bs_bs_banner

doi:10.1111/jog.12617 J. Obstet. Gynaecol. Res. 2014

Comparison of severe pelvic inflammatory disease,

pyosalpinx and tubo-ovarian abscess

Ho Yeon Kim1, Jeong In Yang2 and ChongSoo Moon1

1
Department of Obstetrics and Gynecology, College of Medicine, Hallym University, Seoul and 2Department of Obstetrics and
Gynecology, Ajou University School of Medicine, Suwon, Korea

Abstract
Aim: Inflammation of the upper genital tract causes pelvic inflammatory disease (PID), which may be com-
plicated by pelvic abscesses, such as pyosalpinx and tubo-ovarian abscess (TOA). This study aimed to deter-
mine the clinical differences between pyosalpinx and TOA in patients with PID.
Material and Methods: We retrospectively evaluated 458 female patients who were admitted to Hallym
University Kang Dong Sacred Heart Hospital for a clinical diagnosis of PID from 1 January 2007 to 30 April
2012. Sociodemographic, clinical and laboratory data were compared among the non-abscess, pyosalpinx, and
TOA groups.
Results: We identified 110 patients (24%) diagnosed with pelvic abscess associated with PID, including 34
with pyosalpinx and 76 with TOA. The pyosalpinx group had shorter hospital stays (P = 0.007), lower
C-reactive protein levels (P = 0.015), smaller mass sizes (P < 0.001), and fewer surgical interventions (P < 0.001)
than the TOA group.
Conclusions: Pyosalpinx is a less severe form of PID that leads to shorter hospital stays and more favorable
outcomes than TOA.
Key words: pelvic inflammatory disease, pyosalpinx, tubo-ovarian abscess.

Introduction no reports of the differences between these two forms

Pelvic inflammatory disease (PID) describes inflamma-
tion of the upper genital tract that causes salpingitis,
endometritis, peritonitis, and occasionally pelvic
abscess, including pyosalpinx and tubo-ovarian
abscess (TOA). In the progression of PID, cervicitis and
endometritis come first, followed by salpingitis, which
causes purulent material spillage from the salpinx to
the uterus and nearby peritoneum. Adhesions close the
salpinges, which are filled with pus, form a pyosalpinx.
Ultimately, during the final stage of PID, the nearby
ovaries and tubes become infected, leading to TOA.1,2
There is little clinical evidence suggesting whether
pyosalpinx or TOA is the more severe form of PID, and
of PID. Hence, in this retrospective study, we aimed
to differentiate pyosalpinx and TOA according to
demographic, clinical, and microbiological factors.
Methods
Study population
We retrospectively reviewed the medical records of all
the patients diagnosed with PID at the Hallym Univer-
sity Kang Dong Sacred Heart Hospital, Seoul, Korea,
between 1 January 2007 and 30 April 2012. The Institu-
tional Review Board at the hospital approved this
study. A total of 458 inpatient records with Interna-
tional Classification of Diseases-10 codes N70–N74

Received: January 5 2014.

Accepted: September 10 2014.
Reprint request to: Professor ChongSoo Moon, Department of Obstetrics and Gynecology, College of Medicine, Hallym University,
#150 Seongnae-gil, Gangdong-gu, Seoul 134-701, Korea. Email: obgyn25@hallym.or.kr

© 2014 The Authors 1

Journal of Obstetrics and Gynaecology Research © 2014 Japan Society of Obstetrics and Gynecology
H. Y. Kim et al.

