Bleich, April T. MD; Sheffield, Jeanne S. MD; Wendel, George D. Jr MD; Sigman, Amy MD; Cunningham, F. Gary MD
The prevalence of infections caused by Neisseria gonorrhoeae has steadily declined in the United States since 1941 when tracking was begun by the Centers for Disease Control and Prevention.1 Despite this rather remarkable decrease, gonorrhea remains the second most common reportable bacterial sexually transmitted disease with an estimated 700,000 new cases annually in this country.2 Although in the majority of adults gonococcal infections are limited to mucosal tissues, in some, bacteremia develops to cause dermatitis, tenosynovitis, migratory polyarthritis, and purulent arthritis. Rarely, other more serious sequelae develop and include endocarditis, meningitis, or osteomyelitis.3–8 There are no recent prevalence studies from which to estimate current rates of disseminated infection that arise from untreated mucosal infections, but in the past, this was reported to be 0.5–3%.9,10 Importantly, N gonorrhoeae remains the most common cause of septic arthritis in sexually active young adults who have no pre-existing joint diseases.11
Disseminated gonococcal infections share most risk factors cited for other sexually transmitted diseases.12–14 One exception is a marked predilection for women who are affected by a 4:1 ratio compared with men. This is attributed to the greater likelihood that women with cervical mucosal infections are asymptomatic and thus do not present for treatment.14 For unknown reasons, women are at highest risk for dissemination during or shortly after menstruation and when pregnant. This has been hypothesized to be related to mucosal vascularity and selection of virulent gonococcal strains.15,16 Although not well documented, antepartum infections with N gonorrhoeae may be associated with increased risks for adverse outcomes such as miscarriage, preterm labor, prematurely ruptured membranes, and perinatal morbidity and mortality.9
Most descriptions of disseminated gonococcal infections were reported during times when untreated asymptomatic gonorrhea was much more prevalent. As mucosal infections became less common, most reports of disseminated infections, especially in pregnant women, included either small series or case reports. These publications along with emergence of increasingly resistant gonococcal strains serve as a reminder that disseminated infection remains a clinical concern.17 For these reasons, our purpose now is to review our experiences over the past 34 years with 112 women admitted to Parkland Hospital for treatment of disseminated gonococcal infection.
MATERIALS AND METHODS
This is an analysis of all women admitted to Parkland Memorial Hospital with a diagnosis of disseminated gonococcal infection from January 1, 1975, through December 31, 2008. They were admitted either to the obstetrics or the medicine services, both of which have a policy mandating admission for intravenous antimicrobial therapy for any patient suspected of having a disseminated gonococcal infection. Pregnant women were admitted to the obstetrics service and one or more of the authors participated in their clinical management. Periodically all Parkland Hospital records were interrogated for International Classification of Diseases, 9th Revision diagnostic codes to identify women admitted with disseminated gonococcal infections. These included a search for International Classification of Diseases, 9th Revision codes for gonococcal septicemia (98.89) or gonococcal infection of the joint (98.5), synovium (98.51), bursa (98.52), meninges (98.82), pericardium (98.83), endocardium (98.84), heart (98.85), skin (98.89), or any other specified site (98.8). These medical records were reviewed and those with a confirmed diagnosis of disseminated infection were entered into a database retrospectively. The study was approved by the institutional review boards of the University of Texas Southwestern Medical Center and Parkland Health and Hospital Systems.
Data abstracted included clinical presentation, laboratory findings, treatment, and hospital course. Symptoms were recorded as to type and duration. Clinical findings recorded included temperature, description of skin lesions, location and number of involved joints, and description of vaginal or cervical discharge, mucopurulent cervicitis, or both. For pregnant women, gestational age at diagnosis was ascertained along with follow-up to record pregnancy outcomes. For nonpregnant women, the time within the menstrual cycle was grouped as: 0–7 days, 8–14 days, 15–21 days, 22–28 days, and more than 28 days. Laboratory data collected included white blood cell counts from peripheral blood and joint aspirates. Specimens taken for identification of N gonorrhoeae by Gram staining and for culture usually included blood, skin lesions, joint aspirate, pharynx, cervix, and rectum. The intravenous antimicrobial regimen and its duration, response to treatment, time to improvement, and outpatient treatment were also recorded.
