- Email: [email protected]
National hospital survey of anaerobic culture and susceptibility methods: III
ARTICLE IN PRESS
Anaerobe 14 (2008) 68–72 www.elsevier.com/locate/anaerobe
Clinical medicine
National hospital survey of anaerobic culture and susceptibility methods: III Ellie J.C. Goldsteina,b,, Diane M. Citrona, Pamela J. Goldmana, Ronald J. Goldmana a
b
R M Alden Research Laboratory, Santa Monica, CA 90404, USA The David Geffen School of Medicine at UCLA, Los Angeles, CA 90093, USA Received 9 December 2007; accepted 6 January 2008 Available online 11 January 2008
Abstract To assess the current status of anaerobic bacteriology in the United States, we surveyed, by means of a questionnaire, 150 hospitals selected at random with bed capacities of 200–1000 and we received responses from 98 (65%). Ninety-eight percent processed anaerobic culture specimens with 21% sending them to reference laboratories. Almost all these hospitals processed blood and wound cultures for anaerobes and all used selective media for identification, including BBE (52%), LKV (77%), and PEA (53%) agars. All hospital laboratories attempted identification of blood culture isolates including 80% that attempted speciation. Wound cultures for anaerobic bacteria and sterile site cultures were also processed for anaerobes by almost all labs. Identification of B. fragilis group species to species level was performed only in 56% of labs always and 37% sometimes. Preformed enzyme kits were used by 66% of labs and 30% used special potency disks for identification. Susceptibility testing was performed in-house by 21% of hospital labs and sent out to reference labs an additional 20%. Susceptibility testing was attempted for all blood culture isolates by both hospital (21% of total labs) and reference laboratories, but only performed by 17% for sterile body site and 14% of the time for wound isolates. Etest was used most often followed by broth microdilution. No labs used the agar dilution or disk elution methods. The antimicrobials most often tested in hospital labs, predicated on the commercial panel used, were penicillin/ampicillin and clindamycin (15/18; 83%; 15% of total labs), metronidazole (16/18; 89%; 16% of total labs) and cefotetan and ampicillin/sulbactam (12/18; 67%; 12% of total labs), piperacillin/ tazobactam (7/18; 39%; 7% of total labs), cefoxitin (9/18; 50%), imipenem (8/18; 44%), and chloramphenicol (6/18; 33%). Our current survey suggests that while many labs are processing anaerobic cultures, especially blood cultures, the identification of isolates and the performance of antimicrobial susceptibility testing of isolates are in disarray and in dire need of improvement. r 2008 Elsevier Ltd. All rights reserved. Keywords: Anaerobic susceptibility; B. fragilis; Etest; Methods
1. Introduction While anaerobic bacteria are important clinical pathogens, clinical laboratories vary in their capabilities and interest in the isolation and identification of anaerobes and in their performance of anaerobic susceptibility testing [1,2]. Recently, the controversy about the importance of obtaining anaerobic cultures has reemerged when Lassmann et al. [3] noted a 74% rise in anaerobic bacteremias at the Mayo Clinic and that in today’s complex medical Corresponding author at: 2021 Santa Monica Blvd, Suite 740 East, Santa Monica, CA 90404, USA. Tel.: +1 310 315 1511; fax: +1 310 315 3662. E-mail address: [email protected] (E.J.C. Goldstein).
1075-9964/$ - see front matter r 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.anaerobe.2008.01.001
patients ‘‘the clinical context for anaerobic infections [is] less predictable than it once was.’’ This highlights the need for practice improvement [4]. Numerous studies have demonstrated that appropriate early therapy results in a better clinical outcome with lower mortality and morbidity and shortened length of stay for diverse conditions including bacteremia [5,6]. With the rising resistance to a variety of commonly employed anti-anaerobe agents, including several case reports of clinical metronidazole resistance, also suggests that the availability of susceptibility testing is clinically relevant [7]. However, laboratorians that are often concerned about the cost of testing [2] often treat anaerobic bacteriology as the poor stepsister and as a consequence even guideline committees have
ARTICLE IN PRESS E.J.C. Goldstein et al. / Anaerobe 14 (2008) 68–72
suggested that anaerobic culture of community-acquired intraabdominal infections need not be performed [8]. In order to assess the current status of anaerobic culture and susceptibility methods, we performed a new national survey in randomly selected, large nonteaching hospitals across the United States. 2. Methods In the fall of 2006 a sample of 150 hospitals was randomly selected from the American Hospital Association’s Guide to the Health Care Field, 2006 edition [9]. The sample was limited to general medicine and surgical hospitals with bed capacities of 200–1000. For each selected hospital, the microbiology laboratory was contacted by telephone to identify the appropriate person to discuss anaerobic bacteriology and whether anaerobic cultures were processed in-house or sent to reference laboratories. If processed in-house, the respondent was given the choice to receive the survey questionnaire by e-mail, to complete the survey on-line, or receive it by mail, with a self-addressed, stamped envelop. If the identified lab personnel could not be contacted or declined to participate in the original phone call, a random selection of another hospital in that geographic area was made. Twenty-seven replacement respondents were selected. If cultures were sent to a reference laboratory, the reference lab was called, and the appropriate person was identified to discuss anaerobic bacteriology and offered the same survey completion options. For questionnaires not returned within 2 months, follow-up phone calls were made. The four-page questionnaire, written by the authors, consisted of 16 questions regarding the various elements of clinical anaerobic bacteriology (see Results section for specific questions). Current CLSI guidelines for susceptibility testing [10], Wadsworth-KTL Anaerobe Bacteriology Manual [11] and the Manual of Clinical Microbiology [12] were consulted in composing relevant questions.
