REVIEWS OF INFECTIOUS DISEASES • VOL. 5, NO.5.

SEPTEMBER-OCTOBER 1983
© 1983 by The University of Chicago. All rights reserved. 0162-0886/83/0505-0005$02.00

Tubo-ovarian Abscess: Contemporary Approach to Management

Daniel V. Landers and Richard L. Sweet From the Department of Obstetrics and Gynecology,
San Francisco General Hospital, San Francisco, California

Two hundred and thirty-two patients with tubo-ovarian abscesses (TOAs) were evalu-
ated. Ruptured TOAs were documented in seven (3070) of the patients. One hundred
and seventy-five patients with TOAs were treated with antibiotics alone; for 15 of these
patients, TOAs were confirmed by laparoscopy. The remaining 57 patients required
surgical intervention: drainage (five patients), unilateral salpingo-oophorectomy (19)
and total abdominal hysterectomy and bilateral salpingo-oophorectomy (33). A uni-
lateral TOA was present in 163 patients (70%). Seventy-six patients with TOAs used
intrauterine contraceptive devices, and in this group, 54 (71%) patients had unilateral
TOAs. The most common microorganisms that were recovered from these TOAs were

Downloaded from http://cid.oxfordjournals.org/ at Michigan State University on July 4, 2015

Escherichia coli, Bacteroides fragilis, Bacteroides species, Peptostreptococcus,
Peptococcus, and aerobic streptococci. Sixty-eight percent of the patients treated with
an antimicrobial regimen that included clindamycin had a decrease in the size of the
TOA, while only 36.5% of those receiving antimicrobial regimens without clindamycin
had a decrease in the size of the TOA (P < .01). Long-term follow-up information (two
to 10 years) was available for 58 of the patients treated with antibiotics alone. Eighteen
(31%) required subsequent surgery; 12 had persistent TOAs; and six, chronic salpingo-
oophoritis. Intrauterine pregnancy was documented in eight (13.8%) patients. Of the
19 patients treated with unilateral adnexectomy, two ultimately required hysterectomy
and contralateral adnexectomy, while three patients in this group subsequently became
pregnant (one ectopic and two intrauterine).

Approximately 1 million women are treated for
acute salpingitis (pelvic inflammatory disease) in
the United States each year at an estimated annual
cost of >600 million dollars. Some 250,000 of
these women are treated as inpatients [1]. Tubo-
ovarian abscess (TOA), a known sequela of
salpingitis, has been reported to occur in 3%-16%
of patients hospitalized with salpingitis (2, 3]. It is
generally accepted that rupture of a TOA is an
indication for immediate surgical intervention;
however, there still exists considerable controversy
about the management of the unruptured TOA.
The recommendations range from prompt surgical
intervention with complete removal of the uterus
and adnexa [2, 4] to treatment with iv antibiotics
with surgery reserved for those patients who fail
to respond or in whom rupture is suspected [5, 6].
There does seem to be a general trend in recent
years toward a more conservative approach to
both the timing and extent of surgical intervention
for the unruptured TOA [5, 6]. This retrospective
study was undertaken to evaluate the conservative
medical approach to management, to determine
the usefulness of sonography in the diagnosis and
management of TOAs, to assess the microbiology
of the TOA, to evaluate the therapeutic response
of various antibiotic regimens, and, finally, to lay
the groundwork for the future prospective studies
necessary to elucidate this problem.

Received for publication January 19, 1983.

Informed consent was obtained from the patients, and the
guidelines for human experimentation of the u.s. Department
of Health and Human Services and those of the University of
California, San Francisco, California, were followed in the
conduct of the clinical research.
This work was supported in part by grant no. AI 11136from
the National Institute of Allergy and Infectious Diseases.
Please address requests for reprints to Dr. Richard L. Sweet,
Department of Obstetrics and Gynecology, Room 6J2, San
Francisco General Hospital, 1001 Potrero Avenue, San
Francisco, California 94110.
Materials and Methods
Two hundred and thirty-two patients with a
diagnosis of TOA were admitted between January
1970 and April 1980 to San Francisco General
Hospital. The diagnosis of TOA was based either
on the confirmation at the time of surgical inter-
vention (with histologic evidence of the presence
of an abscess involving the fallopian tube and
ovary in cases where the abscess was surgically

876
Tuba-ovarian Abscess 877

charged with a prescription for oral antibiotics-

usually ampicillin or tetracycline (in those treated
with penicillin-containing regimens) or clindamy-
cin (in those receiving parenteral clindamycin). In
addition to antibiotic therapy, patients were
treated with hydration, bed rest in semi-Fowler's
position with nasogastric suction, and blood re-
placement as needed. Patients were discharged
with recommendation for continued pelvic rest for
six weeks to three months after release from the
hospital.

