Vaginal alpha-lipoic acid shows an anti-inflammatory effect on the cervix, preventing its shortening after primary tocolysis. A pilot, randomized, placebo-controlled study.

INTRODUCTION
Inflammation might be an important underlying cause of preterm birth. Our aim is to explore whether vaginal administration α-lipoic acid reduces cervical inflammation and shortening after primary tocolysis.


MATERIALS AND METHODS
Singleton pregnancies between 24-30 weeks remaining undelivered after hospitalization for preterm labor were randomly allocated to placebo (20 women, 15 analyzed) or vaginal ALA 400 mg (active ingredient 10 mg) daily (20 women, 17 analyzed) for 30 days. A cervical swab to quantify pro-inflammatory (IL1, IL2, IL6, IL8, TNFα) and anti-inflammatory (IL4, IL10) cytokines as well as transvaginal ultrasound cervical length measurement (CL) were performed before and after treatment.


RESULTS
The % changes of pro-inflammatory cytokines do not differ between treatment groups, while IL4 significantly increases by vaginal ALA in comparison to placebo (118.0 ± 364.3% versus 29.9 ± 103.5%, p = 0.012). Combined anti-inflammatory cytokines show same trend (292.5 ± 208.5% versus 64.5 ± 107.4, p = 0.03). CL remains similar in vaginal ALA group (from 23.1 ± 6.6 to 20.80 ± 7.9 mm), while it significantly decreased in placebo group (from 20.4 ± 6.5 to 13.8 ± 7.5 mm, p < 0.001 versus Baseline; p = 0.003 versus vaginal ALA).


CONCLUSION
Vaginal ALA significantly stimulates anti-inflammatory ILs in the cervix of undelivered women after a preterm labor episode. This effect is associated with a stabilization of the CL.


Background
The overall rate of preterm birth (PTB) is steadily increasing in the developed world [1,2]. Recent clinical and experimental evidence supports the syndromic nature of PTB [3] which derives from four pathogenic processes, sharing a common final biological pathway leading to uterine contractions and cervical changes with or without premature rupture of membranes: (1) the decidual-chorioamniotic or systemic inflammation, (2) the decidual hemorrhage, (3) the activation of the maternal or fetal hypothalamic-pituitary-adrenal axis and (4) pathological distention of the uterus [4].
Strategies to prevent preterm labor/PTB can now be taken into account in view of the availability of drugs and devices. Among them, treatment with progestogens to prevent PTB has represented the major research focus in the last years.
Several studies have demonstrated the efficacy of either vaginal progesterone (P) or intramuscular 17a-hydroxyprogesterone caproate for the prevention of PTB in asymptomatic women with singleton pregnancy with a history of a previous preterm birth [5][6][7]. Moreover, evidence support the efficacy of vaginal P in asymptomatic women presenting with an objectively evaluated short cervix at mid pregnancy [8]. On the other hand, the efficacy of progestogens in women remaining undelivered after an episode of PTL remain to be demonstrated, as controversial data have been published [9][10][11][12]. For this reason, new approaches to PTB prevention are needed [13], namely in the field of the so called ''secondary or maintenance tocolysis''.
Alpha-lipoic acid (ALA), also known as thioctic acid, and its reduced form, dihydrolipoic acid (DHLA), are naturally occurring compounds [14] with a well-known anti-oxidant and anti-inflammatory profile [15], probably via the inhibition of nuclear factor k beta (NF-kB)-signaling pathway [16]. NF-kB plays a fundamental role in the expression of various genes involved in inflammatory response and in cell apoptosis processes and it has a central role in labor-associated activation pathways. Its expression is increased in the myometrium in the later stages of pregnancy and NF-kB activity mediates cyclo-oxygenase (COX)-2 expression in the human amnion during labor [17]. Its stimulation is connected to the cell's exposure to lipopolysaccharide, inflammatory cytokines (tumor necrosis factor (TNF)-a and interleukin (IL)-1), lack of progesterone, growth factors, free radicals and many other physiological and nonphysiological stimuli [18]. It has been observed that treatment with ALA was effective both in inhibiting the degradation of the IkB, a protein that inhibits NF-kB, and in directly reducing the expression of NF-kB and of matrix metalloproteinase-9 (MMP-9), an enzyme responsible for the degradation of the extracellular matrix [19]. This enzyme has been associated with PTB as its increased levels in amniotic fluid and its enhanced gene expression in fetal membranes have been associated with preterm rupture of membranes compared with other causes of PTB and TNFa-induced MMP-9 expression results in a significant decrease in the force needed to rupture human term fetal membranes [20]. LA in vitro treatment inhibited TNFa-induced weakening of fetal membranes and cytokineinduced MMP9 and prostaglandins production in both intact fetal membranes and amnion cells suggesting that its supplementation during pregnancy might prove clinically useful in the prevention of preterm premature rupture of fetal membranes [21]. Recently, it was also shown by preliminary trials that ALA, administered by oral or vaginal route, is able to potentiate the effect of progesterone in the healing process of threatened miscarriage and it may speed up the process of subchorionic hematoma resorption, by enhancing the levels of vascular endothelial growth factor (VEGF) and alphasmooth muscle actin (alpha-SMA) and decreasing the expression of NF-kB and MMP-9 [22,23].

