Study selection
A total of 229 articles were identified through an initial search in the databases including PubMed (n = 34), Embase (n = 53), Medline (n = 17), Web of Science (n = 40), Cochrane Library (n = 51), CNKI (n = 25), WanFang Data (n = 4), CBM (n = 5). After the removal of duplicates, 98 articles remained. Subsequently, 72 records were excluded by reviewing the title and abstract while 26 full-text articles were retrieved to assess their eligibility in terms of the inclusion criteria. Based on this assessment, 10 studies were excluded due to the following reasons: 4 were abstract only; 4 did not report relevant outcomes; 1 was study protocol, and 1 did not met the inclusive criteria. Ultimately, 16 relevant trials were included for qualitative analysis. The intervention of 15 trials was sildenafil and 1 was tadalafil. Therefore, only 15 relevant trials investigating the effect of sildenafil were included for meta-analysis, as shown in Fig. 1.
Studies Characteristics
The demographic characteristics of included studies are reported in Table 1. Studies were published between 2012 and 2022. Five studies were conducted in Egypt [29–33], 2 in India[34, 35], 2 in Iran[16, 17], 1 in Brazil [36], 1 in China[37], 1 in Japan[18], 1 in the Netherlands[23], 1 in New Zealand and Australia[21], 1 in the UK[22] and 1 in Canada[38].
The population of the selected trials consisted of 1492 pregnant women, with 747 and 745 pregnancies were in the sildenafil group and control/ comparators group respectively. The gestational age of the initial treatment ranged from 24 to 35 weeks. The minimum and maximum dosage of sildenafil used across studies ranged from 25 to 75 mg daily and the dosage of tadalafil was 20 mg once daily. Treatment duration ranged from single-dose before Doppler assessment in 3 studies[17, 30, 33], up to delivery or fetal death in 10 studies[18, 21–23, 29, 31, 32, 34, 35, 38], up to 28 days in 1 study[37]or was not mentioned in 2 studies[16, 36].
Sildenafil citrate was used as the treatment group, while placebo was used as the control/ comparators group in 10 trials[16, 17, 21, 23, 30–33, 36, 38] and L-arginine as the control/ comparators group in 1 trial[35]. One study[29]used sildenafil citrate in addition to fish oil and zinc supplementation as the treatment group to compared with placebo in addition to the same supplementation as the control/ comparators group. One study[22]used sildenafil citrate in addition to injection of corticosteroids and magnesium sulphate as the treatment group, and placebo in addition to the same injection as the control/ comparators group. One study[34] used sildenafil citrate in addition to injection of betamethasone as treatment group, and no intervention as control/ comparators group. One trial[37] used sildenafil citrate in addition to LMWH as treatment group, and LMWH as control/ comparators group. Tadalafil in addition to conventional treatment for FGR according to Japanese guidelines was used as treatment group, conventional treatment as control/ comparators group in 1 trial[18].
Twelve studies reported data on MCA-PI[16–18, 21–23, 29–33, 36], 13 studies on UA-PI[16–18, 21–23, 29–33, 35, 36] and only 2 studies on MAP-PI[30, 36]. The most commonly reported indicator for efficacy: (1) birth weight (grams); (2) pregnancy prolongation (in days); (3) gestation age at birth (weeks); (4) blood flow (UA-PI and MCA-PI). The most commonly reported indicator for safety: (1) infants admitted to neonatal intensive care unit (NICU); (2) headache, flushing/rush, gastrointestinal side effects in mothers; (3) perinatal mortality or major neonate morbidity; (4) intraventricular hemorrhage (IVH), necrotizing enterocolitis in infants.
Efficacy
Birth weight - overall and subgroup analysis (maternal age in experiment group)
Fetal birth weight was analyzed in 7 trials[16, 21, 22, 31, 32, 34, 35] and included 829 patients, Of them, 420 patients are in sildenafil group and 409 patients are in control/ comparators group. Sildenafil was associated with a statistically significant increase of 164.07 grams (MD:164.07, 95%CI:61.55-266.59, P = 0.002, I2 = 90%; Fig. 2) in birth weight compared with no sildenafil.