Downloaded from http://cid.oxfordjournals.org/ at Michigan State University on July 4, 2015

Table 1. Treatment regimens most frequently used for
patients with tubo-ovarian abscesses during the study
period of January 1970 to April 1980 at San Francisco
General Hospital.
Figure 1. Sonogram of a tubo-ovarian abscess. Note
the presence of a complex mass with cystic and solid Treatment regimen,
components and a well-defined border. antimicrobial agent
(route of administration) Doses

excised) or on the presence of an adnexal mass(es) Penicillin alone

Aqueous penicillin G (iv) 30-60 million units
demonstrated by pelvic examination of patients
per day in 4
with symptoms and signs consistent with a divided doses
diagnosis of acute pelvic inflammatory disease. Penicillin plus an aminoglycoside
Confirmation of clinical findings was obtained in Aqueous penicillin G (iv) 20-30 million units
the majority of patients with the use of ultrasono- per day in 4
divided doses
graphy, laparoscopy, or laparotomy. The pres-
plus
ence of a complex adnexal mass on ultrasound was Tobramycin (iv), gentamicin (iv), 3-5 mg/kg per day
considered sonographically consistent with the in 3 divided
diagnosis of TOA (figure 1). Patients were ex- doses
cluded from the study if they were postmeno- or
Kanamycin (im) 1 g per day in 2
pausal, had recent abdominal or pelvic surgery,
divided doses
had a negative laparoscopy for presumed TOA, or Clindamycin plus an aminoglycoside
were within six weeks postpartum. Clindamycin (iv) 2.4 g per day in 4
The treatment regimens varied considerably divided doses
during the 10-year period. In the earlier years plus
Tobramycin (iv) or gentamicin (iv) 3-5 mg/kg per day
patients were treated with high doses of penicillin
in 3 divided
(n = 35) or with penicillin plus an aminoglycoside doses
(n = 48). In later years antibiotic regimens that in- Triple antibiotic therapy
cluded clindamycin usually were employed (n = Aqueous penicillin G (iv) 20 million units per
78) - either clindamycin plus an aminoglycoside or day in 4 divided
doses
triple therapy with clindamycin, an aminoglyco-
plus
side, and penicillin. Patients allergic to penicillin Tobramycin (iv), gentamicin (iv), See above
were treated with similar regimens except that or
cephalothin was substituted for penicillin. Ten Kanamycin (im) See above
patients were treated with cefoxitin and two pa- plus
Clindamycin (iv) 2.4 g per day in 4
tients, with chloramphenicol plus an aminogly-
divided doses
coside. The exact treatment regimens are outlined Cephalosporin
in table 1. Patients generally were treated for five Cefoxitin (iv) 4-8 g per day in
to 10 days with iv antibiotics and had a minimum 4 divided doses
of 48-72 hr without fever before being switched to Cephalothin (iv) 4-12 g per day in
4-6 divided doses
treatment with oral antibiotics. Patients were dis-
878
Response to therapy was determined on the
basis of improvement in symptoms, absence of
fever, reduction of pelvic tenderness, and decrease
in the size of the TOA. Since all patients who did
not require surgery during their initial hospitaliza-
tion were afebrile with improvement of symptoms
and a decrease in pelvic tenderness, the major
determinant of successful medical management
was the reduction in mass or the resolution of the
TOA. The data were subjected to X2 analysis or
Fisher's exact test.
Results
During the study period, 1,588 patients were ad-
mitted to San Francisco General Hospital with a
diagnosis of salpingitis; of these, 232 (14.6070) had
a diagnosis of TOA. Group I consisted of 175
patients treated with antibiotics alone. TOAs for
15 of these patients were confirmed surgically.
Group II was composed of 57 patients treated with
a combination of antibiotics and surgery. Five
patients underwent surgical drainage; 19, uni-
lateral salpingo-oophorectomy; and 33, total
abdominal hysterectomy and bilateral salpingo-
oophorectomy (TAH-BSO). Exploratory laparo-
tomies were performed on 10 patients suspected to
have a ruptured TOA. Seven of these were con-
firmed at surgery. Thus, the overall incidence of
rupture was 3%.
There was a lower percentage of nulliparous pa-
tients in the surgically confirmed group than in the
clinically diagnosed group. However, this differ-
ence was not statistically significant. A history of
prior gonorrhea, prior pelvic inflammatory dis-
ease, or use of an intrauterine contraceptive device
(IUD) was present for about one-third of the cases
(table 2). There were no significant differences
Table 2. Epidemiologic correlates of tube-ovarian abscesses (TOAs) diagnosed clinically compared with those of
TOAs confirmed surgically.
No. of patients (010) with indicated epidemiologic correlate
Patient category (no. of patients) Nulliparous
TOA clinically diagnosed only (160) 90 (56)
TOA surgically confirmed (72) 30 (42)
TOA not excised (20) 9 (45)
TOA excised (52) 21 (40)
Total (232) 120 (52)
* IUD := intrauterine contraceptive device.
Landers and Sweet
between the surgically diagnosed group and the
clinically diagnosed group in any of these cate-
gories. Among the surgically confirmed group, a
larger percentage of patients (36%) with TOAs
requiring surgical extirpation used IUDs than did
patients with TOAs not requiring surgical treat-
ment (20%). This difference was not statistically
significant. Of the 76 patients with TOAs and a
history of IUD usage, 60 (79%) had the IUD in
situ at the time of admission.
The most frequent presenting symptoms were
acute and/or chronic pain, fever, chills, vaginal
Downloaded from http://cid.oxfordjournals.org/ at Michigan State University on July 4, 2015
discharge, irregular vaginal bleeding, nausea, and
vomiting (table 3). There were no statistically sig-
nificant differences in the incidence of these
symptoms among the medically or surgically
treated groups. The occurrence of fever and leu-
kocytosis on admission was slightly higher in the
surgically treated group, but this was not statis-
tically significant (table 3). Even among the pa-
tients whose TOAs were surgically confirmed,
35% were afebrile and 23070 had a normal WBC
count.
Sonograms were obtained for 98 patients for
confirmation of the presence of their TOA (table
4). Sonography confirmed the clinical finding of
an adnexal mass in 91 patients (93%); 83 patients
(85070) were noted to have complex adnexal
masses suggestive of TOAs; and eight patients
(8%) had a cystic adnexa mass(es). In only seven
instances did sonography not confirm the clinical
finding of an adnexal mass. Thirty-one of the
patients underwent sonography prior to surgical
exploration. In this surgically confirmed group,
sonography demonstrated complex masses in 29
patients (93.6%) and cystic masses in two (6.4%).
There were no false-negative sonograms among
the surgically confirmed cases.
Prior history Prior history
of gonorrhea of salpingitis IUD usage"
48 (30) 52 (32.5) 53 (33)
23 (32) 24 (33) 23 (32)
6 (30) 6 (30) 4 (20)
17 (33) 18 (35) 19 (36)
71 (31) 76 (33) 76 (33)
Tubo-ovarian Abscess 879

