GROUP B STREPTOCOCCUS INFECTION

Abstract

A screening method for confirming that a subject does not have a Group B Streptococcus (GBS) infection, the method comprising: determining if a GBS-volatile organic compound (VOCs) is not present in a sample that has been taken from the genital mucosa of the subject, wherein if a GBS-VOC is not present in the sample, the subject does not have a Group B Streptococcus infection. A method of diagnosing that a subject has a Group B Streptococcus (GBS) infection, the method comprising: determining if a GBS-volatile organic compound (VOCs) is present in a sample that has been taken from the genital mucosa of the subject, wherein if a GBS-VOC is present in the sample, the subject has a Group B Streptococcus infection.

Claims

1. A screening method for confirming that a subject does not have a Group B Streptococcus (GBS) infection, the method comprising: determining if a GBS-volatile organic compound (VOCs) is not present in a sample that has been taken from the genital mucosa of the subject, wherein if a GBS-VOC is not present in the sample, the subject does not have a Group B Streptococcus infection.

2. A method according to claim 1, wherein if a GBS VOC is present in the sample, the subject has a GBS infection.

3. A method of diagnosing that a subject has a Group B Streptococcus (GBS) infection, the method comprising: determining if a GBS-volatile organic compound (VOCs) is present in a sample that has been taken from the genital mucosa of the subject, wherein if a GBS-VOC is present in the sample, the subject has a Group B Streptococcus infection.

4. A method of treating a subject that has a GBS infection, the method comprising: diagnosing a subject that has a Group B Streptococcus infection using the method according to any one of claims 1 to 3, and administering a therapeutic agent to the subject, in order to treat the infection.

5. A method of treating a subject that has a GBS infection, the method comprising: administering a therapeutic agent to a subject, who has been diagnosed with a GBS infection using the method according to any one of claims 1 to 3, in order to treat the infection.

6. A method of determining if a therapeutic agent is effectively treating a GBS infection in a subject, the method comprising: determining the concentration of a GBS-VOC in a test sample that has been taken from the genital mucosa of the subject, and comparing the concentration of the GBS-VOC in the test sample to the concentration in a reference sample, wherein if the concentration of a GBS-VOC in the test sample is lower compared to the concentration in a reference sample, it may be indicative that the therapeutic agent is effectively treating the Group B Streptococcus infection in the subject.

7. The method according to claim 6, wherein the reference sample has been taken from the same subject or a different subject.

8. The method according to claim 6 or 7, wherein reference sample is a sample that has been taken from the same subject but at an earlier time point than the test sample.

9. The method according to claim 8, wherein the reference sample taken at the earlier sample indicated that the subject was colonized with GBS.

10. The method according to any one of claims 6 to 9, wherein the concentration of the GBS-VOC in the test sample is lower by (or reduced by) at least about 5%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 95% or 100% compared to the concentration in the reference sample.

11. The method according to any one of the preceding claims, wherein the sample is liquid lining the genital mucosa of a subject, such as vaginal liquid.

12. Use of a GBS-VOC as a biomarker for determining if a subject has, or does not have, a GBS infection.

13. The method according to any one of claims 1 to 11 or the use according to claim 12, wherein a GBS-VOC is selected from the group consisting of butyraldehyde (butanal), ethyl acetate, methyl acetate and 2-methylpropanal.

Description

FIGURE LEGENDS

[0045] For a better understanding of the invention, and to show how embodiments of the same may be carried into effect, reference will now be made, by way of example, to the accompanying Figures, in which:

[0046] FIG. 1 is a 3D topographic map generated by GC-IMS analysis of the constituents of a GBS positive swab. The x-axis represents the drift time in the IMS and the y-axis represents the retention time of the same chemical in the GC. The non-blue areas are chemical signals be detected by the instrument;

[0047] FIG. 2 is an AUC ROC curve based on the analysis of samples taken from pregnant women that are colonised with GBS versus samples taken from pregnant women that are not colonised with GBS. Samples were analysed using a G.A.S. GC-IMS instrument using the VOCs, butyraldehyde (butanal), ethyl acetate, methyl acetate and 2-methylpropanal;

[0048] FIG. 3 is a box and whiskers plot representing the distribution of level of probabilities of assigning VOCs outputs to patients with and without GBS. The closer the probability is to 1, the more certainty the classification model has to define the sample group. The analysis was performed based on the VOCs, butyraldehyde (butanal), ethyl acetate, methyl acetate and 2-methylpropanal. Each of the swabs was confirmed as positive or negative for GBS by simultaneously performing an enrichment culture analysis of the sample. Dots adjacent to the “whiskers” correspond to outliers. The whiskers indicate 1.5 IQR; and

[0049] FIG. 4 is an intermediate analysis graph that shows the location of VOCs identified by GC-IMS analysis of (A) a GBS positive sample and (B) a GBS negative sample after removal of background VOCs (i.e. non-GBS VOCs), before ROC and probability analysis.

