Group B streptococci (GBS) colonizes the vagina and rectum in 10–30% of pregnant women.1 In the newborn, GBS is a leading cause of neonatal sepsis and a major cause of pneumonia and meningitis.2
In 2002, the Centers for Disease Control and Prevention (CDC) issued guidelines to recommend that all pregnant women be screened at 35–37 weeks of gestation for GBS and, if positive, treated with intrapartum antibiotic prophylaxis. The aims of prophylaxis are 1) to decrease colony counts at the time of delivery; 2) to prevent the organism from ascending and proliferating in the amniotic fluid compartment; and 3) to achieve adequate levels of effective antibiotic in the fetal bloodstream during labor. For penicillin nonallergic patients, the protocol recommends a 5-million-unit infusion of penicillin G, followed by 2.5 million units every 4 hours until delivery.3 At least 4 hours of intrapartum antibiotic prophylaxis are recommended. The origin of this 4-hour time threshold is not clear. A systematic review of published evidence suggests, at best, that in women with established risk factors for early-onset GBS disease of the newborn, greater than 1 or 2 hours of intrapartum antibiotic prophylaxis is effective in reducing neonatal GBS colonization or disease.4–8 Studies evaluating either the pharmacokinetics or the efficacy of the currently recommended penicillin G dosing regimen in a cohort of maternal-fetal dyads have not been performed.
The purpose of this prospective cohort study was to examine the amount of time required after maternal infusion of penicillin G to achieve the minimal inhibitory concentration (MIC) for GBS in fetal serum and to examine the fetal pharmacokinetic profile of this maternal penicillin G–dosing regimen over time, particularly in the first 4 hours after maternal administration of the drug.
MATERIALS AND METHODS
Laboring GBS-positive women who were administered penicillin G by their medical provider according to the 2002 CDC protocol as standard of care were eligible for the study. Yale University Institutional Review Board approval was obtained, and patients were enrolled after obtaining informed consent. Inclusion criteria included pregnancy of 37 weeks or more; singleton gestation; GBS carrier status by rectovaginal or urine culture; and receipt of intravenous intrapartum penicillin G prophylaxis in standard CDC dosing regime doses. The exclusion criteria included hypertensive or renal disease, multiple gestation, penicillin allergy, and current other antibiotic usage. After consent, duration and timing of antibiotic infusions, maternal height and weight, and demographic information were all recorded from patient charts. Based on estimated means and variances from previous work on maternal serum levels,9 we estimated that we required 10 patients in each time interval (less than 1 hour, 1 to less than 2 hours, 2 to less than 3 hours, 3 to less than 4 hours, and 4 hours or more), to achieve more than 80% power to detect penicillin G concentration statistically greater than the MIC (0.1 micrograms/mL), alpha=0.05, using the CDC dosing regimen.
Umbilical cord blood samples were obtained from the hospital blood bank, which stores samples for blood typing of all infants. The blood was stored in glass tubes labeled with patient name and medical record number at 4°C for one week until a research team member collected them. The blood was then centrifuged at 3,000g for 15 minutes and stored in 0.5 mL aliquots at -80°C until high-performance liquid chromatography (HPLC) analysis. This approach was validated by comparing penicillin G levels in fresh cord blood collected at delivery to blood bank samples in 18 subjects. For all 18 samples, the penicillin G level was consistently slightly lower in the blood bank samples when compared with the delivery floor samples (Fig. 1). The percent lost ranged from 0.40% to 16.40%, with a mean of 9.54±4.76%. Using blood bank samples routinely collected at all deliveries facilitated the potential inclusion of all eligible subjects who delivered rapidly (fewer that 4 hours) upon admission to the labor unit. By analyzing the blood bank values, our data contain a degree of underestimation. We used this conservative approach because if the underestimated values from the blood bank were significant, then by extension, values at the time of delivery would also be significant.
