The Emergence of Neisseria gonorrhoeae with Decreased Susceptibility to Ciprofloxacin in Cleveland, Ohio: Epidemiology and Risk Factors

Methods

This study was done at a public clinic for sexually transmitted diseases in Cleveland, Ohio (clinic X), that has participated in GISP since 1991. Clinic X serves a county with a population of 500 000 persons on a walk-in basis 5 days per week, and it receives approximately 5000 visits annually. One laboratory technician immediately read Gram stains of all urethral discharges to assess the presence of polymorphonuclear leukocytes and gram-negative diplococci. From 1986 through 1993, primary treatment for uncomplicated gonorrhea at clinic X was ceftriaxone, 250 mg administered intramuscularly. In 1994, the dose of ceftriaxone was reduced to 125 mg. Spectinomycin, 2.0 g administered intramuscularly, was used as primary therapy for gonorrhea in patients who had a history of allergy to penicillin. All patients suspected of having gonorrhea also received a 7-day course of doxycycline, 100 mg twice daily. Fluoroquinolones have never been used to treat gonorrhea in clinic X.

Urethral isolates from the first 20 men with gonococcal infections who attended clinic X were submitted each month to the regional GISP laboratory for susceptibility testing. Isolates were collected from the first 25 men; isolates 21 through 25 were used to replace any isolates missing from the sequence of isolates 1 through 20. The Section of Microbiology at the Cleveland Clinic Foundation (Cleveland, Ohio) has served as the regional laboratory for clinic X since September 1991.

Definition of Case-Patients and Ascertainment and Selection of Controls

A case-patient was defined as any male patient who attended clinic X between January 1992 and March 1994 (the study period) and had a urethral N. gonorrhoeae isolate that was submitted to GISP and showed decreased susceptibility to ciprofloxacin (mean inhibitory concentration ≥ 0.12 µg/mL and ≤ 0.25 µg/mL). Controls were selected from among male patients who had attended clinic X and who had a urethral N. gonorrhoeae isolate that was submitted to GISP and showed susceptibility to ciprofloxacin (MIC ≤ 0.06 µg/mL).

The records at clinic X for 51 case-patients and 106 controls were abstracted by two investigators. Abstracted information included demographic data, the reason for visiting the clinic, history of sexually transmitted diseases, sexual history, results of clinical examination and assessment, and results of serum rapid plasma reagent assays. Results of serologic tests for human immunodeficiency virus (HIV) were kept in separate medical records that were not reviewed.

Laboratory Methods

The urethral specimens were plated on selective Martin-Lewis agar and processed; this processing included the identification of gonococci, testing for β-lactamase, and the recording and assignment of a GISP number. All isolates were identified as N. gonorrhoeae by established methods [9]. Gonococcal isolates were subcultured from the selective primary medium to a chocolate agar with 1% IsoVitaleX (Becton Dickinson, Sparks, Maryland) supplement. Organisms from the subculture were suspended in tryptic soy broth that contained 20% glycerol. Isolates were then frozen at −20°C and shipped on dry ice to the microbiology laboratory at the Cleveland Clinic Foundation, where they were kept frozen at −70°C. The laboratory determined MICs by using the agar dilution methods recommended by the National Committee for Clinical Laboratory Standards (NCCLS) with GC II agar base (Becton Dickinson) and 1% IsoVitaleX (Becton Dickinson) supplement [10]. The isolates were tested against penicillin, tetracycline, spectinomycin, ceftriaxone, cefixime, ciprofloxacin, erythromycin, and azithromycin using standard sterile powders supplied by the manufacturers.

Susceptibilities were interpreted according to the recommendations of the NCCLS, which define gonococcal isolates as susceptible to ciprofloxacin if the MIC of ciprofloxacin is 0.06 µg/mL or less. In our study, isolates were defined as having decreased susceptibility to ciprofloxacin if the MIC was 0.125 µg/mL or more. The CDC [11] recently proposed interpretative criteria for defining an isolate as resistant to ciprofloxacin if the MIC is 1.0 µg/mL or more.

Culture and DNA Preparation for Pulsed-Field Gel Electrophoresis

The methods used were modifications of those previously published [12]. Isolates were stored at −70°C until needed and were then plated twice onto chocolate agar (Becton Dickinson) at 36.5 °C in 5% CO₂. Bacterial growth 18 to 22 hours after the second chocolate agar transfer was inoculated into sterile saline to a density approximately equal to McFarland number 4, sonicated, and centrifuged. The pellet was resuspended in 1 mL of cold cell-suspension buffer (Bio-Rad, Hercules, California).

