A comparison of aztreonam and two regimens of gentamicin in a rabbit model of intra-amniotic infection and sepsis

OBJECTIVE

To compare aztreonam in a standard dose with two gentamicin doses in the early treatment of experimental intra-amniotic infection in rabbits induced by intracervical inoculation with Escherichia coli.

METHODS

Timed pregnant rabbits on day 21 (70% of gestation) were inoculated intracervically with 10(4)-10(5) colony-forming units of E coli. After inoculation, the animals were treated with one of three regimens: 1) aztreonam at 90 mg/kg/day ("standard" dose in humans), 2) gentamicin at 4.5 mg/kg/day ("standard" dose in humans), or 3) higher-dose gentamicin at 6.0 mg/kg/day, each given in three divided doses daily. Outcomes included fever, delivery, and presence of a live fetus. At necropsy, cultures were taken from endometrium, amniotic fluid, and blood. Data were analyzed by Fisher exact test because the expected cell size was fewer than five.

RESULTS

Compared with rabbits treated with aztreonam, those treated with gentamicin 4.5 mg/kg/day delivered significantly more often (P = .002), had more positive cultures (P < .001), and had significantly fewer live fetuses (P < .001). Compared with rabbits treated with gentamicin 6.0 mg/kg/day, those treated with gentamicin 4.5 mg/kg/day delivered more often (P = .003), had fewer live fetuses (P = .02), and had more positive cultures (P = .02). There were no significant differences between the aztreonam and gentamicin 6.0 mg/kg/day groups.

CONCLUSIONS

This study demonstrates in an animal model that aztreonam and gentamicin at 6.0 mg/kg/day are more effective than gentamicin at 4.5 mg/kg/day (a dose that is widely used empirically in humans) in the early treatment of experimental intra-amniotic infection in rabbits. Aztreonam was as effective as gentamicin at 6.0 mg/kg/day. In this rabbit model, in which intra-amniotic infection is accompanied by maternal sepsis, 4.5 mg/kg/day of gentamicin was not adequate for the treatment of severe maternal infection.

A clinical, histologic, and DNA study of vulvodynia and its association with human papillomavirus

OBJECTIVE

To compare the clinical and histologic characteristics of vulvodynia with or without associated human papillomavirus (HPV) DNA, as determined by polymerase chain reaction (PCR).

METHODS

We conducted a standardized chart review of patients referred for vulvodynia lasting for more than 3 months and systematically reviewed all vulvar biopsy specimens histologically. In addition, specimens were amplified by PCR followed by Southern blot hybridization to detect HPV DNA, and positive cases were typed using the Hybrid Capture system.

RESULTS

Of 55 cases, 48 were evaluable by PCR. Human papillomavirus DNA was detected in 35% (17 of 48), including 44% (four of nine) of normal cases, 25% (eight of 32) with reactive squamous atypia, 67% (four of six) with condyloma/mild dysplasia, and 100% (one of one) with moderate/severe dysplasia. Patients who were positive for HPV DNA (n = 17) were not significantly different from HPV-negative patients (n = 31) for any of 82 clinical or epidemiologic variables. When patients with normal biopsies (n = 9) were compared to those with reactive squamous atypia (n = 39), there were significant differences in only two of 82 variables (duration of symptoms and current sexual activity). Of the 17 HPV-positive cases, 13 were typeable by the Hybrid Capture system. Five (38%) were positive for low-risk HPV types, three (23%) were positive for high-risk HPV types, and five (38%) were positive for both low- and high-risk types.

CONCLUSIONS

Vulvodynia associated with HPV DNA is clinically identical to vulvodynia without HPV DNA, and vulvodynia associated with normal biopsy findings is very similar to that with reactive squamous atypia. These data suggest that HPV does not cause vulvodynia.

Is penicillin G a better choice than ampicillin for prophylaxis of neonatal group B streptococcal infections?

A review of the pharmacokinetics and the narrow spectrum of action of penicillin G favors this antibiotic over ampicillin for the prophylaxis of early neonatal group B streptococcal (GBS) disease. Penicillin G provides good placental transfer and fetal and neonatal tissue levels. Group B streptococcal disease has a narrower range of sensitivities to penicillin G than to ampicillin, and the potential for selecting more resistant organisms may be greater with the broader spectrum of ampicillin than with penicillin G.

