Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Comparative Activity of Ciprofloxacin, Levofloxacin and Moxifloxacin against Klebsiella pneumoniae, Pseudomonas aeruginosa and Stenotrophomonas maltophilia Assessed by Minimum Inhibitory Concentrations and Time-Kill Studies

  • Antoine Grillon ,

    antoine.grillon@chru-strasbourg.fr

    Affiliation Institute of Bacteriology, Faculty of Medicine, University of Strasbourg and Strasbourg University Hospital, Strasbourg, France

  • Frédéric Schramm,

    Affiliation Institute of Bacteriology, Faculty of Medicine, University of Strasbourg and Strasbourg University Hospital, Strasbourg, France

  • Magali Kleinberg,

    Affiliation Institute of Bacteriology, Faculty of Medicine, University of Strasbourg and Strasbourg University Hospital, Strasbourg, France

  • François Jehl

    Affiliation Institute of Bacteriology, Faculty of Medicine, University of Strasbourg and Strasbourg University Hospital, Strasbourg, France

Comparative Activity of Ciprofloxacin, Levofloxacin and Moxifloxacin against Klebsiella pneumoniae, Pseudomonas aeruginosa and Stenotrophomonas maltophilia Assessed by Minimum Inhibitory Concentrations and Time-Kill Studies

  • Antoine Grillon, 
  • Frédéric Schramm, 
  • Magali Kleinberg, 
  • François Jehl
PLOS
x

Abstract

The aim of this study was to compare the in vitro susceptibility of Klebsiella pneumoniae, Pseudomonas aeruginosa and Stenotrophomonas maltophilia to three fluoroquinolones. The minimum inhibitory concentrations (MICs) to ciprofloxacin, levofloxacin and moxifloxacin were examined by E-test® for a total of 40 K. pneumoniae strains, 40 S. maltophilia strains and 40 P. aeruginosa strains. Then, the bactericidal activity of these fluoroquinolones was investigated on five strains of each bacterial species by means of time-kill curves. For K. pneumoniae and P. aeruginosa, the distance of the measured MIC from the clinical break-point is a good indicator of the bactericidal activity for ciprofloxacin and levofloxacin as obtained in our experiments. The lower the MIC, the better the bactericidal activity in term of CFU Log decreases. If MIC of ciprofloxacin and levofloxacin against the considered bacteria are far from clinical breakpoint, these two antibiotics are equivalent. According to our MIC50 and modal MIC, the breakpoints of both ciprofloxacin and levofloxacin seem to be somewhat high and data suggest reducing them. On S. maltophilia, none of the tested antibiotics showed a satisfactory activity.

Introduction

Gram-negative bacteria such as Enterobacteriaceae, P. aeruginosa and S. maltophilia are common nososomial pathogens. Among Enterobacteriaceae, Klebsiella is frequently isolated from hospitalized patients [1]. The Urinary tract is the most common site of infection by this pathogen in immunocompromised patients [2], but other infections such as pneumonia, bacteremia, wound infections, nosocomial infections in intensive care unit or neonatal septicemia are frequent [1]. P. aeruginosa, in particular, is well recognized opportunist pathogen in immunocompromised patients [3], with a mortality rate as high as 50% [4]. It is the most prevalent pathogen among patients with cystic fibrosis [5], and is more common in adults. S. maltophilia is an environmental bacterium that can cause respiratory-tract infections in humans [6]: 2% of hospitalized patients in intensive care units (ICU) develop colonization or infection [7], and at least 30% of patients with cystic fibrosis are colonized [8, 9]. Although S. maltophilia is not a highly virulent pathogen, it can lead to severe infections in immunocompromised patients, with a mortality rate up to 37.5% [10].

Fluoroquinolones are currently among the most heavily prescribed antimicrobials in the world because of their spectrum of activity, their pharmacokinetic profiles, and their generally good tolerance. The older narrow-spectrum ciprofloxacin is usually active against Gram-negative bacteria like Enterobacteriaceae, P. aeruginosa or S. maltophilia, but anaerobic bacteria and some Gram-positive bacteria like Enterococcus spp., Streptococcus spp. and Listeria spp. are naturally resistant. In the 2000s, new fluoroquinolones –levofloxacin and moxifloxacin– were developed, exhibiting enhanced potencies with very low MICs against Gram-positive organisms and/or anaerobes, while maintaining Gram-negative activity [1113]. They also demonstrated improved PK profiles, characterized by a better systemic distribution particularly in respiratory tract tissues and fluids [14].

