Correlation of cervical progesterone levels to plasma progesterone levels in normal pregnancy and preterm labor: A cross-sectional study [version 2; peer review: 2 approved with reservations]

Background: The theory of “functional progesterone withdrawal” explains the role of progesterone prior to delivery. Previous studies mentioned the existence of progesterone regulation in the cervix that plays a role in maintaining the integrity of the cervix and cervical ripening. Cervical progesterone levels relate to activities of progesterone at the cervix, compared to its amount in circulation. The objective of this study was to measure cervical mucus progesterone levels and its correlation to plasma progesterone levels in pregnancy. Methods: This was a cross-sectional study conducted in January-September 2010 at Persahabatan Hospital. The subjects were pregnant women in the 28 th – 34 th weeks of gestational age. In total, 72 subjects who met the criteria were divided into normal pregnancy group and preterm labor group. The cervical and plasma progesterone levels were measured using The Advia Centaur® Progesterone kit, which is a commercial immunoassay with direct chemiluminescence method. Results: There was a positive correlation (r=0.539) between cervical progesterone levels with plasma progesterone levels in the preterm labor group. There was no correlation between cervical progesterone levels with plasma progesterone


Amendments from Version 1
We have made several changes in version 2 of our manuscript. The first two changes are the changes in our title and authors. First, we changed our title to avoid misunderstanding due to language barrier. Second, due to certain reasons and mutually agreed upon between us, Ms. Mariana did not include her name as the author of this study.
Afterward, we have made significant changes to our manuscript. We have added a comparison of both cervical and serum progesterone level between normal pregnancy and preterm labor. However, the results were not significant. We have also added the possible reason and the prior studies about the issue. Moreover, there are some detailed additional description to our method and results. There are also some comparisons with similar research done in other countries. We have acknowledged that there were some limitations to our study, of which we addressed at the later part of our manuscript. We have also acknowledged that our previous conclusions were an overstatement. Thus, we have revised the conclusion according to our results. We would like to thank the editors and reviewers who have given tremendous input to our manuscript. Finally, we would also like to thank the readers who have seen our manuscript and given their opinion directly to the authors.

Introduction
Preterm birth contributes to various long and short-term complications to neonatal health problems. The incidence of preterm birth has not shown a significant decline. About a decade ago, the incidence of preterm birth was reported around 12-13% in United States and 5-11% in developed countries 1,2 . Meanwhile, the rate of preterm birth is 15.5% in Indonesia. Indonesia is also included in the list of 10 countries with the highest preterm births according to the World Health Organization 2013 3 .
The hormone progesterone maintains the continuity of pregnancy in order to prevent preterm labor and preterm delivery. Progesterone is a steroid hormone which has an important role of maintaining uterine quiescence until the pregnancy reaches term. A previous study stated that circulatory progesterone level in humans remains stable until placenta is born. Nevertheless, the role of progesterone in the onset of labor is still believed to occur through an indirect mechanism called "functional progesterone withdrawal" 4,5 .
Circulation of progesterone hormones may enter the peripheral tissue such as salivary glands and cervix. Progesterone regulation also occurs in uterus and cervix. Uterus produces 5α-reductase enzyme and 20α-hydroxysteroid dehydrogenase enzyme (20α-HSD) which converts progesterone into inactive metabolites and decreases the uterine quiescence 6 Mahendroo et al. (1999) showed that in the uterus of full-term rat with 5α-reductase deficiency, the uterus remains to contract although delivery does not happen because there is no cervix maturation process 6 .
Andersson et al. (2008) showed that during pregnancy, cervical glandular epithelial cells produce 17β-hydroxysteroid dehydrogenase (17β-HSD) type 2 enzyme that converts estradiol to estrone and 20α-hydroxyprogesterone (20αP) to progesterone 7 . Approaching the delivery, regulation of 17β-HSD type 2 decreases and creates a state that supports cervical maturation. The data on that study supported the idea that cervical maturation and myometrial contraction in labor involve regulation of progesterone in the cervix and uterus through the complex relationship between cervical epithelial, stromal, and myometrium 7 . Other studies also explored the possibility of tissue progesterone receptor expression and function changes becoming the reason of the functional progesterone withdrawal 8 .
Progesterone plasma level is widely studied to monitor the function of the corpus luteum in secretory phase and the beginning of pregnancy. In later pregnancy, progesterone level monitoring is only done for individuals with a high risk of having abortus or preterm delivery, although the effectivity of its application in clinical practice remains controversial 9 .
Assumption about the reduced quantity of plasma progesterone before delivery are now widely questioned. Currently, the role that progesterone plays before delivery is described as "progesterone functional withdrawal" theory. Prior studies mention that there is progesterone regulation in the cervix, which contributes to preserve cervical integrity as well as cervical maturation 6,7 . It is assumed that cervical progesterone levels reflect the target organ's progesterone activity, i.e. the cervix, better than circulatory progesterone. Therefore, measuring cervical mucus progesterone during normal pregnancy and preterm delivery pregnancy is a feasible study to show the significance of the test.
This study aims to investigate the correlation between cervical progesterone level and plasma progesterone level in both normal pregnancy and pregnancy with preterm labor. Based on previous studies, it was expected that a positive correlation would be found between the two measurements on both study groups.

