Oct 2, 2009 (CIDRAP News) – In findings sure to renew the continuing controversy over respiratory protection for healthcare workers, surgical masks appeared to protect hospital nurses from influenza about as well as N-95 respirators did in a randomized trial conducted in Ontario.
There were only two more confirmed flu cases among a group of more than 200 mask-wearing nurses than in a similar size group of nurses wearing N-95 respirators, according to the report published yesterday by the Journal of the American Medical Association (JAMA). The result met a statistical test for showing that the masks were "noninferior" to the respirators.
But other experts said today that the study has important limitations—including the lack of a control group using no respiratory protection—that cast doubt on the findings.
The report is described as the first randomized trial comparing different forms of respiratory protection against flu to reach publication. It comes a few weeks after the news of a study by Australian researchers in which N-95 respirators were found to be clearly better than surgical masks for preventing flu in healthcare workers. That study was reported at a medical meeting but has not yet been published in a journal.
N-95 respirators are designed to fit closely to the face and filter out at least 95% of airborne particles, whereas surgical masks fit more loosely and were originally designed to prevent the wearer from infecting others.
But health workers say the tight-fitting N-95s are uncomfortable and difficult to wear for long periods, and hospitals sometimes have trouble keeping them in supply. Surgical masks are more comfortable and cheaper, but scientists have not found much evidence that they protect wearers from respiratory pathogens.
Respiratory protection for health workers has been a big issue since the emergence of the novel H1N1 virus. In early September, the National Academy of Sciences' Institute of Medicine (IOM) issued a report affirming the current Centers for Disease Control and Prevention (CDC) guidance on the topic, which recommends use of N-95s by all healthcare workers who enter the rooms of patients with confirmed or suspected H1N1 infection. The same advice goes for emergency medical personnel who come in close contact with such patients.
The Canadian researchers, led by Mark Loeb, MD, MSc, of McMaster University in Hamilton, Ont., recruited 446 nurses who worked in emergency departments and medical and pediatric wards in eight Ontario hospitals.
They were randomly assigned to wear either a surgical mask or a fit-tested N-95 respirator when caring for patients with febrile respiratory illness. The nurses continued to use their assigned respiratory protection during aerosol-generating procedures such as intubation, provided tuberculosis was not suspected.
The team confirmed influenza in the volunteers by either of two methods: detection of viral RNA in patient specimens by reverse-transcriptase polymerase chain reaction (RT-PCR) or at least a four-fold rise in serum antibodies to circulating flu strains at the end of the study period.
To reduce errors, the serologic criteria for infection with seasonal flu strains were used only for nurses who had not received the seasonal flu vaccine. About 30% of the mask group and 28% of the N-95 group had been vaccinated.
To check whether nurses were actually wearing the assigned protection, the researchers called hospital units daily during the peak flu period in March to find out if they had any patients with flu or febrile respiratory illness. If they did, an auditor was sent to the unit to observe for compliance by watching discreetly from outside the patient's room.
In the end, flu was confirmed in 50 of 212 nurses (23.6%) in the mask group and 48 of 210 nurses (22.9%) in the N-95 group, the report says. That signaled a difference in absolute risk of only 0.73% (95% confidence interval [CI], -8.8% to 7.3%, P=.86), indicating the masks were not inferior to the respirators.
Most of the cases were confirmed by serology; confirmation was by RT-PCR in only six in the mask group and four in the N-95 group.
The authors also looked at several other outcomes, including non-flu respiratory infections and influenza-like illness (ILI), and found no significant differences between the two groups. Nine nurses in the mask group had an ILI, versus only 2 in the N-95 group, indicating the risk for the N-95 group was 3.3% lower (95% CI, -6.31% to 0.28%, P=.06). All 11 nurses had lab-confirmed flu.
Serologic evidence of infection with the emerging novel H1N1 virus was also found in the study, in 8.0% of the mask group and 11.9% of the N-95 group. The findings signaled noninferiority for the masks.
There were 18 episodes in which the researchers checked whether nurses were using their assigned protection in the prescribed situations. They found that all 11 nurses in the mask group and 6 of 7 in the N-95 group (100% versus 86%) were complying.