Neonatal group B streptococcal sepsis during 2 years of a universal screening program

OBJECTIVE

To assess the feasibility and efficacy of a protocol for universal screening for group B streptococci combined with selective intrapartum prophylaxis at a teaching hospital.

METHODS

This is a descriptive study of experience with a standardized protocol in which patients were screened at 26-28 weeks with a rectal and genital culture placed directly in selective media. As risk factors, we used clinical chorioamnionitis, preterm birth, and rupture of the membranes greater than 12 hours. Participants were all women receiving prenatal care at our hospital. Major outcomes were compliance and neonatal sepsis due to group B streptococci.

RESULTS

The prevalence of rectal and genital group B streptococci was 18.5% of 3721 screened women. Of culture-positive women, 35% developed risk factors (9% chorioamnionitis, 13% preterm birth, and 13% membrane rupture greater than 12 hours at term). With strict application of criteria, the compliance rate in administering indicated prophylaxis was 80.3%. Of women receiving prophylaxis, 42% had the first dose for 4 hours or less before delivery. There were five cases of group B streptococcal neonatal sepsis, resulting from either protocol violations, protocol failures, or both. Compared to the historic rate of group B streptococcal sepsis of 1.5 per 1000 births at our hospital, the rate in these 2 years was 1.0 per 1000 (1.6 per 1000 in the first year and 0.5 per 1000 in the second).

CONCLUSIONS

It is feasible to conduct such a protocol, but compliance is only moderately good because the algorithm is complex. The protocol is not foolproof in preventing neonatal group B streptococcal sepsis, as there are protocol failures and violations.

Randomized comparison of ampicillin-sulbactam to cefoxitin and doxycycline or clindamycin and gentamicin in the treatment of pelvic inflammatory disease or endometritis

OBJECTIVE

To evaluate the efficacy and safety of ampicillin-sulbactam (3 g every 6 hours) in patients with pelvic inflammatory disease or postpartum endometritis using a randomized, comparative, multicenter study of parallel design.

METHODS

Eligible patients with pelvic inflammatory disease were randomized to receive either ampicillin-sulbactam or cefoxitin (2 g every 6 hours) plus doxycycline (100 mg every 12 hours). Those with endometritis were randomized to ampicillin-sulbactam or clindamycin (900 mg every 8 hours) plus gentamicin (1.5 mg/kg every 8 hours). In the ampicillin-sulbactam group, chlamydia-positive patients also received oral doxycycline.

RESULTS

For pelvic inflammatory disease, the clinical response rates (cure or improvement) were 85.5% (47 of 55) and 89.6% (43 of 48) in the ampicillin-sulbactam and cefoxitin and doxycycline groups, respectively (chi 2 = 0.10, P = .76). For endometritis, the clinical response rates were 88.7% (141 of 159) and 90.8% (139 of 153) in the ampicillin-sulbactam and clindamycin and gentamicin groups, respectively (chi 2 = 0.15, P = .70). The percentages of patients with pelvic inflammatory disease who had adverse experiences were not significantly different in the cefoxitin and doxycycline group (47% [29 of 62]) than in those receiving ampicillin-sulbactam (33% [22 of 66]) (P = .12). These adverse effects were mostly mild or moderate. In the endometritis subjects, the incidence of adverse experiences in the ampicillin-sulbactam group (11% [20 of 179]) was comparable to that during treatment with clindamycin and gentamicin (12% [22 of 180]). These adverse experiences were also mostly mild to moderate.

CONCLUSION

Ampicillin-sulbactam is as effective and well tolerated as combination regimens using cefoxitin plus doxycycline and clindamycin plus-gentamicin for the treatment of pelvic inflammatory disease or endometritis, respectively.

A survey of practices in infectious diseases by obstetrician-gynecologists

OBJECTIVE

To survey current practices among obstetricians and gynecologists concerning a wide range of infectious diseases to guide future efforts in continuing medical education.

METHODS

A survey questionnaire of multiple-choice questions was mailed to 2500 physicians, under age 65, randomly selected from the American Medical Association specialty list of obstetrician-gynecologists. The first 500 returns constituted the data set and were analyzed by computer.

