SSuN Cycle II was a sentinel sexually transmitted disease (STD) surveillance project funded by the CDC from 2008 to 2012 (CDC, PS08–865), through which 12 sites (local and state health jurisdictions) collected population-based data on a randomly selected sample of patients reported with gonorrhea. This cycle of SSuN included 5 city (Baltimore, Chicago, New York City, Philadelphia, and San Francisco) and 7 state health departments (Alabama, California, Colorado, Connecticut, Virginia, and Washington). Data were collected through laboratory and/or provider case reports and interviews with patients and were transmitted to CDC on a quarterly basis by collaborating health departments.

STD Surveillance Network sites each collected enhanced data through interviews from a random sample of 240 or more gonorrhea patients annually per site. Gonorrhea is a reportable infection in all US states. Cases were defined based on criteria established by the Council for State and Territorial Epidemiologists. Patients diagnosed as having multiple anatomical sites of infection were counted as 1 case, and persons diagnosed as having multiple episodes of gonorrhea during the study period may have been contacted and interviewed multiple times. Cases reported more than 60 days after diagnosis or those reported from other jurisdictions were excluded before sampling. The sample fraction varied across sites, with some sites contributing substantially more cases than 240 cases per year, depending on the overall case burden and available local resources.

Data collected included information on diagnosis and treatment, symptoms of gonorrhea, type of provider, demographics, education, and a range of sexual behavior measures, including sex of sex partners. Although treatment data were collected, the information did not include dosage and was limited to the specific drugs prescribed or administered. Data elements were collected through case reports in some sites and via patient interviews in others. Patients in the random sample were contacted and interviewed using standardized interview questions at all sites. In some jurisdictions, patients are contacted and interviewed by local health department Disease Intervention Specialists, whereas in other areas, SSuN-specific interviewers were hired to conduct interviews. Similarly, interview methods vary by site, with some areas conducting patient interviews primarily by telephone, whereas in other areas, interviewers conducted field visits and completed many interviews face-to-face. No interviews were conducted through computer-assisted self-interview.

Provider type was collapsed from 13 possible categories into the following 7: STD clinic, hospital/emergency department, family planning and reproductive health, private provider, public clinic, other provider, and missing provider type. Health department–funded clinics that follow local STD clinic protocols for STD treatment and serve primarily lesbian, gay, bisexual, transgender, and questioning populations were included in the STD clinic provider type for this analysis. Similarly, cases diagnosed through outreach efforts of local health departments and therefore referred to the local STD clinic for treatment were also grouped into the STD clinic category.

Interview response rates varied greatly among SSuN sites. We restricted analyses for this study to the 6 sites (Alabama, Baltimore, California, Colorado, Washington, and San Francisco), which met the following criteria: (1) interview response rate greater than 30%, (2) 100 or more gonorrhea cases reported per year, and (3) fewer than 30% of cases missing treatment information. Transgendered people accounted for approximately 0.06% of interviewed cases and were excluded from analyses.

The study included data for cases reported during the period January 1, 2010, through December 31, 2012. We developed stratification weights to account for varying sampling fractions across sites by year, as well as poststratification weights to adjust for interview nonresponse based on sex, age group, and provider type using the 7 collapsed categories noted above. We produced weighted estimates of the proportion of patients treated with ceftriaxone and other gonorrhea therapies using weighted survey procedures available in SAS 9.3 (SAS Institute Inc, Cary, NC), including Wald χ² tests to test for differences in ceftriaxone treatment. We used Poisson regression, including site as covariate and weighted for nonresponse and sampling fractions, to produce prevalence ratios (PRs) and 95% confidence intervals (CIs) evaluating the relationship between ceftriaxone use and sex, sex of sex partners, race, ethnicity, age, provider type, and SSuN site.²¹ To examine changes in gonorrhea treatment practices over the study period, all analyses examining gonorrhea treatment practices are stratified by year of gonorrhea report.

The protocol for enhanced population-based gonorrhea surveillance through SSuN received a determination of nonresearch from CDC as activities conducted in the process of routine disease surveillance by state and local governments; no human subjects research approval was necessary for the current analysis. Instruments used in data collection were approved for local use (OMB Control Number: 0920–0842).

From January 1, 2010, through December 31, 2012, 135,984 gonorrhea cases from 101 US counties were reported to the 6 sites included in our study. Of these, 15,246 (11.2%) were randomly sampled for interview, and 7851 (51.5%) of these patients were interviewed. The number of cases reported from and interviewed in each site varied substantially (Table 1). Sex of sex partners is not routinely collected via the case report in several sites; therefore, these data were missing for many noninterviewed patients. Interview response rates also varied by site, ranging from 37.1% to 79.6%.

