WO2011098783A1

WO2011098783A1 - Markers for preterm labour and subsequent delivery

Info

Publication number: WO2011098783A1
Application number: PCT/GB2011/050179
Authority: WO
Inventors: Rachel Tribe; Andrew Sherman
Original assignee: Kings College London
Current assignee: Kings College London
Priority date: 2010-02-09
Filing date: 2011-02-03
Publication date: 2011-08-18
Anticipated expiration: 2012-08-09

Abstract

This is a method for identifying whether a pregnant animal has an increased risk of preterm labour and subsequent delivery, cervical shortening or PPROM comprising measuring the concentration of MCP-1 in a sample obtained from the subject, wherein an increased concentration of MCP-1 is indicative of an increased risk of preterm labour.

Description

MARKERS FOR PRETERM LABOUR AND SUBSEQUENT DELIVERY

Field of the invention

The invention relates to a method of identifying pregnant women at risk of preterm labour and poor outcome of the baby. In particular, the invention relates to a diagnostic marker found in cervico-vaginal samples which can be used to identify women likely to deliver preterm.

Background to the invention

The incidence of preterm labour in developed countries has remained unchanged over several decades, complicating approximately 7% of deliveries. Whilst survival rates have improved dramatically, morbidity remains high. Identification of those women who are at greatest risk of preterm labour and delivery and poor outcome of the baby is inaccurate and still largely based on clinical history.

Preterm labour is syndrome of multiple aetiologies (e.g. infection, multiple pregnancy, placental abruption, stress etc), and it is unlikely that a single biomarker test will identifiy all women destined to go into preterm labour. It is generally agreed that in singleton pregnancies, a large proportion (between 40-70 %) of early preterm births (e.g. < 34 weeks') are associated with infection or subclinical infection. Infection can result either from ascending infection from the vaginal cavity, haematogenous spread, or due to the emergence of a pre-pregnancy maternal infection resident in the uterine cavity (e.g in the endometrium). With regard to the former, the structural and chemical defences provided by the cervix, cervical epithelia and mucous are breeched, the cervix shortens and bacteria colonise fetal membranes and amniotic fluid. Thus, preterm labour associated with intraamniotic infection and poor neonatal outcome is proposed to more common in women with a shortened cervix. Tests that accurately assess the inflammatory environment are, therefore, likely to be useful and discriminatory and provide an earlier indication of threatened infection. Infection or inflammation may also cause neurological damage to the fetus, or increase the risk of harm to the baby following delivery, and this maybe independent of gestational age. The focus on inflammatory biomarkers discovery is supported by a series of small studies that have assessed the utility of cytokines measurement in a range of biological fluids [serum, amniotic fluid, cervico-vaginal fluid (CVF) and saliva] from women at high risk and low risk of preterm labour. Overall, these studies implicate inflammatory cytokines in preterm labour and causing fetal morbidity/mortality, and confirm their potential usefulness as biomarkers for prediction of outcome. Serum cytokine measurement, often perceived as the easiest approach, has a poor correlation with tissue and amniotic fluid concentrations. Amniotic fluid biomarkers are promising acute indicators of threatened preterm labour and fetal outcome, but the use of amniocentesis is neither widespread nor acceptable to women in Europe and the UK. In contrast, the use of CVF is highly favoured because of the link between cervical shortening and preterm labour.

Summary of the invention

According to the invention, there is provided a method for identifying whether a pregnant animal has an increased risk of preterm labour and subsequent delivery, shortened cervix or premature preterm rupture of membrane (PPROM) comprising measuring the concentration of GM-CSF or MCP-1 in a sample obtained from the subject, wherein an increased concentration of GM-CSF or MCP-1 is indicative of an increased risk of preterm labour. In particular the method may be used to identify an animal with an increased risk of preterm labour and subsequent delivery, shortened cervix or premature preterm rupture of membrane, and a likely poor outcome for the offspring.

The animal is preferable a mammal, especially a primate. In one embodiment, the animal is a human.

When the animal is a human, the term "preterm labour" preferably means spontaneous labour at less than 37 weeks' completed gestation. Preferably, it means at between 16 and 37 weeks' completed gestation, more preferably between 20 and 37 weeks' completed gestation especially between 24 and 37 weeks' completed gestation. Gestational age may be assessed using the last menstruation date and ultrasound examination.