2. What is the background of your laboratory director? Pathologists directed the laboratories in 87% of hospitals and 80% of reference laboratories. PhDs directed 6% of the hospital laboratories and 13% of the reference laboratories. The remainder were directed by ‘‘other MDs or not specified’’. 3. From which sites do you process anaerobes? One of 76 hospital labs did not answer this question.
Blood Sterile body sites Wounds–Aspirates Wounds–Swabs Other Specimens
4.
5.
6.
3. Results Of the 150 hospital laboratories contacted, 128 (85%) responded to the screening question by saying that they processed anaerobic cultures in-house, and 22 (15%) said that they send selected specimens to reference laboratories, of which 8 were within their hospital system and 12 were to independent laboratories and two did not answer. Of the 128 hospital labs sent the survey instrument, 76 (59%) were returned, and 15 of the 22 reference labs (68%) returned their surveys. Questions and answers: 1. Do you have an Infectious Diseases physician who practices on site? 98% of hospitals had ID physicians on site. Hospitals that used reference labs were not asked this question.
69
7.
Hospital lab (%)
Reference labs (%)
99 99 100 92 31
93 93 93 93 57
3b. Do you reject poor quality specimens? Seventy-eight percent of hospital labs and 86% of reference labs rejected poor quality specimens. Approximately, how many specimens do you process for anaerobes (or send out) each month in your laboratory? Blood cultures: 1–7000 range; 600, median; 1192, average Other specimens: 15–1700, range; 125, median; average 97 Are anaerobes done in-house? This question simply confirms the screening question by ensuring all the hospital questionnaires are from hospitals processing anaerobes in-house. 5b. If done in-house, do you process specimens on all shifts? Eight hospital labs and two reference labs did not answer this question. Seventy-two percent of hospital labs and 62% of reference labs process anaerobes on all shifts. Do you use transport media and batch specimens for processing? Sixty-nine percent of hospital labs and 80% of reference labs used transport media and batched specimens for processing. Do you use selective and differential media? One hundred percent of hospital labs and 93% of reference labs used selective media. 7b. If so, which media? Hospital labs
Reference labs
39 58 40 31
5 13 7 6
(52%) (77%) (53%) (41%)
BBE LKV PEA Other
(36%) (93%) (50%) (43%)
8. Do you identify anaerobes from blood cultures? All (100%) of hospital labs and 93% of reference labs identified anaerobes from blood cultures. All hospital and reference labs reported gram-stain
ARTICLE IN PRESS E.J.C. Goldstein et al. / Anaerobe 14 (2008) 68–72
70
results; 80% of hospital labs and 64% of reference labs reported all isolates to species level, while 17% of hospital labs and 18% of reference labs only reported to species levels ‘‘sometimes’’. Nineteen percent of hospital labs and 18% of reference labs reported isolates to genus level on a regular basis, while 1% of hospital labs and 18% of reference labs reported blood culture isolates to group level. 9. Do you identify anaerobes from wound cultures? Ninty-nine percent of hospital labs and 93% of reference labs identified anaerobes from wound cultures. All these hospital and reference labs report Gram-stain results; 69% of hospital labs and 64% of reference labs reported to species level and 24% and 18% reported to genus level, respectively, with the remainder reporting to group level only. 10. Do you identify B. fragilis group to species level? For hospital labs and reference labs 56% and 53% always reported to species, 37% and 13%, sometimes and 7%and 33% did not attempt this, respectively. 11. What methods do you use for anaerobe identification?
Special potency disks Preformed enzyme kit Fermentation reactions Individual enzyme tests Other biochemical tests Gas liquid chromatography Molecular Other methods
Hospital labs (%)
Reference labs (%)
30 66 21 7 30 4 0 4
47 60 13 0 60 7 0 0
12. Do you do anaerobic susceptibility testing? Overall, 56% of hospitals had access to susceptibility testing, with 41/76 (54%) of hospital labs that processed anaerobes had access to susceptibility testing but only 21/41 (51%) performed them in-house and 20/41 (49%) sent them to reference labs for performance. Of the 15 reference labs that returned the survey, 6/15 (40%) performed anaerobic susceptibility testing in-house, and 4 of these 15 (27%) sent the isolates out to yet another reference lab (5 said that they do not do anaerobic susceptibility testing). Of the 98 hospital labs (76+22 reference labs) that are accounted for in this study, only 21% of hospital labs perform anaerobic susceptibility testing, either inhouse or outsourced. 13. Which organisms would you test for susceptibility? Of those that performed susceptibility studies
Blood isolates: Other isolates all Selected
In-house (N ¼ 20) (%)
Sent out (N ¼ 6) (%)
100 56 44
100 50 50
Sterile body sites All isolates Selected isolates Surgical wounds: All isolates Selected isolates Other wounds: All isolates Selected isolates Other: upon request
85 47 53 70 15 85 50 10 90 40
100 33 67 67 50 50 67 25 75
14. What susceptibility method is used? Sixty-five percent of hospital labs used the Etest and 40% used microbroth dilution methods (all commercially prepared) compared to 17% and 100% for reference labs, respectively. No lab used the agar dilution method, and some used more than one method. 15. Which drugs do you test? Of the 98 hospital labs (76+22 reference labs) that are accounted for in this study, only 41 hospital labs and 10 reference labs report susceptibility testing results and only 21 hospital and 6 reference labs perform them in-house. Three hospital labs did not report on the drugs tested. The percentages are based on those denominators.
