Study selection and data extraction for each included penicillin were performed. With the search strategy and based on title and abstract, 40 studies for amoxicillin (with and without clavulanic acid), 113 for ampicillin, 120 for benzylpenicillin, 0 for pheneticillin, 4 for phenoxymethylpenicillin, 3 for flucloxacillin and 12 for piperacillin (with and without tazobactam) were identified. After reading the full paper, 8 studies for amoxicillin (with and without clavulanic acid), 15 for ampicillin, 4 for benzylpenicillin, 0 for pheneticillin, 1 for phenoxymethylpenicillin, 0 for flucloxacillin and 4 studies for piperacillin (with and without tazobactam) were included. A detailed overview of the study selection is presented in the PRISMA flow diagram in Fig. 2.

An overview of the PK studies of penicillins during pregnancy is presented in the paragraphs below. Drugs are presented in alphabetical order. First, study characteristics of the individual penicillins are described. Second, PK and exposure are discussed according to the absorption, distribution, metabolism, elimination (ADME) sequence. Last, the probability of target attainment is described and, if available, evidence-based dosing regimens of the individual penicillins are discussed.

A total of 8 studies on amoxicillin (with or without clavulanic acid) were included [9,10,11,12,13,14,15,16]. No PK studies reported the PK and exposure parameters of clavulanic acid. The patient characteristics of the included studies are listed in Table 2. The number of participants in the different studies ranged from 17 to 50. The mean age ranged from 29.0 to 31.8 years. The mean body weight throughout pregnancy ranged from 79.0 to 80.9 kg. The mean pre-pregnancy body weight and pre-pregnancy body mass index (BMI) ranged from 63.3 kg to 65.7 kg and from 23.1 to 24.0 kg/m2, respectively. Only one study reported all the different trimesters of pregnancy [9]. For the other 7 studies, gestational age ranged from 29.4 to 42.4 weeks. In four out of 8 studies, amoxicillin was administered intravenously, 3 orally and 1 vaginally.

The PK and exposure parameters of pregnant women (if possible, in comparison with non-pregnant women) of the included studies are summarised in Table 2. When focusing on the absorption phase, none of the 4 oral or vaginal studies reported absorption-related parameters, such as bioavailability (F) [9,10,11,12,13,14,15,16]. Only one study showed that Cmax decreased significantly in the second (21.8% points) and third trimester (33.3% points) of pregnancy compared to postpartum. The same study described the effects of pregnancy on the AUC [9]. At an oral dose of 500 mg, AUC decreased significantly, by 25.5% points and 27.0% points, respectively, in the second and third trimester of pregnancy compared to postpartum. Studies of Buckingham et al. and Zareba-Szczudlik et al. reported concentrations in maternal blood only [10, 14,15,16].

None of the 8 studies described the differences in Vd/F between pregnant and non-pregnant women. Only the intravenous (IV) studies performed by Muller et al. described the effects on Vd during the third trimester of pregnancy [11,12,13]. Results were obtained with the use of a developed population PK model. Two of the studies used a 3-compartment model and the other study described a 5-compartment model. The Vd in the central compartment ranged from 5.59 to 8.7 L. The first peripheral compartment ranged from 5.88 to 11.8 L. The second peripheral compartment ranged from 5.88 to 40.4 L [11,12,13].

Of all 8 studies, 4 reported elimination-related parameters (t1/2 and CL/F) [9, 11,12,13]. The reported mean CL/F in the different studies in pregnant women ranged from 19.7 to 35.5 L/h, corresponding with an increased CL/F with 86.4–123.3% compared to non-pregnant women. In comparison to the t1/2 3 months postpartum, the t1/2 was 18.8–31.3% shorter in pregnant than in non-pregnant women [9, 11, 12].

Of the 8 studies, 4 [9, 11,12,13] reported a desired target concentration and the probability of target attainment with the currently used dosing regimens of amoxicillin in pregnant women. The target varied per study and micro-organism (e.g., Group B Streptococcus, S. agalactiae, Table 2). Andrew et al. reported a probability of target attainment with different dosing simulations [9]. According to the studies of Muller et al. [11,12,13] no dosing adjustments for amoxicillin were necessary for pregnant women in the third trimester. These findings are partly supported by the study of Andrew et al. [9] that stated that dosing adjustments are not necessary for sensitive isolates. However, for less sensitive isolates, shortening the dosage interval is advised to reach the desired target concentration.

