Polyclonal Staphylococcus epidermidis intravascular catheter-related infections

CONCISE COMMUNICATIONS

Brucella detection in blood: comparison of the BacT/Alert standard aerobic bottle, BacT/Alert FAN aerobic bottle and BacT/Alert enhanced FAN aerobic bottle in simulated blood culture B. Su¨merkan, S. Go¨kahmetoglu and D. Esel Department of Clinical Microbiology, Erciyes University, Kayseri, Turkey

The objective of this study was to compare the performances of the standard aerobic bottle (StAe), FAN aerobic (FANAe) and enhanced FAN aerobic (E-FANAe) (the charcoal component of the FANAe was revised recently to improve the feasibility of Gram smear interpretation) blood culture bottles for BacT/Alert system for the detection of Brucella melitensis in simulated blood culture. Triplicate strains of eight clinical isolates of B. melitensis were studied. Each bottle was inoculated with 5 mL of freshly collected human blood at three different targeted bacterial inocula (101, 102 and 103 CFU/bottle). All bottles were monitored for up to 21 days or until they became positive. The results of time to detection (TTD) on the eight B. melitensis samples were as follows: at 101 CFU/bottle, the E-FANAe had a mean TTD significantly shorter than the StAe (48 h vs. 56.2 h, P < 0.05); and at 103 CFU/bottle, the FANAe and E-FANAe had a mean TTD significantly shorter than the StAe (41.2 h and 40 h vs. 45.6 h, P < 0.05). The reproducibilities (no.of positive signals/no.of all bottles) of three bottle systems were as follows: at 101 CFU/bottle, the reproducibilities of StAe, FANAe and E-FANAe were 96, 83 and 58%, respectively. At 103 CFU/bottle, the reproducibilities of StAe, FANAe and E-FANAe were 95, 95 and 91%, respectively. Positive results for the presence of bacteria in Gram smears were confirmed in 68% of StAe, 54% of FANAe and 90% of E-FANAe. In case of suspected brucellosis, the combination of one StAe bottle and one E-FANAe bottle seems to provide the highest and fastest recovery of the organism. Keywords Brucella, BacT/Alert blood culture system Accepted 26 April 2001

Clin Microbiol Infect 2001; 7: 369–393

IN T R ODUC T ION Brucellae are intracellular parasites that cause septicemic febrile illness or localised infection of bone, tissue, or organ systems in humans. Brucellosis in humans has a variable incubation period, an insidious or abrupt onset, and no pathognomonic symptoms or signs [1]. Diagnosis of brucellosis is based on bacteriological and serological tests, and blood cultures should be obtained when brucellosis is suspected. Primary isolation of Brucella spp. is difficult because it is a slow-growing bacterium and conventional diphasic blood culture bottles may require an incubation period of weeks [2]. Automated blood culture systems seem to shorten the time to detect these organisms from blood and other body fluids [1].

¨ niversitesi, Corresponding author and reprint requests: B. Su¨merkan, Erciyes U Tip Faku¨ltesi, Mikrobiyoloji ve Klinik Mikrobiyoloji Anabilim Dali, 38039 Kayseri, Turkey Tel: þ90 3524374910 Fax: þ90 3524375288 E-mail: [email protected]

The BacT/Alert Microbial Detection System (Organon Teknika Corp., Durham, NC, USA) was introduced in 1990 as an automated colorimetric blood culture system consisting of standard aerobic (StAe) and anaerobic blood culture bottles and pediatric aerobic bottles, containing growth sensors, for detecting microbial growth [3]. The manufacturer also developed an aerobic medium (FANAe) with a brain–heart infusion base containing Ecosorb. Ecosorb is a proprietary substance that contains adsorbent charcoal, Fuller’s earth, and other components. FAN media were developed to enhance the recovery of fastidious organisms from blood, as well as to improve the detection of bacteremia and fungemia in patients receiving antimicrobial agents. Recently, the charcoal component of the FAN media was revised to improve the feasibility of Gram smear interpretation. These revised media were termed enhanced FAN (E-FANAe). Brucellosis caused by B. melitensis is endemic in our region. In our laboratory, approximately 50–60 B. melitensis strains are isolated from 8000 blood cultures per year. The aim of this study was to evaluate the performance of the three different bottles (StAe, FANAe and E-FANAe) for the detection of B. melitensis in seeded blood cultures.

ß 2001 Copyright by the European Society of Clinical Microbiology and Infectious Diseases

370 Clinical Microbiology and Infection, Volume 7 Number 7, July 2001

Statistical analysis

M AT ER I AL S A ND ME T HOD S

Data were analyzed by ANOVA and the posthoc Scheffe procedure. P-values lower than 0.05 were accepted as statistically significant.

Study design StAe, FANAe, and E-FANAe blood culture bottles for the BacT/Alert system were used. The bottles were supplied by the manufacturer (Organon Teknika Corp.). Eight recent blood isolates of B. melitensis, all of which were selected arbitrarily, were tested. These were isolated from patients with brucellosis and identified by standard laboratory methods in our laboratory. Brucella strains were inoculated on sheep blood agar and incubated for 48 h at 35 8C in an aerobic atmosphere. Suspensions of these growths were made in 10 mL of trypticase soy (TS) broth (BBL) and adjusted to a McFarland 0.5 standard. These suspensions (containing approximately 108 colony-forming units (CFU)/mL) were subsequently diluted with TS broth to achieve an organism concentration range of between 101 and 103 CFU/mL. Colony counts were performed to verify the actual concentrations. Blood culture bottles were each inoculated with 5 mL of fresh blood collected from known healthy volunteers who had been checked for Brucella antibodies by the Wright agglutination test before the study. Triplicate bottles were inoculated with 1 mL of each the bacterial suspensions containing 101, 102 and 103 CFU/mL. All bottles were transiently vented, placed in the BacT/Alert instrument under continuous agitation and monitored for up to 21 days or until they became positive. When a positive bottle was flagged, a Gram stain of the broth was performed, and a portion of the fluid was subcultured on TS agar medium with 5% sheep blood. In addition, blind subcultures of all terminally negative bottles were performed using sheep blood agar. Gram stain results, time to detection (TTD) of positive results and blind subculture results were recorded. Contaminated bottles were excluded from the study.

R E S ULT S Time to detection analyses of B. melitensis strains from the three different bottles are shown in Table 1. The recovery of B. melitensis strains in the FANAe and E-FANAe bottles was faster than in the StAe bottles. The performances of the three bottle systems are shown in Table 2. False negative (detection negative and subculture positive) results were 3, 4.6, and 7.5% in StAe, FANAe, and E-FANAe, respectively. The reproducibilities (no. of positive signals/no. of all bottles) of the FANAe and E-FANAe bottles were found to be lower when compared with the StAe bottles (Figure 1). Positive results for the presence of

Figure 1 Percentages of recovery of B. melitensis by bottle systems and inocula.

Table 1 Time to detection analyses of B. melitensis strains from the three different bottles Time to detection (h), mean  SD CFU/bottle

n

StAe

n

FANAe

n

E-FANAe

P

101 102 103 Total

23 22 21 66

56.2  7.9 51.3  7.6 45.6  5.3 51.2  8.2

19 21 19 59

51.5  5.8 46.7  5.3a 41.2  3.4a 46.5  6.4a

14 15 20 49

48.0  3.6a 44.2  2.3a 40.0  3.5a 43.6  4.6a

<0.05 <0.05 <0.05 <0.05

a

Significantly different from StAe.

Table 2 Overall performances of three bottle systems (%)

Bottle system/no.of bottles

Detection positive, subculture positive

Detection negative, subculture negative

Detection positive, subculture negative

Detection negative, subculture positive

StAe/69 FANAe/66 E-FANAe/67

95.7 89.4 73.1

1.3 6 19.4

0 0 0

3 4.6 7.5

ß 2001 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 7, 369–393

Concise Communications

bacteria with Gram smears were confirmed in 68% of StAe, 54% of FANAe, and 90% of E-FANAe. DI SCUS S ION A definitive diagnosis of brucellosis is based on the culture of Brucella strains from different samples, mainly blood. Brothbased blood culture systems have gained universal acceptance for the isolation of Brucellae, but the slow growth of the organism has hampered this method. Blood cultures of patients with suspected brucellosis need a long incubation time (30 days) and periodic blind subcultures to avoid false negative results [4]. The time to isolation of the organism has been decreased by using biphasic Castan˜eda medium and a lysis-filtration method [2]. These techniques are time-consuming and labor-intensive and require extensive manipulation of specimens. They also pose a substantial risk for laboratory staff [5]. However, automated blood culture systems have reduced the detection time of these organisms [6–8]. The BacT/Alert is an automated blood culture system in which the growth of organisms produces increased amounts of CO2, which diffuses through a semipermeable membrane in the base of the culture bottle and reacts with water to generate hydrogen ions. This causes a decrease in pH, resulting in a color change of a built-in sensor. In this study we compared the performance of three bottle systems of BacT/ Alert (StAe, FANAe, and E-FANAe) for the detection of B. melitensis in seeded blood cultures. The recovery of B. melitensis strains in FANAe and E-FANAe bottles was faster than in StAe bottles (Table 1). This may be due to the differences between the media types (brain–heart infusion broth vs. tripticase soy broth) as well as to the components found in the bottle systems. However, these data should not be significant in a clinical setting. Solomon et al. [9] demonstrated a mean detection time of 48  1 h in eight of 10 replicates seeded with a stock containing 102 CFU/mL of B. melitensis by using BacT/Alert StAe bottles. Zimmermann et al. [10] performed seeded blood culture studies of a B. abortus isolate with fresh human blood and target inocula of both 5 and 500 CFU/mL with the BACTEC NR 730 system and indicated that the larger (500 CFU/mL) inoculum produced positive instrument detection within 2 days, whereas the smaller (5 CFU/mL) inoculum required 5.5–7.5 days for detection, depending on the medium used. As expected, the number of seeded microorganisms was correlated inversely with the TTD of bacterial growth in all bottle systems. Yagupsky et al. [11] have shown that the magnitude of Brucella bacteremia correlated inversely with the TTD of the organism, using another automated system, BACTEC 9240. The performances of the three bottle systems are shown in Table 2. No false positive (detection positive, subculture negative) result was detected in any bottle system. False negative

371

(detection negative, subculture positive) results were found in 3% of StAe bottles, 4.6% of FANAe bottles, and 7.5% of E-FANAe bottles. StAe bottles showed a higher performance when compared with FANAe and E-FANAe (detection positive, subculture positive were 95.7, 89.4, and 73.1%, respectively), regardless of the magnitude of bacteria. The reproduciblity of the FANAe and E-FANAe bottles was found to be lower when compared with StAe bottles (Figure 1). The amount of sodium polyanethol sulfonate (SPS) in StAe bottles was lower than in the FANAe and E-FANAe bottles. SPS, used as an anticoagulant in many blood cultures, exerts a harmful effect on the outer membrane of the bacteria, making it permeable to hydrophobic substances and thus hindering growth [12]. Gamazo et al. [13] found that growth values of B. melitensis in an automated blood culture system were lower in vials with SPS than in vials without this agent. In our study, positive results for the presence of bacteria with Gram smears were confirmed in 68% of StAe, 54% of FANAe, and 90% of E-FANAe. The Gram stain results from the E-FANAe bottles were better than from the FANAe bottles since the charcoal component of the FAN media was revised to improve the feasibility of Gram smear interpretation. The poor counterstaining quality of brucellae might also preclude detection by Gram stain [14]. Therefore a negative Gram stain should not be used to rule out a signal positive result, especially in blood cultures of patients suspected of having brucellosis. In the case of suspected brucellosis, the combination of one StAe bottle and one E-FANAe bottle seems to provide the highest and fastest recovery of the organism. The results of this study need to be supported by further clinical investigations. ACK NOW L EDGME N T S We thank Prof. Dr Osman Gu¨ nay for helping with statistical analysis. We also thank Organon Teknika for the support for this study.

R EFER E NCE S 1. Mayer NP, Holcomb LA. Brucella. In: Murray PR, Baron JE, Pfaller MA, Tenover FC, Yolken RH, eds. Manual of clinical microbiology. Washington. DC: American Society for Microbiology, 1995: 549–55. 2. Ruiz Casten˜ ada M. Laboratory diagnosis of brucellosis in man. Bull WHO 1961; 24: 73–84. 3. Rohner P, Auckenthaler R. Review on evaluations of currently available blood culture systems. Clin Microbiol Infect 1999; 5: 513–29. 4. Navas E, Guerrero A, Cobo J, Loza E. Faster isolation of Brucella spp. from blood by isolator compared with BACTEC NR. Diagn Microbiol Infect Dis 1993; 16: 79–81. 5. Staszkiewicz J, Lewis C, Colville M, Zervos JM, Band J. Outbreak of Brucella melitensis among microbiology laboratory workers in a community hospital. J Clin Microbiol 1991; 29: 287–90.

ß 2001 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 7, 369–393

372 Clinical Microbiology and Infection, Volume 7 Number 7, July 2001

6. Bannatyne RM, Jackson MC, Memish Z. Rapid diagnosis of Brucella bacteremia by using the Bactec 9240 system. J Clin Microbiol 1997; 35: 2673–4. 7. Ruiz J, Lorente I, Pe´ rez J, Simarro E, Martı´nez-Campos L. Diagnosis of brucellosis by using blood cultures. J Clin Microbiol 1997; 35: 2417–8. 8. Yagupsky P. Detection of Brucella melitensis by BACTEC NR660 blood culture system. J Clin Microbiol 1994; 32: 1899–901. 9. Solomon HM, Jackson D. Rapid diagnosis of Brucella melitensis in blood: some operational characteristics of the BACT/ALERT. J Clin Microbiol 1992; 30: 222–4. 10. Zimmerman SJ, Gillikin S, Sofat N, Bartholomew WR, Amsterdam D. Case report and seeded blood culture study of Brucella bacteremia. J Clin Microbiol 1990; 28: 2139–41.

11. Yagupsky P, Peled N, Press J, Abramson O, Abu-Rashid M. Comparison of BACTEC 9240 Peds plus medium and isolator 1.5 microbial tube for detection of Brucella melitensis from blood cultures. J Clin Microbiol 1997; 35: 1382–4. 12. Lysko PG, Morse SA. Neisseria gonorrhoeae cell envelope: permeability to hydrophobic molecules. J Bacteriol 1981; 145: 946–52. 13. Gamazo C, Vitas AI, Lo´ pez-Goni I, Diaz R, Moriyo´ n I. . Factors affecting detection of Brucella melitensis by BACTEC NR730, a non-radiometric system for hemocultures. J Clin Microbiol 1993; 31: 3200–3. 14. Dunne WA Jr, Nolte FS, Wilson ML. Cumitech 1B. In: Blood cultures III. Hindler JA, Coordinating ed. Washington, DC: American Society for Microbiology, 1997.