Beta-lactamase test only Penicillin/ampicillin Amoxicillin Ampicillin/sulbactam Piperacillin/tazobactam Clindamycin Metronidazole Moxifloxacin Ertapenem Imipenem Meropenem Doxycycline Tigecycline Cefoxitin Cefotetan Chloramphenicol Other drugs
Hospital labs (N ¼ 18) (%)
Reference labs (N ¼ 6) (%)
28 83 6 67 39 83 89 0 0 44 17 0 0 50 67 33 39
17 83 0 83 50 100 100 0 0 33 17 0 17 100 67 67 100
16. How do you test for C. difficile? Eighty-four percent of hospital labs and 92% of reference labs use an Elisa test for Toxin A+B; 11% of hospital labs but no reference labs use Toxin A testing only. Other tests used by hospital and reference labs were: tissue culture assay, 5% and 23%, respectively; stool culture, 1% and 15%, respectively; universal antigen-glutamate dehydrogenase (GDH) ELISA screen test, 3% but by no reference labs.
ARTICLE IN PRESS E.J.C. Goldstein et al. / Anaerobe 14 (2008) 68–72
4. Discussion The typical clinical presentation of patients with anaerobic bacteremia is less predictable then it was a decade ago [3]. In addition, anaerobes have exhibited new antimicrobial resistance patterns [13–15]. Despite technical improvements and increased standardization, our prior two surveys had noted a dramatic decline in the availability of anaerobic bacteriology and susceptibility data available to clinicians [1,2]. Consequently, they often must use surveys and small studies published from a handful of expert and specialized reference laboratories as a basis for empirical antibiotic selection. As a consequence of limited laboratory input, clinicians employ ‘‘top-of-the-line’’ broad-spectrum, often expensive antimicrobials for therapy of all anaerobic infections adding increased selective pressure for the development of new resistance and added cost to medical care [16]. In our last survey [2] performed more than a decade ago, only 65% of hospital (200–1000 beds) laboratories performed some degree of anaerobic bacteriology. Our current survey showed improvement with 76% now performing some degree of anaerobic bacteriology inhouse and an additional 23% send anaerobic specimens to a reference laboratory. Cultures from most clinically relevant sources such as blood, sterile sites and wounds are processed. Some amount of quality control is exerted with 78–86% rejecting poor quality specimens. Of those hospital laboratories that do process anaerobic cultures inhouse 100% use selective media including 52% that use BBE for rapid identification of the Bacteroides fragilis group compared to 31% in our prior 1993 study [2] and 20% in our 1990 survey [1]. The patterns of use of LKV and PEA media did not change over the past 16 years. Procedures for optimizing anaerobic culture methods are available from a variety of sources [11,12,17,18]. Sixty-two percent of laboratories also indicated interest in attending Anaerobic Bacteriology workshops that are offered at the Annual Meeting of the American Society for Microbiology, Anaerobe Society for the Americas meetings, by Anaerobe Systems, and at other venues. The value of performing anaerobic blood cultures has been questioned [19]. The importance of obtaining anaerobic blood cultures was recently highlighted by Lassmann et al. [3] who noted a 74% rise per 100,000 patient days in anaerobic bacteremias at the Mayo Clinic, with 35% not having an obvious primary anaerobic infection. In this survey, as in the prior surveys, all labs that processed anaerobes also processed anaerobic blood culture isolates; 67% of them identified the isolates to species level all the time and an additional 15% some of the time, while 16% tended to identify isolates only to genus level. This result is similar to those of 1990 when 66% of labs identified blood culture isolates to species level and 32% identified to genus level [1]. In addition, 99% of the labs currently also identify anaerobes from wound cultures with a similar speciation pattern compared to 80% in our
71
1990 survey [1]. B. fragilis was identified to species level by 56% and 53% of hospital and reference labs always, and an additional 37% and 13% sometimes. Surprisingly, 33% of reference labs did not speciate B. fragilis group isolates in spite of their known differences in resistance to commonly used antimicrobial agents. Evaluating the response to the performance of anaerobic susceptibility testing was somewhat complex and percentages dramatically change when extrapolated to the total sample. Overall, only 21% (21/98) of hospital labs perform anaerobic susceptibility testing in-house. Currently, 53% (41/75) of hospital labs and 67% (10/15) of reference labs have access to anaerobic susceptibility testing data. These rates continue to decline as in 1990 [1] 70% performed susceptibility testing that declined to 33% in 1993 [2]. Testing, when performed was universally done on blood isolates. It was performed by 85% (17/20) of labs for sterile body site isolates, and by 14/20 (70%) on mostly selected surgical wound isolates. An additional 40% (8/20), would also perform susceptibility testing by special request. Most hospital labs used the Etest (62%; 13/21) for susceptibility testing, while only 17% of reference labs used it. All reference labs used microbroth dilution for susceptibility testing as did 40% (8/20) of hospital labs [all commercially prepared]. No laboratory, hospital or reference, used the agar dilution method, which seems to be reserved to ‘‘centers-of-excellence’’ or research centers with an interest in anaerobes. Previously, in 1993, the microbroth dilution method was predominant in hospital labs (47%) [2] and the use of disk diffusion (33%) and broth disk elution method, which were predominant (56%)in 1990 [1] have disappeared from use. CLSI [10] does not describe ‘‘user friendly’’ methods that laboratories could easily adopt as standard procedures. The broth micro dilution technique set up using bench techniques is limited to the relatively oxygen-tolerant members of the B. fragilis group, although laboratories that have anaerobic chambers are able to use this method for testing more fastidious