American Journal of Emergency Medicine xxx (xxxx) xxx

Contents lists available at ScienceDirect

American Journal of Emergency Medicine

journal homepage: www.elsevier.com/locate/ajem

Assessing the effectiveness of empiric antimicrobial regimens in cases of

septic/infected abortions
Yuval Fouks, MD a,⁎, Ofri Samueloff, MD b, Ishai Levin, MD, BPharm a, Ariel Many, MD a,
Sharon Amit, MD, PhD c, Aviad Cohen, MD a
a
Lis Maternity Hospital, Department of Obstetrics and Gynecology, Tel Aviv Sourasky Medical Center (affiliated to the Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel), Tel Aviv, Israel
b
Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel
c
The Department of Clinical Microbiology and Infectious Diseases, Hadassah University Hospital, Jerusalem, Israel

a r t i c l e i n f o a b s t r a c t

Article history: Introduction: Infected abortion is a life-threatening condition that requires immediate surgical and medical inter-
Received 2 May 2019 ventions. We aimed to assess the common pathogens associated with infected abortion and to test the microbial
Received in revised form 5 August 2019 coverage of various empiric antimicrobial regimens based on the bacteriological susceptibility results in women
Accepted 5 August 2019 with infected abortions.
Available online xxxx
Methods: A retrospective study in a single university-affiliated tertiary hospital. Electronic records were searched
for clinical course, microbial characteristics, and antibiotic susceptibility of all patients diagnosed with an infected
Keywords:
Infected abortion
abortion. The effectiveness of five antibiotic regimens was analyzed according to bacteriological susceptibility re-
Septic abortion sults.
Antibiotics Results: Overall, 84 patients were included in the study. The mean age of patients was 32.3(SD ± 5.8) years, and
Microbiology the median gestational age was 15 (IQR 8–19) weeks. Risk factors for infection were identified in 23 patients
(27.3%), and included lack of medical insurance (n = 12), recent amniocentesis/chorionic villus sampling or
fetal reduction due to multifetal pregnancies (n = 10). The most common pathogens isolated were Enterobacte-
riaceae (35%), Streptococci (31%), Staphylococci (9%) and Enterococci (9%). The combination of intravenous am-
picillin, gentamicin and metronidazole showed significant superiority over all the other tested regimens
according to the susceptibility test results. Piperacillin-tazobactam as an empiric single-agent drug of choice
and provided a superior microbial coverage, with a coverage rate of 93.3%.
Conclusions: A combination of ampicillin, gentamicin, and metronidazole had a better spectrum of coverage as a
first-line empiric choice for patients with infected abortion.
© 2019 Published by Elsevier Inc.

1. Introduction lacking the necessary skills or in an environment that does not conform
to minimal medical standards, or both” [5]. In contrast, infected abortion
Infected abortion refers to a complicated form of any type of abor- is uncommon in developed countries: the estimated case-fatality rate is
tion (spontaneous or induced) accompanied by infection [1]. It is esti- 0.7 per 100,000 of legal abortions in the US [4].
mated that approximately 10% of all maternal deaths worldwide are The diagnosis of infection in the setting of abortion is indicated by
due to a sepsis-related complication of an abortion [2]. Localized infec- the presence of fever or chills, foul smelling vaginal or cervical dis-
tion within the placenta can spread rapidly to the uterus and distant or- charge, abdominal or pelvic pain, prolonged vaginal bleeding or spot-
gans, leading to sepsis, fetal loss, and septic shock [3]. Complications of ting, uterine tenderness, and elevated inflammatory markers [5].
infected abortions are the leading cause of death after either spontane- Patients with infected abortion should be treated promptly by surgical
ous or induced abortions [4]. The majority of infected abortions occurs removal of the conception product and the administration of antimicro-
in developing countries as a result of unsafe abortions, which are de- bials. Broad-spectrum antibiotics should cover genitourinary and fecal
fined by the World Health Organization (WHO) as “a procedure for ter- flora, including Gram-positive, Gram-negative aerobic and anaerobic
minating an unintended pregnancy, carried out either by persons bacteria as well as Mycoplasmataceae. Current antimicrobial regimens
are mostly based on long standing bacteriological studies, and there is
no consensus regarding the most effective empiric antibiotic protocol
⁎ Corresponding author at: Lis Maternity Hospital, Department of Obstetrics and
Gynecology, Tel Aviv Sourasky Medical Center, 6 Weizmann Street, Tel Aviv 6423906,
[6]. Given the rapid emergence of antimicrobial resistant bacteria in
Israel. even young and healthy adults [7], it has become a matter of consider-
E-mail address: Fouksi@gmail.com (Y. Fouks). able importance to re-evaluate the pathogens associated with septic

