- Email: [email protected]
infection in pregnant women with preterm premature rupture of membranes and in their neonates
Journal of Hospital Infection (2003) 54, 158–160
www.elsevierhealth.com/journals/jhin
SHORT REPORT
Emergence of nosocomial Pseudomonas aeruginosa colonization/infection in pregnant women with preterm premature rupture of membranes and in their neonates A. Casettaa, F. Audibertb, F. Briveta, N. Boutrosc, C. Boithiasb, L. Lebrunc,* a
´ d’Hygie `ne Hospitalie `re, AP-HP, Ho ˆpital Antoine Be ´cle `re, Clamart, France Unite ´partement d’Obste ´trique et de Ne ´onatologie, AP-HP, Ho ˆpital Antoine Be ´cle `re, Clamart, France De c ˆpital Antoine Be ´cle `re, Clamart, France Laboratoire de Microbiologie, AP-HP, Ho b
Received 9 October 2002; accepted 18 February 2003
KEYWORDS Pseudomonas aeruginosa; Nosocomial infection; Antibiotic prophylaxis; Premature rupture of membranes
Summary The epidemiology, risk factors, maternal and neonatal outcomes of nosocomial Pseudomonas aeruginosa acquisition in preterm premature rupture of membranes were analysed. Of 63 women receiving antibiotic prophylaxis with coamoxiclav, 11 acquired P. aeruginosa vaginal carriage with a median delay of 15 days (6 –42) i.e. an incidence of 8.94 per 1000 days of expectant management. Five neonates born to 11 positive mothers were colonized or infected, three of whom died of fulminant sepsis. The duration of antibiotic treatment and multiple pregnancy were identified as independent risk factors. The epidemiological investigation revealed a vertical transmission between mothers and neonates, and suggested selective pressure of antibiotic treatment. Q 2003 The Hospital Infection Society. Published by Elsevier Science Ltd. All rights reserved.
Introduction Preterm premature rupture of membranes (pPROM), which often results from a subclinical infection, is a leading cause of prematurity and increases the risk of maternal and neonatal infections. Therefore, pregnant women with pPROM may *Corresponding author. Dr Le ´ a Lebrun, Service de Microbiologie, Ho ˆpital Antoine Be ´cle `re, 157 Rue de la Porte de Trivaux, 92141 Clamart ce ´dex, France. Tel.: þ33-1-45-37-46-29; fax: þ 33-1-46-32-67-96. E-mail address: [email protected]
benefit from an antibiotic prophylaxis under careful bacteriological surveillance when expectant management is required.1 Recently, emergence of bacteria resistant to antibiotic treatment in patients with pPROM and in their neonates has become of special clinical concern, particularly in very low birthweight infants2,3 with higher risk of Pseudomonas aeruginosa infection.4,5 We confirm that P. aeruginosa infection in this selected population is a major concern, especially in our maternity reference center. We report the epidemiology, risk factors and maternal and neonatal outcomes.
0195-6701/03/$ - see front matter Q 2003 The Hospital Infection Society. Published by Elsevier Science Ltd. All rights reserved. doi:10.1016/S0195-6701(03)00121-X
Nosocomial P. aeruginosa and pregnancy
Materials and methods From January to December 1999, all pregnant women directly admitted or referred for pPROM, occurring before 32 weeks gestation and benefiting from an expectant management for more than 48 h, were prospectively included. Lower genital tract samples for bacteriological surveillance cultures were performed at admission, twice per week and at delivery, using routine bacteriological procedures. At birth, cultures of gastric fluid and ear secretions were performed for all neonates. P. aeruginosa strains were identified by use of the API 20 NE identification system (bioMe ´rieux, Marcy l’Etoile, France). Antibiotic prophylaxis using intravenous co-amoxiclav was started at admission until results of cultures. Antibiotic treatment was subsequently adjusted and prescribed for at least five days, when vaginal colonization with high-risk pathogens for perinatal infection, such as Escherichia coli or group B streptococci, was detected. In the absence of pathogens, prophylaxis was discontinued. Other adjuvant treatment included a weekly administration of steroids to accelerate fetal lung maturation. Data on maternal age, multiple pregnancy, duration of antibiotic treatment, gestational age both at rupture and at delivery, length of latent period, cervical cerclage (a procedure used to prevent the consequences of cervical insufficiency), mode of delivery, occurrence of chorioamnionitis and postpartum endometritis were recorded for each mother. To determine the mode of acquisition or transmission of P. aeruginosa vaginal carriage, environmental investigation was carried out on the Maternity Unit water supply, including showers and tap water. All maternal, neonatal and environmental P. aeruginosa isolates were typed using
159 pulsed-field gel electrophoresis (PFGE).6 For PFGE, DNA of P. aeruginosa strains embedded in agarose plugs was digested with the Spe I restriction enzyme. Separation of the DNA fragments was carried out on a 1.2% agarose gel in a CHEF-DR III system (Bio-Rad). The separated fragments were stained with ethidium bromide and visualized by UV transillumination. Profiles with differences of more than three bands were considered unrelated.