EXAMPLES

[0050] Study Design

[0051] An accuracy study was undertaken at one UK hospital (University Hospitals Coventry & Warwickshire) serving a diverse population. The study protocol was approved by the NHS Research Ethics Committee West Midlands Birmingham South on 14th January 2014 (13/WM/0486) and all participants gave written informed consent. The Group B Strep Support charity was consulted prior to the application for funding regarding a patient perspective about the study. Research was carried out according to The Code of Ethics of the World Medical Association (Declaration of Helsinki).

[0052] Participants and Test Methods

[0053] Women between 14-36 weeks gestation were consented during their attendance to a high risk antenatal clinic for women at an increased risk of spontaneous preterm birth. A speculum examination was performed as per patient routine care. A vaginal swab (reference standard) for microbiology culture and sensitivity testing using the enriched culture method was taken and placed into a nonnutritive transport medium, and concurrently two cotton swabs were used to obtain index test vaginal samples.

[0054] The index test swabs were then placed in universal containers and snap frozen in liquid nitrogen and stored at −80° C. Specimens were obtained by gently rotating the swabs across the mucosa of the vagina. Biomedical scientists independently interpreted the reference swab cultures. Demographic data including age at booking pregnancy, BMI, ethnicity and smoking status were collected about each woman (Table 1). Samples were taken in a consecutive series from all women who consented in the clinic, some women consented to samples being taken during every attendance to the clinic.

[0055] Chemical Analyzer

[0056] Chemical vapour analysis of the index swabs was undertaken in the BioMedical Sensors Laboratory, University of Warwick. Here a Gas Chromatograph-Ion Mobility Spectrometer (GC-IMS) was used. This instrument was chosen over more traditional gas chromatograph mass spectrometer (GCMS) as the basic sensitivity of the instrument is much higher than GCMS, it can use nitrogen/air as the carrier gas (so no need for expensive carrier gases such as helium), has a lower purchase/test cost than GCMS and has a much smaller form factor, making it applicable for a ward setting.

[0057] Analysis was undertaken using GC-IMS instrument manufactured by G.A.S. (GC-IMS is also the product name, Dortmund, Germany), which is based on Gas Chromatograph-Ion Mobility Spectrometery principles (GC-IMS). In use, the samples, formed of a mixture of VOCs that emanate from the vaginal swab, are injected into the GC-IMS. These VOCs are preseparated by the GC column (in this case a 30m column), which takes the complex mix of chemicals and separates them based on their interaction with the long column coated with a retentive layer (made of (5%-phenyl)(1%-vinyl)-methylpolysiloxane)). Thus chemicals are eluted from the column at different times (known as the retention time). The pre-separated chemicals exit the GC and enter a drift tube IMS detector. Here the molecules are ionized using a radioactive source (in this case tritium) and then released into the drift tube in a controlled manner. The ions are then moved along the drift tube using an electric field (400 V/cm). At the same time a buffer gas (in this case nitrogen) is fed in the opposite direction to the ions. The resultant impacts between the ions and the buffer gas reduce the velocity of the ions. Thus ions with different sizes and shapes will have different drift times (i.e. the time taken for the ion to be detected by the Faraday plate detector at the end of the drift tube).

[0058] Thus ions achieve different velocities, inversely proportional to their size, mass and charge and then are collected on a Faraday plate, to provide a time-dependent signal corresponding with ion mobility. Thus, the larger the ion, the greater the number of impacts and the slower the ion travels along the tube. The device can measure substances in the low ppb range.

[0059] Chemical Testing and Analysis

[0060] The G.A.S. GC-IMS instrument was used to test 607 samples from 243 women. Samples were briefly stored at −80° C. before being thawed and transferred to a 20 ml glass vial in batches of 20. The vials were then sealed with a crimp top lid fitted with a PTFE septum. In measurement, the index samples were initially placed in a tray and kept refrigerated at 4° C. to reduce unwanted odour emission and sample degradation, whilst other samples were being tested. Prior to measurement samples were heated to 40° C. for 10 minutes. The sample line for the GC-IMS was inserted into the septa of the vial using a needle and 2 mls of sample were then extracted from the vial by syringe and injected into the analytical platform. The machine settings were as follows: E1: 150 ml/min (for the drift tube IMS), E2: 20 ml/min (for the GC column) and the pump at 25%. The total run time was 10 minutes. The temperatures were set to: T1: 45° C., T2: 80° C., and T3: 70° C.

