Rapid detection of urinary tract pathogens using microcalorimetry: principle, technique and first results


Gernot Bonkat, Department of Urology University Hospital Basel, Spitalstrasse 21, CH-4031 Basel, Switzerland. e-mail: bonkatg@uhbs.ch


Study Type – Diagnostic (exploratory cohort)

Level of Evidence 2b

What's known on the subject? and What does the study add?

Microcalorimetry has been applied in several microbiological studies, but never in a clinical urological context. In addition, basic knowledge on the growth of urinary pathogens in urine is still scarce and data regarding the growth rate of many urinary pathogens in urine are still not available.

The study demonstrates that this innovative application of microcalorimetry is useful in (i) investigating the growth of urinary pathogens in sterilized urine and (ii) as a rapid tool for diagnosis of urinary infection as well as for further identification of the causative infectious agent.


  • • To investigate the value of isothermal microcalorimetry (IMC) in the detection and differentiation of common urinary tract pathogens in urine. IMC is a non-specific analytical tool for the measurement of heat in the microwatt range.


  • • A microcalorimeter equipped with 48 channels was used. Detection was accomplished, and growth was monitored for four bacterial strains in sterilized urine at 37 °C by measuring metabolic heat flow (µW =µJ/s) as a function of time.
  • • The strains were Escherichia coli, Proteus mirabilis, Enterococcus faecalis and Staphylococcus aureus.


  • • Bacterial growth was detected after 3.1 to 17.1 h with decreasing inocula.
  • • The detection limit was 1 colony-forming unit (CFU)/mL for E. coli, 10 CFU/mL for P. mirabilis and E. faecalis and 103 CFU/mL for S. aureus.
  • • The total heat was highest in P. mirabilis ranging from 10 to 12 J, followed by E. coli (3–4 J), S. aureus (2–3 J) and E. faecalis (1.3–1.5 J).
  • • The shape of the heat flow curves was characteristic for each species independent of its initial concentration.


  • • IMC allows rapid detection of bacteriuria, much faster than conventional culture. Urinary tract pathogen detection after only 3.1 h is realistic.
  • • Clearly different heat flow patterns enable accurate pathogen differentiation.
  • • Due to expeditious identification of urine samples that contain only low colony counts (i.e. less than 103 CFU/mL), IMC may become a valuable screening tool for detecting the presence of significant bacteriuria.

urine culture


urinary tract pathogen


isothermal microcalorimetry


brain heart infusion


matrix-assisted laser desorption/ionization time of flight


UTI is the second most infection in humans after respiratory infections. The majority of UTIs are community acquired (57.4%), whereas 35.6% are healthcare associated and 7% nosocomial [1]. In the hospital setting, UTIs are the most frequent acquired infection and the second most common cause of bacteremia [2,3]. The majority of cases are caused by a limited number of bacteria of which uropathogenic Escherichia coli is the predominant pathogen [4,5].

Urine culture (UC) confirms the diagnosis of UTI and is still considered the diagnostic ‘gold standard’ for pathogen identification, quantification and susceptibility testing. However, up to 80% of samples will not yield bacterial growth [6], resulting in high laboratory workload and costs for what proves to have been inconsequential testing. Automated urinalysis, semi-automated strip tests and microscopic analysis are used to select urine samples more likely to yield positive results in culture, with the aim of improving the efficiency of handling these samples in the laboratory. However, despite optimal specimen collection, sample processing and interpretation, no method is foolproof for all patient groups [7]. A diagnostic tool that can more rapidly and reliably identify the presence of urinary tract pathogens (UTPs) than existing methods might reduce diagnostic manpower and other costs – provided the required instrumentation and supplies are not more expensive. Such a method could also provide an urgently needed, more evidence-based approach to clinical management of UTIs. Isothermal microcalorimetry (IMC) is a non-specific analytical tool for measurement of heat produced or consumed over time by chemical reactions or physical changes of state in a specimen [8]. With sensitivity of the order of 0.2 µW, IMC can detect the heat produced by a small number of microorganisms. Assuming that a typical single bacterial cell produces ≈2 pW when active, only 100 000 bacteria are required to produce a detectable signal in most commercial isothermal microcalorimeters [9].

Although microcalorimetry has been commonly used in the material sciences, the explosives industry, chemistry and the physical sciences, only few attempts have been made to use it in medical settings [10,11]. The aim of this pilot study was to evaluate the diagnostic potential of IMC in the detection and identification of common UTPs.