strains. The Etest is referred to as the ‘‘gradient method’’ in CLSI documents, but, as a proprietary product, is not mentioned by name. Its usefulness and reliability for testing anaerobes has been well documented in many publications [20–23]. Susceptibility data is now available to a very limited number of hospitals for a very few antimicrobial agents. Five hospital labs (28%) have only beta-lactamase testing results available. Since other mechanisms of beta-lactam resistance may be present, a negative test is not a guarantee of an organism’s susceptibility. Selection of the antimicrobials tested appears to be dependent upon the antimicrobials present on the commercial panels available to laboratories. No hospital lab tested for tigecycline, doxycycline, moxifloxacin or ertapenem susceptibility. The most frequently tested antimicrobials were penicillin/ ampicillin and clindamycin (15/18; 83%), metronidazole (16/18; 89%) and cefotetan and ampicillin/sulbactam (12/18; 67%). Other, more limited testing data was available for piperacillin/tazobactam (39%), cefoxitin
ARTICLE IN PRESS 72
E.J.C. Goldstein et al. / Anaerobe 14 (2008) 68–72
(50%), imipenem (44%), and chloramphenicol (33%). If one uses the base of 100 respondents then the most commonly reported results (to metronidazole) shrinks dramatically to 16% overall. Consequently, clinicians must be using information from the manufacturer, the FDA indications, published study/survey data or just guessing at appropriate empirical as well as directed therapy. Our current survey suggests that while many labs are processing anaerobic cultures, especially blood cultures, the identification of isolates and the performance of antimicrobial susceptibility testing of isolates lack a standard approach and need improvement. Acknowledgments This survey was supported in part by a grant form Merck & Co, Inc. We thank C. Vreni Merriam, Judee Knight and Alice E. Goldstein for various forms of support. References [1] Goldstein EJC, Citron DM, Goldman RJ. National hospital survey of anaerobic culture and susceptibility testing methods: results and recommendations for improvement. J Clin Microbiol 1992;30: 1529–34. [2] Goldstein EJC, Citron DM, Goldman RJ, Claros MC, HuntGerrado S. United States national hospital survey of anaerobic culture and susceptibility methods II. Anaerobe 1995;1:309–14. [3] Lassmann B, Gustafson DR, Wood CM, Rosenblatt JE. Reemergence of anaerobic bacteremia. Clin Infect Dis 2007;44:895–900. [4] Hecht DW. Routine anaerobic blood cultures: back where we started? Clin Infect Dis 2007;45:901–3. [5] Nguyen MH, Yu VL, Morris AJ, McDermott L, Wagener MW, et al. Antimicrobial resistance and clinical outcome of Bacteroides bacteremia: findings of a multicenter prospective observational trial. Clin Infect Dis 2000;30:870–6. [6] McGregor JC, Rich SE, Harris AD, Perenchevich EN, Osih R, et al. A systematic review of the methods used to assess the association between appropriate antibiotic therapy and mortality in bacteremic patients. Clin Inf Dis 2007;45:329–37. [7] Goldstein EJC, Citron DM, Hecht DW. Chapter XXX resistance in anaerobic bacteria. In: Fong IW, Drlica K, editors. Antimicrobial resistance and implications for the 21st century (Chapter 6, pp. 207–29) Berlin: Springer; 2007.
[8] Solomkin JS, Mazuski JE, Baron EJ, Bradley J, Chow A, Dellinger EP, et al., Guidelines for the selection of anti-infective agents for complicated intraabdominal infections. Clin Infect Dis, 2008, in press. [9] American Hospital Association. American Hospital Association guide to the healthcare field. Chicago: American Hospital Association; 2006. [10] Clinical and Laboratory Standards Institute. Methods for antimicrobial susceptibility testing of anaerobic bacteria. Approved standard, 7th ed., CLSI document M11-A7. Wayne, PA: CLSI; 2007. [11] Citron DM. rapid identification of anaerobes in the clinical laboratory. Anaerobe 1999;5:109–13. [12] Murray PR, Baron EJ, Jorgensen JH, Landry ML, Pfaller MA, editors. Manual of clinical microbiology. Washington, DC: American Society for Microbiology Press; 2007. [13] Snydman DR, Jacobus NV, McDermott LA, Ruthazer R, Goldstein EJC, Finegold SM, et al. National survey on the susceptibility of Bacteroides fragilis group: report and analysis of trends for 1997–2004: a US study. Antimicrob Agents Chemother 2007;51: 1649–55. [14] Koeth LM, Good CE, Appelbaum PC, Goldstein EJC, Rodloff AC, Claros M, et al. Surveillance of susceptibility patterns in 1297 European and US anaerobic and capnophilic isolates to co-amoxiclav acid and five other antimicrobial agents. J Antimicrob Chemother 2004;53:1039–44. [15] Golan Y, McDermott LA, Jacobus NV, Goldstein EJC, Finegold S, Harrell LJ, et al. Emergence of high-level fluoroquinolone resistance among Bacteroides species. J Antimicrob Chemother 2003;52:208–13. [16] Finegold SM. A century of anaerobes: a look backward and a call to arms. Clin Infect Dis 1993;16(Suppl. 4):S453–7. [17] Isenberg HD, editor. Clinical microbiology procedure manual. 2nd ed. Washington, DC: ASM Press; 2004. [18] Jousimies HR, Summanen P, Citron DM, Baron EJ, Wexler HM, Finegold SM. Wadsworth-KTL anaerobic bacteriology manual. Belmont, CA: Star Publishing Co.; 2002. [19] Bourgault AM, Harkness JL, Rosenblatt JE. Clinical usefulness of susceptibility testing of anaerobes. Arch Int Med 1978;138:1824–7. [20] Citron DM., Hecht DW. Susceptibility test methods: anaerobic bacteria. In: Manual clinical microbiology (Chapter 76, pp. 1214–22). Washington, DC: American Society Microbiology Press; 2007. [21] Citron DM, Ostovari MI, Karlsson A, Goldstein EJC. Evaluation of the E-test for susceptibility testing of anaerobic bacteria. J Clin Microbiol 1991;29:2197–203. [22] Rosenblatt JE, Gustafson DR. Evaluation of the Etest for the susceptibility testing of anaerobic bacteria. Diagn Microbiol Infect Dis 1995;22:279–84. [23] Rhomberg PR, Biedenbach DJ, Jones RN. Activity of BMS284756 (T-3811) tested against anaerobic bacteria Campylobacter jejuni, Helicobacter pylori, and Legionella spp. Diag Microbiol Infect Dis 2001;40:45–9.