The pellets were molded in gel plugs as recommended by the manufacturer of the Gene-Path reagent kit (Bio-Rad). Solidified blocks were treated with lysozyme and incubated for 1 hour to release nucleic acid. Plugs were washed with buffer for 30 minutes and then soaked in buffer that contained 4 mg of phenylmethyl-sulfonyl fluoride per mL (Sigma, St. Louis, Missouri) at room temperature for 1 hour. This was followed by three washes with buffer at room temperature.

Pulsed-Field Gel Electrophoresis Assay

Gel plugs were digested with 8.3 µL of Spe I (a restriction enzyme with a rare recognition site) overnight in 300 µL of restriction buffer. The digested DNA plugs were placed in wells that contained 1% agarose gel with 1% low-molecular-weight protein agarose. Electrophoresis was done on the digested DNA plugs in a contour-clamped homogenous electric field apparatus (Gene-Path, Bio-Rad). Gels were stained with ethidium bromide and photographed under ultraviolet light. Bands of 15 to 500 kb in 12 to 15 fragments were produced per organism.

Chromosomal DNA restriction patterns produced by pulsed-field gel electrophoresis were interpreted on the basis of criteria proposed by Tenover and colleagues [13]. If the pulsed-field gel restriction patterns were identical, the isolates were considered to be indistinguishable; if the patterns differed by 1 to 3 bands, the isolates were considered to be closely related; and if the patterns differed by more than 3 bands, the isolates were considered to be unrelated.

Expanded Surveillance for Fluoroquinolone-Resistant Neisseria gonorrhoeae in Cleveland

A point prevalence survey was done in the summer of 1994 to determine the prevalence of ciprofloxacin-resistant N. gonorrhoeae urethral isolates obtained from patients who attended two public clinics for sexually transmitted diseases (other than clinic X) in Cleveland (25 isolates) and from patients treated at the Cleveland Clinic Foundation (29 isolates).

Statistical Analysis

All data were entered into a computer database for analysis (Epi Info, version 5; CDC, Atlanta, Georgia). The Student t-test and the Fisher two-tailed exact test were used for univariate analyses of the significance of associations. Multivariable logistic regression was done using PC SAS/STAT (SAS Institute, Cary, North Carolina). Differences were considered statistically significant at a P value of 0.05 or less.

Results

From September 1991 through September 1995, 1378 isolates of N. gonorrhoeae from clinic X were tested; 140 (10.4%) had decreased susceptibility to ciprofloxacin. The incidence of gonococcal infections with these strains increased significantly during this period, from 2% in 1991 to 16% in 1994 (Figure 1). During the study period (1 January 1992 to 31 March 1994), 746 isolates of N. gonorrhoeae from patients in the GISP sample from clinic X were tested; 55 isolates (7.4%) had decreased susceptibility to ciprofloxacin.

Description of Case-Patients

The clinical records of 51 of the 55 patients whose N. gonorrhoeae isolates had decreased susceptibility to ciprofloxacin were available for review and were included in the epidemiologic study. All patients were black and heterosexual men, and most were unmarried (94%). The mean age was 28.7 years (range, 14 to 63 years). All resided in the greater Cleveland metropolitan area; 72% (37 of 51) had one of the six ZIP codes belonging to the area surrounding clinic X. A total of 96% of patients (49 of 51) were symptomatic volunteers who had a history of urethral discharge (92%) or dysuria or both (61% either had a history of urethral discharge and dysuria or a history of dysuria alone) and had urethral discharge (92%) on examination. Only 2 patients (4%) reported having used an antimicrobial agent within the preceding 14 days. The mean interval from the last reported sexual activity to the time of the clinic visit was 7.5 days (range, 1 to 90 days). Patients reported having had a mean of 1.5 sexual partners in the 30 days before their clinic visit and a median of 20 sexual partners in their lifetime (range, 3 to 500 sexual partners). Sixty-three percent (32 of 51) had a history of sexually transmitted diseases, including gonorrhea (57% [29 of 51]) or nongonococcal urethritis or both (10% [5 of 51]) had gonorrhea and nongonococcal urethritis or nongonococcal urethritis alone). Among patients who had previously had episodes of gonorrhea, 20 reported having had a single episode and 9 reported having had at least two episodes.

