Overuse of antibiotics for urinary tract infections in pregnant refugees, Lebanon

Abstract Objective To determine whether adding urine culture to urinary tract infection diagnosis in pregnant women from refugee camps in Lebanon reduced unnecessary antibiotic use. Methods We conducted a prospective, cross-sectional study between April and June 2022 involving pregnant women attending a Médecins Sans Frontières sexual reproductive health clinic in south Beirut. Women with two positive urine dipstick tests (i.e. a suspected urinary tract infection) provided urine samples for culture. Bacterial identification and antimicrobial sensitivity testing were conducted following European Committee on Antimicrobial Susceptibility Testing guidelines. We compared the characteristics of women with positive and negative urine culture findings and we calculated the proportion of antibiotics overprescribed or inappropriately used. We also estimated the cost of adding urine culture to the diagnostic algorithm. Findings The study included 449 pregnant women with suspected urinary tract infections: 18.0% (81/449) had positive urine culture findings. If antibiotics were administered following urine dipstick results alone, 368 women would have received antibiotics unnecessarily: an overprescription rate of 82% (368/449). If administration was based on urine culture findings plus urinary tract infection symptoms, 144 of 368 women with negative urine culture findings would have received antibiotics unnecessarily: an overprescription rate of 39.1% (144/368). The additional cost of urine culture was 0.48 euros per woman. Conclusion A high proportion of pregnant women with suspected urinary tract infections from refugee camps unnecessarily received antibiotics. Including urine culture in diagnosis, which is affordable in Lebanon, would greatly reduce antibiotic overprescription. Similar approaches could be adopted in other regions where microbiology laboratories are accessible.


Introduction
Inappropriate antibiotic use is one of the main drivers of antimicrobial resistance. 1 The standard of care for pregnant women is to give antibiotics to prevent complications, particularly in those with a urinary tract infection and asymptomatic bacteriuria.However, overprescription is common and previous studies have found that antibiotics were overprescribed in up to 96% of pregnant women with suspected urinary tract infections. 2,3n pregnant women, a urinary tract infection can present symptomatically or asymptomatically. 3,4If left untreated or if treated inappropriately, infection can lead to complications for mothers (e.g.pyelonephritis, septicaemia or acute respiratory distress syndrome) and can result in preterm delivery, intrauterine growth restriction or a low-birth weight infant. 5,6herefore pregnant women should undergo screening for urinary tract infections to ensure that infections are accurately diagnosed and treated effectively.This approach can be very challenging during conflicts and among refugees because quality health care may be lacking, drugs may be unavailable, and timely and effective access to diagnostic services, including microbiology laboratories, may be costly and limited or nonexistent. 7,8ternational clinical guidelines recommend that pregnant women should be screened for urinary tract infections during their first antenatal care visit.For those with positive urinary tract infection screening test results, urine culture should be conducted to confirm the diagnosis, and antibiotic sensitivity testing should be performed to select the most appropriate antibiotic therapy. 6,9,10Rapid screening methods for urinary tract infections, such as the urine dipstick test, are often the only options available in low-resource settings and humanitarian situations but they lack reliability.A recent meta-analysis showed that urine dipstick tests had a pooled sensitivity of 73% and a specificity of 89% for detecting asymptomatic bacteriuria during pregnancy. 11Another metaanalysis published in 2004 concluded that urine dipsticks were effective for ruling out infection when both nitrite and leukocyte results were negative, but were not reliable as stand-alone tests to confirm the presence of a urinary tract infection. 12t the Médecins Sans Frontières (MSF) sexual reproductive health clinic in the Burj al-Barajneh refugee camp in south Beirut, Lebanon, pregnant women are screened for urinary tract infections during each antenatal care visit using a repeated CombiScreen urine dipstick test (Analyticon Biotechnologies GmbH, Lichtenfels, Germany) that can detect the presence of nitrites and leukocytes, as specified Bull World Health Organ 2024;102:389-399| doi: http://dx.doi.org/10.2471/BLT.23.291235   Christine Al Kady et al.
Research by MSF guidelines. 13Women who have two positive urine dipstick test results are given either a single 3 g oral dose of fosfomycin trometamol or receive 200 mg oral doses of cefixime twice daily for 5 days.No confirmatory tests are performed.
The aim of our study was to investigate the effect of adding urine culture to the diagnostic algorithm for urinary tract infections on the prescription of antibiotics for pregnant Syrian refugees who attended an MSF antenatal care clinic in Lebanon.In addition, we determined whether the use of a repeated positive urine dipstick test result alone to diagnose urinary tract infections in these women, in accordance with the current MSF protocol, led to inappropriate antibiotic prescription and usage.