In summary, no direct comparison was made between pregnant and non-pregnant women. Little information is available on the effect of pregnancy on amoxicillin absorption. One study showed that AUC and Cmax during pregnancy were significantly lower than postpartum. No studies were performed to investigate the differences in distribution-related parameters between pregnant and non-pregnant women for amoxicillin, but CL/F is higher and the t1/2 is shorter in pregnant compared to non-pregnant women. The 4 studies that investigated target attainment reported that overall, no dosing adjustments for amoxicillin are necessary. However, for less sensitive isolates, shortening the dosage interval is advised to reach the desired target concentration.

In total, 12 studies on ampicillin were included [10,11,12,13,14,15,16,17,18,19,20,21]. Table 3 provides the characteristics of the included PK studies. Among these studies, only one included pregnant women during their first trimester of pregnancy, 3 studies during the second trimester, 10 during their third trimester; from 2 studies the trimester of the included women was not reported. The number of participants in the different studies ranged from 3 to 40 women. The mean age of the patients included in the studies ranged from 20 to 40 years. Only one study [21] reported the BMI (mean 33.0 kg/m2). The other studies [10,11,12,13,14,15,16,17,18,19,20], reported weight which varied from 46.8 to 88.4 kg. Of the 12 studies, 2 investigated the PK of ampicillin after oral administration, 2 after intramuscular administration, 6 after intravenous administration and for 2 studies the route of administration was unknown. The PK and exposure parameters of ampicillin in pregnant women (where possible compared with non-pregnant women) of the included studies are summarised in Table 3.

Of all 12 studies, one reported the F [22]. The mean F in non-pregnant women was 2.5% points higher compared with pregnant women (45.6 vs 48.1%). Six studies reported the Cmax and Cmean. After oral administration, the Cmax values reached in non-pregnant women were significant higher than in pregnant women (difference 40.5% points) [22]. Creatsas et al. [23] also investigated the effects on Cmean, Cmax and tmax of intramuscular administration in pregnant women in their third trimester. The authors found a Cmax of 7.88 μg/mL in mother’s blood 1 h after administration. When focussing on the distribution, Chamberlain et al. [20] stated that the Cmax reached in pregnant women (trimester not specified) was 8.1% points higher than Cmax reached in non-pregnant women after IV administration [20]. Philipson et al. [22] concluded that the Cmax after IV administration is comparable in pregnant (all trimesters) to non-pregnant women (difference of 1.6% points). Bloom et al. [18] and Bray et al. [19] investigated the effects on Cmax of IV administration in pregnant women in their third trimester. Bloom et al. [18] found a Cmax in mother’s blood of 142 μg/mL, 3 min after infusion, where Bray et al. [19] found a mean Cmax of 26.5 μg/mL, directly after infusion.

The Vd/F of ampicillin was investigated in 4 studies [17, 20, 22, 24]. Of these 4 studies, 3 investigated the differences between pregnant and non-pregnant women. One study stated that the Vd in pregnant women is slightly lower (5.0% points) than in non-pregnant women [20]. Contradicting results were mentioned in another study; namely a 38.4%-point higher Vd in pregnant women than in non-pregnant women was reported [22]. Assael et al. [17] supported this increase (45.6%) in Vd/F in pregnant compared to non-pregnant women.

The studies of Assael et al. [17], Chamberlain et al. [20] and Philipson et al. [22] investigated the effects of pregnancy on the AUC, in comparison to non-pregnant women. Assael et al. [17] concluded that the AUC in pregnant women in their second and third trimester was 37.3% points lower compared with non-pregnant women [17]. The studies of Chamberlain et al. [20] and Philipson et al. [22] supported the decrease of AUC in pregnant women, with 20.7% points and 34.9% points, respectively. Explaining the difference in AUC, CL/F in pregnant women was significantly higher than in non-pregnant namely, 179.1% [17] 122.4% [20] and 155.6% [22].

The majority of the studies investigated ampicillin PK for Group B Streptococcus (GBS) prophylaxis. Overall, those studies reported that in 100% of the cases the desired target concentration of 0.25–2.0 ug/mL was achieved and that there was no need to adapt the dosage.

In summary, F is almost unchanged in pregnant women compared with non-pregnant women. After oral administration, one study showed a lower Cmax in pregnant women. In general, Vd/F of ampicillin is increased in pregnant women compared to non-pregnant women. Furthermore, the reached AUC and CL/F values in pregnant women are lower and higher, respectively, than in non-pregnant women. Moreover, desired target concentrations for ampicillin in pregnant women are achieved with the currently used dosage.

