Prediction of the efficacy of dinoprostone slow release vaginal insert (Propess) for cervical ripening: A prospective cohort study

Predictive Factors for the Success of Vaginal Dinoprostone for the Induction of Labour

Objective

To evaluate factors predictive of the success of a slow-release dinoprostone vaginal insert for cervical ripening.

Methods

This retrospective study included 187 women who received dinoprostone vaginal inserts for cervical ripening. The participants were divided into two groups: the transvaginal delivery group (n = 87) and cesarean section termination group (n = 100). The correlation between the parameters present before cervical ripening with dinoprostone slow release and its success, as well as complications and adverse outcomes, was analyzed. Cesarean section predictors and area under the curve (AUC) were compared between the two Groups.

Results

There were statistical differences between the two groups in body mass index (BMI), height, cervical Bishop score, cephalic position, time of medication use, and fetal head position at the time of medication use (P<0.05). The optimal thresholds for identifying cesarean section in dinoprostone vaginal insert for cervical ripening were 162.5 for height (AUC = 0.61), 10.65 cm for amniotic fluid index (AUC = 0.6), S-2.5 for cephalic position (AUC = 0.61), 5.5 for bishop score of cervix (AUC = 0.65). The height, amniotic fluid index, cephalic position, and Bishop score of the cervix were included in the same model. The AUC value of the combined model was higher than the AUC value of the single factor.

Conclusion

The combined model was a better predictor of cesarean section in dinoprostone vaginal inserts for cervical ripening and labor induction. The success of cervical ripening with a dinoprostone slow-release vaginal insert can be predicted by the factors that can be recognized at admission.

Optimal cervical-ripening method for labor induction in Japan after the era of controlled-release dinoprostone vaginal insert

OBJECTIVE

To investigate the predictive value of obstetric findings when using dinoprostone (prostaglandin E2 [PGE2]) vaginal inserts for cervical ripening, and to assess the optimal cervical-ripening method between PGE2 vaginal insert and/or cervical dilators.

METHODS

This prospective observational study enrolled pregnant women who underwent cervical ripening for labor induction in 37-41 week' gestation in 2020. In evaluation 1, optimal obstetric findings predictive of rapid cervical ripening using PGE2 were assessed. In evaluation 2, the duration from PGE2 administration to labor onset and perinatal outcomes were compared between cases in which only PGE2 was used and cases that were treated with PGE2 after mechanical cervical dilators (Dilapan®) for extremely immature cervical ripening (uterine cervical os <2 cm).

RESULTS

In evaluation 1, uterine dilatation before the use of a PGE2 vaginal insert was mostly correlated with the time from PGE2 administration to labor onset (r = -0.428, p < 0.001). When the uterine cervical os dilatation was ≥2 cm, a shorter time-to-labor onset was found. In addition, os dilatation, effacement, and station at the time of PGE2 vaginal insert removal also significantly progressed. In evaluation 2, the median duration from PGE2 administration to labor onset was 1740 min in cases where only PGE2 was used, and 610 min in those where PGE2 was used after mechanical cervical dilators (p = 0.011).

CONCLUSION

PGE2 vaginal inserts are relatively effective when the uterine cervical os is ≥2 cm in diameter. However, in cases of extremely immature cervical-ripening, it was feasible to use PGE2 vaginal inserts before mechanical cervical dilatation.

Dinoprostone tablet versus continuous vaginal insert (Propess®) for elective induction in low-risk nulliparous women at term

OBJECTIVE

To evaluate the efficacy and safety of dinoprostone tablet and continuous vaginal insert (Propess®) in low-risk nulliparous women at term with insufficient cervical ripening receiving elective induction.

MATERIALS AND METHODS

A retrospective study was conducted between March 2020 and February 2022 and included 230 women who underwent elective induction with dinoprostone tablet or vaginal insert. The primary endpoint was failure of induction. Secondary endpoints included time to vaginal delivery, vaginal delivery rate, as well as maternal and neonatal complications and adverse outcomes.

