Improved isolation of Stenotrophomonas maltophilia from the sputa of patients with cystic fibrosis using a selective medium


The prevalence of Stenotrophomonas maltophilia in cystic fibrosis (CF) patients has increased since the early 1980s, reaching over 30% in some centers [1]. The reasons for this are unclear, but may relate to changes in laboratory reporting practices, culture techniques [2], and the selective pressure exerted by antipseudomonal antimicrobials [3]. Although the clinical significance of this bacterium for patients with CF is disputed, there have been anecdotal reports of gradual clinical deterioration in patients chronically colonized with bacterial numbers of ≥ 105 CFU/mL over several years [4]. The use of selective media has been shown to increase the isolation rates of S. maltophilia from clinical and environmental samples [5,6]. The introduction of blood agar containing imipenem for routine sputum culture at one CF center in the UK saw the prevalence of S. maltophilia increase to 16% from 8% the year before [2]. Unfortunately this study did not compare directly the sensitivity of the selective medium with a non-selective medium. For this reason, we studied the sensitivity of a selective medium (VIA medium, incorporating vancomycin, imipenem and amphotericin B as selective agents) [5] for isolating S. maltophilia from sputum samples collected from children with CF in comparison to an existing in-house method that utilized an imipenem disk placed upon bacitracin–chocolate agar (BC medium).

Between April 1994 and January 1995, 814 sputum samples from 87 of the 140 full-time patients who attended the Regional Paediatric Cystic Fibrosis Unit during the study period were analyzed. Each sample was mixed with an equal volume of a sputum homogenizing agent (Sputasol; Oxoid Ltd, Basingstoke, UK) and 10 µL applied to half of a VIA medium plate [5]. Briefly, this consisted of a mannitol agar base (Mast Laboratories, Bootle, UK) containing bromothymol blue as an indicator (S. maltophilia does not produce acid from mannitol), to which 5 mg/L of vancomycin (Eli Lilly, Basingstoke, UK), 32 mg/L of imipenem (MSD, Hoddesdon, UK) and 4 mg/L of amphotericin B (Bristol-Myers Squibb, Hounslow, UK) were added as selective agents [5]. The sputum/homogenizer mixture was diluted 1 : 50 with sterile water; 10 µL was applied to the other half of the VIA plate, and 10 µL to bacitracin (10 000 U/L) chocolate (BC) medium. A 10-µg imipenem disk was also placed onto the surface of the BC agar. All plates were incubated at 37 °C in air for 48 h. Plates that exhibited no growth at 48 h were incubated for a further 24 h. The number of colonies growing on VIA medium was counted to give an approximate number of CFU/mL. These colonies and any exhibiting imipenem resistance on BC medium were subcultured onto non-selective media for purity and identified using the API 20NE system (bioMérieux, Marcy l'Etoile, France) according to the manufacturer's guidelines. Statistical analysis of the results was performed using the chi-squared test.

The comparative performance of VIA medium and BC medium for the detection of S. maltophilia is shown in Table 1. VIA medium detected a significantly greater number of samples that were positive for S. maltophilia than did BC medium with an imipenem disk placed upon it. The relationship between the colony count on VIA medium and successful detection on BC medium is shown in Figure 1. VIA medium was more sensitive than BC medium, particularly when the number of CFU/mL in samples was below 106.

Table 1.  Comparative performance of VIA medium and BC medium for isolation of S. maltophilia from sputum samples taken from CF patients
  • a

    P < 0.0001.

VIA+106129 a235VIA+268 a34
Figure 1.

Samples positive for S. maltophilia on VIA medium: relationship between colony count and detection on BC medium.

Other carbapenem-resistant Gram-negative bacilli were also isolated during the study, although they were clearly distinguishable from S. maltophilia by their appearance on VIA medium. Eighty-nine samples (from 16 patients) grew Pseudomonas aeruginosa, and 21 samples (two patients) grew Burkholderia cepacia.

This study highlights the improved detection of S. maltophilia from CF sputum samples using VIA as a selective medium. In the case of CF sputum, the use of VIA medium inhibits most strains of P. aeruginosa, the most significant respiratory pathogen in this patient group. Although 89 of the samples (from 16 patients) grew imipenem-resistant P. aeruginosa, this organism was easy to distinguish from S. maltophilia on VIA medium by its colonial appearances. The increased sensitivity of VIA medium related to its ability to detect low numbers of S. maltophilia more readily than BC medium ( Figure 1). Of those samples positive on BC medium, 96% were present at ≥ 105 CFU/mL. Although the clinical significance of this organism in CF is uncertain, it has been associated with clinical deterioration in some patients, particularly if counts persistently excede 106 CFU/mL [4,7]. Even at this level, BC medium failed to grow S. maltophilia from 10% of samples that were positive on VIA medium ( Figure 1). The increased sensitivity of VIA medium also allowed earlier detection of S. maltophilia in patient samples. Three patients were positive for S. maltophilia on VIA medium for 1–4 months before the isolation of S. maltophilia on BC medium. Indeed, since the end of the study, the use of enterobacterial repetitive intergenic consensus sequences for typing [8] has confirmed that a strain of S. maltophilia isolated on BC medium had been isolated 3 years earlier on VIA medium during the study period. These findings suggest that patients may be colonized for months or even years with S. maltophilia before the organism is detected with conventional culture methods. The role of S. maltophilia in the natural history of pulmonary disease in patients with CF remains undetermined. VIA medium will be a valuable tool in our attempts to understand the role of this emerging organism in CF.


We would like to thank all the members of the laboratory staff at the Microbiology Department of St James's University Hospital, Leeds, who helped to process the sputum samples during this study. This work was previously presented at the 20th European Cystic Fibrosis Conference, Brussels, Belgium, 18–21 June 1995.