Serodiagnosis of tuberculosis by enzyme immunoassay using A60 antigen 

A. Yu¨ce1 , M. Yu¨cesoy2, S. Genc¸3, M. Sayan2 and E. S. Uc¸an3 Departments of 1Microbiology and Infectious Diseases, 2Chest Diseases and 3Microbiology and Clinical Microbiology, Dokuz Eylu¨ l University, Medical Faculty Incirati, 35340 Izmir, Turkey 

Tel: þ90 232 259 59 59/2579

Fax: þ90 232 259 05 41

E-mail: [email protected]

Accepted 1 December 2000

Tuberculosis continues to be a worldwide public-health problem, with an estimated 90 million new cases and 30 million deaths during the last decade of the 20th century [1]. It has been an ever-present health threat in developing countries. At the same time, it has been increasing in Europe and the USA following the emergence of AIDS in the world [2]. The prevalence of tuberculosis in Turkey is 0.35%, and 30 000– 40 000 new cases are being reported every year [3]. An estimated 2 billion people are currently infected with Mycobacterium tuberculosis and Mycobacterium intracellulare complex throughout the world [4]. The rates of morbidity and mortality are also rising as a result of multidrug-resistant strains [5]. The diagnosis of mycobacterial diseases depends upon identifying the infecting organism in the secretion or tissues. However, there are several limitations of this method. One is that Mycobacterium tuberculosis is usually present in undetectable numbers (for smear 5  103 to 5  104 bacilli/mL, and for culture 10–100 bacilli/mL), so that it is recognized for the most part in advanced cases [6]. Second, most mycobacteria are slow-growing organisms and require long periods of time to culture, even if the most advanced techniques are used [6,7]. Third, negative smears are usually obtained until cavities form [8]. Thus, a fast, easy and reliable method was needed for the diagnosis of tuberculosis. Among several serologic techniques, it has been concluded that enzyme immunoassay (EIA) is a

sensitive, reliable, simple and rapid method [6]. Several purified antigens, such as 38-kDa protein, 85A antigen, lipoarabinomannan, plasma membrane antigen, antigen 5 and antigen 60 (A60), have been used for the serodiagnosis of tuberculosis [8–12]. This study was undertaken to evaluate the usefulness of the EIA method using A60 antigen for the diagnosis of different forms of tuberculosis in Turkish patients. Serum samples were collected from four groups of patients and the control group. Sera from active tuberculosis patients were collected before chemotherapy, and kept at 70 8C until the EIA procedure. Group 1 consisted of 112 patients (74 male, 38 female) with a mean age of 36.32  12.25 years, who were diagnosed as having active lung tuberculosis by positive smear and/or culture and clinical and radiologic findings. They were all anti-HIV negative. Group 2 consisted of 40 patients (29 male, 11 female) with a mean age of 59.47  13.36 years, with the diagnosis of inactive lung tuberculosis by radiologic findings and patient history. Three of the patients had a history of tuberculosis during the previous 2 years, and 26 of them had a history of tuberculosis before that period. Eleven of 40 patients had no tuberculosis history but had findings related to a previous infection in their chest X-rays. None of their sputum smears and cultures were positive.

ß 2001 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 7, 369–393

Concise Communications

Group 3 consisted of 21 patients (9 male, 12 female) with a mean age of 36.0  14.37 years, diagnosed as having miliary tuberculosis by radiologic findings and positive cultures of various specimens (four sputum, one bronchoalveolar lavage, two transbronchial, one larynx, one liver, three lymph node biopsies). All cases had scattered miliary densities on their chest X-rays. Two patients also had cavities in their radiographs. They were all anti-HIV negative. Group 4 consisted of 41 patients (17 female, 24 male) with a mean age of 34.18  12.45 years, with various lung diseases other than tuberculosis. Among these patients, 21 had lung tumors, while seven, four and four of them had bronchiectasis, chronic obstructive pulmonary disease/chronic bronchitis and pneumonia, respectively. The diagnoses of the rest of the cases were: bronchial asthma (one), sarcoidosis (one), idiopathic hemoptysis (one), lung fibrosis (one) and asbestosis (one). Group 5 consisted of 38 heathy individuals (19 male, 19 female) with a mean age of 36.86  14.43 years. The serum samples were kept at 70 8C until tested. Sera of all the patients and healthy individuals were searched for IgG and IgM antibodies against A60 by micro EIA (Eurospital spa, Trieste, Italy). IgA antibodies were only investigated in the sera of 135 patients and in all healthy individuals with the same method, according to the manufacturer’s instructions. Calibrators were used in the IgG and IgA tests, so that the results were quantitated. For the IgM test, the results were given as –, , þ. The cut-off points of EIA for each antibody were calculated from the results obtained with sera from the healthy group using the following formula: the mean value obtained from the healthy control group  2 standard deviation. The sensitivity, specificity and predictive values were estimated with the Epi Info Version 6 Program [13]. Statistical significance was determined by the chi-squared test and chisquared test with Fisher’s correction with the SPSS Program 6.0 [14]. The cut-off value for IgG EIA was calculated as 2 EU (EIA units). The numbers of positive cases for IgG in all groups are shown in Table 1. A comparison of the groups, according to the positivity rates of IgG, revealed a statistical difference (w2 ¼ 42.50, P ¼ 0.0000). Patients with active tuberculosis showed a significantly higher positivity rate of IgG than healthy individuals and the patients with lung diseases other

373

than tuberculosis (w2 ¼ 18.55, P ¼ 0.00002, and w2 ¼ 27.31, P ¼ 0.00000, respectively). At the same time, statistically important differences were observed between the active and inactive tuberculosis groups (w2 ¼ 9.68, P ¼ 0.00186). The sensitivity, specificity and positive predictive and negative predictive values for IgG are shown in Table 2. The number of cases positive for anti-A60 IgM in each group are shown in Table 1. A statistical difference was observed when the positivity rates of all the groups and groups 1 and 4 and 5 were compared (w2 ¼ 49.54, P ¼ 0.0000, w2 ¼ 15.32, P ¼ 0.00009, and w2 ¼ 21.59, P ¼ 0.00000, respectively). The number of positive cases in the active tuberculosis group was significantly higher than the number of positive cases in the inactive tuberculosis group (w2 ¼ 17.21, P ¼ 0.00003). The sensitivity, specificity and positive predictive and negative predictive values for IgG are shown in Table 2. A60 IgA was detected in 50 of 64 cases of active and 14 of nine cases of miliary tuberculosis (Table 1). A statistical difference was detected between the positivity rates of all groups (w2 ¼ 56.93, P ¼ 0.0000). There was also a statistically significant difference between groups 1 and 4 and 1 and 5 (w2 ¼ 35.21, P ¼ 0.00067; w2 ¼ 47.11, P ¼ 0.00000). The positivity rates of groups 1 and 2 were also compared for IgA, and a statistically important difference was observed (w2 ¼ 17.33, P ¼ 0.00003). The sensitivity, specificity and positive predictive and negative predictive values for IgG are shown in Table 2. There was no statistical difference between the antibody responses of the smear-negative and smear-positive cases of groups 1 and 3 (P > 0.05). There was also no statistically important difference between the culture-negative and culture-positive cases of group 3 (P > 0.05). The development of mycobacterium infections entails a strong interaction of components of mycobacteria and the immune system of the host [15]. These mycobacterial components and the antibodies against them are valuable for the diagnostic assays of tuberculosis. One of the best known antigens is A60, which is the major heat-stable component of tuberculin and protein purified derivative (PPD). The sensitivity and specificity for IgG against A60 in pulmonary tuberculosis cases were found to be 36–91% and 68– 98%, respectively [16–24]. In this study, we found the sensitivity and specificity values for IgG A60 to be 53.6% and 94.7%,

Table 1 The positivity rates for IgG, IgM and IgA for the groups Group number

n

IgG positive (%)

IgM positive (%)

n

IgA positive (%)

1 2 3 4 5

112 40 21 41 38

60 (53.6) 10 (25.0) 10 (47.6) 6 (14.6) 2 (5.3)

49 (43.8) 3 (7.5) 1 (4.7) 4 (9.8) 1 (2.6)

64 40 19 12 38

50 (78.1) 15 (37.5) 14 (73.7) 3 (25.0) 3 (7.9)

ß 2001 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 7, 369–393

71.4 (56.5, 83.0) 39.1 (20.5, 61.2) 64.8 (50.6, 77.0) 32.1 (16.6, 52.4) 56.8 (41.1, 71.3) 37.0 (27.7, 47.3) 37.0 (27.7, 47.3) 30.8 (22.9, 40.0) 30.8 (22.9, 40.0) 30.8 (22.9, 40.0) 40.9 (30.7, 51.9) 40.2 (30.0, 51.3) 36.4 (27.1, 46.7) 35.7 (26.5, 46.1) 32.3 (23.1, 42.9) 94.3 (83.4, 98.5) 94.3 (83.4, 98.5) 95.5 (86.6, 98.8) 95.5 (86.6, 98.8) 81.0 (70.3, 88.6) 98.0 (88.0, 99.9) 92.5 (80.9, 97.6) 98.0 (88.0, 99.9) 92.6 (81.3, 97.6) 94.3 (83.4, 98.5) 96.8 (87.8, 99.4) 90.9 (80.6, 96.3) 97.2 (89.4, 99.5) 92.1 (83.0, 96.7) 87.5 (77.8, 93.5) 92.1 (77.5, 97.9) 75.0 (42.8, 93.3) 92.1 (77.5, 97.9) 75.0 (42.8, 93.3) 62.5 (45.8, 76.8) 78.1 (65.7, 87.1) 78.1 (65.7, 87.1) 77.1 (66.3, 85.3) 77.1 (66.3, 85.3) 77.1 (66.3, 85.3) CI, confidence interval.

1 þ3 and 2

1 þ3 and 4

1 þ3 and 5

1and 4

43.8 (34.5, 53.4) 43.8 (34.5, 53.4) 37.6 (29.5, 46.4) 37.6 (29.5, 46.4) 37.6 (29.5, 46.4)

IgM

94.7 (80.9, 99.1) 85.4 (70.1, 93.9) 94.7 (80.9, 99.1) 85.4 (70.1, 93.9) 75.0 (58.5, 86.8) 53.6 (43.9, 63.0) 53.6 (43.9, 63.0) 52.6 (43.8, 61.3) 52.6 (43.8, 61.3) 52.6 (43.8, 61.3) 1and 5

IgM IgG IgG Groups

IgA

(CI) (CI)

97.4 (84.6, 99.9) 90.2 (75.9, 96.8) 97.4 (84.6, 99.9) 90.2 (75.9, 96.8) 92.5 (78.5, 98.0)

IgA IgM IgG IgG

IgM

IgA

(CI)

IgA

Positive predictive value (%)

(CI)

Specificity (%) Sensitivity (%)

Table 2 Sensitivity, specificity, positive predictive values and negative predictive values according to the groups considered for IgG, IgM and IgA EIA

Negative predictive value (%)

374 Clinical Microbiology and Infection, Volume 7 Number 7, July 2001

considering the active tuberculosis group and healthy individuals. These findings are in agreement with other results. Although the sensitivity rate is not as high as desired for a perfect diagnostic test, specificity values are very high. We found that 53.6% of active and 25.0% of inactive tuberculosis cases had IgG A60. These results are similar to or lower than those of other authors, who found the rate of positivity to be between 36% and 85% [16,22,25]. This may be because of the population chosen or the criteria that these authors used for diagnosis. According to our results, the IgG positivity rate was significantly higher than in healthy individuals and in patients with lung diseases other than tuberculosis. Our findings are consistent with the results of Alifano et al [17] and Turneer and Nerom [19]. In the present study, the sensitivity of A60 IgM antibody was found to be 43.8%. The sensitivity value for this antibody was previously reported to be between 24% and 80% [12,16,21,23,24]. This wide range may be explained by the time of diagnosis, because IgM can be detected at the early stages of tuberculosis and after a short period of time it decreases, while IgG antibodies increase [23]. In our study, the specificity of IgM was found to be 97.4%, which is a higher value than the sensitivity. A60 IgM EIA with high specificity could be used as a reliable test in the diagnosis of pulmonary tuberculosis when the result is positive. We found the positivity rates of IgA antibody to be 78.1%, 73.7% and 37.5% for the active, miliary and inactive tuberculosis cases, respectively. These values are in agreement with the results of Gupta et al [21]. Our sensitivity and specificity values according to the IgA test are 78.1% and 92.1%. To our knowledge, this is the first study to consider IgA serologic testing against A60 in various cases of tuberculosis in a Turkish population. Our results indicate that A60 IgA antibody seems to be a useful diagnostic test for active tuberculosis cases. The positivity rates of IgG, IgM and IgA for the miliary tuberculosis group were 47.6%, 4.7% and 73.7%, respectively. Very few data are currently available concerning the usefulness of IgA A60 antibody testing in miliary tuberculosis cases. Zou et al [16] and Gupta et al [21] searched for these antibodies in miliary tuberculosis cases and reported the sensitivities of IgG as 76% and 91.6%, those of IgM as 50% and 45%, and those of IgA as 88.4% and 54%. The specificity values were found to be over 90% for all antibodies. It can be concluded that IgG and IgA A60 antibodies should be considered for miliary tuberculosis cases which are difficult to diagnose and usually require invasive procedures. The positivity rates of the antibodies can be considered to be relatively low when compared to the culture and smear positivity rates in group 1 (100% and 91%, respectively). However, the detection of antibody response can be important, especially in smear-negative cases, since antibody response is not

ß 2001 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 7, 369–393

Concise Communications

statistically different between smear-negative and smear-positive cases. At the same time, the disadvantages of the classical techniques mentioned above should be considered. In our study, we found the rates of IgG, IgM and IgA A60 positivity to be 5.3%, 2.6% and 7.9% for the healthy individuals, and 14.6%, 9.8% and 25.0% for the patients with various lung diseases, respectively. The diagnosis of the patients who gave false-positive results were pneumonia and lung carcinoma in the group of non-tuberculous lung diseases. The rates of IgG A60 for these groups were found to be 0.38–23.8% and 6–55%, respectively, in previous studies [16,17,20,22,23]. Zou et al [16] reported two false-positive results for IgM and IgA antibodies; however, they mentioned one false-positive case in the group of healthy individuals. They also pointed out three instances of false positivity among 205 patients with lung diseases other than tuberculosis. Alifano et al [17] found similar false-positive results. These results may be due either to subclinical infection with environmental non-tuberculous mycobacteria that also express A60, or to the presence in the host of commensal nonpathogenic mycobacteria [26,27]. The dysregulation of the humoral immune response might be another cause of nontuberculous disease [26]. Dhand et al [28] reported that BCG and lung cancer cells shared common antigens, which could explain the false positivity observed for patients with lung cancer. We conclude that measurement of antibodies against A60, especially if molecular and new techniques for the diagnosis of tuberculosis are not available, may be considered a useful diagnostic tool in early active tuberculosis cases, as shown in a Turkish population. IgM antibody against A60 with high specificity could be used as a reliable test in the diagnosis of pulmonary tuberculosis when the result is positive. R EFER E NCE S 1. Pilheu JA. Tuberculosis 2000: problems and solutions. Int J Tuberc Lung Dis 1998; 2: 696–703. 2. Festenstein F, Grange JM. Tuberculosis and the acquired immune deficiency syndrome. J Appl Bacteriol 1991; 71: 19–30. 3. Erbakcy´ A, Du¨ ndar V, C¸etinkaya F, Kazgo¨ l N, Erem AR. Tu¨ berku¨ lozun serolojik tany´ sy´ nda antijen A600 a karþy´ antikorlary´n tany´ dekeri ve c¸aly´þmaya aly´nan popu¨lasyonun o¨nemi. Klimik Dergisi 1992; 5: 56–9. 4. Shinnick TM, King CH, Quinn FD. Molecular biology, virulence and pathogenicity of mycobacteria. Am J Med Sci 1995; 309: 92–8. 5. Koneman EW, Allen SD, Janda WM, Schreckenberger PC, Winn WC. Color Atlas and Textbook of Diagnostic Microbiology. Mycobacteria, 5th edn. Philadelphia: Lippincott-Raven, 1997: 893–952. 6. Daniel TM, Debanne SM. The serodiagnosis of tuberculosis and other mycobacterial diseases by enzyme-linked immunosorbent assay. Am Rev Respir Dis 1987; 135: 1137–51. 7. Kalish LR, Radin RC, Phair JP, Levitz D, Zriss R, Metzger E. Use of an enzyme-linked immunosorbent assay technique in the

8.