Case-Control Study

Case-patients were significantly less likely to have gram-negative diplococci seen on an examination of urethral discharge than were controls (42 of 51 case-patients compared with 101 of 106 controls had gram-positive stains [P ≤ 0.01]; odds ratio, 4.3 [95% CI, 1.2 to 16]) (Table 1). Case-patients were also significantly less likely than controls to have been treated for gonorrhea at the time of their clinic visit (12 of 51 case-patients compared with 5 of 106 controls were not treated [P < 0.001]; odds ratio, 6.2 [CI, 1.9 to 22]). When a logistic regression model was used, the Gram stain result (the absence of intracellular or extracellular organisms) was the only independent risk factor for resistant N. gonorrhoeae (odds ratio, 6.6 [CI, 1.2 to 36]; P < 0.03).

The participants in the two groups did not differ significantly in distribution of ZIP codes, presenting signs and symptoms of urethritis, number of polymorphonuclear leukocytes on Gram stains of urethral exudate, history of sexually transmitted diseases, or number of sexual partners. The prevalences of a reactive serum rapid plasma reagent assay, cocaine use, injection drug use, history of exchanging sex for drugs, or use of antibiotics within 14 days of the clinic visit were low in the study sample and did not differ between case-patients and controls.

The 17 participants (12 case-patients and 5 controls) who were not treated for gonorrhea at the time of their clinic visit were not clustered in time. Four participants (2 case-patients and 2 controls) had intracellular gram-negative diplococci on urethral smears, and 3 participants (2 case-patients and 1 control) had extracellular gram-negative diplococci on urethral smears. Notably, only 2 of the patients (12%) with gonococcal infections who did not receive appropriate therapy at the time of their initial visit to clinic X were successfully contacted and later treated for gonorrhea.

In Vitro Studies of the Susceptibility of Neisseria gonorrhoeae Isolates

The susceptibilities of N. gonorrhoeae isolated from the 51 case-patients and 106 controls are shown in Table 2. The mean inhibitory concentrations for ciprofloxacin were 0.12 µg/mL or greater and 0.25 µg/mL or less for all strains of N. gonorrhoeae from the 51 case-patients, and none of the isolates that had decreased susceptibility to ciprofloxacin were susceptible to penicillin. Three isolates that were susceptible to ciprofloxacin were negative for β-lactamase. None of the isolates that had diminished susceptibility to ciprofloxacin were susceptible to tetracycline. The mean inhibitory concentrations for tetracycline were in the intermediate category, with the exception of that for one resistant isolate that had acquired the 25.2-Md tetracycline-resistant plasmid containing the TetM determinant [7]. All isolates were susceptible to ceftriaxone, spectinomycin, and cefixime.

Pulsed-Field Gel Electrophoresis

Digestion of the N. gonorrhoeae isolates from the controls and case-patients resulted in different patterns on pulsed-field gel electrophoresis for the two groups. Within the ciprofloxacin-resistant isolates, however, patterns differed by only one or two fragments (data not shown).

Expanded Surveillance for Neisseria gonorrhoeae with Decreased Susceptibility to Ciprofloxacin

The prevalences of N. gonorrhoeae with decreased susceptibility to ciprofloxacin were 20% (5 of 25 isolates tested) at the two public clinics in Cleveland and 14% (4 of 29 isolates tested) at the Cleveland Clinic.

Discussion

Recent results from the Gonococcal Isolate Surveillance Project [14] suggest that resistance to fluoroquinolone is not widespread in the United States. In 1994, only 67 of 4996 isolates (1.3%) showed decreased susceptibility to ciprofloxacin, and only two strains (0.04%) required MICs of 1 µg/mL or more. To date, Ohio is the only geographic area in the United States in which the persistence of N. gonorrhoeae isolates with decreased susceptibility to ciprofloxacin has been documented, and Cleveland appears to be the epicenter. Isolates showing decreased susceptibility to ciprofloxacin were not only recovered repeatedly from patients attending clinic X but were recovered with a frequency that increased significantly during a 48-month period (Figure 1). Molecular typing done using pulsed-field gel electrophoresis suggested that a single strain or closely related strains of N. gonorrhoeae accounted for the decreased susceptibility to ciprofloxacin, which was also suggested by auxotype-serotype analysis, as reported elsewhere [7].

