Antibiotic treatment of tuboovarian abscess: Comparison of broad-spectrum ~-lactam agents versus clindamycin-containing regimens Susan D. Reed, MD, Daniel V. Landers, MD, and Richard L. Sweet, MD San Francisco, California One hundred nineteen patients with tuboovarian abscess were evaluated for response to antibiotics. Results were stratified into three groups by antimicrobial regimen. Group 1 consisted of 37 patients treated with a single-agent broad-spectrum intravenous antibiotic and oral doxycycline. Initial clinical response (defined as decreased pain, diminished white blood cell count, or defervescence) in group 1 was 31/37 (84%). Group 2 consisted of 64 patients treated with c1indamycin in combination with an aminoglycoside with or without a penicillin. There was an initial clinical response in 45 of 64 (70%). Group 3 consisted of 18 patients from group 1 who were changed to a clindamycin-containing regimen after 2 to 3 days of initial treatment with a single-agent broad-spectrum antibiotic. The decision to switch antibiotics was not based on treatment failure but occurred when delayed ultrasonography confirmed the diagnosis of tuboovarian abscess. The switch reflected physician preference for clindamycin-containing regimens in the treatment of tuboovarian abscesses. The response rate in this subset of patients was 14 of 18 (78%). Overall initial clinical response rate was 90 of 119 (75%). There were no statistically significant demographic or clinical differences among the three groups. There was no statistical difference in the rate of early and late antibiotic failure rates among the groups. Our study demonstrates that extended-spectrum antibiotic coverage, including single-agent broad-spectrum antibiotics such as cefoxitin, in conjunction with doxycycline has efficacy that is equivalent to that of clindamycin-containing regimens. An overall medical treatment success rate of 75% suggests that conservative treatment of tuboovarian abscesses is warranted. (AM J OSSTET GYNECOL 1991 ;164:1556-62.)
Key words: Tuboovarian abscess, pelvic abscess, pelvic inflammatory disease, tuboovarian complex, antimicrobial therapy Tuboovarian abscesses have long been recognized as a severe form of pelvic inflammatory disease that typically results from inadequate or delayed treatment of a pelvic infection. Tuboovarian abscesses are reported in 3% to 16% of hospitalized patients with acute pelvic inflammatory disease. I Management of tuboovarian abscesses has evolved markedly from the purely surgical approach of the preantibiotic era, which included extraperitoneal drainage with an associated high mortality from ruptured tuboovarian abscesses. When antibiotics were introduced, aggressive intraabdominal surgical approaches were used to treat ruptured tuboovarian abscesses; also antibiotics were used in the treatment of unruptured tuboovarian abscesses to "cool a From the Department of Obstetrzcs, Gynecology and Reproductive Sciences, Unwerslty of California, San FranCIsco, and San FrancIsco General HospItal. Supported in part b_~ NatIOnal InstItutes of Health grant 1P01 AI
24768.
Presented by invItation at the Fifty-seventh Annual Meeting of the Pacific Coast Obstetncal and Gynecologzcal SOClet.~, Sun Valley, Idaho, September 9-14. 1990 Repnnt requests: RIchard L. Sweet, MD, Department of Obstetrics and Gynecology, Room 6D14, San FranCIsco General HospItal, 1001 Potrero Ave., San FrancISco, CA 9-1110.
