Role of Combination Antibiogram in Empirical Treatment of Infection Due to Multidrug-ResistantAcinetobacter baumannii

Role of Combination Antibiogram in Empirical Treatment of Infection Due to Multidrug-Resistant Acinetobacter baumanniiAuthor(s): By Anucha Apisarnthanarak, MD; Linda M. Mundy, MDSource: Infection Control and Hospital Epidemiology, Vol. 29, No. 7 (July 2008), pp. 678-679Published by: The University of Chicago Press on behalf of The Society for Healthcare Epidemiologyof AmericaStable URL: http://www.jstor.org/stable/10.1086/588681 .

Accessed: 14/05/2014 05:20

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

The University of Chicago Press and The Society for Healthcare Epidemiology of America are collaboratingwith JSTOR to digitize, preserve and extend access to Infection Control and Hospital Epidemiology.

http://www.jstor.org

This content downloaded from 194.29.185.86 on Wed, 14 May 2014 05:20:12 AMAll use subject to JSTOR Terms and Conditions

http://www.jstor.org/action/showPublisher?publisherCode=ucpress

http://www.jstor.org/action/showPublisher?publisherCode=shea

http://www.jstor.org/action/showPublisher?publisherCode=shea

http://www.jstor.org/stable/10.1086/588681?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp


l e t t e r s t o t h e e d i t o r

infection control and hospital epidemiology july 2008, vol. 29, no. 7

Role of Combination Antibiogram inEmpirical Treatment of Infection Due to

Standards Institutes to identify A. baumannii and to establish an-timicrobial susceptibility profiles.6 Antimicrobial susceptibilitytesting was performed with conventional susceptibility micro-

007

scep

Multidrug-Resistant Acinetobacter baumannii

To the Editor—We read with interest the report by Mizutaand colleagues1 on the role of a combination antibiogramfor empirical treatment of Pseudomonas aeruginosa infec-tion. The emergence of multidrug-resistant (MDR) gram-negative microorganisms, especially Acinetobacter bauman-nii, has created prescribing dilemmas for physicians tryingto select empirical therapy.2 In Thailand, the national inci-dence of MDR–A. baumannii—which is defined as A. bau-mannii that is resistant to 3 or more classes of antimicrobialagents—peaked at 45% in 2006.3 Although dual therapy iscommonly used when P. aeruginosa infection is suspected,most infectious diseases experts in Thailand also recom-mend dual therapy for suspected MDR–A. baumannii infec-tions. Given reports that infections with MDR–A. bauman-nii were associated with higher mortality,4-5 one potentialoption is to use dual or triple empirical antimicrobial ther-apy. We, therefore, conducted a feasibility assessment todetermine the optimal initial therapy for patients withMDR–A. baumannii infection.

We identified all hospitalized adults who had A. baumanniiisolates recovered at Thammasat University Hospital fromJanuary 1, 2007, through December 31, 2007. If multiple A.baumannii isolates were obtained from the same patient dur-ing the same hospitalization, only the first isolate was evalu-ated. We used criteria suggested by the Clinical and Laboratory

table. Combination Antibiogram for Acinetobacter baumannii, 2

Drug

Percentage of isolates su

GEN AMI NET CEFP CETZ CP-SB

GEN . . . . . . . . . 28 21 39AMI . . . . . . . . . 30 23 42NET . . . . . . . . . 31 34 54CEFP 28 30 31 . . . . . . . . .CETZ 21 23 34 . . . . . . . . .CP-SB 39 42 54 . . . . . . . . .AM-SB 34 36 39 . . . . . . . . .CIP 21 24 31 25 20 35IMI 31 34 40 . . . . . . . . .MER 30 32 39 . . . . . . . . .PIP-TZ 36 35 35 . . . . . . . . .COL 100 100 100 100 100 100CP-SB � NET . . . . . . . . . . . . . . . . . .

note. A total of 560 A. baumannii isolates were identified. The combinatiothis analysis. AM-SB, ampicillin-sulbactam; AMI, amikacin; CEFP, cefepimsulbactam; CP-SB � NET, cefoperazone-sulbactam and netilmicin; GEN, genttazobactam.