https://doi.org/10.1016/j.ajem.2019.158389
0735-6757/© 2019 Published by Elsevier Inc.
2 Y. Fouks et al. / American Journal of Emergency Medicine xxx (xxxx) xxx

abortions and update therapeutic guidelines accordingly. Thus, the outcomes, presenting signs and symptoms, inflammatory markers
main objectives of our study were to survey prevailing pathogens asso- upon admission [white blood cells (WBC) and C-reactive protein
ciated with infected abortion and to assess the effectiveness of various (CRP)] and results of bacteriologic cultures (cervical, blood). Bacterial
empiric antimicrobial regimens based on the antibiograms of the iso- identification and antibiotic susceptibility test results from clinical spec-
lated pathogens. imens were performed at the local microbiology laboratory using the
Vitek MS and Vitek2 systems (bioMerieux, St. Louis, MO) according to
2. Methods the Clinical and Laboratory Standards Institute criteria [8]. Isolates con-
sidered as being contaminants were excluded from the analysis (de-
This retrospective study was conducted in a single university- tailed in Results). Identical isolates in multiple cultures in the same
affiliated tertiary medical center with an annual rate of approximately patient were considered as one.
900 dilation and curettage (D&C) procedures. The institutional review Antibiotic susceptibility patterns of all clinically significant isolates
board approved the study design, protocol, and waiver of informed con- were compared to the initial empiric treatment. Empiric treatment
sent (IRB approval number: 0620–15-TLV). Comprehensive electronic was classified as retrospectively adequate if the cultured pathogen
medical records were reviewed, obtaining all patients who were diag- was susceptible to the antibiotic treatment administered at the time of
nosed with infected abortion between 1/2011 and 12/2017. We identi- admission, based on the antibiotic susceptibility test results. In order
fied potentially eligible cases with International Classification of to determine the most effective antibiotic regimen, a second analysis
Diseases, Ninth Revision codes with a discharge diagnosis of infectious was performed to evaluate the adequacy of several empiric antimicro-
associated abortion. These results were cross matched with our local bial regimens as suggested by Eschenbach et al. [3]: 1) ampicillin, gen-
microbiology laboratory registry based on blood, cervical and tissue cul- tamicin and metronidazole, 2) amoxicillin-clavulanic acid,
ture taken in cases of suspected infected abortion at the time of admis- 3) piperacillin-tazobactam, 4) levofloxacin and metronidazole, and
sion. All health records of the study candidates for objectively confirmed 5) clindamycin and gentamicin. Adequacy of treatment was defined as
or presumed infected abortion were manually reviewed. Three abstrac- the number of cases in which the patients received a protocol covering
tors, all practicing obstetrics and gynecology, were trained for data col- the eventual cultured pathogens. In addition, we sought to determine
lection and use of the electronic data collection instrument. The whether the initial antibiotic regimen should have been changed
principal investigator monitored data collection activity following the based on the culture results. This analysis was further stratified accord-
first 10 cases and by periodic meetings, reviewing each abstractor's per- ing to disease severity (“localized infected abortion” vs. “septic abor-
formance. A standardized data abstraction form was developed follow- tion”) or gestational age (first vs. second trimester).
ing an extensive literature review in accordance with the study
objectives. To ensure the quality of the data extraction, the abstractors 2.1. Statistical analyses were performed using SPSS 24.0 statistical software
were blinded to the study objectives, and any discrepancies during the (Chicago, IL)
study were clarified by the principal investigator. All charts were
