Abstract
Brucella species occasionally cause spontaneous human abortion, but theories regarding whether they do so more frequently than do other infectious pathogens remain controversial. We reviewed 92 pregnant women who presented with acute brucellosis at a Saudi Arabian hospital. From 1983 through 1995, the cumulative incidence of pregnancy and brucellosis was 1.3 cases per 1000 delivered obstetrical discharges. The incidence of spontaneous abortion in the first and second trimesters was 43%, and the incidence of intrauterine fetal death in the third trimester was 2%. Antepartum antimicrobial therapy with cotrimoxazole or cotrimoxazole/rifampin was protective against spontaneous abortion (relative risk, 0.14; 95% confidence interval, 0.06–0.37; P < .0001). The beneficial effect of treatment occurred in women with febrile illness; vaginal bleeding at presentation usually led to spontaneous abortion. This study demonstrated that the incidence of spontaneous abortion among pregnant women with brucellosis is high and that these women should receive prompt therapy with antimicrobial agents when they present for medical care.
Brucellosis, a zoonotic disease of global importance [1], is endemic in Saudi Arabia (national seroprevalence, 15%) [2]. Endemicity in this region results from the persistence of domestic animal reservoirs for Brucella species and the human consumption of unpasteurized dairy products [3, 4].
Various Brucella species are well-known causes of contagious abortion in cattle, sheep, goats, swine, and dogs [5]. There is also evidence that brucellosis can produce spontaneous abortion in humans, which has been demonstrated by rare cases in which Brucella species were isolated from fetal or placental tissues, but it has not been demonstrated that brucellosis causes abortions more frequently than do other bacterial infections [6]. It is believed that brucellosis causes fewer spontaneous abortions in humans than it does in animals because of the absence of erythritol in the human placenta and fetus [7]. Erythritol is a constituent of normal ungulate fetal and placental tissue and, in cases of bovine abortion, promotes overwhelming infection of the placenta and fetus. An additional reason for the lesser role of Brucella infection in human abortion is the presence of anti-Brucella activity in human amniotic fluid [8].
In the present study, we reviewed the experience with brucellosis in pregnant women at a tertiary-care hospital in Saudi Arabia during a 13-year period. We sought to characterize clinical presentations, the incidence of abortion, and the effect of treatment with antimicrobial agents on the occurrence of spontaneous abortion.
Patients and Methods
The King Fahad National Guard Hospital (Riyadh, Saudi Arabia) has served the secondary- and tertiary-level medical needs of Saudi National Guard soldiers and their extended families since 1983. Guardsmen come from bedouin tribes that have retained a nomadic pastoral lifestyle that includes the consumption of unpasteurized milk from goats, camels, and sheep; many of these animals are known to be infected with Brucella species. Because of their nomadic lifestyles, not all pregnant women receive prenatal care, some are lost to follow-up, and some are treated for brucellosis at our hospital but subsequently deliver elsewhere.
The medical record database at the hospital was reviewed, and all patients who had a discharge diagnosis of brucellosis and pregnancy from 1983 through 1995 were identified. The diagnosis of acute brucellosis was based on compatible clinical features and a serum agglutinin titer of ⩾1 : 320 or a positive blood culture result. Patient records were reviewed, and the following information was extracted: patient age, obstetrical history, gestational age at the time of infection, symptoms, results of serological testing and blood cultures for Brucella, antimicrobial therapy, delivery outcome, results of placental cultures and histological findings, neonatal status, and neonatal results of serological testing and cultures for Brucella. The first trimester of pregnancy was defined as a gestational age of ⩽12 weeks; the second trimester of pregnancy, >12 to ⩽24 weeks; and the third trimester of pregnancy, >24 weeks. Patients with brucellosis were considered to have been treated if they received therapy with cotrimoxazole, rifampin, or a tetracycline for at least 4 weeks. Fetal deaths that occurred at ⩽24 weeks' gestation were designated “spontaneous abortions,” while deaths that occurred at >24 weeks' gestation were designated “intrauterine fetal deaths.”