RESULTS

No statistically significant differences were found between the two groups regarding the main outcome measures; however, the high responders had a significant higher proportion of hyperstimulation and non-reassuring fetal status. The high responder in the Propess group was statistically significant younger (31.68 ± 4.73 vs. 33.82 ± 4.39, p = 0.027), while they had a significantly lower BMI at delivery time of the tablet group (24.49 ± 2.24 vs. 27.42 ± 4.32, p = 0.024). Factors associated with success of vaginal delivery within 24 h (p = 0.015, OR = 0.9, 95%CI = 0.82-0.98) and the Cesarean section (p < 0.001, OR = 1.17, 95%CI = 1.08-1.27) was BMI at delivery time.

CONCLUSION

Slow-release vaginal insert and dinoprostone tablet had similar efficacy and safety for elective induction in low risk nulliparous women at term. Women with younger maternal age or lower BMI at delivery time may have a better response to dinoprostone and had a significantly higher proportion of hyperstimulation and non-reassuring fetal status.

Predictors of vaginal delivery following balloon catheter for labor induction in women with one previous cesarean

BACKGROUND

The aim of this study was to estimate predictors for vaginal birth following balloon catheter induction of labor (IOL) in women with one previous cesarean section (CS) and an unfavorable cervix.

METHODS

This 4-year retrospective cohort study was conducted in Longhua District Central Hospital in Shenzhen China, between January 2015 and December 2018. Patients with one previous CS and a current singleton-term pregnancy who underwent balloon catheter cervical ripening and IOL were enrolled. Univariate analysis was used to identify predictive factors associated with vaginal birth after cesarean section (VBAC). Binary logistic regression was further used to identify which factors were independently associated with the outcome measure. The primary outcome was VBAC, which was a successful trial of labor after cesarean delivery (TOLAC) following IOL.

RESULTS

A total of 69.57% (208/299) of the women who planned for IOL had VBAC. In the final binary logistic regression equation, lower fetal weight (< 4000 g) (odds ratio [OR]5.26; 95% confidence interval [CI] 2.09,13.27), lower body mass index (BMI,<30 kg/m²) (OR 2.27; CI 1.21, 4.26), Bishop score after cervical ripening > 6 (OR 1.94; CI 1.37, 2.76) remained independently associated with an increased chance of VBAC.

CONCLUSIONS

The influencing factors of VBAC following IOL were fetal weight, BMI, and Bishop score after cervical ripening. Adequate individualized management and assessment of the IOL may help improve the VBAC rate.

Using dinoprostone slow release vaginal insert for cervical ripening in term-pregnancy with oligohydramnios

AIMS

The study purposed to evaluate the success rate of cervical ripening using dinoprostone controlled-release vaginal insert and reveal some factors relating to successful cervical ripening.

METHODS

This cross-sectional study was conducted at Tu Du Hospital in Vietnam from December 2021 to August 2022. The study enrolled 200 pregnant women with gestational age ≥37 weeks diagnosed with oligohydramnios. These candidates underwent dinoprostone cervical ripening (DCR) according to the local protocol. The Bishop score ≥7 after 24 h was determined for the successful cervical ripening (SCR).

RESULTS

In total, the success rate of DCR achieved at 57.5% and the cesarean delivery rate was 46.5%. None of the severe side-effects and complications was present. Using multivariable logistic regression, the study found that the body mass index ≥25 kg/m² and oxytocin infusion drip related to SCR with adjusted odds ratio (aOR): 3.67 (95% confidence intervals [CI]: 1.78-7.57) and aOR: 4.68 (95% CI: 1.84-11.93), p < 0.001. Using the Kaplan-Meier curve, the present study revealed a significant difference between Bishop <3 and ≥3 following the duration time of cervical ripening, with hazard ratio: 1.38 (95% CI: 1.19-1.59), p < 0.001. The time duration of cervical ripening was not significantly different following amniotic fluid index from 3 to 5 cm.

CONCLUSIONS

Cervical ripening using a dinoprostone vaginal insert is a potentially acceptable method in term pregnancy accompanying with oligohydramnios. The probability of SCR can be predicted on a careful assessment of relative factors by obstetricians. Further studies are required to strengthen these findings.

