More controversy arises over NHANES latex results

Flaws mar phase two; phase one shows no clear link

A survey that federal researchers hoped would shed light on latex allergy among health care workers has instead led to disclaimers and more controversy. The Centers for Disease Control and Prevention in Atlanta tested blood serum collected in the National Health and Nutrition Examination Survey (NHANES III) for the presence of latex specific IgE antibody. CDC researchers had hoped to compare the prevalence of latex sensitivity in health care workers and the general population.

But before it could analyze its first phase of data collection, the CDC learned of serious quality problems with the second phase, which had been coordinated by the Food and Drug Administration (FDA). The CDC now says it will not publish an analysis using either phase one or phase two data, but will wait for new NHANES data that will become available in 2002.

However, using the phase one dataset that is in the public domain, outside researchers published a report in the March 15, 2001, issue of American Journal of Epidemiology asserting that the survey showed no significant difference in latex sensitivity among current health care workers using gloves with those in other occupations. Both groups had prevalence rates of about 18% to 20%.1 The researchers found a weak association between latex sensitization and those who reported their longest-held job was in health care, and strong association between childhood atopy, or allergy, and sensitivity. Certain information, such as the type of glove used (i.e., powdered or powder-free) was not available.

Allegiance Healthcare Corp. of McGaw Park, IL, a leading manufacturer of latex gloves, sponsored the analysis, and many of the conclusions have previously appeared on the Allegiance web site. That fact made it a lightning rod for criticism by some researchers who study latex allergy in health care workers and by advocates for the latex allergic.

All scientific work has either governmental or private sponsorship, and the sponsorship does not affect the objectivity or reliability of the analysis, responds David Garabrant, MD, professor of occupational medicine at the University of Michigan School of Public Health in Ann Arbor, and lead researcher on the latex study.

The NHANES analysis included 174 current health care workers who reported on glove use, 37 of them with latex sensitivity. The comparison group included 5,335 nonhealth care workers. Only adults ages 17 to 60 were surveyed and tested. The findings are important because they offer a rare opportunity to directly compare health care worker sensitivity with the prevalence in the general population, says Garabrant. "There’s no other study published that makes any comparison of the risk of sensitization among health care workers to the risk of sensitization in other occupations."

Garabrant and his colleagues concluded that, "Health care workers are not at clearly increased risk of latex sensitization compared to other workers in the United States."

Differing rates raise questions

CDC researchers involved in the NHANES study design were not asked to review the article and didn’t even know of its existence until it was published. But the journal’s editors published an invited commentary that questioned the limitations of the study and the high prevalence found among both health care workers and those in other occupations.2

Noting that previous reports indicated the prevalence of latex sensitivity among nonhealth care workers ranges from 0 to 12%, the commentary questioned the higher rates found in the NHANES data. "The way you learn from studies is by explaining differences. I didn’t say it was invalid. I’m saying, how do you explain that difference?" says Daniel Wartenberg, PhD, professor of environmental and community medicine at the Robert Wood Johnson Medical School of the University of Medicine and Dentistry of New Jersey in Piscataway. Wartenberg co-authored the commentary with his colleague, Gail Buckler.

"I think they should have given a more thorough discussion of some of the limitations," Wartenberg tells Hospital Employee Health. "That would have been beneficial both in terms of their methodology and the confidence they could have with their results regarding the issue of whether exposure to latex caused the sensitivity."

Garabrant countered that the NHANES prevalence rates are not out of line with some studies of latex sensitivity among the general population.3 Moreover, researchers were able to control for atopy, age, sex, and other demographic factors in the NHANES data. "That’s why the NHANES data are far more important than most of those other subjective comparisons because there’s a direct comparison between health care workers and other sectors of the population in which we can control for differences in age, sex, race, and atopy," says Garabrant.

Quality problems plague phase one

Meanwhile, overshadowing the NHANES analysis are questions about the testing methods and the loss of phase two data. NHANES III was designed as a six-year cycle, spanning 1988 to 1994. CDC personnel tested the first phase sera using a manual method of the AlaSTAT assay, a test developed by Diagnostic Products Corp. of Los Angeles. The second phase used a newer version of the assay, with an automated method. CDC researchers still expected the results to be comparable "since both were from the same manufacturer and used allergens in a liquid format, with the newer assay using a microplate format," according to a published CDC statement.

By the fall of 1998, problems became apparent at the FDA laboratory in Dallas. When phase one samples were retested with the newer test, only half showed similar results between the two laboratories. Then quality control issues emerged at the FDA lab. "There were times when the same serum sample was tested again and the results were not the same on duplicate runs," explains Lawrence D’Hoostelaere, PhD, senior scientific coordinator in the Division of Field Science at the FDA’s Office of Regulatory Affairs. "There was indication that some of the samples had insufficient volume remaining to do the test. The computer reported those out as no IgE or no antigen," he says. "It wasn’t tracked closely enough to distinguish those that had insufficient serum from those that had no antigen or antibody."

With changes in personnel and the closing of the laboratory, some paperwork was discarded. Other documentation was insufficient. Ultimately, an FDA review concluded that the phase two data are "not fit for scientific purposes."

While the validity of the phase one data has not been called into question, the CDC does issue a caution: "Because the AlaStat assay has changed to an automated microplate method from the older manual tube method that was used to produce the latex allergy data from the first phase of NHANES III, analyses of these data should take this into account when comparing with results from other, more recent studies."

CDC also notes that using the test to screen a general population sample "is known to produce more variable results than screening in high prevalence populations."

For those seeking answers about the risk of latex allergy among health care workers, the studies continue. Gordon Sussman, MD, associate professor of medicine at the University of Toronto and a researcher of latex allergy, offers his clinical perspective. While screening tests can produce false positives, he says, "we know health care workers have latex allergy because we see them. You see much more latex allergy in health care workers than in the general public. That’s clinical, not just a lab test."


1. Garabrant DH, Roth HD, Parsad R, et al. Latex sensitization in health care workers and in the U.S. general population. Am J Epidemiol 2001; 153:515-522.

2. Wartenberg D, Buckler G. Invited commentary: Assessing latex sensitization using data from NHANES III. Am J Epidemiol 2001; 153:523-526.

3. Garabrant DH, Roth HD, Parsad R, et al. Garabrant et al. respond to Wartenberg and Buckler. Am J Epidemiol 2001; 153:527-528.