[0061] Statistical Analysis

[0062] The GC-IMS data was first extracted using the L.A.V. software (v2.2.1, G.A.S, Germany), which converts the data from its native file format to a text file. This was followed by a pre-processing step to reduce the dimensionality of the data, making the statistical analysis less computationally intensive. A typical GC-IMS output file (of a single sample) contains typically 11 million data points. Though the number of data points is high, the information content is sparse, with the all of the values containing non-background information being located around the centre of the dataset. Thus, we are able to crop the central section of the data and then apply a threshold to make the background values all be zero. These values are selected by visual inspection of the data using the LAV software and results in around a 500-fold reduction in the number of non-zero data points. Once completed, the data was analysed using a 10-fold cross validation approach. In each fold, the data was split into a 90% training set and a 10% test set. Features with discriminatory power were identified form the training set using a rank-sum test and 50 features with the lowest p-value were taken forward for classification. Here, five different classifiers we used, specifically sparse logistic regression, random forest, Gaussian process classifier, support vector machine and neural network (this set is commonly used within our pipeline). Once the training models had been created, they were applied to the same features in the test set. This process is repeated ten times until all the data has a test result. This process provided test probabilities for each sample and from this, statistical values, including sensitivity and specificity were calculated.

Example 1—Vaginal Swab

[0063] 243 women were swabbed throughout pregnancy corresponding to 607 sets of swabs. The demographics of these women is illustrated in Table 1. The maternal GBS colonisation rates, as defined by a positive enriched culture from vaginal swabs, was 13.6% (corresponding to 33 women).

TABLE-US-00001 TABLE 1 Characteristics of the women taking part in the study, the booking BMI of six patients is not known. Number of women with demographic data 243 Age at booking (years) Mean (SD) 31.3 (4.9) Booking BMI (kg/m.sup.2) Mean (SD) 26.3 (5.7) Parity n (%) 0 80 (32.9) 1 89 (36.6) 2 49 (20.2) 3 12 (4.9) 4 5 (2.1) ≥5 5 (2.1) Unknown 3 (1.2) Ethnic group n (%) White 192 (79.0) Mixed 1 (0.4) Asian 23 (9.5) Black 22 (9.1) Other 5 (2.1)

Example 2—Chemical Identification

[0064] VOC analysis indicates that the presence of butyraldehyde (butanal), ethyl acetate, methyl acetate and 2-methylpropanal can be used to distinguish between samples from subjects that are GBS positive and GBS negative (see FIG. 4A and 4B).

Example 3—Statistical Analysis

[0065] FIG. 1 shows a typical output of the GC-IMS to a positive swab. The output is a 3D topographic map with each point characterised by the retention time in the chromatographic column (in seconds), the drift time in the drift tube (in milliseconds) and the intensity of the ion current signal (in millivolts). The signal intensity is indicated by colour. Each high-intensity area represents a single or combination of chemicals (with the same properties). The long line red line is the RIP (reactive ion peak), which is a background signal. The background is represented in blue with the non-blue areas showing that the instrument is detecting chemicals. The intensity of the peak (with red being the highest intensity) represents the amount of ions (and thus the chemical) detected. In general, each of the circular areas of higher intensity represent a different chemical. Furthermore, it can be see that the majority of the response is in the central section of the output.

[0066] The data from the G.A.S. GC-IMS was analysed, as describe above, and the statistical output is shown in Table 2. The results demonstrate a strong VOCs signal (from butyraldehyde (butanal), ethyl acetate, methyl acetate and 2-methylpropanal) is associated with GBS colonisation.

TABLE-US-00002 TABLE 2 Statistical output from G.A.S. GC-IMS analytical platform using a sparse logistic regression classifier. Statistical parameter Value (95% CI) Area under the curve 0.93 (0.89-0.98) Sensitivity 0.81 (0.71-0.89) Specificity 0.97 (0.91-1) Positive predictive value (PPV) 0.84 Negative predictive value (NPV) 0.97 p-value 2.05E−21

[0067] An area under the curve value of 0.93 indicates that the VOCs, butyraldehyde (butanal), ethyl acetate, methyl acetate and 2-methylpropanal, are extremely good at discerning between subjects that are GBS positive and GBS negative. This is also supported by a small p-value of 2.05.sup.21. A sensitivity of 0.81 indicates that there is a probability of 0.81 that a test result will be positive when the subject is GBS positive. A specificity of 0.97 indicates that there is a probability of 0.97 that a test result will be negative when the subject is GBS negative.