Anaerobe 14 (2008) 68–72 www.elsevier.com/locate/anaerobe
Clinical medicine
National hospital survey of anaerobic culture and susceptibility methods: III Ellie J.C. Goldsteina,b,, Diane M. Citrona, Pamela J. Goldmana, Ronald J. Goldmana a
b
R M Alden Research Laboratory, Santa Monica, CA 90404, USA The David Geffen School of Medicine at UCLA, Los Angeles, CA 90093, USA Received 9 December 2007; accepted 6 January 2008 Available online 11 January 2008
Abstract To assess the current status of anaerobic bacteriology in the United States, we surveyed, by means of a questionnaire, 150 hospitals selected at random with bed capacities of 200–1000 and we received responses from 98 (65%). Ninety-eight percent processed anaerobic culture specimens with 21% sending them to reference laboratories. Almost all these hospitals processed blood and wound cultures for anaerobes and all used selective media for identification, including BBE (52%), LKV (77%), and PEA (53%) agars. All hospital laboratories attempted identification of blood culture isolates including 80% that attempted speciation. Wound cultures for anaerobic bacteria and sterile site cultures were also processed for anaerobes by almost all labs. Identification of B. fragilis group species to species level was performed only in 56% of labs always and 37% sometimes. Preformed enzyme kits were used by 66% of labs and 30% used special potency disks for identification. Susceptibility testing was performed in-house by 21% of hospital labs and sent out to reference labs an additional 20%. Susceptibility testing was attempted for all blood culture isolates by both hospital (21% of total labs) and reference laboratories, but only performed by 17% for sterile body site and 14% of the time for wound isolates. Etest was used most often followed by broth microdilution. No labs used the agar dilution or disk elution methods. The antimicrobials most often tested in hospital labs, predicated on the commercial panel used, were penicillin/ampicillin and clindamycin (15/18; 83%; 15% of total labs), metronidazole (16/18; 89%; 16% of total labs) and cefotetan and ampicillin/sulbactam (12/18; 67%; 12% of total labs), piperacillin/ tazobactam (7/18; 39%; 7% of total labs), cefoxitin (9/18; 50%), imipenem (8/18; 44%), and chloramphenicol (6/18; 33%). Our current survey suggests that while many labs are processing anaerobic cultures, especially blood cultures, the identification of isolates and the performance of antimicrobial susceptibility testing of isolates are in disarray and in dire need of improvement. r 2008 Elsevier Ltd. All rights reserved. Keywords: Anaerobic susceptibility; B. fragilis; Etest; Methods
1. Introduction While anaerobic bacteria are important clinical pathogens, clinical laboratories vary in their capabilities and interest in the isolation and identification of anaerobes and in their performance of anaerobic susceptibility testing [1,2]. Recently, the controversy about the importance of obtaining anaerobic cultures has reemerged when Lassmann et al. [3] noted a 74% rise in anaerobic bacteremias at the Mayo Clinic and that in today’s complex medical Corresponding author at: 2021 Santa Monica Blvd, Suite 740 East, Santa Monica, CA 90404, USA. Tel.: +1 310 315 1511; fax: +1 310 315 3662. E-mail address: [email protected] (E.J.C. Goldstein).
1075-9964/$ - see front matter r 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.anaerobe.2008.01.001
patients ‘‘the clinical context for anaerobic infections [is] less predictable than it once was.’’ This highlights the need for practice improvement [4]. Numerous studies have demonstrated that appropriate early therapy results in a better clinical outcome with lower mortality and morbidity and shortened length of stay for diverse conditions including bacteremia [5,6]. With the rising resistance to a variety of commonly employed anti-anaerobe agents, including several case reports of clinical metronidazole resistance, also suggests that the availability of susceptibility testing is clinically relevant [7]. However, laboratorians that are often concerned about the cost of testing [2] often treat anaerobic bacteriology as the poor stepsister and as a consequence even guideline committees have
ARTICLE IN PRESS E.J.C. Goldstein et al. / Anaerobe 14 (2008) 68–72
suggested that anaerobic culture of community-acquired intraabdominal infections need not be performed [8]. In order to assess the current status of anaerobic culture and susceptibility methods, we performed a new national survey in randomly selected, large nonteaching hospitals across the United States. 2. Methods In the fall of 2006 a sample of 150 hospitals was randomly selected from the American Hospital Association’s Guide to the Health Care Field, 2006 edition [9]. The sample was limited to general medicine and surgical hospitals with bed capacities of 200–1000. For each selected hospital, the microbiology laboratory was contacted by telephone to identify the appropriate person to discuss anaerobic bacteriology and whether anaerobic cultures were processed in-house or sent to reference laboratories. If processed in-house, the respondent was given the choice to receive the survey questionnaire by e-mail, to complete the survey on-line, or receive it by mail, with a self-addressed, stamped envelop. If the identified lab personnel could not be contacted or declined to participate in the original phone call, a random selection of another hospital in that geographic area was made. Twenty-seven replacement respondents were selected. If cultures were sent to a reference laboratory, the reference lab was called, and the appropriate person was identified to discuss anaerobic bacteriology and offered the same survey completion options. For questionnaires not returned within 2 months, follow-up phone calls were made. The four-page questionnaire, written by the authors, consisted of 16 questions regarding the various elements of clinical anaerobic bacteriology (see Results section for specific questions). Current CLSI guidelines for susceptibility testing [10], Wadsworth-KTL Anaerobe Bacteriology Manual [11] and the Manual of Clinical Microbiology [12] were consulted in composing relevant questions.
2. What is the background of your laboratory director? Pathologists directed the laboratories in 87% of hospitals and 80% of reference laboratories. PhDs directed 6% of the hospital laboratories and 13% of the reference laboratories. The remainder were directed by ‘‘other MDs or not specified’’. 3. From which sites do you process anaerobes? One of 76 hospital labs did not answer this question.