Prediction of an effective cervical ripenning in the induction of labour using vaginal dinoprostone

To develop a predictive model for successful cervical ripening in women that undergo induction of labour by means of a vaginal prostaglandin slow-release delivery system (Propess®).  Prospective observational study on 204 women that required induction of labour between February 2019 and May 2020 at "La Mancha Centro" hospital in Alcázar de San Juan, Spain. The main variable studied was effective cervical ripening (Bishop score > 6). Using multivariate analysis and binary logistic regression, we created three initial predictive models (model A: Bishop Score + Ultrasound cervical length + clinical variables (estimated fetal weight, premature rupture of membranes and body mass index)); model B: Ultrasound cervical lenght + clinical variables; and model C: Bishop score + clinical variables) to predict effective cervical ripening. All three predictive models obtained (A, B and C) presented good predictive capabilities, with an area under the ROC curve ≥ 0.76. Predictive model C, composed of the variables: gestational age (OR 1.55, 95% CI 1.18-2.03, p = 0.002), premature rupture of membranes (OR 3.21 95% CI 1.34-7.70, p = 0.09) body mass index (OR 0.93, 95% CI 0.87-0.98, p = 0.012), estimated fetal weight (OR 0.99, 95% CI 0.99-1.00, p = 0.068) and Bishop score (OR 1.49 95% CI 1.18-1.81, p = 0.001), is presented as the model of choice with an area under the ROC curve of 0.76 (95% CI 0.70-0.83, p < 0.001). A predictive model composed of the variables: gestational age, premature rupture of membranes, body mass index, estimated fetal weight and Bishop score upon admission presents good capabilities in predicting successful cervical ripening following administration of prostaglandins. This tool could be useful in making clinical decisions with regard to induction of labour.

Clinical significances of Bishop score and vaginal bleeding to controlled-release dinoprostone delivery system (PROPESS) efficacy for cervical ripening: A retrospective single-center study in Japan

AIMS

To evaluate the effect of vaginal bleeding on the efficacy of controlled-release dinoprostone delivery system (PROPESS) for cervical ripening and the factors affecting the PROPESS efficacy in a Japanese clinical setting.

METHODS

A total of 100 term pregnant women in whom PROPESS was used due to an unfavorable cervix (Bishop score ≤ 6) were enrolled. We retrospectively investigated which factors, including vaginal bleeding, were associated with the success of cervical ripening using logistic regression analysis. Moreover, the effect of vaginal bleeding on vaginal acidity was examined in 24 selected cases (control, 11; rupture of membrane, 4; and vaginal bleeding, 8).

RESULTS

A 25 women successfully ripened the cervix (effective group), and 75 were unsuccessful (noneffective group). Bishop score at insertion (adjusted odds ratio: 1.87; 95% confidence interval: 1.23-2.86; p = 0.004), and vaginal bleeding at PROPESS insertion (adjusted odds ratio 6.63; 95% confidence interval 1.21-36.36; p = 0.029) affected cervical ripening success. The cases with vaginal bleeding showed a significantly higher vaginal pH than the control cases (median value: 6.75 and 5.0, respectively). We identified no obvious adverse outcomes, such as tachysystole, fetal heart rate abnormality, or low Apgar/pH, associated with vaginal bleeding at insertion.

CONCLUSIONS

Our findings suggest that the PROPESS efficacy depends on Bishop score at insertion and that vaginal bleeding at PROPESS insertion might have a significantly positive effect on cervical ripening in term pregnant women.

Maternal and neonatal outcomes following induction of labor for fetal growth restriction: Extra-amniotic balloon versus prostaglandins

OBJECTIVE

To compare perinatal outcomes in pregnancies with fetal growth restriction (FGR) undergoing induction of labor by extra-amniotic balloon (EAB) versus prostaglandin E₂ (PGE₂ ).

METHODS

A retrospective cohort study of women with singleton pregnancies and FGR, undergoing induction at term via EAB, PGE₂ , or both, at a single medical center (2014-2017). Primary outcome was rate of cesarean deliveries (CDs). Secondary outcomes included composite maternal and neonatal outcomes.

RESULTS

Overall, 266 women met the inclusion criteria. Among them, 131 (49.2%) underwent induction by PGE₂ , 116 (43.6%) by EAB, and 19 (7.14%) by both methods. No differences were noted in baseline characteristics. Rate of CD (17.24% vs. 6.11% vs. 10.53%, P = 0.022) and maternal composite outcome (18.97% vs. 6.11% vs. 10.53%, P < 0.01) were higher among women who underwent induction by EAB compared with PGE₂ or both. No difference was noted between groups in neonatal outcomes. In a multivariable logistic regression, rates of cesarean delivery and composite maternal outcome were no longer higher in the EAB group (adjusted odds ratio [aOR] 1.68, 95% confidence interval [CI] 0.68-4.16, P = 0.260; and aOR 1.94, 95% CI 0.84-4.45, P = 0.120, respectively).