Study design
This study was a cross-sectional correlation study conducted in January 2010 until September 2010 at Persahabatan General Hospital Jakarta, Indonesia.

Participants
The participants of this study were women with 28 th -34 th weeks of pregnancy and singleton fetus. They were divided into 2 groups, namely the normal pregnancy and preterm labor groups. The inclusion criteria for this study were as follows: (1) pregnant women on 28 th -34 th weeks of pregnancy with singleton fetus, (2) Preterm labor group subjects had a minimum of 4 contractions in 20 minutes that was proved by cardiotocography, without amniotic membrane ruptured or blood mucus in cervix, (3) Normal pregnancy group subjects were pregnant women with matched age and gestational age who had no contraction history during this pregnancy (4) subject willing to take part in research and signed informed consent. The exclusion criteria for this study were as follows: (1) not willing to be a participant in this study (2) having fetal membrane rupture which was proved by history taking and speculum inspection (3) having uterine of cervical pathology (leiomyoma, cervical cancer) which were proved by history taking and physical examination.
The sample size was derived from the correlations study sample size formula A sample of at least 12 subjects in each subgroup is estimated to be needed in order to detect a moderate correlation (r = 0.75) with 80% power and 5% confidence limit. Based on this calculation, we enrolled 72 participants, 36 subjects in the normal pregnancy group and 36 subjects in the preterm labor group.

Data collection
The written informed consent was signed by the patient and the partner. After the patients signed the approval sheet, history taking and speculum examination were performed to ensure that there was no additional pathology on the cervix. Afterward, venous blood (5 mL) was collected from patients to measure progesterone plasma and cervical. The blood samples were centrifuged on 1,000 g for 15 minutes to obtain the plasma/serum. Cervical mucus was collected using prism-shaped MQA ophthalmic sponge which was supported by previous study by Philip et al. (2004) 10 . The speculum was inserted inside the vagina on lithotomy position to visualize the cervix with enough lighting to inspect. The ophthalmic sponge was hold on the base of the prism using long forceps/tweezers and collection of cervical mucus was performed by inserting the ophthalmic sponge into the external orifice of the uterus and holding still for 60 seconds until the sponge absorbs the mucus. The sponge was removed gently and put inside a storage tube, a neutral vacuum tube filled with 1 mL of 0,9% NaCl to store the specimen. The tube was put inside a cooler (2-8°C) within 30 minutes of collection to prevent contamination and the tube was put inside a freezer within 4 hours after collection. Before processing, samples were centrifuged at 16,000 g to obtain the supernatant (to separate sample from sponge).
The plasma blood and cervical mucus progesterone levels were measured using The Advia Centaur® Progesterone kit, which is a commercial immunoassay with direct chemiluminescence method. The samples were diluted with 1:10 dilution for quantifying the progesterone concentration. This kit had 0.21 µg/L (0.67 nmol/L) analytical sensitivity. The score used in the analysis were between 0.21 µg/L to 60.00 µg/L.

Statistical analysis
After all the necessary data was collected, the data was coded on pre-arranged coding sheets by the principal investigator.
Statistical analyses and data entry were performed using the Statistical Package for the Social Sciences (SPSS), version 20.0. We performed statistical analysis using a Spearman test for measures the correlation or the strength of association between cervical progesterone level and plasma progesterone level in normal pregnancy group and the preterm labor group. The results were presented in tables and figures.