Shortened cervix means a shortening in length of the cervix at an inappropriate time, that is to say earlier than expected during pregnancy.

Premature preterm rupture of membrane means rupture of the amniotic sac at an inappropriate time, that is to say before term. The method of the invention is particularly useful for identifying subjects in which the shortened cervix, PPROM or early delivery is as a result of infection, rather than other causes. If the subject has an infection, it may be better to allow delivery to occur, or to try to treat the infection, rather than taking steps to prevent delivery as the infection may harm the baby. Accordingly, the method can be used to identify subjects in which the baby is at risk from the infection and allow appropriate intervention to be undertaken.

The sample may be any sample obtainable from the subject such as plasma or blood, but is preferably amniotic fluid or CVF.

The concentration of progesterone in the sample may be measured by any known means., such as by ELISA.

The term "increased concentration of GM-CSF or MCP-1" preferably means that the concentration of GM-CSF or MCP-1 in the sample is higher than the expected concentration of GM-CSF or MCP-1 for that stage of gestation.

The expected concentration of GM-CSF or MCP-1 may be identified in a number of ways. For example, the sample may be compared with a comparable sample from another animal. If the concentration of GM-CSF or MCP-1 is higher than that of the comparable sample, the level may be considered to be increased.

The sample may be compared with a standard level or range of levels appropriate for the sample type and the gestation at which the sample is collected. Reference ranges may be established by testing a large number of comparable samples from a number of animals over a range of gestational ages. Such reference ranges may allow the determination of the distribution of concentration of GM-CSF or MCP-1 at a given gestation, and in particular of the centiles of the distribution.

Comparable samples are the same type of sample as each other. For example, comparable samples are taken from animals of the same species. Equally, blood samples should be compared with other blood samples, CVF samples with other CVF samples etc. When compared with a sample from another animal or with a standard level, the increase in the concentration of GM-CSF or MCP-1 is preferably an increase of at least 100%. The GM- CSF concentration is preferably at least twice, more preferably at least five times, even more preferably at least ten times the expected concentration. The MCP-1 concentration is preferably at least twice, more preferably at least three times, even more preferably at least four times the expected concentration. Alternatively, in the instance that the presence of GM- CSF in the sample is not expected, the simple detection of GM-CSF could be considered to be an increase in the terms of the invention. Alternatively, the sample may be compared with a projected or likely concentration of GM- CSF or MCP-1, established by testing the same subject at an earlier date and predicting the concentration of GM-CSF or MCP-1 that should be reached by the relevant stage of gestation. The sample may also simply be compared with earlier samples obtained from the subject to identify a change, especially an increase in concentration of either GM-CSF or MCP-1 or both.

When testing samples from the same woman, the samples may be obtained repeatedly, such as daily, every two or three days, weekly, or fortnightly for example. The method of identifying subjects at risk may be used in combination with other methods that identify subjects at risk, such as checking fetal fibronectin levels, progesterone concentrations in the sample and observing for a shortened cervix.

The method of the invention may also be used to identify subjects that may benefit from treatment with progesterone, cervical stitch or other intervention.

In order to identify subjects at risk and to be able to take action to delay or avoid preterm labour, it is advantageous to test samples taken early on during pregnancy. In human women, it is preferable to test samples taken at between 14 to 16 weeks and 30 to 34 weeks of gestation, more preferably between 16 and 28 weeks, even more preferably between 16 and 24 weeks.

Also provided is a kit for assaying for concentrations of GM-CSF or MCP-1 in samples obtained from patients, especially amniotic fluid or CVF samples. The kit preferably contains an assay for one or both or GM-CSF or MCP-1, that can be used in amniotic fluid or CVF samples. Preferably the assay can be used as a point of care assay, providing results in less than 60 minutes. The invention will now be described in detail, by way of example only, with reference to the figures.

Figure 1. Median longitudinal cervico-vaginal fluid (CVF) concentrations of GM-CSF in women with a short cervix (biophysical indicator of risk of preterm labour) compared to those women with a normal cervical length. Women with a normal cervix have little or undetectable GM-CSF concentrations in CVF samples. Women with a shortened cervix, even following treatment have measurable concentrations of GM-CSF and overall higher median values of GM-CSF in early and mid pregnancy. Figure 2. Median longitudinal cervico-vaginal fluid (CVF) concentrations of MCP-1 in women with a short cervix (biophysical indicator of risk of preterm labour) compared to those women with a normal cervical length. Median MCP-1 CVF concentrations are generally higher in women with a shortened cervix compared to women with a normal cervix from early to mid pregnancy.

