Politics blurs the science of respiratory protection
Two studies at center of debate over N95s
Surgical masks are no worse than respirators in protecting health care workers from influenza. Is this statement based on science or politics?
In the absence of adequate science-based evidence, respirators should be used by health care workers having close contact with patients with a novel influenza strain. Should this policy take into account practical issues, such as cost and supply concerns?
The current pandemic has raised the stakes on a long-simmering dispute between infection control practitioners (who favor the use of surgical masks) and industrial hygienists (who insist that only respirators can be used as personal protective equipment). Two recent studies one yet unpublished have added fuel to the debate.
The result is a kind of power struggle. Although the Centers for Disease Control and Prevention issued guidelines calling for health care workers to use respirators when caring for patients with novel H1N1, and the U.S. Occupational Safety and Health Administration directed its inspectors to enforce those guidelines, professional organizations are still arguing the point.
In a Nov. 5 letter to President Obama, the Society for Healthcare Epidemiology of America (SHEA), the Infectious Diseases Society of America (IDSA), and the Association for Professionals in Infection Control and Epidemiology (APIC) called for an "immediate moratorium" on OSHA enforcement of the respirator guideline. They cited the two studies as evidence that "N95 respirators are not superior to surgical masks in the prevention of transmission of influenza in most patient care settings."
At the same time, Raina MacIntyre, MD, MBBS, FRACP, FAFPHM, MAE, PhD, professor of infectious diseases epidemiology at the University of New South Wales in Sydney, Australia, felt the heat of a political response to her scientific findings, which she said do not actually support the use of surgical masks.
Ideology and bias?
MacIntyre presented results to an Institute of Medicine panel that showed respirators were significantly more effective than surgical masks at preventing influenza-like illness and lab-confirmed influenza. Although the IOM panel said it did not consider any unpublished data in its decision to recommend use of respirators rather than surgical masks, MacIntyre received an angry barrage from infection control practitioners, whom she called "highly emotional and entrenched and ideological."
"I have received hate mail accusing me of being on the payroll of 3M [a major respirator manufacturer]. I have been attacked in the most personal way. You have no idea what's gone through my [e-mail] inbox," says MacIntyre, who says she had no agenda when she approached the research and in fact expected to find no difference between the two. "I have many times been surprised by findings and that the answers are not what I expect them to be, but I go with the truth and not political conditions or ideology."
The other paper at issue was published in the Journal of the American Medical Association, written by lead author Mark Loeb, MD, MSc, FRCPC, director of the infectious diseases department at McMaster University in Hamilton, Ontario, where he also is professor of pathology and molecular medicine and clinical epidemiology and biostatistics. Loeb found that surgical masks were not inferior to N95 respirators.
"I'm not necessarily in one camp or the other. We did the study because we wanted to ask the question," says Loeb, who also decries the politicization. "There are people who are so biased one way or the other that it detracts from moving the field forward."
Loeb takes issue with industrial hygienists who base their position on studies involving inert particles, rather than infectious particles. "What's most important is, 'How do these devices protect health care workers?'"
People who are "dead-set supportive of N95s" aren't looking at the evidence impartially, he contends. "Some people just tend to take whatever data is available and shape it to whatever conclusion they want to see, and I think that's counterproductive," he says.
N95s reduce flu by 75%
So what do the MacIntyre and Loeb studies say?
The MacIntyre study involved 1,936 health care workers at 24 Beijing hospitals who wore masks, respirators, or no facial protection for four weeks during the winter. Researchers tracked the health care workers for five weeks to detect onset of respiratory illness. The doctors and nurses wore the protection throughout their shifts.
Overall, respirators were 42% more protective than masks, although no statistical difference was found between fit-tested and nonfit-tested respirators. The respirators were associated with a 75% reduction in both flu-like illness and laboratory-confirmed flu.
The control group in the study was not randomized because health care workers in China would object to being told not to wear masks. Therefore, MacIntyre and her colleagues used a convenience sample as a control from hospitals where mask use was not routine.
When a journal reviewer said the control group should be removed because it was not randomized, MacIntyre revised her analysis. The p value was also adjusted to take into account the potential for different rates of infection at different hospitals in the study. The trend remained showing N95 respirators as more protective, but the results lost statistical significance, she says. The new analysis was presented at an IDSA meeting.
"The rates of all outcomes in the control arm were higher than in the surgical or N95 groups, so removing over 500 health workers from the analysis contributed to the p values losing significance," MacIntyre says. "What this means is that the study still shows a likely superiority of N95s, with half the rate of infection compared to surgical [masks], but the study was probably underpowered to pick up statistical significance when we removed the control group."
An analysis including the control group is valid, and the two analyses are simply two different ways of looking at the same evidence, she says.
Asymptomatic flu with N95s and masks
The Loeb study compared surgical masks and N95 respirators used by 446 nurses at eight hospitals in Toronto. The nurses wore the masks or respirators when caring for patients with febrile respiratory illness. There was no control group in the study.