Urine from a healthy donor was collected in a sterile container and rapidly filtered through a 0.2 µm pore size Stericup® filter (Millipore, Billerica, MA, USA). The sterile filtered urine was stored at 4 °C until use (within 72 h) or alternatively at −80 °C for longer storage (Fig. 1).

Figure 1.

Experimental setup. Cultures of test organisms are diluted in sterile saline. From this suspension, serial 10-fold dilutions are prepared in sterile saline to attain final concentrations targeted at 105, 103, 101 and 1 CFU/mL. Urine is collected in a sterile container and rapidly filtered. Calorimetry ampoules are filled with sterile filtered urine and the prepared inoculum. A microcalorimetry instrument (Thermal Activity Monitor) equipped with 48 measuring channels is used to measure and record the heat flow.


Four different bacterial strains of the most common UTPs [4,5] were used in this study: Staphylococcus aureus (ATCC 25923), Enterococcus faecalis (JH2-2), Proteus mirabilis (DSM 4479) and E. coli (DSM 10142). Strains were stored in cryovials at −80 °C in brain heart infusion (BHI) containing 20% glycerol before experiments. Precultures of test organisms were grown overnight in BHI at 37 °C. Inoculates were prepared as follows. Overnight cultures were diluted in sterile saline (0.85% NaCl) and adjusted to an optical density (OD600) of ≈0.132 corresponding to a McFarland value of 0.5. From this suspension, serial 10-fold dilutions were prepared in sterile saline to attain final concentrations targeted at 105, 103, 101 and 1 CFU/mL. The accuracy of the inoculum size was confirmed by quantitative cultures on BHI agar plates. Colonies were counted from plates with 10–200 colonies, and the colony-forming units per millilitre were calculated (Fig. 1).


IMC glass ampoules (4 mL) and caps were autoclaved at 121 °C for 20 min. Each calorimetry ampoule was aseptically filled with 2.9 mL of sterile filtered urine and 0.1 mL of the prepared inoculum; the remaining headspace consisted of 1 mL of the ambient air present during aseptic filling. Ampoules were sealed with a standard crimped metal lid with a silicone rubber seal.


A microcalorimetry instrument (Thermal Activity Monitor, TAM48, TA Instruments, New Castle, DE, USA; Fig. 1) equipped with 48 measuring channels was used to measure and record the flow (i.e. heat flow =µW =µJ/s) during the growth of the different strains in urine.

The instrument thermostat was previously set at 37.0 °C. Samples in microcalorimetric ampoules were introduced into the measuring channels using the two-step ampoule temperature equilibration procedure recommended by the manufacturer. Heat flow data were recorded until the signal returned to baseline, or up to a maximum of 5 days.


Detection time was defined conservatively as the time from insertion of an ampoule into the microcalorimeter until exponential growth produced a rising heat flow signal >10 µW (i.e. ≈40 times the effective sensitivity of the calorimeter). Growth rate was calculated by fitting an exponential model (Qt=Q0 exp(µt) where Q represents the heat, i.e. the integral of the heat flow curve) over the exponential part of the curve (Fig. 2). Generation time and minimum generation time were obtained using the relation tg= (ln 2)/µ where µ is the growth rate obtained through curve fitting. Data analysis was accomplished with the manufacturer's software (TAM Assistant) and the R statistical package (R Development Core Team, Vienna, Austria, 2008).

Figure 2.

(A) Relationship between heat flow and detection time; (B) relationship between heat and growth parameters. The detection time is defined as the time from insertion of an ampoule into a microcalorimeter until exponential growth produced a rising heat flow signal >10 µW. Growth rate is calculated by fitting an exponential model (Qt= Qexp(µt) where Q represents the heat, i.e. the integral of the heat flow curve) over the exponential part of the curve. Generation time and minimum generation time are obtained using the relation tg=(ln 2)/µ where µ is the growth rate obtained through curve fitting.