6/6/27945
1556
patient off" in preparation for delayed surgery! We are now evolving toward a more conservative approach, in which potent and broad-spectrum antibiotics are used as definitive first-line therapy and in which surgical intervention occurs only in tuboovarian abscesses that do not respond to initial antibiotic therapy. Research on the medical treatment of tuboovarian abscess indicates that proper antimicrobial treatment of pelvic abscesses should include antibiotics that penetrate abscess cavities and remain stable in the abscess environment. 3 In particular, antimicrobial agents that provide adequate coverage against the resistant gramnegative anaerobes of Bacteroides group should be used. Antibiotic regimens containing clindamycin or metronidazole have generally been considered the best available agents for the treatment of tuboovarian abscesses! 5 In 1983 we reported data on 321 patients with tuboovarian abscesses, only 10 of whom had been treated with one of the now commonly used extendedspectrum second- and third-generation cephalosporins. I The patients in that study were treated with either a clindamycin-containing regimen or a regimen with inadequate anaerobic coverage. The overall initial clinical response rate with antibiotic treatment alone was
Treatment of tuboovarian abscess
Volume 164 ]\; umber 6, Part 1
70%, and the response rate for those treated with clindamycin (89%) was significantly better than that of patients who received non-clindamycin-containing regimens (35%), The current study compares antimicrobial regimens containing a single-agent broad-spectrum antibiotic that is effective against the resistant gramnegative anaerobes with clindamycin-containing regimens, The predominant regimens used were cefoxitin plus doxycycline and clindamycin plus gentamicin, Methods
One hundred twenty-five patients were discharged from San Francisco General Hospital between April 1980 and January 1989 with a diagnosis of tuboovarian abscess (palpable adnexal or pelvic mass confirmed by ultrasonography or computed tomography), Six patients underwent surgical treatment of tuboovarian abscess within 48 hours of admission without an antibiotic trial and thus were excluded, The charts of the remaining 119 patients were retrospectively reviewed, Patients had been enrolled previously in treatment trials of acute pelvic inflammatory disease that were approved by the University of California, San Francisco, Committee on Human Research, In all 119 patients either an initial clinical diagnosis of tuboovarian abscess was confirmed by ultrasonography with the exception of two patients who had computed tomography scan confirmation or ultrasonography after admission demonstrated a tuboovarian abscess in instances in which one had not been clinically appreciated, The ultrasonographic diagnosis of tuboovarian abscess required demonstration of a welldefined complex adnexal mass ~4 cm in size with no apparently normal ipsilateral ovary, In the patients who required surgical intervention, all clinically and ultrasonographically diagnosed tuboovarian abscesses were confirmed at surgery, All masses were at least 4 cm in diameter and were complex in nature by imaging techniques, Cultures were obtained in the majority of patients, Endocervical, endometrial, and abscess specimens were evaluated for Neisseria gonorrhoeae and Chlamydia trachomatis, In addition, the endometrium and abscess were sampled for anaerobic and aerobic bacteria, Cultures of the cervix for N, gonorrhoeae were obtained with a Dacron swab, inoculated directly onto modified Thayer-Martin agar plates, and held at 35 0 C in 5% carbon dioxide, Cervical specimens for C. trachomatis were collected with a sterile calcium alginatetipped swab that was placed immediately in minimal essential media containing antibiotics and 10% fetal calf serum and refrigerated at 4 C until transport to the laboratory for isolation, Most specimens were processed within a few hours of acquisition, and all were processed within 24 hours, 0
1557
A double-lumen, plastic telescoping catheter containing a wire brush and sealed with a gelatin plug was used for collection of the endometrial specimens, Endometrial brushes were divided into two parts under sterile conditions, One was processed for C. trachomatls, and the other for isolation of N, gonorrhoeae and other aerobic and anaerobic bacteria, All specimens for C. trachomatis were inoculated onto monolayers of cycloheximide-treated McCoy cells by centrifugation, The cells were cultured for 65 hours, then examined for inclusions by either iodine or immunofluorescent staining, One blind passage was performed on all negative specimens, Specimens were processed for isolation of facultative organisms by inoculation onto the following media, which were incubated with 5% carbon dioxide at 35 C: trypticase soy agar with 6% sheep blood, chocolate agar, calcium nutrient agar with 6% human blood, modified Thayer-Martin agar, and calcium nutrient agar with 6% sheep blood, Specimens processed for anaerobic growth were inoculated into Brucella agar supplemented with 6% sheep blood, vitamin K, and