Abstract
We pretreated with SDS 71 urine samples with bacterial counts of >105 CFU/ml and matrix-assisted laser desorption ionization–time of flight (MALDI-TOF) identification scores of <2, in order to minimize failure rates. Identification improved in 46.5% of samples, remained unchanged in 49.3%, and worsened in 4.2%. The improvement was more evident for Gram-negative (54.3%) than for Gram-positive (32%) bacteria.
TEXT
Urinary tract infections (UTIs) are among the most common human bacterial infections (1).Tests developed for UTI screening include urine dipstick testing, urinalysis, and Gram staining. The urine culture remains the “gold standard,” but the use of this method, without any previous screening procedure, is time-consuming and expensive, because of the high cost of unnecessary testing of negative samples (2).
Matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF MS) has been shown as a fast and reliable method for bacterial identification both from culture plates (3, 4) and from blood cultures vials (5, 6), and also from some other samples, such as infected urine, especially when Gram-negative bacteria with high bacteria counts are involved (7).
We described recently a procedure for processing urine samples (7) which begins with a centrifugation step (2,000 × g for 30 s) to remove leukocytes. Thus, the 7 to 8% of identification failures reported in culture-positive samples might be associated with the removal of intraleukocytic microorganisms in these first steps of processing. We have developed a further study, including a sample pretreatment with SDS, because this compound can lysate cells and then release microorganisms, thereby increasing method sensitivity.
Processing of urine samples.
We analyzed 71 urine clinical samples with bacterial counts of ≥105 CFU/ml on blood agar after 18 h of incubation in an aerobic atmosphere at 37°C and with a score of <2 in MALDI-TOF MS identification. Samples showing mixed cultures were discarded.
Routine urine samples were initially processed for direct microorganism identification with MALDI-TOF MS, according to previously reported methods (7). An aliquot of each sample was stored at 4°C. When the bacterial count was ≥105 CFU/ml and the MALDI-TOF MS identification score value was <2, the stored aliquot was spread again on blood agar for checking that bacterial count had not changed significantly (count modifications of <5% were considered acceptable) and processed again for MALDI-TOF MS identification, after SDS pretreatment, always before 24 h of storage. Comparison between improvement rates in Gram-positive and Gram-negative bacteria was performed by using the Fisher exact test with the mid-P method. Statistical significance was considered when the P value was <0.01.
Differential procedure. (i) MALDI-TOF MS.
Samples were processed as described before (7). Briefly, urine (3 ml) was centrifuged at 2,000 × g for 30 s to remove leukocytes. The supernatant was centrifuged at 15,500 × g for 5 min to collect bacteria. The pellet was washed once with deionized water.
(ii) MALDI-TOF MS with SDS pretreatment.
A total of 600 μl SDS 10% was added to 3 ml of urine and vortexed for 2 min. After 5 min of repose on the benchtop, the samples were centrifuged at 15,500 × g for 5 min to collect bacteria. The pellet was washed once with deionized water and transferred to a new tube.
Common procedure.
The pellets from samples processed by these two ways were centrifuged at 15,500 × g for 5 min and underwent ethanol-formic acid extraction and MALDI-TOF MS as described previously (7).
No samples showed significant changes of bacterial count between the first spread and the spread of the stored aliquot. Results on whole identification improvement appear in Table 1. Identification reliability increased in 33/71 samples (46.5%). In 22 samples (31%), identification reliability improved by more than one step, from “no reliable identification” (score of <1.7) or “no peaks found” to “identification reliable to the species level” (score of ≥2). Identification level remained unchanged or had changes with no repercussion for identification purposes (from “no peaks found” to “no reliable identification” or vice versa) in 35 samples (49.3%) and decreased in only 3 samples (4.2%).
TABLE 1.
Identification (ID) modifications observed with the SDS method
| Conventional method result | No. (%) of samples with the indicated result by conventional methods with SDS pretreatment |
|||
|---|---|---|---|---|
| No peaks found | ID not reliable | ID genus level | ID species level | |
| No peaks found | 29 (40.8)b | 2 (2.8)b | 7 (9.9)c | 17 (23.9)c |
| ID not reliable | 1 (1.4)b | 1 (1.4)b | 0 | 5 (7)c |
| ID genus level | 2 (2.8)a | 1 (1.4)a | 2 (2.8)b | 4 (5.6)c |
ID level worsened.
