- Email: [email protected]
Conservative management of clostridial endometritis
American Journal of Obstetrics and Gynecology (2004) 191, 266e70
www.elsevier.com/locate/ajog
Conservative management of clostridial endometritis E. Steve Lichtenberg, MD, MPH, Chynel Henning, MD Department of Obstetrics and Gynecology, Northwestern University Medical School, Chicago, Ill; National Cyto-Path Laboratories, Long Beach, Calif Received for publication August 20, 2003; revised November 1, 2003; accepted November 18, 2003
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– KEY WORDS Clostridium endometritis Abortion Dilation and evacuation Infection Cluster
Objective: Clostridial infection during pregnancy may rapidly progress to life-threatening sepsis. This danger could prompt clinicians to consider hysterectomy as a treatment option as soon as clostridial infection becomes highly suspect or is confirmed, irrespective of its clinical severity. We present evidence that conservative management with the use of intravenous antibiotics is a reasonable initial treatment choice in women undergoing induced abortion who show no sign of sepsis. Study design: We describe the conservative treatment and epidemiologic investigation of unsuspected, culture proven clostridial infections in 5 women undergoing dilation and evacuation abortion during a 22-month period and review the existing literature. Results: Prompt administration of broad-spectrum parenteral antibiotics successfully treated 5 healthy, stable women with culture proven uterine or blood stream clostridial infections. Conclusion: Intravenous broad-spectrum antibiotics with close surveillance is a reasonable initial treatment choice after atraumatic uterine evacuation for women with known or suspected clostridial infection that manifests no sign of sepsis. Ó 2004 Elsevier Inc. All rights reserved.
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Clostridial infection during pregnancy may rapidly progress to life-threatening sepsis.1 This danger could prompt clinicians to consider hysterectomy as a treatment option as soon as clostridial infection becomes highly suspect or is confirmed, irrespective of its clinical severity. We present evidence that conservative management with the use of intravenous antibiotics is a reasonable initial treatment choice in women undergoing induced abortion who show no sign of sepsis. The following is a report of 5 cases of culture-proven clostridial infections, including 3 with bacteremia, that occurred in association with late second trimester dilation and evacuation (D&E) abortion (R20 weeks’ gesta-
Reprints not available from the authors. 0002-9378/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.ajog.2003.11.022
tion) during a 22-month period at a single multisite institutional provider. These women were treated successfully with parenteral antibiotics and supportive care. We also describe the epidemiologic investigation undertaken to find a common cause for the 4 clostridial infections that occurred within a 4-month interval at 1 treatment site.
Cases The 5 index cases profiled in the Table took place between May 2000 and March 2002 and describe the complete incidence of clostridial infections of this multisite provider. All 5 women were healthy, ranging from 15 to 41 years. Preoperative hematocrits were normal
267
Lichtenberg and Henning (range 35%-40%) and no patient had a history of risk factors for pelvic infection or unusual vaginal discharge (not shown). In accordance with institutional protocols, they were not screened for sexually transmitted diseases, but were instead prescribed oral doxycycline 100 mg twice daily beginning at the time of the first laminaria insertion. The vagina and laminaria were cleansed with povidone-iodine solution before laminaria insertions performed with sterile instruments. Cervical preparation included laminaria treatments for 1 or 2 days, with the number of tents and treatments determined according to cervical responsiveness (Table). The 4 women operated on at site B (cases 2-5) also underwent fetocidal injection of intra-amniotic digoxin 1 mg on the first day of treatment, and in all cases the demise of the fetus was documented ultrasonographically within 30 to 60 minutes after injection (Table). Inducing fetal death at least 12 hours before surgery has the purposes of avoiding unscheduled delivery of a live fetus and easing instrumental extraction by causing softening of fetal cortical bone.2 Of the 5 women, 3 experienced premature rupture of membranes (Table). Two experienced surgeons performed the D&Es. Overall, operating times were from 3 to 6 minutes, and blood loss was estimated between 100 to 175 mL (not shown). Bolus infusion of rapid-acting analgesic and sedative agents comprised the general anesthetic, administered by licensed airway managers. No halogenated gases were used. Clostridial infection was not clinically suspected; treatment in all cases was predicated on the appearance of a fever. In case 1, the fever did not appear until 29 hours after the D&E. The other 4 patients, cases 2 to 5, became febrile on the day of surgery. They were otherwise asymptomatic and responded immediately to broad-spectrum intravenous antibiotics administered per protocol for occurrence of fever 37.8(C (100(F) or higher and were therefore not hospitalized overnight (Table). These patients were not hospitalized because they fulfilled the institutional protocol for same-day release mandating 3 consecutive hourly oral temperatures 37.2(C (99.0(F) or less after surgery. Clostridial organisms were not identified or suspected in any case until anaerobic cultures were returned at a minimum of 4 days after treatment was begun (cervical gram stains were not performed). All 5 patients were discharged with at least a 5-day course of 2 oral antibiotic agents. The 4 women treated as same-day patients at site B were required to return to the index facility within 24 to 72 hours for postabortion examination. In case 1 (treated at site A), hospital physicians obtained aerobic and anaerobic blood cultures at the time of admission for a high fever and uterine tenderness but no rebound or adnexal tenderness. In cases 2 to 5, the treating surgeon took aerobic and anaerobic cultures of the lower uterus or blood before ordering intravenous antibiotics (Table I). The site of culture was determined