hemin, calcium nutrient agar with 6% sheep blood, kanamycin-vancomycin laked blood agar, and chopped meat carbohydrate broth incubated at 35 C in either a Gas-Pakjar (BBL, Cockeysville, Md,) or an anaerobic chamber (Coy Manufacturing, Ann Arbor), N. gonorrhoeae and facultative bacteria were identified by standard methods, Anaerobic organisms were identified with gas-liquid chromatography and prereduced, anaerobically sterilized media, Patients were placed in three groups according to antimicrobial treatment regimen (Table I), Group 1 consisted of 37 patients treated with a single-agent extended-spectrum antibiotic combined with doxycycline in most instances, Cefoxitin plus doxycycline (n = 27) was the most common regimen, Patients in group 2 (n = 64) were treated with clindamycin, usually in conjunction with an aminoglycoside with or without a penicillin, Patients in group 3 (n = 18) were initially treated with a group 1 regimen for 2 to 3 days and then were switched to a group 2 regimen, The decision to change antimicrobial agents was not due to treatment failure but rather was based on physician bias toward clindamycin-containing regimens when delayed ultrasonography revealed the presence of a tuboovarian abscess, The groups were compared with respect to clinical presentation, initial response to antibiotics, and treatment failure, Initial response was defined as defervescence, decreased pain, decrease in mass size, and discharge from hospital without need for surgical intervention, Early treatment failure was defined as lack of an initial clinical response plus institution of surgical intervention or a switch in antimicrobial therapy or both, 0
0
1558
Reed, Landers, and Sweet
June 1991 Am J Obstet Gynecol
Table I. Antibiotic regimens used for treatment of tuboovarian abscesses Group 1 (n = 37)
Group 2 (n = 64)
Clindamycin (n = 37) Clindamycin (n = 24) Clindamycin (n = I) Clindamycin (n = I) Clindamycin
Cefoxitin and doxycycline (n = 27) Cefotetan and doxycycline (n Ampicillin and sulbactam (n
= =
5) 2)
SCH 34343 and doxycycline (n = I) Ciproftoxacillin (n = 2)
and tobramycin and gentamicin and kanamycin and aztreonam
Group 3 (n = 18)
Cefoxitin and doxycycline: clindamycin and aminoglycoside (n = 15) Cefotetan and doxycycline; clindamycin and aminoglycoslde (n = 2) Ampicillin and sulbactam; clindamycin and aminoglycoside (n = I)
and ciproftoxacillin
(n = I)
Table II. Clinical variables in patients with tuboovarian abscesses Group 1 = 37)
(n
Mean age (yr) Parity History of PID Mean temperature (C) Mean WBC (X 106 / ml) Mean ESR (mm/hr) Mean mass size (cm) Days of intravenous antibiotics
27.3 ± 7.4 l.l ± 1.3 20 (55.0%) 38.3° ± 0.66° 14.6 ± 4.2 63.5 ± 27.1 7.2 ± 2.2 7.8 ± 2.2
Group 2 = 64)
(n = 18)
Group 4 (N = 119)
28.0 ± 8.3 0.85 ± l.l 28 (43.7%) 38.2° ± 0.6° 15.4 ± 11.7 65.0 ± 28.9 6.7 ± 2.5 7.2 ± 2.3
30.3 ± 7.2 0.94 ± l.l 7 (41.1 X) 38.6° ± 0.4° 16.3 ± 5.0 66.5 ± 30.4 7.4 ± 2.0 9.4 ± 2.6
27.7 ± 7.9 0.98 ± 1.2 55 (46.0%) 38.3° ± 0.65° 15.0 ± 9.0 64.3 ± 28.0 6.9 ± 2.4 7.5 ± 2.3
(n
Group 3
P1D. Pelvic inftammatory disease; WBC, white blood cell count; ESR, erythrocvte sedimentation rate.
Follow-up after discharge from hospital consisted of return visits at 2 weeks, with some follow-ups continuing > 1 year after discharge. Clinical response at follow-up was defined as diminished or absent pelvic mass by pelvic examination or ultrasonography and diminished or absent pain. Late treatment failure was defined as a persistent mass or pain at follow-up. Seventy-five patients returned for follow-up. Descriptive statistical analysis was used to compare the groups. Discrete variables were analyzed standard X" test. Continuous variables were analyzed with the unpaired t test. A p value of
Key words: Tuboovarian abscess, pelvic abscess, pelvic inflammatory disease, tuboovarian complex, antimicrobial therapy Tuboovarian abscesses have long been recognized as a severe form of pelvic inflammatory disease that typically results from inadequate or delayed treatment of a pelvic infection. Tuboovarian abscesses are reported in 3% to 16% of hospitalized patients with acute pelvic inflammatory disease. I Management of tuboovarian abscesses has evolved markedly from the purely surgical approach of the preantibiotic era, which included extraperitoneal drainage with an associated high mortality from ruptured tuboovarian abscesses. When antibiotics were introduced, aggressive intraabdominal surgical approaches were used to treat ruptured tuboovarian abscesses; also antibiotics were used in the treatment of unruptured tuboovarian abscesses to "cool a From the Department of Obstetrzcs, Gynecology and Reproductive Sciences, Unwerslty of California, San FranCIsco, and San FrancIsco General HospItal. Supported in part b_~ NatIOnal InstItutes of Health grant 1P01 AI
24768.
Presented by invItation at the Fifty-seventh Annual Meeting of the Pacific Coast Obstetncal and Gynecologzcal SOClet.~, Sun Valley, Idaho, September 9-14. 1990 Repnnt requests: RIchard L. Sweet, MD, Department of Obstetrics and Gynecology, Room 6D14, San FranCIsco General HospItal, 1001 Potrero Ave., San FrancISco, CA 9-1110.