ID level unmodified.
ID level improved.
Improvement was more frequent in urine samples infected with Gram-negative than Gram-positive bacteria. In whole, identification improved in 25/46 (54.3%) samples infected with Gram-negative bacteria and in 8/25 (32%) samples infected with Gram-positive bacteria. Nevertheless, no statistical significance was obtained with the Fisher exact test with mid-P method (P = 0.039). Behavior by species is shown in Table 2.
TABLE 2.
Modification of identification level by species in 71 specimens
| Species (no. of samples) | Conventional method result | No. (%) of samples with the indicated result by conventional methods with SDS pretreatmentd |
No. of samples (%) with whole identification improvementa | |||
|---|---|---|---|---|---|---|
| No peaks found | ID not reliable | ID genus level | ID species level | |||
| E. coli (34) | No peaks found | 13 (38.2)c | 1(2.9)c | 5 (14.7)b | 10 (29.4)b | 19 (55.9) |
| ID not reliable | 0 | 0 | 0 | 0 | ||
| ID genus level | 0 | 0 | 1 (2.9)c | 4 (11.8)b | ||
| E. faecalis (16) | No peaks found | 10 (62.5)c | 0 | 1 (6.3)b | 2 (12.5)b | 6 (37.6) |
| ID not reliable | 0 | 0 | 1 (6.3)b | 2 (12.5)b | ||
| ID genus level | 0 | 0 | 0 | 0 | ||
| Pseudomonas aeruginosa (4) | No peaks found | 0 | 0 | 1 (25)b | 2 (50)b | 4 (100) |
| ID not reliable | 0 | 0 | 0 | 0 | ||
| ID genus level | 0 | 0 | 0 | 1 (25)b | ||
| Streptococcus agalactiae (3) | No peaks found | 1 (33.3)c | 0 | 0 | 0 | 0 (0) |
| ID not reliable | 0 | 0 | 0 | 0 | ||
| ID genus level | 1 (33.3)c | 1 (33.3)c | 0 | 0 | ||
| K. pneumoniae (2) | No peaks found | 0 | 1 (50)c | 0 | 1 (50)b | 1 (50) |
| ID not reliable | 0 | 0 | 0 | 0 | ||
| ID genus level | 0 | 0 | 0 | 0 | ||
| Staphylococcus aureus (2) | No peaks found | 0 | 0 | 0 | 0 | 1 (50) |
| ID not reliable | 1 (50%)c | 0 | 0 | 1 (50)b | ||
| ID genus level | 0 | 0 | 0 | 0 | ||
| Staphylococcus saprophyticus (2) | No peaks found | 2 (100)c | 0 | 0 | 0 | 0 (0) |
| ID not reliable | 0 | 0 | 0 | 0 | ||
| ID genus level | 0 | 0 | 0 | 0 | ||
| Providencia stuartii (1) | No peaks found | 0 | 0 | 0 | 0 | 0 (0) |
| ID not reliable | 0 | 0 | 0 | 0 | ||
| ID genus level | 1 (100)c | 0 | 0 | 0 | ||
| Raoultella ornithinolytica (1) | No peaks found | 0 | 0 | 0 | 0 | 0 (0) |
| ID not reliable | 0 | 0 | 0 | 0 | ||
| ID genus level | 0 | 0 | 1 (100)c | 0 | ||
| Salmonella species (1) | No peaks found | 0 | 0 | 0 | 0 | 1 (100) |
| ID not reliable | 0 | 0 | 1 (100)b | 0 | ||
| ID genus level | 0 | 0 | 0 | 0 | ||
| Klebsiella oxytoca (1) | No peaks found | 1 (100)c | 0 | 0 | 0 | 0 (0) |
| ID not reliable | 0 | 0 | 0 | 0 | ||
| ID genus level | 0 | 0 | 0 | 0 | ||
| Enterobacter cloacae (1) | No peaks found | 1 (100)c | 0 | 0 | 0 | 0 (0) |
| ID not reliable | 0 | 0 | 0 | 0 | ||
| ID genus level | 0 | 0 | 0 | 0 | ||
| P. mirabilis (1) | No peaks found | 0 | 0 | 0 | 0 | 0 (0) |
| ID not reliable | 0 | 1 (100)c | 0 | 0 | ||
| ID genus level | 0 | 0 | 0 | 0 | ||
| Enterococcus faecium (1) | No peaks found | 0 | 0 | 0 | 1 (100)b | 1 (100) |
| ID not reliable | 0 | 0 | 0 | 0 | ||
| ID genus level | 0 | 0 | 0 | 0 | ||
| Streptococcus pyogenes (1) | No peaks found | 1 (100)c | 0 | 0 | 0 | 0 (0) |
| ID not reliable | 0 | 0 | 0 | 0 | ||
| ID genus level | 0 | 0 | 0 | 0 | ||
| Gram-negative bacteria | No peaks found | 15 (41.7)c | 2 (5.6)c | 6 (16.7)b | 13 (36.1)b | 25 (54.3) |
| ID not reliable | 0 | 1 (50)c | 1 (50)b | 0 | ||
| ID genus level | 1 (12.5)c | 0 | 2 (25)c | 5 (62.5)b | ||
| Gram-positive bacteria | No peaks found | 14 (77.8)c | 0 | 1 (5.6)b | 3 (16.7)b | 8 (32) |
| ID not reliable | 1 (20)c | 0 | 1 (20)b | 3 (60)b | ||
| ID genus level | 1 (50)c | 1 (50)c | 0 | 0 | ||
| Total | 33 (46.5) | |||||
Proportion of samples increasing identification level. For this purpose, “no peaks found” and “ID not reliable” were considered as the same level.
ID level improves.
ID level does not improve.
“0” indicates no specimens in the category.