This content downloaded from 194.29.185.8All use subject to JSTOR T

broth dilution trays, and we tested susceptibility for the followingantimicrobials: gentamicin, amikacin, netilmicin, cefepime, cefta-zidime, cefoperazone-sulbactam, ampicillin-sulbactam, cipro-floxacin, imipenem, meropenem, piperacillin-tazobactam, andcolistin. We created a standard antibiogram for A. baumannii iso-lates, along with annual combination antibiograms created in amatrix fashion, which listed the antimicrobial agents tested bothhorizontally and vertically, as had been previously done by Mizutaand colleagues.1 In each matrix box of the combination antibio-gram, we noted the percentage of isolates susceptible to at least 1of the 2 agents (Table). The backbone of noncolistin-based,2-antibiotic regimens for triple antimicrobial agents was selectedfrom the 2 antibiotics to which MDR–A. baumannii had the high-est percentage of susceptibility: cefoperazone-sulbactam (32%)and netilmicin (27%).

There were 560 A. baumannii isolates identified during thestudy period, of which 381 (68%) were recovered from urine,45 (8%) from blood, and 134 (24%) from other sites; 218 iso-lates (39%) were MDR–A. baumannii. The majority of isolates(308 [55%]) were recovered from intensive care unit patients.Antimicrobial susceptibility for A. baumannii as shown by thecombination antibiogram (dual vs triple antibiotics) revealedthat the combinations with the broadest coverage were consis-tently colistin-based regimens (100%), whereas cefoperazone-sulbactam and netilmicin (54%) provided the broadest cov-erage among noncolistin-based regimens (Table). Triplecombinations of imipenem, cefoperazone-sulbactam, and

tible to at least 1 of the 2 agents, by drug

M-SB CIP IMI MER PIP-TZ COL CP-SB � NET

34 21 31 30 36 100 . . .36 24 34 32 35 100 . . .39 31 40 39 35 100 . . .

. . . 25 . . . . . . . . . 100 . . .

. . . 20 . . . . . . . . . 100 . . .

. . . 31 . . . . . . . . . 100 . . .

. . . 31 . . . . . . . . . 100 . . .31 . . . 29 30 24 100 49

. . . 29 . . . . . . . . . 100 65

. . . 30 . . . . . . . . . 100 61

. . . 24 . . . . . . . . . 100 . . .100 100 100 100 100 . . . 100. . . 49 65 61 . . . 100 . . .

the same antibiotic or �-lactam with �-lactam antibiotics were excluded fromETZ, ceftazidime; CIP, ciprofloxacin; COL, colistin; CP-SB, cefoperazone-in; IMI, imipenem; MER, meropenem; NET, netilmicin; PIP-TZ, piperacillin-

A

n ofe; Camic

6 on Wed, 14 May 2014 05:20:12 AMerms and Conditions


netilmicin increased the coverage to 65% among noncolistin-based regimens. Although colistin-based combinations pro-vided the broadest coverage for infection with A. baumannii,

references

1. Mizuta M, Linkin DR, Nachamkin I, et al. Identification of optimal com-binations for empirical dual antimicrobial therapy of Pseudomonas aerugi-

letters to the editor 679

colistin has recognized adverse effects and low tissue pene-tration in lower respiratory tract infections.7 The triplenoncolistin-based regimens provided broader coverage thanthe dual noncolistin-based regimens for MDR–A. baumanniiinfections. These results were not substantially different whenthe analysis was repeated for the following subgroups: (1) iso-lates recovered from sites other than the urinary tract, (2) iso-lates recovered from the urinary tract, (3) isolates recoveredfrom patients in the intensive care unit, and (4) isolates recov-ered from patients outside the intensive care unit.