Blood Sterile body sites Wounds–Aspirates Wounds–Swabs Other Specimens
4.
5.
6.
3. Results Of the 150 hospital laboratories contacted, 128 (85%) responded to the screening question by saying that they processed anaerobic cultures in-house, and 22 (15%) said that they send selected specimens to reference laboratories, of which 8 were within their hospital system and 12 were to independent laboratories and two did not answer. Of the 128 hospital labs sent the survey instrument, 76 (59%) were returned, and 15 of the 22 reference labs (68%) returned their surveys. Questions and answers: 1. Do you have an Infectious Diseases physician who practices on site? 98% of hospitals had ID physicians on site. Hospitals that used reference labs were not asked this question.
69
7.
Hospital lab (%)
Reference labs (%)
99 99 100 92 31
93 93 93 93 57
3b. Do you reject poor quality specimens? Seventy-eight percent of hospital labs and 86% of reference labs rejected poor quality specimens. Approximately, how many specimens do you process for anaerobes (or send out) each month in your laboratory? Blood cultures: 1–7000 range; 600, median; 1192, average Other specimens: 15–1700, range; 125, median; average 97 Are anaerobes done in-house? This question simply confirms the screening question by ensuring all the hospital questionnaires are from hospitals processing anaerobes in-house. 5b. If done in-house, do you process specimens on all shifts? Eight hospital labs and two reference labs did not answer this question. Seventy-two percent of hospital labs and 62% of reference labs process anaerobes on all shifts. Do you use transport media and batch specimens for processing? Sixty-nine percent of hospital labs and 80% of reference labs used transport media and batched specimens for processing. Do you use selective and differential media? One hundred percent of hospital labs and 93% of reference labs used selective media. 7b. If so, which media? Hospital labs
Reference labs
39 58 40 31
5 13 7 6
(52%) (77%) (53%) (41%)
BBE LKV PEA Other
(36%) (93%) (50%) (43%)
8. Do you identify anaerobes from blood cultures? All (100%) of hospital labs and 93% of reference labs identified anaerobes from blood cultures. All hospital and reference labs reported gram-stain
ARTICLE IN PRESS E.J.C. Goldstein et al. / Anaerobe 14 (2008) 68–72
70
results; 80% of hospital labs and 64% of reference labs reported all isolates to species level, while 17% of hospital labs and 18% of reference labs only reported to species levels ‘‘sometimes’’. Nineteen percent of hospital labs and 18% of reference labs reported isolates to genus level on a regular basis, while 1% of hospital labs and 18% of reference labs reported blood culture isolates to group level. 9. Do you identify anaerobes from wound cultures? Ninty-nine percent of hospital labs and 93% of reference labs identified anaerobes from wound cultures. All these hospital and reference labs report Gram-stain results; 69% of hospital labs and 64% of reference labs reported to species level and 24% and 18% reported to genus level, respectively, with the remainder reporting to group level only. 10. Do you identify B. fragilis group to species level? For hospital labs and reference labs 56% and 53% always reported to species, 37% and 13%, sometimes and 7%and 33% did not attempt this, respectively. 11. What methods do you use for anaerobe identification?
Special potency disks Preformed enzyme kit Fermentation reactions Individual enzyme tests Other biochemical tests Gas liquid chromatography Molecular Other methods
Hospital labs (%)
Reference labs (%)
30 66 21 7 30 4 0 4
47 60 13 0 60 7 0 0
12. Do you do anaerobic susceptibility testing? Overall, 56% of hospitals had access to susceptibility testing, with 41/76 (54%) of hospital labs that processed anaerobes had access to susceptibility testing but only 21/41 (51%) performed them in-house and 20/41 (49%) sent them to reference labs for performance. Of the 15 reference labs that returned the survey, 6/15 (40%) performed anaerobic susceptibility testing in-house, and 4 of these 15 (27%) sent the isolates out to yet another reference lab (5 said that they do not do anaerobic susceptibility testing). Of the 98 hospital labs (76+22 reference labs) that are accounted for in this study, only 21% of hospital labs perform anaerobic susceptibility testing, either inhouse or outsourced. 13. Which organisms would you test for susceptibility? Of those that performed susceptibility studies
Blood isolates: Other isolates all Selected
In-house (N ¼ 20) (%)
Sent out (N ¼ 6) (%)
100 56 44
100 50 50
Sterile body sites All isolates Selected isolates Surgical wounds: All isolates Selected isolates Other wounds: All isolates Selected isolates Other: upon request
85 47 53 70 15 85 50 10 90 40
100 33 67 67 50 50 67 25 75
14. What susceptibility method is used? Sixty-five percent of hospital labs used the Etest and 40% used microbroth dilution methods (all commercially prepared) compared to 17% and 100% for reference labs, respectively. No lab used the agar dilution method, and some used more than one method. 15. Which drugs do you test? Of the 98 hospital labs (76+22 reference labs) that are accounted for in this study, only 41 hospital labs and 10 reference labs report susceptibility testing results and only 21 hospital and 6 reference labs perform them in-house. Three hospital labs did not report on the drugs tested. The percentages are based on those denominators.
Beta-lactamase test only Penicillin/ampicillin Amoxicillin Ampicillin/sulbactam Piperacillin/tazobactam Clindamycin Metronidazole Moxifloxacin Ertapenem Imipenem Meropenem Doxycycline Tigecycline Cefoxitin Cefotetan Chloramphenicol Other drugs
Hospital labs (N ¼ 18) (%)
Reference labs (N ¼ 6) (%)
28 83 6 67 39 83 89 0 0 44 17 0 0 50 67 33 39
17 83 0 83 50 100 100 0 0 33 17 0 17 100 67 67 100
16. How do you test for C. difficile? Eighty-four percent of hospital labs and 92% of reference labs use an Elisa test for Toxin A+B; 11% of hospital labs but no reference labs use Toxin A testing only. Other tests used by hospital and reference labs were: tissue culture assay, 5% and 23%, respectively; stool culture, 1% and 15%, respectively; universal antigen-glutamate dehydrogenase (GDH) ELISA screen test, 3% but by no reference labs.