A variety of microbiological methods were used at our hospital during the 13-year period to isolate Brucella species, determine antimicrobial susceptibility, and detect agglutinating antibodies; these methods have been described in detail elsewhere [9–11]. In brief, samples of blood and body fluids were cultured by use of the automated BACTEC System (Becton Dickinson). Before 1987, identification of Brucella organisms to the species level was not done because of technical limitations in the microbiology laboratory. In subsequent years, Brucella isolates were identified to the species level by use of standard biochemical methods. Antimicrobial susceptibility tests were performed according to a standard disk diffusion method. Agglutination tests for antibody to Brucella were performed by means of a microtiter serial dilution technique.
Epidemiological associations were expressed as relative risk with 95% CIs, and statistical significance was assessed by use of the χ2 test or 2-tailed Fisher's exact test where appropriate. All statistical measures were generated by use of Epi-Info, version 6 (Centers for Disease Control and Prevention).
Results
From 1983 through 1995, there were 545 cases of brucellosis at the hospital, of which 92 (17%) occurred in pregnant women (cumulative incidence, 1.3 cases of brucellosis in pregnant women per 1000 delivered obstetrical discharges [0.3 cases of brucellosis in pregnant women per 1000 hospital admissions]). The characteristics of the 92 pregnant women with acute brucellosis are shown in table 1. The mean age of the women was 26.9 years. Seventy-one (81%) of 88 women had 3 or more pregnancies prior to the study period, 62 (72%) of 86 had never spontaneously aborted before, and 44 (48%) of 92 presented during the second trimester of pregnancy.
Characteristics of 92 pregnant women with acute brucellosis in Saudi Arabia.
Overall, 40 (43%) of 92 pregnant women with acute brucellosis had spontaneous abortions during the first or second trimester, and 2 (2%) had intrauterine fetal deaths in the third trimester (table 2). The incidence of abortion was significantly higher during the first and second trimesters than it was during the third trimester. To put these rates in context, we calculated the incidence of spontaneous abortion or intrauterine fetal death for all women for the period from July 1999 through June 2000. During this 1-year interval, the total rate of spontaneous and missed abortions (i.e., early fetal death before completion of 24 weeks of gestation with retention of dead fetus or retained products of conception, not following spontaneous abortion, induced abortion, or delivery) during the first and second trimesters combined was 2.8% (589 spontaneous abortions and 121 missed abortions among the 25,540 women who received antenatal care), and the rate of intrauterine fetal death during the third trimester was 0.3% (66 intrauterine fetal deaths among the remaining 24,830 pregnant women who did not abort during the first or second trimesters).
Occurrence of spontaneous abortion and intrauterine death, according to trimester of pregnancy, in 92 women with acute brucellosis in Saudi Arabia.
Of the 92 pregnant women with brucellosis, 91 had positive serum agglutinin titers that ranged from 1 : 320 to ⩾1 : 20,480; the geometric mean titer was 1 : 2560. Serological testing was not performed for 1 patient; brucellosis in this patient was diagnosed by a blood culture result that was positive for Brucella species. There was no apparent relationship between the magnitude of the titer and the occurrence of bacteremia. Serum agglutinin titers ⩾1 : 2560, compared with those <1 : 2560, were not significantly associated with the occurrence of spontaneous abortion (P = .5).
Fifty-two of the 92 women had blood cultures performed. Twenty-two (42%) of these cultures had positive results: 16 yielded Brucella melitensis, 1 yielded Brucella abortus, and 5 yielded Brucella species that were not identified to the species level. Antimicrobial susceptibility testing for these 22 isolates revealed that 18 were susceptible to cotrimoxazole and 4 were resistant. Eight (36%) of 22 women with brucella bacteremia aborted, and 16 (53%) of the 30 women for whom blood cultures were negative aborted; this difference in the incidence of abortion was not statistically significant (P = .2).