Four trials[22, 31, 32, 34], comprising a total of 409 patients, reported on birth weight in patients under 30, and 2 trials [16, 21]were in patients above 30 (202 patients). In patients under and above 30, sildenafil significantly increases birth weight (MD:198.6, 95%CI:19.95-377.25, P = 0.03, I2 = 92%; MD:82.73, 95%CI:7.14-158.32, P = 0.03, I2 = 0%; respectively; Fig. 3)
Pregnancy prolongation - overall and subgroup analysis (maternal age in experiment group)
Four trials[16, 21, 32, 34] reported on pregnancy prolongation and included 386 pregnant women, 195 in sildenafil group and 191 in control/ comparators group. Sildenafil was associated with a significant increase in pregnancy prolongation for 6.09 days (MD:6.09, 95%CI:2.15–10.03, P = 0.002, I2 = 75%; Fig. 4).
Two trials[32, 34], comprising 184 patients, reported pregnancy prolongation in patients under 30 in the experiment group, 2 trials [16, 21]in patients above 30 (202 patients). Sildenafil-treated pregnant women under 30 showed a significant pregnancy prolongation for 8.04 days (MD:8.04, 95%CI:6.16–9.92, P < 0.00001, I2 = 0%; Fig. 5) compared with control/ comparators. In contrast, pregnant women above 30 did not prolong pregnancy in women compared with control/ comparators group (MD: 2.22, 95% CI: -0.62 -5.06, P = 0.13, I2 = 0%; Fig. 5)
Blood flow—UA-PI
Five studies [17, 29, 30, 33, 36]reported on UA-PI and included 225 pregnant women, 106 in sildenafil group and 119 in control/ comparators group. Sildenafil was associated with a significant decrease of UA-PI (MD: -0.24, 95%CI: -0.32 - -0.15, P < 0.00001, I2 = 55%; Fig. 6).
Blood flow—MCA-PI
Five studies [17, 29, 30, 33, 36]reported on MCA-PI and included 225 pregnant women, 106 in sildenafil group and 119 in control/ comparators group. Sildenafil was not associated with an increase of MCA-PI (MD:0.23, 95%CI: -0.24 -0.70, P = 0.35, I2 = 98%; Fig. 7). No subgroup analysis would be analyzed in the overall forest plots with no statistical significance.
Gestation age at birth
Eight studies[22, 29, 31, 32, 34, 35, 37, 38] reported on gestation age at birth and included 809 pregnant women, 409 in sildenafil group and 400 in control/ comparators group. Sildenafil was not associated with an increase of gestation age at birth (MD: 0.44, 95%CI: -0.29 -1.17, P = 0.24, I2 = 89%; Fig. 8). No subgroup analysis would be analyzed in the overall forest plots with no statistical significance.
Safety
Infants admitted to NICU
Infants admitted to NICU were reported in 6 studies[16, 21, 22, 29, 32, 35], totaling 593 pregnant women, 305 in sildenafil group and 288 in control/ comparators group. There was no difference between the groups in the occurrence of infants admitted to NICU (OR:0.63, 95%CI:0.34–1.19, P = 0.16, I2 = 56%; Fig. 9).
Headache in mothers
Nine trials[17, 23, 29–34, 36]reported on events of headaches in mothers and included 714 patients, 351 in sildenafil group and 363 in control/ comparators group. Sildenafil was associated with a statistically significant increase in events of headache in mothers (OR:5.57, 95%CI:2.89–10.72, P < 0.00001, I2 = 0%; Fig. 10) compared with no sildenafil.
Flushing/rash in mothers
Nine trials[17, 23, 30–34, 36, 37] reported on events of flushing /rash in mothers and included 784 patients, 386 in sildenafil group and 398 in control/ comparators group. Sildenafil was associated with a statistically significant increase of flushing /rash in mothers (OR:5.11, 95%CI:2.08–12.53, P = 0.0004, I2 = 0%; Fig. 11) compared with no sildenafil.