Conclusions and limitations
The authors conclude that the incidence of flu in the two groups was similar. "Surgical masks had an estimated efficacy within 1% of N-95 respirators," they state.
They acknowledge several limitations of their study, including an inability to assess compliance by all participants and no monitoring of hand hygiene or the use of gowns and gloves. They say it is impossible to know whether the volunteers contracted flu because of exposure in the hospital or in the community, but add that their data on household exposure suggest that it was similar between the two groups.
"We acknowledge that not surveying participants' coworkers about influenza-like illness was a limitation," they state. They also caution that their findings apply to routine care and should not be generalized to settings where flu viruses can be aerosolized, such as intubation, where use of N-95s is recommended.
Experts note drawbacks
Other experts who commented on the study today cited the inability to assess possible differences in flu exposure between the two groups as a drawback.
Kristine Moore, MD, MPH, medical director of the University of Minnesota Center for Infectious Disease Research and Policy, publisher of CIDRAP News, called the study an important step in understanding respiratory protection for health workers, but said she was concerned about possible differences in flu exposures.
"It may be that the exposure risks within the healthcare facilities were different for the two groups (ie, if one group had more direct contacts with influenza patients than the other group)," she commented by e-mail.
"Also, unlike [for] a pathogen such as drug-resistant Mycobacterium tuberculosis, much of the exposure to influenza occurs in the community setting rather than in a healthcare facility," Moore added. "The investigators had no way to determine the attributable risk of community exposures versus the attributable risk of exposures within the healthcare facility.
"Even though they found comparable histories of ILI among household members, this information may not have accurately reflected true influenza rates in household members and does not address other community exposures. A high attributable risk for community exposure could have accounted for the similar infection rates among the respirator and mask groups.”
Lisa Brosseau, ScD, MS, an associate professor of environmental science at the University of Minnesota School of Public Health and veteran researcher in respiratory protection, voiced similar concerns. Calling the finding of no difference in protection surprising, she said the limitations acknowledged by the authors and other problems may account for it.
In particular, there was no control group of nurses working in the same hospitals but not using respiratory protection—unlike in the recently reported study by Australian researchers, she said. "Thus, it is impossible to say whether either 'mask' made a difference in healthcare worker infection rates," she said.
Brosseau also noted that the limited auditing in the study indicated only 86% compliance in the N-95 group, versus 100% in the surgical mask group. Lower compliance in the N-95 group could lead to higher direct exposure to influenza in that group, distorting the results, she said.
Also, she said, because the observations of compliance were very limited in nature and number, it's not possible to know such things as whether the nurses used respirators that matched their fit test, checked the seal, or wore them the whole time they were in the patient's room.
"I don't think this is the definitive study everyone is looking for, but rather illustrates the great difficulties involved in conducting studies of personal protective equipment effectiveness in the 'real world,'" Brosseau said.
Editorial cites strengths, weaknesses
An accompanying JAMA editorial by two other experts praised the study on several counts. The authors, Arjun Srinivasan, MD, of the CDC and Trish M. Perl, MD, MSc, of Johns Hopkins University, said the two study groups "were well balanced with similar risk factors for influenza infection, including vaccination and febrile respiratory illness among household members, and participants were evenly distributed across study hospital wards."
On the other hand, the writers note that the researchers "did not directly assess exposure risks by collecting data on the number of nurse contacts with patients" who had febrile respiratory illness.
Srinivasan and Perl also comment that the controversy over respiratory protection has distracted attention from the need to use other strategies to prevent flu transmission in healthcare settings—chief among them, annual vaccination of healthcare workers.
Loeb M, Dafoe N, Mahony J, et al. Surgical mask vs N95 respirator for preventing influenza among health care workers. JAMA 2009; early online publication Oct 1 [Full text]
Srinivasan A, Perl TM. Respiratory protecton against influenza. (Editorial) JAMA 2009; early online publication Oct 1
Sep 17 CIDRAP News story "Study on respirators versus masks hailed as landmark"