Ethical considerations
This study was approved by The Research Ethics Committee of Persahabatan General Hospital (approval number 01/Diklit-RSP/Kom.Etik/II/2010). The research proposal was submitted and reviewed by the Research Committee, whereby permission was granted to conduct the research. A written letter of consent was submitted to the health institution to seek permission to conduct this study. Privacy and confidentiality of the subjects' information was done through the use of data collection with coded identification numbers. All prospective subjects received an explanation from the main researcher and additional researchers regarding the procedures for conducting research. The decision to follow or refuse to follow the research was taken by informed consent. All data will be kept confidential and the subject had the right to know all the results of the examination carried out.

Results
This study enrolled a total of 72 pregnant women who met the study criteria. They were divided into two groups: 36 women in the normal pregnancy group and 36 women in the preterm labor group. Table 1 showed that there were no significant differences in age distribution, gestational age, education level, and working status between groups. Table 2 shows that the median of the cervical progesterone level of normal pregnancy group was 1.74 ng/ml, which is lower than preterm labor group (median: 1.91 ng/ml). The range of progesterone levels in the preterm labor group was greater than in normal group. Plasma progesterone levels in both groups had normal data distribution. Mean value of normal pregnancy group was 174.52 ± 59.11 ng/ml, lower than preterm labor group (195.10 ± 82.21 ng/ml). However, the differences of both cervical and plasma progesterone level between normal pregnancy and preterm labor subjects were not significant (p > 0.05).
The correlation between cervical progesterone levels and plasma progesterone levels in normal pregnancy group were shown using a scatter plot graph ( Figure 1). Although schematically there was a trend of increasing relationship between cervical and plasma progesterone levels, the statistical test proved that there was no correlation between the two tests (p=0.251; r=0.196). The correlation between cervical progesterone levels and plasma progesterone levels was significant in the preterm labor group compared to the normal group ( Figure 2). The test showed a moderate correlation between the two tests (p=0.001; r=0.539). This result indicated that elevated level of cervical progesterone was directly proportional to plasma progesterone level.

Discussion
As a pregnancy hormone, progesterone is not only metabolized in circulation, but also in the uterus. Previous studies have proven the existence of local regulation of progesterone in uterus and cervix. Full-term uterus produces an enzyme that convert progesterone into inactive metabolites and decreases the uterine quiescence 6,7 . In this study, the cervical progesterone level could be measured through the cervical mucus, which was collected using a prism-shaped ophthalmic sponge. However, the range of the data was quite large, especially in the preterm labor group with a median of 1.  11 . Some of the possible causes of this abnormal data distribution were race difference and bias which could occur in data collection and analysis. The presence of cervical local factors (such as clinical or subclinical infection) might affect the regulation of progesterone. However, it was not explored in this study.
Progesterone is a hormone that plays a dominant role to keep the uterus quiescence during pregnancy. Circulatory progesterone level in humans remains elevated until birth, which refers to the notion of a "functional progesterone withdrawal", which occurs before delivery. The definition may also apply to cervical progesterone. Progesterone is inactivated by local regulation of the cervix, not by the quantity of the progesterone hormone. The study by Andersson et al. was one of the studies that proved the existence of progesterone inactivation in the uterus and cervix before delivery 7 . Using immunohistochemical methods and quantitative real-time PCR, this study proved the existence of elevation of 20α-hydroxyprogesterone (20αP), an inactive metabolite of progesterone, just before delivery. This elevation was allegedly caused by decreasing concentrations of 17β-HSD type 2 enzyme in endocervical epithelial cells 7 . However, the method that we used was different from the methods above. We tried to measure progesterone level from cervical mucus by using competitive immunoassay with direct chemiluminescence method.
In the present study, cervical progesterone levels in the normal group were lower than the preterm labor group. These results were inversely related to the expected hypothesis where cervical progesterone levels in preterm labor group were expected to be lower than the normal pregnancy group. However, the difference between groups were not significant. It is in contrast with other studies as ones performed by Romero et al. (2016) or EPPPIC group (2021), in which the progesterone was used as a prophylaxis for preterm labor 12,13 . It is suspected that the small sample size and other factors presenting during the examination influenced the results of this study.
The hypothesis of our study, which was based on the assumption that the cervical mucus was similar with saliva and the study conducted by Connor et al. in which it was stated that salivary progesterone concentrations between 24 and 34 weeks of pregnancy were lower in pregnant women who experienced delivery before 34 weeks of gestation compared to pregnant women who experienced delivery after 37 weeks, were not proven in this study 14 . The salivary progesterone levels in the study by Connor et al. were measured serially, while in this study progesterone level were measured one-time. Therefore, to get better data, serial measurements of cervical progesterone level using a good method in the pregnancy group might be needed 14 .
High cervical progesterone levels in the preterm labor group in the present study might be caused by an increase of progesterone metabolites in the cervix due to infection process. Infection rates contributed more than 30% as a cause of preterm labor. Mahendroo et al. showed conversion of progesterone into inactive metabolites by cervical enzymes before delivery in pregnant rats 6 . Using radioimmunoassay examination technique (RIA), this study detected conversion of progesterone into 5a-pregnan-3,20-dione by steroid 5a-reductase, into 4-pregnen-20a-ol-3-one by 20a-hydroxysteroid dehydrogenase and into 5a-pregnan-3a, 20a-diol by the combined action of two enzymes with 3a-hydroxysteroid dehydrogenase. This present study used a competitive immunoassay technique with direct chemiluminescence method which did not differentiate active progesterone to its inactive metabolite produced in the cervix. Detected progesterone level in cervical mucus possibly includes progesterone and its inactive metabolites and therefore did not reflect the actual activity of cervical progesterone. However, an additional assays with other kit was not performed in this study. Therefore, additional examination with different kits would be beneficial for future studies. Another explanation to the higher progesterone level in our preterm group was the increased metabolism of progesterone in the uterus and cervix 15 . However, it could not be proven in our study as the metabolism was not measured.
In the normal pregnancy group, although schematically there was a trend of increasing relationship between cervical and plasma progesterone level, the result generated by the correlation test showed that there was no significant correlation with weak strength. Different results were obtained in the preterm labor group: the relationship between cervical progesterone and plasma progesterone were more visible and the result obtained from the test showed a significant relationship as well as moderate correlation strength (r=0.539). This result indicated that the elevation of cervical progesterone level in the preterm labor group was directly proportional to plasma progesterone levels. Differences occurred in both groups require further study.
The limitation of this study was its small sample size and several possible confounding variables which might affect the results of the study. Was the result influenced by the regulation of local progesterone that occurs in the cervix? As known from previous studies, the regulation of cervical progesterone during pregnancy was different than before delivery. Would inflammation/infection in the cervix influences cervical and plasma progesterone level? There were some factors which might be controlled in future studies. Moreover, a one-time sample collection was also thought to be possible factor influencing the results of both groups.