according to the following protocol started at site B in 1999: A temperature between 37.8(C and 38.4(C (100101(F) mandates a culture of the lower uterus; whereas above 38.4(C (101(F), blood cultures are required. Lower uterine cultures are obtained after vaginal cleansing with povidone-iodine solution using a specially designed kit with long, thin swabs to avoid vaginal contamination. In aggregate, during the period between November 8, 2001, and March 5, 2002, when the 4 index clostridial cases occurred at site B, 23 uterine and 14 blood cultures were obtained for reason of fever. These 37 cultures were taken from 34 D&E patients (3 patients had cultures from both uterus and blood) yielding a clostridia species positivity rate of 1 in 12 (8%) of febrile patients. During the same 4-month interval, about 2650 D&Es were performed at site B, more than 75% of them at 19 weeks or more from last menses, yielding an overall clostridial infection incidence of 1:662 D&Es (w0.15%).
Epidemiologic investigation After identification and review of the first 2 clostridial cases, the Infection Control Committee (ICC) at site B sent a notice to the medical staff exhorting renewed attention to proper aseptic technique. When the latter 2 cases were identified, the ICC launched an intensive investigation. The committee observed the antiseptic technique of the 2 advanced level practitioners who performed laminaria treatments and digoxin injections, as well as the practices of the sole physician who performed all index D&Es at hospital site B. They identified no breach of sterile technique. Instrument sterilization procedures at a clinic referral site where some laminaria treatments were performed and at the proprietary hospital (site B) were subjected to random inspections and found to be satisfactory on each occasion. At this point, the laminaria themselves became the focus of intense scrutiny. The hospital’s Director of Nursing reported that in November 2001 several laminaria packages from hospital stock bearing lot number 0301 were found to have had broken seals, thereby rendering them nonsterile. Additional unsealed packages from lot number 0301 were then discovered at the hospital. No tents from lot number 0301 were located at referral sites. The committee submitted 10 unused tents with broken seals from lot number 0301 for aerobic (environmental) and anaerobic testing to the same bacteriology laboratory that had identified clostridial species in cases 2 to 5. Anaerobic cultures from the 10 laminaria showed no growth after 7 days. One of the 10 aerobic cultures grew a bacillus species. The laboratory log recorded no increase in the incidence of clostridial isolates from other sources during several months bounding the study period, ruling out laboratory contamination as a possible source of infection. The search for a common strain and for toxin production
268 Table
Lichtenberg and Henning Gestational age, surgical preparation, clinical signs, treatment, and culture results of 5 women with clostridial infection
Case no.
Weeks’ gestation by menstrual age
No. of laminaria treatments
No. of laminaria inserted
Digoxin 1 mg injected intraamniotically
Membrane rupture
Fever first noted
1
20.0
2
5, 15
No
No
29 h after the D&E
2 3 4
23.0 20.5 20.5
2 1 2
5, 18 8 5, 20
Yes Yes Yes
Preop on DOS Preop on DOS Day 2
5 h before the D&E 90 min before the D&E 1 min after the D&E
5
22.0
2
4, 16
Yes
No
1 min after the D&E
Preop, Preoperatively; DOS, day of surgery; Tmax before antibiotics, maximum oral temperature before intravenous antibiotics; WBC, white blood cell count; NA, not ascertained. * Patient refused overnight hospitalization.
was not feasible because the infections occurred from 3 to 9 weeks apart and were therefore not identified as a possible common-source outbreak until after the first 3 samples had been discarded. Also, antigens of Clostridium perfringens are extremely variable, making identification of strains highly challenging.3 The Norwegian manufacturer reported that each tent is sealed by hand at 200(F and recorded no customer complaints in the past 7 years. The US company responsible for sterilization uses a vacuum-primed ethylene oxide gas system at high humidity. They documented that all 8 cartons containing devices from lot number 0301 had been sterilized in the standard way. However, they noted that 2 of these cartons had been opened by the US Customs Service and postulated that inspectors may have unsealed some of the tents. Neither the Norwegian manufacturer nor the US sterilizing company is aware of prior clostridial cases associated with their devices. In conclusion, the committee failed to find a common source of infection. No clostridial infections have occurred before or since the study period.