Table 3. Presenting symptoms and findings among Escherichia coli from 10 (37070); aerobic strepto-
232 patients with tubo-ovarian abscess. cocci, 5 (18.5070); Bacteroides fragilis, 6 (22%);
No. of patients in indicated Bacteroides species, 7 (26%); Peptococcus, 3
group (0J0) with presenting (11%); and Peptostreptococcus, 5 (18.5070). While
symptom or finding gonococci were recovered from the cervices of
Medically Surgically 31% of patients, this organism was recovered
treated treated from only two of the abscess aspirates.
Symptom/finding (n = 175) (n = 57) Seven of 10 patients undergoing exploratory
Acute pain 158 (90) 48 (84) laparotomy within 72 hr of admission for suspi-
Chronic pain 29 (17) 14 (25) cion of rupture had ruptured TOAs; the remaining
Fever/ chills 86 (49) 31 (53) three patients had unruptured TOAs. Five patients
Vaginal discharge 53 (30) 11 (19)
responded to initial antibiotics and colpotomy

Downloaded from http://cid.oxfordjournals.org/ at Michigan State University on July 4, 2015

Abnormal uterine bleeding 37 (21) 11 (19)
Nausea 44 (25) 17 (30) drainage. Forty-two (19.4%) of the remaining 217
Vomiting 23 (13) 13 (23) patients required surgical extirpation during their
Temperature, >100 F 102 (58) 37 (65) initial hospitalization because they failed to
WBC, >10,OOO/mm3 114 (72) 44 (77) respond to antimicrobial therapy alone.
Twenty-seven (64%) of these patients had been
treated solely with regimens that did not contain
The occurrence of a unilateral TOA was noted clindamycin. Fifteen (36%) had received clinda-
in 164 (71070) of the 232 patients (table 5). There mycin; in seven of these patients with TOAs, ther-
was no statistically significant difference between apy with clindamycin had been instituted after
the incidence of unilateral abscesses in those who these patients failed to respond to an initial com-
did use an IUD and in those who did not or bination regimen of penicilllin and aminogly-
between the clinically diagnosed and the surgically coside. Thus, eight (19070) of the patients requiring
diagnosed groups. However, among the 42 pa- intervention were treated initially with clinda-
tients who failed to respond to antibiotic therapy mycin. In contradistinction, 34 patients (81%) in
and who required surgical intervention, 18 (43070) the group who required surgical intervention re-
had bilateral TOAs, whereas 24 (57070) had uni- ceived initial therapy with regimens that did not
lateral TOAs. Although there was a trend toward contain clindamycin. Whereas only 19% of the
surgical intervention with patients who had bi- patients with TOAs requiring surgical intervention
lateral TOAs, this trend was not statistically sig- had initially been treated with clindamycin, 71
nificant. '(33070) of 217 patients who did not undergo im-
Specimens for culture were obtained directly mediate surgical intervention and 63 (36070) of 165
from the abscess by needle aspiration in 53 pa- of the patients with TOAs who responded to anti-
tients of whom all but three had received pre- microbial therapy alone received clindamycin.
operative antibiotics. Organisms were recovered in These differences are statistically significant (P
51% of these patients. Four (15070) of the cultures < .05).
from which organisms were isolated yielded only
facultative bacteria; 10 (37070), only anaerobes;
and 13 (48%), a mixture of facultative bacteria Table 4. Results of ultrasonography in 98 patients
and anaerobes. Thus, anaerobic bacteria were re- with suspected tubo-ovarian abscesses (TOAs).
covered from 85070 of the abscesses yielding micro- No. of patients (070) with
organisms in culture. No bacteria were recovered indicated finding
in the remaining 49% of patients from whom Complex Cystic
specimens for culture were taken. The majority of Patient category adnexal adnexal No
cultures yielding no growth were obtained prior to (no. of patients) mass mass masses
the establishment of the anaerobic research TOA clinically
laboratory at San Francisco General Hospital. diagnosed (67) 54 (81) 6 (9) 7 (10)
There were 58 isolates from the 27 positive cul- TOA surgically
tures, for an average of 2.15 species per abscess. documented (31) 29 (93.6) 2 (6.4) o (0)
Total (98) 83 (85) 8 (8) 7 (7)
The most common organisms recovered were
880 Landers and Sweet

