Antimicrobial Resistance: Issues and Options (1998)

There is an urgent need for effective domestic and global surveillance of antibiotic resistance in animals and humans. There is also an urgent need for more prudent use of antibiotics in both human and veterinary medicine, particularly as it relates to food production. Of equal urgency is the need for better hospital infection control and implementation of guidelines to reduce the spread of infection and antibiotic-resistant pathogens in the hospital environment. There is a great need for strengthening the curriculum of human and veterinary health care professionals in the areas of sterilization and disinfection, mechanisms of antibiotic resistance, and factors contributing to its spread, including inappropriate antibiotic usage. There is also a need for patient education regarding appropriate uses of antibiotics. More basic research is needed to more clearly delineate mechanisms of antibiotic resistance and to identify new antimicrobial targets. Lastly, greater emphasis must be placed upon research related to rapid, reliable diagnostic tests and vaccines for prevention and control of infectious diseases.

The most compelling indication for a national surveillance system is that no such system currently exists. Recent advances now make it possible for clinical

and reference laboratories to accurately assess the susceptibility of a wide variety of bacterial and fungal pathogens via standardized methods. Medical informatics and computer technology are now available for accurate collection, efficient transmission, and timely analysis of surveillance data which will allow the information to be disseminated in a site-specific manner. Implementation of a surveillance system with necessary quality assurances and fiscal support will allow the generation of antimicrobial resistance data needed for decision-making regarding therapy and/or prophylaxis. Data will be forthcoming to predict emerging resistances among available therapeutic drugs, leading to effective interventions that could control the dissemination of resistance.

1. Focus on the most prevalent bacterial and fungal pathogens (not viral) that concern human health. This will assess isolates from clinical disease cases and routine isolates so that no bias from one center testing only the "problem" isolate or more resistant isolates compromises the results. Attention will be given to the trend in upward "creep" of MICs [minimum inhibitory concentrations]. There is also a need to monitor food sources such as animal products at the supermarket level as well as imported fruits, vegetables, and other products that may carry colonizing, drug-resistant bacteria and colonizing fecal flora in some patient populations. Salmonella and Shigella both should be monitored. Salmonella gives the best window into the impact of uses of antibiotics in the animal world, and the fraction of Shigella that is imported gives us an excellent view of the impact of antibiotic uses in the developing world.^* Monitoring of soil waste in farms should also be considered.

2. Establish a baseline of antimicrobial in vitro efficacy to which the following can be compared: earlier data from similar surveillance studies found in medical literature reviews, especially if these studies utilized comparable methodologies and surveillance techniques; subsequent surveillance data resulting from the establishment of a national surveillance system analyzed in a longitudinal manner; non-U.S. data to assess the international risks of resistance.

3. Accumulate concurrent demographic profile information to assess the relationships between organisms emerging in hospitals of various sizes or disease therapy focuses and pathogens prevalent among ambulatory patients in the community and animals housed in various environments. The role of drug use in these environments shall be addressed.

4. Establish a mechanism whereby organisms possessing certain phenotypic and genotypic resistance patterns will be referred to adequately funded laboratories for detailed study. Various molecular typing and investigative procedures can lead to earlier understanding of developing resistance mechanisms and spread of epidemic clones.

5. Allow for the future assessment of the encountered resistant pathogens' effect on patient outcomes, general community health, and the costs of health care delivery. Surveillance will target areas for specific intensive interventions for prevention (like vaccine campaigns and antimicrobial use reduction programs). Surveillance will also identify areas where epidemiologic investigations are needed to improve understanding of spread of drug-resistant strains and to identify ways to interfere with spread.

6. Maximize the possibility that data will lead to significant forms of professional health care intervention to reduce the probability that resistance to the drug will be spread widely and have an adverse impact on the national quality of health care outcomes. Interventions ideally should be focused at the local level, but regional and national interventions could also provide great benefits.

7. Provide expert federal agencies and societies the information to modify recommendations regarding therapy or prophylaxis of diseases or regarding testing procedures. These can be implemented at various levels related to patient or institutional demographics or by geography (local, regional, national).

8. Provide a compatible system in which subsets of participants could be grouped for common benefits. Examples include federal hospitals (Veterans Administration [VA], military, etc.), animal care facilities (university-based, U.S. Department of Agriculture [USDA], etc.), recognized HMO-like programs, and academic institutions such as university teaching hospitals.

9. Provide the accumulated data to pharmaceutical manufacturers, thus providing the validations of contemporary drug spectrums. This will be valuable in establishing meaningful organism coverage indications in antimicrobial agent package inserts.

10. Provide a system that can be modified to address any discovered area of concern related to the effective therapy of infectious organisms. This could allow expansion to cover fungi, viruses, cell-associated organisms, and some parasites.