reviewed through the electronic hospital information system and man- Descriptive statistics were expressed as mean, median, standard de-
ual review of paper charts. In case of missing data, the variable was de- viation, and minimum and maximum values. The Student t-test was
leted from the analysis. used for parametric variables that are normally distributed. All tests
The patients included in the analysis were those presenting to the were two-sided, and P values of b0.05 were considered significant.
emergency department between 6 + 0/7–19 + 6/7 weeks of gestation
due to incomplete, missed abortion or with retained products of preg- 3. Results
nancy and a discharge diagnosis of infected abortion. Patients were di-
agnosed as having either localized infected or septic abortion 3.1. Incidence
according to the severity of the infection. Those with presenting signs
and symptoms of localized pelvic infection (uterine or cervical motion During the 8-year study period, 7298 patients underwent uterine
tenderness and foul-smelling vaginal discharge) were diagnosed as suction curettages at our institution. Eighty-four of them (1.1%) were di-
having “localized infected abortion”, whereas patients with clinical agnosed with infected abortion, of whom 51 (60.7%) were considered as
signs and symptoms suggesting generalized inflammation (chills or being “septic” according to the above-mentioned criteria.
sweats, fever, rebound tenderness, hypotension, and tachycardia)
were diagnosed as having “septic abortion” [7]. Blood and cervical cul- 3.2. Demographic data and risk factors
tures were obtained immediately upon admission where swabs from
products of conception were taken following uterine evacuation. Paren- Demographic data and risk factors of the study cohort are presented
teral antibiotics therapy was initiated at time of admission and uterine in Table 1. The median gestational age at diagnosis was 15 weeks (IQR
evacuation was accomplished within 12 h of admission. A uterine evac- 8–19). Risk factors for infection were identified in 23 patients (27.3%)
uation was performed with a rigid cannula connected to a manual vac- as follows: 12 had no medical insurance, 10 had a recent intrauterine
uum system. An ultrasonographic scan of the uterine cavity was manipulation (five had amniocentesis/chorionic villus sampling and
performed at the end of each procedure to verify the complete removal five with multifetal pregnancies had fetal reduction), and one was diag-
of products of pregnancy. nosed with infected abortion in the presence of an intrauterine device.
There was no uniform antibiotic protocol upon the diagnosis of lo- There was no record of any patient in our study having undergone an
calized infected or septic abortion during the 8-year study period. Em- unsafe surgical abortion prior to admission.
piric treatment included IV ceftriaxone 1 g qd plus IV metronidazole
500 mg bid and oral doxycycline 100 mg bid, IV amoxicillin-clavulanic 3.3. Clinical characteristics
acid 1 g tid, or IV ampicillin 2 g qid plus IV gentamicin 5 mg/kg IV qid,
according to the admitting physician's preference. In cases of positive Lower abdominal pain and/or cervical motion tenderness were the
cultures, the initial empiric treatment was replaced by a definitive treat- most common symptoms documented at initial presentation (n = 57,
ment as determined by the antibiogram of the isolated pathogen. 67.8%). Fever was present in 26 (30.9%) patients, of whom 15 (57.6%)
The retrieved data included demographics, medical history, risk fac- had a fever ≥39 °C. Ten patients (11.9%) were admitted to the intensive
tors for infection [recent (b7 days) intrauterine manipulation (amnio- care unit (ICU) following uterine evacuation due to severe sepsis. Four
centesis, chorionic villus sampling and fetal reduction), presence of of the patients who were admitted to the ICU had disseminated intra-
intrauterine device and absence of medical insurance], adverse vascular coagulation, three had adult respiratory distress syndrome,
 Y. Fouks et al. / American Journal of Emergency Medicine xxx (xxxx) xxx 3

