Zika

by Donald G. McNeil

  Jenner’s discovery stands as a medical milestone, but in those days, most people still believed diseases were caused by bad air or an imbalance of humors. The “germ theory”—that disease was caused by creatures too small to be seen—was new, counterintuitive, and controversial, and many people rejected it. And the idea of deliberately sticking diseased pus into a child offended many average people, including many clergy, who railed against it as disgusting in itself and as defiance of God’s will.

  Blaming vaccines has become so routine that those rumors largely fell on deaf ears this time. There was no evidence that Brazil had bought bad rubella vaccine, because there had been no rubella outbreak nine months earlier. It was true that Brazil had relatively recently introduced a new diphtheria/tetanus/pertussis shot. But that was a change many countries, including the United States, had made years before. The new “acellular” component—made from broken-up pertussis bacteria instead of weakened whole-cell bacteria—was developed because the old vaccine had been blamed for occasionally causing seizures. It had nothing to do with microcephaly.

  The rumor that it was all just an overcount was the tough one.

  I’m not sure how it started—probably among scientists in Brazil, because some of them, even at the Cruz Foundation, definitely believed it.

  I heard it first from Jeffrey R. Powell, a highly respected Yale professor of evolutionary biology who, among other pursuits, studies the genetics of Aedes mosquitoes. His lab did research in Brazil and had just shifted its focus to include Zika.

  On January 28, he sent Simon Romero and me a draft op-ed piece he had written arguing that the microcephaly epidemic was a fiction.

  The editors had declined it, he said, but he thought we might be interested in his thinking.

  It was clearly a scientist’s work, concise and packed with evidence. It noted that the virus had been in Africa and Asia for decades, apparently without ever causing a microcephaly surge. It noted that Zika antibody tests were unreliable in anyone who had had dengue or yellow fever, which many Brazilians had. And its core tenet was that the Brazilian health ministry had, in the middle of its counting process, expanded its definition of a microcephalic head from one of 32 centimeters or less in diameter to one of 33 centimeters or less.

  “This change in definition,” he wrote, “increases by five-fold what is classified as microcephaly.” His piece ended, “If we are lucky, in a year or so we may look back and conclude that the panic now occurring, most acutely in Brazil, was not warranted.”

  If he was right, Simon and I and the New York Times would look pretty stupid. We had been featuring the epidemic on the front page for a month, pushing it harder than other media outlets.

  Even when they turned something down, the op-ed editors sometimes mentioned provocative ideas like that to newsroom editors. Also, submissions they rejected on occasion ended up in the Wall Street Journal or elsewhere. One way or the other, I was going to get quizzed about this.

  A month earlier, when I’d started on the story, I’d read all the PAHO reports. I thought I remembered reading that Brazil’s health ministry had changed its definition in mid-investigation. But what I remembered was that it had been changed in the opposite direction.

  I dug through old reports until I found what I remembered, and sent Dr. Powell a note: “Unless I am misreading this WHO/PAHO page, the change in definition of microcephaly that Brazil made last December was in the opposite direction: Previously, newborns with heads less than 33 centimeters were considered microcephalic, now they must be below 32 cm. Normally, that would mean that far fewer children would be found to be microcephalic, no?”

  Dr. Powell’s first reply was that it contradicted what he had heard from Brazil, and he wanted to double-check. He later wrote back saying, graciously, “Well, seems I was wrong, and I thank you very much, Donald, for correcting me.”

  But it didn’t end there. Wherever it had begun, the rumor was off and running. I was actually in California at the time, seeing my stepmother, who was declining from bone marrow disease, and then taking a break by driving down the coast. I was harder than usual for the desk to reach.

  Then, on February 3, Brazil released the results of its first analysis of thousands of reported microcephaly cases. Another colleague in Brazil, Vinod Sreeharsha, covered it.

  The results looked pretty damning:

  Since the previous October, 4,783 cases of microcephaly had been reported.

  The health ministry had thus far investigated 1,113 of them.

  Of those, only 404 had been confirmed as microcephaly.

  Of those 404, only 17 had tested positive for the Zika virus.

  One might easily conclude that the skeptics were right: it was all a miscount.

  I emailed Vinod. He was unhappy—his usual beat was business and political stories and, as a stringer, he lacked clout with the desk. Editors had read the report and worried that the earlier rumor was right. He had felt pressured to be cautious and emphasize the possibility that Brazil had overreported cases. Other news outlets were being even more emphatic in saying the numbers implied it might all be a mistake.

  Vinod had explained—correctly—how Brazil had tightened its microcephaly definition and had quoted both a Brazilian and an American expert calling that a medically sound decision. But those paragraphs were down near the end of the story, after a lot of copy stating an overcount was possible. Overall, the story served as a brief for the doubters.

  The headline emphasized the skepticism: “Birth Defects in Brazil May Be Overreported Amid Zika Fears.”

  The next morning, Dr. Powell wrote me again, saying the article was exactly what he had suspected. Ambiguous microcephaly definitions and bad testing, he said, had “conflated to set off the whole hullabaloo.”