Results None of the 63 patients included was detected positive for P. aeruginosa in the vaginal flora at the time of admission. Eleven patients (17.5%) acquired P. aeruginosa vaginal colonization during hospitalization. The rate of obstetric unit-acquired P. aeruginosa vaginal colonization was 8.94 per 1000 days of expectant management. The median delay for detection of P. aeruginosa was 15 days (6 – 42). Positive patients did not differ significantly from negative patients except for duration of antibiotic treatment and multiple pregnancy (P , 0:001 and P ¼ 0:02; respectively; Table I). In multivariate analysis (logistic regression), multiple pregnancy (OR: 6.3, 95% CI 1.0 –39.4) and duration of antibiotic treatment (OR: 1.6, 95% CI 1.2 –2.2) were the two independent risk factors for acquisition of P. aeruginosa vaginal carriage (P ¼ 0:04 and P ¼ 0:003; respectively). The vaginal carriage of P. aeruginosa had no significant impact on maternal outcome as occurrence of chorioamnionitis or post-partum endometritis were not significantly different between carrier and non carrier-patients. In contrast, among the five P. aeruginosa-positive neonates born from colonized mothers, three very low birthweight infants (, 1000 g) died from early onset P. aeruginosa fulminant sepsis. The two others were only colonized.
Table I Pregnancy characteristics and outcome according to vaginal Pseudomonas aeruginosa acquisition P. aeruginosa in vaginal cultures Variablesa
Positive patients ðN ¼ 11Þ mean(SD )
Negative patients ðN ¼ 52Þ mean(SD )
Pb
Maternal age (years) Multiple pregnancy (N) Duration of prophylaxis (days) Gestational age at rupture (weeks) Gestational age at delivery (weeks) Length of latent period (days) Cerclage Caesarean section Chorioamnionitis Post-partum endometritis
32.6 (1.3) 5 8.7 (2.1) 28.4 (2.3) 30.9 (0.9) 18.4 (3.9) 3 5 7 1
32.0 (0.7) 8 4.6 (3.1) 26.9 (3.9) 29.5 (0.5) 18.7 (2.9) 8 21 17 7
0.7 0.02 ,0.001 0.2 0.3 0.9 0.3 0.5 0.06 0.6
a b
Chi2 test, Fisher’s exact test, and Mann–Whitney U test were used when appropriate. P ,0.05 was considered significant.
160
Figure 1 Representative PFGE patterns of P. aeruginosa isolates after Spe I digestion. Lanes 1–3: environmental strains; lanes 4– 12: maternal strains; lanes 13 –16: two mother – baby pairs; lane 17 molecular weight standards.
Environmental investigation revealed the presence of P. aeruginosa in three faucets. All the maternal and the environmental P. aeruginosa strains were found to be unrelated. However patterns of each mother –baby pairs were indistinguishable (Figure 1).
Discussion To date, P. aeruginosa, naturally resistant to many b-lactam antibiotics, such as co-amoxiclav, is recognized as an infrequent cause of intrapartum or intrauterine infection.4,5 However, it seems that antibiotic prophylaxis in pPROM may select resistant bacteria and cause neonatal sepsis.2 Although the risk of nosocomial infection with resistant bacteria is recognized, no study until now has evaluated the level or precise nature of this risk. Our findings show the high rate of nosocomially acquired P. aeruginosa vaginal carriage in mothers with pPROM receiving co-amoxiclav, and its relationship with the duration of antibiotic treatment and multiple pregnancy. Molecular typing excluded mother-to-mother cross-transmission or a source of contamination