CONCLUSION

EAB and PGE₂ have comparable maternal and neonatal outcomes when used for induction of labor due to FGR.

The Association between Cervical Length and Successful Labor Induction: A Retrospective Cohort Study

This study aimed to determine whether transvaginal sonographic measurement of cervical length before labor induction can predict successful induction. This retrospective study recruited 138 pregnant women who underwent labor induction at 37-41 weeks of gestation. Cervical length was measured using transvaginal ultrasonography before labor induction. Labor was induced according to the hospital protocol. Age, gestational age (GA), parity, body mass index (BMI), Bishop score, hemoglobin level, maternal disease, and epidural anesthesia were also recorded. Labor induction outcomes, including cesarean section for failed induction, time of induction, and the three labor stages, were assessed. From December 2018 to December 2021, 138 women were recruited for our study, including 120 and 18 women with successful and failed labor induction, respectively. Shorter cervical length (≤3.415 cm, OR = 6.22, 95% CI = 1.75-22.15) and multiparity (OR = 17.69, 95% CI = 2.94-106.51) were associated with successful induction. Higher BMI was associated with failed induction (OR = 0.87, 95% CI = 0.75-0.99). Age, GA, Bishop score, and fetal birth weight were not associated with successful labor induction. The ROC curve showed a cervical length cutoff value of 3.415 cm, revealing 76.8% of the area under the curve. In conclusion, a shorter cervical length (≤3.415 cm) was associated with a higher chance of successful labor induction (76.8%). This parameter might be used to predict the chance of successful labor induction. This information could help better inform clinician discussions with pregnant women concerning the chance of successful labor induction and consequent decision-making. Nevertheless, further large-scale clinical trials are warranted.

Comparison of the Dinoprostone Vaginal Insert and Dinoprostone Tablet for the Induction of Labor in Primipara: A Retrospective Cohort Study

This retrospective study aimed to compare the safety and efficacy of Prostin E2 and Propess for the induction of labor (IOL) in nulliparous women between January 2018 and October 2021. The inclusion criteria were nulliparous, singleton, >37 weeks’ gestation, cephalic presentation with an unfavorable cervix (Bishop score ≤ 6), no signs of labor, and use of one form of dinoprostone (Prostin E2 or Propess) for IOL. The cesarean section (C/S) rate and induction-to-birth interval were the main outcome measures. In total, 120 women were recruited. Sixty (50%) patients received Propess and 60 (50%) received repeated doses of Prostin E2. The Prostin E2 and Propess groups had similar patient characteristics, but the Bishop score was significantly higher in the Propess group than in the Prostin E2 group; therefore, multivariate analysis was conducted, and the Bishop score was not associated with the induction-to-birth interval. The C/S rate was not significantly different between the two groups, but the Propess group achieved a shorter induction-to-birth interval, a higher rate of vaginal delivery in 24 h, and a lower number of vaginal examinations than the Prostin E2 group. Propess was effective and safe in IOL and could be an option for cervical ripening in nulliparous pregnancy.

The efficacy and safety of second dinoprostone pessary or balloon catheter after unsuccessful primary ripening with dinoprostone pessary

We investigated which treatment should be applied if primary cervical ripening with a dinoprostone pessary is unsuccessful. We included 281 women who experienced unsuccessful cervical ripening with a dinoprostone pessary and continued on induction of labour (IOL). Of the 281 women recruited, 177 were given a second dose of dinoprostone; 104 women received a balloon catheter. The second dinoprostone pessary was successful in achieving vaginal delivery in 88 of the 177 (48.6%) women, while the balloon catheter was successful in 42 of the 104 women (40.4%); there was no significant difference between the two treatments with regards to successful vaginal delivery. However, of the women who experienced successful vaginal delivery, the delivery rate in the dinoprostone group was significantly higher than that in the balloon catheter group 12, 24, 36, or 48 h after insertion (p = .0094, .0005, .0258, .0483, respectively). The neonatal outcomes, the proportion of maternal infection and postpartum haemorrhage were similar between the two groups.IMPACT STATEMENTWhat is already known on this subject? Labour induction is a common procedure in obstetrics in a bid to achieve vaginal delivery in China, because vaginal delivery is more beneficial and associated with a better quality of life as compared to a Caesarean delivery. There is consensus relating to the preferred method of IOL after unsuccessful IOL with a dinoprostone pessary.What do the results of this study add? This is the first study in a Chinese population to compare the dinoprostone pessary and balloon catheter for women with no response to dinoprostone for cervical ripening with a sample size greater than 100. We found that a second dose of dinoprostone can reduce the time from the re-initiation of IOL to vaginal delivery compared with the balloon catheter. Our data also indicated that all other outcomes relating to the mother and infant were similar.What are the implications of these findings for clinical practice and/or future research? A second dose of dinoprostone is a superior choice for women who experience unsuccessful IOL with dinoprostone to further accelerate vaginal delivery.

