given under certain conditions and if that proved to be incorrect, the CDC’s reputation would be soiled. Then they would ask, “Do any of the studies reach statistical significance?”
We proceeded with plans to publish a review in the MMWR. The issue would also contain a joint statement with the FDA, issuing a stronger warning about the use of aspirin.
The night before publication of the MMWR article, I received a call from the FDA. That agency said that the aspirin manufacturers had come late that day with new information on the use of aspirin and Reye syndrome. FDA felt obligated to review that information and so could not sign on to a statement until that had been completed.
The aspirin manufacturers anticipated that the government would have to delay publication of a recommendation, and they may have gone to bed that night content that they had bought more time. Such stalling tactics could have continued on for an extended period. But the CDC took them by surprise. The next day, February 12, 1982, the MMWR article (6) was published without the FDA’s signoff.
The MMWR reviewed what had happened since the original Arizona report on seven cases and indicated that a fourth study had now been reported. Michigan had conducted a second study during the 1979–1980 influenza season, concentrating on all medications used with attempts to get dosage figures and frequency of administration; this study had one to three controls for each patient. Interviews were conducted within days of the report of the case to improve the accuracy of reporting. All twelve case-patients had used salicylates versus 41 percent of the twenty-nine control children (6).
The CDC then went on to note, “After reviewing the data from all four studies and discussing the various epidemiologic and analytic methods and results, the CDC consultants concluded that it was unlikely for the limitations of the studies, either singly or in combination, to explain totally the strength and consistency of the observed association between Reye syndrome and salicylates.” The consultants felt that there was “sufficient evidence to support the cautionary statements on salicylate usage that had been published previously by the Centers for Disease Control.” Furthermore, it was the consensus of the consultants that “until the nature of the association between salicylates and Reye syndrome is clarified, the use of salicylates should be avoided, when possible, for children with varicella infection and during presumed influenza outbreaks.” The report continued: “In summary, these studies indicate to CDC that salicylates may be a factor in the pathogenesis of Reye syndrome, although the observed epidemiologic association does not prove causality. The exact pathogenesis of this disease and the possible role of salicylates in its pathogenesis remain to be determined. Additional well-controlled studies are also needed. Until definitive information is available, CDC advises physicians and parents of the possible increased risk of Reye syndrome associated with the use of salicylates for children with chickenpox or influenza-like illness” (6).
It may not seem so, but this was a clear step beyond saying to exercise caution. It was a recommendation to withhold salicylates. The reaction was swift. The aspirin manufacturers went to my supervisor, Dr. Ed Brandt, assistant secretary for health at HHS. He could have referred them to me, saying he had not been involved in the statement. Instead, he told them that he fully supported my authority to publish the article and that he agreed with the conclusions.
The aspirin manufacturers then went to the secretary of HHS, Richard Schweiker, who also gave the CDC his full support. The manufacturers next went to the White House. And the White House ordered the CDC to “cease and desist” and to start a new study of the association of Reye syndrome and aspirin. This was done. But “cease and desist” orders from the White House were of little interest to pediatricians, and the order from the White House could not take away the words of that publication. Pediatricians stopped using aspirin during flu outbreaks and for chickenpox treatment.
Four months later, a Surgeon General Advisory was issued in the June 11, 1982, MMWR (7). This, of course, was an official statement and therefore cleared by the White House. The report said, “The Surgeon General notes that the matter has been reviewed recently by several groups from within and outside government” and went on:
On January 11, 1985, the MMWR reported that “during 1982–1984, the annual incidence of RS [Reye syndrome] was the lowest reported since the initiation of national surveillance” (9). The report went on to describe the new study ordered by the White House:
A pilot study was conducted between February and March 1984 to determine the study feasibility and establish methodology. The pilot study included 29 RS cases and 143 controls consisting of children admitted to the same hospital (IP) or emergency room (ER), attending the same school, or identified by random-digit dialing (RDD). Ninety-seven percent of case children were reported to have received salicylates during the respiratory or chickenpox illness before a clinically defined onset of RS, compared with 28% (ER), 23% (IP), 59% (school), and 55% (RDD) at any time during their matched illnesses. The risk defined in the pilot study was comparable to or greater than that determined in the previous studies. The Institute of Medicine (IOM), National Academy of Sciences, served to advise and critique the protocol, monitor the study progress, and review study analysis and results.
The MMWR went on to state that, after the IOM reviewed the methods of data collection, in July 1984, and the data analysis, in December 1984, the IOM stated, on January 8:
The MMWR article concluded with the following: “A report of the pilot study is currently being prepared for publication. In view of these preliminary findings, physicians, parents, and older children who self-medicate should continue to be advised of the probable increased risk of RS associated with the use of salicylates for children, including teenagers, with influenza-like illness or chickenpox [reported by the Division of Viral Diseases, Center for Infectious Diseases, CDC; the Reye Syndrome Task Force, consisting of members from the FDA, NIH, and the Office of the Assistant Secretary of Health, and the CDC]” (9).
The Impact on Reye Syndrome
On February 7, 1986, the MMWR reported on the full 1985 Reye syndrome experience (10):
Following the results of the pilot phase of the U.S. Public Health Service study on RS and medications, the U.S. Food and Drug Administration has proposed that oral over-the-counter medicine containing aspirin add a label reading: WARNING: Children and teenagers should not use this medicine for chickenpox or flu symptoms before a doctor is consulted about Reye syndrome, a rare but serious disease.
For the 1985 surveillance year, 91 cases of Reye syndrome (RS) meeting CDC’s case definition were reported. Although delayed reports through June 1986 may increase the number of cases for 1985, the provisional 1985 total is less than half the lowest annual total [emphasis added] reported through the National Reye Syndrome Surveillance System (NRSSS) since its initiation in December 1973.
In previous years, the RS incidence, at least in part, has reflected the intensity and type of influenza activity. By all surveillance parameters, 1985 influenza activity was comparable in intensity to 1984, and the activity was greater than in 1982 or 1983.
And so we saw the beginning of a solution to the Reye syndrome problem. It involved all of the fears of parents; the suffering of children; the mystery of causation; the use of surveillance systems; epidemiology, risk calculations, and understanding the implications of government statements; the clash of the marketplace and public health; both the beneficial and problematic roles of government; and, finally, a solution communicated to medical workers and the general public to alter behaviors. The foundations of the solution were science and a systematic reduction of uncertainty until the association of aspirin in causing this problem could not be ignored. The system worked once again but not before many had their patience tested and many others suffered an unnecessary illness or death.
a The surgeon general used no reference but noted that the FDA would notify health professionals through its Drug Bulletin and would take the steps necessary to establish new labeling requirements for drugs containing salicylates.