The incidence of abortion according to pregnancy trimester, clinical presentation, and treatment is shown in table 3. Overall, of the 92 pregnant women with brucellosis, 41 (45%) received treatment. With all pregnancy trimesters represented, 100% of 38 women who presented with preterm febrile illness received antimicrobial therapy, and there was 1 intrauterine fetal death (3%). Of the women in the first and second trimesters who presented with vaginal bleeding and either with or without fever, 3 (8%) of 40 were treated with antimicrobial agents, and all aborted. None of 14 women with febrile illness at term or postpartum were treated, and there was 1 fetal death. Overall, 60 (65%) of the 92 women with brucellosis were febrile.
Incidence of spontaneous abortion and intrauterine death, according to pregnancy trimester and clinical presentation, among 92 pregnant women with acute brucellosis in Saudi Arabia.
The details of antepartum antimicrobial treatment are presented in table 4. Cotrimoxazole was administered as 1 double-strength tablet (800 mg or sulfamethoxazole and 160 mg of trimethoprim), given orally b.i.d., and rifampin was administered as 600–900 mg, given orally q.d. Antepartum treatment with antimicrobial agents was significantly more protective against the occurrence of abortion than was inadequate treatment or no treatment (RR, 0.14; 95% CI, 0.06–0.37; P < .0001). However, the incidence of abortion among 22 patients who received treatment with cotrimoxazole was not significantly different from that of 17 patients who received treatment with a combination of cotrimoxazole and rifampin (P = .6). Follow-up data were available for 36 infants of the 41 women who were treated: 33 (92%) were normal and well at follow-up, and 3 (8%) delivered prematurely (2 of the premature infants died).
Antepartum antimicrobial treatment and pregnancy outcomes for pregnant women with acute brucellosis in Saudi Arabia.
Serum agglutinin titers were measured for 7 of the infants born to women with brucellosis, and 4 of them had titers ⩾1 : 320. Cultures of blood specimens from 6 newborns (⩽1 week of age) were performed, but only 1 had a result that was positive for Brucella species; culture of blood from the mother of this infant with bacteremia was not done. The mothers of 4 of the 6 newborns for whom cultures were performed had brucella bacteremia, and all presented during the second trimester.
Placentas and fetuses were not systematically studied in this case series. Cultures of 11 placental specimens and 1 fetal specimen did not yield Brucella species. Histological examination of 28 placental samples revealed nonspecific inflammatory changes in 22 of them.
Discussion
We found that the incidence of spontaneous abortion and intrauterine death among a retrospective cohort of 92 pregnant women with acute brucellosis due primarily to B. melitensis was 46%. As shown in the Results section, the rates of spontaneous abortion and intrauterine death among women with active brucellosis substantially exceeded the rates of spontaneous abortion and intrauterine death among the general population of pregnant women at our institution during the corresponding trimesters. The frequency of fetal loss among patients with brucellosis was comparable in the first and second trimesters, but intrauterine fetal death during the third trimester was uncommon. Most of these women were older than 20 years of age, were not primigravid, and most had not had a spontaneous abortion prior to the study period. Our rate of incidence exceeds the 10% rate reported by Criscuolo and di Carlo [12] in 200 cases of B. melitensis infection in pregnant women; however, it is similar to the rate of abortion of 31% among 35 women who were in the first trimester of pregnancy and who had brucellosis, which was reported by Lulu et al. [13], and a 40% rate among 30 pregnant women with brucellosis, which was reported by Madkour [14]. Conversely, Sarram et al. [15] observed that, among 51 women who had spontaneous abortions during their second trimester, 11.6% had B. melitensis infection, while Makhseed et al. [16] diagnosed acute or chronic brucellosis in 7% of 29 women who had spontaneous abortions and 10% of 51 women who had intrauterine fetal deaths.
In his classic monograph, Spink [17] stated that “… the passage of time has produced no definitive evidence that the Brucellae produce abortions any more frequently than do other species of bacteria.” This statement is difficult to verify because data on the incidence of abortion caused by other microorganisms are lacking or are not directly comparable. For example, Campylobacter species, like Brucella species, are well-described causes of abortion in animals; however, in a series of 10 pregnant women with Campylobacter jejuni infection who were at a single hospital, only 1 (10%) had premature labor at 28 weeks' gestation that resulted in a neonatal death [18]. Other reports of Campylobacter infection in pregnant women include a report of a series of cases that were selected on the basis of the occurrence of abortion [19] and single case reports of fetal loss in association with septic shock, bacteremia, and diarrheal illness [20–22].