Comparison of perinatal outcomes between controlled-release dinoprostone vaginal delivery system (PROPESS) and metreurynter for cervical ripening in labor induction: A retrospective single-center study in Japan

AIM

This study aimed to compare the efficacy and safety of a controlled-release dinoprostone vaginal delivery system (PROPESS) and a metreurynter for labor induction.

METHODS

This retrospective case-controlled study included 117 pregnant women (51 and 66 in the PROPESS and metreurynter groups, respectively) who required labor induction after >37 weeks' gestation at Mie University Hospital between January 2018 and September 2020. The primary outcome was the success rate of vaginal delivery. The secondary outcomes were changes in the Bishop score from the first insertion of PROPESS or the metreurynter to removal, uterine hyperstimulation and non-reassuring fetal status during the first insertion, proportion of pregnant women who needed pre-delivery oxytocin after removal, time to vaginal delivery after the first insertion, proportion of pregnant women who delivered vaginally within 12 or 24 h after the first insertion, and neonatal outcomes.

RESULTS

The proportion of pregnant women, especially primiparas, who delivered vaginally was significantly higher in the PROPESS group (26/34 [76.5%]) than in the metreurynter group (25/52 [48.1%]; p = 0.01). Moreover, among multiparas in the PROPESS group who delivered vaginally, nine (56.3%) out of 16 women delivered vaginally within 3 h of labor onset.

CONCLUSIONS

PROPESS for cervical ripening may reduce the risk of undergoing cesarean section in pregnant women requiring labor induction, especially primiparas. It is important to consider the possibility of precipitate labor when using the PROPESS in multiparas.

Using dinoprostone vaginal insert for induction of labor: A single institute experience

OBJECTIVE

To evaluate the factors associated with the successful induction of labor (IOL) in women treated with dinoprostone slow-released vaginal insert.

MATERIAL AND METHODS

A retrospective study was conducted between June 2017 and December 2017, enrolling 65 patients who underwent dinoprostone slow-released vaginal insert-induced labor. The correlation between the characteristics of the pregnant women and its success, as well as perinatal complications and adverse outcomes were analyzed.

RESULTS

Fifty-three of 65 (81.5%) achieved a successful vaginal delivery. Only multi-parous pregnant women were an independent predictor factor for successful induction after dinoprostone slow-released vaginal insert, since all of them succeeded after this strategy treatment (100%, n = 18), compared to 74.5% (35/47) of successful rate in the nulliparous pregnant women with a statistically significant difference (p = 0.018). There were no adverse events occurred in both mothers and their offspring.

CONCLUSION

Dinoprostone slow-released vaginal insert is a good choice for multiparous pregnant women who need IOL, regardless of which reasons are indicated. For nulliparous women, more studies might be needed to evaluate the effectiveness of PGE2 for IOL.

The Successful Use of Nitroglycerin for Uterine Hyperstimulation with Fetal Heart Rate Abnormality Caused by a Controlled-Release Dinoprostone Vaginal Delivery System (PROPESS): A Case Report

The PROPESS, a controlled-release dinoprostone vaginal delivery system, is a pharmacological cervical ripening intervention and promotes cervical change causing uterine contraction. During insertion of the PROPESS, uterine hyperstimulation could occur and result in fetal heart rate (FHR) abnormality. We report a case of uterine hyperstimulation accompanied with FHR abnormality caused by the PROPESS in a pregnant woman. Postural change, oxygenation, fluid infusion, and the immediate PROPESS removal were ineffective to address the adverse event, so we administered nitroglycerin for acute uterine relaxation. The nitroglycerin resulted in uterine relaxation, and the FHR abnormality was resolved immediately, thereby preventing an emergency cesarean section. Therefore, nitroglycerin could be considered an effective option for uterine hyperstimulation accompanied with FHR abnormality caused by the PROPESS.