Among interviewed patients, 4471 (57.0%) were men, and among the men, 1587 (35.5%) reported having male sex partners in the past 2 to 3 months, 2065 (46.2%) had only female sex partners, and sex partner sex was unknown for the remaining 819 (18.3%; Table 1). All jurisdictions except Washington used a 3-month time frame; Washington collected data about partner sex in the prior 2 months. Approximately one-third of the cases were reported during each of the 3 years in the study period. More interviewed patients were diagnosed in STD clinics (22.9%) than in any other provider type group (Table 1).

Treatment data were missing for 487 (6.2%) patients, and 124 (1.6%) of patients were untreated at the time of interview; these patients were excluded from further analysis, leaving 7240 patients with complete treatment data. After weighting the data for interview nonresponse and different sampling fractions across sites, 76.8% (95% CI, 73.3%–80.3%) of these patients had been treated with ceftriaxone, either alone or in combination with another medication. Overall, 44.9% of persons received dual therapy including ceftriaxone and azithromycin, 24.8% were treated with ceftriaxone alone, 7.0% received ceftriaxone and doxycycline, 9% received an oral cephalosporin alone, 7.4% received an oral cephalosporin in combination with azithromycin or doxycycline, and 6.9% received another regimen, primarily azithromycin (Table 2). We also examined the proportion of patients receiving any cephalosporin; 93.1% (95% CI, 92.0%–94.2%) received a cephalosporin and 59.3% (95% CI, 56.0%–62.5%) received any cephalosporin in combination with azithromycin or doxycycline (data not shown). Ceftriaxone treatment increased each year during the study period, from 64.1% of cases reported in 2010 to 85.4% in 2012 (P = 0.0001; Table 3). The use of dual therapy, including ceftriaxone and either azithromycin or doxycycline, increased from 34.2% of patients in 2010 to 64.6% in 2012 (Table 2). Although the frequency of ceftriaxone use varied widely by SSuN site, the observed increase in ceftriaxone treatment over time was consistent across all SSuN sites, except one (Fig. 1).

Given the observed increase in ceftriaxone use over time, remaining analyses were stratified by year of gonorrhea report. Including only patients with complete treatment data, all covariates examined in bivariate analysis, except age, were associated with ceftriaxone therapy in 2010, but in 2012, only sex of sex partners, race, and patient age were related to receipt of ceftriaxone (Table 3). There was a wide variation in the rate at which ceftriaxone use increased across each covariate. For example, among family planning clinic patients, ceftriaxone treatment increased from 41.4% in 2010 to 74.2% in 2011 and then to 91.8% in 2012, whereas the proportionate increase was smaller in STD clinics, where ceftriaxone already was used in 74.6% of cases in 2010. Across the entire study period, factors significantly associated with receipt of ceftriaxone treatment included being male (P = 0.008), being a man who has sex with men (MSM; vs. heterosexual men or men with missing data on sex of sex partners; P < 0.0001), being of native American and Asian/Pacific Islander race (P = 0.0002), Hispanic ethnicity (P = 0.0009), and older age (P = 0.01; Table 3).

In models adjusted for sex, sex of sex partners, race, age, ethnicity, provider type, and SSuN site, MSM patients were more likely than women to be treated with ceftriaxone in 2010 to 2011; however, MSM were less likely than women to be treated with ceftriaxone in 2012 (Table 4). In contrast, men who have sex with women were less likely than women to be treated with ceftriaxone in 2011 and 2012 (Table 4). Differences by provider type were large in 2010, with most provider types being associated with less ceftriaxone treatment when compared with STD clinics. However, these differences were smaller in 2012, and patients diagnosed in family planning, public clinics, and STD clinic in 2012 were equally likely to be treated with ceftriaxone. Substantial changes were also observed over time across SSuN sites. In 2010, compared with the reference site, Colorado, ceftriaxone use was statistically significantly higher in 3 other sites. In contrast, in 2011 and 2012, ceftriaxone use was higher in all of the SSuN sites when compared with the reference site.

We found that 76.8% of randomly sampled gonorrhea cases from geographically diverse areas of the United States were treated with the currently recommended primary drug (ceftriaxone) and that approximately half of all patients were treated with dual therapy including ceftriaxone in combination with azithromycin or doxycycline. In addition, the percentage of patients receiving dual therapy increased from 34.3% of patients in 2010 to 64.6% in 2012, and clinicians’ use of ceftriaxone significantly increased after changes in CDC STD treatment guidelines in 2010. Although we observed differences in ceftriaxone use by patient sex, sex of sex partners, provider type, and site, many of these differences were attenuated over time as ceftriaxone use became more widespread. Overall, we believe that these findings suggest that changes in CDC treatment guidelines can lead to large and relatively rapid changes in clinician’s medical practice.