A MEDLINE search of the literature published from 1966 through 2000 that used the terms “E. coli” and “abortion” did not yield any reports of studies of patients with Escherichia coli bacteremia or sepsis who experienced spontaneous abortion. A similar search for an association between Salmonella species and spontaneous abortion yielded 1 report of a study of 30 pregnancies that were complicated by typhoid fever and that resulted in 3 spontaneous abortions (10%) [23].
Although comparative data are limited, an incidence of abortion from 10% to 42% in patients with active brucellosis exceeds that observed in patients infected with other organisms, such as Campylobacter species and Salmonella species, and suggests that Brucella species may indeed produce human abortions more frequently than do other bacterial pathogens. Clearly, there is a need for better evidence to verify this suggestion, which would come from case-control studies; however, the selection of an appropriate control group is problematic, because our understanding of the pathogenesis of spontaneous abortion in women with brucellosis is poor.
Bacteremia is mentioned as a risk factor for spontaneous abortion in the literature, but our study found no association between maternal bacteremic status and spontaneous abortion. Furthermore, although it was not systematically studied, 22 of 28 placental samples that we examined had only nonspecific inflammation. The lack of evidence of uteroplacental involvement in some cases of abortion has been noted by other investigators [7, 14]. Collectively, these data reveal that women with active brucellosis may abort despite the absence of erythritol in the human placenta and fetus, the absence of bacteremia, and the absence of specific histological abnormalities in the placenta. Endotoxin has been implicated in the pathogenesis of brucellosis [24], and it is possible that its promotility effect on uterine smooth muscle plays a role in abortion [25].
An important observation from the present study is that the natural history of brucellosis in pregnant women may be altered by early institution of antimicrobial therapy. When pregnant women with febrile brucellosis are treated with antimicrobial agents before vaginal bleeding occurs, most deliver normally. However, vaginal bleeding at presentation is an ominous sign: it rarely happens in women who receive treatment with antimicrobial agents, and it is associated with a high rate of spontaneous abortion. Whether the low rate of treatment in cases of vaginal bleeding is a marker for failure to diagnose brucellosis and whether it plays a “permissive” role in subsequent abortion or intrauterine fetal death is difficult to determine from our data.
The most important reservation against treating brucellosis in pregnant women is the possibility that the patient would experience adverse effects of the antimicrobial agents [26]. Although the number of patients in this study was not large, we were still able to extract from our findings that cotrimoxazole and rifampin appear to be safe agents for treating brucellosis in pregnant women, because there were no specific drug-related adverse effects in the 36 newborns for whom there were follow-up data. Other researchers have argued that the potential adverse effects of cotrimoxazole in pregnant women may be overstated [27], and that the duration of treatment with rifampin is longer for pregnant women with tuberculosis than it is for pregnant women with brucellosis [28, 29].
In conclusion, this study demonstrated a remarkably high incidence of first and second trimester spontaneous abortions among pregnant women with active brucellosis. Occurrence of abortion was not associated with either the magnitude of the serum agglutinin titer or the presence of brucella bacteremia. However, antimicrobial therapy with either cotrimoxazole or cotrimoxazole plus rifampin had a strong protective effect against abortion. Future controlled studies are required before one can establish the pathogenetic role of brucellosis in causing human spontaneous abortion and whether Brucella species are more likely to cause abortion than are other infectious pathogens. However, even in the absence of such studies, the necessity of treating pregnant patients who have brucellosis immediately is clear, because brucellosis “can cause fetal death at any stage of pregnancy, whether maternal infection is mild or severe. Therefore, treatment must not be withheld” [30].
Acknowledgments
We thank Dr. Frank Kiel for his valued collaboration in the brucellosis research conducted at King Fahad National Guard Hospital (Riyadh, Saudi Arabia).
References
Author notes
Current affiliation: Department of Medicine, Maricopa Medical Center, Phoenix, Arizona.
Current affiliation: Department of Medicine, University of Calgary, Canada.
Comments