Disseminated infections in these women were grouped according to conventional terminology.18 Proven infection included those with N gonorrhoeae recovered from cultures or identified by Gram staining of specimens from blood, synovial fluid, skin lesions, or any other sterile site. Probable infection defined those with clinical features of disseminated infection and typical response to antimicrobial treatment and from whom gonococci were identified from cultures of one or more mucosal sites that included the cervix, urethra, anorectum, or pharynx. Possible infection included those with typical clinical features and response to antimicrobial treatment but from whom gonococci were not identified from any specimens.
Statistical analyses were performed using SAS 9.2. These included χ2, Student's t test, and Wilcoxon rank-sum and for all of these; P<.05 was considered significant.
A total of 112 women satisfied criteria for the diagnosis of disseminated gonococcal infection. Figure 1 details the marked decrease in disseminated gonococcal infections both nationally and at our institution. Specifically, during the decade of the 1980s, a total of 48 nonpregnant women were treated, and this contrasted to a total of only 14 admitted during the first decade of 2000. We have also noted a decreased frequency of pregnant women hospitalized for these infections during this same time period. Although this rate was 11 per 100,000 deliveries before 1980, it had decreased to approximately five per 100,000 deliveries by 1985 and since has remained stable.
Selected demographic factors are compared in Table 1. There were 80 (71%) nonpregnant women admitted to the medicine service. Although they were more likely to be diagnosed within the first 7 days of the menstrual cycle (41%), this was not significant (P=.12). There were 32 (29%) pregnant women admitted to the obstetrics service: eight in the first trimester, 13 in the second, and 11 in the third. The only significant difference was that pregnant women had a mean age that was approximately 5 years younger than those not pregnant (P=.007). No woman was human immunodeficiency virus-infected nor was immunosuppressed for another reason.
Shown in Table 2 are some clinical findings in these women at the time of presentation for care. For most of these, there were no significant differences in their incidence when the pregnant cohort was compared with nonpregnant women. One important exception was that pregnant women sought care significantly sooner than those who were not pregnant (median 2 compared with 5 days, P=.003). Various combinations of common presentations are illustrated in Figure 2. Arthralgia of one or more joints was universal, and this was frequently accompanied by chills, fever, cervicitis, or skin lesions such as the one shown in Figure 3.
Joints most commonly involved are shown in Table 2 with a distribution similar for pregnant and nonpregnant women. Findings of a joint effusion was more likely in women in the nonpregnant cohort when compared with pregnant women (70% compared with 50%, P=.05). Moreover, in those undergoing arthrocentesis, the mean leukocyte count in synovial fluid was significantly higher in the nonpregnant women compared with those who were pregnant (43,200 compared with 11,500/microliter, P=.009). One pregnant woman who presented 8 days after the onset of unilateral hip pain and fever had radiographic evidence of articular cartilage destruction, shown in Figure 4. Arthrocentesis confirmed pyoarthrosis and gonococci were recovered from the aspirate.
The various sites from which N gonorrhoeae was recovered by bacteriologic culture are listed in Table 3. With one exception, there were no significant differences in the distribution of culture results between the two cohorts. Specifically, cervical specimens from pregnant women were more likely to be culture-positive for gonococci compared with specimens taken from nonpregnant women (79% compared with 49%, P=.007).
The clinical course and response to intravenous antimicrobial treatment of these women were not significantly different between the pregnant and nonpregnant women. The median duration of fever was 2 days (range, 1–4 days) and 3 days (range, 1–6 days), respectively. The median duration of intravenous antimicrobial treatment in the pregnant women was 3 days (range, 1–9 days) and 4 days (range, 1–14 days) in the nonpregnant women. Finally, clinical improvement was noted in 3 days (range, 1–8 days) in the pregnant cohort and 3 days (range, 1–6 days) in the nonpregnant group.
Perinatal outcomes were available for 28 of the pregnant women. After exclusion of one woman who underwent an elective abortion, the remaining 27 women presented with disseminated infection at a median of 27 weeks (range 6–36 weeks). They were delivered at a median of 40 weeks (range, 27–41 weeks) and their median treatment-to-delivery interval was 9 weeks (range, 0–33 weeks). Two women had spontaneous preterm labor and delivery that followed treatment for disseminated infection at intervals of 7 and 12 weeks. Another was delivered of a 27-week fetus that died after labor induction for severe preeclampsia that developed remote from her gonococcal infection. One fetal death at 31 weeks was identified at the time of admission for disseminated infection. Labor was induced and the cause of death for the 1,595-g fetus could not be identified at autopsy. Specifically, no evidence for chorioamnionitis, funisitis, or fetal infection was found. Finally, there were two women who developed puerperal uterine infection, both remote from the diagnosis of gonococcal infection.