(a) (b)

Figure 1 Pyosalpinx. (a) A fluid-

filled dilated tube (mass) with a
low-level echo beside the normal
ovary (arrow). (b) A large dilated
tubular structure filled with
debris/purulent material and
with incomplete septation (cali-
per) where the tube folds on
itself.

were selected. According to the 2010 guidelines of the

Centers for Disease Control and Prevention, PID was
diagnosed when a combination of abdominal tender-
ness, cervical motion tenderness, and bilateral adnexal
tenderness and at least one of the following minor
diagnostic criteria were observed in sexually active
women: documentation of a cervical infection with
Chlamydia trachomatis or Neisseria gonorrhoeae, mucopu-
rulent cervicitis, body temperature (BT) > 38.3°C,
elevated erythrocyte sedimentation rate (ESR) or
C-reactive protein (CRP) level, or presence of an
inflammatory mass on pelvic sonography.3 Of the 458
hospitalized female patients, 348 were designated as
the non-abscess group and 110 as the abscess group
associated with PID. The abscess group was further
divided into subgroups of pyosalpinx in 34 patients
and TOA in 76. We excluded patients if they were preg-
nant, had similar pain from other gynecologic causes,
such as an endometriotic cyst, refused treatment, and
had pelvic abscesses of non-gynecologic origin.

Clinical analysis
Pyosalpinx and TOA abscess were initially diagnosed
using transvaginal ultrasound and/or computed
tomography (CT). Ultrasound was performed by one
of the presenting authors, who all have >5 years’ expe-
rience with gynecologic ultrasound. A CT scan was
reviewed by experienced radiologists with no prior
knowledge of the clinical and surgical data, other than
the presence of acute lower abdominal pain. In patients
who underwent surgery, gross findings and histologic
examination confirmed the diagnosis. A pyosalpinx
diagnosis was confirmed by ultrasound when a tubular
structure filled with debris/purulent material and
fluid/fluid layering was observed and when incom-
plete septation where the tube folds on itself was iden-
tified (Fig. 1).2 Despite the pyosalpinx, the ovarian
tissue showed no signs of infection but was slightly
Figure 2 Tubo-ovarian abscess. A complex multilocular
cystic mass (calipers) with thick irregular septations,
mural nodules, and low-level echoes. A separate ovary
could not be seen.
enlarged. We differentiated hydrosalpinx from pyosal-
pinx by ultrasound as the former appears as a tubular
anechoic adnexal structure usually with a thin wall and
incomplete septation.2 The CT scan showed a tubular
juxta-uterine mass with complex internal fluid and
thick enhancing walls.4 For TOA, the sonographic
finding was an irregularly marginated solid or cystic
mass with uniform internal septa with usually no dis-
crepancy between the tube and the ovary (Fig. 2).2,4 The
ovarian capsule and tissue were not definitively iden-
tified. A contrast-enhanced CT scan demonstrated a
solid-cystic adnexal mass with enhancement of the

2 © 2014 The Authors

Journal of Obstetrics and Gynaecology Research © 2014 Japan Society of Obstetrics and Gynecology
 Pyosalpinx and TOA

wall and septa.4 We categorized 28 cases with one side
TOA and the other side pyosalpinx into the TOA
group.
Cases were analyzed with respect to demographic
factors, including age, parity, abortion, marital status,
occupation, previous pelvic surgery, previous PID
history, types of contraception, smoking, and abnormal
Pap smear test result. We also assessed clinical factors,
including BT, length of hospital stay, symptom dura-
tion, white blood cell (WBC) counts, segmented neu-
trophil counts, ESR, CRP level, surgical procedures,
and mass size and laterality; and microbiological
factors, including N. gonorrhoeae, C. trachomatis,
Ureaplasma urealyticum, Mycoplasma hominis, Actinomy-
ces israelii, and others.
All culture specimens were obtained from the cervix,
an extracted intrauterine device, a blood sample,
and/or intra-abdominal pus. The specimens were
tested for C. trachomatis by enzyme-linked fluorescent
assay (Seegene; sensitivity 98–100%, specificity
95–100%), N. gonorrhoeae by culture, and U. urealyticum
and M. hominis using the mycoplasma IST 2 kit
(Biomerieux; sensitivity 98%, specificity 98%).
All patients received an intravenous antibiotic treat-
ment that consisted of a second- or third-generation
cephalosporin plus metronidazol and/or aminogly-
coside or levofloxacin plus metronidazole, if the patient
was hypersensitive to cephalosporin.
Continuous variables were analyzed using an
independent t-test for parametric variables and the
Mann–Whitney U-test for non-parametric variables.
Categorical variables were evaluated using the χ2 and
Fisher’s exact tests. Data distribution was tested for
normality using the Kolmogorov–Smirnov test. In
addition, a receiver–operator curve (ROC) was used to
evaluate cut-off, sensitivity, and specificity values. A
two-sided P < 0.05 was considered to indicate signifi-
cance. All the statistical analyses were performed using
spss 17.0.
Results
Sociodemographic factors
A total of 458 hospitalized PID cases were encountered
in our institution during the study duration of 5 years,
4 months. There was no difference in age between the
pyosalpinx and TOA groups. There was no significant
difference in parity, abortion, marital status, and previ-
ous Cesarean delivery between the pyosalpinx and
TOA groups. Only 20.7% of the patients were using
contraception. No significant difference in intrauterine
device use was observed between the pyosalpinx and
TOA groups. The patients in the non-abscess group
were less likely to use contraception than those in the
TOA group (Table 1).
Clinical and laboratory factors
Clinical and laboratory factors are reported in Table 2.
The mean BT was similar between the pyosalpinx and
TOA groups, although the patients with TOA had

