Introduction
Methods
Results
Discussion
Limitations
Conclusion

INTRODUCTION

Urinary tract infection (UTI) is one of the most common bacterial infections diagnosed in the emergency department (ED).^1,2 Symptoms of UTI are the reason for approximately 1% of all ambulatory visits and result in 2–3 million ED visits in the US each year.² Urine culture results can take more than a day, and the urinalysis findings can cause diagnostic uncertainty about the existence of a bacterial UTI. Adding to the problem is that the incidence of some sexually transmitted infections (STI) such as gonorrhea, chlamydia, and trichomoniasis is increasing in the US,^3,4 and clinical manifestations of UTIs and STIs may overlap. These overlapping signs and symptoms may lead to underdiagnosing STIs in patients with urinary concerns and overtreating for STIs in patients with genital concerns.^5–9 Previous study findings have shown that STIs are associated with sterile pyuria and other non-specific findings on urinalysis.^5,9,10 Diagnostic confusion may be most common when trichomoniasis is identified in the ED by urinalysis or wet mount and the clinician must consider whether the urine inflammatory changes are caused by Trichomonas vaginalis only or by a concurrent bacterial UTI.

In this analysis, we sought to determine whether infection with gonorrhea, chlamydia, and trichomoniasis was associated with specific urine culture results. Specifically, we attempted to determine the frequency of STIs and having a urine culture yield 10,000 or more colony-forming units per milliliter (CFU)/mL) of bacteria. The research question we sought to answer was as follows: For a woman suspected of having or found to have gonorrhea, chlamydia, or trichomoniasis during the ED encounter who has genitourinary concerns, are the inflammatory changes observed on urinalysis most likely caused only by the STI, or is concurrent bacteriuria (eg, UTI) contributing?

METHODS

Dataset

We used an existing dataset of 75,000 ED encounters of patients ≥18 years in age from a single health system.^11–22 All patients in the dataset received testing for gonorrhea, chlamydia, or trichomoniasis or underwent both urinalysis and urine culture. Patients undergoing only urinalysis, regardless of STI testing, were not included in the dataset. All ED encounters took place April 18, 2014–March 7, 2017. The dataset was created by the institution’s information technology team who extracted retrospective data from the electronic health records (EHR). For our analysis, we included women who were not admitted to the hospital and who had a nucleic acid amplification test (NAAT) for gonorrhea, chlamydia, or trichomoniasis or had a vaginal wet mount. Data on the NAAT swab site were not available. Individual patients could have more than one ED encounter. Our project was approved by the institutional review board with an exemption from full review, and informed consent was waived. Articles have been published using the original dataset.^11–22

Patients in the dataset were identified as having trichomoniasis if the parasite was seen with urine microscopy (a method having very low sensitivity but high specificity), vaginal wet mount (moderate sensitivity and high specificity), or NAAT (high sensitivity and specificity).^22–24 To avoid multicollinearity in the multivariable analysis, we consolidated findings from vaginal wet mount and urine microscopy for T vaginalis into a single variable labeled T vaginalis infection status known during the ED encounter. The T vaginalis NAAT (Aptima, Hologic, Inc, Marlborough, MA) result, or the Neisseria gonorrhoeae or Chlamydia trachomatis NAAT (Aptima), was considered separately because the result was not obtained until after the ED visit. Women may have tested positive for T vaginalis by more than one test during their encounter, and any patient with a positive T vaginalis test was considered to be infected with T vaginalis. All STI testing was performed at the discretion of the treating clinician.

We reported the vaginal wet mount as not performed if the patient had no results from the vaginal wet mount for white blood cells (WBC), yeast, T vaginalis, or clue cells. The vaginal wet mount WBCs were analyzed as 0–10/11 or more cells per high-power field (HPF).¹⁶ For the vaginal wet mount, yeast, clue cells, and T vaginalis were reported by the clinical laboratory as present or absent.

We considered a urinalysis to have been performed if any component test from the urinalysis was reported. The urine sample was reported to have been collected by the following methods: clean catch/voided; missing or not documented by nursing; or “other” (eg, bladder catheter, straight catheter, ileostomy, nephrostomy, suprapubic, or urostomy). From the urinalysis, we considered the following variables: bacteria (0–4+); blood (0–3+); glucose (positive or negative); ketones (positive or negative); leukocyte esterase level (0–3+); mucus (0–4+); nitrites (positive or negative); protein (positive or negative); red blood cells (RBC) (0–101 cells/HPF); Trichomonas (positive or negative); WBC clumps (positive vs negative); WBCs (0–101 cells/HPF); and yeast (present or absent). If a range of urine RBCs and WBCs was reported, we used the median of the range in the analysis, and if more than 100 cells/HPF were reported, we used the result “101 cells/HPF” for analysis. All urine tests were ordered at the discretion of the treating clinician.