Although antibiograms are often used by clinicians to assesslocal antimicrobial susceptibility rates, as an aid in selectingempirical antibiotic therapy, and in monitoring resistancetrends over time in an institution, antibiograms do not revealadditional information concerning microbial isolates, such asthe time the isolate was obtained relative to the time of thepatient’s hospital admission (to determine whether an infec-tion was community acquired or healthcare acquired). In ad-dition, an antibiogram cannot be used to select empirical ther-apy for a patient who develops an infection subsequent to aprevious one, because a patient’s particular infection history,including past antibiotic use, must be considered.

Limitations of our study include the restricted analysis of A.baumannii isolates, instead of an effort to empirically target a va-riety of gram-negative pathogens. Our findings would requiremodification if the process was repeated in other institutions,given the wide local and regional variations in antimicrobial sus-ceptibility data. In addition, the ultimate choice of empirical an-timicrobial regimen will also rest on other factors, such as sus-pected pathogens, likely site of infection, drug allergies andintolerance, drug penetration into different tissue sites, and drugtoxicities. Nonetheless, the selection of empirical dual or triplecombinations via antibiogram provides a useful tool to guide phy-sicians in their initial decision making when MDR–A. baumanniiinfection is suspected in at-risk patients in endemic settings.

acknowledgmentsPotential conflicts of interest. Both authors report no conflicts of interest rel-evant to this article.

Anucha Apisarnthanarak, MD; Linda M. Mundy, MD

Division of Infectious Diseases, Faculty of Medicine, Thammasat UniversityHospital, Pratumthani, Thailand (A.A.); St. Louis University School of PublicHealth, St. Louis, Missouri (L.M.M.).

Address reprint requests to Anucha Apisarnthanarak, MD, Division of In-fectious Diseases Faculty of Medicine Thammasat University Hospital Pra-tumthani, Thailand, 12120 ([email protected]).

Infect Control Hosp Epidemiol 2008; 29:678 – 679© 2008 by The Society for Healthcare Epidemiology of America. All rightsreserved. 0899-823X/2008/2907-0018$15.00. DOI: 10.1086/588681


nosa infection: potential role of a combination antibiogram. Infect ControlHosp Epidemiol 2006;27:413-415.

2. Abbo A, Navon-Venezia S, Hammer-Muntz O, Krichali T, Siegman-Igra Y,Carmeli Y. Multidrug-resistant Acinetobacter baumannii. Emerg Infect Dis2005;11:22-29.

3. National Antimicrobial Resistant Surveillance Center. Annual epidemiologyand surveillance report, 2007. Available at: http://narst.dmsc.moph.go.th/. Ac-cessed January 1, 2008.

4. Abbo A, Carmeli Y, Navon-Venezia S, Siegman-Igra Y, Schwaber MJ. Im-pact of multi-drug-resistant Acinetobacter baumannii on clinical outcomes.Eur J Clin Microbiol Infect Dis 2007;26:793-800.

5. Lee NY, Lee HC, Ko NY, et al. Clinical and economic impact of multidrugresistance in nosocomial Acinetobacter baumannii bacteremia. Infect Con-trol Hosp Epidemiol 2007;28:713-719.

6. National Committee for Clinical Laboratory Standards (NCCLS).Methods for dilution: antimicrobial susceptibility testing for bacteriathat grow aerobically. 4th ed. NCCLS document. Villanova, PA: NCCLS;1997:M7-A3.

7. Falagas ME, Kasiakou SK. Colistin: the revival of polymyxins for the man-agement of multidrug-resistant gram-negative bacterial infections. Clin In-fect Dis 2005;40:1333-1341.

Relationship Between Pathogenic andColonizing Microorganisms Detectedin Intensive Care Unit Patients andin Their Family Members and Visitors

To the Editor—Recent data have demonstrated the usefulnessof an unrestricted visiting policy in the intensive care unit(ICU), the so-called “open ICU.”1-4 One of the most frequentobjections to the open ICU, despite the lack of empirical evi-dence, is an increased risk of patient infection.2,3,5 It is generallyargued that the transmission of microorganisms responsiblefor infections—so-called “cross-pollination” from visitors2—results from the presence of relatives in the ICU. Visitors andrelatives also run the risk of acquiring infection.5

We designed a prospective, observational, pilot study to testthe hypothesis that patients’ family members are healthy car-riers (reservoirs) of pathogens, which are, in turn, transmittedto patients, causing colonization or nosocomial infection. Thisstudy was conducted in an 8-bed, mixed medical-surgical ICU,with a nurse-to-patient ratio of 1:2. Patients in this ICU weretreated in 1 room with 4 beds and in 2 rooms with 2 beds.