The penicillin G concentration was measured by high-performance liquid chromatography and ultraviolet detection (HPLC-UV). The ESA reverse-phased HPLC system (Chelmsford, MA) was equipped with two dual-piston pumps (Model 582), a refrigerated autosampler (Model 542), high pressure mixer, and an ESA Model 528 UV-VIS detector. It was controlled and data acquired using ESA CoulArray for Windows software. An Intersil ODS-3 C18 column (150×4.6 I.D.), 5 micrometer particle size (GL Science Inc, Cambridge, MA) was protected with a Platinum C18 (7.5×4.6 mm I.D.) 5-micrometer particle size guard column (AllTech GmbH, Nicholasville, KY). The mobile phase was prepared using 0.05 dihydrogen phosphate (99.99% purity, pH 5.0; VWR International, West Chester, PA) and acetonitrile (90:10 vol/vol; Fisher Scientific, Fair Lawn, NJ). Serum samples were analyzed at a gradient condition with mobile phase A of 0.05 mol/L sodium dihydrogen phosphate in 10% acetonitrile (pH 5.0) and B of 100% acetonitrile. The internal standard was prepared by dissolving 2.5 mg of ampicillin sodium salt (potency 98%; Sigma Chemical Company, St. Louis, MO) in 5 mL of HPLC grade water (Fisher Scientific) to a final stock concentration of 40 micrograms/mL of internal standard for each sample. Penicillin sodium salt (5 mg, potency 99%; Sigma Chemical Company) was dissolved in 5 mL of HPLC grade water to a final stock concentration of 1 mg/mL. Serum standards for penicillin G were prepared in blank serum at 2.5, 5, 10, 20, and 40 micrograms/mL to create a standard curve for quantification (r=.9997). Both patient serum samples and serum standard curve samples were deproteinated by adding 100 microliters of sample to 108 microliters acetonitrile/ampicillin (internal standard) solution for a final concentration of 40 micrograms/mL of internal standard for each sample. Samples were vortexed for 15 seconds, placed at 4°C for 10 minutes, and then centrifuged at 3,000g for 10 minutes. One hundred microliters of aqueous phase from each serum sample was transferred to a clean autosampler vial. A 25-microliter sample was injected onto the column at a flow rate of 1.0 mL/min. Both penicillin G and ampicillin (internal standard) were detected at 200 nm. Ampicillin and penicillin G had elution times of 2.28 and 13.2 minutes, respectively. The coefficient of variation (CV) was calculated by running three samples in three separate runs (CV=2.9637; accuracy range: ±3%). Additionally, every single run contained a serum standard curve point of known concentration. The calculated CV for those samples was 2.12%, verifying accuracy for all samples on each run. The lower limit of detection of penicillin G sodium was 0.192 micrograms/mL. A blank serum sample was also run and showed no evidence of a penicillin G peak, ruling out the possibility of carryover from run to run. The concentration of penicillin G sodium salt was quantified by comparing peak height ratio (penicillin G/internal standard) from the unknown cord serum samples and those obtained from the penicillin G standard curve. Statistical analysis was performed using Student t test, analysis of variance, and multivariable linear regression using SAS 9.1 software (SAS Institute Inc., Cary, NC).
One hundred ten eligible patients were approached at Yale-New Haven Hospital from June 6, 2007, until August 14, 2007. Ninety-eight patients consented to participate, yielding a participation rate of 89%. Reasons for nonparticipation included lack of time or desire to discuss the study or to read and sign the consent form. The cohort was representative of the population served by our urban, academic medical center; demographic and clinical characteristics of the patients enrolled are listed in Table 1.
For all patients in the study, the number of hours from first dose administration until delivery ranged from 0.53 to 24.5 hours, with a median of 3.48 hours. The penicillin G concentrations for all patients ranged from 1.02 to 17.93 micrograms/mL, with a mean of 6.25±4.15 micrograms/mL. Fetuses exposed to fewer than 4 hours prophylaxis had higher penicillin G levels than those exposed to greater than 4 hours (P=.003). Penicillin G levels were examined for each time interval; those who failed to receive the scheduled second dose and those who received more than 8 hours of prophylaxis were examined separately. All groups achieved penicillin G levels significantly above the MIC for GBS (0.1 micrograms/mL),10,11P<.002. Furthermore, penicillin G levels observed in each individual cord blood sample were 10–179-fold above the MIC (Table 2).
Each of the seven groups was also compared with one another using analysis of variance. Penicillin G levels for those patients receiving less than 1 hour of prophylaxis (group 1) were statistically significantly greater (P<.05) than all other groups of patients receiving more than 2 hours of prophylaxis (groups 3, 4, 5, 6, and 7). Rather than requiring 4 hours to reach levels of penicillin G above the MIC in the fetal bloodstream, we found statistically significantly higher levels during the 0–2 hour time points when compared with longer durations, even after subsequent redosing with 2.5 million units at 4-hour intervals.