Comment We conducted a MEDLINE inquiry of English-language citations for the years 1965 through 2002 using the search phrases ‘‘clostridium and abortion,’’ ‘‘gas gangrene and abortion,’’ and ‘‘clostridium and septic abortion.’’ This is the first report to describe a cluster of uterine and blood stream clostridial infections from a single-organization provider of legally induced abortion. Clostridia are rod-shaped, gram-positive anaerobic bacteria found in up to 10% of vaginal and cervical cultures in asymptomatic women.4 C perfringens (formerly Welchii) and C sordelli are well-known human pathogens, but in a series of 540 C perfringens isolates, only
about 5% actually caused a clinical illness indicative of exotoxin formation.5 Toxin-producing strains elaborate enzymes that can variably lyse smooth muscle, erythrocytes, renal cells, or hepatocytes and can cause fatal sepsis. As a result, hallmarks of sepsis include marked uterine myonecrosis, hemolytic anemia, jaundice, hemoglobinuria, and cardiorespiratory collapse. In clinical practice, serious infections have been reported in connection with spontaneous and induced abortion, amniocentesis, and delivery. The spectrum of severity with clostridial organisms is wide, ranging from asymptomatic vaginal and cervical culture-positive women who require no special treatment to those with clostridial sepsis, exhibiting deterioration of vital signs and major organs, who may succumb despite customary treatment with hysterectomy and intensive supportive care.1 What is less clear is the optimum treatment for otherwise stable women having positive uterine or blood cultures (bacteremia), but as yet, no signs or symptoms suggesting sepsis. Despite the presence of clostridial organisms in the uterus (cases 2 and 3) or blood stream (cases 1, 4, and 5), none of these 5 women had clostridial sepsis develop. Several factors may account for these fortunate outcomes. First, all women in this report received parenteral antibiotics effective against clostridial species within 1 to 2 hours after becoming febrile. Although C perfringens can elaborate more than a dozen exotoxins, their release requires at least 24 to 48 hours6 even in the presence of necrotic tissue and compromised local circulation, conditions that have been reported with spontaneous,1 and more often, with illegal abortions.6-8 Second, in contrast to the illegal abortion events that have typified prior case reports, the operations in this report were performed sterilely and atraumatically in otherwise healthy women. Third, the laboratories did not report colony counts, and only 2 of the 4 cases at
269
Lichtenberg and Henning
ection Tmax before antibiotics (Celsius)
Highest WBC (mm)
Intravenous antibiotic regimen
40.0
19,000
37.8 38.8 39.3
NA NA 14,700
Ampicillin C gentamicin C clindamycin Cefoxitin Ampicillin/sulbactam Ampicillin/sulbactam, then gentamicin
39.9
12,500
Ampicillin/sulbactam, then gentamicin
site B had clostridial organisms present in both culture bottles. Colony counts may have been low, thereby increasing the chance of cure. Fourth, oral antibiotics were routinely administered preventively before fevers appeared. Specifically, each patient received doxycycline twice a day beginning at the time of the first laminaria placement and continuing for 7 days after surgery with instructions to contact the facility in the event of fever. Pathogenic clostridia are susceptible to tetracyclines in vitro,3 and in 1 series, in which disk sensitivities were carried out on 19 clostridial blood stream isolates, 18 of 19 (95%) were sensitive.7 Finally, multiple effective agents were used. In each case, the parenteral agents randomly chosen belonged to a class of drugs known to be effective against clostridial species.3,9 Thus, each woman was treated with at least 2, and in all but 1 case, at least 3 effective antibiotics beginning at the time of the first spike (Table). This broad, redundant coverage may account for the mild elevations of the white blood cell count observed in 2 of 3 patients at the time parenteral antibiotics were started (Table). The level and progression of leukocytosis is a key prognosticator of patient response.8 Early, broad, and effective antibiotic coverage was probably instrumental in limiting the spread of these infections. However, in the past, some clostridial patients with bacteremia have recovered without antibiotic treatment. In Ramsey’s large series8 of isolates after septic abortion and reported in 1949, 28 febrile women had blood cultures positive for C perfringens. Of these, 13 received no antibiotics with 10 survivors; 11 received only parenteral sulfonamides and 10 survived; and 4 who were treated with penicillin plus sulfonamides all survived. These survival rates are all the more noteworthy because sulfonamides are not recognized as effective