Rationale
ALA, acting on different pathways involved in the pathogenesis of PTB [24], could be a molecule of interest. Thus, we have decided to investigate the vaginal administration of ALA looking at its effects on the cervix, in women remaining undelivered after an episode of preterm labor.

Materials and methods
This was a pilot, prospective, randomized, placebo-controlled, parallel group, monocenter trial. The study was conducted from January 2015 to February 2016 in the Azienda Ospedaliero-Universitaria Policlinico of Modena in the Unit of Obstetrics and Gynecology. The randomization allocation was 1:1 (vaginal ALA:placebo) and was accomplished using an electronic randomization list.

Subjects
A written informed consent was collected from all patients prior to inclusion in the study. The inclusion criteria were: women with a singleton pregnancy, at a gestational age ranging 24-30 weeks, hospitalized for a first preterm labor episode (regular uterine contractions with cervical change documented by transvaginal ultrasound. Cervical change was defined as a length 525 mm or a documented cervical length decrease of more than 10 mm within the previous four weeks of gestation. Exclusion criteria included the following: previous cervical surgery (conization, cervical cerclage), cervical length 55 or435 mm, cervical dilatation 42 cm, use of progestogens or other vaginal suppositories in the last 30 days.
At hospital discharge, women were randomized to receive either alpha-lipoic acid vaginal hard tablets 400 mg (active ingredient 10 mg), 1 tablet per day before sleep (Dav Õ , LO. LI. Pharma, Rome, Italy) per 30 days or apparently similar tablets composed by the very same excipients as the active drug product except the active ingredient (placebo).
During hospitalization, patients were treated according to the standard practice of our Institution, that is, intravenous fluids, tocolytic therapy using atosiban, steroid administration for respiratory distress syndrome prophylaxis, if clinically indicated.

Outcomes of the study
The primary outcome of the study was the change in the profile of pro-inflammatory and anti-inflammatory cytokines in cervical fluid. Monthly change of the cervical length (CL) was considered as a second outcome of the trial. Other outcome included the frequency of adverse events in patients allocated to placebo or ALA group. All outcomes were determined and the database was locked prior to the unsealing of the randomization code.

Determination of pro-and anti-inflammatory cytokines
All the samples for cytokine assays were obtained from the cervical fluid. Both sides of a preweighted dacron swab were applied for 10 s. Swab was then weighed again, stirred in 1.5 ml saline solution, centrifuged and stored at À20 C, until assay. Cytokine concentrations were measured following the manufacturer's instructions by the Ciraplex Õ immunoassay kit (Aushon BioSytem Inc, MA), a multiplex sandwich enzymelinked immunosorbent assay (ELISA). Each well of the 96well microplate is prespotted with protein-specific antibodies. After the removal of excess detection antibody, streptavidinhorseradish peroxidases (SA-HRP) is added and the luminescent signal detected by Cirascan Õ Imaging System (Aushon BioSytem Inc, Billerica, MA). Lower limit of quantification (LLOQ) and upper limit of quantification (ULOQ) (pg/ml) with back-calculated concentration 520% and relative error Changes in pro-and anti-inflammatory cytokines were considered as concentrations (pg/ml), mean of the percent changes (%) and the mean of the percent changes of the pro-(IL1, IL2, IL6, IL8, TNFa) and anti-inflammatory (IL4, IL10) cytokines.