Table 1. Descriptive data demonstrating increased cervico-vaginal fluid values of GM-CSF in women destined to develop a short cervix compared to those women whose cervix remained normal between 14 and 23 weeks of gestation.

Gestation (w eeks)

^Patient GM-CSF 16 18 20 22 24 group

N 16 31 35 36 1 7

% measurable

50 35.5 37.1 30.5 35.3 concentrations

Women .. ,.

... Median

wⁱth a inn/ml) .3 0 0 0 0 normal IP / J

cervix Lower 25^th

0 0 0 0 0 centile

Upper 25^th

9.6 4.8 3.2 0.5 4.1 centile

Women N 13 20 18 9 4

Table 2. Descriptive data demonstrating increased cervico-vaginal fluid values of MCP-1 in women destined to develop a short cervix compared to those women whose cervix remained normal between 14 and 23 weeks of gestation.

Study protocol

Women were recruited from two London based preterm surveillance clinics (St Thomas' Hospital and Queen Charlotte and Chelsea Hospital) and women with at least one preterm delivery between 14 and 34 weeks were eligible for recruitment. The participant was then seen initially on a two-weekly basis between 14 and 28 weeks' gestation. At each visit, symptoms were assessed, a high vaginal swab and blood samples obtained and a transvaginal ultrasonographic cervical length measurement was taken. In a small cohort of women, endocervical swabs were obtained. If the cervical length measurement was found to be short, i.e. <2.5 cm and the participant was <24 weeks' gestation, she was randomised to receive either of two routinely used interventions: cerclage (Group 2) or vaginal progesterone (Group 3) using an internet-based randomisation system. Following the random allocation of intervention, the participant was scanned and sampled on a weekly basis.

Participants did not know which treatment they would be allocated to prior to recruitment. Those who did not exhibit cervical shortening and thus did not receive an intervention were also followed for the duration of the study and samples stored for analysis. These women were designated the control group (Group 1). Data including patient demographic details, obstetric history, symptoms, cervical length measurements, hospital admissions and medications were recorded on a data capture sheet and this was later transferred to the computer database (see below for further details of the database).

Three groups of women were studied:

Group 1: Control group - women who did not exhibit cervical shortening on transvaginal ultrasound surveillance before 24 weeks' gestation.

Group 2: Short cervix with cerclage - women exhibiting cervical shortening prior to 24 weeks' gestation randomised to receive a cervical cerclage.

Group 3: Short cervix with progesterone - women exhibiting cervical shortening prior to 24 weeks' gestation randomised to receive progesterone.

There is very little published data on the behaviour of inflammatory markers or other biomarkers over time during pregnancy and a lack of information concerning the correlation of repeated measurements and cervical length in the same women at different gestations. Nonetheless, the power calculations for recruitment numbers were performed on published data from Tanaka et al. (1998) who report IL-8 concentrations (one inflammatory mediator measured in the current study) in preterm labour as 2318 pg/ml (SD 1192) and preterm not in labour as 496 pg/ml (SD 323). Inspection of their data showed that a log transformation is appropriate. Based on these values, it was possible to estimate the mean and SD of logioilL- 8) in the 2 groups as 3.37 (SD 0.30) and 2.70 (SD 0.32). We surmised that randomised treatment with progesterone might lower or inhibit the rise in IL-8 levels by 33%, as compared to cerclage, a difference of 0.176 in log₁₀(IL-8). We also assumed that there would be at least one pre-treatment measurement available for all subjects, with a pre-post correlation of 0.5 or more. 40 subjects per group would give us 84% power to detect such a difference. For the comparison of inflammatory markers between groups (prior to cervical change/intervention), with 80 subjects in the 'cervical change' group (Group 2 and 3) versus 40 subjects in the 'no cervical change' group (Group 1), it was thought possible to detect a 50% higher concentration of IL-8 in the first group with a power of 84%. These power calculation were conservative as in both cases, the collection of repeated measurements (approximately 6 per patient) would increase the power to detect a difference, as well as providing temporal information of inflammatory markers throughout gestation in relation to changes in cervical length. Because of the lack of detailed data, no allowance was made for this in the calculations above. On completion of study recruitment and full data entry, data stored on the internet database was transferred to the statistical package Stata (StataCorp, College Station, Texas) for analysis. Standard checks were carried out for non-normality (Normal distribution plots, Shapiro Wilks test) and appropriate transformations carried out. Estimates and tests of differences, both between and within subjects, used linear regression adjusting for repeated measures and non-normality through the use of generalised estimating equations. Where concentrations were below the limit of detection of the test, interval regression was used, treating the true concentration as being an unknown value between 0 and the lowest non-zero concentration detected in any sample. Initial results have been obtained through the use of Microsoft Excel and Sigma Plot vlO.O (SPSS Inc, 2006)."