9.

10.

11.

12.

13.

14. 15. 16.

17.

18.

19.

20.

21.

22.

23.

24.

375

differential diagnosis of active pulmonary tuberculosis in humans. J Infect Dis 1983; 147: 523–30. Charpin D, Herbault H, Gevaudan MJ et al. Value of ELISA using A60 antigen in the diagnosis of active pulmonary tuberculosis. Am Rev Respir Dis 1990; 142: 380–4. Julian E, Matas L, Ausina V, Luquin M. Detection of lipoarabinomannan antibodies in patients with newly acquired tuberculosis and patients with relapse tuberculosis. J Clin Microbiol 1997; 35(10): 2663–4. Wilkinson RJ, Haslov K, Rappuoli R et al. Evaluation of the recombinant 38-kilodalton antigen of Mycobacterium tuberculosis as a potential immunodiagnostic reagent. J Clin Microbiol 1997; 35(3): 553–7. Daniel TM, Debanne SM, van der Kuyp F. Enzyme-linked immunosorbent assay using Mycobacterium tuberculosis antigen 5 and PPD for the serodiagnosis of tuberculosis. Chest 1985; 88(3): 388–92. Simonney N, Molina JM, Molimard M, Oksenhendler E, Lagrange H. Comparison of A60 and three glycolipid antigens in an ELISA test for tuberculosis. Clin Microbiol Infect 1996; 2(3): 214–22. Dean AG, Dean JA, Coulombier D et al. Epi Info, Version 6: a word processing, database, and statistics program for epidemiology on microcomputers. Atlanta, Georgia: Centers for Disease Control and Prevention, 1994. SSPS, Inc. SSPS for windows release 6.0. Chicago Ill: SPSS Inc., 1993. Chaparas SD. The immunology of mycobacterial infections. CRC Crit Rev Microbiol 1982; 9: 139–97. Zou YL, Zhang JD, Chen MH, Shi GQ, Prignot J, Cocito C. Serological analysis of pulmonary and extrapulmonary tuberculosis with enzyme-linked immunosorbent assays for anti-A60 immunoglobulins. Clin Infect Dis 1994; 19: 1084–91. Alifano M, De Pascalis R, Sofia M, Faraone S, Del Pezzo M, Covelli I. Detection of IgG and IgA against the mycobacterial antigen A60 with extrapulmonary tuberculosis. Thorax 1998; 53: 377–80. Caminero JA, Rodriguez de Castro F, Carillo T, Lafargo B, Diaz F, Cabrera P. Value of ELISA using A60 antigen in the serodiagnosis of tuberculosis. Respiration 1994; 61: 283–6. Turneer M, Nerom EV. Serological comparison of purified antigens 60 and 85A (P32) of Mycobacterium bovis BCG, and purified protein derivative in active pulmonary tuberculosis. Eur J Epidemiol 1993; 9: 541–6. Qadri SMH, Smith KK. Nonspecificity of the Anda A60-tb ELISA test for serodiagnosis of mycobacterial disease. Can J Microbiol 1992; 38: 804–6. Gupta S, Kumari S, Banwalikar JN, Gupta SK. Diagnostic utility of the estimation of mycobacterial antigen A60 specific immunoglobulins IgM, IgA and IgG in the sera of cases of adult human tuberculosis. Tubercle Lung Dis 1995; 76: 418–24. Baelden M, Vanderelst B, Dieng M, Priignot J, Cocito C. Serological analysis of human tuberculosis by an ELISA with mycobacterial antigen 60. Scand J Infect Dis 1990; 22: 63–73. Maes R. Clinical usefulness of serological measurements obtained by antigen 60 in mycobacterial infections: development of a new concept. Klin Wochenschr 1991; 69: 696–709. Li LF, Lin MC, Chen NH, Hsieh MJ, Lee CH, Tsao TC. Serodiagnosis of tuberculosis by enzyme-linked immunosorbent assay for anti-A60 and anti-A38. Chang Keng I Hsueh Tsa Chih 1998; 21(3): 258–64.

ß 2001 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 7, 369–393

376 Clinical Microbiology and Infection, Volume 7 Number 7, July 2001

25. Landron de Guevara MC, Gonzalez A, Ortega A, Saz JV. Serological diagnosis of pulmonary tuberculosis using ELISA and the A60 antigen. Enferm Infect Microbiol Clin 1992; 10: 17–19. 26. Cocito C. Properties of the mycobacterial antigen complex A60 and its applications to the diagnosis and prognosis of tuberculosis. Chest 1991; 100: 1687–93.

27. Alifano M, Sofia M, Mormile M et al. IgA immune response against the mycobacterial antigen A60 in patients with active pulmonary tuberculosis. Respiration 1996; 63: 292–7. 28. Dhand R, Ganguly NK, Vaishnavi C, Gilhotra R, Maly´ k SK. False-positive reactions with enzyme-linked immunosorbent assay of Mycobacterium tuberculosis antigens in pleural fluid. J Med Microbiol 1988; 26: 241–3.

Effect of step-down therapy of ceftriaxone plus loracarbef versus parenteral therapy of ceftriaxone on the intestinal microflora in patients with community-acquired pneumonia F. Vogel1, H. R. Ochs2, K. Wettich3, S. Kalich3, I. Nilsson-Ehle4, I. Odenholt5 and C. E. Nord6 1

Medizinische Klinik III, Krankenhaus Hofheim, Hofheim, 2Innere Abteilung, Marienkrankenhaus Soest, Soest, 3Lilly Deutschland GmbH, Bad Homburg, Germany, 4Department of Infectious Diseases, University Hospital, Lund, 5 Department of Infectious Diseases, University Hospital, Malmo¨ and 6Department of Microbiology, Pathology and Immunology, F82, Huddinge University Hospital, SE-141 86 Stockholm, Sweden 

Tel: þ46 8 58587838

Fax: þ46 8 7113918

E-mail: [email protected]

Accepted 27 March 2001

Many patients with community-acquired pneumonia can be treated on an outpatient basis, but a significant proportion of patients require hospitalization for initiation of parenteral antibiotic therapy [1]. Patients with pneumonia may be converted from parenteral antibiotic treatment to oral treatment after 3–5 days. Although outpatient parenteral treatment has been shown to be effective in the therapy of community-acquired pneumonia, early conversion to oral antibiotic treatment may result in significantly fewer adverse events and reduction of major costs. With the availability of newer oral b-lactam agents with favorable pharmacokinetic and pharmacodynamic properties, the use of step-down therapy for lower respiratory tract infections has increased in clinical medicine [2]. For most cases of community-acquired pneumonia requiring hospitalization, the recommended regimens consist of a cephalosporin such as cefuroxime, cefotaxime or ceftriaxone, or a penicillin plus a b-lactamase inhibitor with or without erythromycin [1]. The most often isolated pathogenic bacteria are Streptococcus pneumoniae, Haemophilus influenzae and Moraxella catarrhalis. The recommended agents are active against these pathogens. Loracarbef, a synthetic b-lactam agent of the carbacephem class for oral administration, has favorable pharmacokinetic, pharmacodynamic and microbiological properties, which suggests that this agent may be useful for early conversion from parenteral treatment to oral treatment of community-acquired pneumonia [3].

The most common and significant cause of disturbances in the normal intestinal microflora is the administration of antimicrobial agents [4]. When the number of microorganisms is reduced during therapy, resistance to colonization is decreased, which may lead to several unwanted effects. One is overgrowth of already present microorganisms with natural resistance, such as yeasts, which may cause systemic infections in immunocompromised patients, and Clostridium difficile, which may lead to diarrhea and/or colitis. A second consequence is the establishment of new resistant pathogenic bacteria, which may also colonize other areas of the host. A third effect is the fact that bacterial overgrowth also encourages the transfer of resistance factors among bacteria. Several factors influence the extent to which a given antimicrobial agent will reduce the normal microflora. Predominant among these factors is the incomplete absorption of agents perorally administered. Poorly absorbed drugs can reach the colon in active form, where they suppress susceptible microorganisms and disturb the ecological balance. Antimicrobial drugs that are secreted in the bile or from the intestinal mucosa also tend to affect the normal intestinal microflora. Loracarbef is approximately 90% absorbed from the gastrointestinal tract after oral administration, and only a minor impact on the intestinal microflora has been reported [5,6]. In contrast, parenteral ceftriaxone is excreted in the bile in high concentrations, resulting in major changes of the intestinal microflora [7–9].

ß 2001 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 7, 369–393

Concise Communications

The aim of the present investigation was to compare the ecological effect of step-down therapy of ceftriaxone plus loracarbef versus parenteral therapy of ceftriaxone on the intestinal microflora in patients with community-acquired pneumonia. Twenty-four newly hospitalized patients (16 males and eight females; mean age 62 years; age range 32–84 years) with community-acquired pneumonia participated in this prospective, randomized, multicenter study in Germany (two hospitals) and Sweden (two hospitals). Five patients in Sweden received ceftriaxone plus loracarbef and five patients ceftriaxone. In Germany, seven patients were treated with ceftriaxone plus loracarbef and seven patients with ceftriaxone. Patients on antibiotic treatment within 6 days before enrollment were excluded. The inclusion criteria were age >18 years, informed consent and disease diagnostic criteria. The patient should have both presence of a radiographic pulmonary infiltrate and acute onset of at least one clinical finding: (1) cough, (2) dyspnea, (3) pleuritic chest pain, (4) sputum production, (5) fever >38 8C, and (6) altered mental status. Clinical examinations were made pre-therapy, during therapy (days 2 and 3), post-therapy (within 72 h of therapy completion), late post-therapy (10–14 days after end of therapy), and very late post-therapy (21–28 days after end of therapy). Sputum and blood cultures were obtained at the same visits from all patients when indicated. Routine laboratory tests were taken for hematology and blood chemistry. All adverse events were reported to the investigators. Informed consent was obtained from all patients, and the study was approved by the local Ethics Committee. Twelve patients received step-down therapy with ceftriaxone (Roche, Basel, Switzerland) 1 g intravenously every 24 h for 2 days, followed by oral loracarbef (Eli Lilly, Indianapolis, USA) 400 mg every 12 h for 8 days. Twelve patients were given only ceftriaxone 1 g intravenously every 24 h for 10 days. Stool specimens were collected prior to antibiotic treatment (day 0), during treatment (day 3), at the post-therapy visit (day 12) and at the late post-therapy visit (day 24). The specimens were collected in sterile plastic containers. The specimens were frozen within 1 h and stored at 70 8C until assayed [10]. The microbiological analysis of the specimens was performed as described by Edlund et al [10]. The stool specimens (1 g of feces) were suspended in pre-reduced peptone–yeast extract medium, diluted to 107 and inoculated on non-selective and selective media. The following agar media were used: blood agar (LabM Kemila, Bury, UK) for total aerobes and anaerobes, CLED agar (Merck, Darmstadt, Germany) for detection of Enterobacteriaceae, Enterococcosel agar (BBL, Cockeysville, MD, USA) for detection of enterococci, Sabouraud agar (Difco, Detroit, MI, USA) for detection of yeasts, Rogosa agar (Difco) for cultivation of lactobacilli, BL agar (Bifidobacterium selective agar; Difco) for cultivation of bifidobacteria, BKV (kanamycin– vancomycin–blood agar) for cultivation of Bacteroides and