RESULTS

We found many clinical areas in which practice patterns were deemed appropriate, including antibiotic selection, universal screening for hepatitis B, and follow-up of urinary tract infection in pregnancy. In other areas, marked by controversy among "experts," practice patterns varied widely. These areas included management of premature rupture of the membranes and premature labor, and universal screening for group B streptococci. Areas in need of further continuing medical education efforts include management of perinatal viral infections and diagnosis and treatment of sexually transmitted diseases.

CONCLUSIONS

Although this survey indicated that practice patterns of American obstetricians and gynecologists are appropriate in many clinical areas relevant to infectious diseases, there are other clinical conditions requiring future efforts in continuing medical education.

Adverse perinatal outcome and resistant Enterobacteriaceae after antibiotic usage for premature rupture of the membranes and group B streptococcus carriage

OBJECTIVE

To report a case series of adverse perinatal outcomes associated with resistant Enterobacteriaceae after antibiotic usage for premature rupture of the membranes (PROM) and group B streptococcus.

METHODS

Maternal and neonatal records were reviewed of four cases in which adverse perinatal outcomes occurred from resistant Enterobacteriaceae after antibiotic usage for either PROM or positive group B streptococcal cultures. Information on clinical setting, antibiotic usage, maternal and neonatal complications, and maternal and neonatal cultures was noted.

RESULTS

All four cases were complicated by PROM at 25-35 weeks' gestation. Ampicillin or amoxicillin was used in several clinical settings, including therapeutically for the presence of group B streptococcus, presumptively for PROM, and prophylactically pending the results of group B streptococcal cultures. Clinical chorioamnionitis subsequently developed in all four cases, and in two cases the maternal course was prolonged and complicated by persistent fever and the need for therapy for pelvic vein thrombophlebitis. Two neonates died from fulminant clinical sepsis. A third infant, one of a twin gestation, was stillborn, presumably because of sepsis. In three cases, neonatal and placental cultures revealed Escherichia coli resistant to ampicillin; in the fourth case, Klebsiella pneumoniae was identified, with only intermediate sensitivity to ampicillin.

CONCLUSION

Resistant Enterobacteriaceae associated with adverse perinatal outcomes may result from the use of antibiotics, such as ampicillin or amoxicillin, after PROM. In deciding whether antibiotic therapy for PROM or group B streptococcal prophylaxis is appropriate, the value of purported benefits must be weighed against presumably infrequent but serious outcomes, including neonatal sepsis and death due to selection or overgrowth of resistant organisms.

A rabbit model for ascending infection in pregnancy: intervention with indomethacin and delayed ampicillin-sulbactam therapy

OBJECTIVE

In a modified pregnant rabbit model using intracervical inoculation of Escherichia coli we investigated the effects of administration of delayed antibiotics and indomethacin on outcomes.

STUDY DESIGN

We inoculated 10(5) colony-forming units of Escherichia coli or saline solution bilaterally in the cervix of New Zealand White rabbits at 70% of gestation and assigned animals to ampicillin-sulbactam therapy beginning at 0, 4, 8, 12, and 16 hours after inoculation with Escherichia coli or to no antibiotic therapy. We alternated indomethacin pretreatment in rabbits receiving no antibiotic therapy and rabbits starting ampicillin-sulbactam 4 hours after inoculation.

RESULTS

Compared with saline solution inoculated control animals, those inoculated with Escherichia coli (and given no antibiotic therapy) had significant increases in fetal loss, fever, bleeding at 24 hours, and positive cultures (100%, 92%, 76%, 98% versus 0%, respectively, all p < 0.01). In Escherichia coli-inoculated animals receiving no antibiotic therapy pretreatment with indomethacin significantly decreased bleeding and delivery within first 24 hours compared with those not treated with indomethacin (p < 0.05) but did not significantly improve fetal survival. Ampicillin-sulbactam treatment stated at 0, 4, 8, and 12 hours after inoculation resulted in improved fetal survival compared with the untreated group (100%, 56%, 50%, 50% versus 0%, respectively, all p < 0.05). Treatment initiated at 16 hours resulted in outcomes similar to Escherichia coli-inoculated animals receiving no antibiotic therapy.

CONCLUSION

Intracervical Escherichia coli inoculation produced infection in the uterus and uniform pregnancy loss. Pretreatment with indomethacin did not result in improved fetal survival. Ampicillin-sulbactam therapy, initiated as long as 12 hours after Escherichia coli inoculation, resulted in significant improvement in fetal survival compared with antibiotic therapy. We believe this model mimics ascending infection in pregnancy more closely than do previous animal models.

