Advice
Evidence review
Evidence review
Clinical and technical evidence
Regulatory bodies
A search of the Medicines and Healthcare Products Regulatory Agency (MHRA) website revealed no manufacturer Field Safety Notices or Medical Device Alerts for this device. No reports of adverse events were identified from a search of the US Food and Drug Administration (FDA) database: Manufacturer and User Device Facility Experience (MAUDE).
Clinical evidence
There is a substantial body of published evidence describing the use of the Xpert GBS test for intrapartum identification of GBS. A literature search identified 10 fully published studies on the Xpert GBS. Two diagnostic accuracy studies were excluded because they did not include a suitable reference standard. This briefing focuses on 1 randomised controlled trial and 1 diagnostic accuracy study, which investigated the feasibility of using intrapartum antibiotic prophylaxis based on the results of polymerase chain reaction (PCR) testing. The results of 6 further diagnostic accuracy studies are also summarised.
The randomised controlled trial by Hakansson et al. (2014) had 2 phases. The first phase used an earlier version of the Xpert GBS test system (when reagents were added manually by the user); however, an upgraded version was used during the second phase of the study and this is likely to be the current Xpert GBS system. The first phase was a multicentre randomised controlled trial that primarily aimed to determine whether introducing the Xpert GBS test into a labour ward could direct appropriate use of prophylactic antibiotics in women considered to be at increased risk of carrying GBS. A secondary aim was to evaluate the efficacy of intrapartum antibiotic prophylaxis to prevent GBS colonisation in the baby. During this phase, 229 women across 6 delivery units in Sweden were randomised into 2 groups: 112 to group A and 117 to group B. Group A were swabbed for both Xpert GBS PCR (using the older version of the device) and conventional bacterial culture, and intrapartum antibiotic prophylaxis was administered if the PCR assay result was positive or invalid for GBS. Group B were swabbed for conventional bacterial culture only and were treated with intrapartum antibiotic prophylaxis according to the Swedish recommended guidelines, which advocate a risk‑based approach. In both groups, the babies were swabbed for conventional culture to determine their GBS colonisation status.
The second phase was non‑randomised and aimed to evaluate the performance of an upgraded version of the Xpert GBS test in 94 women across 3 of the 6 centres from the first phase. All of the women were tested for GBS using both Xpert GBS PCR (using the upgraded version of the device) and conventional bacterial culture. Intrapartum antibiotic prophylaxis was administered if the PCR result was either positive or invalid.
The results of phase 1 reported that the sensitivity and specificity of the Xpert GBS test when conclusive were 89% and 90% respectively, when compared with bacterial culture as the reference standard. However, the Xpert GBS PCR assay results were invalid in 44% (47/106) of cases with complete data (6 test results were reported missing). The authors reported a statistically significant 39% reduction in the use of intrapartum antibiotic prophylaxis in this phase (p<0.001), from 92% (107/117) in the recommended guideline group to 53% (59/112) in the PCR group.
The results of phase 2 showed a reduction in the proportion of invalid results to 15% (14/94) using the upgraded system. There was also a further reduction of intrapartum antibiotic prophylaxis administration in phase 2, from 53% of women (59/112) using the earlier version of Xpert GBS to 33% of women (31/94) using the upgraded system.
The authors concluded that the Xpert GBS test was a promising tool for improving intrapartum antibiotic prophylaxis, which would allow a significant reduction in the use of these antibiotics. A description of the study and its results are included in table 1.