A. Casetta et al.
mediated by the environment and revealed vertical transmission of P. aeruginosa in all positive neonates. Interestingly, the polyclonal origin of all vaginal and environmental strains and the long course of treatment as a risk factor, suggested that P. aeruginosa vaginal carriage was acquired from an endogenous route mediated by antibiotic selective pressure. This vaginal carriage had no significant impact on maternal morbidity or progress of pregnancy, but might threaten the neonate, particularly in very low birthweight infants (, 1000 g) in whom P. aeruginosa is recognized as a highly virulent nosocomial pathogen.4,5 Recently, the ORACLE study7 has shown an increased incidence of necrotizing enterocolitis after antenatal use of co-amoxiclav. Our study adds further concerns about possible adverse effects of a prolonged course of broad-spectrum b-lactam prophylaxis for pPROM, particularly in the preterm population. Indeed, selection pressure for resistant strains might be one of the mechanisms responsible for the occurrence of neonatal complications. The benefit/risk ratio of systematic antibiotics for pPROM should be discussed in each specific case. Further studies are needed to evaluate the impact of shorter courses with narrow-spectrum antibiotics.
References 1. ACOG Practice Bulletin, Premature rupture of membranes: clinical management guidelines for obstetrician—gynecologists. Int J Gynecol Obstet 1998;63:75—84. 2. Edwards RK, Locksmith GJ, Duff P. Expanded-spectrum antibiotics with preterm premature rupture of membranes. Obstet Gynecol 2000;96:60—64. 3. Stoll BJ, Hansen N, Fanaroff AA, et al. Changes in pathogens causing early-onset sepsis in very-low-birth weight infants. N Engl J Med 2002;347:240—247. 4. Leigh L, Stoll BJ, Rahman M, McGowan Jr. J. Pseudomonas aeruginosa infection in very low birth weight infants: a casecontrol study. Pediatr Infect Dis J 1995;14:367—371. 5. Kyle P, Turner DPJ. Chorioamnionitis due to Pseudomonas aerginosa: a complication of prolonged antibiotic therapy for premature rupture of membranes. Br J Obstet Gynaecol 1996; 103:181—183. 6. Luk WK. An outbreak of pseudobacteraemia caused by Burkholderia pickettii: the critical role of an epidemiological link. J Hosp Infect 1996;34:59—69. 7. Kenyon SI, Taylor DJ, Tarnow-Mordi W, ORACLE Collaborative Group. Broad-spectrum antibiotics for preterm, prelabour rupture of fetal membranes: the ORACLE I randomised trial. Lancet 2001;357:979—988.
www.elsevierhealth.com/journals/jhin
SHORT REPORT
Emergence of nosocomial Pseudomonas aeruginosa colonization/infection in pregnant women with preterm premature rupture of membranes and in their neonates A. Casettaa, F. Audibertb, F. Briveta, N. Boutrosc, C. Boithiasb, L. Lebrunc,* a
´ d’Hygie `ne Hospitalie `re, AP-HP, Ho ˆpital Antoine Be ´cle `re, Clamart, France Unite ´partement d’Obste ´trique et de Ne ´onatologie, AP-HP, Ho ˆpital Antoine Be ´cle `re, Clamart, France De c ˆpital Antoine Be ´cle `re, Clamart, France Laboratoire de Microbiologie, AP-HP, Ho b
Received 9 October 2002; accepted 18 February 2003
KEYWORDS Pseudomonas aeruginosa; Nosocomial infection; Antibiotic prophylaxis; Premature rupture of membranes
Summary The epidemiology, risk factors, maternal and neonatal outcomes of nosocomial Pseudomonas aeruginosa acquisition in preterm premature rupture of membranes were analysed. Of 63 women receiving antibiotic prophylaxis with coamoxiclav, 11 acquired P. aeruginosa vaginal carriage with a median delay of 15 days (6 –42) i.e. an incidence of 8.94 per 1000 days of expectant management. Five neonates born to 11 positive mothers were colonized or infected, three of whom died of fulminant sepsis. The duration of antibiotic treatment and multiple pregnancy were identified as independent risk factors. The epidemiological investigation revealed a vertical transmission between mothers and neonates, and suggested selective pressure of antibiotic treatment. Q 2003 The Hospital Infection Society. Published by Elsevier Science Ltd. All rights reserved.