This enhanced PK profiles result in large areas under the serum concentrations versus time curves (AUCs) and high peak concentrations. That, in combination with their low MICs, allows them to achieve optimal PK/PD parameters for both efficacy such as AUC0-24/MIC (AUC0-24 is the area under the concentration-time curve from 0 to 24 h), or prevention of de novo resistance like Cmax/MIC (Cmax is the maximum plasma concentration). AUC0-24/MIC ratio should be superior to 125: if this ratio is lower than 125, treatment failure increase by three to four, and if it is greater than 250, recovery time is enhanced [15, 16]. The Cmax/MIC ratio must reach at least 10 to 12 in order to prevent emergence of de novo resistant mutants [15, 16].

The purpose of the present work was to assess the respective potency of either ciprofloxacin, levofloxacin or moxifloxacin against K. pneumoniae, P. aeruginosa and S. maltophilia by (i) comparing their MICs against a large set of clinical strains of K. pneumoniae, P. aeruginosa and S. maltophilia; (ii) comparing their bactericidal activities through time-kill curves performed on selected strains of each bacteria.

Materials and Methods

Bacterial strains

120 clinical strains were obtain from the archived collection of laboratory of clinical microbiology of the University Hospital of Strasbourg was constituted as follows: 40 strains of K. pneumoniae, 40 strains of P. aeruginosa and 40 strains of S. maltophilia. All strains were stored at -80°C using MAST CRYOBANK (Mast diagnostic, Amiens, France).

MICs measurements

MICs were determined by the gradient strips method using E-test® (bioMérieux, France) on Mueller-Hinton agar inoculated with a standard inoculum (105 to 106 CFU/ml) according to The European Committee on Antimicrobial Susceptibility Testing (EUCAST) guidelines [17] (Table 1; Fig 1).

thumbnail
Fig 1. MICs distributions for K. pneumonia (A), P. aeruginosa (B), S. maltophilia (C) against ciprofloxacin, levofloxacin and moxifloxacin.

https://doi.org/10.1371/journal.pone.0156690.g001

Time-kill studies

Antibiotics were supplied by their respective manufacturers as pure titrated powders (ciprofloxacin and moxifloxacin by Bayer Healthcare SAS, levofloxacin by Sanofi-Aventis, France). Five strains of each species were selected for time-kill studies, based on their different susceptibility patterns to the antibiotics (Table 2). For each strain, time kill studies were performed for the three molecules at concentrations equal to the theoretical plasma peak (ciprofloxacin = 4 μg/mL; levofloxacin = 10 μg/mL; moxifloxacin = 3 μg/mL), then at concentrations equal to one fold and two fold the MIC of the antibiotic used. Bacteria were initially cultured for 24 h at 37°C on Mueller-Hinton agar. These cultures were then considered as being on stationary growth-phase and used to prepare exponential growth phase at standard inoculum (106 CFU/mL) in Mueller-Hinton broth (MHB, bioMérieux, France). The inoculum of 106 CFU/ml was obtained by standardizing optical density at 550 nm to 0.125 followed by a 1:100 dilution. Final suspensions of bacteria were supplemented with ciprofloxacin, levofloxacin or moxifloxacin at different concentrations and cultured for 24 h at 37°C. Culture aliquots of 100 ml were removed at 2, 4, 6 and 24 h and plated on agar for colony counts.

thumbnail
Table 2. MICs (μg/mL) of the bacteria strains tested by time-kill studies.

https://doi.org/10.1371/journal.pone.0156690.t002

Results

Strains MICs

MICs –as determined by E-test® for ciprofloxacin, levofloxacin and moxifloxacin– are presented in Fig 1. For all species and for each antibiotic tested, MIC50, MIC90 and modal MIC were calculated (Table 1). Clinical breakpoints used for categorization was those provided by EUCAST [17].

Among all K. pneumoniae strains, 75% were susceptible to ciprofloxacin and levofloxacin, and 67.5% to moxifloxacin. All ciprofloxacin-resistant strains were also resistant to the other quinolones tested (n = 9). The MICs distributions were bimodal for each antibiotic tested (Fig 1A). For each antibiotic, the MIC50 and the modal MIC were in the same order of magnitude, but the MIC90 was at least four-fold greater (Table 1).