Table 1 Summary of the Hakansson et al. (2014) randomised controlled trial
|
Phase 1 – Randomised |
|||
|
Group 1A – |
Group 1B – conventional culture + recommended guidelines (Control arm) |
Analysis (chi squared test or Fisher's exact test) |
|
|
Randomised |
n=112 |
n=117 |
|
|
Efficacy |
Not reported |
Not reported |
|
|
Primary outcome: Use of IAP |
53% (59/112) |
92% (107/117) |
p<0.001 |
|
Secondary outcomes: Efficacy of IAP (based on culture) to prevent GBS colonisation in the baby |
IAP given for a duration of at least 2 hours = 5/43 GBS positive infants (12%) 32 of these 43 women were given IAP for at least 4 hours = 4/32 GBS positive infants (13%) |
No p value recorded |
|
|
Secondary outcomes: Sensitivity analysis of PCR compared with culture |
Sensitivity: 89% Specificity: 90% |
Control arm |
|
|
Phase 2 – Non‑randomised (n=94) |
|||
|
Phase 1 data (original Xpert GBS PCR assay) |
Phase 2 data (upgraded Xpert GBS PCR assay) |
||
|
Primary outcome: Invalid PCR result |
44% (47/106) |
15% (14/94) |
p<0.001 |
|
Secondary outcome: Use of IAP |
53% (59/112) |
33% (31/94) |
p<0.01, OR 0.44 (95% CI 0.24–0.81). |
|
Safety |
None reported |
None reported |
|
|
Patients reporting serious adverse events |
None reported |
None reported |
|
|
Abbreviations: CI, confidence interval; GBS, group B streptococcus; IAP, intrapartum antibiotic prophylaxis; n, number of patients; OR, odds ratio; PCR, polymerase chain reaction |
|||
The diagnostic accuracy and feasibility study by de Tejada et al. (2011) was conducted in the labour ward of a hospital in Switzerland. A run‑in phase was carried out over 1 month to train the midwives to perform the Xpert GBS PCR assay on the GeneXpert system; this phase included 132 women. A further 563 women were included in the second phase of the study, resulting in a total of 695 women included for the diagnostic accuracy study. The feasibility of intrapartum testing to guide intrapartum antibiotic prophylaxis was evaluated using only the results from the second phase of the study, and included 557 women. The women included in this study were not considered to be at increased risk of GBS and were excluded if they had previously had a baby with GBS sepsis, or a positive urinary culture for GBS during pregnancy.
The study compared the results for each patient from an intrapartum Xpert GBS PCR test and antenatal bacterial culture with the results from an intrapartum bacterial culture, which was considered the reference standard. The samples used for both the bacterial cultures were from rectovaginal swabs. The sensitivities of the intrapartum Xpert GBS PCR and antenatal culture were recorded as 85.0% and 81.0% respectively. The proportion of invalid results from the Xpert GBS test was 8.4% (58/695). The authors did not report specificities of the tests because they stated that if culture is used as the reference standard, those patients considered to be false positive by PCR could in fact be positive for GBS, because the PCR assay may be more specific than culture. The feasibility study showed that 4 or more hours of intrapartum antibiotic prophylaxis could be given to 68.2% (73/107) of women whose PCR results were positive for GBS, compared with 63.6% (68/107) women whose antenatal culture results were positive for GBS (p=0.54). The intrapartum PCR assay correctly identified GBS colonisation status in 74.4% (32/43) of women who were delivering pre‑term. Of these 32 women, antenatal culture only correctly identified GBS status in 31.3% (10/32).
The authors concluded that intrapartum GBS testing is feasible and is at least as accurate as antenatal testing, but did not state whether Xpert GBS testing had any benefit over conventional bacterial culture. A description of the study and its results are included in table 2.