377

Prevotella spp., BNV agar (neomycin–vancomycin–blood agar) for cultivation of fusobacteria, Veillonella agar (Difco) for cultivation of veillonella cocci, egg yolk agar (Oxoid, Basingstoke, UK) for cultivation of clostridia, and TCCFA agar (Cycloserine Cefoxitin Fructose agar; Peptone from Casein/Proteose PeptonIII 40 mg/mL, sodium hydrogen phosphate 5 mg/mL, potassium dihydrogenphosphate 1 mg/mL, sodium chloride 2 mg/mL, sodium sulfate 0.2 mg/mL, Bacto Agar/Agar-Agar 20 mg/mL, taurochloric acid 1 mg/mL, neutral red 0.03 mg/ mL, 15% fructose, Clostridium difficile supplement D-cycloserine, cefoxitin) for detection of Clostridium difficile. The aerobic agar plates were incubated for 48 h at 37 8C and the anaerobic plates for 48 h at 37 8C in anaerobic jars (GasPak; BBL). After incubation, different colony types were counted, isolated in pure culture and identified to genus level. All isolates were analyzed according to Gram reaction and cell and colony morphology, followed by different biochemical tests. API 20E test kit (BioMe´ rieux SA, Marcy l’Etoile, France) was used for the identification of Enterobacteriaceae. The anaerobic microorganisms were identified by gas–liquid chromatography of metabolites from glucose. The lower limit of detection was 102 microorganisms/g feces. The 12 patients treated with ceftriaxone had the following respiratory findings: cough 10 patients, sputum production nine patients, dyspnea eight patients, tachypnea nine patients, and pleuritic chest pain six patients. The 12 patients treated with ceftriaxone–loracarbef showed the following findings: cough 12 patients, sputum production seven patients, dyspnea eight patients, tachypnea nine patients, and pleuritic chest pain six patients. There were no differences in the respiratory diagnosis criteria between the two treatment groups, and no differences in clinical efficacy between the two groups. The patients receiving ceftriaxone had the following bacterial strains isolated from sputum on admission: Streptococcus pneumoniae two strains, H. influenzae four strains, H. parainfluenzae four strains, Staphylococcus aureus two strains, and Staphylococcus epidermidis one strain. Three patients in the ceftriaxone group had two bacterial isolates (Streptococcus pneumoniae and H. parainfluenzae; H. influenzae and Staphylococcus epidermidis; H. parainfluenzae and Staphylococcus aureus). In two sputum samples at the first visit, no microorganisms were isolated. In the ceftriaxone–loracarbef group, Streptococcus pneumoniae (one strain), H. influenzae (two strains), H. parainfluenzae (six strains), Moraxella catarrhalis (one strain), Staphylococcus aureus (one strain) and Staphylococcus epidermidis (two strains) were recovered from the sputum samples at the first visit. In the ceftriaxone–loracarbef group, three patients had two bacterial isolates (H. parainfluenzae and Staphylococcus aureus; H. parainfluenzae and Staphylococcus epidermidis; H. parainfluenzae and Staphylococcus epidermidis). No microorganisms were found in five samples, which may be due to previous antibiotic treatment or sampling failure. Streptococcus pneumoniae was

ß 2001 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 7, 369–393

378 Clinical Microbiology and Infection, Volume 7 Number 7, July 2001

isolated from blood in one patient treated with ceftriaxone and in one patient treated with ceftriaxone plus loracarbef. Seventeen blood cultures in 24 patients were negative at the first visit. All isolated bacterial strains were considered to be pathogenic for the patients’ pneumonia. Patients with Chlamydia pneumoniae, Mycoplasma pneumoniae and Legionella pneumophila were not included in the study. Figure 1 shows the impact of ceftriaxone on the aerobic intestinal microflora. The numbers of Escherichia coli decreased markedly, while the numbers of enterococci and Candida

increased during the therapy and post-therapy visits. The microflora was normalized in most patients at the very late post-therapy visit. The anaerobic intestinal microflora was also affected (Figure 2). Lactobacilli, bifidobacteria and clostridia but not Bacteroides strains were suppressed during and after ceftriaxone treatment but returned to normal levels after 24 days. The aerobic intestinal microflora was affected in a similar way by the step-down treatment with ceftriaxone and loracarbef (Figure 1). Figure 2 presents the effect on the anaerobic intest-

Figure 1 Impact of ceftriaxone ( ) and ceftriaxone plus loracarbef ( ) treatment on the intestinal aerobic microflora in 24 patients. Solid line, mean value of the logarithmic number of microorganisms/g feces in patients receiving ceftriaxone. Dashed line, mean value of the logarithmic number of microorganisms/g feces in patients receiving ceftriaxone plus loracarbef.

Figure 2 Impact of ceftriaxone ( ) and ceftriaxone plus loracarbef ( ) treatment on the intestinal anaerobic microflora in 24 patients. Solid line, mean value of the logarithmic number of microorganisms/g feces in patients receiving ceftriaxone. Dashed line, mean value of the logarithmic number of microorganisms/g feces in patients receiving ceftriaxone plus loracarbef.

ß 2001 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 7, 369–393

Concise Communications

inal microflora. Lactobacilli, bifidobacteria and clostridia were suppressed significantly, while Bacteroides strains were not markedly affected by the ceftriaxone–loracarbef treatment. Three patients had Clostridium difficile strains during treatment. The intestinal microflora was normalized after 24 days. When comparing the two treatment groups, no differences were observed during the first 2 days of treatment. No statistical differences in the effect on the intestinal microflora between the two treatment regimens were observed. Loracarbef is a b-lactam antibiotic for oral administration. It is 90% absorbed from the gastrointestinal tract after oral administration with or without food. As a result of the absorption, minor changes in the intestinal microflora after loracarbef administration have been observed [5]. The stool concentrations of loracarbef are also low, 0–0.9 mg/kg after administration of loracarbef 200 mg twice daily for 7 days [5]. In contrast, ceftriaxone is excreted in high concentrations in the bile (500 mg/L), and pronounced effects on the aerobic and anaerobic microflora have been reported [7,11]. Ceftriaxone is used in the outpatient treatment of community-acquired pneumonia in children and adults with clinical success [12,13]. However, in order to reduce the ecological impact of ceftriaxone on the intestinal microflora and the duration of hospitalization, step-down therapy with an oral b-lactam agent may be preferable. In the present trial, loracarbef was used as the oral agent. There was no difference in the clinical efficacy between the two 12-patient treatment groups, indicating that step-down therapy of ceftriaxone plus loracarbef for treatment of community-acquired pneumonia may be acceptable. However, it should be mentioned that only a small number of patients were participating in the trial. In conclusion, step-down therapy with ceftriaxone plus loracarbef caused similar ecological alterations in the intestinal microflora to those of parenteral ceftriaxone therapy in patients with community-acquired pneumonia.

379

R EFER E NCE S 1. Brown RB. Community-acquired pneumonia: diagnosis and therapy of older adults. Geriatrics 1993; 48: 43–50. 2. Hitt CM, Nightingale CH, Quintiliani R, Nicolau DP. Streamlining antimicrobial therapy for lower respiratory tract infections. Clin Infect Dis 1997; 24: S231–7. 3. Hyslop DL, Jacobson K, Guerra FJ. Loracarbef (LY163892) versus amoxicillin/clavulanate in bronchopneumonia and lobar pneumonia. Clin Ther 1992; 14: 254–67. 4. Nord CE, Edlund C. Impact of antimicrobial agents on human intestinal microflora. J Chemother 1990; 2: 218–37. 5. Nord CE, Grahne´ n A, Eckerna¨s S. Effect of loracarbef on the normal oropharyngeal and intestinal microflora. Scand J Infect Dis 1991; 23: 255–60. 6. Kato N, Kato H, Tanaka Y, Tanaka K, Watanabe K, Ueno K. Impact of loracarbef, a new oral carbacephem, on human fecal flora. Chemotherapy 1993; 41: 152–9. 7. Arvidsson A, Leijd B, Nord CE, Angelin B. Interindividual variability in biliary excretion of ceftriaxone: effects on biliary lipid metabolism and on intestinal microflora. Eur J Clin Invest 1988; 18: 261–6. 8. Lejko-Zupanc T, Zakelj J, Strle F, Janc M, Pleterski-Rigler D. Influence of ceftriaxone on emergence of Clostridium difficile. Antimicrob Agents Chemother 1992; 36: 2850–1. 9. Welling GW, Meijer-Severs GJ, Helmus G et al. The effect of ceftriaxone on the anaerobic bacterial flora and the bacterial enzymatic activity in the intestinal tract. Infection 1991; 19: 313–16. 10. Edlund C, Beyer G, Hiemer-Bau M, Ziege S, Lode H, Nord CE. Comparative effects of moxifloxacin and clarithromycin on the normal intestinal microflora. Scand J Infect Dis 2000; 32: 81–5. 11. Nilsson-Ehle I, Nord CE, Ursing B. Ceftriaxone. pharmacokinetics and effect on the intestinal microflora in patients with acute bacterial infections. Scand J Infect Dis 1985; 17: 77–82. 12. Karachalios GN, Georgiopoulos AN, Kanatakis S. Treatment of various infections in an outpatient practice by intramuscular ceftriaxone: home parenteral therapy. Chemotherapy 1989; 35: 389–92. 13. Leibovitz E, Tabachnik E, Fliedel O et al. Once-daily intramuscular ceftriaxone in the outpatient treatment of severe community-acquired pneumonia in children. Clin Pediatr 1990; 29: 634–9.

Adherence to polystyrene of clinically relevant isolates of Candida species A. Sa´nchez-Sousa, D. Tarrago´, J. Velasco, M. E. Alvarez and F. Baquero Department of Microbiology and Parasitology, Mycology Unit, Ramo´ n y Cajal Hospital, 28034 Madrid, Spain 

Tel: þ34 1 336 83 30

Fax: þ 34 1 336 88 09

E-mail: [email protected]

Accepted 26 March 2001

Infection of implanted plastic devices is one of the most serious complications in hospitalized patients [1]. The genus Candida comprises important nosocomial pathogens involved in biofilm

production on catheter and prosthetic materials. The ability of many strains of the genus Candida to adhere to inert polymeric surfaces may allow these microorganisms direct access into the

ß 2001 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 7, 369–393

380 Clinical Microbiology and Infection, Volume 7 Number 7, July 2001

human host. Yeast adherence appears to be an important step in the pathogenesis of superficial and systemic candidosis [2]. Candida albicans, C. tropicalis, C. parapsilosis, C. glabrata and C. lusitaniae have clearly been associated with prolonged catheterization [3,4]. In addition, the clinical relevance of Candida species other than C. albicans is increasing in compromised patients [5–9]. The presence of an indwelling catheter contributes to the local enrichment of Candida cells adhering to plastic material. A catheter-mediated selection occurs with adherent yeast-phase cells, and both microbial (slime) and host (fibrin and fibronectin) factors contribute to the formation of the biofilm, essential for adherence and maintenance of colonization and infection [10]. Therefore, a relationship between adherence in vitro, biofilm production and Candida invasiveness may be suspected. In this study, intraspecific differences in adherence, mostly related to more invasive strains, were noted. The SYDT method was demonstrated to be an inexpensive, easy and accurate tool for quantifying adherence to plastic in vitro. This fact would help in identifying some isolates producing a biofilm in vivo. A connection between adherence in vitro, the ability to produce a biofilm in vivo and catheter-related infection is suggested here. In total, 103 yeast clinical isolates belonging to the genus Candida were studied. Eighty-two of them were Candida species other than C. albicans: C. parapsilosis (37), C. tropicalis (17), C. glabrata (20), and C. lusitaniae (8). All of them were significant isolates of these species obtained during the last 2 years in the Mycology Unit at Ramo´ n y Cajal Hospital (Madrid). Within the C. parapsilosis group, two subsets were analyzed: 20 isolates recovered from deep-seated locations in bone marrow transplant recipients (15 strains from blood cultures, three from indwelling catheters, one from an implanted mitral valve, and one from cerebrospinal fluid), and 17 strains isolated from epidermal or mucosal sites. Seven isolates of C. tropicalis were obtained from significant (>100 000 CFU/mL) urine specimens, six from deep respiratory tract secretions, two from vaginal secretions, and two from wound infections. Nine C. glabrata isolates were obtained from vaginal infections, six from significant urine specimens, four from deep respiratory tract secretions, and one from a blood culture. One isolate of C. lusitaniae was obtained from an indwelling catheter, and seven isolates from vaginal and significant urine specimens. All C. albicans isolates were isolated from vaginal and significant urine specimens. Adherence was evaluated by measuring the dye from stained Candida organisms attached to sterile polystyrene tubes. Isolates were cultured in Sabouraud Agar plates for 48 h at 30 8C; then, three colonies were gently suspended in test tubes containing Sabouraud Broth (Institute Pasteur Production, Marnes la-Coquette, France) with 8% glucose. The inoculum size was estimated in about 1.5  107 CFU/mL by immediate plate counting. Tubes (3  0.5 inches) were incubated at 35 8C with

mild shaking (180 rev/min) for 24 h. Then, the Sabouraud broth was completely removed, and tube walls were gently washed (overflow, without shaking) with deionized water and stained with 2 mL of safranine (4 g/L) alcohol solution (20%). After 30 s of staining, the safranine solution was removed. Tubes were allowed to dry for about 2 h in a closed chamber with air extraction. Direct visual detection (DVD) of stained tubes was obvious at this stage (three independent observers). A quantitative measurement of adherence was performed by direct spectrophotometric reading at 560 nm (Bausch and Long Spectronic 20) of stained yeasts in dried tubes (SYDT); for every tube, the mean optical density (OD) from three readings corresponding to 1208 rotating positions was calculated. Alternatively, before tubes became dry, 3 mL of deionized water was added, followed by intensive vortexing for enough time to allow the stained yeast to be completely incorporated in the fluid. Subsequently, the stained yeast suspension was spectrophotometrically read at 610 nm (SSYS). The test was considered to be positive when the OD ¼ 0.1 by each procedure. The correlation between DVD and SYDT was excellent: of 50 DVD-positive strains, 45 showed an OD  0.1, and of 53 DVD-negative strains, 51 showed an OD < 0.1. A good correlation between DVD-positive strains and SSYS-positive strains was also found. Of 50 DVD-positive strains, 49 showed an OD  0.1. In contrast, of 53 DVD-negative strains, only 16 showed an OD < 0.1. When an OD cut-off of 0.2 was used by SSYS, 21 DVD-positive strains of 50 were correctly classified. Given the better correlation of DVD and SYDT, a positive adherence test was considered when an OD  0.1 by the SYDT method was obtained, which corresponded to 45.6% of all studied strains (47/103). The detailed data of adherence for the different Candida species by SYDT are shown in Table 1. The spectrophotometric assays used in this study were useful for quantifying and comparing adherence to plastic material of different Candida species and revealed significant differences within the genus. A low frequency of adherent C. glabrata isolates was found (10%; 2/20), slightly higher than for C. albicans (4.8%; 1/21). A high degree of adherence was detected in all C. tropicalis isolates tested (100%; 17/17), and it was the species that most consistently produced adherence under our experimental conditions. A better adherence of some strains of C. tropicalis compared with strains of C. albicans has previously been described [11]. C. tropicalis is more hydrophobic than C. albicans in water/hexadecane partition tests, and the better adherence to plastic surfaces can be related to this fact. A large number of C. tropicalis isolates (76.5%; 13/17) showed an OD  0.3, which represents the highest proportion of isolates with a high adherence level observed with any Candida species tested. Most C. lusitaniae isolates showed a positive test (87.5%; 7/8). Surprisingly, one C. lusitaniae isolate had an OD greater than 0.8, suggesting that some isolates may be particularly adherent. The ability of C. lusitaniae to produce biofilms on