Among all P. aeruginosa strains, 65% were susceptible to ciprofloxacin, 57.5% to levofloxacin, and 37.5% to moxifloxacin. Aside from one strain, all strains resistant to levofloxacin were also resistant to ciprofloxacin and moxifloxacin (n = 13). One strain was resistant to moxifloxacin (MIC = 3 μg/mL), but susceptible to ciprofloxacin (CMI = 0.19 μg/mL) and levofloxacin (MIC = 0.75 μg/mL). The MICs distributions were bimodal for each antibiotics tested (Fig 1B). Modal MICs and MICs90 of each antibiotic tested were all ≥32 μg/mL (Table 1).

Among all S. maltophilia strains, 15% were susceptible to ciprofloxacin, 35% to levofloxacin, and 52.5% to moxifloxacin. All strains intermediate or resistant to moxifloxacin were also resistant to the other quinolones tested (n = 19). Ciprofloxacin and levofloxacin MIC50 (4 and 1.5 μg/mL respectively) were higher than clinical lower breakpoint for susceptibility. Only moxifloxacin had a MIC50 (0.75 μg/mL) that was lower than the cut-off for resistance. MICs distribution was almost identical for each antibiotics tested (Fig 1C).

Time kill studies at plasma peak concentrations

K. pneumoniae.

Three strains among the 5 tested were susceptible to ciprofloxacin (strains 1, 3 and 4). Nevertheless, strains 1 and 4 had not have far lower MICs than strain 3 (50 and 15 times lower, respectively, Table 2). For these both strains, ciprofloxacin at peak concentration exerted a bactericidal activity, with a 2-log CFU/mL inoculum reduction in the first 6 h. A 2–6 h bacteriostatic activity was observed for the others strains. Then, regrowth occurred up to 24 h for all strains (Fig 2).

thumbnail
Fig 2. Time-kill studies with 5 strains of K. pneumoniae, P. aeruginosa, S. maltophilia with concentration equal to the theoretical plasma peak (ciprofloxacin = 4 μg/mL; levofloxacin = 10 μg/mL; moxifloxacin = 3 μg/mL).

https://doi.org/10.1371/journal.pone.0156690.g002

Four strains among the 5 tested were susceptible to levofloxacin (strain 1–4). Again, strains 1 and 4 had lower MICs than strains 2 and 3 (5 and 15 times lower, respectively, Table 2). A 4-5-log decrease was observed at 6 h for these strains. Bacteriostatic activity was observed for strain 1. A bacterial growth re-occurred at 24 h for strains 2 and 3 (resistant) (Fig 2).

Moxifloxacin exhibited only a 2–6 h bacteriostatic activity, whatever strains categorization, followed by regrowth up to 24 h. (Fig 2).

P. aeruginosa.

Four strains among the 5 tested were susceptible to ciprofloxacin (strains 1–4) and one was intermediate (strain 5) (Table 2). At least a 24 h 5-log decrease occurred for strains 1 and 3, which are the most susceptible according to their MICs. A bactericidal activity, with 2.5-log inoculum reduction at 6 h, against strains 2 and 4 was observed. A bacterial growth re-occurred at 24 h for strain 5 (resistant) (Fig 2).

Two strains among five were susceptible to levofloxacin (strains 1 and 3) and three were intermediate (strains 2, 4 and 5) (Table 2). All strains except one were killed by levofloxacin at 5-log decrease level. Strain 5, which is resistant (MIC = 1.5 mg/L) was subject to a regrowth starting at the 4th hour.

Two strains were susceptible to moxifloxacin (strains 1 and 3) and three were resistant to moxifloxacin (strains 2, 4 and 5) (Table 2). A bactericidal activity was observed for only one strain (strain 1), with a 2 log inoculum reduction at 24 h. For the others strains, whatever their MICs, a bacteriostatic activity was observed until 6 h, followed by regrowth (Fig 2).

S. maltophilia.

No bactericidal activity was observed excepted for strain 1 with levofloxacin (Fig 2). All antibiotics exhibited a static effect to all strains at 6 h.