Table 2 Summary of the de Tejada et al. (2011) diagnostic accuracy and feasibility study
|
Study component |
Description |
|
Objectives/hypotheses |
Primary objective: to evaluate the diagnostic accuracy of the intrapartum PCR using the Xpert GBS system and antenatal culture (at 35–37 weeks) compared with the results of intrapartum culture. Secondary objective: to investigate the feasibility of implementing intrapartum antibiotic prophylaxis based on the results of the intrapartum PCR. |
|
Study design |
Prospective diagnostic accuracy study A run‑in phase was performed between December 2007 and January 2008 to train the midwives to perform the PCR assay on the GeneXpert system. Women were recruited in this phase provided they met the eligibility criteria. During the second phase, from January to April 2008, all women delivering in the labour ward were included, provided they met the same eligibility criteria. During the feasibility study, only results from the second period were used. |
|
Setting |
Labour ward in the University Hospital of Geneva, Switzerland. Women were recruited in this phase if they met the eligibility criteria. During the second phase, from January to April 2008, all women delivering in the labour ward were included, if they met the same eligibility criteria. During the feasibility study, only results from the second period were used. |
|
Inclusion/exclusion criteria |
Inclusion criteria: women considered not to be at risk of GBS colonisation. Exclusion criteria: elective caesarean section, a previous infant with GBS sepsis or a positive urinary culture for GBS during pregnancy. |
|
Primary outcomes |
Diagnostic accuracy of intrapartum PCR Feasibility of implementing IAP |
|
Statistical methods |
A power analysis calculated that 700 women would be needed to recruit 140 GBS culture‑positive women and to show a statistically significant increase in assay sensitivity between antenatal culture and intrapartum PCR testing (alpha = 0.05 and power = 90%). McNemar's test was applied to estimate the sensitivity with a 95% confidence interval for the diagnostic accuracy study. Differences in proportions and continuous variables were analysed using Fisher's exact test and student t‑test, respectively. A significance level of p<0.05 was used. |
|
Participants |
n=695 women for the diagnostic accuracy study n=557 women for the feasibility study (a subset of the 695) |
|
Results |
Diagnostic accuracy of intrapartum Xpert GBS PCR: Colonisation rate:
Of the 695 women in the study, 66 antenatal cultures were not done, and in 58, no result was obtained from the intrapartum PCR after 2 attempts. Sensitivity (reference standard: intrapartum culture):
The proportion of indeterminate test results from the intrapartum PCR was 8.4% (58/695). There were discordant results between antenatal culture and intrapartum culture in 6.9% (48/695) women. PCR was unable to resolve the result in 5 cases and was in agreement with the intrapartum culture in 62.8% (27/43) of the remaining women. Feasibility of implementing IAP: In 76.5% (426/557) of women, the results of the PCR were obtained at least 4 hours before delivery. Intrapartum antibiotic prophylaxis was administered to 26.0% (145/557) of women, 83.4% (121/145) for early onset GBS disease. Of these 121 women, 71.1% (86/121) received IAP for at least 4 hours. Intrapartum culture identified 107 GBS colonised women. In 68.2% (73/107) of these women, administration of IAP was feasible for at least 4 hours when based on PCR, and in 63.6% (68/107) when based on antenatal cultures (p=0.54). Nine of 43 women delivering pre‑term were identified as colonised with GBS. Intrapartum PCR correctly identified GBS status at least 4 hours before delivery in 74.4% (32/43) and antenatal culture correctly identified GBS status in 31.3% (10/32) of these women. |
|
Conclusions |
The authors concluded that the sensitivity of the intrapartum PCR using Xpert GBS to detect GBS during labour was slightly superior to antenatal culture techniques; however, the difference was not statistically significant. The intrapartum approach would therefore allow the identification of high‑risk groups for neonatal sepsis, such as women who are delivering pre‑term, or women who were not followed throughout their pregnancy. |
|
Abbreviations: CI, confidence interval; GBS, group B streptococcus; IAP, intrapartum antibiotic prophylaxis; n, number of patients; PCR, polymerase chain reaction |
|
Six other diagnostic accuracy studies reporting performance characteristics of the Xpert GBS test against a reference standard were identified. Each of these studies explicitly described the characteristics of the Xpert GBS test; however, it is not clear from the reports whether each study used the current version of the Xpert GBS test. Sensitivity rates ranged from 83.3% to 98.5% and indeterminate result rates ranged from 2.1% to 23.5%. These studies are briefly summarised in table 3.