6/6/27945
1556
patient off" in preparation for delayed surgery! We are now evolving toward a more conservative approach, in which potent and broad-spectrum antibiotics are used as definitive first-line therapy and in which surgical intervention occurs only in tuboovarian abscesses that do not respond to initial antibiotic therapy. Research on the medical treatment of tuboovarian abscess indicates that proper antimicrobial treatment of pelvic abscesses should include antibiotics that penetrate abscess cavities and remain stable in the abscess environment. 3 In particular, antimicrobial agents that provide adequate coverage against the resistant gramnegative anaerobes of Bacteroides group should be used. Antibiotic regimens containing clindamycin or metronidazole have generally been considered the best available agents for the treatment of tuboovarian abscesses! 5 In 1983 we reported data on 321 patients with tuboovarian abscesses, only 10 of whom had been treated with one of the now commonly used extendedspectrum second- and third-generation cephalosporins. I The patients in that study were treated with either a clindamycin-containing regimen or a regimen with inadequate anaerobic coverage. The overall initial clinical response rate with antibiotic treatment alone was
Treatment of tuboovarian abscess
Volume 164 ]\; umber 6, Part 1
70%, and the response rate for those treated with clindamycin (89%) was significantly better than that of patients who received non-clindamycin-containing regimens (35%), The current study compares antimicrobial regimens containing a single-agent broad-spectrum antibiotic that is effective against the resistant gramnegative anaerobes with clindamycin-containing regimens, The predominant regimens used were cefoxitin plus doxycycline and clindamycin plus gentamicin, Methods
One hundred twenty-five patients were discharged from San Francisco General Hospital between April 1980 and January 1989 with a diagnosis of tuboovarian abscess (palpable adnexal or pelvic mass confirmed by ultrasonography or computed tomography), Six patients underwent surgical treatment of tuboovarian abscess within 48 hours of admission without an antibiotic trial and thus were excluded, The charts of the remaining 119 patients were retrospectively reviewed, Patients had been enrolled previously in treatment trials of acute pelvic inflammatory disease that were approved by the University of California, San Francisco, Committee on Human Research, In all 119 patients either an initial clinical diagnosis of tuboovarian abscess was confirmed by ultrasonography with the exception of two patients who had computed tomography scan confirmation or ultrasonography after admission demonstrated a tuboovarian abscess in instances in which one had not been clinically appreciated, The ultrasonographic diagnosis of tuboovarian abscess required demonstration of a welldefined complex adnexal mass ~4 cm in size with no apparently normal ipsilateral ovary, In the patients who required surgical intervention, all clinically and ultrasonographically diagnosed tuboovarian abscesses were confirmed at surgery, All masses were at least 4 cm in diameter and were complex in nature by imaging techniques, Cultures were obtained in the majority of patients, Endocervical, endometrial, and abscess specimens were evaluated for Neisseria gonorrhoeae and Chlamydia trachomatis, In addition, the endometrium and abscess were sampled for anaerobic and aerobic bacteria, Cultures of the cervix for N, gonorrhoeae were obtained with a Dacron swab, inoculated directly onto modified Thayer-Martin agar plates, and held at 35 0 C in 5% carbon dioxide, Cervical specimens for C. trachomatis were collected with a sterile calcium alginatetipped swab that was placed immediately in minimal essential media containing antibiotics and 10% fetal calf serum and refrigerated at 4 C until transport to the laboratory for isolation, Most specimens were processed within a few hours of acquisition, and all were processed within 24 hours, 0
1557