ß 2001 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 7, 369–393

Concise Communications

381

Table 1 Number of isolates for the different Candida species OD

C. albicans

C. parasilopsis

C. glabrata

C. tropicalis

C. lusitaniae

Total

< 0.1 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 Total

20 0 0 1 0 0 0 0 0 0 21

(7) þ 10 (6) þ 4 (4) þ 3 (1) 0 (1) (1) 0 0 0 (20) þ 17

(3) þ 15 0 0 (1) 0 0 0 0 (1) 0 (5) þ 15

0 3 1 1 2 2 1 1 (5) (1) (6) þ 11

1 5 1 0 0 0 0 0 0 (1) (1) þ 7

(10) þ 46 (6) þ 12 (4) þ 5 (2) þ 2 2 (1) þ 2 (1) þ 1 1 (6) (2) (32) þ 71

An optical density value (OD) ¼ 0.1was a positive result by the SYDTadherence test. Bold numbers in parentheses represent numbers of invasive strains.

polystyrene surfaces has never been explored before. The adherence of this species to plastic may be a frequent trait. This finding might account for the frequent involvement of C. lusitaniae in infected catheters. Most C. parapsilosis isolates adhered to plastic (54%; 20/37). These results obtained with C. parapsilosis isolates, mostly isolated from bone marrow transplant recipients, are consistent with those reported by Branchini et al [12], who found a positive adherence test (estimated by visual reading) in 80% of the studied isolates. In our study, if only strains from bone marrow transplant recipients were considered, a higher value would have been obtained (65%; 13/20). Depending on the recovery site, a significant difference in adherence was found (Table 1), thus indicating intraspecific variability in all studied species. Interestingly, a higher frequency of invasive strains (68.7%; 22/32) than those recovered from epidermal sites (35.2%; 25/71) produced a positive adherence test. In addition, within each Candida species, the highest OD values were obtained with the most invasive strains: one C. lusitaniae from blood culture; three C. parasilopsis from indwelling catheters, one from cerebrospinal fluid, and three from blood culture; one C. glabrata from blood culture; and five C. tropicalis from deep respiratory tract secretions. Many studies have demonstrated the importance of the basal yeast layer in anchoring the biofilm to the plastic surface [2,13]. A positive correlation between adherence of these organisms to host epithelial surfaces and adherence to plastic surfaces has previously been described [14]. In addition, a lower susceptibility of Candida in biofilms to antifungal agents has previously been reported [10]. All these findings suggest a connection between the degree of colonization, adherence to polystyrene, biofilm production in vivo, susceptibility to antifungal agents, invasiveness and pathogenicity of yeast isolates involved in infected indwelling catheters and other implanted plastic devices. Although one of the strengths of our data is that the study includes an extremely large number of strains and species

of clinical relevance, further studies are obviously warranted to establish a direct connection between adherence and pathogenicity. ACK NOW L EDGME N T S We are grateful to Dr L. De Rafael from the Microbiology Department, Ramo´ n y Cajal Hospital, Madrid for his kind help in preparing this manuscript.

R EFER E NCE S 1. Raad I, Darouiche R, Hachem R, Sacilowski M, Bodey GP. Antibiotics and prevention of microbial colonization of catheters. Antimicrob Agents Chemother 1995; 39: 2397–400. 2. Baillie GS, Douglas LJ. Role of dimorphism in the development of Candida albicans biofilms. J Med Microbiol 1999; 48: 671–9. 3. Komshian SV, Uwaydah AK, Sober JD, Crane LR. Fungemia caused by Candida species and Torulopsis glabrata in the hospitalized patient: frequency, characteristics, and evaluation of factors influencing outcome. Rev Infect Dis 1989; 11: 379–90. 4. Lowder JN, Lazarus HM, Herzig RH. Bacteriemias and fungemias in oncologic patients with central venous catheters: changing spectrum of infection. Arch Intern Med 1982; 142: 1456–9. 5. Lecciones JA, Lee JW, Navarro EE et al. Vascular catheterassociated fungemia in a patient with cancer: analysis of 155 episodes. Clin Infect Dis 1992; 14: 875–83. 6. Girmenia C, Martino P, De Bernardis F et al. Rising incidence of Candida parapsilosis fungemia in patients with hematologic malignancies: clinical aspects, predisposing factors, and differential pathogenicity of the causative strains. Clin Infect Dis 1996; 23: 506–14. 7. Levin AS, Costa SF, Mussi NS et al. Candida parapsilosis fungemia associated with implantable and semi-implantable central venous catheters and the hands of healthcare workers. Diagn Microbiol Infect Dis 1998; 30: 243–9. 8. Weems JJ. Candida parapsilosis: epidemiology, pathogenicity, clinical manifestations and antimicrobial susceptibility. Clin Infect Dis 1992; 14: 756–66.

ß 2001 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 7, 369–393

382 Clinical Microbiology and Infection, Volume 7 Number 7, July 2001

9. Baker JG, Nadler HL, Forgacs P, Kurtz SR. Candida lusitaniae: a new opportunistic pathogen of the urinary tract. Diagn Microbiol Infect Dis 1984; 2: 145–50. 10. Baillie GS, Douglas LJ. Candida biofilms and their susceptibility to antifungal agents. Methods Enzymol 1999; 310: 644–56. 11. Rotrosen D, Calderone RA, Edwards JE. Adherence of Candida species to host tissues and plastic surfaces. Rev Infect Dis 1986; 8: 73–85. 12. Branchini MI, Pfaller MA, Rhine-Chalberg J, Frempong T, Isenberg HD. Genotypic variation and slime production among

blood and catheter isolates of Candida parapsilosis. J Clin Microbiol 1994; 32: 452–6. 13. San-Millan R, Ezkurra PA, Quindos G, Robert R, Senet JM, Ponton J. Effect of monoclonal antibodies directed against Candida albicans cell wall antigens on the adhesion of the fungus to polystyrene. Microbiology 1996; 142: 2271–7. 14. Samaranayake YH, Wu PC, Samaranayake LP, So M, Yuen KY. Adhesion and colonisation of Candida krusei on host surfaces. J Med Microbiol 1994; 41: 250–8.

Brucella infective endocarditis: a report of four successfully treated patients H. Alsoub Hamad Medical Corporation, Department of Medicine, Infectious Diseases Section, PO Box 3050, Doha, Qatar Fax: þ974 4392273

E-mail: [email protected]

Accepted 23 March 2001

Brucellosis is a zoonosis with a worldwide distribution, especially in the Mediterranean basin, the Arabian peninsula, the Indian subcontinent, Mexico and Central and South America [1]. It is a systemic disease, and almost every organ can be affected. The infection usually manifests itself as a febrile syndrome with no apparent focus, chills, sweating, arthralgia, and myalgia. About 30% of patients suffer from some localization, most commonly bone and joint involvement [2]. Cardiovascular complications are rare, occurring in less than 2% of patients with brucellosis; they include endocarditis, myocarditis, and pericarditis [3]. Brucella endocarditis accounts for the majority of deaths related to brucellosis [4]. The treatment for brucella endocarditis has not been well established, basically because of the low number of reported cases. We herein report our experience with four cases of brucella endocarditis, discuss treatment options, and review the pertinent literature. Between 1987 and 1999, four patients were diagnosed as having brucella endocarditis at Hamad Medical Corporation. Diagnosis was based on the presence of vegetations on echocardiography, and further supported by positive blood culture, culture of infected tissue growing Brucella organisms, or a high brucella serologic titer of over 1 : 320 dilution. The following data were collected: age, sex, nationality, occupation, raw milk ingestion, symptoms, physical findings at presentation, total white blood cell count, erythrocyte sedimentation rate, serology, echocardiography, blood culture, culture of infected tissue, treatment, complications, and outcome. The age of the patients ranged between 29 and 40 years. All patients complained of fever. Other symptoms included

anorexia, cough, and night sweats. One patient had hemiparesis. The duration of symptoms ranged between 2 weeks and 9 months. All patients had a history of raw milk ingestion. Three patients had a history of valvular heart disease, and had mechanical valvular prostheses. Physical findings on admission included fever in all patients, and splenomegaly and right-sided hemiparesis in one patient each. None of the patients had evidence of heart failure on presentation or during follow-up. All patients had normal white blood cell counts. The erythrocyte sedimentation rate was elevated in all patients (range 33–96 mm/h). Blood culture grew Brucella organisms in three patients. The brucella antibody titer was elevated in three patients (range 1280–5120). In one patient, brucella serology was negative in spite of repeated testing; however, the blood culture was positive in this patient. Transesophageal echocardiography showed vegetations in all patients. It also showed paravalvular leak in two patients and an abscess in one patient. One patient with native valve brucella endocarditis was treated with antibiotics alone. The other three with prosthetic valve brucella endocarditis were treated with antibiotics and valve replacement. Details of antibiotic therapy are shown in Table 1. The indications for valve replacement were failure of medical treatment in one patient (number 3; vegetations developed while he was receiving antibiotics), and brucella prosthetic valve endocarditis in the other two. The duration of antibiotic treatment was between 10 weeks and 20 weeks. Surgery on the three patients was carried out 4 days, 10 days, and 11 weeks, respectively, after starting antibiotic treatment.

ß 2001 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 7, 369–393

Co-trimoxazole/doxycycline/rifampin Doxycycline/rifampin 10 weeks 10 months Cured M, male; ND, not done.

Rifampin/doxycycline/co-trimoxazole Rifampin/doxycycline/co-trimoxazole 20 weeks 10 months Cured Streptomycin/doxycycline Rifampin/doxycycline 12 weeks 15 months Cured

Doxycycline/streptomycin, rifampin/doxycycline Rifampin/doxycycline/co-trimoxazole 13 weeks 12 months Cured

35/M Raw milk ingestion 3 months Fever, cough Yes Mitral valve prosthesis Negative Negative 1:1280 39/M Raw milk ingestion 2 weeks Fever, night sweats, cough Yes Aortic valve prosthesis Brucella species Negative Negative 29/M Raw milk ingestion 9 months Fever, anorexia, right hemiparesis Yes Aortic valve prosthesis Brucella melitensis Positive 1: 5120 40/M Raw milk ingestion 2 months Fever, palpitation No Native aortic valve Brucella melitensis ND 1: 2560

Age (years)/Sex Risk factor Duration of fever Clinical features Known heart disease Valve involved Blood culture Culture of valve Brucella titer Treatment Acute phase Postoperative Duration of treatment Duration of follow-up Outcome

2 1

Patient no.

Table 1 Clinical findings, treatment and outcome of patients with brucella endocarditis

3

4

Concise Communications

383

All patients were followed up in the clinic. The duration of follow-up ranged between 10 months and 15 months. During follow-up, no recurrences were observed. Endocarditis is a rare complication of brucellosis, occurring in less than 2% of cases [3]. In a recent review of the English and French language literature since 1900, only 108 cases were reported [5]. Brucellosis is endemic in Qatar; however, despite this, only four cases of brucella endocarditis have been diagnosed during the last 15 years. Brucella melitensis and B. abortus are the most frequently isolated species. Brucella endocarditis produces highly destructive lesions of the valve structure [3]; this might explain the high fatality rate for brucella endocarditis. Although the mortality rate for brucellosis is less than 1%, endocarditis accounts for 80% of these deaths [4]. The complication that is responsible for the majority of deaths is heart failure [6]. Although published reports indicate that brucella endocarditis usually involves normal native valves, and predominantly the aortic valve in 75% of cases [3], it is interesting that three of our patients had prosthetic valve brucella endocarditis, and only one had native valve endocarditis. Brucella prosthetic valve endocarditis is a very rare disease, with only 11 cases being reported in the English literature [7–14]. Three of our patients were males, in keeping with other reports [5]. The clinical features in our patients were similar in many respects to those reported by others [5,15]. Although heart failure is reported to be common in these patients, being either present initially or developing later on during the course of illness [5,7,15,16], none of our patients had this complication. One of our patients had hemiparesis, most probably secondary to an embolus originating from valve vegetation. Diagnosis of brucella endocarditis may be difficult. Routine laboratory tests are not helpful. Diagnosis depends on isolation of Brucella from blood culture or cardiac tissue [15]. Blood cultures were positive in three of our patients; B. melitensis grew in two, and in the third the species was not specified. Culture of cardiac tissue taken during surgery may be positive, despite prior antibiotic therapy for up to 6 weeks [8]. In our series, cultures of cardiac tissue were done in three patients. Brucella grew in the culture from one patient who had only had 5 days of antibiotic treatment before surgery. The two patients whose cultures were negative had received 10 days and 11 weeks of antibiotic treatment before surgery. The use of serologic tests to measure antibodies against Brucella is helpful in the diagnosis of brucellosis. The standard tube agglutination test (STA) is sensitive and specific. A titer of 1 : 160 or more is presumptive evidence of brucella infection [15]. Three patients in this series had a positive STA test of 1 : 1280 or more. The fourth patient had repeatedly negative serology. This patient’s blood grew Brucella species. A false-negative STA test in patients with brucellosis can occur if serum is not diluted beyond 1 : 320 owing to a prozone, or if the infection is caused by B. canis, which does not react to the antigen routinely used in the STA test [17]. It is unlikely that

ß 2001 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 7, 369–393

384 Clinical Microbiology and Infection, Volume 7 Number 7, July 2001

the negative test in our patient is related to a prozone, since serum is routinely tested to dilution beyond 1 : 320 in our laboratory. Echocardiography, especially transesophageal echocardiography, was very helpful in establishing the diagnosis of brucella endocarditis in our patients by showing vegetations in all patients, paravalvular leak in two patients, and an abscess in one patient. Uncertainty exists regarding many aspects of the treatment of brucella endocarditis, including the appropriate antibiotic combination and its duration, and the indications for surgery and its timing [5,15,18,19]. A combination of antimicrobial agents and valve replacement is the most accepted treatment for brucella endocarditis [4,5]. Our first patient with native valve brucella endocarditis was cured with medical treatment alone. Antimicrobial agents can achieve sterilization of valve vegetations, and several cases of cure of native valve brucella endocarditis with medical therapy alone have been reported [15,20– 22]; however, because of the risk of relapse and the catastrophic consequences of a relapsing infection, it is of the utmost importance to carefully monitor the patient so that complications necessitating valve replacement are detected as early as possible [18]. Medical therapy for brucella endocarditis consists of a combination of antibiotics used for a long period. The optimal combination and duration are unknown. In our patients, different combinations of doxycycline, rifampin, streptomycin and co-trimoxazole were used, so definite recommendations regarding the best combination could not be made. A prolonged course of antibiotics to prevent relapse seems to be advisable [15]. In different reports, the duration of antibiotic therapy ranged from 6 weeks to 12 months, and in our series from 10 to 20 weeks. The decision to stop treatment must be determined in each case after thorough clinical observation [15]. The other three patients with prosthetic valve endocarditis were treated with a combination of antibiotics and valve replacement. Prosthetic valve endocarditis caused by Brucella is cited as a primary indication for surgery [8]; there is only one case report of cure of brucella prosthetic valve endocarditis with antibiotics alone [13]. Valve replacement may be performed during the active phase of the disease, as the risk of recurrent infection in the new prosthesis is low [5,23], as was the case in two of our patients. The outcome was good in our series, which may be because none of our patients had heart failure during the course of their illness. In conclusion, brucella endocarditis remains a rare but serious complication of brucellosis. In our region, brucella endocarditis should be seriously considered in the differential diagnosis of patients with abnormal valves presenting with endocarditis, especially in those at risk for brucellosis. Brucella endocarditis of native valves may be treated with antibiotics alone, with careful monitoring for complications. A prolonged course of antibiotics is necessary. Brucella prosthetic valve endocarditis

should be treated with a combination of antibiotics and valve replacement. Early surgical intervention during the active phase of infection is safe, since the risk of infection of the new prosthesis is low.