Table 1 Sociodemographic characteristics

Non-abscess (n = 348) Pyosalpinx (n = 34) TOA (n = 76) P-value
Age (years) 31.6 ± 9.9 35 ± 10.2 37.8 ± 10.6* <0.001
Parity 1.8 ± 1.8 2.2 ± 1.7 2.5 ± 1.8* 0.019
Abortion 1.1 ± 1.2 1.3 ± 1.2 1.3 ± 1.3 NS
Married 180 (51.9) 16 (47) 48 (63.2) NS
Occupation 193 (55.9) 12 (35.3)* 34 (44.7) 0.023
Cesarean section 57 (16.4) 4 (11.8) 17 (22.4) NS
Previous pelvic surgery 62 (17.8) 12 (35.3)* 11 (14.5)** 0.028
Previous PID 50 (14.4) 3 (8.8) 3 (3.9)* 0.034
Contraception
IUD 26 (7.5) 6 (17.6)* 23 (30.3)* <0.001
Oral pills 18 (5.4) 2 (5.9) 0* NS
Tubal ligation 7 (2) 1 (3) 3 (3.9) NS
Others 16 (4.6) 1 (2.9) 2 (2.6) NS
No contraception 286 (82.1) 25 (73.6) 50 (65.8)* <0.001
Smoking 42 (12.9) 4 (11.8) 8 (11.8) NS
Abnormal PAP 20 (7.2) 0 3 (4.5) NS
*P < 0.05 compared to non-abscess. **P < 0.05 compared to pyosalpinx. Missing observations: abortion, n = 3; marital status, n = 1; occupation,
n = 3; history of PID, n = 1; oral pills, n = 13; smoking, n = 30; abnormal Pap smear, n = 83. IUD, intrauterine device; NS, not significant; PAP,
Papanicolaou test; PID, pelvic inflammatory disease; TOA, tubo-ovarian abscess.

© 2014 The Authors 3

Journal of Obstetrics and Gynaecology Research © 2014 Japan Society of Obstetrics and Gynecology
H. Y. Kim et al.

Table 2 Clinical and laboratory characteristics

Non-abscess (n = 348) Pyosalpinx (n = 34) TOA (n = 76) P-value
Body temperature (°C) 36.9 ± 0.7 37.5 ± 1.0* 37.6 ± 1.1* <0.001
Hospital day 8.7 ± 4.0 10.9 ± 3.3* 13.9 ± 6.3*,** <0.001
Symptom duration 9.1 ± 13.7 6 ± 6.7 9.7 ± 15.6 NS
WBC (×109/L) 9.5 ± 4.0 12.5 ± 4* 14.6 ± 6.9* <0.001
Segmented neutrophil (%) 68.9 ± 13.5 77.9 ± 11.1* 79.5 ± 10* <0.001
ESR (mm/h) 21 ± 20 38.5 ± 34.8* 48.7 ± 31.4* <0.001
CRP (mg/L) 29.7 ± 49.6 93.4 ± 116.9* 139.3 ± 115.8*,** <0.001
Surgical management 30 (8.6) 14 (41.2)* 61 (80.2)*,** <0.001
Size of mass (cm) — 3.8 ± 1.6 6.1 ± 2.3** <0.001
Laterality
Unilateral — 23 (67.7) 49 (64.5) NS
*P < 0.05 compared to non-abscess. **P < 0.05 compared to pyosalpinx. Data are presented as mean ± standard deviation, absolute numbers
(%). CRP, C-reactive protein; ESR, erythrocyte sedimentation rate; NS, not significant; TOA, tubo-ovarian abscess; WBC, white blood cells.