Chorioamnionitis and bacterial vaginosis

An improved understanding of bacterial vaginosis and of clinical intraamniotic infection and histologic chorioamnionitis has produced data showing strong associations among these conditions. It has recently been shown that the microorganisms in both bacterial vaginosis and clinical intraamniotic infection are similar, of which anaerobes, Gardnerella vaginalis, and Mycoplasma hominis are the predominant organisms in both. Furthermore, in the amniotic fluid of women with intraamniotic infection, strong associations among anaerobes, G. vaginalis, and M. hominis have recently been observed. In two epidemiologic studies (one in a high-risk group of women in labor and another in a lower risk group of antepartum women), the presence of bacterial vaginosis has been associated with the development of intraamniotic infection. Additional recent studies have provided new evidence that histologic inflammation of the placental membranes is associated with both clinical intraamniotic infection and positive cultures of the placenta. Multiple logistic regression analysis has shown a relationship between isolation of organisms from the chorioamnion and bacterial vaginosis.

Amniotic fluid tumor necrosis factor-alpha and interleukin-1 in a rabbit model of bacterially induced preterm pregnancy loss

OBJECTIVE

The objective of this study was to determine whether the cytokines tumor necrosis factor-alpha, interleukin-1 alpha, and interleukin-1 beta were produced in the amniotic fluid of the rabbit after intracervical inoculation with Escherichia coli.

STUDY DESIGN

Timed pregnant rabbits on day 21 (70% of gestation) were inoculated with a hysteroscope intracervically with 10(4) to 10(5) colony-forming units Escherichia coli or sterile saline solution. Escherichia coli-inoculated animals (N = 16) were put to death at 4, 8, 12, and 16 hours after inoculation. Control animals (N = 6) were put to death at similar intervals. At death, cultures were taken from endometrium, amniotic fluid, peritoneum, and blood. Amniotic fluid was collected and assayed for tumor necrosis factor bioactivity by a modified fibroblast cytotoxic assay in L929 cells, for interleukin-1 alpha, and interleukin-1 beta with a specific radioimmunoassay, and for prostaglandin E2 and prostaglandin F2 alpha by radioimmunoassay.

RESULTS

Levels of amniotic fluid tumor necrosis factor-alpha, interleukin-1 alpha, and interleukin-1 beta were elevated as early as 4 hours after inoculation in some animals and by 12 to 16 hours after inoculation in all. Levels of all three cytokines correlated significantly with time from intracervical inoculation with Escherichia coli (p < 0.05). Levels of amniotic fluid prostaglandin E2 and prostaglandin F2 alpha correlated significantly with time from intracervical inoculation with Escherichia coli (p < 0.05).

CONCLUSIONS

Levels of tumor necrosis factor-alpha, interleukin-1 alpha, interleukin-1 beta, prostaglandin E2 and prostaglandin F2 alpha are elevated in the amniotic fluid of rabbits after intracervical inoculation with Escherichia coli. Similarity exists between elevations of amniotic fluid cytokines in this model and in cases of intraamniotic infection and preterm labor unresponsive to tocolytics in humans. Modulation of cytokines may offer a strategy for improvement of outcome in this experimental model of infection-induced pregnancy loss.

A retrospective review of the efficacy and safety of prostaglandin E2 with premature rupture of the membranes at term

OBJECTIVE

The purpose of this study was to assess rates of endometritis, clinical chorioamnionitis, cesarean delivery, and neonatal sepsis from the records of patients with premature rupture of the membranes (PROM) and an unfavorable cervix treated with vaginal prostaglandin (PG) E2 in comparison with those in the literature.

METHODS

Using a computer data base at Denver General Hospital, we identified 146 women with PROM and cervical dilatation of 2 cm or less at term who were treated with PGE2 gel or suppositories. The records were reviewed to identify rates of maternal and neonatal infection and complications, as well as cesarean delivery.

RESULTS

The cesarean rate was 12%. Chorioamnionitis developed in 6.8% of the study group and endometritis in 2%. Neonatal complications were limited to two with low Apgar scores (less than 7 at 5 minutes), one with microbiologically confirmed sepsis, and two with positive urine counterimmunoelectrophoresis for group B streptococcus. The only instance of neonatal sepsis occurred in a patient with rupture of membranes longer than 24 hours.