1. Bacterial pathogens considered important in human and animal infections should be monitored. The choice of pathogens should be made and updated periodically based on the frequency with which these agents cause disease, the human or animal morbidity and economic impact of resistance in that species if it occurs or increases, the perceived threat of genomic mutation, and the need to confirm the continued efficacy of important therapeutic antimicrobial agents.

2. Some strains/species shall be tested on a regular schedule (quarterly) and others as dictated by the needs of the surveillance oversight panel or requirements for spectrum validations as part of the interaction with industry colleagues (spectrum data for drug package inserts).

3. The numbers of organisms tested should be significant (final counts depending on number/species/site and the numbers of participant sites) and should be finalized with input from medical statisticians.

4. Some isolates will require the choice of participants that routinely test unusual species (referral centers) or, in the case of some animal pathogens, those geographic specialty laboratories having expertise with a single pathogen or a group of pathogens specific for a single animal species, such as swine, cattle, or sheep.

5. Some species may be added that represent unusual organisms from which documented resistance genome have been transferred to prevalent human pathogens (examples: viridans group streptococci and oral Neisseria spp.).

1. A representative sample of organisms can be achieved only by a broad sampling of geographically dispersed laboratory isolates. Preliminary ideas for selection include the following.

1. Laboratories should be chosen according to U.S. population density by state. Local and state public health laboratories should be involved in data collection but should not be the sole participants.

2. States should be grouped into regions that should not be significantly different than those used in the CDC Morbidity and Mortality Weekly Report.

3. Participants should represent demographic populations of patients and organisms within their states or regions. Examples would include laboratories from large and small (<250 beds) hospitals plus samples from microbiology laboratories servicing outpatient clinic practices. Distribution of centers based on hospital bed size should be encouraged.

2. The above ideas will focus on human pathogen surveillance. Smaller numbers of participants will contribute the animal pathogens which will also be geographically diverse and hopefully representative for the animal species/pathogen population data.

3. In general there is a need for one monitoring site per 1,000,000 to 2,000,000 people, supplemented by animal pathogen participants and important demographic subsets (VA medical centers, etc.). Distribution of centers based upon hospital bed size should be encouraged, with regional centers being considered in areas where patients are more widely dispersed. Not all locations will have to be recruited in the first year, but representative participants from each region should be sought as early as possible.

1. Demographic profiles of the hospital/laboratory will be required for participation (updated each year). This would allow initial classification of participants by various parameters such as hospital size, services offered, formulary practices, antimicrobial use patterns, infection control practices, teaching status, etc.

2. Particularly important surveillance groups will be selected by:

1. Hospital size.

2. Services or teaching offered.

3. Health care delivery classification (HMO, federal, VA, private, etc.).

4. Clinical practice.

5. Hospital-acquired pathogens versus pathogens from other sources.

6. Prior antimicrobial therapy and type.

7. Others.

3. Given the limited testing done by commercial laboratories, a systematic approach for measuring community-acquired antibiotic-resistant bacteria must be established. This task might best be performed by the local and state health departments.

1. All methods used should be of reference quality and should closely follow the documents published by the NCCLS [National Committee for Clinical Laboratory Standards].

2. Where possible, initial screening should use the disk diffusion method (NCCLS M2-A5) for those pathogen—antimicrobial combinations that can be accurately tested.

3. A subset of organism/antimicrobial combinations will require special tests that will need referral to a reference laboratory or laboratories. Alternatively, these generally fastidious species can be tested at each location by some newer technologies, if costs and reagent availability dictate.

4. Method choices should be focused through an advisory panel of experts in the field of antimicrobial susceptibility testing.

5. Rigid quality controls (QC) will be necessary for each participant (local) laboratory, each monitoring reference laboratory, and any

laboratory used for special contracted tests or molecular studies. Only data obtained by accepted methods and accompanied by validating QC should be entered into the antimicrobial surveillance database. Participants regularly not complying with QC guidelines should be replaced.

1. The dominant, quarterly testing should be performed at the laboratory participant locations.

2. Specific pathogens (identified by species, resistance phenotype, isolate origin, etc.) should be routinely forwarded to secondary, reference (monitor) laboratories. These centers should perform contracted studies by specified methods (protocol) and report results to the U.S. antimicrobial surveillance data processing location (i.e., CDC [see below]).

3. Molecular typing, resistance mechanism studies, and other molecular-level techniques will be necessary on an annual basis. These studies should be assigned to appropriate reference laboratories as required by the surveillance administrators.

1. All data should be expressed as quantitative endpoints regardless of method. This dictates measurement of disk diffusion tests by calipers to the nearest whole millimeter and the use of MIC endpoints in micrograms per milliliter for dilution methods. Such measurements will facilitate the recognition of susceptibility changes within categories (qualitative) established by the NCCLS.