  And that was where things stood, for a while. The data was the data. The WHO had declared a global emergency just two days before. If the skeptics were right, it too would look foolish.

  But I was sure the microcephaly was real, for one simple reason: in order to believe it was just a counting error, one had to assume that all the neonatal intensive care clinicians in at least four cities in Brazil’s northeast were mistaken. In interviews that Simon and Sabrina had sent, and many others I’d read, they had all said the same thing: For years they had been seeing at most two or three microcephalic babies a year. Now they were caring for a dozen at a time in their wards. No neonatal specialist just fails to notice a deformed head. Also, the babies in the pictures didn’t have heads just a centimeter or two below normal. They were truly tiny, and there were many pictures of them.

  In retrospect, the Brazilian health ministry may have erred in reporting the results of its investigation so early. It was a public relations disaster. The ministry was opting for transparency, but the small percentage of confirmed cases and tiny number with detectable Zika virus made it look as if the agency had cried wolf.

  A couple of weeks later, the ministry compounded its PR problems when it stopped reporting the number of unconfirmed cases. It made the change in an effort to squelch the rumors, but it looked like a cover-up.

  Since then, the confirmed count has climbed to over 1,400.

  It was no doubt true that Brazil had historically underreported microcephaly. With a population of 200 million people, it reported an average of 163 a year. In Europe and the United States, prevalence rates were at least two and maybe four times higher, depending on what definition of microcephaly you used.

  Even so, that didn’t come close to accounting for what had happened. Northeast Brazil, which is more sparsely populated than the south, normally reported 40 cases a year, a quarter of the national total. In just the six months from October 2015 to March 2016, the northeast states together reported 876 confirmed cases, nearly 90 percent of the national total.

  And there was a sensible explanation for why the virus was found in only 4 percent o
f the confirmed cases. Most of the mothers would have been infected in their first trimesters, six to nine months earlier. Antibodies usually wipe out live virus within two weeks. I was surprised there was live virus in any babies. It has since been noted, in blood tests on women and in the work on pregnant monkeys by Dr. O’Connor at the University of Wisconsin, that it sometimes does persist. How it does so is another medical mystery.

  Those were the rumors as of early February 2016—and the answers. People would have to wait for more evidence, and then decide whether they found it persuasive.

  10

  The Proof

  ON FEBRUARY 1, 2016, when the WHO declared its emergency based on the possibility that Zika caused microcephaly, reporters asked WHO officials exactly what evidence was needed to be sure it did.

  Initially, both Dr. Bruce Aylward, who was in charge of the response, and Dr. Heymann, the advisory committee chairman, gave the same answer: a large case-control study.

  Scientists in Latin America, they said, were already recruiting pregnant women into one. They had signed up about 5,000, mostly in Colombia, some in Brazil, some elsewhere; they had to be pregnant and to have come up positive on tests for the virus. Those were the “cases.” They were also signing up “controls”: a roughly equal number of pregnant women who did not have Zika. They would try to match the two cohorts as closely as possible: same ages, same races, same neighborhoods, same income levels, same medical histories, especially regarding previous dengue or chikungunya infections. (Obviously, if a woman in the control group got Zika during the study, she would be shifted to the case group.)

  They would monitor the two groups until their babies were born, and compare the results. If the Zika group had far more babies with microcephaly than the control group, they could definitively say Zika was the cause.

  This was a “prospective cohort study,” the gold standard in epidemiology. The women enrolled first were due to start giving birth in May and June, Dr. Aylward said, so final proof would have to wait until that data was ready.

  In fact, the science moved forward much faster.

  On March 31, without any fanfare, the WHO made a subtle but important change to one sentence on the face of its weekly Zika situation report. It read:

  “Based on observational, cohort and case-control studies there is strong scientific consensus that Zika virus is a cause of GBS, microcephaly and other neurological disorders.”

  “Strong scientific consensus” marked a shift from previous reports, which said it was “highly likely” Zika was a cause.

  Then on April 13, the CDC made it definitive. Its director, Dr. Frieden, scheduled an afternoon press conference with the leaders of his Zika team and declared unequivocally, “It is now clear: the CDC has concluded that Zika does cause microcephaly.”

  It was “an unprecedented association” in medicine, he added. “Never before in history has there been a situation where a bite from a mosquito can result in a devastating malformation.”

  What led the agencies to change their minds about waiting for the big study?

  A series of small studies.

  The number of cases of confirmed microcephaly in Brazil had just kept growing. Before the CDC announcement, it had passed the 1,000 mark, with nearly 900 clustered in the northeast.

  Microcephalic babies were by then being born not just in Brazil but in Colombia, in Panama, in Martinique, and in the Cape Verde Islands, not to mention the cluster discovered retrospectively in French Polynesia. Each cluster had followed a Zika outbreak about nine months earlier.