Table 1 3.5. Adequacy of antibiotic treatment

Patient demographics and clinical characteristics (N = 84)

Mean age at abortion (y) 32.3 (±5.8) The most commonly prescribed empiric antibiotic regimen was in-
Gestational age, week (range) 15 (8–19) travenous (IV) ceftriaxone 1 g qd plus IV metronidazole 500 mg bid
Gravidity, n (range) 2 (1–3) and oral doxycycline 100 mg bid (n = 39, 47.5%) followed by IV
Parity, n (range) 1 (0–2)
amoxicillin-clavulanic acid 1 g tid (n = 15, 18.2%), IV ampicillin 2 g
Patients with N1 past miscarriages, n (%) 23 (28.3)
Conceived via assisted reproductive technology, n (%) 17 (21.9) qid plus IV gentamicin 5 mg/kg qd (n = 11, 13.0%), and IV clindamycin
History of cesarean section, n (%) 12 (14.8) 900 mg tid plus IV gentamicin 240 mg qd (n = 2, 2.4%).
Risk factors, n (%) 23 (27.3) According to bacterial identification and antimicrobial susceptibility
No medical insurance 12
tests, 33 (71.7%) patients received appropriate empiric treatment fol-
Amniocentesis/chorionic villus sampling b5 days 5
Fetal reductions b5 days 5
lowing the diagnosis of infected abortion. The pathogens resistant to
Pregnancy with Intrauterine device, n 1 the empiric treatment that had been given were: Enterobacteriaceae (n
Symptoms and signs in admitted patients, n (%) = 5, including one extended-spectrum beta-lactamase-producing bac-
Abdominal pain and/or premature contractions 30 (37) terium), Enterococcus faecalis (n = 4), Listeria monocytogenes (n = 2),
Cervical motion tenderness 27 (33.3)
and Pseudomonas aeruginosa (n = 1). The mean interval between ad-
Fever [PO N38.0 °C] 26 (30.9)
Malaise 14 (17.2) mission to bacterial identification and specific antibiotic susceptibility
Leukocyte count of N15,000 cells/mm3 or b3000 cells/mm3 n (%) 18 (21.4) result was 3.7 ± 1.5 days. Patients with isolates resistant to empiric
C-reactive protein at admission (mg/L) 68.1 (±43.2) treatment had significantly longer hospitalization compared to patients
Length of hospitalization, days (range) 5.8 (3–14)
with sensitive isolates (8.2 Vs 5.1 days, P = .04 CI 95% 5.74–0.05).
Transferred to the intensive care unit, n (%) 10 (11.9)
Fig. 2 demonstrates the presumed adequacy of several empiric anti-
Data are presented as mean (±SD), median (interquartile ranges) and n (%). microbial regimens as suggested by Eschenbach et al. [3]. The adequacy
rates were based on the antibiograms of all pathogenic isolates. It
emerged that the combination of ampicillin, gentamicin, and metroni-
dazole would have yielded substantially higher adequacy rates in com-
three had acute liver failure, and four had acute renal injury. None of the parison to all other regimens and to the empiric antibiotic treatment
patients required a hysterectomy and there were no maternal deaths. actually prescribed, with not so much as a single resistant isolate. In a
(Table 1). subanalysis of 23 bacteremic patients, a combination regimen of ampi-
cillin plus gentamicin and metronidazole or single agent piperacillin-
tazobactam were superior to the prescribed regimens (8 vs. 0 resistant
3.4. Microbial profile isolates, P = .001).

Of the 109 positive cultures obtained from 46 patients (42.2%), 11

cultures were excluded from analysis due to presumably nonpathogenic 3.6. Gestational age and disease severity
microorganisms, such as Gardnerella vaginalis, Candida albicans and
coagulase-negative staphylococci. The remaining 98 isolates were con- A subanalysis of the microbial diversity of first-trimester vs. second-
sidered as pathogens. The microbial profiles of the pathogenic isolates trimester infection-associated abortions showed similar rates of the var-
are shown in Fig. 1. The most frequently encountered isolates were En- ious pathogens (Table 2): Enterobacteriaceae (27.5% vs. 36.6%), Strepto-
terobacteriaceae (35%), Streptococci (31%), Staphylococci (9%) and En- cocci (27.5% vs. 23.3%) and Staphylococci (13.7% vs. 16.6%). Listeria
terococci (9%). A single organism was isolated in 35 patients (42.8%), monocytogenes were not identified in first trimester abortions in this co-
two organisms in 10 patients (11.9%) and three organism in 1 patient hort. Conversely, second-trimester infection-associated abortions were
(1.1Twenty-three blood cultures (27%), 22 cervical cultures (26%) and associated with a more severe presentation [mean WBC 13.6 ± 6.3 Vs
14 product of conception cultures (16%) showed bacterial growth. 20.8 ± 6.5 (95% CI -4.3-10.0) P = .0001, CRP 46.2 ± 49 Vs 70.2 ± 57

1(2%) 1(2%)

2( 4%)
2(4%)
Enterobacteriaceae
2(4%)
Streptococci

16(35%) Staphylococci
4(9%) Enterococci
Listeria monocytogenes
Bacteroides fragilis
4(9%) Haemophilus influenza
Diphtheroids
Pseudomonas aeruginosa

14(31%)

Fig. 1. Pathogen isolated from cervicovaginal, tissue and blood cultures.