We included the following demographic and triage information if it was available during the ED encounter: method of ED arrival; marital status; age; race; and the triage Emergency Severity Index. Age in years was converted to a categorical variable to account for the nonlinear relationship with STIs.²⁵

Women were considered to have a UTI diagnosis if they had a specific ED code on the International Classification of Diseases, 9^th or 10^th Rev (ICD-9/ICD-10) (Supplement 1). Women were considered pregnant if they had a documented positive pregnancy test or a specific ICD-9 or ICD-10 code (Supplement 1).

Statistical Analysis

We summarized continuous variables as median and interquartile range, with analysis of variance F tests used to test associations. We reported categorical variables as counts and percentages, with a χ² test used to test associations. We perfomed multinomial logistic regression analysis accounting for multiple demographic, clinical, and diagnostic testing variables, with the Wald test used to determine P values. Multivariable analyses were performed for patients who had complete data for all model covariates. Odds ratios and 95% confidence intervals were calculated from the multivariable model. A P value less than .05 was considered significant. We conducted statistical analyses with statistical software JMP Pro 14 (JMP Statistical Discovery, LLC, Cary, NC) and SAS version 9.4 (SAS Institute, Inc, Cary, NC).

RESULTS

Among the 75,000 ED encounters in the original dataset, 16,755 women met our inclusion criteria. A summary of the clinical encounters is shown in Table 1. Among the 1,631 patient encounters with a positive test result for gonorrhea, chlamydia, or both, 1,443 (88.5%) had urinalysis, 443 (27.2%) had urine culture, and 438 (26.9%) had both urinalysis and urine culture. Among the 1,354 women with T vaginalis identified on vaginal wet mount and 418 women with a positive NAAT result for T vaginalis, 1,203 (88.8%) and 374 (89.5%) patients, respectively, had urinalysis. Table 2 shows encounters with a positive STI test result and the results of the urine culture. Among the 443 patients with gonorrhea, chlamydia, or both who had a urine culture result, 341 (77.0%) had less than 10,000 CFU/mL of bacteria, and 102 (23.0%) had 10,000 or more CFU/mL of bacteria in the urine culture.

In total, 1,804 patient encounters had a positive test result for Trichomonas by urine microscopy, vaginal wet mount, or NAAT. Of these, 1,612 (89.4%) had a urinalysis test result, 548 (30.4%) had a urine culture performed, and 538 (29.8%) had both a urinalysis and urine culture result. A total of 9,650 clinical encounters had complete observations for all model covariates and were included in the multivariable analysis (Table 3). This number included 2,414 patient encounters with less than 10,000 CFU/mL of bacteria, 722 patients with 10,000 or more CFU/mL of bacteria, and 6,514 patients with no urine culture performed.

The following variables were significantly more likely to be associated with a urine culture with 10,000 or more CFU/mL compared with less than 10,000 CFU/mL: higher bacteriuria on urinalysis; higher amount of blood in the urine nitrite-positive urine presence of urinary WBC clumps; higher urinary WBC count; and fewer WBCs on the vaginal wet mount (all P ≤ 0.01; Table 3). These variables had a significantly lower likelihood of being associated with a urine culture with ≥10,000 CFU/mL: no T vaginalis NAAT result (compared with a negative T vaginalis NAAT) and protein in the urine (both P ≤ 0.01; Table 3). The following variables were significantly more likely to be associated with ≥10,000 CFU/mL of bacteria in the urine culture (compared with no urine culture performed): married/life partner (vs single); higher bacteriuria on urinalysis; higher urine leukocyte esterase level; nitrite-positive urine, protein in the urine, presence of urinary WBC clumps, UTI diagnosed in the ED, and higher urinary WBC count (all P ≤ 0.02; Table 3). These variables had a significantly lower likelihood of being associated with a urine culture with ≥10,000 CFU/mL (compared with no urine culture performed): no T vaginalis NAAT result (compared with a negative T vaginalis NAAT), protein in the urine, and no vaginal wet mount clue cells (compared with present) (all P ≤ 0.01; Table 3). Neisseria gonorrhoeae or C trachomatis detected by NAAT, or known T vaginalis infection in the ED, was not associated with a urine culture yielding 10,000 or more CFU/mL. Additionally, UTI diagnosed in the ED was not associated with a urine culture yielding 10,000 or more CFU/mL compared with less than 10,000 CFU/mL.