Family members (2 visitors per patient) were admitted inthe afternoon from 12:30 pm–2:00 pm and from 6:30 pm–8:00pm. If the patient awakened or regained consciousness, the sec-ond afternoon visit can be extended from 4:00 pm–8:00 pm. Forpediatric patients, an unrestricted visiting policy was applied.

The visitors were required to wash their hands and wear adisposable gown; shoe-covers, gloves, and masks were not re-quired. Another hand washing was required on departure.

Using Margherita software (Istituto Mario Negri),6 we per-



table 1. Microbiological Data for Patients With Intensive Care Unit (ICU)–Acquired Infection and for Their Family Membersand Other Visitors

Patients Family members

680 infection control and hospital epidemiology july 2008, vol. 29, no. 7

formed routine surveillance of infections already present at thetime of ICU admission as well as ICU-acquired infections.Routine surveillance of tracheobronchial tree colonization byculture of an unprotected tracheal aspirate sample obtainedfrom intubated patients and/or those with tracheotomies wasperformed twice weekly, even if clinical signs of pulmonaryinfection were absent. If clinical signs of infection were present,diagnostic culture samples were collected from all possible sitesof infection prior to beginning empirical antibiotic therapy.

From May through August 2007, patients’ family members andother visitors were screened for bacterial and fungal contamina-tion or colonization on entry to the ICU and following handwashing. Swab samples from the nares and palmar skin were ob-tained. Family members and other visitors were screened only forpatients who stayed in the ICU for more than 6 days. Each day,according to a bed rotation, samples from 1 family member orvisitor (always the same person for each patient) of each of 2 dif-ferent patients were obtained. Once a week, samples from 2 familymembers or visitors were obtained prior to departure and beforehand washing to see if cross-contamination had occurred be-tween the patients and their family members and visitors. We

PatientType or site of

infection Causative organism

1 Pneumonia MRSA

2 Peritoneum, CVC Enterobacter cloacae, Pseudaeruginosa, Acinetobacterbaumanii, Klebsiella spec

3 Pneumonia, BSI, Skin A. baumanii, Klebsiella speEscherichia coli

4 Urinary tract Klebsiella species

5 Pneumonia Klebsiella species, A. bauma

6 Pneumonia, CVC E. coli, Klebsiella species

7 Pneumonia Klebsiella species

8 Pneumonia, CVC MSSA, Stenotrophomonas s9 Pneumonia, fever of

unknown originKlebsiella species

10 Pneumonia Proteus species

11 Skin A. baumanii12 Pneumonia Klebsiella species

13 Pneumonia Klebsiella species14 Cerebrospinal fluid MSSA

note. BSI, bloodstream infection; CoNS, coagulase-negative stapcatheter without bacteremia); MRSA, methicillin-resistant Staphylococca A. baumanii detected at departure from ICU, prior to hand washing.


collected all microorganisms recovered from routine surveillancecultures of tracheal aspirate samples and samples from other sites(urine, skin, cerebrospinal fluid from ventricular devices, andwound drainage) and from clinical cultures performed for ICUpatients.

Overall, 180 swab samples were obtained from 90 familymembers and visitors; a total of 16 nasal swab samples and 20skin swab samples were positive for pathogens. Twenty isolateswere recovered from 90 skin swab samples (13 coagulase-negative Staphylococcus isolates, 5 methicillin-susceptibleStaphylococcus aureus isolates, and 2 Acinetobacter speciesisolates). Sixteen isolates were collected from 90 nasal swabsamples (7 coagulase-negative staphylococci, 8 methicillin-susceptible Staphylococcus aureus, and 1 Aspergillus species).During that period, 19 ICU-acquired infections and 1 fever ofunknown origin occurred in 14 patients who stayed in the ICUfor longer than 6 days.