patients, with the TOA group being more likely to be

infected than the non-abscess group.

Figure 3 Receiver–operator curve of C-reactive protein
in the diagnosis of tuboovarian abscess. The area under
the curve was 0.649.
longer hospital stays than those in the pyosalpinx
group (P = 0.007). The CRP levels were significantly
higher in patients with TOA than in pyosalpinx
patients (P = 0.014). The ROC curve indicated the
utility of CRP as a diagnostic marker of TOA compared
to pyosalpinx (Fig. 3). CRP had a sensitivity of 75% and
a specificity of 54.5% with a cut-off value of 56.5 mg/L.
Microbiological factors
There were no significant differences in the isolation of
N. gonorrhoeae and/or C. trachomatis, U. urealyticum
and/or M. hominis, and other bacteria (Escherichia coli,
Bacteroides fragilis, Staphylococcus species, group B
Streptococcus, and Klebsiella species) among the three
groups. Actinomycosis was detected in only 1.6% of the
Treatment
Of the patients, 353 (77.1%) were treated medically
with broad-spectrum antibiotics. Among the 110
patients diagnosed with abscess, 35 (31.8%) had con-
servative treatment with antibiotics and 75 (68.2%)
underwent surgical treatment, consisting primarily of
laparoscopic abscess drainage and removal of the
abscess cavity and associated necrotic tissues. The
patients in the TOA group were more likely to have
undergone surgeries than those in the pyosalpinx
group (P < 0.001). There were 14 patients (13.3%) who
underwent surgery due to treatment failure and 17
(16.2%) who did so due to incidental detection of the
pelvic mass, which was initially considered unrelated
to the PID based on an imaging study. Three patients
with TOA underwent percutaneous drainage proce-
dures. The mean mass size of the TOA group was sig-
nificantly larger than that of the pyosalpinx group
(P < 0.001) (Table 2).
Discussion
This is the first study comparing the clinical features of
pyosalpinx and TOA. It is apparent that pyosalpinx
develops into TOA when the disease progresses1,2 but
no previous clinical information has defined the sever-
ity of pyosalpinx and TOA. Of the numerous baseline
sociodemographic, clinical, laboratory, and microbio-
logical data analyzed in this study, hospital stay, CRP
level, mass size, and rate of surgical intervention dif-
fered significantly between the pyosalpinx and TOA
groups. These variables suggest that pyosalpinx is

4 © 2014 The Authors

Journal of Obstetrics and Gynaecology Research © 2014 Japan Society of Obstetrics and Gynecology