2. Qualitative interpretations shall be applied objectively by computer programs based on current NCCLS tables. Similarly, QC guidelines found in the NCCLS tables should also establish the validity of each participant's/referee's data.

1. Surveillance studies of all types can greatly benefit from well-structured computer systems. Prior studies performed by the CDC (NNIS [National Nosocomial Infection Surveillance System] and Public Health Laboratory Information System), WHO [World Health Organization] (WHONET), and private parties have effec

tively utilized relatively simple data input programs. Such software is available and/or programmable without the significant risk of problems. Timeliness, flexibility, and the ease of expanding the pyramidal reporting structure are important considerations.

2. All input should be simplified to include minimal transcriptions via digital reading, bar-coding, disk transfers, or modem networking.

3. A single data analysis location would be preferred. Programming would be under the direction of the surveillance oversight panel, which should also periodically review the analysis results.

4. Simple in-laboratory work forms should also be standardized for all locations as a hard-copy backup. These forms should not greatly differ from the clinically used forms at each location.

5. Previously organized, computerized networks can be used as models. Possible collaborations might include CDC—NNIS. CDCSTD regional surveillance in Neisseria gonorrhoeae, VA networks, SCOPE (University of Iowa program), and international programs through various medical specialty societies.

1. Preliminary guidelines for the surveillance are as follows.

1. Surveillance of basic pathogen group by each participant at the local site, QUARTERLY.

2. Analysis of special (usually fastidious) pathogens by a selected number of reference or special skill laboratories, QUARTERLY.

3. Expanded list of antimicrobial agents and pathogens that will satisfy annual validation of drug package insert spectrum, YEARLY (winter quarter).

4. Special studies as directed by the oversight panel, VARIABLE.

2. The program must establish the perception that participation is a benefit. To accomplish this goal the following items might be considered.

1. Quarterly reports to all participants.

2. Annual newsletter summary of results.

3. Publications generated from surveillance data acknowledging the participation of all sites.

4. Participant interaction with the database. Confidentiality should be maintained by coding locations and limiting laboratory-defined access to the participant center only. However, data would otherwise be available but audited by the oversight panel or study administrators.

5. Periodic meeting of the surveillance participants, at a national meeting or as a free-standing annual symposium.

1. Funding should be sought from all parties that would derive a significant benefit from the system. These include, but would not be limited to (a consortium approach), the following.

1. Federal and state agencies: CDC, Food and Drug Administration (FDA) (several components), National Institutes of Health (NIH), USDA, state health departments, VA, and Department of Defense.

2. Industry: Pharmaceutical manufacturers, reagent manufacturers (antimicrobial tests and microbiology media), drug delivery systems, health care delivery and hospital corporations, and contract research organizations.

3. Academic institutions, professional societies, and university medical centers.

2. An oversight panel should be established for the surveillance study network. This panel should include representatives from the principal federal agencies, members of the scientific community at large who are experts in antimicrobial resistance, specialists in infectious disease and in vitro antimicrobial testing in animals and humans, experienced personnel in multilaboratory surveillance, hospital/health care epidemiologists, computer and statistical analysts conversant with antimicrobial issues, and representatives of industry as dictated by consortium participation. The Antimicro

bial Resistance Surveillance Program should be located within the National Center for Infectious Diseases at the CDC. This would allow integration with other surveillance activities for infectious diseases by experts in microbiology and epidemiology. Because of the CDC's track record and proven expertise, this is the logical location for a national surveillance system.

3. If this program is successful in meeting the cited goals and objectives, cost savings to the national health care system could easily be in the hundreds of millions of dollars. A few areas of greatest impact are (i) the focusing of pharmaceutical research as early as possible on emerging drug resistance problems; (ii) the reduction of drug development costs by placing new compounds into the system to establish in vitro spectrums (cost rates would be less than the multiple individual grants to independent investigators); (iii) validation of current compound spectrums annually, reducing manufacturers' costs for independent vendor contracts (like item ii); (iv) availability of contemporary, localized data to federal and state agencies, enabling them to identify emerging resistances or pathogen frequencies and thus develop early interventions and selections of effective therapeutic regimens or prophylactic modalities; and (v) other interventions stimulated by the data derived from the surveillance, leading to greater cooperation among government, industry, and professional components of the health care system at large.

1. Convene an expert panel to develop surveillance protocol and establish an annual budget. The above outline could be used as a preliminary or tentative plan. Federal funding should be immediately identified for the national surveillance system, and several agencies (CDC, FDA, NIH, VA, USDA, etc.) as well as other sources should be involved in funding decisions.

2. Seek appropriate expertise in statistical analysis for determination of organism sample sizes, population-based participant selection, and computer support for analyses.

3. Initiate an extensive search for earlier surveillance databases on resistance generated from U.S. multicenter investigations performed with NCCLS or compatible methods. Also expand that literature search to worldwide surveillance data, if available.