ARTICLE IN PRESS E.J.C. Goldstein et al. / Anaerobe 14 (2008) 68–72
4. Discussion The typical clinical presentation of patients with anaerobic bacteremia is less predictable then it was a decade ago [3]. In addition, anaerobes have exhibited new antimicrobial resistance patterns [13–15]. Despite technical improvements and increased standardization, our prior two surveys had noted a dramatic decline in the availability of anaerobic bacteriology and susceptibility data available to clinicians [1,2]. Consequently, they often must use surveys and small studies published from a handful of expert and specialized reference laboratories as a basis for empirical antibiotic selection. As a consequence of limited laboratory input, clinicians employ ‘‘top-of-the-line’’ broad-spectrum, often expensive antimicrobials for therapy of all anaerobic infections adding increased selective pressure for the development of new resistance and added cost to medical care [16]. In our last survey [2] performed more than a decade ago, only 65% of hospital (200–1000 beds) laboratories performed some degree of anaerobic bacteriology. Our current survey showed improvement with 76% now performing some degree of anaerobic bacteriology inhouse and an additional 23% send anaerobic specimens to a reference laboratory. Cultures from most clinically relevant sources such as blood, sterile sites and wounds are processed. Some amount of quality control is exerted with 78–86% rejecting poor quality specimens. Of those hospital laboratories that do process anaerobic cultures inhouse 100% use selective media including 52% that use BBE for rapid identification of the Bacteroides fragilis group compared to 31% in our prior 1993 study [2] and 20% in our 1990 survey [1]. The patterns of use of LKV and PEA media did not change over the past 16 years. Procedures for optimizing anaerobic culture methods are available from a variety of sources [11,12,17,18]. Sixty-two percent of laboratories also indicated interest in attending Anaerobic Bacteriology workshops that are offered at the Annual Meeting of the American Society for Microbiology, Anaerobe Society for the Americas meetings, by Anaerobe Systems, and at other venues. The value of performing anaerobic blood cultures has been questioned [19]. The importance of obtaining anaerobic blood cultures was recently highlighted by Lassmann et al. [3] who noted a 74% rise per 100,000 patient days in anaerobic bacteremias at the Mayo Clinic, with 35% not having an obvious primary anaerobic infection. In this survey, as in the prior surveys, all labs that processed anaerobes also processed anaerobic blood culture isolates; 67% of them identified the isolates to species level all the time and an additional 15% some of the time, while 16% tended to identify isolates only to genus level. This result is similar to those of 1990 when 66% of labs identified blood culture isolates to species level and 32% identified to genus level [1]. In addition, 99% of the labs currently also identify anaerobes from wound cultures with a similar speciation pattern compared to 80% in our
71
1990 survey [1]. B. fragilis was identified to species level by 56% and 53% of hospital and reference labs always, and an additional 37% and 13% sometimes. Surprisingly, 33% of reference labs did not speciate B. fragilis group isolates in spite of their known differences in resistance to commonly used antimicrobial agents. Evaluating the response to the performance of anaerobic susceptibility testing was somewhat complex and percentages dramatically change when extrapolated to the total sample. Overall, only 21% (21/98) of hospital labs perform anaerobic susceptibility testing in-house. Currently, 53% (41/75) of hospital labs and 67% (10/15) of reference labs have access to anaerobic susceptibility testing data. These rates continue to decline as in 1990 [1] 70% performed susceptibility testing that declined to 33% in 1993 [2]. Testing, when performed was universally done on blood isolates. It was performed by 85% (17/20) of labs for sterile body site isolates, and by 14/20 (70%) on mostly selected surgical wound isolates. An additional 40% (8/20), would also perform susceptibility testing by special request. Most hospital labs used the Etest (62%; 13/21) for susceptibility testing, while only 17% of reference labs used it. All reference labs used microbroth dilution for susceptibility testing as did 40% (8/20) of hospital labs [all commercially prepared]. No laboratory, hospital or reference, used the agar dilution method, which seems to be reserved to ‘‘centers-of-excellence’’ or research centers with an interest in anaerobes. Previously, in 1993, the microbroth dilution method was predominant in hospital labs (47%) [2] and the use of disk diffusion (33%) and broth disk elution method, which were predominant (56%)in 1990 [1] have disappeared from use. CLSI [10] does not describe ‘‘user friendly’’ methods that laboratories could easily adopt as standard procedures. The broth micro dilution technique set up using bench techniques is limited to the relatively oxygen-tolerant members of the B. fragilis group, although laboratories that have anaerobic chambers are able to use this method for testing more fastidious strains. The Etest is referred to as the ‘‘gradient method’’ in CLSI documents, but, as a proprietary product, is not mentioned by name. Its usefulness and reliability for testing anaerobes has been well documented in many publications [20–23]. Susceptibility data is now available to a very limited number of hospitals for a very few antimicrobial agents. Five hospital labs (28%) have only beta-lactamase testing results available. Since other mechanisms of beta-lactam resistance may be present, a negative test is not a guarantee of an organism’s susceptibility. Selection of the antimicrobials tested appears to be dependent upon the antimicrobials present on the commercial panels available to laboratories. No hospital lab tested for tigecycline, doxycycline, moxifloxacin or ertapenem susceptibility. The most frequently tested antimicrobials were penicillin/ ampicillin and clindamycin (15/18; 83%), metronidazole (16/18; 89%) and cefotetan and ampicillin/sulbactam (12/18; 67%). Other, more limited testing data was available for piperacillin/tazobactam (39%), cefoxitin