Table S. The incidence of bilateral and unilateral tubo- group not receiving clindamycin to 4.8070 in the
ovarian abscesses (TOAs) according to patient's use of group receiving clindamycin. Five patients re-
an intrauterine contraceptive device (IUD), clinical sponded to cefoxitin alone with a reduction in
diagnosis, surgical confirmation of TOA, and patient's
response to therapy. mass size. Five patients treated with cefoxitin
remained febrile with no reduction in mass size,
No. of patients (%) with
three responded to a change to a clindamycin-
indicated TOAs
aminoglycoside regimen, and two patients who
Patient category Bilateral Unilateral failed to respond to medical therapy required
(no. of patients) TOAs TOAs
surgical extirpation of their unruptured TOAs.
Use of IUD (76) 22 (29) 54 (71) One hundred and thirty-four (77070) of the pa-
No use of IUD (156) 46 (29.5) 110 (70.5) tients treated with antibiotics alone returned for
Clinically diagnosed TOA (160) 41 (26) 119 (74)
follow-up examination two to four weeks after

Downloaded from http://cid.oxfordjournals.org/ at Michigan State University on July 4, 2015

Surgically confirmed TOA (72) 27 (37.5) 45 (62.5)
Failure to respond to antibiotics
discharge (table 7). In 61070 of the patients, the
alone; required surgical removal mass had decreased in size or disappeared, while
of TOA (42) 18 (43) 24 (57) in 39070 the mass was unchanged or had enlarged.
All patients with TOAs (232) 68 (29) 164 (71) A decrease in mass size was noted in 86070 of pa-
tients treated with clindamycin as compared with
46.4070 of patients treated with regimens that did
Among patients treated only with regimens that not include clindamycin. This difference was
did not contain clindamycin, 27 (19.4070) of 139 re- statistically significant (P < .01). An opposite
quired surgical intervention. Only eight (11070) of pattern was seen in the patients who did not
71 patients with TOAs initially treated with respond to therapy.
regimens including clindamycin required surgical Long-term follow-up data were available for 58
treatment. This difference is statistically significant of the patients treated with antibiotics alone. This
(P < .(05). Of the 146 patients whose initial anti- follow-up period ranged from two to 10 years
microbial regimen did not include clindamycin, 34 after discharge. Subsequent intrauterine pregancy
(23070) underwent surgical intervention. This was documented in eight (13.8070) of 58 patients.
finding demonstrated a trend toward a better A subsequent laparotomy for pelvic pain and/or
response with clindamycin but was not statistically the persistence of the mass was required in 18
significant (P = .06). Overall, 15 (18070) of 78 pa- (31070) of 58 patients. Twelve patients were again
tients receiving clindamycin at some time during
their antimicrobial treatment underwent surgical
intervention. Table 6. Clinical response at time of discharge from
All patients treated with antibiotics alone were San Francisco General Hospital of patients with tubo-
afebrile with marked decrease in pain prior to ovarian abscesses (TOAs) treated with antimicrobial
discharge. Of the 175 patients treated with anti- agents only.
biotics alone, 167 were reexamined prior to dis- No. of patients (070)
charge for evaluation of the size of the TOA. The with indicated clinical
response
results are summarized in table 6. Eight patients
who were treated with antibiotics alone were im- Mass Mass
proved on discharge from the hospital, but they Antibiotic regimen of TOA of TOA
(no. of patients) decreased increased
were not reexamined for evaluation of TOA size.
A reduction in mass size was noted in nearly one- Antimicrobial regimens without
half of the patients. Patients receiving a treatment clindamycin (104) 38 (36.5)* 11 (10.6)
Penicillin only (32) 8 (25) 5 (15.6)
regimen that included clindamycin were more
Penicillin plus an amino-
likely to have a decrease in mass size (68.3070) than glycoside (47) 23 (49) 3 (6.4)
were patients treated with regimens not including Other (25) 7 (28) 3 (12)
clindamycin (36.5070). This difference was statis- Antimicrobial regimens with
tically significant (P< .01). On the other hand, an clindamycin (63) 43 (68.3)* 3 (4.8)
Total (167) 81 (48.5) 14 (8.4)
increase in mass size was noted in 8.5070 of
patients, an incidence ranging from 10.6070 in the * p < .01.
Tubo-ovarian Abscess 881