In this report we describe the clinical courses of 112 women admitted to Parkland Hospital for treatment of disseminated gonococcal infections. From these experiences we are able to make a number of observations. First, we confirmed that the frequency of complicated infections caused by N gonorrhoeae decreased markedly and in parallel with the decreasing prevalence of gonorrhea reported for women in the United States during the past 50 years. Second, because one-third of the women now described with disseminated infections were pregnant, we are able to compare their clinical findings and response to antimicrobial treatment with those of women who were not pregnant. Finally, we provide observations concerning any adverse maternal and perinatal outcomes possibly related to gonococcemia and disseminated infection.
The first finding is the markedly decreased frequency of disseminated gonococcal infections in women residing in Dallas County during the years 1975–2008. The policy for mandatory admission for treatment of women with these complicated infections has not changed, the number of women seen in our system has continued to increase, yet the absolute number of women hospitalized for this indication has substantively decreased. As shown in Figure 4, these observed decreases parallel those reported for uncomplicated gonococcal infections by the Centers for Disease Control and Prevention. Most of the women now described had classical findings associated with the gonococcal dermatitis–arthritis syndrome. With one exception, all had joint pain that was attributable to tenosynovitis, arthritis, or both. Although many had migratory polyarthralgia that began with tenosynovitis, this frequently became centered on one or two joints. In some, however, pain began and persisted in the involved joint(s). As shown in Figure 2, arthralgia was usually accompanied by chills and fever, skin lesions, and cervicitis. As previously reported by others, symptoms in nonpregnant women more likely began during the first week of menstruation.11,13,19 In pregnant women, there appeared to be no predilection for a particular gestational age for symptoms to manifest.
Findings associated with these disseminated infections are similar when women in the pregnant and nonpregnant cohorts are compared. For example, other than slightly different mean ages, demographic factors are similar. There is a significant disparity of distribution within the three categories of disseminated infections. In particular, there are significant differences between the two cohorts for infections classified as either probable or possible infections. Both of these depend on recovery of N gonorrhoeae from mucosal sites, which may explain the disparity. It is likely that the higher gonococcal recovery rate from mucosal sites in pregnant women is explicable in that more than half of this cohort was studied during the time in which an obstetrical bacteriology research laboratory was operational. Thus, multiple specimens were taken from each site and especial care given to recovery of the fragile gonococci.
Clinical findings for the two groups are compared in Table 2, and although most are similar, there are important exceptions. The first is that pregnant women sought care a mean of 3 days sooner than nonpregnant women (2 compared with 5 days, P=.003). We believe this is related to the finding that nonpregnant women significantly more often had joint effusions (P=.05) from which the aspirates had a higher mean leukocyte count when compared with pregnant women (P=.009). These observations lend some credence to conventional teaching that disseminated gonococcal infections often have two stages.14,18 In this scheme, there is an initial bacteremic stage characterized by systemic symptoms of the sepsis syndrome, which are accompanied by polyarthralgia from tenosynovitis and hemorrhagic and pustular skin lesions from gonococcemia. This stage may present similarly to other infectious diseases manifesting systemically, although the dermatologic and joint complications are often more severe. In some cases, especially when there is delayed antimicrobial treatment, a septic-joint stage follows in which N gonorrhoeae causes pyoarthrosis and can be recoverable from joint aspirates. Prompt treatment before the septic joint phase will usually prevent joint destruction.
The antimicrobial regimens used over the 3½ decades of this study were those contemporaneously recommended by the Centers for Disease Control and Prevention. With these regimens, women in both cohorts had similarly rapid clinical responses with joint pain and tenderness resolution by a mean of 3 days. Involvement of a weightbearing joint most often accounted for pain of longer duration and two extreme examples are cited. One obese pregnant woman had persistent pain for almost 2 weeks after treatment was begun for gonococcal pyoarthrosis in a knee previously affected by traumatic arthritis. The other was a pregnant woman who had septic arthritis with articular cartilage destruction at presentation. There was no instance in which antimicrobial resistance was shown to cause a delayed clinical response.