All the bacterial strains used were able to grow in urine. Bacterial growth was readily detected after 3.1 to 17.1 h (Table 1, Fig. 3). E. coli and P. mirabilis at initial inoculum concentrations of 105 CFU/mL were detected within 4 h. The same initial concentrations of S. aureus and E. faecalis were detected in ≈5 and 6 h respectively. A 100-fold decrease in the inoculate concentration resulted in ≈3 h delay in detection for the different strains considered here. This is because lower initial inoculum concentrations naturally require additional time for growth to increase the number of bacteria to the point that their aggregate metabolic heat is above the instrument's detection limit. Furthermore, the four bacterial strains tested show clearly different heat flow patterns (Fig. 3). P. mirabilis exhibited a single sharp peak with a long tail, most probably related to remaining urease activity releasing heat even after growth stopped. P. mirabilis produced the highest heat flow (≈300 µW). The E. coli heat flow pattern showed mostly two peaks of similar height (between ≈90 and 120 µW) followed by a short tail. E. faecalis generated a heat flow pattern composed of a sharp first peak (≈60 µW) followed by a smaller broader peak (≈30 µW) and a final very small and broad peak (≈10 µW). Finally, S. aureus also produced a two-peak pattern followed by a long tail. In S. aureus the first peak was the highest (≈70–80 µW) and the second peak was slightly lower in intensity (50–60 µW) and broader. Finally, the total heat produced by the different bacterial strains was highest in P. mirabilis ranging from 10 to 12 J, followed by E. coli (3–4 J), S. aureus (2–3 J) and E. faecalis (1.3–1.5 J). Calculated growth rates determined using the calorimetric data (heat over time curves) were between 0.54 h−1 and 1.21 h−1 in urine (Table 2). Similarly the generation times in urine were between 0.58 h and 1.41 h. For each species tested, the estimated generation time found using microcalorimetry was in close agreement with the generation time determined using classical culture [12–14].

Table 1.  Mean detection time in hours (±sd) of microorganisms using different initial inoculates
MicroorganismsInoculum concentration (CFU/mL)
  1. ND, not detectable. Measured microbial concentrations at targeted concentrations of 105 CFU/mL were 1.12 × 105 CFU/mL for E. coli, 1.71 × 105 CFU/mL for E. faecalis, 1.41 × 105 CFU/mL for P. mirabilis and 0.81 × 105 CFU/mL for S. aureus.

E. coli 3.92 ± 0.177.05 ± 0.1712.42 ± 2.5917.16
E. faecalis 6.19 ± 0.0510.27 ± 0.2712.5 ± 0.44ND
P. mirabilis 3.11 ± 0.105.55 ± 0.057.89 ± 0.41ND
S. aureus 4.78 ± 0.059.89 ± 0.65NDND
Figure 3.

Heat flow curves of UTPs. The four bacterial strains tested show clearly different heat flow patterns. E. coli generates a heat flow pattern composed of mostly two peaks of similar height followed by a short tail. The E. faecalis heat flow shows a sharp first peak followed by a smaller broader peak and a final very small and broad peak. P. mirabilis exhibits a single sharp peak with a long tail. In addition, P. mirabilis produces the highest heat flow (≈300 µW). S. aureus also shows a two-peak pattern followed by a long tail.

Table 2.  Bacterial growth parameters detected by IMC and their comparison with classical urine culture
MicroorganismsIMC growth rate (h−1)IMC Td (h)Classical culture Td (h)Reference
  1. T d, doubling time.

E. coli 0.63 ± 0.121.16 ± 0.220.75 ± 1.67 [14]
E. faecalis 0.64 ± 0.151.13 ± 0.240.81 ± 0.05 [16]
P. mirabilis 1.21 ± 0.080.58 ± 0.040.40 ± 1.02 [15]
S. aureus 0.54 ± 0.071.41 ± 0.21


UTIs are among the most common bacterial infections and lead to a significant workload in microbiology laboratories [15]. Although molecular diagnostic approaches or DNA hybridization and amplification techniques are being applied to the diagnosis of many infections, UTIs are still generally diagnosed – as they have been for decades – by UC. However, the majority of UCs sent to the laboratory are negative [6]. In many instances, the clinician rapidly wants to rule out UTI. However, the turnaround time is especially long for negative cultures as they require prolonged incubation. At the authors' institution, ≈17 000 urine cultures are ordered per year, of which only 25% show bacterial growth. Long-time monitoring of what turn out to be negative urine samples results in an unnecessary workload and increased consumption of laboratory resources – important factors in times of ever-increasing cost restraints. A variety of screening methods, e.g. dipstick testing [16], automated urine flow cytometry [17] and microscopic sediment analysis, have been developed to reduce the number of urine specimens that will produce a negative UC. However, these tests are not designed to replace UC because they neither identify the causal pathogen nor establish antibiotic susceptibility to guide the choice of antibiotic treatment [18].

Another major drawback of UC is the time required from sample collection to the delivery of pathogen identification and susceptibility results.

