CHAPTER 4. Application of Dx-FS in POCT and EPOCT settings
4.1 UTI diagnosis using patient sample
Application of Dx-FS in POCT and EPOCT settings to elaborate on the how Dx-FS fits clinical needs in a resource-limited setting, we performed a field trial of Dx-FS in Tiruchirappalli, India, where the majority of patients come from rural areas (Figure 4.1). Predominantly patients seeking intervention with symptoms of UTI visit primary care centers, and they are prescribed medication based on symptoms rather than a clinical diagnosis. Selected cases are subjected to traditional clinical diagnosis like microscopic examination, dipstick assay and cultures, in a laboratory which are located in urban settings. Generally, the primary care administers empirical antimicrobials as per the guidelines for community accrued UTI.19 Studies on a large population of UTI-affected patients, concluded that symptom-based UTI diagnosis place a large number of patients at risk by prescribing unnecessary antimicrobials.60,95–97 Though the practice is well-established and convenient for treating many patients who predominantly visit primary care centers, it significantly contributes to antimicrobial resistance due to the lack of diagnostic tool.98 To put our device in practical use we set out to perform a field test in India by recruiting patients who have clinical symptoms UTI .
4.1.1 Field trials at Kauvery Hospitals in India.
For testing our device with patient samples in India, devices were first fabricated under sterile conditions. Every device was individually inspected for damage and irregularities to avoid failure during the experiment. The devices were then treated with an oxygen plasma and vacuum-sealed to prevent any contamination during transfer. For testing every sample, a new device was opened, the bottom liquid was injected (Figure 2.8), and the device was then used. The study was performed under the approval of the Kauvery Hospital Institutional Ethics Committee (IEC) (Ref No:
KHEC/018/August/Meeting No-002/Proposal No: 002). The samples from the patients were obtained according to the IEC protocol of Kauvery Hospitals India, Pvt., Ltd. (# ECR/966/Inst/TN/2017). The samples were obtained from Kauvery Hospital, Trichy, and tests were performed at a microbiology lab within the hospital. The entire study was registered in the Clinical Trial Registry India (CTRI/2018/09/015678).99 The samples were obtained from the patients after obtaining consent for the study. Urine samples were obtained from patients suspected with UTI (based on symptoms). The presence of bacteria and their count were confirmed in the microbiology lab at Kauvery Hospital.
72
Figure 4.2. Proof-of-concept testing using Dx-FS. (a) Urine samples from UTI suspects tested by culturing or Dx-FS. Bacterial cell counts measured by culturing or by Dx-FS are shown in the heat map.
Diagnosis results from medical doctors are indicated by green or red circles. * Urinalysis of Patient 29 showed 15 pus cells/high power field (HPF) in urine. ** Urinalysis of Patient 38 showed 15 RBCs/HPF and 5 pus cells/HPF in urine. (b) Receiver operating characteristic (ROC) curves of UTI diagnoses from culturing and Dx-FS. The Dx-FS results are highly comparable to the conventional culture-based UTI diagnosis in terms of sensitivity and specificity. (c) Average turnaround times. A culture takes over 120 times longer than Dx-FS. (d) Pie charts showing the results of our simulation of antibiotic dosages.
Results indicate that 54 % of UTI suspects misused or were over treated with antibiotics, 5 % of UTI patients were undertreated (* and ** from (a)), and 41 % of UTI patients were treated correctly. In contrast, Dx-FS could aid in treating all patients with the correct treatment.
73
Figure 4.3. Clinical validation of UTI with Dx-FS. A clinical examination of UTI using Dx-FS was performed. The patients were split into two groups based on the bacterial cell culture results, i.e., < 103 CFU/mL or ≥ 103 CFU/mL. Here, the bacterial cells detected by Dx-FS were compared based on different criteria: (a) gender (left panel, p = 0.0826; right panel, p = 0.1122), (b) age under or over 50 (left panel, p = 0.4681; right panel, p = 0.2386), and (c) the presence of red blood cells (RBCs) (left panel, p = 0.3214; right panel, p = 0.7497). (d) For samples with bacterial cells ≥ 103 CFU/mL, measurements are compared for bacterial cell strains that are gram-negative or gram-positive (left panel, p = 0.3720) and Escherichia coli or Klebsiella pneumoniae (right panel, p = 0.3435). For all cases, the measurements were not statistically different (NS, p > 0.05).