4 Y. Fouks et al. / American Journal of Emergency Medicine xxx (xxxx) xxx

Ampicillin, gentamicin and metronizadole

combinaon therapy

Piperacillin/Tazobactam

Levofloxacin, metronidazole combinaon therapy

Clindamycin, gentamicin combinaon therapy

P< 0.1 P< 0.01 P< 0.002 P< 0.002

Amoxicillin/clavulanate

100
93.3
Rate of adequat covrege

86.6
82.2 82.2

Fig. 2. Adequacy of empiric antibiotic treatment.

(95% CI -0.4-47.9), P = .04] and a longer hospitalization [4.6 ± 2.7 Vs 7.1 into consideration the polymicrobial nature of this infection [3]. How-
± 2.7 (95% CI -1.2-3.7) days, P = .0001]. ever, these recommendations are based on small studies that had
mostly been performed three decades ago [12-14]. Antimicrobial resis-
4. Discussion tance has become a major contemporary health concern that is reported
in healthy childbearing age patients, thereby questioning the adequacy
Infected abortion is a life-threatening condition that carries signifi- of previously recommended empiric protocols.
cant morbidity. The infection may be localized to the placenta and pro- The reported rate of positive cultures in women with septic abortion
duce local symptoms, or the bacteria may gain access to the maternal was shown to be low in previous studies [10,15]. In the study by Dham
intervillous space, resulting in systemic infection and potential distant et al., identification of pathogenic bacteria was possible in 57% of the pa-
organ damage [9]. The prevalence of infection among patients hospital- tients [15]. Similarly, Finkielman et al. showed that only 24% of the pa-
ized for abortion varies between 3 and 15% in developing countries [2]. tients with septic abortion admitted to the intensive care unit had a
In contrast, infected abortion is far less frequent in developed countries, positive culture [10]. In their study, a single organism was isolated in
especially where illegal abortions are uncommon and qualified surgical 80% of the case, of which Enterobacteriaceae were the most common
treatment is available. In the present study, 1.1% of the patients who identified pathogens. Similar to these results, in the present study path-
underwent surgical abortions were diagnosed with infection, however, ogenic bacteria were identified in 42% of our patients, of which Entero-
this may represent an overestimation of the actual prevalence, as many bacteriaceae were the most frequently encountered isolates. Although
patients were referred to our tertiary medical center. in the majority of our patients a single pathogen was isolated,
The clinical presentation of infected abortion varies and depends on polymicrobial etiology was detected in 24%. The rate of polymicrobial
the extent of infection (i.e., local vs. systemic). Fever is a common pre- infection in previous reports [12,15] ranged between 20 and 50%, em-
senting sign in patients with septic abortion, reportedly involving phasizing the need for broad spectrum antibiotic coverage.
48%–57% of those patients [10,11]. In our study, lower abdominal pain Empiric antibiotic coverage was adequate in only 70% of the patients
and/or cervical motion tenderness were the most common presenting with positive cultures. The length of hospital stay was longer among pa-
signs (70.3%), followed by fever (30.9%). These differences can be ex- tients with inadequate empiric coverage (8.2 Vs 5.1 days, P = .04). In
plained by the proportion of patients in our study that were diagnosed search for the optimal antibiotic protocol, the adequacy of the initial
as having localized infected abortion (presumably due to rapid referral empiric treatment was compared to putative and empiric regimens as
and treatment), with the infection still localized to the uterus and in suggested by Eschenbach et al. in their recent review [3]. A combination
the absence of any systemic manifestation. of ampicillin, gentamicin, and metronidazole was universally efficient
The principal treatment for infected abortion is IV antibiotic therapy (i.e., no resistant isolates), and significantly more efficient than the
followed by prompt uterine evacuation. The associated pathogens orig- four other regimens to which it was compared. The most effective
inate mainly from the vaginal flora, and are comprised of a variety of single-drug option was piperacillin-tazobactam, covering 93.3% of all
aerobic and anaerobic bacteria [12] Current recommendations regard- isolates.
ing empiric antibiotic treatment are broad-spectrum coverage, taking The presence of Clostridium perfringens has been reported in up to 5%
of all septic abortions, especially in relation with unsafe ones [3]. Clos-
Table 2 tridium sordellii was recently associated with cases of abortion-related
Microbial diversity in infected abortions during first and second trimester. deaths after medical abortions [16,17]. Despite its rarity, infection with
Clostridium sordellii can be lethal and therefore calls for a high level of
Pathogen 6–12 weeks gestation 13–19 weeks gestation P value
suspicion. Although these two pathogens were not isolated in the pres-
(n = 29) (n = 30) ent study, they are known to be susceptible to a number of antibiotics,
n (%) n (%) including beta-lactams, clindamycin, and metronidazole, which are
Enterobacteriaceae 8 (27.5) 11 (36.6) 0.4 part of the empiric antibiotic that were evaluated [18]. The relative mi-
Streptococci 8 (27.5) 7 (23.3) 0.7 crobial similarity between isolates in cases of localized infected (local-
Staphylococci 4 (13.7) 5 (16.6) 0.7 ized infection) and septic abortion (systemic infection) support the
Enterococci 5 (17.2) 4 (13.3) 0.6
concept that infected abortion is an early stage of the disease, and not
Listeria monocytogenes 0 (0) 2 (6.6) 0.1
Diphtheroids 1 (3.4) 0 (0) 0.3 a representative of a different infection process involving less virulent
Bacteroides fragilis 1 (3.4) 1 (3.3) 0.9 pathogens.
Pseudomonas aeruginosa 1 (3.4) 0 (0) 0.3 The impact of gestational age on disease severity has been previ-
Haemophilus influenzae 1 (3.4) 1 (3.3) 0.9 ously investigated. In the study by Bartlett et al., the mortality ratio
(n) cultured pathogens. per 100,000 induced abortions increased substantially as gestational
 Y. Fouks et al. / American Journal of Emergency Medicine xxx (xxxx) xxx 5