The site(s) of the patients’ infections and the contaminationand/or colonization status of family members and visitors isdetailed in Table 1. None of the microorganisms responsiblefor infection in patients was found on the skin or in the nares of

Cultureresults

Site ofcolonization Bacteria

1 positive, 5negative

Skin MSSA

nas 14 negative None None

2 negative None None


Skin MSSA


Skin A. baumaniia


Skin MSSA


Skin, nares MSSA from both sites

es 2 negative None None2 positive, 6

negativeSkin, nares CoNS from skin; MSSA

from nares2 positive, 8

negativeSkin, nares CoNS and A. baumaniia

from skin; CoNS fromnares

4 negative None None1 positive, 7

negativeNares MSSA

6 negative None None4 negative None None

occi; CVC, central venous catheter (infection due to central venousreus; MSSA, methicillin susceptible S. aureus.

omo

iescies,

nii

peci

hylocus au



family members or visitors. No correlation was found betweenisolates recovered from routine surveillance cultures done forpatients and isolates found to be colonizing or contaminating

unit: results from a pilot, randomized trial. Circulation 2006;113:946 –952.

5. Quinio P, Savry C, Deghelt A, Guilloux M, Catineau J, De Tinteniac A. A


patients’ family members or visitors.The duration of family visits in our ICU (3 or more hours

per day) is no shorter than the visiting times reported in othercountries.5,7,8 In addition, 18 of 33 patients had a length of stayof 15 days or more. Thus, limited exposure time is not a goodexplanation for the lack of correlation between isolates recov-ered from patients and isolates recovered from their respectivefamily members and visitors.

Hand washing is recommended as the most effective meansto prevent transmission by direct contact, because it reducesthe concentration of contaminants on the skin.9 We did detectcontamination with A. baumanii on the hands of visitors priorto hand washing at departure, an organism that was also iso-lated from the patients that they visited. After that, these visi-tors were monitored for the presence of A. baumanii prior toentry to the ICU and after hand washing; A. baumanii was notisolated again from these individuals. We believe the lack ofcorrelation between the isolates recovered from patients andthose recovered from family members may relate to our handwashing policies.

Fumagalli et al.4 have shown that an unrestricted visitationpolicy, despite imposing a greater microbial burden andgreater environmental contamination, does not increase therisk of infectious complications in cardiac ICU patients. Po-tential pathogens isolated from patients do not appear to be thesame as those carried by their family members and visitors, nordoes exposure to these pathogens increase the risk of infectionin the ICU if appropriate hand hygiene is enforced.

acknowledgmentsPotential conflicts of interest. All authors report no conflicts of interest rele-vant to this article.

Paolo Malacarne, MD; Silvia Pini, MD; Nunzio De Feo, RN

From the Unitá Operativa, Anestesia e Rianimazione Azienda Ospedaliera–Universitaria Pisana, Pisa, Italy (all authors).

Address reprint requests to Paolo Malacarne, UO Anestesia e RianimazioneAzienda Ospedaliera–Universitaria Pisana, Via Roma 2, Pisa, Italy([email protected]).

Infect Control Hosp Epidemiol 2008; 29:679 – 681© 2008 by The Society for Healthcare Epidemiology of America. All rightsreserved. 0899-823X/2008/2907-0019$15.00. DOI: 10.1086/588703

references

1. Berwick DM, Kotagal M. Restricted visiting hours in ICUs: time to change.JAMA 2004;292:736 –737.

2. Giannini A. Open intensive care units: the case in favour. Minerva Anesth2007;73:299 –305.

3. Burchardi H. Let’s open the door. Intensive Care Med 2002;28:1371–1372.4. Fumagalli S, Boncinelli L, Lo Nostro A, et al. Reduced cardiocircula-

tory complications with unrestrictive visiting policy in an intensive care


multicenter survey of visiting policies in French intensive care units. Inten-sive Care Med 2002;28:1389 –1394.