Hospital stay
Culture source
Pathogens identified
8d
Blood
Clostridium perfringens
8h 4h 9 h*
Lower uterus Lower uterus Blood
9h
Blood
C perfringens; Enterococcus C perfringens; Escherichia coli Clostridium species, not perfringens; hemolytic Streptococcus C perfringens; group B betahemolytic Streptococcus; Enterococcus
therapy for clostridia.3 Though remarkably only 4 of these 28 bacteremic women in Ramsey’s series died, they all had notable signs of clinical infection, namely, endometritis (13/28), pelvic peritonitis (6/28), or frank sepsis (9/28), and all experienced protracted convalescences. In contrast, among our 5 cases, only case 1 with uterine (but no rebound) tenderness had a clinical sign other than fever or required an extended hospital stay. Periabortal antibiotic prophylaxis has been shown to reduce by 42% the incidence of postabortion infection in women undergoing induced abortion up to 15 weeks’ gestation.10 Current practice among North American abortion providers reflects this evidence. A recent survey11 found that 91% of responding centers use routine prophylaxis for first trimester abortion, and of these, 81% favor doxycycline. During second trimester abortion by induction method, Spence et al12 found a statistically significant reduction in postabortion endometritis among 198 women randomly assigned to cephalothin versus placebo. Because clostridial organisms commonly inhabit the vaginal flora of healthy women, and ascending infections, though uncommon, can be life-threatening, there is good reason to extend first trimester prophylaxis into the second trimester for women undergoing induced and therapeutic abortion. Finally, although ascending pelvic infection is infrequent among healthy women undergoing induced first and second trimester abortion under safe circumstances,2 rapid treatment of fever spikes with broad-spectrum parenteral antibiotics may prevent the development of serious infections regardless of the specific pathogen(s) at their source. Conservative management with close surveillance is a reasonable consideration as initial therapy for women with known or suspected clostridial infection that manifests no signs of sepsis.
270
References 1. Barrett JP, Whiteside JL, Boardman LA. Fatal clostridial sepsis after spontaneous abortion. Obstet Gynecol 2002;99:899-901. 2. Lichtenberg ES, Grimes DA, Paul M. Abortion complications: prevention and management. In: Paul M, Lichtenberg ES, Borgatta L, Grimes DA, Stubblefield PG, editors. A clinician’s guide to medical and surgical abortion. New York: Churchill Livingstone; 1999. p. 197-216. 3. Bittner J. The clinical significance, taxonomy and special methodological problems of the pathogenic clostridia. Infection 1980;8:117-22. 4. Dylewski J, Wiesenfeld H, Latour A. Postpartum uterine infection with Clostridium perfringens. Rev Infect Dis 1989;11:470-3. 5. Butler HM. Bacteriological studies of Clostridium perfringens infections in man: with special reference to the use of direct smears for rapid diagnosis. Surg Gynecol Obstet 1945;81: 475-86.
Lichtenberg and Henning 6. O’Neill RT, Schwarz RH. Clostridial organisms in septic abortions: report of 7 cases. Obstet Gynecol 1970;35:458-61. 7. Rathbun HK. Clostridial bacteremia without hemolysis. Arch Intern Med 1968;122:496-501. 8. Ramsey AM. The significance of Clostridium welchii in the cervical swab and blood-stream in postpartum and postabortum sepsis. J Obstet Gynaecol Br Commonw 1949;56:247-58. 9. Ledger WJ, Hackett KA. Significance of clostridia in the female reproductive tract. Obstet Gynecol 1973;41:525-30. 10. Sawaya GF, Grady D, Kerlikowske K, Grimes DA. Antibiotics at the time of induced abortion: the case for universal prophylaxis based on a meta-analysis. Obstet Gynecol 1996;87:884-90. 11. Lichtenberg ES, Paul M, Jones H. First trimester surgical abortion practices: a survey of National Abortion Federation members. Contraception 2001;64:345-52. 12. Spence MR, King TM, Burkman RT, Atienza MF. Cephalothin prophylaxis for midtrimester abortion. Obstet Gynecol 1982;60: 502-5.