Table 7. Clinical response two to four weeks after during their initial hospitalization was long-term
discharge from San Francisco General Hospital of pa- follow-up data available. The lack of gynecologic
tients with tubo-ovarian abscesses (TOAs) treated with problems reported posthysterectomy may reflect
antimicrobial agents only.
this lack of follow-up in the group of patients who
No. of patients (%) underwent hysterectomies.
with indicated
clinical response

Mass
Discussion
Mass increased
Conservative medical management of the un-
Antibiotic regimen decreased or un-
(no. of patients) or absent changed
ruptured TOA, particularly in young, nulliparous
patients, has been gradually gaining acceptance.
Antimicrobial regimens exclusive of
Although there is general agreement that prompt

Downloaded from http://cid.oxfordjournals.org/ at Michigan State University on July 4, 2015

clindamycin (84) 39 (46.4)* 45 (53.6)
Penicillin only (26) 10 (38.5) 16 (61.5)
Penicillin plus an amino-
glycoside (37) 20 (54)
Other (21) 9 (43)
Antimicrobial regimens including
clindamycin (50) 43 (86)*
Total (134) 82 (61)
*p < .01.
noted to have TOAs (well-encapsulated mass
involving tube and ovary containing fluid or
purulent material) and six patients, chronic
salpingo-oophoritis. Similar long-term follow-up
data were available for all 19 of the patients who
underwent unilateral adnexectomy for their TOA.
Total hysterectomy and contralateral adnexec-
tomy were subsequently required in only two
patients (10.5070). There were three subsequent
pregnancies (15.8070) in this group-one ectopic
pregancy and two intrauterine pregancies. Of the
last two, one patient delivered at term and the
other patient underwent an elective termination of
her pregnancy. No comparable follow-up data
were available for those treated during their initial
hospitalization with hysterectomy and salpingo-
oophorectomy (unilateral or bilateral). For only
eight of the women who underwent hysterectomy
surgical intervention is indicated for suspected
rupture, some investigators have advocated this
17 (46) approach in the treatment of the unruptured TOA
12 (57) as well [2, 4]. Kaplan et al. [4] treated 71 patients
with unruptured TOAs with TAH-BSO after 24-
7 (14)
52 (39) 72 hr of antibiotic therapy. With this aggressive
approach, bowel injury (serosal tears) occured in
8.4070 of the patients. Several investigators since
then have reported favorable results with more
conservative management aimed at the preserva-
tion of the patients' reproductive potential [5, 6]
(table 8). Franklin et al. [5] reported their results
for 120 patients who were treated initially with a
conservative approach. Eighty-five patients were
treated with antibiotics alone and 35 patients, with
antibiotics plus colpotomy drainage. The overall
rate of failure was 26%. In 1980Ginsberg et al. [6]
published the results for 160 patients initially
treated with antibiotics alone; early treatment
failure was noted in 31% and later failure, in
36070. In the present study, there was a 20070 rate
of early failure and a 31% rate of late failure for
patients treated with antibiotics alone.
These three studies indicate that response to
medical therapy is successful in 33070-74% of
patients. The rate of subsequent intrauterine preg-
nancy ranged from 9.5% to 13.8%. Future repro-
ductive capability is a significant concern to most

Table 8. Summation of the results in recent studies advocating conservative management of tubo-ovarian abscesses.
No. (010) of patients with
No. of early treatment failures/ No. of late treatment failures/ long-term follow-up
no. of patients initially treated no. of patients with long-term who subsequently had
Reference with antibiotics (%) follow-up (%) intrauterine pregnancies

Present study 421217 (19.4) 18/58 (31.0) 8 (13.8)