For graphical pharmacokinetic analysis, the data for those patients who received only one dose of 5 million units of penicillin G (Fig. 2) were analyzed by using time elapsed since most recent dose. As can be seen in Figure 2, the relationship between penicillin G concentration and time elapsed since dose of 5 million units of penicillin G was not strictly linear. The concentration rose linearly (R2=.40) (the equation is penicillin=0.255 (minutes)+1.2718), as the penicillin G made its way across the placenta and into the fetal circulation, until the 1-hour time point. After 1 hour, the penicillin G concentration decreased according to a power-decay model (R2=.67) determined by optimizing r2 (concentration penicillin=0.255[minutes]+1.2718) and (concentration penicillin=2,745.5[minutes]−1.2503). This period represents the combined efforts of both maternal and fetal clearance, as well as maternal and fetal metabolism. To accommodate the nonlinearity of penicillin G levels over time, the time variable was transformed according to the power-decay model equation above for values greater than 1 hour. Multiple linear regression analysis performed on the cohort of all patients showed that fetal penicillin G levels were associated with duration of exposure to penicillin, time since last dose, dosage, and number of doses, but not maternal BMI. Assuming maternal BMI is an adequate, although imperfect, marker for maternal size and maternal volume of distribution, the maternal volume of distribution did not appreciably alter the fetal cord blood concentrations of penicillin.
As seen in figure 2, the highest value of penicillin G concentration in the cord blood was observed at approximately 1 hour after administration of the loading dose of 5 million units. The lowest value observed for all patients was 1.02 micrograms/mL seen in a patient who delivered 5 hours and 34 minutes after receiving her initial dose. Additionally, the group with the lowest mean (2.28±0.89 micrograms/mL) was represented by six patients who failed to receive their additional 2.5 million units at the 4-hour time point. For patients receiving maintenance doses of 2.5 million units every 4 hours, levels remained consistently above the MIC and decayed in the same fashion as the loading dose. Of note, penicillin G levels did not accumulate after repeated maintenance dosing.
The data presented here demonstrates that shorter durations of exposure to intrapartum antibiotic prophylaxis are effective in attaining levels of penicillin G in the neonatal bloodstream significantly above the minimum inhibitory concentration for group B Streptococcus. Fetal serum penicillin levels peak within 1 hour of maternal intravenous administration of penicillin G. Levels recorded until 2 hours after initiation of chemoprophylaxis remain significantly above all later time intervals. In this cohort, individual samples were 10–179-fold above the MIC at all time points examined. The declining levels of penicillin G levels in the six patients who failed to receive the protocol-recommended 2.5 million units at the 4-hour time point supports the 4-hour dosing interval. Additionally, patients who received up to six additional doses of 2.5 million units every 4 hours had similar levels to those who received two additional doses. Fetal serum penicillin G levels do not build with time, rather they nadir at the end of each 4-hour interval. Therefore, adherence to dosing every 4 hours, independent of the duration of the intrapartum prophylaxis should be a priority.
Up until this point, little has been documented regarding the pharmacokinetic properties of penicillin G in the pregnant woman and her fetus. Similar studies on ampicillin have shown that ampicillin levels rise rapidly in the fetal serum following maternal intravenous administration.12 To date, studies on penicillin G pharmacokinetics have generally been performed on nonpregnant women or men or neonates themselves (dose given to the neonate after delivery).13–15 One recent study examined serum concentrations of penicillin G in pregnancy, but examined only the maternal circulation and not the fetal.9 Another study measured maternal and fetal penicillin G levels after one intramuscular dose of penicillin G benzathine. This study showed levels above the MIC at 30 days after injection.16 Our current investigation documents penicillin G levels in the fetus using the current intravascular CDC dosing regimen.
Several limitations in our study should be considered. Due to natural variability in duration of labor, we cannot control duration of prophylaxis in relation to timing of umbilical cord blood sampling. This investigation was therefore limited to observational study and a single sample per fetus at the time of birth. Traditional pharmacodynamic modeling, in which multiple serum samples are collected at designated time points after drug administration, is therefore not possible.
Furthermore, our methodology used umbilical blood samples retrieved from the blood bank. The focus of our investigation was rapidly progressing, less than 4-hour deliveries. The challenge of recruiting and consenting women in precipitous labor motivated us to evaluate the use of blood bank samples. We could therefore approach mothers after delivery and, if they consented, retrieve the blood bank sample. Our analysis of 18 fresh umbilical cord blood samples compared with their matched blood bank samples showed that this approach was legitimate. The blood bank samples represent a slight underestimation of penicillin G levels contained in cord blood at time of delivery.