Methods
Since 1949, more than 30 000 refugees are estimated to have settled in the overcrowded Shatila and Burj al-Barajneh camps in the southern suburbs of Bei-rut, 14 where they face extremely challenging living conditions, including high unemployment, limited infrastructure and unpredictable security. 15The main provider of free primary health care in both camps remains MSF, which has, since 2014, offered antenatal, postnatal and family planning services through an outpatient centre that caters for a monthly average of 1300 antenatal care consultations (Médecins Sans Frontières, unpublished data, 2023).
We performed an analytical, crosssectional study involving prospectively collected data on pregnant women aged 18 years or older who: (i) received routine antenatal care at an MSF clinic in south Beirut between 27 April and 30 June 2022; and (ii) had two consecutive, positive, urine dipstick test results during a single clinic visit.Each woman was included only once.We excluded women who had received antibiotics in the week before their visit, or who were eligible for urine culture according to the current MSF clinical algorithm (i.e. three consecutive, positive, urine dipstick test results at three different antenatal care visits).
Each woman enrolled in the study provided two midstream urine samples on the same day.The first sample was used for the first dipstick test.If either the nitrite or leukocyte result was positive, a second urine sample was collected after precleaning with soap and water.The second sample was used for a repeat dipstick test and for urine culture.For our study, we recorded the nitrite and leucocyte results from the second positive dipstick test.In women with symptoms of a urinary tract infection, women provided both urine samples before starting antibiotic treatment (Fig. 1).
If a woman reported any symptoms of sexually transmitted infection or vaginitis during the visit, she was offered an inspection of the vulva and a speculum exam checking for cervical and/or vaginal inflammation or discharge, as part of her routine care.Before examination, the patient had to give verbal consent, as per usual practice in the clinics.Urine cultures were processed at the Saint Marc laboratory in Beirut, a microbiology laboratory whose quality had previously been validated for urine testing by MSF.The laboratory reported culture results within 48 hours.Bacteria were isolated and enumerated and bacterial growth was assessed using a cystine, lactose, electrolyte-deficient agar and 1 or 10 µL loops.Bacteria were identified primarily using a stain test (i.e.Gram staining) supplemented by tests that assessed whether the bacteria contained or had the ability to produce certain enzymes (e.g.catalase and oxidase) and an analysis of each bacterium's characteristic colony morphology on the agar.Antibiotic susceptibility testing was interpreted according to the guidelines of the European Committee on Antimicrobial Susceptibility Testing (EUCAST; version 12.0, valid from 1 January 2022) and an antibiogram was generated for each positive urine culture to identify each organism's susceptibility to different antimicrobials. 16e extracted sociodemographic and routine clinical data from electronic health records and additional data were retrieved from patients' medical files and from a pre-drafted study questionnaire that asked about urinary tract infection history, recent antibiotic use and urinary tract infection symptoms.We obtained microbiology through an online laboratory information system.