4.1.2 Routine Urine culture test of patient sample.
The urine samples obtained from the patient were deidentified, and a routine culture was performed at the hospital laboratory. In brief, 1–10 μL of urine from a patient was spread onto a BBL Trypticase Soy Agar plate with 5% blood (Kauvery Hospital, India) and, in parallel, a MacConkey plate.
The plates were incubated at 37 °C overnight in ambient air and were checked for growth on consecutive days, and the colony count was confirmed depending on the growth levels. The microbes were transferred to glass slides, stained using an acid-fast stain, and visualized under a microscope.
74 4.1.3 Outcomes of the study patient sample.
In our approach, we attempted to minimize the time and associated cost and prevent the misuse of antimicrobials by providing clinical information regarding bacterial load within 50 minutes (Figure 2.1a, Dx-FS). We tested urine samples from 39 patients who enrolled during August 1–17, 2018 (Table 4.3). The samples were collected and processed according to the regulations of the Clinical Trial Registry-India, and all experiments were performed in a microbiology lab within the hospital (Figure 4.1). Samples were processed with two different methods: conventional (urinalysis and culturing at a hospital laboratory) and Dx-FS (Figure 1a). Here, patients that had bacterial cells ≥103 CFU/mL in their urine were considered UTI+ according to the hospital standards (European urinalysis guidelines).100 The measurements from all 39 patients are shown in Figure 4.2a. Twenty-three UTI suspects (59%) tested negative according to the bacterial cell cultures (Patients 1–21, 29, and 38, Figure 4.2a, culture), whereas 21 UTI suspects (54%) had ≤103 CFU/mL in their urine according to the Dx-FS test (Patients 1–21, Figure 4a, Dx-FS). For Patient 29, over 15 pus cells (leukocytes) per high-power field (HPF) were found, and for Patient 38, over 15 red blood cells (RBCs)/HPF and over 5 pus cells/HPF were found from the subsequent urinalysis (Figure 4.2a, denoted as * and **, respectively). Such pus cell or RBC counts in urine indicate infection and urinary tract damage; therefore, antimicrobial treatments are needed21. The results confirm that the UTI diagnosis from the culture and Dx-FS were comparable based on the ROC curves (Figure 4.2b). Dx-FS provided a robust and versatile UTI diagnosis devoid of gender, age, and strains of bacterial cells (Figure 4.3). The culture results, 54% of the UTI suspects did not have UTI. Moreover, 5% of the UTI suspects that tested negative according to the culture still needed antimicrobials (Figure 4.2d). In the Dx-FS-based POCT scenario, bacterial cell occurrence in urine can be identified within 50 minutes (Figure 4.2c, Dx-FS); therefore, not all the UTI suspects with clinical symptoms but only those who had pathogen detected may be treated with antimicrobials. The empirical antimicrobial treatment of non-UTI patients could have undesirable side effects, paving way to drug resistant.
75 Table 4.3 Summary of UTI patient information.
Characteristic No %
Total patients enrolled in the study = 39 Symptomatically identified
as UTI Age (Years)
Age <16 2 5%
16 ≤ Age < 25 4 10%
25 ≤ Age < 50 10 26%
50 ≤ Age < 80 23 59%
Gender
Male 21 54%
Female 18 46%
Inpatient/outpatient
Inpatient 25 64%
Outpatient 14 36%
Bacterial load (CFU/mL)
Bacterial load ≥105 6 15%
104 ≤ Bacterial load <105 4 10%
103 ≤ Bacterial load <104 8 21%
102 ≤ Bacterial load <103 5 13%
No growth (NG) 16 41%
Organism *Percentile within bacterial load ≥103
Escherichia coli 10 56%
Enterococcus faecalis 1 6%
Klebsiella pneumoniae 6 33%
Streptococcus 1 6%
Pus cell (#/HPF*)
Many (Pus cell ≥ 15) 2 5%
Moderate (5 ≤ Pus cell <15) 4 10%
Few (1 ≤ Pus cell <5) 18 46%
No cell 15 38%
RBC (#/HPF*)
Many (RBC ≥15) 5 13%
Moderate (5 ≤ RBC <15) 0 0%
Few (1 ≤ RBC < 5) 15 38%
No cell 19 49%
*HPF: High-Power Field.
76