Introduction Preterm premature rupture of membranes (pPROM), which often results from a subclinical infection, is a leading cause of prematurity and increases the risk of maternal and neonatal infections. Therefore, pregnant women with pPROM may *Corresponding author. Dr Le ´ a Lebrun, Service de Microbiologie, Ho ˆpital Antoine Be ´cle `re, 157 Rue de la Porte de Trivaux, 92141 Clamart ce ´dex, France. Tel.: þ33-1-45-37-46-29; fax: þ 33-1-46-32-67-96. E-mail address: [email protected]
benefit from an antibiotic prophylaxis under careful bacteriological surveillance when expectant management is required.1 Recently, emergence of bacteria resistant to antibiotic treatment in patients with pPROM and in their neonates has become of special clinical concern, particularly in very low birthweight infants2,3 with higher risk of Pseudomonas aeruginosa infection.4,5 We confirm that P. aeruginosa infection in this selected population is a major concern, especially in our maternity reference center. We report the epidemiology, risk factors and maternal and neonatal outcomes.
0195-6701/03/$ - see front matter Q 2003 The Hospital Infection Society. Published by Elsevier Science Ltd. All rights reserved. doi:10.1016/S0195-6701(03)00121-X
Nosocomial P. aeruginosa and pregnancy
Materials and methods From January to December 1999, all pregnant women directly admitted or referred for pPROM, occurring before 32 weeks gestation and benefiting from an expectant management for more than 48 h, were prospectively included. Lower genital tract samples for bacteriological surveillance cultures were performed at admission, twice per week and at delivery, using routine bacteriological procedures. At birth, cultures of gastric fluid and ear secretions were performed for all neonates. P. aeruginosa strains were identified by use of the API 20 NE identification system (bioMe ´rieux, Marcy l’Etoile, France). Antibiotic prophylaxis using intravenous co-amoxiclav was started at admission until results of cultures. Antibiotic treatment was subsequently adjusted and prescribed for at least five days, when vaginal colonization with high-risk pathogens for perinatal infection, such as Escherichia coli or group B streptococci, was detected. In the absence of pathogens, prophylaxis was discontinued. Other adjuvant treatment included a weekly administration of steroids to accelerate fetal lung maturation. Data on maternal age, multiple pregnancy, duration of antibiotic treatment, gestational age both at rupture and at delivery, length of latent period, cervical cerclage (a procedure used to prevent the consequences of cervical insufficiency), mode of delivery, occurrence of chorioamnionitis and postpartum endometritis were recorded for each mother. To determine the mode of acquisition or transmission of P. aeruginosa vaginal carriage, environmental investigation was carried out on the Maternity Unit water supply, including showers and tap water. All maternal, neonatal and environmental P. aeruginosa isolates were typed using
159 pulsed-field gel electrophoresis (PFGE).6 For PFGE, DNA of P. aeruginosa strains embedded in agarose plugs was digested with the Spe I restriction enzyme. Separation of the DNA fragments was carried out on a 1.2% agarose gel in a CHEF-DR III system (Bio-Rad). The separated fragments were stained with ethidium bromide and visualized by UV transillumination. Profiles with differences of more than three bands were considered unrelated.
Results None of the 63 patients included was detected positive for P. aeruginosa in the vaginal flora at the time of admission. Eleven patients (17.5%) acquired P. aeruginosa vaginal colonization during hospitalization. The rate of obstetric unit-acquired P. aeruginosa vaginal colonization was 8.94 per 1000 days of expectant management. The median delay for detection of P. aeruginosa was 15 days (6 – 42). Positive patients did not differ significantly from negative patients except for duration of antibiotic treatment and multiple pregnancy (P , 0:001 and P ¼ 0:02; respectively; Table I). In multivariate analysis (logistic regression), multiple pregnancy (OR: 6.3, 95% CI 1.0 –39.4) and duration of antibiotic treatment (OR: 1.6, 95% CI 1.2 –2.2) were the two independent risk factors for acquisition of P. aeruginosa vaginal carriage (P ¼ 0:04 and P ¼ 0:003; respectively). The vaginal carriage of P. aeruginosa had no significant impact on maternal outcome as occurrence of chorioamnionitis or post-partum endometritis were not significantly different between carrier and non carrier-patients. In contrast, among the five P. aeruginosa-positive neonates born from colonized mothers, three very low birthweight infants (, 1000 g) died from early onset P. aeruginosa fulminant sepsis. The two others were only colonized.