6 50/160 (31.3) 34/95 (35.8) 9 (9.5)
5 12/120 (10) 16/97 (16.5) 10 (10.3)
Total 104/497 (20.9) 681250 (27.2) 27 (10.8)
882
patients with TOAs and plays a major role in the
desire for a more conservative approach to ther-
apy. Although definitive cure can be achieved in
all patients treated with hysterectomy and BSO,
the physician must keep in mind that surgical
morbidity does occur and that the patient loses
reproductive capability and hormone production.
In the present study early failure was defined as
the persistence of fever and pain or the enlarge-
ment of the TOA. Early failure was considered an
indication for surgical intervention. The type and
extent of surgical intervention was decided on the
basis of both the desire for the maintenance of
reproductive potential and/or hormone function
and the size and extent of the abscesses. Drainage
via colpotomy was reserved for patients with
fluctuant TOAs dissecting in the midline of the
upper one-third of the rectovaginal septum. In the
absence of these criteria, transabdominal explora-
tion and surgical extirpation of the abscess were
performed.
Because only 72 (30%) of 232 patients had
surgical confirmation of the TOA diagnosis, we
might possibly have included other inflammatory
masses, such as pyosalpinx or loops of bowel
adhesed to the adnexa, in our analysis. However,
the sonographic confirmation of TOAs was con-
sistent for both the surgically documented and
clinically diagnosed groups. In addition, the pre-
senting symptoms and signs, the epidemiologic
correlates, and the incidence of unilateral ab-
scesses were similar in both groups. Furthermore,
when we limited our analysis to those patients who
underwent laparoscopy (without drainage or ex-
cision) and who had visual confirmation of a
TOA, the results and our conclusions were un-
changed.
Whether unilateral adnexectomy is appropriate
therapy or TAH -BSO is necessary in the surgical
management of unilateral TOAs remains contro-
versial. Some authors recommend the complete
removal of the uterus and adnexa [2, 4]. In 1977,
Rivlin and Hunt [7] reported their results on the
use of conservative surgery with peritoneal lavage
in 113 patients with ruptured TOAs. Hysterec-
tomy was required in only 3% of the cases. In our
study, 19 patients were treated with unilateral
adnexectomy and only two of 19 patients required
subsequent TAH and contralateral adnexectomy.
In addition, three of 19 patients became pregnant.
As Rivlin and Hunt [7] pointed out, this conserva-
Landers and Sweet
tive surgical approach to TOAs offers the advan-
tages of a hope for future fertility, maintenance of
hormonal and menstrual function, and avoidance
of the physiologic and psychologic effects of cas-
tration and hysterectomy. Perhaps patients with a
unilateral TOA, who respond to initial antibiotic
therapy but in whom the mass persists, would
benefit from unilateral adnexectomy in terms of
the prevention of future flare-ups and of the en-
hancement of the prognosis for future fertility.
This issue is particularly important when we
consider the frequency with which unilateral
Downloaded from http://cid.oxfordjournals.org/ at Michigan State University on July 4, 2015
TOAs occur. In our series, rv71% of the TOAs
were unilateral in both those who used IUDs and
those who did not. These data confirm the high in-
cidence of unilateral TOAs, regardless of whether
or not an IUD is used, that has been reported in
prior studies [6, 8] Ginsberg et al. [6], reporting on
160 patients with TOAs (diagnosed either clin-
ically or surgically), noted that 90 (56070) patients
had unilateral TOAs; 46 (61070) of 75 of those
who used an IUD had unilateral TOAs, and 44
(52%) of 85 of those who did not use an IUD had
unilateral TOAs. Golde et al. [8] reported that 37
(44% ) of 85 TOAs diagnosed surgically were uni-
lateral; however, unilateral TOAs occurred in 20
(62.5%) of 32 patients who used IUDs as com-
pared with 17 (32 %) of 53 who did not. In addi-
tion' unilaterality of a TOA seems to be helpful in
predicting which patients are more likely to re-
spond to conservative medical management. In
this series, 43% of patients with TOAs who re-
quired surgery after failing to respond to conser-
vative medical management had bilateral TOAs as
compared with 26% of those who responded to
antibiotics alone. Ginsberg et al. [6] likewise
showed bilaterality of abscesses to be predictive of
failure to respond to medical therapy.
One of the most serious complications associ-
ated with TOAs is intraabdominal rupture, a
surgical emergency for which the mortality rate is
rapidly increased by unnecessary delay. Many
investigators feel that when a TOA ruptures, the
treatment of choice is TAH-BSO [9-12]. How-
ever, our data support those of Rivlin and Hunt
[7] for their large series in which conservative