In a previous study by our group (7), MALDI-TOF MS identified correctly the etiologic microorganism to the species level in 91.8% of urine samples when the bacterial count was >105 CFU/ml, while reliability dropped sharply with lower bacterial counts.
Studies with experimental inocula confirmed this threshold. Minimal bacterial counts ranging between 8 × 104 and 1.5 × 105 CFU/ml were necessary for obtaining good protein profiles. A recent study with a very similar design reported similar bacterial thresholds (8). Inocula required for obtaining MALDI-TOF MS score values of ≥2.0 ranged between 6 × 104 CFU/ml (Proteus mirabilis and Klebsiella pneumoniae) and 1 × 106 (Enterococcus faecalis). Obtaining scores of ≥2.0 for Escherichia coli required bacterial counts of ≥5 × 105. Results on MALDI-TOF MS identification from real urine samples were also around 90%, and the authors also concluded that samples with bacterial counts of <105 CFU/ml are not reliable for MALDI-TOF MS identification.
Sample pretreatment with SDS, a compound that lysates cell membranes, might release microorganisms and increase method sensitivity. Similar methods have been proposed for MALDI-TOF MS direct identification from blood cultures, based on similar theoretical considerations, yielding identification improvement rates of 11% to the genus level and 6% to the species level (9).
The new method allowed a correct identification of 46% of samples that could not be identified with the first described method. Accepting a 92 to 95% level of correct identifications using the standard method, according to our previous results (7), the new method applied on the preliminary identification failures would mean a 2.5 to 4% level of correct identifications, reaching figures around 96 to 97.5%. As happened with the standard method, the results are better for Gram-negative microorganisms. Nevertheless, 32% of urine samples harboring Gram-positive bacteria that had not been identified by the standard method were identified with the SDS pretreatment.
According to our whole results, urine samples reported as presumptively positive by the screening might be directly processed by MALDI-TOF MS, and those for which MALDI-TOF MS does not give a reliable identification are tested again with SDS extraction. This scheme would report the identification of urine tract pathogens soon after receiving the sample in >95% of cases. This might allow for better adjusted empirical treatment in many community-acquired UTIs. Obviously, moderate to severe infections would still require antibiogram in the usual way. Even so, the fast identification of the pathogen may be useful in these patients. The repercussion on the laboratory workflow, the impact in the patient care, and the cost/benefit rate are, in our opinion, the points that need to be studied in the future.
The study did not include any mixed samples. In a previous study (5) of three mixed samples, two did not lead to any reliable identification by MALDI-TOF MS, but one led to E. coli identification. Results obtained with mixed samples probably depend on two factors: the bacterial count of each single population and the proportion between populations. Populations with low bacterial counts would not be detected, as happens in monobacterial infections. Thus, these populations would be ignored, both if they are contaminant and if they are potential pathogens. On the other hand, these populations probably would not interfere with identification of microorganisms with high bacterial count. When we have two or more populations with high bacterial counts, it leads to bizarre protein profiles that would not match with any profile in the database. Nevertheless, newer software versions seem to be able to identify mixed cultures, and thus this problem might be avoided, should both populations attain bacterial counts high enough to be reliably detected and identified.