In total, 4 studies on benzylpenicillin were included in this study (Table 4) [25,26,27,28]. One study investigated second trimester pregnant women, while the other studies included only women during their third trimester of pregnancy. The number of included women in each study varied from 15 to 60. The mean age of the woman ranged from 23.5 to 29.5 years. Mean BMI of the women included in the studies varied from 29.2 to 31.6 kg/m2 and weight varied from 50.9 to 151.7 kg. Mean gestational age ranged from 28.6 to 39.4 weeks. The PK and exposure parameters of benzylpenicillin in pregnant women (where possible, in comparison with non-pregnant women) of the included studies are summarised in Table 4.

No absorption-related PK parameters were reported. Only Johnson et al. reported a Vd in pregnant women during their third trimester, namely 0.27 L/kg [25]. No comparison was made between pregnant and non-pregnant women. The same study also reported CL. The mean total CL in pregnant women during their third trimester was 0.25 L/h/kg [25].

The study of Weeks et al. [28] mentioned a desired target concentration (MIC 0.06 μg/mL) and reported that with their currently used dosing scheme the target concentration will be reached in pregnant women. While Nathan et al. [23] mentioned that in only 40% of the pregnant women, the 0.018 μg/mL target concentration was achieved on Day 7 for congenital syphilis. No adapted evidence-based dosing regimen was provided.

In summary, as limited information is available on the PK changes of benzylpenicillin during pregnancy, no statement can be made on changes in PK of benzylpenicillin in pregnant compared with non-pregnant women. The same holds for the attainment of the target concentration.

One study on phenoxymethylpenicillin was included in this study (Table 5) [29], which included second and third trimester pregnant women (n = 12), which were compared to non-pregnant women (n = 6). Table 2 provides the characteristics of the included PK study and the PK-related changes of oral phenoxymethylpenicillin used throughout pregnancy.

When focusing on absorption related parameters (Table 5); the Cmax was 33.9% and 3.5% lower in the second and third trimester, respectively, in pregnant women compared to non-pregnant women. When focusing on Vd; this parameter has not been reported. However, when calculating Vd based on the reported CL and t1/2, the Vd was 55.2% and 14.4% lower in the second and third trimester of pregnancy compared to non-pregnant women. As for the elimination-related parameters, AUC was 52.1% and 39.2% lower and consequently CL was 68% and 47.4% higher in the second and third trimester compared to non-pregnant women.

The authors stated that to reach an adequate target concentration in the second and third trimester of the pregnancy, either the dosing interval must be shorter, namely from 8 to 6 h with a dose of 1 million IU, or dosages should be increased. The authors did not provide an increased evidence-based dosing regimen administered every 8 h.

In summary, Cmax is lower in pregnant and non-pregnant women. In general, Vd of phenoxymethylpenicillin is lower, whereas AUC was lower and consequently CL is higher during pregnancy compared to non-pregnant women.

In total, 4 studies were included in this study (Table 6) 1 of which investigated piperacillin with tazobactam, and 3 studies without tazobactam. All studies investigated second and third trimester pregnant women or pregnant women during delivery. Bourget et al. [30] and Heikkela et al. [29] compared pregnant women to non-pregnant women, while Voigt et al. [31] only studied PK in pregnant women and compared the results with PK data of non-pregnant women from literature. The number of included women in each study varied from 3 to 12. The mean age of the women ranged from 19 to 37 years. Mean weight of the women included in the studies varied from 60 to 73.5 kg. The mean gestational age ranged from 22–40 weeks. Table 6 provides the characteristics of the included PK study and the PK-related changes of piperacillin (with and without tazobactam) used throughout pregnancy.

When considering distribution-related parameters for piperacillin and tazobactam; Cmax (43.5–49.2% and 47.4%) was lower in second, and third trimester pregnant women, respectively, compared to non-pregnant women [29, 30]. Volume of distribution of piperacillin and tazobactam was 45.5–84.9% and 67.5% higher in second and third trimester pregnant women compared to non-pregnant women. As for the elimination-related PK parameters, AUC (13.9–41.1% and 53.9%) of piperacillin and tazobactam was lower, while CL was 80%–184.8% and 127.1% higher, respectively, in second and third trimester compared to non-pregnant women.

Of the 4 studies, 2 [29, 30] reported a desired target concentration (dependent on the micro-organism, Table 2) and the probability of target attainment with the currently used dosing regimens of piperacillin with or without tazobactam in pregnant women. All 4 studies reported the need of a higher dose to increase the probability of target attainment in pregnant women. However, studies did not report the required evidence-based dosing to obtain target attainment.

In summary, Vd is lower during pregnancy compared to the non-pregnant state resulting in a higher Cmax. Finally, AUC was lower in pregnant compared to non-pregnant women, caused by a higher CL in pregnant women.