Table 3 Brief summary of 6 additional Xpert GBS diagnostic accuracy studies
|
Reference |
Study aims |
Setting, number of patients and study dates |
Selected results and conclusions |
|
Tanaka et al. (2015) |
To analyse the diagnostic accuracy of the Xpert GBS assay for identification of group B streptococcus, using intrapartum culture as the reference standard. |
Hospital in Japan n=79 Japanese women January ‑ May 2014 |
Results were obtained for 73 women. The performance characteristics of the Xpert GBS assay compared with conventional culture were:
Proportion of indeterminate results: not reported. The authors concluded that intrapartum real‑time PCR for GBS screening has a similar accuracy to the antepartum conventional culture technique. |
|
Mueller et al. (2014) |
To compare the Xpert GBS assay performed in the laboratory and in the labour ward, using selective culture as a reference standard. |
Hospital in Switzerland n=300 women, (150 for phase I and 150 for phase II) January 2007 – August 2010 |
Phase I – swabs were analysed by selective culture and the Xpert GBS assay in the laboratory. Performance characteristics of the rapid PCR were:
Phase II – swabs were analysed by selective culture and the Xpert GBS assay in the labour ward. Performance characteristics of the rapid PCR were:
After initiating a 2 hour training period for operating personnel in the labour ward, the proportion of indeterminate results decreased to 13.4%. The authors concluded that the Xpert GBS assay in the labour ward yields adequate results to identify GBS colonisation; however, a short training period would be necessary to reduce the number of indeterminate results. |
|
Abdelazim (2013) |
To compare the Xpert GBS assay with the standard antepartum culture for detecting group B streptococcus colonisation. |
Hospital in Kuwait n=445 women Study dates not reported |
The sensitivity and specificity of the Xpert GBS assay to diagnose GBS colonisation were 98.3% and 99.0% respectively, compared with 73% sensitivity and 95.5% specificity for antepartum culture (p>0.05). The accuracy of the intrapartum PCR test was 98.8% for detecting GBS colonisation in comparison to 90.0% for antepartum culture. Proportion of indeterminate results: 4.5% The authors concluded that the Xpert GBS assay is an accurate test when used at the bedside and has the potential to be used as a screening test for GBS colonisation to allow for appropriate management. |
|
Young et al. (2011) |
To evaluate the Xpert GBS assay for detecting group B streptococcus. |
Labour unit in a US medical centre n=559 women January – June 2010 |
The prevalence of GBS was 19.5% by antepartum culture and 23.8% by intrapartum culture. The performance characteristics of the Xpert GBS assay (using intrapartum culture as the reference standard) were:
Proportion of indeterminate results: 2.1%. The performance characteristics of antepartum culture (using intrapartum as a reference standard) were:
The authors concluded that the Xpert GBS assay may be superior to antepartum culture for detecting intrapartum GBS which could allow for more accurate management of labour and reduce the incidence of neonatal GBS sepsis. |
|
El Helali et al. (2009) |
To evaluate the diagnostic accuracy of the Xpert GBS assay at the onset of labour using intrapartum culture as a reference standard and also compare it with screening by culture at 35–37 weeks gestation. |
Maternity ward in a hospital in France n=968 April 2007 – March 2008 |
Using intrapartum culture as the reference standard, the performance characteristics of the Xpert GBS assay were:
The authors concluded that the Xpert GBS assay is a highly accurate test that is able to determine intrapartum GBS status at point of care. The technology could improve the use of intrapartum antibiotic prophylaxis, including in women with pre‑term labour or pre‑term rupture of the membranes. |
|
Gavino and Wang (2007) |
To evaluate the Xpert GBS assay for detecting GBS in active labour in the delivery unit. |
Delivery unit of a hospital in the USA n=55 women Study dates not reported |
Performance of the Xpert GBS assay was analysed using intrapartum culture results as the reference standard. The performance characteristics of the Xpert GBS assay were:
Proportion of indeterminate results: not reported. The authors concluded that the Xpert GBS assay was highly sensitive for GBS detection in the sample population they had tested. However, corroboration of these data would be needed in a large population. |
|
Abbreviations: CI, confidence intervals; GBS, group B streptococcus; n, number; NPV, negative predictive value; PCR, polymerase chain reaction; PPV, positive predictive value |
|||
Recent and ongoing studies
One ongoing or in‑development trial of Xpert GBS was identified in the preparation of this briefing:
-
Group B Streptococcus (GBS) Polymerase Chain Reaction (PCR) Concordance Study (NCT00972894) – The manufacturer has stated that this trial is independent from Cepheid but the current status is unknown.
Costs and resource consequences
The Xpert GBS assay could be used as an intrapartum test for women considered to be at increased risk for transmission of GBS to their baby. As it does not replace an existing test, it would represent an additional acquisition cost to the NHS. However, resource‑use savings could potentially be made through the offset of staff time and by costs avoided because of a reduction in unnecessary intravenous antibiotics and potential associated complications.
Two studies were identified that investigated the cost‑effectiveness of different intra‑partum screening approaches.