surgery and peritoneal lavage were advocated for
patients with a ruptured TOA. In our study, there
were seven ruptured TOAs, of which four were
treated with unilateral adnexectomy and copious
irrigation. None of these patients required further
Tubo-ovarian Abscess
surgery, and one carried a subsequent intrauterine
pregnancy to term.
The advent and improvement of grey-scale ul-
trasound has resulted in a useful tool for diagnosis
and management of patients with TOAs. In the
current study a mass was correctly identified by
our sonographers in all patients with surgically
confirmed abscesses, and the masses were sono-
graphically characteristic of a TOA in 85070 of the
patients. Although the frequency of false-negative
results was low, it is disturbing that they all oc-
curred in the medically treated group. This fact
raises a question about the accuracy of the
diagnosis when the presence of a TOA was not
established at surgery. However, the accuracy of
sonography as demonstrated in the surgical group
is reassuring. In addition Taylor et al. [13] re-
ported on their experience with sonography in 220
patients with abdominal and/or pelvic abscesses.
In their series 36 of 40 abdominal abscesses were
correctly identified, and the diagnosis was ex-
cluded in 112 of 113 patients. Among patients
with pelvic abscesses, 32 of 33 abscesses were
correctly diagnosed, and the diagnosis was ex-
cluded in 33 of 34 patients.
During the past two years, the improved resolu-
tion of and knowledge about grey-scale ultra-
sound have made its accuracy far superior to that
in our study. More recent data are being compiled
now, and the results are even more encouraging.
Ultrasound is also useful in assessing the response
to therapy and detecting changes in the size of the
masses. It is not yet clear whether the computer-
ized tomography (CT) scan will improve this ac-
curacy enough to justify the expense. The prospect
of percutaneous drainage of abscesses may also be
enhanced when it is directed by sonography or CT
scan.
TOAs contain predominantly a mixed flora of
anaerobes and facultative or aerobic organisms [2,
8, 12]. Anaerobic organisms are particularly
prevalent in these abscesses, having been isolated
from 63070 to 100% of adnexal abscesses for which
appropriate anaerobic microbiological technology
was used [14, 15]. In our series the predominant
organisms isolated were E. coli, B. fragilis, other
Bacteroides species, Peptococcus, and Peptostrep-
tococcus. These are the same organisms involved
in a biphasic aerobic-anaerobic animal model
associated with abscess formation [16, 17]. Al-
though 85% of our positive cultures yielded
883
anaerobes, 49070 of the cultures of abscess speci-
mens yielded no growth. We feel that "sterile
abscess" is a misnomer, and the lack of success in
recovering organisms was more likely related to
the lack of proper anaerobic collection, transport,
and culture techniques. Unfortunately, during the
earlier years of our study, anaerobic techniques
were not utilized in the transport and culture of
our abscess aspirates. The major role played by
anaerobes in TOAs was initially demonstrated by
Altemeier [14] in the early 1940s when he isolated
anaerobic organisms from 92070 of TOA speci-
Downloaded from http://cid.oxfordjournals.org/ at Michigan State University on July 4, 2015
mens that had previously been reported by the
clinical laboratory as having yielded no growth.
The abscess environment is characterized by a
low level of oxygen tension and, thus, a low redox
potential. It is this low redox potential that allows
the anaerobes to proliferate. This proliferation
leads to tissue destruction and circulatory com-
promise, thus preventing many antibiotics from
reaching the area. The combination of these forces
and the poor phagocytosis by neutrophils in this
environment are all important factors in the re-
sistance of these infections to antimicrobial ther-
apy. The high levels of enzymes produced by
bacteria within the abscess aid in the destruction
of many antibiotics like penicillin, ampicillin,
first-generation cephalosporins, ticarcillin, car-
benicillin, and chloramphenicol.
In our study, the regimens found to be most
successful were those that included clindamycin.
A decrease in mass size was noted in 68.3% of the
patients treated with regimens that include clin-
damycin, as compared with 36.5070 of the patients
treated with regimens that did not include clin-
damycin. Such an improved response in the group
receiving clindamycin probably reflects the ac-
tivity of this agent against gram-negative anaero-
bic rods, especially B. fragilis, which commonly
are associated with abscess formation [15, 18].
A second explanation for this increased re-
sponse in patients treated with clindamycin may