There was no evidence that these disseminated infections with N gonorrhoeae had any deleterious effects on pregnancy outcomes. Although four women were delivered preterm, in only one was delivery temporally related to disseminated infection. She presented for care at 31 weeks of gestation with a dead fetus and with symptoms of several days' duration. Induction and delivery were accomplished and there was no evidence of fetal or placental infection found at autopsy. These observations are consistent with recent reports that disseminated gonococcal infections are seldom associated with adverse neonatal outcomes.3,6,20–22
We conclude that although rates of uncomplicated mucosal infections with N gonorrhoeae have decreased dramatically in this country, the risk for disseminated infection persists. This report serves as a reminder that gonococcal infection remains at the forefront in the differential diagnosis for any sexually active woman with septic arthritis. With prompt recognition and treatment, the more serious complications of gonococcemia can be prevented and a favorable outcome anticipated.
1. Centers for Disease Control and Prevention. Sexually transmitted disease surveillance 2009. Atlanta (GA): US Department of Health and Human Services; 2010.
2. Centers for Disease Control and Prevention. Sexually transmitted diseases treatment guidelines, 2010. Atlanta (GA): US Department of Health and Human Services; 2010.
3. Burstein H, Sampson MB, Kohler JP, Levitsky S. Gonococcal endocarditis during pregnancy: simultaneous cesarean section and aortic valve surgery. Obstet Gynecol 1985; 66 (suppl): 48S–51S.
4. Burgis JT, Nawaz H 3rd. Disseminated gonococcal infection in pregnancy presenting as meningitis and dermatitis. Obstet Gynecol 2006; 108: 798–801.
5. Mofredj A, Baraka D, Madec Y, Lemaitre P. Disseminated gonococcal infection and meningitis. Am J Med 2000; 109: 71–2.
6. Akkinepally S, Douglass E, Moreno A. Tricuspid valve gonococcal endocarditis: fourth case report. Int J Infect Dis 2010; 14 (suppl 3): e196–7.
7. Jain S, Win HN, Chalam V, Yee L. Disseminated gonococcal infection presenting as vasculitis: a case report. J Clin Pathol 2007; 60: 90–1.
8. Angevine CD, Hall CB, Jacox RF. A case of gonococcal osteomyelitis. A complication of gonococcal arthritis. Am J Dis Child 1976; 130: 1013–4.
9. Handsfield HH, Sparling PF. Neisseria gonorrhea. In: Mandell GL, Bennett JE, Dolin R editors. Principles and practice of infectious diseases. 6th ed. Philadelphia (PA): Churchill Livingstone; 2005. p.2498–529.
10. Moran JS. Gonorrhoea. Clin Evid (Online) 2007; pii: 1604.
11. Rice PA. Gonococcal arthritis (disseminated gonococcal infection). Infect Dis Clin North Am 2005; 19: 853–61.
12. Ohl CA. Bone and joint infections. In: Mandell GL, Bennett JE, Dolin R editors. Principles and practice of infectious diseases. 6th ed. Philadelphia (PA): Churchill Livingstone; 2005. p. 1311–22.
13. Shirtliff ME, Mader JT. Acute septic arthritis. Clin Microbiol Rev 2002; 15: 527–44.
14. Holmes KK, Counts GW, Beaty HN. Disseminated gonococcal infection. Ann Intern Med 1971; 74: 979–93.
15. O'Brien JP, Goldenberg DL, Rice PA. Disseminated gonococcal infection: a prospective analysis of 49 patients and a review of pathophysiology and immune mechanisms. Medicine (Baltimore) 1983; 62: 395–406.
16. Brown D. Gonococcal arthritis in pregnancy. South Med J 1973; 66: 693–8.
17. Centers for Disease Control and Prevention. Consultation on cephalosporin-resistant N. gonorrhoeae outbreak response plan. Atlanta (GA): US Department of Health and Human Services; 2011.
18. Hook EW III, Handsfield HH. In: Holmes KK, Sparling PF, Stamm WE, Piot P, Wasserheit JN, Corey L, et al. editors. Sexually transmitted diseases. 4th ed. New York (NY): McGraw Hill; 2008. p. 634–5.
19. Al-Suleiman SA, Grimes EM, Jonas HS. Disseminated gonococcal infections. Obstet Gynecol 1983; 61: 48–51.
20. O'Leary AJ, Tejura H, Latibeaudiere M, Edwards G. Gonorrhoea infection presenting in pregnancy with septic arthritis of the sternoclavicular joint. J Obstet Gynaecol 2006; 26: 373–4.
21. Mahendran SM. Disseminated gonococcal infection presenting as cutaneous lesions in pregnancy. J Obstet Gynaecol 2007; 27: 617–8.
22. Phupong V, Sittisomwong T, Wisawasukmongchol W. Disseminated gonococcal infection during pregnancy. Arch Gynecol Obstet 2005; 273: 185–6.