Loeb and colleagues tested for laboratory-confirmed influenza based on either PCR from "nasopharyngeal and flockednasal specimens or at least a fourfold increase in serum antibodiesto circulating influenza strain antigens." About 30% of the surgical mask group and 28% of the N95 group had been vaccinated against seasonal influenza. The groups were well-balanced in terms of vaccination as well as household exposure to influenza-like illness.
The researchers found a fairly high rate of asymptomatic influenza in both groups 24% in the surgical mask group and 23% in the N95 group. That represented nonsuperiority of N95 respirators, Loeb concluded.
Eleven of the nurses had influenza-like illness and laboratory-confirmed influenza. Nine of them were in the surgical mask group and two in the respirator group. With a p value of 0.06, this did not meet Loeb's standard of statistical significance of p = 0.05. In fact, the difference was "probably more a function of how we defined influenza-like illness," Loeb says.
Interestingly, the study found a substantial amount of 2009 H1N1, although the final serology was completed by mid-May. The novel H1N1 strain was first identified in Mexico in April and was just beginning to circulate in North America. Eight percent of the surgical mask group and 11.9% of the N95 respirator group had serologic evidence of 2009 H1N1 infection. "We suspect there had been pandemic H1N1 circulating earlier on [than suspected]," Loeb says.
The result also could have been influenced by some cross-reactivity to seasonal strains. Only results related to the seasonal antigens were included in the study, he says.
Everyone agrees that more research is needed on the efficacy of surgical masks and N95 respirators, including their benefits in preventing infection. But in the midst of a pandemic, public policy must proceed.
The Institute of Medicine's stance mirrored what has been known as "the precautionary principle." Respirators should be used "until or unless further evidence can be provided to the effect that other forms of protection or other guidelines are equally or more effective" at protecting health care workers.
The IOM panel noted that it considered the experimental work that compared respirators and surgical masks, not studies on "their effectiveness in the clinical setting due to the fact that the availability of data is quite limited on clinical effectiveness."
That conclusion in favor of respirators was supported by the National Institute for Occupational Safety and Health (an arm of CDC), the American Public Health Association and the American Industrial Hygiene Association (AIHA), as well as labor unions representing health care workers. "We believe that there is clear information, based upon the evidence-based science, that N95 and higher types of respiratory protection provide a superior benefit than surgical masks," says Steve Derman, past chair of the Health Care Working Group of AIHA and president of Medishare Environmental Health and Safety Services in Coopertino, CA.
However, the infection control community strenuously argued in favor of practical considerations and questioned the added benefit of respirators. "When you take the body of evidence that is available currently in evaluating whether N95 respirators or surgical masks were appropriate for health care workers, we think the evidence favors surgical masks for routine care of patients," says Mark Rupp, MD, president of SHEA and medical director of health care epidemiology and infection control at the University of Nebraska Medical Center in Omaha.
Among the practical considerations: lack of comfort and tolerability, cost, supply, and communication difficulties. The increase in H1N1 vaccination and evidence that 2009 H1N1 has a lower fatality rate than seasonal influenza further fuel their argument.
"We feel very strongly that this is a real waste of resources and it puts an unnecessary burden on the health care system at a time when [hospitals] are already strained," Rupp says.
The American College of Occupational and Environmental Medicine (ACOEM) has been largely silent on this contentious issue. Occupational medicine physicians differ in their opinions about the need for respirators versus surgical masks, says Robert K. McLellan, MD, MPH, FACOEM, chief of occupational and environmental medicine at Dartmouth-Hitchcock Medical Center in Lebanon, NH.
However, ACOEM would like to shift attention to other measures that should be taken to protect health care workers from novel H1N1 and other infectious diseases. For example, through "administrative controls," patients with febrile respiratory illness should be identified swiftly and sick leave policies should encourage ill health care workers to stay at home. Barriers can be used in triage areas to protect workers, and vaccination can reduce the risk of transmission. Ultraviolet germicidal irradiation may have promise in reducing the risk of infectious diseases as well.
"There's been this kind of evolution of the use of respirators where they have taken too much prominence in the world of protecting health care workers," McLellan says.
Health care workers need to be educated about infection control practices and respirator use, he says. Improper use of respirators can actually increase the risk of infection, he adds.
"We're a strong proponent of well-defined infection control policies. Educate, vaccinate, and put in place good infection control policies," says McLellan.
"We need better science on transmission of flu. . . . Let's look at all the issues and problems we have in the use of anything [protective] over your mouth and let's try to push the science of respirators forward," he says.
1. Loeb M, Dafoe N, Mahony J, et al. Surgical mask vs. N95 respirator for preventing influenza among health care workers. JAMA 2009; 302:1,865-1,871. Published online Oct. 1, 2009 (doi:10.1001/jama.2009.1466). Accessed on December 14, 2009.