Repeated dose of prostaglandin E2 vaginal insert when the first dose fails

Objective

To compare the obstetric and neonatal outcomes of patients treated with repeated-dose prostaglandin E2 (dinoprostone) vaginal insert when the first dose fails.

Materials and Methods

This retrospective study included 1.043 pregnant women who received dinoprostone for labor induction between November 2012 and August 2015. Pregnant women were divided into two groups according to the number of dinoprostone administrations: group 1, single-dose dinoprostone (n=1.000), and group 2, repeated-dose dinoprostone (n=43). Intrapartum, postpartum, and neonatal outcomes of the pregnant women were compared.

Results

Vaginal delivery rate was 65% in group 1 and 30.2% in group 2 (p=0.001). The need for the neonatal intensive care unit was found in 44 pregnant women (4.4%) in group 1 and 6 pregnant women (13.6%) in group 2 (p=0.006).

Conclusion

When obstetric and neonatal data were evaluated in our study, we observed that dinoprostone administration was associated with increased cesarean rates and adverse neonatal outcomes with repeated-dose dinoprostone when the first dose failed.

Predictor variables in the success of slow-release dinoprostone used for cervical ripening in intrauterine growth restriction pregnancies

OBJECTIVE

This study aims to evaluate the consequences of a trigger by vaginal Dinoprotone on outcome of pregnancies with Intrauterine growth restriction (IUGR).

MATERIALS AND METHODS

This retrospective study included 161 induced IUGR fetuses (35-39 weeks). Consecutive patients who were evaluated formed the basis of the clinical outcomes. The penalized maximum likelihood estimation (PMLE) method was used instead of traditional logistic regression in order to reduce the risk of overfitting.

RESULTS

Of the 25,678 deliveries that occurred during the study period, 161 (0.6%) women underwent IUGR delivery; of these, 117 (73%) succeeded and 44 (27%) failed to achieve cervical ripening using the dinoprostone slow-release vaginal insert. Two predictors were associated with dinoprostone vaginal delivery success: Parity (OR:1.4([0.89-2.3]), and Bishop score (OR:1.54[1.23-1.94]). The PMLE model correctly classified 78% participants (c-index: 0.78).

CONCLUSION

Basic parameters such as parity and Bishop score can be used to predict successful vaginal birth following dinoprostone slow-release vaginal insert administration.

Is there an interest in repeating the vaginal administration of dinoprostone (Propess®), to promote induction of labor of pregnant women at term? (RE-DINO): study protocol for a randomized controlled trial

BACKGROUND

Labor is induced in over 20% of women in France. Prostaglandins, especially intravaginal dinoprostone (Propess®), are widely used to initiate cervical ripening. If labor does not start within 24 h, there is uncertainty about whether to administer a second dinoprostone pessary or to use oxytocin to induce labor in order to achieve a vaginal delivery.

METHODS

RE-DINO is a prospective, open-label, multicenter, randomized superiority trial with two parallel arms running in six French hospitals. A total of 360 patients ≥ 18 years of age at > 37 weeks of gestation who exhibit unfavorable cervical conditions (Bishop score < 6) 24 h after placement of the first Propess®, with fetuses in cephalic presentation, will be included. Patients with premature membrane rupture, uterine scars, or multiple pregnancies will be excluded. Our principal objective is to determine whether placement of a second Propess® (followed by oxytocin [Syntocinon®], if necessary) in women for whom the first Propess® failed to induce cervical ripening increases the vaginal delivery rate compared to direct oxytocin injection. The vaginal delivery rate is therefore the primary outcome. The secondary outcomes are the induction failure rates and maternofetal morbidity and mortality.

DISCUSSION

This study may help in determining the optimal way to induce labor after failure of a first Propess®, an unresolved problem to date. This trial explores the effectiveness and safety of placing a second Propess® and may contribute to development of an obstetric consensus.

TRIAL REGISTRATION

Registered on 2 September 2016 at clinicaltrials.gov (identification number NCT02888041).