Sexually transmitted infections and prior antibiotic use as important causes for negative urine cultures among adults presenting with urinary tract infection symptoms to primary care clinics in Zimbabwe: a cross-sectional study

Objective Urinary tract infections (UTIs) are common in primary care. The yield of urine cultures in patients with UTI symptoms can be considerably different between high-income and low-income settings. This study aimed to explore possible causes of negative urine cultures in patients presenting with symptoms of UTI to primary health clinics in Harare. Design Cross-sectional study. Setting Nine primary health clinics in Harare, Zimbabwe. Participants Adults presenting with symptoms of UTIs between March and July 2020. Primary outcome measures Urine samples underwent dipstick testing, microscopy, culture, and testing for sexually transmitted infections (STIs) using GeneXpert and for the presence of antibiotic residues using an antibiotic bioassay. The primary outcomes were the number and proportion of participants with evidence of STIs, prior antibiotic exposure, leucocyturia and UTIs. Results The study included 425 participants with a median age of 37.3 years, of whom 275 (64.7%) were women. Leucocyturia was detected in 130 (30.6%, 95% CI 26.2% to 35.2%) participants, and 96 (22.6%, 95% CI 18.7% to 26.9%) had a positive urine culture for a uropathogen. Chlamydia trachomatis, Neisseria gonorrhoeae and Trichomonas vaginalis were detected in 43/425 (10.1%, 95% CI 7.4% to 13.4%), 37/425 (8.7%, 95% CI 6.2% to 11.8%) and 14/175 (8.0%, 95% CI 4.4% to 13.1%) participants, respectively. Overall, 89 (20.9%, 95% CI 17.2% to 25.1%) participants reported either having taken prior antibiotics or having had a positive urine bioassay. In 170 (40.0%, 95% CI 35.3% to 44.8%) participants, all of the tests that were performed were negative. Conclusions This study found a high prevalence of STIs and evidence of prior antimicrobial use as possible explanations for the low proportion of positive urine cultures.

P5 line 31: selection bias? Can the lowest income groups in low income countries afford this user fee or is an essential part excluded of this type of primary health care? General question about discussion: What is common management in Zimbabwe and in sub-Saharan Africa? Is there a (local) guideline? The European guidelines cannot be used as the results of urine analysis are completely different. Could this be an issue of the discussion or conclusion? Considering these results, what would be a reasonable management? Also when taking into account the probably limited resources. What to do with the 40 % negatieve results? How could these patients be best taken care of?
Taking into account the underdiagnosis of STI in patients consulting with urinary complaints it could be rational to first test for STI and in a second step, when STI test is negative to further examine for UTI (which is in essence a self limiting disease). In doing so a cost saving could be made. In High-income countries urinary culture is not needed in women with suggested uncomplicated UTI, but as 40% of these patients do not have a pos STI nor a pos culture, in sub-Saharan Africa it could be desirable to perform a culture, to avoid overtreatment with antibiotics. # Symptom resolution or improvement was reported by 74/89 (83.1%) of participants with a positive urine culture. Among participants with a positive urine culture, symptoms improved or resolved in 64/67 (95.5%) of those who reported having taken their prescribed treatment and in 9/20 (45%) of those who had not. The numbers of patients in the two sentences do not correspond with each other.

REVIEWER
# In figure 2, severe symptoms were not clearly defined.

REVIEWER 1 Comment #1
P5 line 31: selection bias? Can the lowest income groups in low income countries afford this user fee or is an essential part excluded of this type of primary health care?

Response to comment #1
This is an excellent point and we agree that clinic fees might have impacted on healthcare access which may have led to selection bias. This is now included as part of the limitations "Potential participants who did not seek care because of not being able to afford the clinic fees would have been missed." An alternative to evaluating the prevalence of STIs would have been to conduct the study at community-level. However, the parent-study was focused on the prevalence of antimicrobial resistance among patients presenting with symptoms of urinary tract infection and thus required a clinic-setting.

Comment #2
P5 line 53: what is the diameter used to be considered positive? Response to comment #2 Because the discs did not contain any antimicrobials (they were filter paper discs onto which the urine samples from the participants were inoculated), any growth inhibition was considered relevant for the presence of antimicrobial residues in the urine sample. For each batch of tests, we used positive and negative controls to ensure the quality of testing results. This was also further clarified in the manuscript text: "Antibacterial activity was considered to be present in the sample if a growth inhibition zone of any size around the disc was observed following incubation for 24 hours at 37°C."