A Health Technology Assessment set in 2 large UK obstetric units investigated the cost effectiveness of rapid testing for GBS in women during labour (Daniels et al. 2009). The study comprised 2 parts: a diagnostic accuracy study to establish the efficacy of the Cepheid Smart GBS system (a precursor to the present Xpert GBS system) and an optical immunoassay; both were individually compared with a bacterial culture reference test. A decision analytic model was created to assess the cost effectiveness of the technologies in various scenarios. The study found that the Smart GBS system had better diagnostic accuracy than the optical immunoassay, although neither technique was cost‑effective for managing high risk of colonisation in women during labour. The most cost‑effective option was providing intravenous antibiotic prophylaxis for all women, without screening.
The costs of implementing intrapartum PCR screening for GBS, using the Xpert GBS assay in pregnant women, were compared with the previously used antenatal culture in a study by El Helali et al. (2012). The authors concluded that the newer PCR screening method provided superior diagnostic sensitivity, and that there was a greater than 50% chance that PCR dominated antenatal culture (was more effective and less expensive).
Strengths and limitations of the evidence
There is a reasonably large evidence base for using Xpert GBS for intrapartum identification of GBS. The evidence for this technology is primarily in the form of diagnostic accuracy studies that focus on the diagnostic performance characteristics of the test, and most did not investigate the potential changes to clinical practice. This briefing has focused on 1 randomised controlled trial (Hakansson et al. 2014) and 1 diagnostic accuracy study (de Tejada et al. 2011) which described outcomes that could impact on the NHS.
The main limitation of the Hakansson et al. (2014) study was that the first phase involved a randomised controlled trial of an earlier version of the Xpert GBS test. This may have caused the high number of test results that were invalid in this phase of the study (44%). The authors reported a reduction in invalid results to 15% during the second phase of the study when an upgraded version of the Xpert GBS test, likely to be the current version, was evaluated. A large number of women were excluded for not fulfilling the eligibility criteria (106/335), which opens up the possibility of selection bias and may limit the generalisability of the study. The randomisation in this study appears to have been effective because both groups indicated similar numbers of each obstetric risk factor. The study does not appear to have been blinded to the researchers, clinicians, or participants. However, this is unlikely to have introduced significant bias because of the objective nature of the outcomes reported. The primary eligibility criteria were set to include all women in labour, who had at least 1 of the selected obstetric risk factors for GBS, but these were not the same as the risk factors used in the UK and this introduces issues of generalisability. The diagnostic analyses in this study were poorly reported, and were not replicable by the authors of this briefing, introducing the possibility of mistakes in the calculations.
Of the 7 diagnostic accuracy studies identified, the study by de Tejada et al. (2011) was the only one to assess the feasibility of implementing intrapartum antibiotic prophylaxis based on the results of the intrapartum PCR to detect GBS, using the Xpert GBS system. However, in this study healthy women in labour were screened for GBS, and the exclusion criteria included 2 obstetric risk factors for GBS colonisation, and therefore may not be generalisable to current NHS practice. A strength of the study was that the researchers included a large number of participants for both the diagnostic accuracy study and the feasibility study (695 and 557 respectively), which was based on a suitable power calculation to determine sample size. It is not clear whether the midwifery staff were blinded to the results of the antenatal culture before performing the intrapartum Xpert GBS PCR test, giving rise to potential bias, although this should be minimal as only objective outcomes were reported.
The remaining 6 diagnostic accuracy studies did not describe the potential impact or feasibility of Xpert GBS testing at the point of care, or the potential clinical and resource implications of this. The range of sensitivity rates and indeterminate result rates reported in these studies may be because of different versions of the Xpert GBS test and different culture methods used as the reference standard. The studies were also done in various international hospital settings, therefore results and indications may not be generalisable to NHS practice.
The Health Technology Assessment by Daniels et al. (2009) should be considered with caution because it used a different version of the technology, with statistically significantly worse diagnostic accuracy than has since been reported for the current Xpert GBS system.
The cost‑effectiveness study by El Helali et al. (2012) compared screening techniques and because it did not include consideration of treatment using a risk stratification algorithm, it is not generalisable to the NHS. This study was a 'before and after' study that did not feature a control arm or reference standard, and accordingly the results may be subject to significant bias. Additionally, the study was done in France and costs were reported in US dollars, further limiting generalisability.