Data analysis
We defined a suspected urinary tract infection as two consecutive, positive, urine dipstick test results (i.e. the presence of nitrites or a leukocyte level of 2 or higher) during a single clinic visit.A confirmed urinary tract infection was defined as positive urine culture findings, as indicated by bacterial growth of at least 10 000 colony-forming units (CFUs) per mL, irrespective of the presence of clinical signs of infection.To avoid undertreatment, we used this cut-off value combined with the presence of a maximum of two uropathogens to confirm the need for urinary tract infection treatment.
We defined overprescription of antibiotics as the prescription of antibiotics for a suspected urinary tract infection when it was unnecessary, as indicated by negative urine culture findings.An inappropriate antibiotic was defined as a prescribed antibiotic whose use was not consistent with antibiotic susceptibility findings among women in the study with confirmed urinary tract infections.
To accurately assess the overprescription of antibiotics, we determined that a conservative sample of 450 pregnant women was needed.We assumed that 50% of women with two positive dipstick results on the same visit would have a confirmed urinary tract infection on urine culture, and considered a 95% confidence interval (CI) with a 5% level of precision and a 15% nonresponse rate (OpenEpi Calculator, OpenEpi).We report findings using descriptive statistics and consider a P-value of 0.05 or less as significant.
From the microbiology data, we calculated the proportion of women with a confirmed urinary tract infection and estimated the proportion of antibiotics overprescribed.We also identified individual bacterial species isolated from urine cultures.In addition, information from antibiotic susceptibility testing was used to estimate the proportion of antibiotics used inappropriately.Susceptibility results reported as "I -susceptible, increased exposure" were grouped with results reported as "S -susceptible, standard dosing regimen", except for antibiotics without clinical breakpoint concentrations in their respective EU-CAST categories. 16All analyses were performed using Stata v. 14.0 (StataCorp LLC, College Station, United States of America).
In estimating the additional cost of urine culture for suspected urinary tract infections, we considered the cost of: (i) the urine dipstick tests, urine culture tests (including urine cups) and antibiotics that would have been provided hypothetically under routine conditions for all enrolled women; and (ii) the antibiotic susceptibility tests used and antibiotics actually prescribed under research conditions for women with positive urine culture findings.We calculated the mean cost per patient with and without urine culture.Estimated costs were calculated in euros (€) and based either on local pricing in Lebanon (e.g. for urine cultures and a Two urine cups (one for each dipstick) are needed in the absence of urine culture; with urine culture, an additional cup is used and sent to the laboratory for testing.b The costs shown are for full courses of each antibiotic: without urine culture, the costs were for antibiotics provided hypothetically under routine conditions for all enrolled women; with urine culture, the costs were for antibiotics actually prescribed following antibiotic susceptibility testing.c Other antibiotics were used when the causative organism was resistant to both fosfomycin and cefixime.

Research
antibiograms) or on MSF pricing (e.g. for imported urine dipsticks and antibiotics).Table 1 summarizes the costs included.
The study protocol was approved by the MSF ethics review board (ID: 2155, 6 December 2021) and the Lebanese American University institutional review board (LAU.SOM.AF2, 7/ Apr/2022).All women provided written informed consent.
Of the 449 women, 81 (18.0%) had their urinary tract infection diagnosis confirmed on urine culture, whereas 368 (82.0%) had negative urine culture findings (Fig. 1).Among the 81 with confirmed urinary tract infections, 59 (72.8%) had a uropathogen concentration greater than 10 5 CFU/mL, whereas 22 (27.2%) had a concentration between 10 4 and 10 5 CFU/mL.Of the 22 with a concentration less than or equal to 10 5 CFU/mL, 14 (63.6%) had at least one clinical symptom indicative of a urinary tract infection at the time of the visit, whereas eight (36.4%) had no symptoms.Compared with women who had negative urine culture findings, women with a confirmed urinary tract infection were significantly more likely to be in their first pregnancy or first trimester, and to have a urinary tract infection for the first time during their current pregnancy (Table 2).
At enrolment, 43.9% (197/449) of women had at least one symptom of a urinary tract infection (Fig. 1); dysuria was the most common (82.2%, 162/197).Among the 81 with a confirmed urinary tract infection, 28 (34.6%) had asymptomatic bacteriuria.Women with positive urine culture findings were significantly more likely to present with a symptom than those without: the proportions were 65.4% (53/81) and 39.1% (144/368), respectively (P < 0.001; Table 2).Thirty of the 144 (20.8%) women who had symptoms of a uri-nary tract infection but negative urine culture findings were diagnosed with a sexually transmitted infection based on speculum observations.Almost all women (96.4%, 27/28) with a positive nitrite result on their urine dipstick test had a confirmed urinary tract infection (Table 2).Nitrite results on the dipstick test had a specificity of 99.7% (367/368) and a sensitivity of 33.3% (27/81) for positive urine culture findings.