DISCUSSION

Both UTIs and STIs can have overlapping signs and symptoms and can cause inflammatory changes in the urine. Distinguishing between UTI and STI can be challenging in the ED.^5,6,26 We sought to assess the relationship between bacteriuria and STIs. Our research question was as follows: For a woman suspected of having or found to have gonorrhea, chlamydia, or trichomoniasis during the ED encounter who has genitourinary concerns, are the inflammatory changes observed on urinalysis most likely caused only by the STI, or is concurrent bacteriuria (eg, UTI) contributing? Our results show that infection with gonorrhea, chlamydia, or trichomoniasis was not associated with also having a urine culture yielding ≥10,000 CFU/mL of bacteria compared with <10,000 CFU/mL or no urine culture performed. An important finding was that when T vaginalis was identified during the ED encounter on urine microscopy or vaginal wet mount, there was no significant association with bacteria in the urine culture. When an emergency clinician is evaluating a woman with genitourinary concerns and an STI is suspected or actually identified, as is the case on urine microscopy or vaginal wet mount for T vaginalis, bacteriuria is not more likely to coexist. Our findings support recommendations for screening for both UTIs and STIs in appropriate patients.^7,8 For instance, women undergoing pelvic examination who were also diagnosed with a UTI in the ED were subsequently found to have high rates of STIs.⁷ However, emergency clinicians frequently do not screen for STIs in women with dysuria who are diagnosed with a UTI.⁸ Furthermore, our findings support those of smaller studies.^6,27

A study by Prentiss et al showed that among adolescent girls with urinary tract symptoms in the ED, 9% had an STI, 57% had a UTI, and 6% had both an STI and a UTI.⁶ Clinician accuracy was 83% for STIs and 71% for UTIs, whereas only 23% correctly diagnosed patients with both UTI and STI.⁶ Shapiro et al²⁷ found that among 92 women with urinary tract symptoms, STI rates were not different between women with a positive vs a negative urine culture (10² CFU/mL). Additionally, a retrospective study of ED patients found that patients who were treated for a UTI, tested positive for gonorrhea, chlamydia, or trichomoniasis, and had pyuria were significantly more likely to have a negative urine culture than a positive urine culture.⁹ Reliance on positive urine nitrite and pyuria to treat for UTI in patients with confirmed or suspected STI may result in overtreatment with antibiotics. However, patients with an STI and a positive urine culture had significantly higher urine leukocytes than those with negative culture results.⁹

We also found that clinical encounters in which patients were diagnosed with a UTI in the ED were not more likely to have a urine culture of ≥10,000 CFU/mL of bacteria compared with <10,000 CFU/mL Possibly, patients who were diagnosed with a UTI but who had <10,000 CFU/mL were more likely to have an STI, but this association was not examined in the current study. Because the diagnosis of a UTI was not part of our inclusion criteria, not all women with a UTI diagnosis are represented in our cohort. We were able to study only women who had both a urinalysis and urine culture, not just a urinalysis. Therefore, the association between a UTI diagnosis and bacteriuria deserves further investigation.

LIMITATIONS

Although our study used a large dataset, it has some limitations. First, not all women from our dataset underwent urinalysis, urine culture, vaginal wet mount, and NAAT for STIs. Furthermore, not all women diagnosed with a UTI underwent STI testing or a vaginal wet mount. Second, modeling T vaginalis in the ED has inherent limitations because the urinalysis and vaginal wet-mount results are available to the clinician during the encounter, but they lack high sensitivity, whereas NAAT is highly sensitive and specific but typically does not yield results during the patient encounter. Third, women undergoing STI testing who also had a urine culture may have been more likely to be concerned about urinary symptoms, which could have biased our analysis to those women with genitourinary concerns. Because we were unable to include history and physical examination findings in our analysis, we could not differentiate between patients with more genital concerns and those having more urinary symptoms. Alternatively, some women included in the analysis may have had asymptomatic bacteriuria or an asymptomatic STI, although this possibility is thought to be less likely. We did not attempt to differentiate between contaminated urine cultures and those yielding classic uropathogens.⁵

Fourth, our dataset represented a limited geographical area in the US, and the patients were predominantly Black; therefore, our results may not be generalizable to other locations and races. Fifth, the data did not include pediatric patients or men; so our results cannot be extrapolated to those groups. Sixth, patients who were treated presumptively for STIs without specific testing were excluded from analysis, and this could have resulted in selection bias. Finally, inherent limitations exist to using a dataset extracted from the institution’s EHR and ICD codes.

CONCLUSION

In our regression analysis, positive gonorrhea, chlamydia, and trichomoniasis test results were not associated with bacteriuria yielding ≥10,000 CFU/mL compared with <10,000 CFU/mL or no urine culture obtained.

Footnotes

Section Editor: Ioannis Koutroulis, MD, MBA, PhD

Full text available through open access at http://escholarship.org/uc/uciem_westjem

Address for Correspondence: Johnathan M. Sheele, MD, MHS, MPH, Mayo Clinic, Department of Emergency Medicine, 4500 San Pablo Rd., Jacksonville, FL 32224. Email: sheele.johnathan@mayo.edu
05 / 2024; 25:358 – 367

Submission history: Revision received February 1, 2023; Submitted December 13, 2023; Accepted January 16, 2014

Conflicts of Interest: By the WestJEM article submission agreement, all authors are required to disclose all affiliations, funding sources and financial or management relationships that could be perceived as potential sources of bias. No author has professional or financial relationships with any companies that are relevant to this study. There are no conflicts of interest or sources of funding to declare.

REFERENCES

1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

Share on Facebook

Tweet

Share via email