Time-kill studies at one and two fold MICs

At concentrations equal to one (Fig 3A) or two fold (Fig 3B) MICs, the same profile was observed for all strains tested. A bacteriostatic activity was observed up to 6 h, followed by a regrowth at 24 h. No significant difference was observed between the antibiotics tested.

thumbnail
Fig 3. Time-kill studies with 5 strains of K. pneumoniae, P. aeruginosa, S. maltophilia with concentration equal to one fold (A) or two fold (B) the MIC of the strain against the considered antibiotic.

https://doi.org/10.1371/journal.pone.0156690.g003

Discussion

In a general manner, the bactericidal activity of ciprofloxacin and levofloxacin was inversely proportional to the MICs.

K. pneumoniae

For K. pneumonia and fluoroquinolones, the break points are well separated from the ecoffs (epidemiological cut-off). In our population, for each antibiotic tested, MIC50 and modal MICs were similar, meaning that the majority of the strains were susceptible. In fact, our results clearly show two distinct sub-populations: a susceptible-one with MICs close to MIC50, and a resistant-one with MICs closer to MIC90. Interestingly our data seems to fit with EUCAST-ones. E-coff and modal MIC provided by EUCAST are 0.125 μg/mL and 0.03 μg/mL for ciprofloxacin, 0.25 μg/mL and 0.06 μg/mL for levofloxacin/moxifloxacin, respectively (EUCAST data). The ecoff of our K. pneumoniae population is about 0.38 μg/mL, indicating that our strains may be representative of general population. The MIC50 of the three antibiotics are very low and close one to the other (0,047, 0,094 and 0,064 μ g/mL, respectively), suggesting a similar in vitro activity against our susceptible strains. Thus, on a bacteriostatic point of view, the three antibiotics are equivalent on K. pneumoniae. We confirm this through time-kill curves at concentrations of each antibiotic of one or two fold the MICs resulting in a bacteriostatic activity. Nevertheless, in terms of bactericidal activity at higher concentration, equal to peak concentrations, it has been shown that the population density expressed in log-CFU is inversely proportioned to MICs values.

Nevertheless, levofloxacin seems to be somewhat more bactericidal than ciprofloxacin. It is not likely that this difference may have clinical relevance. The bactericidal activity of ciprofloxacin was high for both strains with low MIC, when it was much lower for. strain 3 the MIC of which (0.38 μg/mL; susceptible) is close to the clinical breakpoint (0.5 μg/mL). This should be a key point when using MICs. A MIC value close to clinical breakpoint could indicate that the antibiotic could be less effective than what can be expected when the MIC is far below from breakpoint. The same holds true with levofloxacin. Among the susceptible 4 strains, both strains with very low MICs undergo a deep bactericidal effect, whereas the two strains with MICs values closer to clinical breakpoint do not. Another limit of MICs is highlighted by the results obtained for moxifloxacin. The three antibiotics had very close and very low MIC50 values (0.047, 0.094 and 0.064 μg/mL respectively), suggesting an equivalent in vitro good activity against susceptible strains. But, on the contrary, time-kill studies indicated an absence of bactericidal activity against K. pneumoniae.

P. aeruginosa

In our study, the modal MICs and MICs90 of our total population were identical (>32 mg/mL) for all molecules, showing that the majority of strains included were resistant to these antibiotics. In fact, our results show two distinct subpopulations: a susceptible one, the smallest, with modal MICs for ciprofloxacin, levofloxacin and moxifloxacin equal to 0.19, 0.38 and 1 to 1.5 μg/mL respectively and a larger resistant one. EUCAST e-coff and modal MIC are 0.5 μg/mL and 0.12 μg/mL for ciprofloxacin, 2 μg/mL and 0.5 μg/mL for levofloxacin, and 4 μg/mL and 1 μg/mL for moxifloxacin, respectively. In our study, the MICs50 in susceptible population are close to the modal MICs provided by EUCAST data. The lower clinical breakpoints for ciprofloxacin and levofloxacin are 1 μg/mL, being far higher from our modal MIC of the susceptible population. But moxifloxacin breakpoints at 1 μ g/mL is very close to modal MICs of our susceptible population and cover a part of wild strains. This correlates with EUCAST data, with an e-coff equal to 4 μg/mL, higher than clinical breakpoints.