Overprescription of antibiotics
We found that, if the urine dipstick test was the only tool used for diagnosis of a urinary tract infection (standard practice in many low-and middleincome countries and routinely used by the MSF south Beirut clinic), 368 (144 with urinary tract infection symptoms plus 224 without) of the 449 women with two positive urine dipstick test results would have received antibiotics unnecessarily because urine culture subsequently showed they did not have a urinary tract infection (Fig. 1).This finding corresponds to an overprescription of 82.0% (368/449).Moreover, 43.9% (197/449) of women enrolled in our study received empirical antibiotics before their urine culture findings were available because they presented with at least one urinary tract infection symptom: of these, 42 (21.3%)received cefixime and 155 (78.7%) received fosfomycin.However, 144 of the 197 (73.1%) subsequently had negative urine culture findings and, hence, were unnecessarily prescribed antibiotics.If only women with symptoms who had two positive urine dipstick results were given empirical antibiotic treatment (i.e.under research conditions), 144 of all 368 women with negative urine culture findings would have received antibiotics unnecessarily, which corresponds to an overprescription rate of 39.1% (144/368; Fig. 1).

Microbiology findings
Among the 81 women with positive urine culture findings, the most frequently isolated organism was Escherichia coli, which was present in 64 (80.0%) cultures.The other organisms detected were Proteus species in nine women (11.2%),Klebsiella pneumoniae in seven (8.7%) and Enterococcus faecalis in one (1.2%).Although the presence of extended spectrum β-lactamases (ESBLs) was not assessed, we estimated that 20.0% (16/80) of the Gram-negative bacteria responsible for urinary tract infection were ESBL producers because they were resistant to third-generation cephalosporins (e.g.ceftriaxone) but susceptible to cefoxitin and β-lactamase inhibitors.Almost half of all bacterial isolates were resistant to cefixime, and one E. coli isolate showed fosfomycin resistance (Table 3).Cefixime would have been used inappropriately for three of 13 isolates (23.1%) that would have been treated with the drug under routine conditions, because the isolates were found to be resistant on susceptibility testing and fosfomycin would have been used inappropriately for one E. coli isolate.

Costs
The estimated total cost of diagnosis and treatment for all 449 women included in the study under routine conditions (i.e.without urine culture) was €1400 compared to €1620 with urine culture, which corresponds to a mean cost per woman of €3.12 (95% CI: 3.07-3.17)and €3.60 (95% CI: 3.40-3.80),respectively.The mean difference was €0.48 per woman.Not administering antibiotics to the 368 women with negative urine culture findings could have saved €992.70.