Perinatal mortality or major neonatal morbidity
The number of perinatal mortality or major neonatal morbidity was evaluated in 3 trials[21, 22, 29], with a total of 284 pregnant women, 149 in sildenafil group and 135 in control/ comparators group. Overall, there was no clear difference identified in perinatal mortality or major neonatal morbidity between control/ comparators group or sildenafil group (OR: 1.02, 95%CI:0.54–1.90, P = 0.96, I2 = 0%; Fig. 12).
IVH in infants
Three trials[22, 23, 36]reported on IVH in infants and included 259 patients, 136 in sildenafil group and 123 in control/ comparators group. Overall, there was no clear difference identified in IVH between control/ comparators and sildenafil group (OR:1.46, 95%CI:0.62–3.46, P = 0.39, I2 = 0%; Fig. 13).
Necrotizing enterocolitis in infants
Three trials[22, 23, 36]reported on necrotizing enterocolitis in infants and included 279 pregnant women, 146 in sildenafil group and 133 in control/ comparators group. There was no significant difference in events of necrotizing enterocolitis in infants between pregnancy with sildenafil or no sildenafil (OR:0.60, 95% CI: 0.29–1.23, P = 0.16, I2 = 0%; Fig. 14).
Gastrointestinal side effects in mothers
Eight trials[23, 29–34, 37] reported gastrointestinal side effects in mothers and included 765 patients, 383 in sildenafil group and 382 in control/ comparators group. Sildenafil was not associated with a statistically significant increase of gastrointestinal side effects in mothers (OR:1.68; 95%CI: 0.89–3.16; P = 0.11, I2 = 0%; Fig. 15) compared with no sildenafil.
Pregnancy hypertension
Five trials [21–23, 34, 38]reported on events of pregnancy hypertension and included 622 patients, 317 in sildenafil group and 305 in control/ comparators group. Overall, there was no clear difference identified in pregnancy hypertension between control/ comparators and sildenafil group (OR:1.11, 95%CI:0.78–1.58, P = 0.57, I2 = 0%; Fig. 16).
Stillbirth
The number of stillbirths was evaluated in 3 trials[23, 29, 38], with a total of 281 pregnant women, 142 in sildenafil group and 139 in control/ comparators group. Overall, there was no clear difference identified in perinatal mortality or major neonatal morbidity between control/ comparators group and sildenafil group (OR:1.24, 95%CI: 0.24–6.41, P = 0.79, I2 = 52%; Fig. 17).
Neonate death
Six studies [21–23, 29, 32, 38] reported on neonate death and included 521 pregnant women, 272 in sildenafil group and 249 in control/ comparators group. There was no significant difference in neonate death between pregnancy with sildenafil or no sildenafil (OR:1.58, 95%CI:0.91–2.76, P = 0.11, I2 = 0%; Fig. 18).
Pulmonary hypertension in infants
Two trials [23, 38] reported on events of pulmonary hypertension in infants and included 183 patients, 96 in sildenafil group and 87 in control/ comparators group. Sildenafil was associated with a statistically significant increase in events of pulmonary hypertension in infants (OR:4.37, 95%CI: 1.49–12.80, P = 0.007, I2 = 0%; Fig. 19) compared with no sildenafil.
Study Appraisal
Among RCTs, overall ROB was recorded as high for 5 trials[17, 18, 35, 37, 38], some concerns for 7 trials[16, 23, 30, 31, 33, 34, 36], and low for 4 studies[21, 22, 29, 32]. There was no appropriate randomization process in approximately quarter of the studies[17, 35, 37, 38]. Bias due to deviations from intended interventions was found in 4 trials[18, 34, 35, 38]. Losses more than 5% of the sample data occurred in 4 studies[17, 35, 37, 38]. The method of measuring the outcome was inadequately explained in 1 study[18]. The reported results were selected in around 56% of studies[16–18, 23, 30, 31, 33–38]. The risk analysis of bias is outlined in Fig. 20 which shows the risks of each bias between each trial separately. Figure 21 presents the percentage of each bias among all trials.