CONCLUSIONS

The use of vaginal PGE2 suppositories for induction of labor and cervical ripening in term patients with PROM was accompanied by a high rate of vaginal delivery and a low rate of maternal and neonatal complications in a city hospital setting.

Life-threatening puerperal infection due to group A streptococci

We describe two patients with life-threatening puerperal infection due to group A beta-hemolytic streptococcus. Each patient had bacteremia, shock, and multi-organ involvement. Both cases were compatible with a recently described streptococcal toxic shock-like illness. Both women failed to improve despite vigorous medical and antibiotic therapies, and each required hysterectomy. Obstetricians should be alert to the importance of early diagnosis and treatment of this potentially lethal infection.

A review of premature birth and subclinical infection

Premature birth causes high rates of neonatal morbidity and mortality. There are multiple causes of preterm birth. This article reviews the evidence linking subclinical infection and premature birth. Although maternal genital tract colonization with specific organisms has been inconsistently associated with preterm birth and/or premature rupture of membranes, some infections have been consistently associated with preterm delivery. The association of histologic chorioamnionitis with prematurity is a consistent finding, but the mechanisms require further study. The relationship between histologic chorioamnionitis infection and the chorioamnionitis of prematurity requires additional research. A varying number of patients in "idiopathic" preterm labor have positive amniotic fluid cultures (0% to 30%), but it is not clear whether infection preceded labor or occurred as a result of labor. Evidence of subclinical infection as a cause of preterm labor is raised by finding elevated maternal serum C-reactive protein and abnormal amniotic fluid organic acid levels in some patients in preterm labor. Biochemical mechanisms for preterm labor in the setting of infection are suggested by both in vitro and in vivo studies of prostaglandins and their metabolites, endotoxin and cytokines. Some, but by no means all, antibiotic trials conducted to date have reported decreases in prematurity. These results support the hypothesis that premature birth results in part from infection caused by genital tract bacteria. In the next few years, research efforts must be prioritized to determine the role of infection and the appropriate prevention of this cause of prematurity.

Polymerase chain reaction-based genotyping for allotypic markers of immunoglobulin kappa shows allelic association of Km with kappa variable segment

Allotypic markers of immunoglobulin kappa (Km) may be determined using a novel method of amplification of the constant segment (C kappa) (IGKC) by polymerase chain reaction (PCR) followed by restriction enzyme digestion. Restriction sites in the C kappa PCR product correlate with allotypic differences among Km(1), Km(1,2), and Km(3) alleles. An AccI site in the PCR product correlates with Km(3); and presence or absence of a MaeII site correlates with the Km(1) or Km(1,2) allele, respectively. Km allelic frequencies were determined in a Caucasian population and compared to genotypic frequencies of nearby polymorphic markers. Among unrelated individuals with rheumatoid arthritis (RA) and controls, there is no evidence of allelic association between CD8A and polymorphic markers of the immunoglobulin kappa region [a V kappa (IGKV) BglII polymorphism about 24 kb centromeric to C kappa, Km allotype, and a SacI polymorphism 3.5 kb telomeric to the C kappa segment]. Similarly, there is no allelic association of the SacI C kappa polymorphism with Km or with the BglII V kappa polymorphism. However, there is evidence of allelic association of V kappa B3 and Km, specifically between the V kappa BglII 2.2-kb allele and Km(3) and also between the V kappa 3.5-kb allele and Km(1,2). Therefore, Km typing by PCR-based methods suggests the presence of allelic association between polymorphisms within the coding region of the C kappa segment and the nearest V kappa segment.