www.elsevier.com/locate/ajog
Conservative management of clostridial endometritis E. Steve Lichtenberg, MD, MPH, Chynel Henning, MD Department of Obstetrics and Gynecology, Northwestern University Medical School, Chicago, Ill; National Cyto-Path Laboratories, Long Beach, Calif Received for publication August 20, 2003; revised November 1, 2003; accepted November 18, 2003
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– KEY WORDS Clostridium endometritis Abortion Dilation and evacuation Infection Cluster
Objective: Clostridial infection during pregnancy may rapidly progress to life-threatening sepsis. This danger could prompt clinicians to consider hysterectomy as a treatment option as soon as clostridial infection becomes highly suspect or is confirmed, irrespective of its clinical severity. We present evidence that conservative management with the use of intravenous antibiotics is a reasonable initial treatment choice in women undergoing induced abortion who show no sign of sepsis. Study design: We describe the conservative treatment and epidemiologic investigation of unsuspected, culture proven clostridial infections in 5 women undergoing dilation and evacuation abortion during a 22-month period and review the existing literature. Results: Prompt administration of broad-spectrum parenteral antibiotics successfully treated 5 healthy, stable women with culture proven uterine or blood stream clostridial infections. Conclusion: Intravenous broad-spectrum antibiotics with close surveillance is a reasonable initial treatment choice after atraumatic uterine evacuation for women with known or suspected clostridial infection that manifests no sign of sepsis. Ó 2004 Elsevier Inc. All rights reserved.
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Clostridial infection during pregnancy may rapidly progress to life-threatening sepsis.1 This danger could prompt clinicians to consider hysterectomy as a treatment option as soon as clostridial infection becomes highly suspect or is confirmed, irrespective of its clinical severity. We present evidence that conservative management with the use of intravenous antibiotics is a reasonable initial treatment choice in women undergoing induced abortion who show no sign of sepsis. The following is a report of 5 cases of culture-proven clostridial infections, including 3 with bacteremia, that occurred in association with late second trimester dilation and evacuation (D&E) abortion (R20 weeks’ gesta-
Reprints not available from the authors. 0002-9378/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.ajog.2003.11.022
tion) during a 22-month period at a single multisite institutional provider. These women were treated successfully with parenteral antibiotics and supportive care. We also describe the epidemiologic investigation undertaken to find a common cause for the 4 clostridial infections that occurred within a 4-month interval at 1 treatment site.
Cases The 5 index cases profiled in the Table took place between May 2000 and March 2002 and describe the complete incidence of clostridial infections of this multisite provider. All 5 women were healthy, ranging from 15 to 41 years. Preoperative hematocrits were normal
267
Lichtenberg and Henning (range 35%-40%) and no patient had a history of risk factors for pelvic infection or unusual vaginal discharge (not shown). In accordance with institutional protocols, they were not screened for sexually transmitted diseases, but were instead prescribed oral doxycycline 100 mg twice daily beginning at the time of the first laminaria insertion. The vagina and laminaria were cleansed with povidone-iodine solution before laminaria insertions performed with sterile instruments. Cervical preparation included laminaria treatments for 1 or 2 days, with the number of tents and treatments determined according to cervical responsiveness (Table). The 4 women operated on at site B (cases 2-5) also underwent fetocidal injection of intra-amniotic digoxin 1 mg on the first day of treatment, and in all cases the demise of the fetus was documented ultrasonographically within 30 to 60 minutes after injection (Table). Inducing fetal death at least 12 hours before surgery has the purposes of avoiding unscheduled delivery of a live fetus and easing instrumental extraction by causing softening of fetal cortical bone.2 Of the 5 women, 3 experienced premature rupture of membranes (Table). Two experienced surgeons performed the D&Es. Overall, operating times were from 3 to 6 minutes, and blood loss was estimated between 100 to 175 mL (not shown). Bolus infusion of rapid-acting analgesic and sedative agents comprised the general anesthetic, administered by licensed airway managers. No halogenated gases were used. Clostridial infection was not clinically suspected; treatment in all cases was predicated on the appearance of a fever. In case 1, the fever did not appear until 29 hours after the D&E. The other 4 patients, cases 2 to 5, became febrile on the day of surgery. They were otherwise asymptomatic and responded immediately to broad-spectrum intravenous antibiotics administered per protocol for occurrence of fever 37.8(C (100(F) or higher and were therefore not hospitalized overnight (Table). These patients were not hospitalized because they fulfilled the institutional protocol for same-day release mandating 3 consecutive hourly oral temperatures 37.2(C (99.0(F) or less after surgery. Clostridial organisms were not identified or suspected in any case until anaerobic cultures were returned at a minimum of 4 days after treatment was begun (cervical gram stains were not performed). All 5 patients were discharged with at least a 5-day course of 2 oral antibiotic agents. The 4 women treated as same-day patients at site B were required to return to the index facility within 24 to 72 hours for postabortion examination. In case 1 (treated at site A), hospital physicians obtained aerobic and anaerobic blood cultures at the time of admission for a high fever and uterine tenderness but no rebound or adnexal tenderness. In cases 2 to 5, the treating surgeon took aerobic and anaerobic cultures of the lower uterus or blood before ordering intravenous antibiotics (Table I). The site of culture was determined