Venous blood sample collection and preparation

Collection of blood samples

Venous blood samples were required for analysis of inflammatory markers and gene polymorphisms. Venepuncture using a vacutainer system to obtain 24 ml of blood evenly distributed in the 4 collection tubes was successfully achieved in the antecubital fossa according to standard protocols. The tubes were inverted gently, avoiding shaking, and kept at room temperature for 5 min before being stored on ice until centrifugation.

Preparation of blood samples Sample preparation was completed within 1 hour of sample collection. Samples were centrifuged at 2,600 g (4,000 rpm) for 10 minutes at 4°C on a standard centrifuge (Centrifuge 5702R, Eppendorf, UK). The resultant plasma from the EDTA blood sample was pipetted into 750 μΐ aliquots. The remaining buffy coat (leucocytes) from the one EDTA blood sample was pipetted into a 2 ml plastic microfuge tube. The remaining pellet of erythrocytes from one tube was pipetted into another tube. The erythrocytes and the buffy coat in the other tube were left in situ and the tube covered for genetic analysis. The resultant serum from the SST-gel bottles was pipetted into 750 μΐ aliquots. The tubes were labelled with the participant's trial number (e.g. 1001), date of birth, date of sample collection and type of sample using a label maker (Bradyprinter BP 1344, Datamax Corp, UK). The samples were stored at -80°C.

Vaginal swab sample collection and preparation

Collection of vaginal swabs

The subject was asked to lie on the couch with a sheet placed over her abdomen. The subject was asked to place the soles of her feet together and draw her knees up towards her abdomen. The labia were gently parted and the vaginal speculum was gently inserted into the vagina. Swabs for clinical use were taken first if indicated, followed by the high vaginal swab and the endocervical swab (if required). When swabs were taken for research purposes, they were rotated in the posterior vaginal fornix (HVS) or the endocervix (ECS) for 10 seconds to achieve saturation. Once this was complete, the speculum was gently removed and excess lubricant was wiped away.

Two pre-labelled pre-prepared vials of 750 μΐ of standard phosphate -buffered saline solution containing protease inhibitor [1 protease inhibitor cocktail tablet (Complete, Roche Diagnostics GmbH, Germany) dissolved in 50 ml standard phosphate-buffered saline solution (Sigma- Aldrich Company, Ayrshire, UK] were gently warmed to room temperature for a few seconds. The swab tips were placed in the vials and then on ice for immediate transport to the laboratory or placed in the centrifuge and prepared prior to being stored at -80°C. Following the procedure, all used equipment was disposed of appropriately.

Preparation of swab samples from high vagina and endocervix

The swab tip was removed and placed into a clean empty tube. The tube containing the vaginal or cervical mucus in phosphate-buffered saline solution with protease inhibitor was vortexed for 30 s and placed on ice. The tube containing the Dacron swab tip was centrifuged at 2,600 g (4,000 rpm) for 10 minutes at 4°C on a standard centrifuge (Centrifuge 5702R, Eppendorf, UK) to collect any extra mucus that was adherent to the tip. The extra sample was collected and pipetted into the eppendorf tube containing the cervical mucus sample. The resultant fluid was vortexed for 30 s and then 110 μΐ was aliquotted into 2 ml plastic microfuge tubes (high vaginal sample and endocervical samples). The samples were frozen at -80°C until analysis.