[0068] A PPV of 0.84 indicates that there is a 0.84 probability that the subject is GBS positive when the test is positive. An NPV of 0.97 indicates that there is a probability of 0.97 that the subject is GBS negative when the test is negative. These results particularly indicate that the VOCs, butyraldehyde (butanal), ethyl acetate, methyl acetate and 2-methylpropanal, are good at identifying subjects that are not GBS positive (in accordance with the first aspect of the invention).

[0069] An AUC receiver operator characteristic (ROC) curve is shown in FIG. 2.

[0070] FIG. 3 shows a box and whisper plot with the predicted probability that a sample is positive for GBS. The samples in Group 1 correspond to those taken from pregnant subjects that are confirmed as positive for GBS (by enrichment culture), and the samples in Group 2 correspond to those taken from pregnant subjects that are confirmed as negative for GBS (by enrichment culture). The small degree of overlap between the interquartile range (IQR) of the Group 1 and Group 2 indicates (visually) that butyraldehyde (butanal), ethyl acetate, methyl acetate and 2-methylpropanal,are good biomarkers for ensuring that a sample/subject is correctly identified as GBS positive or GBS negative.

[0071] Discussion

[0072] Main Findings

[0073] The results presented here demonstrate the diagnostic value of vaginal VOCs in the detection of colonisation with GBS. Such a test may be used as a point of care test for women intrapartum, reducing the incidence of EOGBS disease by appropriate administration of antibiotic prophylaxis.

[0074] VOC profiles from samples taken from pregnant women colonised with GBS could be discriminated from those who were not colonised with GBS with a high sensitivity and specificity. The results indicate that women who are colonised with GBS have chemically different vaginal swabs to those who are not colonised. The vagina has its own varied microbiome, but the data suggests that despite this, there are differences in VOCs from high vaginal swabs in those who are colonised with GBS. These GBS associated differences in VOCs were demonstrated and are detectable with this novel technology. The inventors believe, although they do not wish to be bound by the theory that the detected VOCs are gaseous waste products that occur as a result of the complex interactions in the vagina between the vaginal and cervical epithelial cells, the vagina flora and the invading GBS pathogens. The data presented shows that GBS colonisation produces a unique VOC fingerprint.

[0075] The results presented here demonstrate that the G.A.S. GC-IMS instrument has a very high specificity and negative predictive value for the detection of GBS colonisation. This technology can now be developed as a bedside test for GBS. In the acute intrapartum scenario, women could have a swab taken and analysed in a hand held device in minutes. Where the results are positive, this could guide clinicians to prompt and appropriate administration of intrapartum antibiotics, reducing the risk of EOGBS. The high negative predictive value of the test could be used to counsel families about the low likelihood of colonisation with GBS and bring into question whether administration of antibiotic prophylaxis is necessary, reducing unnecessary antibiotic exposure to both the mother and infant. Furthermore, a large number of women need to be tested for a screening program and the cost of this test is minimal.

[0076] Strengths of the results include the large number (n=607) of swabs analysed using the G.A.S. GC-IMS instrument and compared to the reference standard. STARD statement was complied with bias was minimised as far as possible. However, there were a few limitations. The prevalence of colonisation with GBS was lower than expected at 13.6%, compared with the global average of 17.9% and European average 19.0%. In the clinic, women have a vaginal swab only (as part of their screening for risk of spontaneous preterm birth), this is not in keeping with recommendations for specimen collection for detection of colonisation of GBS. Swabbing both the lower vagina and rectum increases the culture yield when compared to sampling the vagina only. Previous studies sub-analysis has illustrated that colonisation from low vaginal swabs only had a mean prevalence of 14.2%, this has more similarity to the cohort used in the Examples.

[0077] In conclusion, EOGBS disease remains the leading infectious cause of morbidity and mortality amongst neonates. Preventative efforts have reduced the burden of this disease over time but at present worldwide no universal screening tool or pathway can be agreed. This study has shown that the VOC signature present in vaginal swabs of pregnant women distinguished those swabs from which GBS was detected. Using the G.A.S. GC-IMS analytical platform with a sensitivity and specificity for GBS colonisation of 0.81 and 0.97 respectively. Development of this technology has the potential to provide clinically useful and cost-effective universal screening intrapartum for colonisation with GBS.