This study did not evaluate levels of penicillin G in amniotic fluid. Evaluating amniotic fluid would have required performing an invasive procedure, amniocentesis, in the setting of rapidly progressing deliveries or would have required limiting our cohort to women undergoing caesarean deliveries, which would have limited the generalizability of our findings. Because we did not examine amniotic fluid, we are unable to document whether sufficiently high levels of penicillin G were obtained in amniotic fluid to prevent transmission by that route. However, other studies on intrapartum ampicillin and cefazolin have shown that when adequate concentrations were achieved in the cord blood, they were also reported in the amniotic fluid.17,18
Ultimately, studies to assess the optimal duration of intrapartum prophylaxis will require correlation with clinical outcomes. However, early onset GBS sepsis is an infrequent event and even more rare in the setting of intrapartum prophylaxis. Only a well designed case–control or a large cohort study could examine this outcome. We must also consider that most cases of early onset GBS sepsis may represent fetal exposure to GBS in utero, in the setting of ruptured membranes, well before the initiation of intrapartum antibiotic prophylaxis. The fetal serum penicillin level and the duration of exposure to intrapartum antibiotic may be irrelevant in these cases.
Even with the most valiant of efforts, there will frequently be GBS-positive mothers who arrive at the labor floor and deliver in fewer than 4 hours. Obstetric providers have little control over the time patients arrive at the hospital to begin prophylaxis and likewise little control over the progression of labor and the ultimate timing of delivery. Providers may believe that 4 hours of prophylaxis are necessary to achieve adequate levels in the fetal bloodstream to prevent GBS transmission and therefore, may choose not to begin penicillin G dosing during precipitous labor. Likewise, providers may attempt to prolong labor to reach the 4-hour recommendation. Our data indicate that fetal serum levels far exceed the MIC at durations well under 1 hour, suggesting that antibiotic prophylaxis should be pursued even in the most precipitous of deliveries and that it is unnecessary to prolong labor to achieve adequate levels in fetal serum. Likewise, prolonging labor beyond 1 hour will not achieve higher penicillin levels in fetal serum.
Knowledge about the dosing regimen has implications beyond the labor and delivery floor. Preliminary studies as well as data from a large health maintenance organization demonstrate that 40–50% of GBS-colonized women do not receive antibiotics at least 4 hours before delivery due the rapidity of their labors.19 This is especially notable for multiparous women. According to the 2002 CDC guidelines, the newborns of all GBS-positive women who present to labor units and deliver before receiving 4 hours of intrapartum antibiotic prophylaxis are deemed as “at risk” and recommended to undergo blood cultures, complete blood count, and 48 hours of observation.3 At some institutions, these infants have been placed in designated observation units for up to 6 hours after delivery to monitor for signs of sepsis, often causing great angst for parents and care providers. These interventions have not been proven to reduce or detect more cases of GBS sepsis.20 Knowledge that fetal serum penicillin G levels are far above the MIC within 1 hour raises the possibility that these interventions and testing may be, at best, superfluous and, at worst, expensive and deleterious.
This study shows that fetal serum penicillin G levels far exceed the MIC even for short durations of maternal intrapartum prophylaxis. However, studies that correlate duration of prophylaxis with clinical outcomes are needed. Much of the current literature has examined neonatal GBS colonization, but the usefulness of using this as a surrogate for risk of early onset GBS sepsis does not have much, if any, supporting evidence. Therefore, studies investigating duration of chemoprophylaxis in relation to incidence of early onset GBS sepsis are necessary. If those studies are in line with the evidence presented here, the results may alter GBS sepsis protocols, so that in appropriate circumstances, shorter durations of intrapartum prophylaxis may be considered adequate.
1. Schuchat A. Epidemiology of group B streptococcal disease in the United States: shifting paradigms. Clin Microbiol Rev 1998;11:497–513.
2. Gibbs RS, Schrag S, Schuchat A. Perinatal infections due to group B streptococci. Obstet Gynecol 2004;104:1062–76.
3. Schrag S, Gorwitz R, Fultz-Butts K, Schuchat A. Prevention of perinatal group B streptococcal disease. Revised guidelines from CDC. MMWR Recomm Rep 2002; 51:1–22.