Our findings suggest the introduction and endemic spread of these isolates, but the origin of these strains in Ohio remains unclear. Our study was limited because it was retrospective and examined a relatively small number of cases. We could not directly assess specific factors, such as previous exposure to fluoroquinolones or sexual activity with partners from outside of the Cleveland area. However, our findings suggest at least one possible explanation for why these strains have become endemic in Cleveland. At clinic X, 25% of the men (12 of 51) who had gonococcal urethritis with organisms that showed decreased susceptibility to ciprofloxacin were not treated for gonorrhea when they visited the clinic (5% of controls [5 of 106] with ciprofloxacin-susceptible isolates were also not treated). Most of these patients were thought to have nongonococcal urethritis at the time of their clinic visit, and thus they received doxycycline therapy alone. The absence of gram-negative diplococci on examination of urethral discharge was significantly associated with infection with N. gonorrhoeae isolates that had decreased susceptibility to ciprofloxacin. Notably, only 12% of study participants (2 of 17) who were discharged from their initial clinic visits with untreated gonorrhea were subsequently contacted and adequately treated.

Studies have shown that resistance to fluoroquinolones develops during therapy with these agents as well as through the in vitro selection of mutants by stepwise exposure to increasing concentrations of quinolones [15]. Although the use of antimicrobial agents is considered to be a predisposing factor for the emergence and spread of resistant bacteria, we were unable to directly investigate the relation between the use of fluoroquinolones and the development of N. gonorrhoeae infection with decreased susceptibility to fluoroquinolones. It is possible that quinolones may have been used before the 2-week period that preceded the clinic visit, because self-medication is common among attendees of sexually transmitted disease clinics and may not be reported [16]. However, none of the public clinics for sexually transmitted diseases in Cleveland (nor any of the other sexually transmitted disease clinics in Ohio from which we received isolates of N. gonorrhoeae that had decreased susceptibility to ciprofloxacin) used quinolones in the primary treatment of gonorrhea.

The relation between diminished susceptibility to ciprofloxacin in vitro and the clinical efficacy of fluoroquinolone therapy has not been determined. Serum levels of ciprofloxacin achieved with currently recommended dosages exceed the MICs for isolates from Cleveland, suggesting that infection with these isolates should respond to therapy with ciprofloxacin [17]. However, we could not assess the clinical significance of these infections because no patients were treated with quinolones. Several reports have described clinical failures in patients infected with ciprofloxacin-resistant isolates of N. gonorrhoeae (mean inhibitory concentration ≥ 1 µg/mL) who received ciprofloxacin or other fluoroquinolones as definitive therapy [18-20]. One report of unsuccessful therapy with ciprofloxacin, 250 mg, in a patient infected with an isolate that had decreased susceptibility to ciprofloxacin (mean inhibitory concentration ≥ 0.125 and ≤ 0.25 µg/mL) has been published [21]. The persistence of isolates with decreased susceptibility to quinolones may be a harbinger of the emergence of clinically significant resistance in the future, which could limit the usefulness of these agents for the primary treatment of gonorrhea in the United States. This point has been underscored by the recent isolation of resistant strains of N. gonorrhoeae in Colorado and Washington and by the first documentation in the United States of a gonococcal infection (acquired in the Philippines) caused by an isolate strain with in vivo resistance to fluoroquinolone therapy [14].

In conclusion, resistance of N. gonorrhoeae to antibiotics continues to be an emerging and expensive public health problem. Gonococcal isolates with decreased susceptibility to ciprofloxacin now seem to be endemic in Cleveland and possibly in other areas of Ohio. The appearance of such isolates provides the potential for the emergence of clinically important resistance that could preclude the future empirical use of ciprofloxacin or ofloxacin as alternative therapies for uncomplicated gonorrhea. This report emphasizes the need to monitor the susceptibility of gonococcal isolates to fluoroquinolones and should heighten awareness about the potential for the emergence of widespread, clinically important resistance.

Presented in part at the 11th Meeting of the International Society for Sexually Transmitted Disease Research held in New Orleans, Louisiana, on 27 to 30 August 1995.

Dr. Carlyn: 739 Thimble Shoals Boulevard, Suite 101, Newport News, VA 2360.

Drs. Washington and Hall, Ms. Doyle, and Ms. Knapp: Department of Pathology and Laboratory Medicine, Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195.