R EFER E NCE S 1. Young E. An overview of human brucellosis. Clin Infect Dis 1995; 21: 283–9. 2. Ariza J. Brucellosis. Curr Opin Infect Dis 1996; 9: 126–31. 3. Colmenero JD, Reguera JM, Martos F et al. Complications associated with Brucella melitensis infection: a study of 530 cases. Medicine (Baltimore) 1996; 75: 195–211. 4. Peery TM, Belter LF. Brucellosis and heart disease: II. Fatal brucellosis. Am J Pathol 1960; 36: 673–97. 5. Jacob F, Abramowicz D, Vereerstraeten P, Le Clerc JL, Zech F, Thys JP. Brucella endocarditis. The role of combined medical and surgical treatment. Rev Infect Dis 1990; 12: 740–4. 6. Romero-Vivas J, Garcia-Fernandez F, Munoz-Sanz A, Buzon LM, Nunez L, Bouza E. Successful treatment of Brucella endocarditis [abstract 875]. In: Program and Abstracts of the 23rd Interscience Conference on Antimicrobial Agents and Chemotherapy. Washington, DC: American Society for Microbiology, 1983. 7. Al-Kassab S, Al-Fagih MR, Al-Yousef S et al. Brucella infective endocarditis. successful combined medical and surgical therapy. J Thorac Cardiovasc Surg 1988; 95: 862–7. 8. Fernandez-Guerrero ML, Martinell J, Aguado JM, del Carmen Ponte M, Fraile J, de Rabago G. Prosthetic valve endocarditis caused by Brucella melitensis: a report of four successfully treated with tetracycline, streptomycin, and sulfamethoxazole and trimethoprim plus valve replacement. Arch Intern Med 1987; 147: 1141–3. 9. O’Meara JB, Eykyn S, Jenkins BS, Braimbridge MV, Phillips I. Brucella melitensis endocarditis: successful treatment of an infected prosthetic mitral valve. Thorax 1974; 29: 377–81. 10. Saadeh AM, Abu-Farsakh NA, Omari HZ. Infective endocarditis and occult splenic abscess caused by Brucella melitensis infection: a case report and review of the literature. Acta Cardiol 1996; 3: 279–85. 11. Iglesias A, Nunez L. Brucella melitensis endocarditis on a porcine heterograft bioprosthesis. Chest 1981; 80: 747–9. 12. Lezuan R, Teruel J, Maitre MJ, Artaza M. Brucella endocarditis on double valvular prosthesis. Post Grad Med 1980; 56: 119–20. 13. Cakalagaoglu C, Keser N, Alhan C. Brucella mediated prosthetic valve endocarditis with brachial artery mycotic aneurysm. J Heart Valve Dis 1999; 8(5): 586–90. 14. Arslan H, Korkaz ME, Kart H, Gul C. Management of brucella endocarditis of a prosthetic valve. J Infect 1998; 37(1): 70–1. 15. Fernandez-Guerrero ML. Zoonotic endocarditis. Infect Dis Clin North Am 1993; 7: 135–41. 16. Jeroudi M, Halim MA, Harder JE, Al-Sibai MB, Ziady G, Mercer EN. Brucella endocarditis. Br Heart J 1987; 58: 279–83. 17. Salata RA, Ravdin JI. Brucella species (brucellosis). In: Mandell GL, Douglas RD, Bennet JE, eds. Principles and Practice of Infectious Diseases. New York: John Wiley, 1985: 1283–90. 18. Cisneros JM, Pachon J, Cuello JA, Martinez A. Brucella endocarditis cured by medical treatment. J Infect Dis 1989; 160: 907.

ß 2001 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 7, 369–393

Concise Communications

19. Pazderka E, Jones JW. Brucella abortus endocarditis: successful treatment of an infected valve. Arch Intern Med 1982; 142: 1567–8. 20. Hudson RA. Brucella abortus endocarditis. A case. Circulation 1957; 16: 411–13. 21. Quinn RW, Brown JW. Bacterial endocarditis. Arch Intern Med 1954; 94: 629–48.

385

22. Al-Mudallal DS, Mousa ARM, Marafie AA. Apyrexic Brucella melitensis aortic valve endocarditis. Trop Geogr Med 1989; 41: 372–6. 23. Baumgartner WA, Miller DC, Reitz BA et al. Surgical treatment of prosthetic valve endocarditis. Ann Thorac Surg 1983; 35: 87–104.

Ventriculoperitoneal shunt infection caused by Staphylococcus lugdunensis J. A. T. Sandoe and C. M. Longshaw Division of Microbiology, School of Biochemistry & Molecular Biology, University of Leeds, Leeds, UK 

Tel: þ44 113 233 5634

Fax: þ44 113 233 5638

E-mail: [email protected]

Accepted 11 April 2001

Staphylococcus lugdunensis is a coagulase-negative staphylococcus (CoNS) with the capacity to cause opportunistic infections in humans. However, infections involving S. lugdunensis typically resemble those associated with S. aureus in terms of the virulence of the organism and the clinical course of infection, which is often highly destructive [1,2]. Clinical case reports of infection caused by this organism include native valve endocarditis, peritonitis, vertebral osteomyelitis and catheter-related infection but it has not previously been reported as a cause of ventriculoperitoneal (VP) shunt infection [1–3]. S. epidermidis is the most common cause of cerebrospinal fluid (CSF) shunt infections, causing more than 50% of cases in one series [4]. The affinity of S. epidermidis for medical devices, such as VP shunts, has been attributed to its ability to adhere to plastics, with the subsequent formation of biofilm which can reduce efficacy of treatment and cause persistence of infection [5]. Biofilm formation is believed to be mediated via the intercellular adhesin gene cluster (ica), which is found more frequently in invasive S. epidermidis isolates than contaminating strains [5,6]. This report describes a case of VP shunt infection caused by S. lugdunensis. In contrast to the usual indolent, subacute presentation of S. epidermidis VP shunt infection, this case presented acutely in a manner akin to a S. aureus infection. The association of S. lugdunensis with a VP shunt infection prompted a search for biofilm-forming ability which could be important for the pathogenicity of this organism. Since S. lugdunensis can cause aggressive infections, and shares some virulence factors with S. aureus, the causative strain was also investigated for hemolysin production and homologues of the global regulatory locus agr (accessory gene regulator) which regulates virulence factor production in S. aureus [7].

A 74-year-old male was referred for management of normal pressure hydrocephalus. The patient had experienced increasing problems with walking, urinary incontinence, memory loss and loss of inhibition for 1 year prior to presentation. A computed tomography (CT) scan was performed and showed ventriculomegaly out of proportion to his cerebral atrophy. In 1996, the patient underwent insertion of a VP shunt and was discharged 7 days after the procedure, without any immediate postoperative problems. Seven days later the patient was readmitted with abdominal pain and sweating. The pain was right-sided, affecting the upper quadrant and was exacerbated by deep inspiration. On admission, the patient appeared ill and was pyrexic (37.9 8C). Peripheral white cell count was 17.5  109/L. Intra-abdominal sepsis was suspected and a CT scan confirmed the presence of a small subphrenic collection and associated pleural effusion. The shunt was removed and the patient started empirically on intravenous vancomycin, cefuroxime and rifampicin. At the time of surgery, the scalp incision was opened first, and the ventricular tip and a specimen of CSF sent for microscopy and culture. The abdominal wound was then opened and, on removal of the peritoneal catheter, purulent fluid emerged from the abdominal cavity. Specimens of pus and the peritoneal catheter were sent for culture. Postoperatively, the patient made a good recovery and intravenous vancomycin with rifampicin was continued for 1 week. A further 1 week of oral ciprofloxacin and rifampicin was prescribed and the patient was discharged well. A penicillin was not prescribed because of a history of allergy. Initial microscopy and Gram stain of the CSF showed 6/mm3 red blood cells, 6/mm3 white blood cells and scanty

ß 2001 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 7, 369–393

386 Clinical Microbiology and Infection, Volume 7 Number 7, July 2001

Gram-positive cocci. The peritoneal pus contained >3000/mL red blood cells, > 3000/mL white blood cells but no organisms were seen. Cultures of CSF, intraperitoneal pus, ventricular catheter tip and peritoneal catheter tip subsequently grew a catalase-positive Gram-positive coccus which was clumping factor-positive, tube coagulase-negative (rabbit plasma) and DNAse-negative. The organism also demonstrated synergistic b-hemolysis when plated alongside S. aureus 8325 on sheep blood agar. The isolate was identified as S. lugdunensis by API 32 Staph identification kit (bioMerieux, Mary-l’Etoile, France) [API profile 467114600] and conventional tests. All four isolates were biochemically indistinguishable. The ability to adhere to plastic and subsequent biofilm formation was quantified using a microtitre plate assay. The optical density of S. lugdunensis biofilm at 590 nm was 0.245 compared with 0.458 shown by S. epidermidis S9 (a biofilm-producing strain), where values above 0.12 were considered to be positive for biofilm production. S. epidermidis and S. aureus adhesion to biomaterials and production of biofilm has been associated with production of extracellular polysaccharide adhesins PS/A and PIA, encoded by the intercellular adhesin (ica) locus [8–10]. Production of extracellular polysaccharide by the clinical isolate described here was initially indicated by production of black pigmented colonies when grown on Congo red agar (CRA). Because previous studies of S. epidermidis have shown an association between black pigmentation of colonies on CRA and a functional ica locus [9], it was decided to screen the genomic DNA of the isolate for the presence of a homolog of the icaA gene using PCR. Oligonucleotide primers: ICAF (GATGGAAGTTCTGATAATAC) and ICAR (CCTCTGTCTGGGCTTGACC), were designed towards regions of high sequence conservation between the icaA genes of S. epidermidis and S. aureus. Genomic DNA was prepared from S. lugdunensis and a PCR reaction using the above primers generated a single 973 (bp) DNA product. This correlated well with the expected size, predicted from the icaA sequence from S. epidermidis and S. aureus. The DNA product was subsequently sequenced from both DNA strands and was found to have high homology to the icaA genes of both S. epidermidis and S. aureus, with 80% and 92% identical residues, respectively. During infection, S. aureus produces a number of extracellular proteins that are important for the pathogenesis of the organism. Synthesis of many of these factors, including protein A, hemolysins, enterotoxins and toxic shock syndrome toxin – 1, is regulated by an untranslated mRNA molecule, RNAIII, whose expression level is modulated by the global regulatory locus, agr [11]. agr-Related sequences have been demonstrated in S. lugdunensis previously and were investigated in this clinical isolate [12]. Oligonucleotide primers: RNAF (GACAGCTTAGTGCCACATTT) and RNAR (GCCTAACTGTTTAAT TCGCATA) were designed toward the S. lugdunensis

RNAIII homologue using the published sequence (GenBank accession no. LI3334). PCR was carried out using S. lugdunensis genomic DNA and amplified a single 600 base pair (bp) DNA product which correlated with the expected size of 597 bp. The DNA product was subsequently identified as the S. lugdunensis RNAIII homolog by sequencing from both strands. CoNS are among the commonest causes of VP shunt infections [4]. It is the ability of these organisms to adhere to indwelling medical devices and form biofilms that allows them to colonize and persist on VP shunts. CoNS are usually low virulence pathogens and the VP shunt infections they cause are mostly indolent and subacute in presentation. In contrast, S. aureus shunt infections frequently manifest a systemic inflammatory response, signs of meningism and involve the shunt and surrounding tissues. The case described suggests that this strain of S. lugdunensis was able to adhere to shunt material and also cause an acute infection similar in clinical picture to S. aureus infection. This isolate of S. lugdunensis could adhere to plastic and form biofilms in vitro, and possessed pathogenicity factors homologous to those which are known to be important for similar infections in S. aureus. S. lugdunensis is usually susceptible to a wide range of antimicrobials, although treatment of CoNS VP shunt infections usually requires shunt removal in addition to antibiotics, as was necessary in this case, to achieve complete clearance of the infecting organism. Infections caused by this organism have been described as resembling those of S. aureus rather than the less virulent CoNS, and further evidence is provided here for the aggressive nature of S. lugdunensis and its significance as an opportunistic human pathogen [2]. The ability of S. lugdunensis to cause infections of indwelling medical devices via biofilm formation and the presence of an icaA gene homolog warrants further investigation, as does the presence of an agr-related sequence whose role in this organism is unclear but which signifies the potential for coordinated virulence gene regulation. S. lugdunensis produces clumping factor but can readily be distinguished from S. aureus in the clinical diagnostic laboratory by negative tube coagulase and DNAse reactions. It is our opinion that isolation of S. lugdunensis should not be disregarded as a contaminant when recovered from a normally sterile body site and should prompt a rapid clinical evaluation by attending clinicians. Infections of VP shunts by S. lugdunensis can be added to the growing list of site-specific infections caused by this emerging opportunistic pathogen. R EFER E NCE S 1. Vandenesch F, Etienne J, Reverdy ME et al. Endocarditis due to Staphylococcus lugdunensis. report of 11 cases and review. Clin Infect Dis 1993; 17: 871–6.3. 2. Schnitzler N, Meilicke R, Conrads G et al. Staphylococcus lugdunensis: report of a case of peritonitis and an easy-to-perform screening strategy. J Clin Microbiol 1998; 36: 812–3.