A double-lumen, plastic telescoping catheter containing a wire brush and sealed with a gelatin plug was used for collection of the endometrial specimens, Endometrial brushes were divided into two parts under sterile conditions, One was processed for C. trachomatls, and the other for isolation of N, gonorrhoeae and other aerobic and anaerobic bacteria, All specimens for C. trachomatis were inoculated onto monolayers of cycloheximide-treated McCoy cells by centrifugation, The cells were cultured for 65 hours, then examined for inclusions by either iodine or immunofluorescent staining, One blind passage was performed on all negative specimens, Specimens were processed for isolation of facultative organisms by inoculation onto the following media, which were incubated with 5% carbon dioxide at 35 C: trypticase soy agar with 6% sheep blood, chocolate agar, calcium nutrient agar with 6% human blood, modified Thayer-Martin agar, and calcium nutrient agar with 6% sheep blood, Specimens processed for anaerobic growth were inoculated into Brucella agar supplemented with 6% sheep blood, vitamin K, and hemin, calcium nutrient agar with 6% sheep blood, kanamycin-vancomycin laked blood agar, and chopped meat carbohydrate broth incubated at 35 C in either a Gas-Pakjar (BBL, Cockeysville, Md,) or an anaerobic chamber (Coy Manufacturing, Ann Arbor), N. gonorrhoeae and facultative bacteria were identified by standard methods, Anaerobic organisms were identified with gas-liquid chromatography and prereduced, anaerobically sterilized media, Patients were placed in three groups according to antimicrobial treatment regimen (Table I), Group 1 consisted of 37 patients treated with a single-agent extended-spectrum antibiotic combined with doxycycline in most instances, Cefoxitin plus doxycycline (n = 27) was the most common regimen, Patients in group 2 (n = 64) were treated with clindamycin, usually in conjunction with an aminoglycoside with or without a penicillin, Patients in group 3 (n = 18) were initially treated with a group 1 regimen for 2 to 3 days and then were switched to a group 2 regimen, The decision to change antimicrobial agents was not due to treatment failure but rather was based on physician bias toward clindamycin-containing regimens when delayed ultrasonography revealed the presence of a tuboovarian abscess, The groups were compared with respect to clinical presentation, initial response to antibiotics, and treatment failure, Initial response was defined as defervescence, decreased pain, decrease in mass size, and discharge from hospital without need for surgical intervention, Early treatment failure was defined as lack of an initial clinical response plus institution of surgical intervention or a switch in antimicrobial therapy or both, 0
0
1558
Reed, Landers, and Sweet
June 1991 Am J Obstet Gynecol
Table I. Antibiotic regimens used for treatment of tuboovarian abscesses Group 1 (n = 37)
Group 2 (n = 64)
Clindamycin (n = 37) Clindamycin (n = 24) Clindamycin (n = I) Clindamycin (n = I) Clindamycin
Cefoxitin and doxycycline (n = 27) Cefotetan and doxycycline (n Ampicillin and sulbactam (n
= =
5) 2)
SCH 34343 and doxycycline (n = I) Ciproftoxacillin (n = 2)
and tobramycin and gentamicin and kanamycin and aztreonam
Group 3 (n = 18)
Cefoxitin and doxycycline: clindamycin and aminoglycoside (n = 15) Cefotetan and doxycycline; clindamycin and aminoglycoslde (n = 2) Ampicillin and sulbactam; clindamycin and aminoglycoside (n = I)
and ciproftoxacillin
(n = I)
Table II. Clinical variables in patients with tuboovarian abscesses Group 1 = 37)
(n
Mean age (yr) Parity History of PID Mean temperature (C) Mean WBC (X 106 / ml) Mean ESR (mm/hr) Mean mass size (cm) Days of intravenous antibiotics
27.3 ± 7.4 l.l ± 1.3 20 (55.0%) 38.3° ± 0.66° 14.6 ± 4.2 63.5 ± 27.1 7.2 ± 2.2 7.8 ± 2.2
Group 2 = 64)
(n = 18)
Group 4 (N = 119)
28.0 ± 8.3 0.85 ± l.l 28 (43.7%) 38.2° ± 0.6° 15.4 ± 11.7 65.0 ± 28.9 6.7 ± 2.5 7.2 ± 2.3
30.3 ± 7.2 0.94 ± l.l 7 (41.1 X) 38.6° ± 0.4° 16.3 ± 5.0 66.5 ± 30.4 7.4 ± 2.0 9.4 ± 2.6
27.7 ± 7.9 0.98 ± 1.2 55 (46.0%) 38.3° ± 0.65° 15.0 ± 9.0 64.3 ± 28.0 6.9 ± 2.4 7.5 ± 2.3
(n
Group 3
P1D. Pelvic inftammatory disease; WBC, white blood cell count; ESR, erythrocvte sedimentation rate.
Follow-up after discharge from hospital consisted of return visits at 2 weeks, with some follow-ups continuing > 1 year after discharge. Clinical response at follow-up was defined as diminished or absent pelvic mass by pelvic examination or ultrasonography and diminished or absent pain. Late treatment failure was defined as a persistent mass or pain at follow-up. Seventy-five patients returned for follow-up. Descriptive statistical analysis was used to compare the groups. Discrete variables were analyzed standard X" test. Continuous variables were analyzed with the unpaired t test. A p value of