according to the following protocol started at site B in 1999: A temperature between 37.8(C and 38.4(C (100101(F) mandates a culture of the lower uterus; whereas above 38.4(C (101(F), blood cultures are required. Lower uterine cultures are obtained after vaginal cleansing with povidone-iodine solution using a specially designed kit with long, thin swabs to avoid vaginal contamination. In aggregate, during the period between November 8, 2001, and March 5, 2002, when the 4 index clostridial cases occurred at site B, 23 uterine and 14 blood cultures were obtained for reason of fever. These 37 cultures were taken from 34 D&E patients (3 patients had cultures from both uterus and blood) yielding a clostridia species positivity rate of 1 in 12 (8%) of febrile patients. During the same 4-month interval, about 2650 D&Es were performed at site B, more than 75% of them at 19 weeks or more from last menses, yielding an overall clostridial infection incidence of 1:662 D&Es (w0.15%).
Epidemiologic investigation After identification and review of the first 2 clostridial cases, the Infection Control Committee (ICC) at site B sent a notice to the medical staff exhorting renewed attention to proper aseptic technique. When the latter 2 cases were identified, the ICC launched an intensive investigation. The committee observed the antiseptic technique of the 2 advanced level practitioners who performed laminaria treatments and digoxin injections, as well as the practices of the sole physician who performed all index D&Es at hospital site B. They identified no breach of sterile technique. Instrument sterilization procedures at a clinic referral site where some laminaria treatments were performed and at the proprietary hospital (site B) were subjected to random inspections and found to be satisfactory on each occasion. At this point, the laminaria themselves became the focus of intense scrutiny. The hospital’s Director of Nursing reported that in November 2001 several laminaria packages from hospital stock bearing lot number 0301 were found to have had broken seals, thereby rendering them nonsterile. Additional unsealed packages from lot number 0301 were then discovered at the hospital. No tents from lot number 0301 were located at referral sites. The committee submitted 10 unused tents with broken seals from lot number 0301 for aerobic (environmental) and anaerobic testing to the same bacteriology laboratory that had identified clostridial species in cases 2 to 5. Anaerobic cultures from the 10 laminaria showed no growth after 7 days. One of the 10 aerobic cultures grew a bacillus species. The laboratory log recorded no increase in the incidence of clostridial isolates from other sources during several months bounding the study period, ruling out laboratory contamination as a possible source of infection. The search for a common strain and for toxin production
268 Table
Lichtenberg and Henning Gestational age, surgical preparation, clinical signs, treatment, and culture results of 5 women with clostridial infection
Case no.
Weeks’ gestation by menstrual age
No. of laminaria treatments
No. of laminaria inserted
Digoxin 1 mg injected intraamniotically
Membrane rupture
Fever first noted
1
20.0
2
5, 15
No
No
29 h after the D&E
2 3 4
23.0 20.5 20.5
2 1 2
5, 18 8 5, 20
Yes Yes Yes
Preop on DOS Preop on DOS Day 2
5 h before the D&E 90 min before the D&E 1 min after the D&E
5
22.0
2
4, 16
Yes
No
1 min after the D&E
Preop, Preoperatively; DOS, day of surgery; Tmax before antibiotics, maximum oral temperature before intravenous antibiotics; WBC, white blood cell count; NA, not ascertained. * Patient refused overnight hospitalization.
was not feasible because the infections occurred from 3 to 9 weeks apart and were therefore not identified as a possible common-source outbreak until after the first 3 samples had been discarded. Also, antigens of Clostridium perfringens are extremely variable, making identification of strains highly challenging.3 The Norwegian manufacturer reported that each tent is sealed by hand at 200(F and recorded no customer complaints in the past 7 years. The US company responsible for sterilization uses a vacuum-primed ethylene oxide gas system at high humidity. They documented that all 8 cartons containing devices from lot number 0301 had been sterilized in the standard way. However, they noted that 2 of these cartons had been opened by the US Customs Service and postulated that inspectors may have unsealed some of the tents. Neither the Norwegian manufacturer nor the US sterilizing company is aware of prior clostridial cases associated with their devices. In conclusion, the committee failed to find a common source of infection. No clostridial infections have occurred before or since the study period.