Conclusion
Cervical progesterone levels can measured through the cervical mucus. A significant positive correlation was only found between cervical progesterone level with plasma progesterone level in the preterm labor group.

Open Peer Review
Despite having high serum and cervical progesterone, some women may be at increased risk of preterm labour due to increased metabolism of progesterone in the target tissues (Patil et al., 2020 3 ) or other mechanisms. expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.

Author Response 30 Sep 2022
Yuyun Lisnawati, Persahabatan General Hospital, Jakarta, Indonesia 1. The aim of the study and the hypothesis were added to the study.
2. We thank you for the comment toward our manuscript. We are also aware to the problem and made the changes accordingly 3. The additional statistical test was performed in our study. However, there were no significant difference of both cervical and serum progesterone between study groups. We have also added the discussion to our discussion part.
4-5. Added to the respective section.
6. We are also aware to the issue. The statement has been omitted.
7. Various grammatical errors and confusing sentences were altered accordingly.
We hope that our response will sufficiently address your concern toward our manuscript. If you have any other comment or recommendation, please let us know. We thank you once again for your cooperation.

Yuyun Lisnawati
Competing Interests: There was no competing interests to be disclosed.

Pei F. Lai
Division of Reproductive and Developmental Biology, Imperial College London, London, UK In this article, the authors highlight that cervical progesterone action during human preterm labour is best characterised by profiling the local (rather systemic) availability of this steroid in pregnant women, which is attributed to the concept of functional progesterone withdrawal. Their aim was to determine whether the correlation in progesterone concentration between maternal peripheral blood plasma and cervical mucus is different between non-labouring and preterm labouring women, when analysed at 28-34 weeks of gestation.
The effort made to obtain preterm gestation-matched samples for comparison of non-labouring and labouring women was good. However, the authors need to present more measurements with statistical analyses for their sample groups to sufficiently establish the extent to which local (cervical) progesterone is more important than circulating (blood) progesterone in the manifestation of preterm labour. The relevance of their findings to the clinical efficacy of progesterone prophylaxis for preterm labour prevention (evaluated by several published clinical studies, refer to Romero et al., 2016 1 and EPPPIC Group, 2021 2 ) also requires more detailed consideration to improve the article.
Additionally, comments for each section of the article are as follows:

Introduction:
Localised regulation of progesterone metabolism is only one possibility for how functional progesterone withdrawal is mediated. The other (relatively more explored) possibility involves changes in tissue progesterone receptor expression/function irrespective of ligand concentration; this should be acknowledged with the inclusion of relevant references, though only needs to be mentioned briefly because it was not the focus of the authors' study.
○ It would be helpful to explicitly state the hypothesis of the study in the Introduction (rather than give it first mention in the Discussion).