In terms of bactericidal activity, at plasma peak concentration, ciprofloxacin and levofloxacin are equivalent on 2/5 strains tested (6-log reduction), levofloxacin is superior to ciprofloxacin on 2/4 strains tested (6-log vs 2-log reduction), and ciprofloxacin and levofloxacin have no bactericidal activity on 1/5 strain tested. As has been shown with K. pneumoniae, the lower the MIC the better the bactericidal activity. For ciprofloxacin at 24 h, the best (6 log-decrease) killing activity was observed for both strains with the lower MICs. The same occurs for levofloxacin, with a 6-log decrease for strain 1, 2 and 4. Ciprofloxacin, as already shown in the literature, seems to have a good efficiency against susceptible P. aeruginosa strains [1821]. However, levofloxacin seems to have a better activity against P. aeruginosa than ciprofloxacin. For the five strains tested by time-kill studies, a bactericidal effect of levofloxacin was observed at 6 h, although some of these strains (strain 2, 4 and 5) were intermediate to levofloxacin, and a total bacterial killing was observed at 24 h for 4/5 strains. It is not likely that this difference may have clinical relevance. Gillespie et al. have shown that levofloxacin less frequently favours resistant mutants appearance than ciprofloxacin, but no difference between the two molecules was observed in their time-kill studies [18]. The few number of tested strains by our time kill studies (n = 5) does not allow us to affirm a real superiority of levofloxacin on ciprofloxacin against P. aeruginosa.

S. maltophilia

In our study, levofloxacin and moxifloxacin MICs50 and modal MICs are significantly lower than those of ciprofloxacin. Moxifloxacin MIC90 was significantly lower than levofloxacin and ciprofloxacin. Moxifloxacin seems therefore to have a better in vitro activity. EUCAST does not provide e-coff for S. maltophilia and quinolones. The only available data are modal MICs: ciprofloxacin 2 μg/mL, 1 μg/mL for levofloxacin and 0.25 μg/mL for moxifloxacin (EUCAST data). In comparison, our strains are more resistant to ciprofloxacin but have the same profile for levofloxacin and moxifloxacin as worldwide strains. For those three antibiotics, clinical breakpoints are very close to modal MICs of population and cover a part of wild strains.

Time kill studies show a little activity of fluoroquinolones against S. maltophilia. A bactericidal activity was only observed for one strain (2-log reduction) with levofloxacin. For the other strains, a bacteriostatic activity, followed or not by regrowth was observed.

Conclusion

In our study, the bactericidal activity has been shown to be inversely proportioned to MICs values. A “susceptible” MIC value close to clinical breakpoint could indicate that the antibiotic could be less effective than what can be expected from a strain with a much lower MIC far from the clinical breakpoint. Our data go in the same meaning as Torres et al. who have recently shown that “highly susceptible” isolates are associated with higher clinical cure rates than “borderline isolates”[22].This consideration should be taken into account when choosing between different antibiotic that are all “susceptible”.

Author Contributions

Conceived and designed the experiments: FJ. Performed the experiments: AG FS MK. Analyzed the data: AG FS MK FJ. Contributed reagents/materials/analysis tools: AG FS MK FJ. Wrote the paper: AG FS FJ.