In conclusion, we have developed a new method, based on SDS pretreatment of urine samples, which allows the identification of urinary tract pathogens directly by MALDI-TOF MS in samples that, while having high bacterial counts, had not been identified by using the standard method. This method allows for direct identification in 46% of the samples with bacterial counts of >105 CFU/ml in which MALDI-TOF MS had not reached a reliable pathogen identification. The improvement is higher for Gram-negative bacteria but affects both Gram-positive and Gram-negative microorganisms. Since >90% of urinary tract pathogens with high bacterial counts are correctly identified by using conventional MALDI-TOF MS as previously described (5), in our opinion this method should be restricted to screening-positive, conventional MALDI-TOF MS procedure-negative samples.
Footnotes
Published ahead of print 13 November 2013
REFERENCES
- 1.Foxman B, Barlow R, D'Arcy H, Gillespie B, Sobel JD. 2000. Urinary tract infection: self-reported incidence and associated costs. Ann. Epidemiol. 10:509–515. 10.1016/S1047-2797(00)00072-7 [DOI] [PubMed] [Google Scholar]
- 2.Deville WL, Yzermans JC, van Duijn NP, Bezemer PD, van der Windt DA, Bouter LM. 2004. The urine dipstick test useful to rule out infections. A meta-analysis of the accuracy. BMC Urol. 4:4. 10.1186/1471-2490-4-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Seng P, Drancourt M, Gouriet F, La Scola B, Fournier PE, Rolain JM, Raoult D. 2009. Ongoing revolution in bacteriology: routine identification of bacteria by matrix-assisted laser desorption ionization time-of-flight mass spectrometry. Clin. Infect. Dis. 49:543–551. 10.1086/600885 [DOI] [PubMed] [Google Scholar]
- 4.Ferreira L, Vega S, Sánchez-Juanes F, Gonzalez M, Herrero A, Muñiz MC, González-Buitrago JM, Muñoz-Bellido JL. 2010. Identifying bacteria using a matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) mass spectrometer. Comparison with routine methods used in clinical microbiology laboratories. Enferm. Infecc. Microbiol. Clin. 28:492–497. 10.1016/j.eimc.2009.12.009 [DOI] [PubMed] [Google Scholar]
- 5.Ferreira L, Sánchez-Juanes F, Porras-Guerra I, García-García MI, García-Sanchez JE, González-Buitrago JM, Muñoz-Bellido JL. 2011. Microorganisms direct identification from blood culture by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. Clin. Microbiol. Infect. 17:546–551. 10.1111/j.1469-0691.2010.03257.x [DOI] [PubMed] [Google Scholar]
- 6.La Scola B, Raoult D. 2009. Direct identification of bacteria in positive blood culture bottles by matrix-assisted laser desorption ionisation time-of-flight mass spectrometry. PLoS One 4:e8041. 10.1371/journal.pone.0008041 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Ferreira L, Sánchez-Juanes F, González-Ávila M, Cembrero-Fuciños D, Herrero-Hernández A, González-Buitrago JM, Muñoz-Bellido JL. 2010. Direct identification of urinary tract pathogens from urine samples by matrix-assisted laser desorption ionization-time of flight mass spectrometry. J. Clin. Microbiol. 48:2110–2115. 10.1128/JCM.02215-09 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Wang XH, Zhang G, Fan YY, Yang X, Sui WJ, Lu XX. 2013. Direct identification of bacteria causing urinary tract infections by combining matrix-assisted laser desorption ionization-time of flight mass spectrometry with UF-1000i urine flow cytometry. J. Microbiol. Methods 92:231–235. 10.1016/j.mimet.2012.12.016 [DOI] [PubMed] [Google Scholar]
- 9.Saffert RT, Cunningham SA, Madrekar J, Patel R. 2012. Comparison of three preparatory methods for detection of bacteremia by MALDI-TOF mass spectrometry. Diagn. Microbiol. Infect. Dis. 73:21–26. 10.1016/j.diagmicrobio.2012.01.010 [DOI] [PubMed] [Google Scholar]