age increased [4]. Similarly, in the present study, patients in their second Writing assistance
trimester had a more severe disease, as indicated by increased inflam-
matory markers at admission (WBC and CRP) and longer hospitaliza- None to declare.
tion. Microbiologic subanalyses showed that the main pathogenic
bacterial groups were essentially similar throughout pregnancy. Thus, Funding disclosure
we consider that the differences in the disease course and severity asso-
ciated with gestational age might be attributed to larger placental vol- This research did not receive any specific grant from funding agen-
ume and vascularity that facilitate rapid bacteria spread. cies in the public, commercial, or not-for-profit sectors.
In a recent meta-analysis aimed at evaluating the effectiveness of
various antibiotic regimens in the treatment of septic abortions, the au- Declaration of Competing Interest
thors concluded that the existing data are limited to old studies and are
therefore insufficient to provide guidelines for changes to the existing All authors report no conflict of interest and that this research was
treatment for septic abortion [6]. non-funded.
As far as we know, this is the first study to compare antibiotic ade-
quacy based on the actual antimicrobial susceptibility of clinical isolates References
in order to establish a culture-guided treatment regimen. Additionally,
there were no data regarding the associated pathogens in relation to [1] Stubblefield PG, Grimes DA. Septic abortion. N Engl J Med 1994;331(5):310–4.
[2] Adler AJ, Filippi V, Thomas SL, Ronsmans C. Quantifying the global burden of morbid-
disease severity and gestational age. ity due to unsafe abortion: magnitude in hospital-based studies and methodological
As a retrospective single center study, it may not represent the issues. Int J Gynaecol Obstet 2012;118 Suppl 2:S65–77.
whole spectra of flora in other countries and regions, limiting generaliz- [3] Eschenbach DA. Treating spontaneous and induced septic abortions. Obstet Gynecol
2015;125(5):1042–8.
ability of our results. The fact that not all of our patients had a positive [4] Bartlett LA, Berg CJ, Shulman HB, Zane SB, Green CA, Whitehead S, et al. Risk factors
culture results may lead to underestimation of the reported pathogens. for legal induced abortion-related mortality in the United States. Obstet Gynecol
For example, anaerobic bacteria are underestimated in this study: they 2004;103(4):729–37.
[5] In: nd, ed. Safe abortion: technical and policy guidance for health systems. Geneva.
are only partially reported from abortion products due to local labora- 2012.
tory practices of collection methods, although these pathogens are [6] Udoh A, Effa EE, Oduwole O, Okusanya BO, Okafo O. Antibiotics for treating septic
targeted in all the above-mentioned antimicrobial regimens. Similarly, abortion. Cochrane Database Syst Rev 2016;7:CD011528.
[7] Clinical management of abortion complications: a practical guide Authors: World
Mycoplasmataceae, which are usually detected by molecular methods,
Health Organization, Department of Reproductive Health and Research Organiza-