Technique of transvaginal ultrasound for cervical length assessment

Cervical length was assessed transvaginally in order to obtain accurate measurements (after the swabs had been taken). This is preferable to the abdominal route because in 50% of cases, the cervix cannot be visualised unless the bladder is full and a full bladder can artifactually increase the true cervical length (Podobnik et al. 1988; Andersen 1991; To et al.

2000) . On the whole, women find transvaginal ultrasound assessment acceptable (Clement et al. 2003). It is also highly reproducible and in 95% of observers the difference between two measurements by the same observer and by two observers is 3.5 mm or less and 4.2 mm or less, respectively (Heath et al. 1998).

A sagittal view of the cervix was obtained and the cervical length measured using optimal magnification and zoom settings without causing undue pressure on the cervix with the probe. The echogenic endocervical mucosa allowed identification of both the internal and external os. The probe was placed in the anterior fornix and then advanced until the cervix was seen to indent. By withdrawing the probe to avoid exerting undue pressure, artificial lengthening of the cervix was avoided.

There was an area of echodensity that represents the external os and a 'V -shaped notch in the internal os that demonstrated the points at which to place the calipers. When the cervical canal was not straight, it was thought reasonable to take a straight line between these two points, as a curved appearance usually indicates a long and normal canal and the length 'missed' is not important. If the cervix is short, the appearance tends to be linear (To et al.

2001) . The linear distance between the external and internal os was recorded 3 times over a minimum of 3 minutes in order to recognise any dynamic changes. In 1% of cases, dynamic changes related to uterine activity are present and the shortest measurement can be recorded (Nicolaides et al. 2004). Funnel depth and width with concomitant changes exerted with pressure were recorded. The total closed length and the closed length cranial to a cerclage were recorded in women who received cervical cerclage. The closed length above the cerclage is the best predictor of outcome: cervical lengths of less than 10 mm are associated with the best prognosis (Groom et al. 2002).

Insertion of transvaginal cervical cerclage

Following admission to the delivery suite, the woman was given a regional block and taken to theatres. After being placed in the lithotomy position, the cervix was adequately exposed with the aid of an experienced assistant who facilitated optimal exposure of the operative field throughout the procedure.

The cervix was exposed by using a Sim's speculum and grasped using two sponge forceps, one on the anterior lip and one on the posterior lip. The cervix was then gently drawn downwards or from side to side to expose all parts so that they could be reached with the needle. Mersilene tape attached to two large needles was used. The first large needle with Mersilene tape threaded onto it was passed through the substance of the cervix as high as possible around the cervix, while avoiding the bladder. On average, 4 to 6 'bites' were taken. The first 'bite' was taken by commencing at the 3 o'clock position and passing the needle, convex to the body of the cervix to the 12 o'clock position and secured by the assistant. The second needle was used to take a 'bite' from 3 to 6 o'clock, 6 to 9 o'clock and 9 to 12 o'clock. The suture was then secured anteriorly and a tail was left, facilitating ease of removal when necessary. Practice determined the correct tension which should ideally be tight enough to close the os, but as cervical tissue is dynamic, it is thought that it would be difficult to cause adverse effects with firm tension (Chandiramani and Shennan 2007b). The original description by McDonald involved multiple 'bites' (McDonald 1957) but in practice, clinicians usually adjust their technique depending on factors such as access.

Early complications of cerclage insertion include abdominal pain, vaginal bleeding, premature prelabour rupture of membranes (PPROM) and pregnancy loss. Late complications include chorioamnionitis, suture displacement, maternal infection, preterm delivery, miscarriage, uterine rupture, increased use of tocolytics and cervical dystocia (Rush et al. 1984; Drakeley et al. 2003; Jongen and van Roosmalen 1997). Participants with a cerclage in situ were advised to attend the hospital early if in threatened preterm labour in order to ensure ample time was available for cerclage removal, thereby avoiding cervical trauma secondary to contractions and resultant increased risk of preterm birth in subsequent pregnancies.

The cerclage was removed electively between 37 and 38 weeks' gestation usually without an anaesthetic. If the participant presented in symptomatic preterm labour, the cerclage was removed as soon as possible, before labour became established. In order to remove the cerclage, the cervix was exposed using a Cuscoe's speculum in order to visualise the points where the Mersilene tape was visible on the exterior of the cervix. The 'tail' left at insertion facilitated removal, allowing the appropriate point to be visualised and cut with subsequent removal of the suture material. The incidence of difficulty in removal of cerclage is 1% (MRC/RCOG Working Party on Cervical Cerclagel993).