References

  1. 1. Podschun R, Ullmann U. Klebsiella spp. as nosocomial pathogens: epidemiology, taxonomy, typing methods, and pathogenicity factors. Clinical microbiology reviews. 1998 Oct;11(4):589–603. pmid:9767057
  2. 2. Lye WC, Chan RK, Lee EJ, Kumarasinghe G. Urinary tract infections in patients with diabetes mellitus. The Journal of infection. 1992 Mar;24(2):169–74. pmid:1569307
  3. 3. Silby MW, Winstanley C, Godfrey SA, Levy SB, Jackson RW. Pseudomonas genomes: diverse and adaptable. FEMS microbiology reviews. 2011 Jul;35(4):652–80. pmid:21361996
  4. 4. Wisplinghoff H, Bischoff T, Tallent SM, Seifert H, Wenzel RP, Edmond MB. Nosocomial bloodstream infections in US hospitals: analysis of 24,179 cases from a prospective nationwide surveillance study. Clin Infect Dis. 2004 Aug 1;39(3):309–17. pmid:15306996
  5. 5. Sousa AM, Pereira MO. Pseudomonas aeruginosa Diversification during Infection Development in Cystic Fibrosis Lungs-A Review. Pathogens (Basel, Switzerland). 2014;3(3):680–703.
  6. 6. Brooke JS. Stenotrophomonas maltophilia: an emerging global opportunistic pathogen. Clinical microbiology reviews. 2012 Jan;25(1):2–41. pmid:22232370
  7. 7. Nseir S, Di Pompeo C, Brisson H, Dewavrin F, Tissier S, Diarra M, et al. Intensive care unit-acquired Stenotrophomonas maltophilia: incidence, risk factors, and outcome. Critical care (London, England). 2006;10(5):R143.
  8. 8. Ballestero S, Virseda I, Escobar H, Suarez L, Baquero F. Stenotrophomonas maltophilia in cystic fibrosis patients. Eur J Clin Microbiol Infect Dis. 1995 Aug;14(8):728–9. pmid:8565998
  9. 9. Steinkamp G, Wiedemann B, Rietschel E, Krahl A, Gielen J, Barmeier H, et al. Prospective evaluation of emerging bacteria in cystic fibrosis. J Cyst Fibros. 2005 Mar;4(1):41–8. pmid:15752680
  10. 10. Falagas ME, Kastoris AC, Vouloumanou EK, Rafailidis PI, Kapaskelis AM, Dimopoulos G. Attributable mortality of Stenotrophomonas maltophilia infections: a systematic review of the literature. Future microbiology. 2009 Nov;4(9):1103–9. pmid:19895214
  11. 11. Hooper DC. Mechanisms of action and resistance of older and newer fluoroquinolones. Clin Infect Dis. 2000 Aug;31 Suppl 2:S24–8. pmid:10984324
  12. 12. North DS, Fish DN, Redington JJ. Levofloxacin, a second-generation fluoroquinolone. Pharmacotherapy. 1998 Sep-Oct;18(5):915–35. pmid:9758306
  13. 13. Balfour JA, Wiseman LR. Moxifloxacin. Drugs. 1999 Mar;57(3):363–73; discussion 74. pmid:10193688
  14. 14. O'Donnell JA, Gelone SP. Fluoroquinolones. Infectious disease clinics of North America. 2000 Jun;14(2):489–513, xi. pmid:10829268
  15. 15. Schentag JJ. Clinical pharmacology of the fluoroquinolones: studies in human dynamic/kinetic models. Clin Infect Dis. 2000 Aug;31 Suppl 2:S40–4. pmid:10984327
  16. 16. Lode H, Borner K, Koeppe P. Pharmacodynamics of fluoroquinolones. Clin Infect Dis. 1998 Jul;27(1):33–9. pmid:9675446
  17. 17. EUCAST. Clinical breakpoints. 2015 [cited 2015 August 06]; Available from: http://www.eucast.org/clinical_breakpoints/
  18. 18. Gillespie T, Masterton RG. Investigation into the selection frequency of resistant mutants and the bacterial kill rate by levofloxacin and ciprofloxacin in non-mucoid Pseudomonas aeruginosa isolates from cystic fibrosis patients. International journal of antimicrobial agents. 2002 May;19(5):377–82. pmid:12007845
  19. 19. Hodson ME, Roberts CM, Butland RJ, Smith MJ, Batten JC. Oral ciprofloxacin compared with conventional intravenous treatment for Pseudomonas aeruginosa infection in adults with cystic fibrosis. Lancet (London, England). 1987 Jan 31;1(8527):235–7.
  20. 20. Hoiby N, Pedersen SS, Jensen T, Valerius NH, Koch C. Fluoroquinolones in the treatment of cystic fibrosis. Drugs. 1993;45 Suppl 3:98–101. pmid:7689459
  21. 21. Valerius NH, Koch C, Hoiby N. Prevention of chronic Pseudomonas aeruginosa colonisation in cystic fibrosis by early treatment. Lancet (London, England). 1991 Sep 21;338(8769):725–6.
  22. 22. Torres E, Delgado M, Valiente A, Pascual A, Rodriguez-Bano J. Impact of borderline minimum inhibitory concentration on the outcome of invasive infections caused by Enterobacteriaceae treated with beta-lactams: a systematic review and meta-analysis. Eur J Clin Microbiol Infect Dis. 2015 Sep;34(9):1751–8. pmid:26032669