ß 2001 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 7, 369–393

Concise Communications

3. Murdoch DR, Everts RJ, Chambers ST et al. Vertebral osteomyelitis due to Staphylococcus lugdunensis. J Clin Microbiol 1996; 34: 993–4. 4. Bayston R. Hydrocephalus shunt infections. J Antimicrob Chemother 1994; 34 (Suppl. A): S75–S84. 5. Frebourg NB, Lefebvre S, Baert S et al. PCR-based assay for discrimination between invasive and contaminating Staphylococcus epidermidis strains. J Clin Microbiol 2000; 38: 877–80. 6. Von Eiff C, Heilmann C, Peters G. New aspects in the molecular basis of polymer-associated infections due to staphylococci. Eur J Clin Microbiol Infect Dis 1999; 18: 843–6. 7. Novick RP, Ross HF, Projan SJ et al. Synthesis of staphylococcal virulence factors is controlled by a regulatory RNA molecule. EMBO J 1993; 12: 3967–75.

387

8. Heilmann C, Schweitzer O, Gerke C et al. Molecular basis of intercellular adhesion in the biofilm-forming Staphylococcus epidermidis. Mol Microbiol 1996; 20: 1083–91. 9. Ziebuhr W, Heilmann C, Go¨ tz F et al. Detection of the intercellular adhesion gene cluster (ica) and phase variation in Staphylococcus epidermidis blood culture strains and mucosal isolates. Infect Immun 1997; 65: 890–6. 10. Crampton SE, Gerke C, Schnell NF et al. The intercellular adhesion (ica) locus is present in Staphylococcus aureus and is required for biofilm formation. Infect Immun 1999; 67: 5427–33. 11. Morfeldt E, Tegmark K, Arvidson S. Transcriptional control of the agr-dependent virulence gene regulator, RNAIIIS, in Staphylococcus aureus. Mol Microbiol 1996; 21: 1227–37. 12. Vandenesch F, Projan SJ, Kreiswirth B et al. Agr-related sequences in Staphylococcus lugdunensis. FEMS Microbiol Lett 1993; 111: 115–22.

Bartonella henselae neuroretinitis F. Messina1, R. Doria1, G. Gabriellini2, M. S. Sartini2, C. Tascini1 and F. Menichetti1 1

U. O. Malattie Infettive, Azienda Ospedaliera Pisana and 2Clinica Oculistica, Universita` di Pisa, Via Paradisa 2, 56100, PISA, Italy Tel: þ39 050 996756

Fax: þ39 050 571021 E-mail: [email protected]

Accepted 27 March 2001

Uveitis may be an important cause of severe visual impairment and blindness, and causes significant morbidity in the economically active young adult population [1]. Infectious processes play an important role in the pathogenesis of uveitis; specifically, herpesviruses and Toxoplasma gondii are common causative agents in immunocompetent as well as in immunocompromised patients [2]. The recognition of an infectious etiology of uveitis is pivotal in choosing the right management. While immunosuppressive medication is essential for the majority of patients with non-infectious uveitis, this treatment modality may induce a fulminant course of infection without specific antimicrobial treatment [3]. A 28-year-old HIV-negative female patient, who presented with left eye intraocular inflammatory disease (neuroretinitis type) is described. Five days before the referral, she had fever and, on the day of admission, a sudden decrease of visual activity of her left eye. Ophthalmological examination revealed visual acuity of 10/10 in the right eye and 6/10 in the left eye, and papillitis with macular edema and retinal vasculitis in both eyes. On general examination, no abnormalities were noticed; specifically, no signs of systemic vasculitis or systemic cat scratch disease (CSD) were present. The results of laboratory evaluations for erythrocyte and leukocyte counts and renal and liver function tests were within the normal limits. The patient’s erythrocyte sedimentation rate was 86 mm/h. There was no serological evidence of infection with herpes

simplex virus (HSV), varicella-zoster virus (VZV), adenovirus, cytomegalovirus (CMV), Epstein–Barr virus (EBV), Borrelia burgdorferi, T. gondii, Coxiella burneti, Brucella spp., Francisella tularensis or syphilis. Tuberculin skin test was negative. Ophthalmoscopic analysis showed signs of optic disc edema, with swelling of the disk in the left eye. Visually evoked potential (VEP) revealed a normal VEP amplitude and a prolongation of the P2 latency. Standard electroretinogram (ERG), macular ERG and cone responses were normal. Fluoro-angiography confirmed a diagnosis of left eye papillitis: a leakage from the optic disk. The indirect fluorescence assay (IFA), with B. henselae as the antigen [4], revealed an IgG titer of 1 : 512 and an IgM titer of 1 : 32 in the serum samples. In our laboratory, the cut-off values for positive serology are 1 : 64 for IgG and 1 : 32 for IgM [4]. A diagnosis of ocular bartonellosis was made, and the patient was treated with azithromycin at 600 mg/day and rifampin at 600 mg/day for 8 weeks. After 4 weeks, IgG titer and IgM titer were 1 : 1024 and 1 : 256, respectively. Upon post-treatment ophthalmic examination, the intraocular inflammation appeared to have subsided. At the same time, IFA IgG titer was 1 : 1024 and IgM titer negative (Table 1). Domestic cats are a major reservoir for B. henselae. Studies suggested that not only fleas but also the human body louse, the sand fly and ticks are possible vectors in the transmission of Bartonella species [5]. Our patient owned several cats and

ß 2001 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 7, 369–393

388 Clinical Microbiology and Infection, Volume 7 Number 7, July 2001

Table 1 Serological and clinical evolution of the patient described Time

IgG

IgM

Manifestation

Therapy

0 4 weeks 8 weeks

1: 512 1:1024 1:1024

1: 32 1: 256 Negative

Left eye retinitis Left eye retinitis Resolved

Azithromycin plus rifampin Azithromycin plus rifampin Stop

presented with cat scratches on her neck, as a consequence of which she suffered from ocular bartonellosis without the typical systemic symptoms of CSD. In immunocompetent patients, ocular disease usually develops after systemic involvement has subsided, suggesting that ocular involvement represents a late complication of the disease [6,7]. However, sometimes the acute phase cannot be recognized and Bartonella DNA has been demonstrated in the ocular fluids as long as 1 year after the start of the symptoms. Ocular bartonellosis is mainly associated with B. henselae infections. This association is predominantly based on serological grounds [6,8]. The characteristic evolution of IgG and IgM serology against B. henselae in the serum of this patient supports the hypothesis of acute infection with ocular localization. Previous studies based on serology were performed mainly by IFA, and no specific IgM titers were determined for these patients [6,9]. In this case, the diagnosis of B. henselae ocular infection was based on the demonstration of the presence of IgM. Therapy with rifampin and azithromycin was effective. Prospective clinical studies evaluating this antibiotic association in B. henselae infection are warranted.

R EFER E NCE S 1. Suttorp MSA, Rotova A. The possible impact of uveitis in blindness: a literature survey. Br J Ophthalmol 1996; 80: 844–8. 2. de Boer JH, Verhagen C, Bruinenberg M, Rothova A. Serologic and polymerase chain reaction analysis of intraocular fluids in the diagnosis on infectious uveitis. Am J Ophthalmol 1996; 121: 650–8. 3. Sabates R, Pruett RC, Brockhurst RJ. Fulminant ocular toxoplasmosis. Am J Ophthalmol 1981; 92: 497–503. 4. Regnery RL, Olson JG, Perkins BA, Bibb W. Serological response to Rochalimaea henselae antigen in suspected cat-scratch disease. Lancet 1992; 339: 1443–5. 5. Maurin M, Birtles RJ, Raoult D. Current knowledge of Bartonella species. Eur J Clin Microbiol Infect Dis 1997; 16: 487–506. 6. Golnik KC, Marotto ME, Fanous MM et al. Ophthalmic manifestations of Rochalimaea species. Am J Ophthalmol 1994; 118: 145–51. 7. Ormerod D, Skolnick KA, Menesky MM, Pavan PR, Pon DM. Retinal and choroidal manifestations of cat-scratch disease. Ophthalmology 1998; 105: 1024–31. 8. Kerkhoff FT, de Ossevaarde JM, Loss WS, Rothova A. Presumed ocular bartonelosis. Br J Ophthalmol 1999; 83: 270–5. 9. Cohen SM, Davis JL, Gass DM. Branch retinal arterial occlusion in multifocal retinitis with optic nerve edema. Arch Ophthalmol 1995; 113: 1271–6.

Polyclonal Staphylococcus epidermidis intravascular catheter-related infections B. J. A. Rijnders1, E. Van Wijngaerden1, J. Van Eldere2 and W. E. Peetermans1 1

Inwendige geneeskunde and 2Labo Bacteriologie, Universitair Ziekenhuis Gasthuisberg Leuven, Herestraat 49, B-3000 Leuven, Belgium 

Fax: þ32 16 346275

Tel: þ32 16 346274

E-mail: [email protected]

Accepted 4 May 2001

During a 4-month period we prospectively investigated the frequency of polyclonal catheter infections with Staphylococcus epidermidis. Of each catheter with pure growth of S. epidermidis, six colonies were genotypically analyzed with pulsed-field gel electrophoresis. Two out of 12 patients with catheter infection had a polyclonal infection. Both clones of each catheter had a clearly different antibiotic susceptibility. This study shows that polyclonal catheter infections are not exceptional. Further studies are needed to define the clinical consequences of polyclonal catheter infection.

ß 2001 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 7, 369–393

Concise Communications

IN T R ODUC T ION Nosocomial bacteremia is frequently caused by catheter-related infections and Staphylococcus epidermidis is the single most frequently isolated pathogen [1]. These infections usually are considered monoclonal (originating from a single clone), and antibiotic susceptibility testing is carried out using a subculture of a single colony taken from the catheter culture. In 1997, we described a patient with polyclonal S. epidermidis prosthetic valve endocarditis [2]. More recently the polyclonal nature of joint prosthesis infections with S. epidermidis was described [3] and the genetic variability of an infecting S. epidermidis strain during prosthetic valve endocarditis was shown [4]. Polyclonality of infections may result in inadequate and misleading antibiotic susceptibility testing, because only one colony is tested, in general. However, the prevalence of polyclonality in catheter-related infections with S. epidermidis has never been determined. In the present observational study we investigated the prevalence of polyclonal S. epidermidis catheter-related colonization and infection. ME T HOD S In the medical microbiology laboratory of a 1900-bed university teaching hospital, bacterial isolates from catheters with a pure growth of S. epidermidis were collected prospectively. The catheters were quantitatively cultured using the sonication technique [5]. For inclusion in the study, >1000 colonyforming units (cfu)/catheter had to be present on the catheter. Staph-Zym (ROSCO, Taastrup, Denmark) was used for the phenotypic identification of S. epidermidis. The treating physician and medical records were consulted both at catheter removal and 48 h later to collect the clinical data needed. From each catheter culture, six colonies were taken randomly, inoculated in brain–heart infusion broth, grown for 4 h and then frozen until further analysis. Genotypic analysis of these isolates was performed using SmaI digestion (Gibco-BRL, Gaithersburg, MD, USA) of chromosomal DNA, and restriction fragments were separated by pulsed-field gel electrophoresis (PFGE) using the CHEF Mapper System (Bio-Rad, Hercules,

389

CA, USA) in 1% (wt/vol) chromosomal grade agarose gels (Bio-Rad). Electrophoresis was performed at 200 V with alternating pulses at a 1208 angle in a 5–15 s pulse time gradient for 10 h and then a 15–45 s gradient for the next 12 h [6]. Polyclonality was defined as described by Tenover et al. [7]. Catheters in which the investigated clones showed a difference of four bands or more were considered polyclonal. If clones differed in fewer than four bands, they were considered genotypically closely related. Antibiotic susceptibility testing was done by the agar dilution method at breakpoint concentrations. Chloramphenicol and doxycycline susceptibility was tested using the disk-diffusion method (neoSensitabs, ROSCO, Taastrup, Denmark). Methicillin resistance was confirmed by the presence of the mecA gene, determined by PCR, as described by Murakami [8]. Catheter-related infection was defined as the presence of a catheter with >1000 cfu/catheter retrieved from a patient with fever and in whom the fever disappeared within 48 h following catheter removal. Because fever is rare in neonates, the prompt disappearance of signs (apnea, bradycardia) after the catheter removal was accepted for the definition of catheter-related infection in this population. Catheter-related bloodstream infection was defined as a catheter with >1000 cfu/catheter and two peripheral blood cultures positive for S. epidermidis. Sepsis had to resolve within 48 h after removal of the catheter. Because it is difficult to obtain two peripheral blood cultures in neonates, one positive peripheral blood culture with S. epidermidis was considered significant in this population. Colonization of a catheter was defined as a catheter growing >1000 cfu/ catheter in a patient without documented bacteremia in whom the symptoms of sepsis did not resolve after catheter removal. R E SULT S During a 4-month period (two periods of 2 months) all 15 catheters with a pure growth of S. epidermidis obtained from 15 patients were studied. Fourteen of these 15 catheters were central venous catheters and six were from neonates. Nine patients had catheter-related bloodstream infection, three had a catheter-related infection without bloodstream infection and three had catheter colonization without infection. SmaI

Table 1 Antibiotic susceptibility of different clones of patients 4 and 13 with polyclonal catheter infection-colonization Patient 4 13

Clone

Pen

Oxa

Ery

Clin

Dox

Co

Ofl

Gen

Tob

Am

Chl

Van

a b a b

R R R R

R S R S

R S R S

I S S S

S S S S

S S S S

S S S S

S S S S

S S I S

S S S S

S S S S

S S S S

Chl, chloramphenicol; Co, cotrimoxazole; Gen, gentamicin; Tob, tobramycin; Am, amikacin; Ofl, ofloxacin; Pen, penicillin; Oxa, oxacillin; Ery, erytromycin; Clin, clindamycin; Dox, doxycycline; Van, vancomycin. Differences in antibiotic susceptibility between clones a and b are in italic.