Comment We conducted a MEDLINE inquiry of English-language citations for the years 1965 through 2002 using the search phrases ‘‘clostridium and abortion,’’ ‘‘gas gangrene and abortion,’’ and ‘‘clostridium and septic abortion.’’ This is the first report to describe a cluster of uterine and blood stream clostridial infections from a single-organization provider of legally induced abortion. Clostridia are rod-shaped, gram-positive anaerobic bacteria found in up to 10% of vaginal and cervical cultures in asymptomatic women.4 C perfringens (formerly Welchii) and C sordelli are well-known human pathogens, but in a series of 540 C perfringens isolates, only
about 5% actually caused a clinical illness indicative of exotoxin formation.5 Toxin-producing strains elaborate enzymes that can variably lyse smooth muscle, erythrocytes, renal cells, or hepatocytes and can cause fatal sepsis. As a result, hallmarks of sepsis include marked uterine myonecrosis, hemolytic anemia, jaundice, hemoglobinuria, and cardiorespiratory collapse. In clinical practice, serious infections have been reported in connection with spontaneous and induced abortion, amniocentesis, and delivery. The spectrum of severity with clostridial organisms is wide, ranging from asymptomatic vaginal and cervical culture-positive women who require no special treatment to those with clostridial sepsis, exhibiting deterioration of vital signs and major organs, who may succumb despite customary treatment with hysterectomy and intensive supportive care.1 What is less clear is the optimum treatment for otherwise stable women having positive uterine or blood cultures (bacteremia), but as yet, no signs or symptoms suggesting sepsis. Despite the presence of clostridial organisms in the uterus (cases 2 and 3) or blood stream (cases 1, 4, and 5), none of these 5 women had clostridial sepsis develop. Several factors may account for these fortunate outcomes. First, all women in this report received parenteral antibiotics effective against clostridial species within 1 to 2 hours after becoming febrile. Although C perfringens can elaborate more than a dozen exotoxins, their release requires at least 24 to 48 hours6 even in the presence of necrotic tissue and compromised local circulation, conditions that have been reported with spontaneous,1 and more often, with illegal abortions.6-8 Second, in contrast to the illegal abortion events that have typified prior case reports, the operations in this report were performed sterilely and atraumatically in otherwise healthy women. Third, the laboratories did not report colony counts, and only 2 of the 4 cases at
269
Lichtenberg and Henning
ection Tmax before antibiotics (Celsius)
Highest WBC (mm)
Intravenous antibiotic regimen
40.0
19,000
37.8 38.8 39.3
NA NA 14,700
Ampicillin C gentamicin C clindamycin Cefoxitin Ampicillin/sulbactam Ampicillin/sulbactam, then gentamicin
39.9
12,500
Ampicillin/sulbactam, then gentamicin
site B had clostridial organisms present in both culture bottles. Colony counts may have been low, thereby increasing the chance of cure. Fourth, oral antibiotics were routinely administered preventively before fevers appeared. Specifically, each patient received doxycycline twice a day beginning at the time of the first laminaria placement and continuing for 7 days after surgery with instructions to contact the facility in the event of fever. Pathogenic clostridia are susceptible to tetracyclines in vitro,3 and in 1 series, in which disk sensitivities were carried out on 19 clostridial blood stream isolates, 18 of 19 (95%) were sensitive.7 Finally, multiple effective agents were used. In each case, the parenteral agents randomly chosen belonged to a class of drugs known to be effective against clostridial species.3,9 Thus, each woman was treated with at least 2, and in all but 1 case, at least 3 effective antibiotics beginning at the time of the first spike (Table). This broad, redundant coverage may account for the mild elevations of the white blood cell count observed in 2 of 3 patients at the time parenteral antibiotics were started (Table). The level and progression of leukocytosis is a key prognosticator of patient response.8 Early, broad, and effective antibiotic coverage was probably instrumental in limiting the spread of these infections. However, in the past, some clostridial patients with bacteremia have recovered without antibiotic treatment. In Ramsey’s large series8 of isolates after septic abortion and reported in 1949, 28 febrile women had blood cultures positive for C perfringens. Of these, 13 received no antibiotics with 10 survivors; 11 received only parenteral sulfonamides and 10 survived; and 4 who were treated with penicillin plus sulfonamides all survived. These survival rates are all the more noteworthy because sulfonamides are not recognized as effective