Discussion
We found that less than one fifth of pregnant women who had a suspected urinary tract infection based on positive urine dipstick results alone had their infection confirmed on urine culture.Hence, adding reliable urine culture to the diagnostic algorithm would avoid most of these women being overprescribed antibiotics.We also found that, if only women with symptoms and two positive urine dipstick results received antibiotics, about two out of five women would be overprescribed antibiotics.
Promoting antimicrobial stewardship during the outpatient treatment of urinary tract infections is particularly challenging in low-resource settings due to diagnostic uncertainties, the limitations of existing guidelines and rising antimicrobial resistance. 17Previous studies reported antibiotic overuse rates above 96% for urinary tract infections in pregnant women in some contexts, 2,[17][18][19] often because treatment was based on clinical suspicion or because rapid screening methods were used rather than microbiological confirmation. 17n our study, around half of confirmed urinary tract infections involved a Bull World Health Organ 2024;102:389-399| doi: http://dx.doi.org/10.2471/BLT.23   20 This finding highlights the need: (i) to reconsider the use of urine dipsticks for diagnosing urinary tract infections to decrease the overuse of Watch antibiotics; (ii) to improve monitoring of appropriate antibiotic prescriptions with systematic point-prevalence surveys; and (iii) to propose alternative treatments for use in Lebanon, such as nitrofurantoin -an Access antibiotic to which E. coli is highly susceptible.Although nitrofurantoin is suggested to be a safe and reasonable first-line treatment for urinary tract infections in pregnancy, 21 concerns exist about the risk of congenital malformation in the first trimester and about neonatal haemolytic anaemia near delivery. 22itrofurantoin remains an option to consider, but implementing a tailored prescribing algorithm for each trimester in humanitarian settings might be challenging.
The levels of antibiotic overuse and inappropriate use we found in this study contribute to antimicrobial resistance, which is particularly concerning during pregnancy because antibioticresistant bacteria may be transmitted to neonates and cause adverse neonatal outcomes. 23,24Therefore, ensuring that antibiotics are used rationally and appropriately during pregnancy is a must, particularly given the increase in antimicrobial resistance both globally and among Syrian refugees. 25,26r estimate that 20% of organisms found in our study participants produced ESBLs may have been too low as previous studies have reported higher proportions: for example, 32% in urinary tract infection cases in Lebanon and 52% in urine samples in the Syrian Arab Republic. 27,28However, the proportion may have been low in our study because the women included were young on average and very few had a history of diabetes.[30][31][32][33][34]

Characteristic
Women with a suspected urinary tract infection b P-values for the difference in positive and negative results were calculated using a χ 2 or Fisher's exact test for categorical variables or using an unpaired t-test for continuous variables.c The Shatila and Burj al-Barajneh refugee camps in south Beirut.d Abortions included spontaneous and induced terminations of pregnancy.e Percentage of women with a history of urinary tract infection during the current pregnancy.f Percentages were calculated for 154 women with a history of urinary tract infection during the current pregnancy because data on antibiotic use for the previous urinary tract infection were not available for six women.g Percentage of women with current symptoms.h Each woman may have had more than one symptom.i No woman with urinary tract infection symptoms had haematuria.j The number of women prescribed an antibiotic (i.e.199) is higher than the number with symptoms (i.e.197) because two patients without symptoms were prescribed an antibiotic based on the medical team's judgment before the urine culture findings were available, which deviated from the protocol.k No woman had a pre-existing immune deficiency disease, sickle cell disease or a urinary tract malformation.m Sexually transmitted infections (STIs) were suspected using a syndromic approach based on speculum observations.n Percentage of women with a diagnosed STI or abnormal vaginal discharge at enrolment.o Each woman may have had more than one diagnosis at the same visit.Antibiotic overuse for urinary tract infection, Lebanon 6][37][38] However, a negative nitrite result does not exclude the possibility of positive urine culture findings; in particular, the relationship between the two may be influenced by urinary tract infection epidemiology. 38n Lebanon, if urine culture cannot be added to the diagnostic algorithm, the overprescription of antibiotics in pregnant women with urinary tract infections could possibly be reduced by requiring either a positive nitrite result on the urine dipstick test or the presence of two or more urinary tract infection symptoms.This approach should be considered cautiously as urine dipstick tests for leukocytes and nitrites are less accurate than urine culture. 39ccording to WHO recommendations, in settings where there is limited access to urine culture, on-site Gram staining of midstream urine samples is preferred for diagnosing asymptomatic bacteriuria in pregnant women unless financially unviable. 40In this situation, dipstick tests should be considered despite the higher likelihood of a false-positive result than with Gram staining. 40Our results were with previous reports that E. coli, Proteus species and Klebsiella species isolates from samples (including urine) collected in the Syrian Arab Republic were resistant to cefixime. 41Antibiotic resistance profiles may vary over time as the epidemiology of antimicrobial resistance evolves.Therefore, considering local changes in microbiological findings is important when choosing antibiotics to treat urinary tract infections in pregnancy. 42e found that adding urine culture slightly increased the average cost of urinary tract infection diagnosis and treatment.However, the difference of € 0.48 per woman we observed is unlikely to be a substantial barrier to accessing urine culture in Lebanon.In settings where urine culture is costly, not easily accessible or time-consuming, it is essential to acknowledge its utility and to explore the possibility of incorporating it into an algorithm tailored to the specific context.This approach can help reduce antibiotic overprescription.Although no diagnostic algorithm can accurately detect and appropriately treat all cases of urinary tract infection, adapting the al-gorithm to the context can help mitigate the negative impact of overprescription and inappropriate antibiotic use.
Our estimate of the prevalence of urinary tract infections in pregnant women may have been higher than the actual prevalence in the wider community of pregnant women because our study included only women who had a suspected urinary tract infection based on two positive urine dipstick test results.Furthermore, as our study included only women with positive urine dipstick test results, we could not calculate the overall sensitivity and specificity of the dipstick test used.Moreover, the sensitivity and specificity of the nitrite results for identifying a confirmed urinary tract infection were calculated only for women who had two positive urine dipstick test results.In addition, as our definition of a positive urine culture included a low uropathogen concentration (i.e. 10 4 to 10 5 CFU/mL), a quarter of infections detected and treated as urinary tract infections may have been due to skin flora contamination.Despite these limitations, we believe the overprescription of antibiotics under routine conditions is higher than the proportion we observed.Although the quality of the urine dipstick readings may have affected the accuracy of our results, we are confident that nurse training and the validation processes in place were sufficient to address uncertain dipstick readings.The costs we considered were based primarily on the price of MSF supplies -there may be slight local variations.No conclusions could be drawn about the cost-effectiveness of adding urine culture to the diagnostic process as this was outside of scope of the study.
Here we show that confirmatory urine culture was available, affordable and of a high quality.Therefore cultures should be incorporated into the diagnostic algorithm.We found that it was important to match urinary tract infection diagnostics and treatment to the local bacterial epidemiology.Such approach could be extrapolated to other settings to improve antimicrobial stewardship and the quality of care, and to help decrease the overall burden of antimicrobial resistance.■