The introduction of the matrix-assisted laser desorption/ionization time of flight (MALDI-TOF) [19] system will expedite identification of species when cultures turn positive, but susceptibility testing will still delay streamlined antibiotic therapy. Empirical therapy of UTIs results in an overuse of antibiotics and has contributed to a dramatic increase in resistant pathogens [20,21]. The Antimicrobial Resistance Epidemiological Survey on Cystitis study has shown considerable resistance rates of E. coli to cotrimoxazole (28.5%) and to ciprofloxacin (8.2%) in uncomplicated female cystitis. Other authors report resistance rates of E. coli to fluoroquinolones greater than 50% in some patient populations [22,23]. A further threat is the emergence of extended-spectrum beta-lactamase-producing Enterobacteriaceae [24].

A clinically simple, accurate and rapid method for detection and identification of UTPs would (i) lead to improvement in the quality of patient care, (ii) free up laboratory resources, (iii) reduce the empirical initiation of antibiotic therapies and therefore (iv) help to prevent the development of resistance. The results of our exploratory study showed that, with IMC, UTP detection and quite possibly identification can be achieved within a few hours. Due to the limited number of common UTPs, identification based on typical heat flow patterns can be achieved using numerical methods such as cross-correlation analysis and parametric comparison [25]. Although even faster, emerging diagnostic approaches such as MALDI-TOF [19], multiplex real-time PCR techniques (SeptiFast®) [26] and biosensor technology [27] lack the ability to evaluate drug susceptibility (i.e. determine minimal inhibitory concentration for a specific antibiotic). This is something easily done with IMC as described below. Also PCR does not distinguish between living and dead bacteria.

No data are yet available demonstrating the IMC evaluation of UTP antibiotic resistance. However, IMC has already been used to rapidly detect S. aureus bacterial resistance [11] and to determine minimum inhibitory concentrations for various antibiotics for various bacterial strains including E. coli and S. aureus[28]. The close agreement observed between growth parameters obtained here by IMC and those reported using conventional methods for the different bacterial strains used indicate the accuracy of the results that microcalorimetry could provide. Since microcalorimetry is simple and provides real-time continuous data, no additional analyses (e.g. cell counts) have to be performed over and over again to document growth. This could drastically reduce workload and other related laboratory costs. With respect to the application of IMC in the clinical setting of UTI, the bacterial strains investigated serve as a valid proof of concept. Indeed, E. coli is the causative agent in 70–95% of community acquired UTI and 50% of all cases of nosocomial infection. E. faecalis, P. mirabilis and S. aureus are among the most common UTI causative agents in female patients with uncomplicated cystitis [5] as well as in nosocomial UTIs, respectively [4].

Since only a small number of bacterial species represent a large proportion of UTI causative agents, the clearly different heat flow patterns for these common pathogens could possibly serve as an additional means of UTP strain identification. In addition, from such patterns it is possible to obtain baseline data about UTP metabolism [9]. The pattern observed for an individual strain is not dependent on the inoculate size but depends on a strain's metabolic proclivities for consuming nutrients. This was previously demonstrated for the detection of contamination in platelet products [29].

Regarding costs of consumables IMC is expected to be comparable with culture methods since only simple ampoules and synthetic urine are required. Current microcalorimeters allow multiple samples to be processed in parallel (from 1 to 48). The potential compatibility of IMC with techniques currently developed for antibiotic resistance detection makes it a highly promising tool. Microcalorimetry is not limited to liquid media. Therefore, the use of solid media such as CLED agar could also be considered, allowing not only detection and presumptive identification but also confirmation using the colorimetric properties of this medium. Furthermore, the use of specifically designed microbiology media might further improve the detection time of UTP. In addition, with respect to slow growing less common UTP such as Mycobacterium spp. (including Mycobacterium tuberculosis), the sensitivity and rapidity of IMC would be an invaluable advantage [30]. Since IMC is completely passive, specimens are undisturbed, and after any period of IMC measurement the ampoule contents (media, bacteria etc.) can be analysed by any other method desired. The continuous IMC data are amenable to mathematical treatment, and the IMC technique generally lends itself to future automation. Finally, it has to be noted that microcalorimetry is a fast growing field. New instruments with miniaturized design such as a microcalorimeter on a chip are emerging and will most probably lead to further improvement in the field of UTI diagnosis due to higher throughput and bedside monitoring.

In conclusion, IMC is a promising novel diagnostic method for UTP detection. IMC is simple to perform, highly sensitive and reproducible. Quantitative evaluation of multiple independent urine samples can be achieved simultaneously. The system allows drug susceptibility testing which cannot be achieved using molecular or spectrometric methods. Finally, IMC offers means to reduce the laboratory workload and related costs due to rapid identification and elimination of negative urine samples submitted to the microbiology laboratory for culture.


None declared.