Characteristics of three vaginal flora patterns assessed by gram stain among pregnant women. Vaginal Infections and Prematurity Study Group

This study was undertaken to define the characteristics and persistence of vaginal flora in 7918 pregnant women at 23 to 26 weeks' gestation. Vaginal smears were categorized as normal (predominant lactobacilli), intermediate (reduced lactobacilli), or positive for bacterial vaginosis. The women with normal flora were least likely to have elevated vaginal pH, amine odor, milky discharge, or colonization by Gardnerella, Bacteroides, or genital mycoplasmas. Women with intermediate vaginal flora had intermediate frequencies of these clinical signs and microorganisms. Group B streptococci and yeast were associated with normal or intermediate flora, whereas Neisseria gonorrhoeae and Chlamydia trachomatis were recovered more frequently from women with intermediate flora or bacterial vaginosis. Trichomonas vaginalis was most associated with intermediate flora. At follow-up, 81% of the women with normal flora had remained normal. Of the women with intermediate flora, 32% acquired bacterial vaginosis and 30% shifted to normal flora. Only 12% of the women with bacterial vaginosis had shifted to normal flora. We conclude that there are two primary stable vaginal flora patterns (normal flora or bacterial vaginosis) and a third less distinct transitional flora pattern between these two.

The effects of pH and osmolality on bacterial growth in amniotic fluid in a laboratory model

In studying the effects of amniotic fluid on bacterial growth in a laboratory model, we noted that the pH of the fluid appeared to exert an independent effect. This study was designed to test the ability to control the growth of Escherichia coli in amniotic fluid simply by controlling two important growth conditions, pH and osmolality. The effects of pH and osmolality on growth of E. coli were systematically studied in a standard media and in amniotic fluid. Optimal ranges in standard media were pH of 5.6 to 6.6 and osmolality of 150 to 215 mOsm. When the results of growth at 24 hours were corrected for pH by analysis of covariance, the presence of amniotic fluid or phosphate had no effect. We found pH to be the only variable predictive of bacterial growth in amniotic fluid in this laboratory model.

A rabbit model for bacterially induced preterm pregnancy loss: intervention studies with ampicillin-sulbactam

We conducted experiments with a previously described rabbit model of Escherichia coli-induced preterm pregnancy loss. Does at 70% gestation were inoculated hysteroscopically with 0.2 ml of Escherichia coli (10(5) colony-forming units per milliliter) or saline solution. Animals were randomly assigned to either receive treatment with ampicillin-sulbactam (begun 1 to 2 hours before inoculation and continued for up to 7 days) or to receive no therapy. Animals were killed after delivery or after 7 days. Saline solution-inoculated animals had no pregnancy loss. Of the Escherichia coli-inoculated animals, those treated with ampicillin-sulbactam had significantly fewer deliveries, fewer positive cultures, and more live fetuses than the untreated animals (p less than or equal to 0.001). Cultures from multiple sites, amniotic fluid prostaglandin levels, and maternal progesterone levels were obtained, and the placenta, uterus, and fetal lung were histologically evaluated. In the second phase of the study, the Escherichia coli-inoculated animals were treated with ampicillin-sulbactam at one of three times: at inoculation or 2 or 4 hours after inoculation. The Escherichia coli-inoculated does treated with ampicillin-sulbactam at or before inoculation had significantly fewer deliveries, fewer positive cultures, and more live fetuses than the Escherichia coli-inoculated does in which treatment was delayed 4 hours (p less than or equal to 0.01).

Progress in pathogenesis and management of clinical intraamniotic infection

In the past decade, gratifying progress has been achieved in our understanding of clinical intraamniotic infection. With a usual incidence of 1% to 4%, clinical intraamniotic infection mainly develops as an ascending process after prolonged rupture of the membranes and labor, but other cases may be hematogenous in origin whereas still others complicate intrauterine procedures. The most common organisms isolated in amniotic fluid of cases of intraamniotic infections are anaerobes, genital mycoplasmas, group B streptococci, and Escherichia coli. The latter two are found most commonly in maternal or neonatal bacteremia complicating intraamniotic infection. Although the diagnosis remains largely a clinical one, laboratory tests have been suggested to confirm the diagnosis in women with symptoms. These include amniotic fluid Gram stain, gas-liquid chromatography, and leukocyte esterase measurement. Maternal treatment consists of antibiotic therapy and delivery. Studies to date have used a penicillin plus an aminoglycoside, with some authors advocating the addition of clindamycin after cesarean delivery. Other broad-spectrum regimens may be equally effective. Complications of clinical intraamniotic infections include an increase in cesarean section rate and in maternal and neonatal bacteremia. Poor neonatal outcomes in intraamniotic infection are more likely in the following cases: (1) when E. coli or group B streptococci are present in the amniotic fluid; (2) when the infant has a low birth weight; (3) when maternal antibiotic therapy is delayed until after delivery.