  Separate teams of doctors—in Brazil, in the United States, even in Slovenia—had found Zika virus in the brain tissue or amniotic fluid of babies who had been born with microcephaly, had died in the womb with microcephaly, or had been aborted because microcephaly was detected on ultrasound. One particularly grim case was described in the New England Journal of Medicine on March 30. It involved a 33-year-old newly pregnant Finnish woman living in Washington, DC, who had taken a quick trip through Belize, Guatemala, and Mexico over Thanksgiving 2015. She had a routine sonogram on December 5, and her baby was fine. But she was having odd symptoms—a rash, eye pain, and a fever. Over New Year’s, she was in Finland—presumably home for Christmas—and she must have read the news about Zika, which was just emerging then, and recognized her symptoms. She had a blood test and an ultrasound there. The ultrasound was normal. But her blood was positive for Zika virus. She went back to the United States and had the same two tests on January 5. Same results. Then, over the next three weeks, two things happened. Her blood remained positive for the virus, which was abnormal. And, horribly, her baby’s brain began to dissolve. On her next MRI and ultrasound, at 19 and 20 weeks, the skull is the right size, but the surface of the brain has thinned out, the hollow spaces in the frontal lobes are larger than they should be, and the white matter that connects the two hemispheres is far smaller than it should be. At week 21, she decided to terminate the pregnancy. On autopsy, the brain was found to be teeming with viral particles. That was solid evidence.

  There was also “biological plausibility,” the CDC said. Biologists at Florida State University had tested the virus in several types of fetal cells that grow into the various parts of a baby. It barely infected some, such as the prekidney cells. But it homed in on the neural progenitor cells—the ones that ultimately turn into the brain—and destroyed them.

  (Also, in results that were then still unpublished, when injected into immune-deficient mice, the virus did not kill adults but did kill fetal ones, and it was found in their brains.)

  But the most convincing and most frightening piece of evidence was a miniature case-control study published by the New England Journal of Medicine on March 4. It was done by doctors at the Cruz Foundation in Rio working with a team from the David Geffen School of Medicine at UCLA. They described a group of 88 pregnant women whom they had started to enroll in September 2015, when the reports of microcephaly began coming out of the northeast. (Although the epidemic’s initial epicenter was in that region, there was a simultaneous smaller outbreak in some parts of the Rio region.) The researchers were already in the middle of a dengue study, and they had noticed a few months earlier that they were getting many patients with fevers and rashes who tested negative for dengue. So they started testing them for Zika, and then began a substudy that looked just at the pregnant ones.

  The study wasn’t nearly over when they published the results. They had rushed it into print because what they were finding was so alarming that it needed to serve as a warning.

  Rather than test all their subjects, they had chosen rashes as the recruitment factor. Whenever they saw a pregnant woman with a rash, they asked whether she would agree to participate. Eighty-eight had said yes. Of those 88, 72 tested positive for Zika. The other 16 became the “controls.” Of the 72, 2 had miscarriages almost immediately. That didn’t necessarily mean anything—miscarriages in early pregnancy are common. Of the 70 left, 42 agreed to have ultrasounds every few weeks. The other 28 refused. Some said the ultrasound clinic was too far away. But some “declined because of fear of abnormalities.” That is, they preferred not to know whether their babies were deformed. They would find out at birth.

  By the time the authors published their preliminary results, 12 of the 42 mothers having ultrasounds were showing evidence of “grave outcomes.” Two babies had had normal ultrasounds, and then had suddenly died in the womb. Both of those mothers had been infected late in their pregnancies, not in the first trimester. The rest had ultrasounds revealing serious defects: some had microcephaly, some had white spots—brain calcifications—suggesting inflammation or cell death, some babies were much too small for their gestational age, some had almost no blood flow in their umbilical cords. By the time the study was published, 8 of the women had given birth, and the ultrasounds had proven accurate.

  Twelve damaged babies out of 42 was a 29 perc
ent “grave outcome” rate. The 16 Zika-free women acting as controls had zero problems. A difference of 29 percent versus 0 percent is more than “statistically significant.” It’s overwhelming. Among other things, those results forced experts to stop saying that the danger was all in the first trimester. Clearly, Zika could kill babies at any point.

  “We were just blown away by that,” Dr. Karin Nielsen-Saines, one of the authors, said. “We weren’t expecting to find problems in all trimesters.”

  (A study done by the CDC that came out later, on May 25, in the New England Journal of Medicine looked at Brazil’s northeastern Bahia state during the height of its epidemic and found that a Zika-infected woman’s risk of having a microcephalic child was between one and 13 percent.)

  The stack of evidence piled up by all these disparate studies, the CDC said, fulfilled “Shepard’s criteria.” That was a set of conditions published in 1994 by a professor of pediatrics at the University of Washington, Dr. Thomas H. Shepard, for determining whether a particular insult to a fetus caused a particular birth defect. (They were different from Koch’s postulates, a better-known set of criteria published by the pioneering German microbiologist Robert Koch in 1890. But Koch’s postulates are for concluding whether a particular germ causes a particular disease. Shepard’s criteria relate to birth defects and incorporate nondisease causes like poisons or radiation.)