ARTICLE IN PRESS 72
E.J.C. Goldstein et al. / Anaerobe 14 (2008) 68–72
(50%), imipenem (44%), and chloramphenicol (33%). If one uses the base of 100 respondents then the most commonly reported results (to metronidazole) shrinks dramatically to 16% overall. Consequently, clinicians must be using information from the manufacturer, the FDA indications, published study/survey data or just guessing at appropriate empirical as well as directed therapy. Our current survey suggests that while many labs are processing anaerobic cultures, especially blood cultures, the identification of isolates and the performance of antimicrobial susceptibility testing of isolates lack a standard approach and need improvement. Acknowledgments This survey was supported in part by a grant form Merck & Co, Inc. We thank C. Vreni Merriam, Judee Knight and Alice E. Goldstein for various forms of support. References [1] Goldstein EJC, Citron DM, Goldman RJ. National hospital survey of anaerobic culture and susceptibility testing methods: results and recommendations for improvement. J Clin Microbiol 1992;30: 1529–34. [2] Goldstein EJC, Citron DM, Goldman RJ, Claros MC, HuntGerrado S. United States national hospital survey of anaerobic culture and susceptibility methods II. Anaerobe 1995;1:309–14. [3] Lassmann B, Gustafson DR, Wood CM, Rosenblatt JE. Reemergence of anaerobic bacteremia. Clin Infect Dis 2007;44:895–900. [4] Hecht DW. Routine anaerobic blood cultures: back where we started? Clin Infect Dis 2007;45:901–3. [5] Nguyen MH, Yu VL, Morris AJ, McDermott L, Wagener MW, et al. Antimicrobial resistance and clinical outcome of Bacteroides bacteremia: findings of a multicenter prospective observational trial. Clin Infect Dis 2000;30:870–6. [6] McGregor JC, Rich SE, Harris AD, Perenchevich EN, Osih R, et al. A systematic review of the methods used to assess the association between appropriate antibiotic therapy and mortality in bacteremic patients. Clin Inf Dis 2007;45:329–37. [7] Goldstein EJC, Citron DM, Hecht DW. Chapter XXX resistance in anaerobic bacteria. In: Fong IW, Drlica K, editors. Antimicrobial resistance and implications for the 21st century (Chapter 6, pp. 207–29) Berlin: Springer; 2007.
[8] Solomkin JS, Mazuski JE, Baron EJ, Bradley J, Chow A, Dellinger EP, et al., Guidelines for the selection of anti-infective agents for complicated intraabdominal infections. Clin Infect Dis, 2008, in press. [9] American Hospital Association. American Hospital Association guide to the healthcare field. Chicago: American Hospital Association; 2006. [10] Clinical and Laboratory Standards Institute. Methods for antimicrobial susceptibility testing of anaerobic bacteria. Approved standard, 7th ed., CLSI document M11-A7. Wayne, PA: CLSI; 2007. [11] Citron DM. rapid identification of anaerobes in the clinical laboratory. Anaerobe 1999;5:109–13. [12] Murray PR, Baron EJ, Jorgensen JH, Landry ML, Pfaller MA, editors. Manual of clinical microbiology. Washington, DC: American Society for Microbiology Press; 2007. [13] Snydman DR, Jacobus NV, McDermott LA, Ruthazer R, Goldstein EJC, Finegold SM, et al. National survey on the susceptibility of Bacteroides fragilis group: report and analysis of trends for 1997–2004: a US study. Antimicrob Agents Chemother 2007;51: 1649–55. [14] Koeth LM, Good CE, Appelbaum PC, Goldstein EJC, Rodloff AC, Claros M, et al. Surveillance of susceptibility patterns in 1297 European and US anaerobic and capnophilic isolates to co-amoxiclav acid and five other antimicrobial agents. J Antimicrob Chemother 2004;53:1039–44. [15] Golan Y, McDermott LA, Jacobus NV, Goldstein EJC, Finegold S, Harrell LJ, et al. Emergence of high-level fluoroquinolone resistance among Bacteroides species. J Antimicrob Chemother 2003;52:208–13. [16] Finegold SM. A century of anaerobes: a look backward and a call to arms. Clin Infect Dis 1993;16(Suppl. 4):S453–7. [17] Isenberg HD, editor. Clinical microbiology procedure manual. 2nd ed. Washington, DC: ASM Press; 2004. [18] Jousimies HR, Summanen P, Citron DM, Baron EJ, Wexler HM, Finegold SM. Wadsworth-KTL anaerobic bacteriology manual. Belmont, CA: Star Publishing Co.; 2002. [19] Bourgault AM, Harkness JL, Rosenblatt JE. Clinical usefulness of susceptibility testing of anaerobes. Arch Int Med 1978;138:1824–7. [20] Citron DM., Hecht DW. Susceptibility test methods: anaerobic bacteria. In: Manual clinical microbiology (Chapter 76, pp. 1214–22). Washington, DC: American Society Microbiology Press; 2007. [21] Citron DM, Ostovari MI, Karlsson A, Goldstein EJC. Evaluation of the E-test for susceptibility testing of anaerobic bacteria. J Clin Microbiol 1991;29:2197–203. [22] Rosenblatt JE, Gustafson DR. Evaluation of the Etest for the susceptibility testing of anaerobic bacteria. Diagn Microbiol Infect Dis 1995;22:279–84. [23] Rhomberg PR, Biedenbach DJ, Jones RN. Activity of BMS284756 (T-3811) tested against anaerobic bacteria Campylobacter jejuni, Helicobacter pylori, and Legionella spp. Diag Microbiol Infect Dis 2001;40:45–9.