ß 2001 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 7, 369–393

390 Clinical Microbiology and Infection, Volume 7 Number 7, July 2001

Figure 1 Pulsed-field gel electrophoresis pattern of one polyclonal catheter (patient 13), one with genotypically closely related but distinguishable clones (patient 8) and one strictly clonal infection (patient 3).

digestion of bacterial DNA and PFGE were performed on 90 colonies (six colonies of 15 catheters). Cultures from two of the 15 catheters (patients 4 and 13) proved to be polyclonal. On each of these catheters, two distinct clones were found. In patient 13, a neonate with catheter-related bloodstream infection, one clone was found in two of the six colonies tested and the other clone was found in the remaining four colonies. In patient 4, an adult patient with catheter-related infection, one clone was isolated in only one of the six colonies tested. The antibiotic susceptibilities of these four different clones are shown in Table 1. Although phenotypic methicillin resistance could be demonstrated in only one of the clones from patient 13, the mecA gene was demonstrated in both clones. In patient 4, the mecA gene could only be found in the methicillinresistant clone. In one patient (patient 8), two closely related but not identical clones were found. However, the antibiotic susceptibility of both clones was exactly the same. Figure 1 shows the pulsed-field pattern of the distinct clones of patient 13, as well as the closely related but not identical isolates of patient 8 and the identical isolates of catheter 3. DI SCUS SION Although catheter-related infection with S. epidermidis is considered to be monoclonal, several observations suggest the polyclonal nature of foreign-body infections with S. epidermidis [2,3]. In 15 consecutive catheters infected or colonized with S. epidermidis, we found two cases of polyclonal infection. If only the 12 patients with catheter infection or catheter-related bloodstream infection are taken into account, polyclonal infection was found in two of these 12 patients. We performed a genotypic analysis of only six colonies of each catheter. It is possible that more polyclonal infections would have been identified if more colonies had been analyzed but PFGE is

very labor-intensive and we arbitrarily decided to limit the number to six. In the two polyclonal catheter infections, we found a good correlation between the antibiogram of the different clones and the PFGE pattern. Thus, performing an antibiogram on several clones might have given the same clinically relevant clonal differentiation. The colonies tested were selected after 48 h of growth and, at that time, no differences in colony morphology were observed. It is impossible to verify whether each of the different strains are equally responsible for the catheter-related infection, as theoretically some strains may colonize the catheter and others may actually infect the catheter. As no difference in colony morphology was demonstrated when the strains were recovered for further PFGE, it was impossible to demonstrate quantitatively the polyclonality any further because it is not possible to genotype all colonies from each patient. As a consequence we cannot say with certainty which of the genotypes was predominant and if all genotypes were present in >1000 cfu/catheter. Polyclonal infections with S. epidermidis were previously demonstrated in a case of prosthetic valve endocarditis [2] and in prosthetic joints infections [3]. C ONCL US ION S Our study suggests that polyclonality of catheter infections, a frequent foreign-body infection, is not exceptional. S. epidermidis foreign-body infections are difficult to treat for several reasons. Polyclonality might be one of the explanations. Although it might be tempting to use vancomycin for all S. epidermidis catheter infections because of concerns about polyclonality, antistaphylococcal penicillins are preferable when a single sensitive clone causes the infection. However, it might be wise to use vancomycin in those patients who have no clinical response to treatment with antistaphylococcal penicillins when catheter removal is not possible (e.g. operatively implanted catheter). Our study also suggests that in S. epidermidis foreign-body infections, the finding of a genotypically different clone during a second episode of infection does not necessarily mean reinfection. It might just as well be a relapse of an (inadequately treated) polyclonal infection. ACK NOW L EDGME N T Supported in part by research funds from the R. von Furth Chair in Infectious Diseases at the K.U. Leuven (to W.E.P.).

R EFER E NCE S 1. Edmond MB, Wallace SE, McClish DK, Pfaller MA, Jones RN, Wenzel RP. Nosocomial bloodstream infections in United States hospitals: a three-year analysis. Clin Infect Dis 1999; 29 (2): 239–44.

ß 2001 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 7, 369–393

Concise Communications

2. Van Wijngaerden E, Peetermans WE, Van Lierde S, Van Eldere J. Polyclonal Staphylococcus endocarditis [see comments]. Clin Infect Dis 1997; 25 (1): 69–71. 3. Galdbart JO, Morvan A, Desplaces N, el Solh N. Phenotypic and genomic variation among Staphylococcus epidermidis strains infecting joint prostheses. J Clin Microbiol 1999; 37 (5): 1306–12. 4. Van Eldere J, Peetermans WE, Struelens M, Deplano A, Bobbaers H. Polyclonal Staphylococcal endocarditis caused by genetic variability. Clin Infect Dis 2000; 31 (1): 24–30. 5. Sherertz RJ, Raad II, Belani A et al. Three-year experience with sonicated vascular catheter cultures in a clinical microbiology laboratory. J Clin Microbiol 1990; 28 (1): 76–82.

391

6. George CG, Kloos WE. Comparison of the SmaI-digested chromosomes of Staphylococcus epidermidis and the closely related species Staphylococcus capitis and Staphylococcus caprae. Int J Syst Bacteriol 1994; 44 (3): 404–9. 7. Tenover FC, Arbeit RD, Goering RV et al. Interpreting chromosomal DNA restriction patterns produced by pulsed-field gel electrophoresis: criteria for bacterial strain typing. J Clin Microbiol 1995; 33 (9): 2233–9. 8. Murakami K, Minamide W, Wada K, Nakamura E, Teraoka H, Watanabe S. Identification of methicillin-resistant strains of staphylococci by polymerase chain reaction. J Clin Microbiol 1991; 29 (10): 2240–4.

Recurrent bacteremia by Chryseobacterium indologenes in an oncology patient with a totally implanted intravascular device E. Nulens1, B. Bussels2, A. Bols2, B. Gordts1 and H. W. Van Landuyt1 1

Laboratory of Microbiology and 2Oncologisch Centrum, Algemeen Ziekenhuis Sint-Jan, Ruddershove 10, B-8000 Brugge, Belgium 

Tel: þ32 50 45 26 00

Fax: þ32 50 45 26 09

E-mail: [email protected]

Accepted 11 May 2001

Chryseobacterium indologenes was isolated from the blood cultures of an oncological patient with a totally implantable device. Because a catheter-related infection was suspected, the Port-A-Cath1 was removed after a 10-day course of piperacillin–tazobactam. Differences in susceptibility may exist if either the criteria for either Pseudomonas or Enterobacteriaceae are used.

IN T R ODUC T ION Totally implantable intravascular devices are used to administer drugs to patients with a malignant disease. These devices offer the advantage of improved patient image and obviate the need for routine catheter-site care. However, pathogenic and opportunistic bacteria can contaminate indwelling catheters, especially in the hospital environment. The device can then become a reservoir for further dissemination of the bacteria. We report a case of an oncology patient with a totally implantable intravascular device (Port-A-Cath1, PAC) that became contaminated with Chryseobacterium indologenes. C A SE R E P OR T In October 1999, a 38-year-old woman was diagnosed with bone and liver metastases of an already locally and regionally advanced breast cancer. Because of the severity of the hypercalcemia and associated symptoms (the patient underwent a successful resuscitation because of ventricular fibrillation),

urgent renal dialysis was performed. Two weeks later a chemotherapeutic regimen of CAF (cyclophosphamide– adriamycine5-fluorouracyl) together with biphosphonates was started in a day-patient setting. The drugs were administered through the PAC on days 1 and 8, and repeated every 28 days. On day 8 of this first cycle the patient was urgently admitted to the hospital because of a relapsing hypercalcemia. Laboratory investigations revealed no neutropenia. Within half an hour after manipulation of the PAC she experienced a fever of 39 8C. Apart from the tumor in the right breast, physical examination revealed no abnormalities. The patient was started on a combination of intravenous piperacillin and tazobactam (4  4 g IV daily). One day after blood cultures were taken, bacterial growth was detected with a Bactec 9240 (Becton Dickinson Diagnostic Instruments System, Sparks, MD, USA) in the aerobic blood culture bottles (BactecTM Plus Aerobic) collected from two different samples. The first sample was through a venous puncture; the second sample was through the PAC of the patient. The anaerobic blood bottles cultured negative.

ß 2001 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 7, 369–393

392 Clinical Microbiology and Infection, Volume 7 Number 7, July 2001

The Gram stain showed Gram-negative bacilli. On the second day, two types of colonies appeared: the first strain was identified as Acinetobacter lwoffii; the second strain grew on conventional agar media but did not grow on MacConkey agar. Chryseobacterium was suspected but the conventional tests could not make a differentiation to the species level after 72 h of incubation. With the BBL1 CrystalTM Enteric/nonfermenter identification system (Becton Dickinson Diagnostic Instruments System), C. indologenes was identified (code 2301110112; oxidase-positive, indole-negative) with a probability of 70.3%. The strain was sent to a reference laboratory (Prof. Wauters, Cliniques Universitaires, St-Luc, Woluwe, Belgium) for further identification and confirmation. Antimicrobial susceptibility testing was performed with the disc diffusion method according to NCCLS [1]. For interpretation we used the zone diameters for Pseudomonas aeruginosa. The strain was resistant to ampicillin, amoxycillin–clavulanate, cefazolin, cefuroxime, aztreonam, ceftriaxone, amikacin, doxycycline, colistin and meropenem. The strain was susceptible to piperacillin, piperacillin–tazobactam, ceftazidime, ofloxacin and trimethoprim–sulfamethoxazole. Additional E test susceptibility testing was performed, in accordance with the manufacturer’s directions, for piperacillin, piperacillin–tazobactam, ceftazidime and ciprofloxacin. The minimum inhibitory concentration (MIC) breakpoints of the NCCLS for ‘other nonEnterobacteriaceae’ were used. The strain was susceptible to the four antimicrobials tested, with a MIC of 4 mg/L for piperacillin (S 16 mg/L) and piperacillin–tazobactam (S 16 mg/ L), 3 mg/L and 0.5 mg/L for ceftazidime (S 8 mg/L) and ciprofloxacin (S 1 mg/L), respectively. After 2 days of intravenous antimicrobial treatment, the fever subsided totally. Intravenous antimicrobial treatment was given for 10 days. However, 6 days after the completion of antimicrobial treatment, the fever recurred. Bacterial growth was detected in the aerobic culture blood bottle, taken through the PAC, while blood taken from a venous puncture cultured negative. The Gram stain showed Gram-negative bacilli. This strain had the same morphological and biochemical characteristics and susceptibility results and was identified as C. indologenes. In view of the suspicion of catheter-related septicemia, the PAC device was removed. The PAC cultured negative, probably because only the external part of the device was rinsed with tryptic soy broth. Thereafter the patient was given an additional 7-day course of pefloxacin (2  400 mg IV daily). To date she has not experienced any recurrence of fever. The chemotherapeutic regimen was continued and we noticed a good response. DI SCUS SION C. indologenes is not part of the human flora but is found in soil, water, plants and foodstuffs [2]. In the hospital environment, the bacterium is found in water and on wet surfaces. Indwelling

vascular catheters, feeding tubes and other fluid-associated equipment may become reservoirs for chryseobacteria [2]. The bacteria can easily survive in the biofilm and resist antimicrobial treatment [3]. Our patient was treated in a day-patient setting, where hygienic measures may be less stringent. The drugs were administered through the PAC, which probably favored the infection. Chryseobacteria are non-fastidious, oxidase-positive Gram-negative rods that do not ferment glucose. They are readily distinguished from other non-fermenters by their ability to produce indole in tryptophan broth, but the reaction often is weak and difficult to demonstrate [4]. Despite their low virulence, chryseobacteria are inherently resistant to many antimicrobial agents, which makes them potential candidates for nosocomial infections [5]. Piperacillin, ofloxacin and ciprofloxacin are potential drugs of choice for the treatment of infections caused by chryseobacteria [6]. It has been shown that susceptibility disc diffusion testing is inaccurate for these organisms, since there is often a poor correlation between the disc diffusion results and the MIC values [5,7]. The reason for this remains unclear. Hence, the susceptibility of the organism cannot be predicted accurately from the disc diffusion method alone. Usually the zone diameters for Pseudomonas aeruginosa or Enterobacteriaceae for the disc diffusion results described by the NCCLS are used, but differences in susceptibility may occur [1,5,6]. An attempt to determine disc diffusion zone diameters for chryseobacteria has been made by Chang et al. [6]. According to the zone diameters used in their study, our strain would have been susceptible to piperacillin and piperacillin–tazobactam. The MIC breakpoints for chryseobacteria have not yet been established by the NCCLS [1,6]. We used the MIC breakpoints for ‘other non-Enterobacteriaceae’ [1]. The E test may be an alternative to the standard agar dilution method for testing susceptibility to ciprofloxacin and ceftazidime but not to piperacillin, since there may be discrepancies between the E test result and the standard agar dilution MIC for piperacillin [5,7]. C ONCL US ION S The reappearance of the bacterium in the blood cultures 6 days after completion of the first antimicrobial treatment was suggestive of indwelling device-related infection. The blood culture taken through the PAC cultured positive, while the venous blood puncture cultured negative. Generally it is accepted that the device should be removed if clinical symptoms do not improve after appropriate antimicrobial therapy [3]. However, successful treatment of indwelling device-related infections caused by C. indologenes without removal of the device have been reported [3]. An alternative treatment option is the antibiotic lock technique in which a high concentration of an antibiotic is delivered and locked into the device. This technique

ß 2001 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 7, 369–393

Concise Communications

may be beneficial to treat totally implantable intravascular device-related septicemia, but few reports have been published so far [8]. R EFER E NCE S 1. National Committee for Clinical Laboratory Standards. Methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically. Approved standard M7–A4. Villanova, PA: NCCLS, 2000. 2. Hsuenh PR, Hsiue TR, Wu JJ et al. Flavobacterium indologenes bacteremia: clinical and microbiological characteristics. Clin Infect Dis 1996; 23: 550–5. 3. Hsuenh PR, Teng LJ, Ho SW, Hsieh WC, Luh KT. Clinical and microbiological characteristics of Flavobacterium indologenes infections associated with indwelling devices. J Clin Microbiol 1996; 34: 1908–13.

393

4. The non-fermentative Gram-negative bacilli. In: Koneman EW, Allen SD, Janda WM, Schreckenberger PC, Winn WC Jr, eds. Color atlas and textbook of diagnostic microbiology, 5th edn. Philadelphia: Lippincott, 1997; 253–320. 5. Fraser SL, Jorgensen JH. Reappraisal of the antimicrobial susceptibilities of Chryseobacterium and Flavobacterium species and methods for reliable susceptibility testing. Antimicrob Agents Chemother 1997; 41: 2738–41. 6. Chang JC, Hsueh PR, Wu JJ, Ho SW, Hsieh WC, Luh KT. Antimicrobial susceptibility of flavobacteria as determined by agar dilution and disk diffusion methods. Antimicrob Agents Chemother 1997; 41: 1301–6. 7. Hsueh PR, Chang JC, Teng LJ et al. Comparison of E test and agar dilution method for antimicrobial susceptibility testing of Flavobacterium isolates. J Clin Microbiol 1997; 35: 1021–3. 8. Bregenzer T, Widmer AF. Bloodstream infection from a porta-cath: successful treatment with the antibiotic lock technique. Infect Contr Hospital Epidemiol 1996; 12: 772.

ß 2001 Copyright by the European Society of Clinical Microbiology and Infectious Diseases, CMI, 7, 369–393