6. Malacarne P, Langer M, Nascimben E, et al. Building a continuous multi-center infection surveillance system in the intensive care unit: findings fromthe initial data set of 9493 patients from 71 Italian intensive care units. CritCare Med 2008;36:1105–1113.

7. Berti D, Ferdinande P, Moons P. Beliefs and attitudes of intensive carenurses toward visits and open visiting policy. Intensive Care Med 2007;33:1060 –1065.

8. Lee M, Friedenberg A, Mukpo D, Conray Kayla, Palmisciano A, Levy M.Visiting hours policies in New England intensive care units: strategies forimprovement. Crit Care Med 2007;35:497–501.

9. Boyce JM, Pittet D. Guideline for hand hygiene in health-care settings:recommendations of the Healthcare Infection Control Practices AdvisoryCommittee and the HICPAC/SHEA/APIC/IDSA Hand Hygiene TaskForce. MMWR Recomm Rep 2002;51(RR-16):1– 45.

Whole-Blood Interferon-Gamma ReleaseAssay for Baseline Tuberculosis Screeningof Healthcare Workers at a SwissUniversity Hospital

To the Editor—In countries with low tuberculosis incidencerates, such as Switzerland,1 targeted testing for latent tubercu-losis infection (LTBI) among risk groups such as healthcareworkers (HCWs) is an important measure for preventing tu-berculosis disease.2-4 We studied the prevalence of LTBI and itsrisk factors among hospital employees at the University Hos-pital of Berne, Switzerland, in a retrospective cohort study us-ing a whole-blood interferon-gamma release assay (IGRA).5

From June 1, 2005, through May 31, 2006, we screened 777employees for tuberculosis infection with an IGRA on com-mencement of employment. The following data were collectedfor each HCW: age, sex, bacille Calmette-Guérin (BCG) vacci-nation status (documented or reported), country of origin,place of work, and profession. The mean age of subjects was 32years (75% were aged 20 – 40 years). The majority (70.8%) ofemployees were female (Table). The overall BCG vaccinationrate was 87.4% (90.4% among employees of Swiss origin, ofwhom 12.1% had multiple BCG vaccinations). The IGRA usedfor screening (QuantiFeron-TB Gold In-Tube assay; Cellestis)was performed according to the manufacturer’s instructions.6

Data were evaluated by univariate analysis as well as multiplelogistic regression analysis. GraphPad Prism 4, version 4.01(GraphPad Software), and StatView, version 5.0 (SAS Insti-tute), were used for all data evaluations.

A positive IGRA result was found for 59 (7.6%) of the 777participants (Table). Tuberculosis disease was ruled out ineach case by a careful consideration of the medical history, thesymptoms, and the chest X-ray findings. The overall rate ofLTBI in our study population was 7.6%, which concurs with



table. Results of Univariate Analysis of Risk Factors for Latent Tuberculosis (TB) Infection Among 777Healthcare Workers (HCWs) Screened

No. (%) of HCWs with risk

682 infection control and hospital epidemiology july 2008, vol. 29, no. 7

published data for countries with an incidence of tuberculosiscomparable to that of Switzerland. A recent German study7

found positive IGRA results for 25 (9.6%) of 261 participants.There was no association between the IGRA result and pro-

fession, age, or sex (Table). Employees were stratified for risk ofLTBI into 2 groups, according to their country of origin. Anannual incidence of tuberculosis disease of 10 cases per 100,000population was chosen as the cutoff 1; employees from a coun-try with a higher rate were placed in group B, and all otherparticipants were placed in group A. The IGRA positivity ratefor employees from group A was 6.0%; the rate was 18.9% foremployees from group B. There was a statistically significantpositive association between a positive IGRA result and mem-bership in group B (odds ratio [OR], 3.65 [95% confidenceinterval {CI}, 2.0 – 6.7]; P � .001).

The work place–associated risk of contracting tuberculosiswas distributed as follows among the HCWs screened: low risk,217 HCWs (27.9%); moderate risk, 398 (51.2%); high risk, 162

FactorNo. (%) ofall HCWs

fact

Posit(n �

Age, years, mean � SD 33.7 �Sex

Female 550 (70.8) 44 (8.Male 227 (29.2) 15 (6.