Comment #3
P6 line 51: concerning fig 1: maybe it could be made more clear to mention the total (between brackets) of STI+, LE+, urine culture + en those who had prior antibiotics.

Response to comment #3
This information was added to the manuscript: " Figure 1  Version%20with%20Signatures.docx). These include recommendations for the management of sexually transmitted and urinary tract infections. A clarification with regards to the guidelines used for prescribing decisions was added in the manuscript. "Treatment for STIs was prescribed during the initial visit by the clinic nurse according to routine practice following the Zimbabwean national guidelines and independent of study procedures." Generally, STIs in Zimbabwe are treated using a syndromic approach because diagnostic testing is not available. Within this study, patients received routine care when they presented to clinic and were prescribed STI treatment if an STI was suspected by the clinic nurse. Patients who had a positive STI test and who were not prescribed appropriate treatment during the initial visit, were called back to the clinic for treatment. Given the high prevalence of STIs identified in this study, a reasonable approach would be to increase the availability of STI testing. This issue is discussed in the manuscript. We also fully agree that using European guidelines would not have been appropriate in this context.

Comment #5
What to do with the 40 % negative results? How could these patients be best taken care of? Response to comment #5 Although a high proportion of participants remained undiagnosed, for most (80%) symptoms were transitory and had resolved or improved at follow up. We hypothesize that in some of these patients, symptoms could have been explained by urinary tract infections and other STIs for which testing was not conducted. The delayed presentation (with a median of 7 days of symptoms until healthcare was sought) may have led to the spontaneous resolution of urinary tract infections and negative urine cultures. Potentially, increasing healthcare access by for example decreasing clinic fees would reduce the delay in seeking care and improve patient management.

Comment #6
Taking into account the underdiagnosis of STI in patients consulting with urinary complaints it could be rational to first test for STI and in a second step, when STI test is negative to further examine for UTI (which is in essence a self limiting disease). In doing so a cost saving could be made. In Highincome countries urinary culture is not needed in women with suggested uncomplicated UTI, but as 40% of these patients do not have a pos STI nor a pos culture, in sub-Saharan Africa it could be desirable to perform a culture, to avoid overtreatment with antibiotics.

Response to comment #6
Because of the economic hardships in Zimbabwe and other low-resource settings, we do not think that a two-step testing approach would be ideal. Patients face great challenges in accessing care by having to pay out-of-pocket for consultation fees, prescribed medicines and transport to the clinic. Therefore, if multiple clinic visits were required, patients may not come back for further assessment after their initial clinic visit. Ideally, rapid point of care testing for STIs would identify patients requiring STI treatment, while in patients with negative STI tests and suggestive symptoms, an UTI can be considered. Despite a high proportion of urine cultures being negative we would not recommend routine urine cultures. We are aware that a large proportion of UTIs are self-limiting and hence urine cultures in women presenting with symptoms of UTIs for the first time in primary care settings are not needed.

REVIEWER 2 Comment #1
The frequency of leukocyturia was very low (30.6%). In participants without leukocyturia, positive urine cultures may not indicate that symptoms were due to UTI. Other causes may contribute to UTI symptoms. Furthermore, evidence of prior antimicrobial use does not mean that the UTI caused the symptoms in participants without positive other tests. These could underestimate the proportion of other causes except for UTI in UTI symptoms.

Response to comment #1
Thank you, we fully agree that positive urine cultures in symptomatic patients with low bacterial loads and no leukocyuria might be difficult to interpret especially in the context of a high STI prevalence and frequent prior antibiotic use. We also fully agree that patients may have taken antibiotics for reasons other than their current symptoms although taking antibiotics for their symptoms would be the most likely explanation.

Comment #2
Symptom resolution or improvement was reported by 74/89 (83.1%) of participants with a positive urine culture. Among participants with a positive urine culture, symptoms improved or resolved in 64/67 (95.5%) of those who reported having taken their prescribed treatment and in 9/20 (45%) of those who had not. The numbers of patients in the two sentences do not correspond with each other.

Response to comment #3
Thank you for pointing out our error. The numbers were corrected. "Among participants with positive urine cultures, symptoms improved or resolved in 65/69 (94.2%) of those who reported having taken their prescribed treatment and in 9/20 (45%) of those who had not."

Comment #3
In figure 2, severe symptoms were not clearly defined.

Response to comment #3
Severe symptoms were defined as symptoms limiting daily activities. The definition was added in the figure legend.