were not included in the analysis. Lastly, despite showing significant tion. WH. Published 1994.
clinical variations among the different groups, our treatment-guiding [8] Humphries RM, Ambler J, Mitchell SL, Castanheira M, Dingle T, Hindler JA, et al. CLSI
methods development and standardization working group best practices for evalu-
conclusions are based on in-vitro susceptibility and not on clinical re-
ation of antimicrobial susceptibility tests. J Clin Microbiol 2018;56(4).
sponse which greatly depends on other factors, such as rapidity of refer- [9] Studdiford WE, Douglas GW. Placental bacteremia: a significant finding in septic
ral and prompt surgical intervention. abortion accompanied by vascular collapse. Am J Obstet Gynecol 1956;71(4):
842–58.
[10] Finkielman JD, De Feo FD, Heller PG, Afessa B. The clinical course of patients with
4.1. In conclusion, an infected abortion is a life-threatening condition that septic abortion admitted to an intensive care unit. Intensive Care Med 2004;30(6):
requires immediate surgical and medical interventions 1097–102.
[11] Savaris RF, de Moraes GS, Cristovam RA, Braun RD. Are antibiotics necessary after 48
hours of improvement in infected/septic abortions? A randomized controlled trial
Although in the majority of the cases a single isolate was identified, followed by a cohort study. Am J Obstet Gynecol 2011;204(4):301 e301–305.
most commonly Enterobacteriaceae, polymicrobial infection was not an [12] Rotheram Jr EB, Schick SF. Nonclostridial anaerobic bacteria in septic abortion. Am J
uncommon finding. Our study suggests that ampicillin, gentamicin and Med 1969;46(1):80–9.
[13] Ostergard DR. Comparison of two antibiotic regimens in the treatment of septic
metronidazole should be the first-line empiric choice for patients with abortion. Obstet Gynecol 1970;36(3):473–4.
infection-associated abortion, irrespective of the gestational age. [14] Chow AW, Marshall JR, Guze LB. A double-blind comparison of clindamycin with
penicillin plus chloramphenicol in treatment of septic abortion. J Infect Dis 1977;
135:S35–9 Suppl.
Authors' contribution to the manuscript [15] Dahm Jr CH, Ostapowicz F, Cavanagh D. Use of cephalothin in septic abortion. Obstet
Gynecol 1973;41(5):693–6.
Y Fouks- Protocol/project development, Data collection or manage- [16] Cohen AL, Bhatnagar J, Reagan S, Zane SB, D'Angeli MA, Fischer M, et al. Toxic shock
associated with Clostridium sordellii and Clostridium perfringens after medical and
ment, Manuscript writing/editing, data analysis O Samueloff - Data col-
spontaneous abortion. Obstet Gynecol 2007;110(5):1027–33.
lection or management I Levin- Manuscript writing/editing, data [17] Fischer M, Bhatnagar J, Guarner J, Reagan S, Hacker JK, Van Meter SH, et al. Fatal toxic
analysis A Many- Protocol/project development, Manuscript writing/ shock syndrome associated with Clostridium sordellii after medical abortion. N Engl
editing S Amit- Infectious diseases consultant, Manuscript writing/ J Med 2005;353(22):2352–60.
[18] Dempsey A. Serious infection associated with induced abortion in the United States.
editing A Cohen- Protocol/project development, Manuscript writing/ Clin Obstet Gynecol 2012;55(4):888–92.
editing, data analysis.