Hospital stay
Culture source
Pathogens identified
8d
Blood
Clostridium perfringens
8h 4h 9 h*
Lower uterus Lower uterus Blood
9h
Blood
C perfringens; Enterococcus C perfringens; Escherichia coli Clostridium species, not perfringens; hemolytic Streptococcus C perfringens; group B betahemolytic Streptococcus; Enterococcus
therapy for clostridia.3 Though remarkably only 4 of these 28 bacteremic women in Ramsey’s series died, they all had notable signs of clinical infection, namely, endometritis (13/28), pelvic peritonitis (6/28), or frank sepsis (9/28), and all experienced protracted convalescences. In contrast, among our 5 cases, only case 1 with uterine (but no rebound) tenderness had a clinical sign other than fever or required an extended hospital stay. Periabortal antibiotic prophylaxis has been shown to reduce by 42% the incidence of postabortion infection in women undergoing induced abortion up to 15 weeks’ gestation.10 Current practice among North American abortion providers reflects this evidence. A recent survey11 found that 91% of responding centers use routine prophylaxis for first trimester abortion, and of these, 81% favor doxycycline. During second trimester abortion by induction method, Spence et al12 found a statistically significant reduction in postabortion endometritis among 198 women randomly assigned to cephalothin versus placebo. Because clostridial organisms commonly inhabit the vaginal flora of healthy women, and ascending infections, though uncommon, can be life-threatening, there is good reason to extend first trimester prophylaxis into the second trimester for women undergoing induced and therapeutic abortion. Finally, although ascending pelvic infection is infrequent among healthy women undergoing induced first and second trimester abortion under safe circumstances,2 rapid treatment of fever spikes with broad-spectrum parenteral antibiotics may prevent the development of serious infections regardless of the specific pathogen(s) at their source. Conservative management with close surveillance is a reasonable consideration as initial therapy for women with known or suspected clostridial infection that manifests no signs of sepsis.
270
References 1. Barrett JP, Whiteside JL, Boardman LA. Fatal clostridial sepsis after spontaneous abortion. Obstet Gynecol 2002;99:899-901. 2. Lichtenberg ES, Grimes DA, Paul M. Abortion complications: prevention and management. In: Paul M, Lichtenberg ES, Borgatta L, Grimes DA, Stubblefield PG, editors. A clinician’s guide to medical and surgical abortion. New York: Churchill Livingstone; 1999. p. 197-216. 3. Bittner J. The clinical significance, taxonomy and special methodological problems of the pathogenic clostridia. Infection 1980;8:117-22. 4. Dylewski J, Wiesenfeld H, Latour A. Postpartum uterine infection with Clostridium perfringens. Rev Infect Dis 1989;11:470-3. 5. Butler HM. Bacteriological studies of Clostridium perfringens infections in man: with special reference to the use of direct smears for rapid diagnosis. Surg Gynecol Obstet 1945;81: 475-86.
Lichtenberg and Henning 6. O’Neill RT, Schwarz RH. Clostridial organisms in septic abortions: report of 7 cases. Obstet Gynecol 1970;35:458-61. 7. Rathbun HK. Clostridial bacteremia without hemolysis. Arch Intern Med 1968;122:496-501. 8. Ramsey AM. The significance of Clostridium welchii in the cervical swab and blood-stream in postpartum and postabortum sepsis. J Obstet Gynaecol Br Commonw 1949;56:247-58. 9. Ledger WJ, Hackett KA. Significance of clostridia in the female reproductive tract. Obstet Gynecol 1973;41:525-30. 10. Sawaya GF, Grady D, Kerlikowske K, Grimes DA. Antibiotics at the time of induced abortion: the case for universal prophylaxis based on a meta-analysis. Obstet Gynecol 1996;87:884-90. 11. Lichtenberg ES, Paul M, Jones H. First trimester surgical abortion practices: a survey of National Abortion Federation members. Contraception 2001;64:345-52. 12. Spence MR, King TM, Burkman RT, Atienza MF. Cephalothin prophylaxis for midtrimester abortion. Obstet Gynecol 1982;60: 502-5.