Table I Pregnancy characteristics and outcome according to vaginal Pseudomonas aeruginosa acquisition P. aeruginosa in vaginal cultures Variablesa
Positive patients ðN ¼ 11Þ mean(SD )
Negative patients ðN ¼ 52Þ mean(SD )
Pb
Maternal age (years) Multiple pregnancy (N) Duration of prophylaxis (days) Gestational age at rupture (weeks) Gestational age at delivery (weeks) Length of latent period (days) Cerclage Caesarean section Chorioamnionitis Post-partum endometritis
32.6 (1.3) 5 8.7 (2.1) 28.4 (2.3) 30.9 (0.9) 18.4 (3.9) 3 5 7 1
32.0 (0.7) 8 4.6 (3.1) 26.9 (3.9) 29.5 (0.5) 18.7 (2.9) 8 21 17 7
0.7 0.02 ,0.001 0.2 0.3 0.9 0.3 0.5 0.06 0.6
a b
Chi2 test, Fisher’s exact test, and Mann–Whitney U test were used when appropriate. P ,0.05 was considered significant.
160
Figure 1 Representative PFGE patterns of P. aeruginosa isolates after Spe I digestion. Lanes 1–3: environmental strains; lanes 4– 12: maternal strains; lanes 13 –16: two mother – baby pairs; lane 17 molecular weight standards.
Environmental investigation revealed the presence of P. aeruginosa in three faucets. All the maternal and the environmental P. aeruginosa strains were found to be unrelated. However patterns of each mother –baby pairs were indistinguishable (Figure 1).
Discussion To date, P. aeruginosa, naturally resistant to many b-lactam antibiotics, such as co-amoxiclav, is recognized as an infrequent cause of intrapartum or intrauterine infection.4,5 However, it seems that antibiotic prophylaxis in pPROM may select resistant bacteria and cause neonatal sepsis.2 Although the risk of nosocomial infection with resistant bacteria is recognized, no study until now has evaluated the level or precise nature of this risk. Our findings show the high rate of nosocomially acquired P. aeruginosa vaginal carriage in mothers with pPROM receiving co-amoxiclav, and its relationship with the duration of antibiotic treatment and multiple pregnancy. Molecular typing excluded mother-to-mother cross-transmission or a source of contamination
A. Casetta et al.
mediated by the environment and revealed vertical transmission of P. aeruginosa in all positive neonates. Interestingly, the polyclonal origin of all vaginal and environmental strains and the long course of treatment as a risk factor, suggested that P. aeruginosa vaginal carriage was acquired from an endogenous route mediated by antibiotic selective pressure. This vaginal carriage had no significant impact on maternal morbidity or progress of pregnancy, but might threaten the neonate, particularly in very low birthweight infants (, 1000 g) in whom P. aeruginosa is recognized as a highly virulent nosocomial pathogen.4,5 Recently, the ORACLE study7 has shown an increased incidence of necrotizing enterocolitis after antenatal use of co-amoxiclav. Our study adds further concerns about possible adverse effects of a prolonged course of broad-spectrum b-lactam prophylaxis for pPROM, particularly in the preterm population. Indeed, selection pressure for resistant strains might be one of the mechanisms responsible for the occurrence of neonatal complications. The benefit/risk ratio of systematic antibiotics for pPROM should be discussed in each specific case. Further studies are needed to evaluate the impact of shorter courses with narrow-spectrum antibiotics.
References 1. ACOG Practice Bulletin, Premature rupture of membranes: clinical management guidelines for obstetrician—gynecologists. Int J Gynecol Obstet 1998;63:75—84. 2. Edwards RK, Locksmith GJ, Duff P. Expanded-spectrum antibiotics with preterm premature rupture of membranes. Obstet Gynecol 2000;96:60—64. 3. Stoll BJ, Hansen N, Fanaroff AA, et al. Changes in pathogens causing early-onset sepsis in very-low-birth weight infants. N Engl J Med 2002;347:240—247. 4. Leigh L, Stoll BJ, Rahman M, McGowan Jr. J. Pseudomonas aeruginosa infection in very low birth weight infants: a casecontrol study. Pediatr Infect Dis J 1995;14:367—371. 5. Kyle P, Turner DPJ. Chorioamnionitis due to Pseudomonas aerginosa: a complication of prolonged antibiotic therapy for premature rupture of membranes. Br J Obstet Gynaecol 1996; 103:181—183. 6. Luk WK. An outbreak of pseudobacteraemia caused by Burkholderia pickettii: the critical role of an epidemiological link. J Hosp Infect 1996;34:59—69. 7. Kenyon SI, Taylor DJ, Tarnow-Mordi W, ORACLE Collaborative Group. Broad-spectrum antibiotics for preterm, prelabour rupture of fetal membranes: the ORACLE I randomised trial. Lancet 2001;357:979—988.