Determination of cervical length
All the ultrasound evaluations of the CL were performed by the same operator in the Preterm Birth Clinic (LP) by using a 25 Gold Mylab Õ (Esaote SpA Õ , Genoa, Italy) with an endovaginal probe model equipped for color Doppler imaging and power Doppler angiography (Model EC 1123).
With the woman's bladder emptied, the vaginal transducer was inserted into the anterior fornix of the vagina and positioned so that the endocervical canal is visualized. The ultrasound probe should be gradually withdrawn until the image is just visible to ensure there is not excessive pressure on the probe. A minimum of three CL measurements should be obtained by placing calipers at the internal and external os. The shortest and best measurement was recorded [25].

Statistical analysis
Data were analyzed by one of the authors (GG) who was blinded to the specific treatment of each group. Statistical analysis was performed by the statistical package StatView (version 5.01.98, SAS Institute Inc., Cary, NC). Within-group comparison was performed with the t-test for paired data and by the Wilcoxon signed-rank test for normal and non-normal data distribution, respectively. For all analyses, the null hypothesis was rejected at two-tailed p values 50.05. Results are expressed as the mean ± standard deviation (SD).

Results
The flowchart of the study is reported in Figure 1. A total of 57 women were eligible according to the inclusion/exclusion criteria of the study and they were asked to participate. Of these, 40/57 (70.2%) subjects accepted, signed the informed consent and were then randomized to the treatment groups (vaginal ALA versus placebo) in a a ratio of 1:1 (20 subjects per group). No significant differences were found in subjects that reach final follow-up visit (17/20 (85%) in vaginal ALA group, 15/20 (75%) in placebo group, p ¼ 0.72).
The baseline features of the women included in the study are reported in Table 1 and comparable between groups, except that subjects in the vaginal ALA were more frequently included presenting both contractions and cervical changes at admission (p ¼ 0.03).

Changes in the concentrations of cytokines in cervical fluid
Baseline and after treatment values of cytokines are reported in Table 2. Baseline levels were not different between groups.
The percent changes of pro-and anti-inflammatory cytokines were then considered. No significant different changes were observed between groups as far as proinflammatory cytokines are concerned (vaginal ALA versus placebo) (Figure 2a and c). On the contrary, vaginal ALA induced a significant increase of IL4 in comparison to the placebo group (118.0 ± 364.34% versus 29.9 ± 103.5%, p ¼ 0.012) (Figure 2b). Once combined with IL10 changes, the increase of both anti-inflammatory cytokines was significantly more pronunced in vaginal ALA group in comparison to placebo group (292.5 ± 208.5% versus 64.5 ± 107.4%, p ¼ 0.03) (Figure 2c).

Changes of cervical length
Overall, cervical length remains similar in vaginal ALA group (from 23.1 ± 6.6 to 20.8 ± 7.9 mm), while it significantly decreased in the placebo group (from 20.4 ± 6.5 to 13.8 ± 7.5 mm, p 0.001 versus baseline; p ¼ 0.003 versus vaginal ALA).

Tolerability
The only adverse effect reported in the vaginal ALA group was transient vaginal discomfort in 2/17 (11.8%) of the  subjects while none in the placebo group. The incidence of adverse effects was not different between groups (p ¼ 0.170).

Principal findings of the study
In this pilot study, we have shown that vaginal ALA administration is associated with a specific anti-inflammatory effect at the cervical level in comparison to placebo, namely stimulating anti-inflammatory cytokines. Vaginal ALA is more effective than placebo in limiting the monthly cervical shortening observed in women remaining undelivered after a preterm labor episode. Such effect is present both in women with a short cervix (520 mm) and in those with a cervical length above this cutoff. However, whether this biological property of vaginal ALA could lower the risk of PTB (thus affecting neonatal morbidity/mortality) or not has to be evaluated in large randomized clinical trial.

