Urinary tract infections (UTIs) are a common infection that can affect the urethra, urinary bladder, ureter, or kidneys. The majority of UTIs are not serious, but some can lead to a potentially life-threatening complication such as sepsis. The treatment of UTIs is complicated by the increasing prevalence and spectrum of antimicrobial resistance [1
]. In current practice, empirical antibiotics are used before a bacterial-culture result is available. The used antibiotic is adjusted according to the identified pathogen and the obtained antibiotic susceptibility test. However, a conventional bacterial culture takes at least 24 h to return results, and even longer to obtain the result of an antibiotic susceptibility test [4
]. Therefore, a more efficient way for bacteria identification and antibiotic susceptibility is necessary for patient treatment. A quick and accurate pathogen identification and antibiotic use can shorten treatment course and avoid bacteremia formation.
The serology method, genotyping, matrix-assisted laser desorption-time of flight mass spectrometry (MALDI-TOF-MS), and nanochips have been exploited recently to shorten the time needed for pathogen identification [5
]. However, when bacterial concentrations in the specimens are too low for pathogen detection, these methods are restricted. Some of the above-mentioned bacterial identification methods [5
] may require a bacterial culture before bacteria identification. Few of these methods can directly verify antibiotic susceptibility.
Raman spectroscopy is a spectroscopic technique that uses observe-vibrational or -rotational modes to provide information on molecular vibrations and crystal structures [7
]. Raman spectroscopy uses a laser light source to irradiate the sample and then generates Raman-scattered light [8
]. The shifting energy gives information about vibrational modes in the system. The Raman spectrum (Raman scattering intensity) depends on the vibrational and rotational states of the molecules. Surface-enhanced Raman spectroscopy (SERS) is a surface-sensitive technique that enhances Raman scattering by molecule adsorbed on rough metal surface or nanostructure [9
]. SERS increases the Raman scattering signal intensity by an enhancement factor of 1010
. SERS using nanoparticles as Raman substrates for molecule detection is an application of Nano Chip technology.
Recently, nanochip technology has applied successfully in antibiotic susceptibility. Liu et al. used SERS to detect isolated bacteria from blood culture, which was subsequently cultured in a tripticase soy broth (TSB) medium [10
]. They were able to find the minimal inhibitory concentrations of antibiotics for these bacteria. The SERS method is also limited by its disturbance of sample impurity substances, such as protein or white blood cells (WBCs) that can compromise detection sensitivity. Human urine samples, similar to peritoneal dialysate from peritoneal dialysis patients, have the advantage of including much fewer proteins than blood, which can prevent the interference of detection from these impurities [11
]. As concentrations of bacterial pathogens in biosamples may be lower than the detection limit, a larger amount of samples and repeated centrifuge are necessary to increase bacterial concentration. Urine also has the advantage of a sufficient sample amount to be exempt from the procedure to enrich bacterial amounts. For example, the bacterial culture test is more time-consuming than SERS chip detection.
In this study, we used a cylindrical SERS made up of silver nanoparticles coated on the tip of a 2 mm polymethylmethacrylate rod (AC). We used these SERS chips to detect pathogens from the urine samples of UTI patients.
In this study, we used the Raman SERS technique to detect pathogens in the urine of UTI patients. We could identify bacterial pathogens in most urine samples after simple urine-sample centrifugation (i.e., 700 rpm for 10 min). In some cases, urine bacteria could be found only after repeated sample concentration.
Raman SERS is a culture-free method for pathogen identifications. SERS can enhance Raman scattering intensity 1010
- to 1011
-fold, which makes it possible to detect bacteria in samples without a predetection culture [13
]. Therefore, the SERS technique has been used to detect bacterial infections in several studies [14
] and has the advantage of quick, within-minutes diagnosis of the pathogen. Fast and correct pathogen diagnosis can shorten the treatment time, avoid unnecessary patient complications, and reduce treatment costs [18
]. Urine is an aqueous sample, which frequently leads to the indeterminate spreading of samples loaded on a plate-shaped SERS chip, and causes poor reproducibility of study results [19
]. Water in the fluid has to be evaporated to allow the contact of bacteria with the Raman substance coated on the chips [19
]. In this study, we used cylindrical SERS chips that can easily make spontaneous contact of urine bacteria with the SERS substrate, and enhance detection sensitivity and reproducibility [6
By comparing the Raman spectrum of standard bacteria with the Raman spectrum obtained from patient urine, we can predict the urine pathogen. With the assistance of the recognition software, pathogen identification becomes much easier and time-saving [11
]. The amplitude of the Raman signal peak can be affected by bacterial concentration [20
]. We cannot discriminate between different bacteria by SERS intensities as was shown in Figure 4
; E. coli
and E. coli ESBL
cannot be differentiated simply by the signal intensities of similar peaks.