likely to be a less severe form of PID than TOA. More-
over, in the presence of pyosalpinges the hospital stay
will likely be shorter and fewer surgical interventions
required.
To delineate pyosalpinx from TOA, thorough
manual palpation of the abdominal mass can be per-
formed, although ultrasound and CT imaging modali-
ties are often used. Ultrasound is the initial choice
because it is inexpensive and readily available, has
35–93% sensitivity, and can differentiate ovarian and
tubal disease involvement.2,5,6 The key feature we used
to differentiate pyosalpinx from TOA by ultrasound
was an intact pelvic structure, particularly intact
ovarian parenchyma and dilated tubes with low-level
echoic fluid collection.
More than half of the cases of pyosalpinx and TOA
were managed surgically. As we performed laparos-
copy in >80% of the patients, we could clearly grossly
discriminate whether a true pyosalpinx or TOA was
present. The strength of this study was that all patients
were diagnosed initially by ultrasound and/or CT, fol-
lowed by gross confirmation in 68% of the cases.
However, we acknowledge that the rate of surgical
intervention in our study was high compared to reports
that ∼25% of cases required surgery or drainage when
antibiotic treatment failed.1 Several recent studies rec-
ommended medical management with antibiotics
and/or drainage procedures as the primary treatment
for PID.7 Early drainage was associated with shorter
hospital stays and decreased morbidity.8,9 In contrast,
others reported that immediate laparoscopy had
advantages of accurate diagnosis and effective
treatment under magnification with minimal compli-
cations, with higher response rates, shorter hospitaliza-
tion times, and decreased infertility.10 However, the
optimal approach to management of TOA remains a
matter of debate.
Our study had several limitations. These include its
retrospective nature and the composition of the study
population as participants were from urban and local
populations. Thus, it remains unclear whether these
results can be generalized to other populations. In
addition, CRP is a well known and widely used inflam-
matory marker. Although our results demonstrate that
the mean CRP level in the TOA group was significantly
higher than that in the pyosalpinx group, CRP levels
© 2014 The Authors
Journal of Obstetrics and Gynaecology Research © 2014 Japan Society of Obstetrics and Gynecology
Pyosalpinx and TOA
exhibit considerable interindividual variation. Thus,
there is overlap between the two groups and CRP may
ultimately not be useful in individual cases.
In conclusion, pyosalpinx is likely to be a less severe
form of pelvic abscess, with a decreased CRP level,
mass size, hospital stay duration, and frequency of
invasive procedures. Practitioners should be aware
when admitting patients with pyosalpinx that they
may require less aggressive management and exhibit a
more benign clinical course than those with TOA.
Disclosure
No actual or potential conflict of interest in relation to
this article exists.
References
1. Chappell CA, Wiesenfeld HC. Pathogenesis, diagnosis, and
management of severe pelvic inflammatory disease and
tuboovarian abscess. Clin Obstet Gynecol 2012; 55: 893–903.
2. Cicchiello LA, Hamper UM, Scoutt LM. Ultrasound evalua-
tion of gynecologic causes of pelvic pain. Obstet Gynecol Clin
North Am 2011; 38: 85–114.
3. Workowski KA, Berman S, Centers for Disease Control and
Prevention. Sexually transmitted diseases treatment guide-
lines, 2010. MMWR Recomm Rep 2010; 59: 1–110.
4. Rezvani M, Shaaban AM. Fallopian tube disease in the non-
pregnant patient. Radiographics 2011; 31: 527–548.
5. Romosan G, Bjartling C, Skoog L, Valentin L. Ultrasound for
diagnosing acute salpingitis: A prospective observational
diagnostic study. Hum Reprod 2013; 28: 1569–1579.
6. Boardman LA, Peipert JF, Brody JM, Cooper AS, Sung J.
Endovaginal sonography for the diagnosis of upper genital
tract infection. Obstet Gynecol 1997; 90: 54–57.
7. Gjelland K, Ekerhovd E, Granberg S. Transvaginal
ultrasound-guided aspiration for treatment of tubo-ovarian
abscess: A study of 302 cases. Am J Obstet Gynecol 2005; 193:
1323–1330.
8. Perez-Medina T, Huertas MA, Bajo JM. Early ultrasound-
guided transvaginal drainage of tubo-ovarian abscesses: A
randomized study. Ultrasound Obstet Gynecol 1996; 7: 435–
438.
9. Levenson RB, Pearson KM, Saokar A, Lee SI, Mueller PR,
Hahn PF. Image-guided drainage of tuboovarian abscesses of
gastrointestinal or genitourinary origin: A retrospective
analysis. J Vasc Interv Radiol 2011; 22: 678–686.
10. Rosen M, Breitkopf D, Waud K. Tubo-ovarian abscess man-
agement options for women who desire fertility. Obstet
Gynecol Surv 2009; 64: 681–689.
5