Cytokine profile multiplex methodology

Multiplex assay for experiment of cytokine breakdown using 3-plex fluid-phase immunoassay and for samples of a cohort of women using 17-plex fluid-phase immunoassay

Using 3-plex fluid-phase immunoassay, the plasma samples taken at different time intervals with and without cellular antigen stabilising reagent were simultaneously analysed for the following cytokines: IL- Ιβ, Π-6 and IL-8. Cervicovaginal mucus samples from nine women at 4 time points from 14 to 28 weeks of gestation were simultaneously analysed for the following cytokines: IL-Ιβ, 11-2, IL-4, IL-5, IL-6, IL-7, IL-8, IL-10, IL-12, IL-13, IL-17, G- CSF, GM-CSF, IFNy, MCP-1, ΜΠ>-1β and TNFa, using 17-plex fluid immunoassay. This was done in order to determine which cytokines would be analysed in the final analysis. Custom kits (BioRad Laboratories) were used and run on a Luminex-100 cytometer (Luminex Corporation, Austin, Texas), equipped with StarStation software (Applied Cytometry Systems, Dinnington, UK).

Assays were performed following manufacturer's instructions. Briefly, the cytokine standard was reconstituted and serially diluted. The multiplex bead working solution was prepared. The 96-plate well was prewetted with the assay buffer and then removed by vacuum filtration. This was repeated using the multiplex bead working solution and subsequently, the assay buffer. Plasma (50 μΐ) was pipetted into each well, covered with aluminium foil and incubated on a microplate shaker for 30 min at room temperature. At the end of the first incubation, the buffer was removed by vacuum filtration and each well underwent 3 washes with assay buffer ensuring vacuum filtration to remove buffer after each wash. Detection antibody solution (25 μΐ) was added to each well, covered and incubated for 30 min at room temperature. Buffer was removed by vacuum filtration and the plate washed 3 times with assay buffer ensuring vacuum filtration to remove buffer after each wash. Streptavidin-PE (50 μΐ) was added to each well, covered and incubated for 10 min at room temperature. Buffer was removed by vacuum filtration and the plate washed 3 times with assay buffer ensuring vacuum filtration to remove buffer after each wash. The beads were re-suspended in each well with assay buffer (125 μΐ), covered and incubated for 30 s before the plate was read on the Bio-Plex system. All samples were analysed in duplicate. Cytokine levels were expressed in picogrammes per millilitre.

Results

Women destined to develop a short cervix (n=35), considered to be indicative of the likelihood of pre-term delivery, exhibited significantly higher CVF concentrations of GM- CSF (16 times greater, CI: 1.7 to 148; p=0.014) and generally higher MCP-1 (4.1, CI 0.9 20.0; p=0.077) concentrations at <24 weeks' gestation than controls (n=43). Both GM-CSF and MCP-1 CVF concentrations were significantly higher (15.7 and 4.5 times greater) at the time of cervical shortening compared to controls.

Progesterone treatment had little effect on CVF cytokine concentrations, whereas the concentration of five cytokines were higher in women randomised to cerclage versus progesterone (p<0.05). Cerclage, but not progesterone treatment, was followed by an increase in cervical length of 11.4 mm (CI: 7.3 to 15.4; p<0.001).

Claims

A method for identifying whether a pregnant animal has an increased risk of preterm labour and subsequent delivery, cervical shortening or PPROM comprising measuring the concentration of MCP-1 in a sample obtained from the subject, wherein an increased concentration of MCP-1 is indicative of an increased risk of preterm labour.

The method of claim 1, wherein the concentrations of both MCP-1 and GM-CSF are measured and wherein an increased concentration of both MCP-1 and GM-CSF is indicative of an increased risk of preterm labour.

The method of claim 1 or claim 2, wherein the sample is CVF.

The method of any preceding claim, wherein the animal is a human.

The method of any preceding claim, further comprising assaying for the presence of another indicator of pre-term delivery.

A kit for assaying for concentrations of MCP-1 and GM-CSF in a sample obtained from a subject, comprising an assay for both MCP-1 and GM-CSF, that can be used in a CVF sample.

A kit according to claim 6, wherein the assay is a point of care assay.