Country of origin, byincidence of TBa

Group A (low) 682 (87.8) 41 (6.Group B (high) 95 (12.2) 18 (18

Degree of risk ofexposure to TB

By place of workLow 217 (27.9) 11 (5.Moderate 398 (51.2) 30 (7.High 162 (20.8) 18 (11

By professionLow 221 (28.4) 15 (6.Moderate 198 (25.5) 14 (7.High 358 (46.1) 30 (8.

Receipt of BCGvaccination

No 63 (8.1) 6 (9.Yes 428 (55.1) 28 (6.Unknown 286 (36.8) 25 (8.

note. BCG, Bacille Calmette-Guérin; CI, confidencea The cutoff for TB incidence was 10 cases per 100,000 popGermany, Switzerland, Finland, Denmark, the Netherlannated in Algeria, Argentina, Austria, Bosnia and HerzegEthiopia, Great Britain, Greece, Croatia, Hungary, India, Iformer Yugoslav Republic of Macedonia, Philippines, PoSudan, Senegal, Spain, Thailand, Tunisia, Turkey, Uruguab By the �2 test for trend.


(20.8%) (Table). The work places considered to have a highrisk were the Departments of Thoracic Surgery, Pneumology,Rheumatology, Infectious Diseases, Anesthesiology, IntensiveCare, and Emergency Medicine. Nonmedical places of work,such as the laundry, the restaurant, the main office, and theadministrative office, were considered to have low risk.

The high-risk category of professions was considered tocomprise nurses, physiotherapists, radiographers, and mem-bers of the maintenance, cleaning, and transportation services.All other HCWs, such as medical doctors or social workers,were considered to be in the moderate-risk category. Profes-sions without direct patient contact, such as secretaries, com-puter and technical specialists, scientists, pharmacists, and se-curity staff, were considered to have low risk.

The IGRA positivity rate was 5.0% for HCWs who worked inlow-risk work places, 7.5% for HCWs in moderate-risk workplaces, and 11.1% for HCWs in high-risk work places (Table).We detected a statistically significant trend for association of a

y IGRA result

Odds ratio(95% CI) P

Negative(n � 718)

31.6 � 9.6 1.02 (0.99–1.04) .100.81 (0.44–1.49) .50

506 (92.0)212 (93.4)

3.65 (2.00–6.67) �.001641 (94.0)

77 (81.1)

.03b

206 (95.0)368 (92.5) 1.52 (0.74–3.11) .24144 (88.9) 2.34 (1.07–5.10) .03

.46d

206 (93.2)184 (93.0) 1.04 (0.49–2.22) .91328 (91.6) 1.25 (0.66–2.39) .49

57 (90.5)400 (93.5) 0.66 (0.26–1.67) .39261 (91.3)

val; IGRA, interferon-� release assay.ion (see text for details). Patients in Group A originated inweden, and the United States. Patients in Group B origi-a, Brazil, Czech Republic, Dominican Republic, Egypt,

Iraq, Japan, Republic of Korea, Sri Lanka, Luxemburg, the, Portugal, Russian Federation, Serbia and Montenegro,iet Nam, and Yemen.

or, b

ive59)

11.1

0)6)

0).9)

0)5).1)

8)0)4)

5)5)7)

interulat

ds, Sovin

ran,landy, V



positive IGRA result with work in a place with a high risk (�2

test for trend, P � .029). Both the country of origin and place ofwork were independent risk factors associated with a positive

Potential conflicts of interest. All authors report no conflicts of interestrelevant to this article.

Ariane Stebler, MD; Patricia Iseli, MD;


IGRA result, according to multiple logistic regression analysis.For employees originating from a group B country, the OR was3.65 (95% CI, 2.00 – 6.67); the P value was less than .001. Foremployees working at a place with a high risk of tuberculosisexposure, the OR was 2.34 (95% CI, 1.07–5.10); the P value was.03.