Methods:
For the participant inclusion/exclusion criteria, please state whether tocolytics or steroids were administered prior to sample collection. Please also describe the method used to confirm absence of fetal membrane rupture. ○ Further detail should be added to Table 1 to make the data more valuable for potential future integration with similar datasets from other sources. For example, gravida, parity, ethnicity, body mass index, history of preterm birth, type of preterm labour (e.g. idiopathic, placental abruption, chorioamnionitis), Bishop score at time of sampling, gestation at fetal delivery, and mode of fetal delivery.

○
Please include a citation for the power calculation formula, to explain what each factor in the calculation represents (to help with interpretation by those who are unfamiliar with this specific formula).

○
The "300 rpm" centrifugation speed (in first paragraph of the 'Data collection' section) needs to be converted to standard relative centrifugal force (g unit).
○ Please state whether the blood serum or cervical mucus samples needed to be diluted prior to use at the immunoassay for quantifying their progesterone concentrations; also include mention of any control samples that were used to evaluate assay accuracy.

Results:
It is stated in the Methods section that "To anticipate differences in progesterone levels every week of gestational age, subgroup analysis was carried out based on gestational age" -this subgroup consideration is not shown in the data presented, despite it being a good approach. The authors need to either include their findings from this subgroup analysis or explain why they decided not to proceed with it.
○ Figures 1 and 2 show that progesterone concentrations, in both blood plasma and cervical mucus, for the two participant groups of interest were within very similar ranges to each other; this is in agreement with median values presented in Table 2. Can the authors comment on whether the wider min-max range for cervical mucus samples from the preterm labour group (shown in Table 2) was mostly attributed to the six samples that contained >6 ng/mL progesterone (shown in Figure 2)? Was there anything clinically distinct about these six women, when compared to the other 30 women in the preterm labour group?

Discussion & Conclusion:
Possible explanations for abnormal data distribution for both participant groups are good.
Can the authors add further comment with regards to possibility of infection for both groups of participants, especially whether it was a consideration for their inclusion/exclusion criteria for participant selection? ○ Progesterone concentration in blood plasma from preterm pregnant women has been quantified in previous studies (e.g. Smith et al., 2009 3 ). The authors should add depth to their discussion by commenting on how comparable their dataset is to these previous studies; with consideration of any differences in participant demographics, clinical observations and sampling methodology.

○
In the third paragraph, the authors should be more cautious about their interpretation of differences in cervical progesterone concentration between their preterm labour and normal pregnancy groups, especially because they did not present statistical analysis outcome for the corresponding data ( Table 2) that shows notable overlap of min-max ranges.

○
The authors should provide any technical information regarding cross-reactivity of the immunoassay (in the Methods section), which would help to judge to what extent we should be concerned about "Detected progesterone level in cervical mucus possibly includes progesterone and its inactive metabolites and therefore did not reflect the actual activity of cervical progesterone" (fourth paragraph). Additional assays (ideally using either a different immunoassay kit or methodology) with control samples (of metabolites) should have been undertaken to ascertain accuracy of progesterone quantification, where high crossreactivity with its metabolites was likely to occur.

○
The following sentence needs correcting: "Thus, the elevation of plasma progesterone levels was not followed by elevation in plasma progesterone level" (fifth paragraph).

○
The last sentence of the Conclusion is an overstatement until more observations and findings are added to the article. If samples are still available for use, the authors should consider acquiring data for quantification of specific progesterone metabolites (such as those identified in references 6 and 7) to assess whether preterm labour is associated with ○ changes in their concentrations more at the cervix than maternal blood.
The benefits of publishing with F1000Research: Your article is published within days, with no editorial bias • You can publish traditional articles, null/negative results, case reports, data notes and more • The peer review process is transparent and collaborative • Your article is indexed in PubMed after passing peer review • Dedicated customer support at every stage • For pre-submission enquiries, contact research@f1000.com