Fig. 1 .
Fig. 1.Diagnosis and treatment flowchart, study of antibiotic overuse for urinary tract infection during pregnancy, Lebanon, 2022

Table 1 . Diagnosis and treatment costs, study of antibiotic overuse for urinary tract infection during pregnancy, Lebanon, 2022 Item Cost of urinary tract infection diagnosis and treatment Without urine culture With urine culture, negative culture findings With urine culture, positive culture findings
€: Euro; NA: not applicable.
.291235 Christine Al Kady et al.Antibiotic overuse for urinary tract infection, Lebanon

Table 3 . Antibiotic susceptibility of organisms isolated, study of antibiotic overuse for urinary tract infection during pregnancy, Lebanon, 2022 Antibiotic treatment No. of isolates susceptible to antibiotic (%)
The total number of isolates was actually 81; one additional organism, Enterococcus faecalis, is not presented because its susceptibility to the antibiotics listed was not tested.bAccordingtoMédecinsSansFrontières' clinical guidelines, cefixime and fosfomycin are first-line treatments for women with urinary tract infections.cTworesults for cefixime (one for E. coli and one for K. pneumonia), three for cefoxitin (two for E. coli and one for K. pneumonia) and two for nitrofurantoin (both for E. coli) were excluded from the calculations because they were reported as "I -susceptible, increased exposure" although their clinical breakpoint concentrations were not listed in the European Committee on Antimicrobial Susceptibility Testing guidelines used.16dAsclinicalbreakpointconcentrationsfor fosfomycin and nitrofurantoin were available only for E. coli, the two antibiotics were not tested on other isolates.eThedenominatorwas the number of E. coli isolates only.Bull World Health Organ 2024;102:389-399| doi: http://dx.doi.org/10.2471/BLT.23.291235 a