4. Boyer KM, Gotoff SP. Prevention of early-onset neonatal group B streptococcal disease with selective intrapartum chemoprophylaxis. N Engl J Med 1986;314:1665–9.
5. Lin FY, Brenner RA, Johnson YR, Azimi PH, Philips JB 3rd, Regan JA, et al. The effectiveness of risk-based intrapartum chemoprophylaxis for the prevention of early-onset neonatal group B streptococcal disease. Am J Obstet Gynecol 2001;184:1204–10.
6. Pylipow M, Gaddis M, Kinney JS. Selective intrapartum prophylaxis for group B streptococcus colonization: management and outcome of newborns. Pediatrics 1994;93:631–5.
7. de Cueto M, Sanchez MJ, Sampedro A, Miranda JA, Herruzo AJ, Rosa-Fraile M. Timing of intrapartum ampicillin and prevention of vertical transmission of group B streptococcus. Obstet Gynecol 1998;91:112–4.
8. de Cueto M, Sánchez MJ, Moltó L, Miranda JA, Herruzo AJ, Ruiz-Bravo A, et al. Efficacy of a universal screening program for the prevention of neonatal group B streptococcal disease. Eur J Clin Microbiol Infect Dis 1995;14:810–2.
9. Johnson JR, Colombo DF, Gardner D, Cho E, Fan-Havard P, Shellhaas CS. Optimal dosing of penicillin G in the third trimester of pregnancy for prophylaxis against group B Streptococcus. Am J Obstet Gynecol 2001;185:850–3.
10. de Cueto M, Sánchez MJ, Moltó L, Miranda JA, Herruzo AJ, Ruiz-Bravo A, et al. Efficacy of a universal screening program for the prevention of neonatal group B streptococcal disease. Eur J Clin Microbiol Infect Dis 1995;14:810–2.
11. Boyer KM, Gadzala CA, Burd LI, Fisher DE, Paton JB, Gotoff SP. Selective intrapartum chemoprophylaxis of neonatal group B streptococcal early-onset disease. I. Epidemiologic rationale. J Infect Dis 1983;148:795–801.
12. Colombo DF, Lew JL, Pedersen CA, Johnson JR, Fan-Havard P. Optimal timing of ampicillin administration to pregnant women for establishing bactericidal levels in the prophylaxis of Group B Streptococcus. Am J Obstet Gynecol 2006;194:466–70.
13. Ebert SC, Leggett J, Vogelman B, Craig WA. Evidence for a slow elimination phase for penicillin G. J Infect Dis 1988;158:200–2.
14. Metsvaht T, Oselin K, Ilmoja ML, Anier K, Lutsar I. Pharmacokinetics of penicillin g in very-low-birth-weight neonates. Antimicrob Agents Chemother 2007;51:1995–2000.
15. Weeks JW, Myers SR, Lasher L, Goldsmith J, Watkins C, Gall SA. Antimicrobial agents: penicillins, cephalosporins, and other beta-lactam antibiotics. In: Gilman AG, Nies AS, Taylor P, editors. Goodman and Gilman’s the pharmacological basis of therapeutics. 8th ed. New York (NY): Pergamon Press; 1990. p. 1065–75.
16. Weeks JW, Myers SR, Lasher L, Goldsmith J, Watkins C, Gall SA. Persistence of penicillin G benzathine in pregnant group B streptococcus carriers. Obstet Gynecol 1997;90:240–3.
17. Bloom SL, Cox SM, Bawdon RE, Gilstrap LC. Ampicillin for neonatal group B streptococcal prophylaxis: how rapidly can bactericidal concentrations be achieved? Am J Obstet Gynecol 1996;175:974–6.
18. Fiore Mitchell T, Pearlman MD, Chapman RL, Bhatt-Mehta V, Faix RG. Maternal and transplacental pharmacokinetics of cefazolin. Obstet Gynecol 2001;98:1075–9.
19. Davis RL, Hasselquist MB, Cardenas V, Zerr DM, Kramer J, Zavitkovsky A, et al. Introduction of the new Centers for Disease Control and Prevention group B streptococcal prevention guideline at a large West Coast health maintenance organization. Am J Obstet Gynecol 2001;184:603–10.
20. Ottolini MC, Lundgren K, Mirkinson LJ, Cason S, Ottolini MG. Utility of complete blood count and blood culture screening to diagnose neonatal sepsis in the asymptomatic at risk newborn. Pediatr Infect Dis J 2003;22:430–4.