Our findings are consistent with reports by other groups. AJapanese study8 found no association between the IGRA resultand the profession of the subjects tested. An analysis of infec-tion risk related to the place of work was not done, however. Arecently published German study9 reported an increased infec-tion risk for HCWs engaged in geriatric care, but this studyincluded no other hospital units. It can be argued that “place ofwork” is a surrogate for risk of exposure to contagious tuber-culosis related to particular patient populations, whereas “pro-fession” represents the intensity of contact during patient care.Since infectious tuberculosis is highly contagious, the lattermay not be as important as the risk of being exposed at all. Theplace of work therefore seems to be a better surrogate thanprofession for the risk of contracting tuberculosis infection. Incontrast to our study, both the Japanese and the German stud-ies8,9 reported a strong association between a positive IGRAresult and increasing age among HCWs. The narrow age dis-tribution in our study population (mean � SD, 32 � 9.7 years)might have reduced our ability to detect such an association.(In the German study, the mean age [�SD] was 40.0 � 10.4years9; in the Japanese study, it was 41.4 � 11.2 years.8)

In conclusion, the IGRA was a useful tool for baselinescreening for LTBI in this population of HCWs with a highbackground of BCG vaccination. Hospital employees had ameasurable extra risk of tuberculosis infection that was associ-ated with certain places of work and with their country of ori-gin but not with their profession.

acknowledgmentsWe thank Kathrin Franz and Susanna Bigler for their helpful discussions,Monika Gimmel for assistance with data collection, and the laboratory staff fortechnical support. We thank W. J. Looney for reading the manuscript.


Kathrin Mühlemann, MD, PhD; Thomas Bodmer, MD

From the Institute for Infectious Diseases, University of Berne (A.S., K.M.,T.B.), and the Occupational Health Care Service, University Hospital of Berne(P.I.), Berne, Switzerland.

Address reprint requests to Thomas Bodmer, MD, University of Berne, In-stitute for Infectious Diseases, Friedbühlstrasse 51, CH-3010 Berne, Switzer-land ([email protected]).

Infect Control Hosp Epidemiol 2008; 29:681– 683© 2008 by The Society for Healthcare Epidemiology of America. All rightsreserved. 0899-823X/2008/2907-0020$15.00. DOI: 10.1086/588587

references

1. World Health Organization. Global TB database. Available at: http://www.who.int./tb/country/en/index.html. Accessed October 2006.

2. American Thoracic Society. Targeted tuberculin testing and treatment oflatent tuberculosis infection. Am J Respir Crit Care Med 2000;161: S221–S247.

3. Jensen PA, Lambert LA, Iademarco MF, Ridzon R; Centers for DiseaseControl. Guidelines for preventing the transmission of Mycobacterium tu-berculosis in health-care settings. MMWR Morb Mortal Wkly Rep 2005;54:1–141.

4. Jasmer RM, Nahid P, Hopewell PC. Latent tuberculosis infection. N EnglJ Med 2002;347:1860 –1866.

5. Pai M, Riley LW, Colford JM Jr. Interferon-gamma assays in the immuno-diagnosis of tuberculosis: a systematic review. Lancet Infect Dis 2004;4:761–776.

6. Cellestis. QuantiFeron-TB Gold In-Tube: technical information; packageinsert. Accessed January 2008.

7. Nienhaus A, Schablon A, Bâcle CL, Siano B, Diel R. Evaluation of the inter-feron-�–release assay in healthcare workers. Int Arch Occup Environ Health2008;81:295–300.

8. Harada N, Nakajima Y, Higuchi K, Sekiya Y, Rothel J, Mori T. Screening fortuberculosis infection using whole-blood interferon-� and Mantoux test-ing among Japanese healthcare workers. Infect Control and Hosp Epidemiol2006;27:442– 448.

9. Nienhaus A, Schablon A, Loddenkemper A, Hauer B, Wolf N, Diel R. Prev-alence of latent tuberculosis infection in healthcare workers in geriatriccare. Pneumologie 2007;61:613– 616.



Documents

Role of Combination Antibiogram in Empirical Treatment of Infection Due to Multidrug-ResistantAcinetobacter baumannii