Interpretation
The etiology of PTB is complex and unfortunately still not understood. After an episode of threatened preterm labor with intact membranes, which is initially treated with tocolysis and corticosteroids for 48 h, women remain at risk for preterm delivery, as they deliver before 37 weeks at a rate ranging 38-75% [26][27][28].
Accumulated research indicates that infection and inflammation (most often subclinical) is an important underlying cause of PTB. In particular, a recent systematic review has shown that IL6 and C-reactive protein (CRP) in amniotic fluid, but not in plasma, are strongly associated with increased risk of spontaneous PTB among asymptomatic women, indicating that inflammation at the maternal-fetal interface rather than a systemic one, may play a major role in the etiology of such spontaneous PTB. These data are confirmed by the elevation of IL6 levels in the cervicovaginal fluid, which are also strongly associated with spontaneous PTB, suggesting that the local level presents the most important and subtle changes in women with PTB. However, there was insufficient evidence for the association between other inflammatory cytokines, like IL1, IL2, IL8 and TNF-a and spontaneous PTB among asymptomatic women [29]. The role of these mediators has not yet been entirely elucidated, but it is likely that inflammatory cytokines bring about parturition through increasing prostaglandin levels by affecting prostaglandins production, metabolism or both.
In this study, we focused on the activity of vaginal ALA at the cervical level. In ongoing experiments in female rats, addressed to verify ALA vaginal absorption, it was demonstrated by immunohistochemistry that the endometrial epithelial and stromal cells and the cervical squamous epithelium and muscular tissue were stained with intense amount of Table 2. Baseline and after treatment concentration (pg/ml) in the cervico-vaginal fluid of pro-inflammatory (IL1, IL2, IL6, IL8, TNFa) and antiinflammatory (IL4, IL10) cytokines of the subjects randomized for the two different treatment groups (placebo (n ¼ 15) and vaginal ALA (n ¼ 17).

Cytokines
Baseline After treatment anti-ALA antibody, proving this route of administration very effective for ALA [30]. In our study, the active treatment did not affect proinflammatory cytokines, including IL6. On the other hand, vaginal ALA stimulated the cervical secretion of IL4, an antiinflammatory cytokine that induces differentiation of naive helper T cells to Th2 cells. It is well known that a shift in the bias of overall maternal reactivity to the more conducive Th2 type, or at least a reduction in strong Th1 bias, may provide the best milieu to favor pregnancy [31].
Moreover, polymorphisms of IL4 have been associated with spontaneous PTB [32], while its dysregulation has found to be correlated with Toll-Like receptor activity in cases of preterm neonates [33].
Our findings are in agreement with the growing body of evidences suggesting the clinical use of LA when the abnormal regulation of the inflammatory response is recognized as an important component of the disease as in the case of diabetes, Alzheimer's disease and multiple sclerosis [15]. In such conditions, LA exhibits anti-inflammatory properties by inhibiting IL6 and IL17 production and T-cell proliferation and activation while possibly maintaining IL10 synthesis [15,16,30].
On the other hand, CL is a marker of clinical value in the prediction of PTB, the shorter is the cervix, the higher is the risk of PTB [34,35]. However, a progressive shortening of the cervix has been reported throughout physiological pregnancy. Namely between 30 and 34 weeks, the cervix shortens by 3 mm [36], the same change we observed in women receiving vaginal ALA. Considering that the cervix shorten the double in our women receiving placebo, the active treatment, slowing down cervical remodeling, reduces the theoretical lower risk of PTB.

Implications for further research
The use of ALA represents a safe and well-tolerate treatment in many circumstances up to doses of 2400 mg/day [37] also in pregnant women. We confirm tolerability in our small trial and demonstrated that vaginal ALA exerts a direct antiinflammatory effect on the cervix, preventing its shortening after primary tocolysis. Such premise allows the planning of a clinical trial to explore the efficacy of this treatment in the prevention of PTB.

Strengths and limitations
The major limitation of this study is the small sample size, and thus, it should be considered as a pilot study. Furthermore, the study was conducted in a single center and the results cannot be generalized. For all these reasons, results should be considered preliminary, although worthy to be explored in larger comparative investigations.

Conclusion
To the best of our knowledge, this is the first study investigating the effects of ALA administration through vaginal route after primary tocolysis. Such treatment induces significant changes in the cervical ILs pattern in undelivered women after a preterm labor episode, namely stimulating antiinflammatory cytokines. Moreover, the treatment is associated with a stabilization of the cervix, which otherwise undergoes to a shortening.