Methicillin-resistant Staphylococcus aureus
, vancomycin-resistant Enterococci
), and extended-spectrum beta-lactamase (ESBL)-producing bacterial are the common multiple-drug-resistant bacteria seen in UTIs [21
]. Multiple drug resistance (MDR) increases the mortality and morbidity of UTIs [22
], and also makes killing bacteria more difficult and challenging than before. A quick diagnosis of antibiotic susceptibility is therefore crucial for UTI treatments. SERS can reveal antibiotic susceptibility, though not as quickly as bacterial identifications. It is still much faster than the traditional disc-diffusion method. The gradual or time-dependent disappearance of a specific signal peak in the Raman spectrum of bacteria after antibiotic treatment indicates antibiotic susceptibility. The persistent existence of a specific signal peak after antibiotic treatments indicates an antibiotic-resistant pathogen. These two phenomena can be seen in Figure 6
PCA analysis can also help differentiate an antibiotic-susceptible bacterial strain from antibiotic-resistant bacteria. E. coli
spots clustered in the upper part of the chart, but E. coli ESBL
grouped in the lower part of the chart (Figure 5
C); Enterococcus faecalis
in the left part, but VRE
in the right part of the chart (Figure 5
G). There were three E. coli
spots located in the upper-right corner of the chart, and these spots appeared to be outliers (Figure 5
C). Signal intensities of these three outliers were lower than those of the grouped spots in the left part of the chart (data not shown) and the lower signal intensities deviated these three spots from the main E. coli
cluster. The lower signal intensities of the three samples may be due to low bacterial concentration in urine. The other possible causes of low signal intensities may result from inappropriate sample processing, for example, cell debris from pyuria attached to the bacterial cell wall, which can lead to poor light emitting and a subsequent low Raman signal.
PCA can also be used to facilitate the diagnosis of mixed-flora infections. The spot of mixed infections with two kinds of bacteria tends to be located near the spots of the two bacteria identified by conventional bacterial culture rather than near the other two unrelated bacteria. This tendency was also seen in all seven cases in which two kinds of bacteria were identified. The reason for this tendency is not clear.
The SERS technique, however, has some limitations. First, it may need an expensive apparatus to perform the experiment, such as confocal microscopy. We used a cylindrical SERS chip illuminated by a portable Raman spectrometer, which was relatively lower in price compared with that of the expansive confocal microscope used in other studies [6
The second limitation of Raman SERS is that the concentration of bacteria in samples may be too low to be detected [20
]. The limit of the lowest-detectable specimen bacteria concentration range was from 103
]. In this study, we chose bacteriuria urine samples from febrile patients with a UTI as our experiment model. Most of these patients were admitted via emergent department without a prehospital antibiotics treatment, and a routine hospital urine examination revealed the presence of bacteria by light microscope. The presence of pyuria and bacteria in routine urinalysis results suggested the presence of bacteria in the urine samples. Therefore, bacteria can be found in most samples without repeated centrifugation for sample concentration. Four urine samples without enough bacterial concentration for SERS detection and bacterial concentration can be increased by the centrifugation method [26
]. Removing cells or cell debris in urine by 700 rpm centrifugation for 10 min, followed by concentration, can increase bacterial concentration (centrifugation of 13,000 rpm for 5 min) and enhance Raman signal intensity. Repeated concentration (centrifugation of 13,000 rpm for 10 min, twice) can further increase urine-sample bacterial concentration and improve the Raman spectrum resolution.
The third problem is mixed-flora infections, which can be found in some UTIs. Mixed-flora infections make bacterial identification difficult. We failed to find all single-pathogen-specific signals in the Raman spectrum of the mixed-flora infections. A Raman spectrometer with higher resolution may help us find some pathogen-specific signals of low intensity. We also need to establish a broad mixed-flora spectrum library to facilitate the recognition of software-assisted pathogen identifications.