March 13, 2003
Kent D. McClure, DVM, JD
Animal Health Institute
1325 G. Street, NW, Ste. 700
Washington, DC 20005-3104
RE: Request for Correction of Information; letters and accompanying materials dated December 6, 2003, and January 8, 2003, to Julie Gerberding, MD, MPH, Director of the Centers for Disease Control and Prevention (CDC) and Administrator for the Agency for Toxic Substances and Disease Registry (ATSDR) from Kent D. McClure, DVM, JD, General Counsel for the Animal Health Institute (amendment dated January 8, 2003)
Dear Dr. McClure:
Thank you for your request for correction. CDC program and scientific staff who are experts in foodborne infectious disease have carefully reviewed your letters and accompanying materials dated December 6, 2002, and January 8, 2003. In your letters, you made several statements indicating that CDC Campylobacter data and analyses have not been consistent with CDC's Information Quality Guidelines. We would like to respond to your comments in the order that you presented them.
CDC seeks to accomplish its mission by working with partners throughout the nation and world to monitor health, detect and investigate health problems, conduct research to enhance prevention, develop and advocate sound public health policies, implement prevention strategies and programs, promote healthy behaviors, foster safe and healthful environments, and provide leadership and training. Monitoring is accomplished through public health surveillance, the ongoing, systematic collection, analysis, and interpretation of data essential to the planning, implementation, and evaluation of public health practice. For surveillance efforts to be complete, data should be synthesized, disseminated in a timely way, and integrated closely with decision making about prevention and control.
In general, bacterial infectious disease surveillance data have depended on collection of information and bacterial strains from established clinical laboratory sources, tested using appropriate and standardized methods, with established laboratory quality control procedures, and analyzed with standard statistical techniques. Before such information is disseminated, it is reviewed for medical, scientific and public health accuracy, soundness and utility by agency experts.
Comment 1. "Extended illness" etc. is not supported by "CDC Dataset" or by "Smith, et al."
CDC Response: The CDC/FoodNet case-control study of campylobacteriosis ("CDC Dataset") and by Smith KE, et al. (Quinolone-resistant Campylobacter jejuni infections in Minnesota, 1992-1998. New England Journal of Medicine. 340(20):1525-32, 1999 May 20) show that persons infected with Campylobacter that is resistant to fluoroquinolones have a longer illness than do persons infected with susceptible strains of Campylobacter.
CDC Dataset: Analysis of data collected in 1998-1999 as part of a case-control study of Campylobacter infections demonstrated a statistically significant difference in duration of illness, supporting the statement. The analysis included persons who had recently traveled outside the United States (probably acquiring infection in foreign countries). Because there are no demonstrated differences in the virulence or clinical impact of primary Campylobacter infection among countries, there is no scientific reason to exclude travelers from the dataset or analysis. The analysis excluded persons who had taken strong antidiarrheal medications because the duration of their diarrhea, the dependent variable, could be modified by such treatment, confounding the effect of antimicrobial resistance.
Smith, et al.: Data and analysis published in the report by Smith et al., in the New England Journal of Medicine, a peer-reviewed scientific journal, supports the statement. Data collection and analytic methods as described in the article appear to be appropriate and support the conclusions reached by the authors. Reviewing original data of peer-reviewed journal articles before citation of these articles is not standard practice.
Comment 2: "Rapid rise in resistance attributable to poultry" is not supported by data.
CDC Response: The statement that a rise occurred is supported by analysis of the data from the National Antimicrobial Resistance Monitoring System for Enteric Bacteria (NARMS-EB). The NARMS-EB surveillance program for Campylobacter is a sentinel clinical laboratory program that began in 5 sites in 1997, when it was added to Salmonella and E. coli O157 surveillance that began in 1996. At that time, Salmonella and E. coli O157 isolates were routinely referred to State Public Health Laboratories for further characterization from the clinical laboratories that isolated them, while Campylobacter isolates usually were not referred. As a result, the sampling framework for Campylobacter was different from that used for the other enteric bacterial pathogens. There have been consistent sampling and testing procedures used for Campylobacter surveillance through NARMS-EB, which allows for analysis and interpretation of data for trends.
Consistent sampling, testing and analytic methods allow for analysis and interpretation for trends, even in the presence of apparent seasonal variation in the proportion of strains that are resistant. Examination of the raw data without appropriate statistical analysis could lead to misinterpretation. The proportion of strains resistant to fluoroquinolones has increased since the inception of the surveillance in 1997. Using a logistic regression model, accounting for the increasing number of reporting sites, the increase from 1997 to 2001 is statistically significant.