Our conclusion is consistent with recent experience in the United Kingdom. Gonorrhea treatment guidelines changed in the United Kingdom in 2011 to recommend that all persons with gonorrhea receive ceftriaxone and azithromycin.²² Among patients treated in genitourinary clinics participating in the UK Gonococcal Resistance to Antimicrobials Program, treatment with ceftriaxone increased from 23% in 2007 to 93% in 2011, and the proportion of patients treated with both ceftriaxone and azithromycin increased from 9% to 79%.²³

In the same study, Ison and colleagues²³ observed a decline in the proportion of gonococcal isolates with decreased susceptibility to cefixime from 2010 to 2011, the period during which both ceftriaxone and dual therapy of ceftriaxone with azithromycin or doxycycline grew most rapidly. This follows a period during which reports of gonococcal isolates obtained from genitourinary clinics with decreased minimum inhibitory concentrations to cefixime were relatively common.^{5,7,11,14,24–27} In 2012, the proportion of isolates with decreased susceptibility to cefixime decreased in the Gonococcal Isolate Surveillance Project, a sentinel surveillance project in the United States, from 1.4% in 2010 and 2011 to 1.0% in 2012.²⁸ Although the declines in isolates with decreased susceptibility to cefixime in Gonococcal Resistance to Antimicrobials Program and Gonococcal Isolate Surveillance Project cannot be causally attributed to changes in treatment practices, they are consistent with the hypothesis that widespread use of highly active treatment may reduce selection pressure for antimicrobial resistance in N. gonorrhoeae. Similar conclusions have been drawn by Evans et al.²⁹ and Olsen et al.³⁰

Previous studies have found both high levels of adherence to new local gonorrhea treatment guidelines³¹ and variation across sites in the uptake of revised national treatment guidelines.³² Dowell et al.³² reviewed communications from local health departments to medical providers about the national update and found that both the timing and format of these communications varied widely. Although we did not complete a similar review of communications, our results demonstrate that responses to national changes in treatment guidelines continue to be somewhat heterogeneous and may be influenced by local program response.

Our findings highlight both the limitations of routine STD surveillance and the value of the supplementary surveillance undertaken as part of SSuN. Treatment information is not routinely reported to state and local health departments as a part of STD case reporting in many US states, making it difficult to evaluate the degree to which US providers follow recommended guidelines. Our findings demonstrate that purposeful sampling of cases and weighting of data are feasible and can be used to monitor the adequacy of treatment and changes in provider practices in the face of new guideline recommendations.

Our study has several limitations. First, although our data were collected in geographically diverse sites, they may not be generalizable to gonorrhea treatment practices in other health jurisdictions. Second, the interview response rate in some sites was less than 50%, raising concerns regarding nonresponse bias. To address this issue, we completed an additional analysis in which we restricted our data to the 3 sites with the highest response rates and found that our results did not change (data not shown). We used design weights to adjust for different sample fractions across sites and poststratification weighting to adjust for differences in interview response versus all cases reported from the participating jurisdictions; however, there may be minor sampling bias because of differing sampling methods used in SSuN sites. Although information on patient self-report of pelvic inflammatory disease (PID) diagnoses is available in SSuN interview data, other complications of gonorrhea such as epididymitis or disseminated gonococcal infection were not collected. We did not exclude the small number of women self-reporting PID because there was no way to assess the validity of those diagnoses. Treatment regimens may differ for these patients from those with uncomplicated gonorrhea; our inclusion of patients with PID and other unreported complications may have led to a slight underestimation of the true portion of patients with uncomplicated infections receiving ceftriaxone. Finally, 7.8% of interviewed gonorrhea cases were missing treatment data. Some of these cases were almost certainly treated; treatment data may have been missing because the provider did not include the treatment regimen on the case report form, or the patient could not recall the medication they received. Our results may be biased if a large or very small proportion of these patients actually received ceftriaxone.

As evidence accumulates regarding the susceptibility of N. gonorrhoeae to oral cephalosporins and with no other class of drugs currently recommended to treat gonococcal infections, it becomes increasingly important to monitor gonorrhea treatment practices to ensure that medical providers use the most effective therapies available. Prevention and control of gonorrhea are the first line of public health defense against cephalosporin-resistant gonorrhea. Although we cannot definitively attribute the recent declines observed in gonococcal isolates with decreased susceptibility to cefixime to changes in gonorrhea treatment practices, it is plausible that increased use of ceftriaxone could contribute the slowing of the development of resistance to cephalosporins, which would in effect extend the period during which this class of antimicrobials remains effective against N. gonorrhoeae. Our findings suggest that medical providers are indeed adapting to changes in treatment recommendations and using ceftriaxone as their first-line treatment of gonorrhea. At the same time, some local health jurisdictions continue to use oral cephalosporins substantially more frequently, and in 2012, approximately one-third of patients did not receive recommended dual therapy. Our findings highlight the need to promote greater adherence with treatment guidelines, understand instances in which guidelines are not being followed, and sustain and improve surveillance of representative STD cases with the goal of monitoring the populations affected by gonorrhea and the treatment of this common infection in an era of increasing antimicrobial resistance.