be due not only to clindamycin's extracellular an-
timicrobial activity but also to its recently dis-
covered ability to penetrate the human neutrophil
[19]. It has been shown that some aerobic orga-
nisms survive within neutrophils and are protected
from the lethal factors in serum and the extracel-
lular antibiotic levels. This may be true for
anaerobic organisms as well. Furthermore, in an
animal model, clindamycin entered infected en-
884
capsulated sc abscesses in mice at a higher concen-
tration (43%-630/0 of peak serum levels) than did
other antimicrobial agents, including metronida-
zole, cefoxitin, and moxalactam [20]. However,
these other antimicrobial agents that are also
active against B. fragilis did enter the abscesses.
More studies are needed to assess further the activ-
ity of all these antibiotics known to be active
against resistant gram-negative anaerobes. On the
other hand, Simon et al. [21] have shown that a
significant amount of the clindamycin that entered
an experimentally infected capsule in rabbits was
inactive against B. fragilis. This inactivity may
contribute in part to some of the failures seen with
clindamycin.
In summary, an initial conservative antimi-
crobial approach to the management of the un-
ruptured TOA is appropriate if the antibiotics
used can penetrate abscesses, are stable in ab-
scesses, and are active to some degree in the ab-
scess environment and against anaerobes, includ-
ing the resistant gram-negative anaerobic rods
such as B. fragilis and Bacteroides bivius. How-
ever, if the patient does not begin to show a re-
sponse within a reasonable amount of time, i.e.,
48-72 hr, surgical intervention should be under-
taken. Suspicion of rupture should remain an indi-
cation for immediate surgery. Once surgery is
undertaken, a conservative approach, i.e., uni-
lateral adnexectomy should be used so that re-
productive potential and hormonal production are
preserved whenever possible and/or appropriate
for the individual circumstances.
References
1. Eschenbach DA. Acute pelvic inflammatory disease:
etiology, risk factors, and pathogenesis. Clin Obstet
Gynecol 1976;19:147-69
2. Nebel WA, Lucas WE. Managment of tubo-ovarian
abscess. Obstet Gynecol 1968;32:382-6
3. Clark JFJ, Moore-Hines S. A study of tubo-ovarian
abscess at Howard University Hospital (1965 through
1975). J Nat Med Assoc 1979;71:1109-11
Landers and Sweet
4. Kaplan AL, Jacobs WM, Ehresman JB. Aggressive
management of pelvic abscess. Am J Obstet Gynecol
1967;98:482-7
5. Franklin EW III, Hevron JE lr, Thompson JD. Manage-
ment of the pelvic abscess. Clin Obstet Gynecol 1973;
16:66-79
6. Ginsburg DS, Stern JL, Hamod KA, Genadry R, Spence
MR. Tube-ovarian abscess: a retrospective review. Am
1 Obstet Gynecol 1980;138:1055-8
7. Rivlin ME, Hunt JA. Ruptured tuboovarian abscess. Is
hysterectomy necessary? Obstet Gynecol 1977;50:519-
22
8. Golde SH, Israel R, Ledger WJ. Unilateral tuboovarian
abscess: a distinct entity. Am J Obstet Gynecol 1977;
Downloaded from http://cid.oxfordjournals.org/ at Michigan State University on July 4, 2015
127:807-10
9. Collins CG, Jansen Wl. Treatment of pelvic abscess in-
fections. Clin Obstet Gynecol 1959;2:152-9
10. Pedowitz P, Bloomfield RD. Ruptured adnexal abscess
(tuboovarian) with generalized peritonitis. Am 1 Obstet
Gynecol 1964;88:721-9
11. Vemeeren 1, TeLinde RW. Intra-abdominal rupture of
pelvic abscesses. Am 1 Obstet Gynecol 1954;68:402-9
12. Mickal A, Sellmann AH. Management of tubo-ovarian
abscess. Clin Obstet Gynecol 1969;12:252-64
13. Taylor KJW, Wasson JF, DeGraaf C, Rosenfield AT,
Andriole VT. Accuracy of grey-scale ultrasound diag-
nosis of abdominal and pelvic abscesses in 220 patients.
Lancet 1978;1:83-4
14. Altemeier WA. The anaerobic strpetococci in tubo-ovarian
abscess. Am J Obstet Gynecol 1940;39:1038-42
15. Sweet RL. Anaerobic infections in the female genital tract.
Am J Obstet Gynecol 1975;122:891-901
16. Sweet RL. Treatment of mixed aerobic-anaerobic infec-
tions of the female genital tract. 1 Antimicrob
Chemother 1981;8(Suppl D):105-14
17. Weinstein WM, Onderdonk AB, Bartlett lG, Gorbach SL.
Experimental intraabdominal abscess in rats: develop-
ment of an experimental model. Infect Immun 1974;10:
1250-5
18. Weinstein WM, Onderdonk AB, Barlett lG, Louie 'rr,
Gorbach SL. Antimicrobial therapy of experimental in-
traabdominal sepsis. J Infect Dis 1975;132:282-6
19. Klempner MS, Styrt B. Clindamycin uptake by human
neutrophils. 1 Infect Dis 1981;144:472-9
20. loiner KA, Lowe BR, Dzink JL, Bartlett lG. Antibiotic
levels in infected and sterile subcutaneous abscesses in
mice. J Infect Dis 1981;143:487-94
21. Simon GL, Richmond DM, Tally FP, Barza M, Gorbach
SL. Penetration of clindamycin into experimental in-
fections with Bacteroides fragilis. 1 Antimicrob
Chemother 1981;8:59-64.