That resistant infections are linked to poultry is supported by a case-control study conducted in several FoodNet sites, a study that was part of a larger FoodNet case-control study of Campylobacter infections. A preliminary analysis of this study was reported by Kassenborg, indicating that resistant infections were associated with eating chicken or turkey cooked at commercial establishment, and with foreign travel; fewer than half of the persons reported foreign travel. It is also supported by the frequent finding of resistant Campylobacter in poultry at retail, and by the observation that in studies done in many countries, consumption of undercooked poultry is the most often identified specific risk factor for Campylobacter infections.
Several of the examples of statements cited that were made by CDC staff refer to preliminary data, which may differ from the final set of data used for analysis and interpretation. Abstracts for scientific meetings are usually based on preliminary data. In these settings and in other reports and presentations, it is standard practice to identify data as partial or preliminary. CDC scientists are expected to be aware of the need for such identification of preliminary data.
CDC has, and continues to, release data and other information concerning its studies in accordance with the provisions of the Freedom of Information Act, and your FOIA request is being handled in this process.
Sentinel County Study, 1989-90: This survey of isolates of Campylobacter from a panel of sentinel counties was undertaken in 1989-1990. The results have not been published in full. As you correctly indicate, the cited abstract by Sobel et al. (Exhibit 14) does not describe the results of susceptibility testing. However, the results of that testing appear in a chapter written by Dr. Fred Tenover (Tenover FC, et al.Antimicrobial resistance in Campylobacter species. In: Nachamkin I, Blaser MJ, Tompkins LS, editors. Campylobacter jejuni: current status and future trends. Washington: American Society for Microbiology; 1992: 66-73) In summary, 1 of 332 isolates tested was resistant to ciprofloxacin. That report is also incomplete, as it was subsequently determined that the one quinolone resistant strain was a Campylobacter lari, a species that is inherently resistant to quinolones (Patton, CM, et al. Common somatic O and heat-labile serotypes amongCampylobacter strains from sporadic infections in the United States. J Clin Microbiol 31: 1525-1530.) C. lari strains are extremely rare.
With the exception of that C. lari strain, the isolates of Campylobacter collected through the survey in 1989-1990 were susceptible to fluoroquinolones while 13-19% of those collected from 1997-2001 through NARMS-EB were resistant. This demonstrates that resistance emerged during the years 1990-1997.
In 2002, CDC provided an electronic database and study methods for the 1989-1990 survey. More recently, copies of the protocol and study questionnaire and information on the chronology of laboratory testing and testing methods that would have been used in 1990 were also made available.
Smith et al., New England Journal of Medicine, 1999. As we previously stated, data collection and analytic methods as described in the article appear to be appropriate and support the conclusions reached by the authors.
Comment 3. Estimate of campylobacteriosis incidence in the U.S. is outdated.
This is the most recently published estimate for the annual incidence of Campylobacter infections. It was published in 1999, by Mead et al, as part of a series of estimates of foodborne illnesses, hospitalizations and deaths (Mead PS et al. Food-related illness and death in the United States. Emerging Infectious Diseases. 5(5):607-25, 1999 Sep-Oct). However, no revised estimate of national incidence has been published by CDC or by others; therefore, the 1999 estimate continues to be cited. Data are being collected and are being reviewed at CDC for revising the estimate. This is a labor intensive task, however, requiring not only data but considerable effort to analyze the data on Campylobacter and other agents causing gastroenteritis.
A declining trend in Campylobacter and several other foodborne bacterial infections was first detected by FoodNet in 2002 (noting a significant decline since 1996).
Comment 4: Places where statements are being disseminated
The Campylobacter information that CDC currently disseminates is consistent with the requirements of CDC's Information Guidelines. However, four documents that you cited were from the mass media, including articles that were published in the Atlanta-Journal Constitution, Food and Chemical News, and the Wall Street Journal. The Guidelines only cover material disseminated by CDC.
In summary, upon careful review of your complaint, we do not find that the statements made in CDC abstracts or presentations to which you refer misrepresent the available data or require any corrective action. However, we agree that the presentation should clearly state what data are preliminary and that final results may differ from preliminary reports. In addition, CDC plans to include a discussion of the limitations inherent to surveillance and to NARMS-EB in particular, in future reports from that system. We also plan to review the documents related to this issue on our website to ensure that they are accompanied by an appropriate discussion of data limitations.
We appreciate your comments and hope that the information we provided helps clarify the status of our work on Campylobacter and our efforts to communicate it to the public.
A written appeal or electronic request may be submitted for reconsideration within 30 days of receipt of the agency's decision. The appeal must state the reasons why the agency response is insufficient or inadequate. You must attach a copy of your original request and the agency's response to it. Also, clearly mark the appeal with the words, 'Information Quality Appeal' and send the appeal:
Centers for Disease Control and Prevention
Management Analysis and Services Office
1600 Clifton Road, N.E.
Atlanta, Georgia 30333
Rima Khabbaz, MD
Associate Director for Epidemiologic Science
National Center for Infectious Diseases
Centers for Disease Control and Prevention
Last Revised: August, 2004