Unmet Need of Antenatal Screening for Asymptomatic Bacteriuria: A Risk Factor for Adverse Outcomes of Pregnancy.

Recommended urine culture is unsuitable for screening pregnant women for asymptomatic bacteriuria due to long turn-around time, unaffordability, and user-unfriendliness. The objective of this review was to evaluate the suitability of various tests for this purpose. A PubMed-based systematic review of published articles irrespective of year and language was done. Search terms included asymptomatic bacteriuria, screening test, urinary tract infection, and diagnostic test. Diagnostic accuracy studies conducted on human populations comparing tests with urine culture were included. One author extracted predefined data fields, including quality indicators, another validated it. Of 78 records, 25 studies describing 15 tests were included. All tests were rapid, seven were valid and two of them were affordable and easy-to-use. No test provided comprehensive identification with antibiotic susceptibility. Despite publication bias, no test was found suitable for screening asymptomatic bacteriuria antenatally and providing evidence-based prescription. Further research is needed to develop tests which suit this purpose. Copyright:

INTRODUCTION

During pregnancy, a woman is predisposed to asymptomatic bacteriuria, as a result of anatomical and functional changes in the urinary tract and reduction in immunity.[12] Additional risk is posed, especially to pregnant women in low-income countries due to low socioeconomic status, advanced maternal age, multiparity, increasing period of gestation, poor sanitation, lack of general hygiene, early marriage, first pregnancy at a younger age, lack of adequate treatment, and failure to attend antenatal clinic.[12] Nearly, 1.9%–9.5%[3] pregnancies are affected by asymptomatic bacteriuria worldwide, more commonly due to Gram-negative bacteria like Escherichia coli, Klebsiella, Proteus, and Pseudomonas, than Gram-positive bacteria like Streptococcus, Staphylococcus, and Enterococcus.[4]

Asymptomatic bacteriuria in pregnancy, if undetected and untreated, can lead to various adverse outcomes, either insidiously or by progressing to symptomatic urinary tract infections (UTIs). In mother, it causes preterm labor, hypertensive disorders of pregnancy, preterm premature rupture of membranes, anemia, postpartum endometritis, and acute pyelonephritis, while in the baby, it leads to intrauterine growth retardation, prematurity, low birth weight, and perinatal mortality.[5] Prematurity affects 15 million babies every year globally and is further responsible for at least 50% of all the neonatal deaths worldwide.[6] Even if the children survive it, they may suffer lifelong physical, neurodevelopmental, behavioral, or learning disability.[6] The cost of long care burdens the families economically, emotionally, and psychosocially.[6]

Screening for asymptomatic bacteriuria and its subsequent treatment has shown reduction in its adverse outcomes, especially in the risk of pyelonephritis,[37] preterm delivery, and low birth weight babies.[38] Based on the current evidence and expert opinions, Infectious Diseases Society of America[3] and US Preventive Service Task Force Recommendation[9] recommend urine culture for screening every pregnant woman for asymptomatic bacteriuria at least once between 12 and 16 weeks’ gestation or during the first prenatal visit. If the test results come positive, the woman should be treated with antimicrobials.

An estimated 213 million pregnancies occur in a year worldwide[10] and 60% of these attend antenatal checkups.[11] The recommended test for screening this large population for asymptomatic bacteriuria, urine culture, is hardly suitable for this task. It is resource intensive, unaffordable, and not available[12] in low-resource settings where the majority of pregnant women come for antenatal checkups. It needs a laboratory setup[13] with incubator, autoclave, laminar airflow and dedicated air-conditioned space, resources for streaking, performing reactions for identification, and conducting disk-diffusion for antibiotic susceptibility. It is not user-friendly and needs trained personnel to execute the test, making it even more difficult for grass-root staff to provide this test. Further, it takes 48–72 h to give results when done manually and 24 h when expensive automated systems are used. It is difficult for a woman to come back to the hospital on the 2nd or 3rd day, when the culture and sensitivity reports come. For these reasons, screening each pregnant woman for asymptomatic bacteriuria requires a test which is rapid, valid, inexpensive, user-friendly, and which provides antibiotic susceptibility so as to facilitate evidence-based prescription for positive cases without needing a repeat visit.

Rationale

The rationale of this mini-review was that it was needed to know whether we have a suitable test for screening asymptomatic bacteriuria in large population with potential for providing evidence-based prescription quickly or there is a gap in research, so as to suggest a basis for further research.

Objective

The objectives of the review were to find out:

What are the tests researched for screening asymptomatic bacteriuria or diagnosing UTI in published literature?

Are they rapid, valid, affordable, and user-friendly for use in large population?

Do they screen or diagnose multiple bacteria and provide antibiotic susceptibility?

METHODS

This review is designed as a narrative review. Methods of conducting the review and eligibility criteria were documented in advance. Diagnostic accuracy studies researching diagnostic tests for UTI or screening tests for asymptomatic bacteriuria in humans as compared to urine culture were included. All types of published articles were considered irrespective of year or language of publication. Participants of any age group screened for asymptomatic bacteriuria or diagnosed with UTI in any level of care hospital were considered. The primary outcome measure was cases of UTI or asymptomatic bacteriuria correctly diagnosed by the tests.

Studies were identified by searching electronic database of Medline. Two searches were conducted on PubMed using the following search terms:

“Urinary tract infection” and “diagnostic test”

“Asymptomatic bacteriuria” and “screening test.”

No filters were applied. Eligibility assessment was done by two reviewers independently in an unblinded standardized manner by screening the titles and abstracts. Full-text publications were reviewed for selected papers. Disagreement between reviewers was resolved by consultation with other coauthors. Two studies required arbitration about selection, the remaining were agreed independently by the reviewers. The last search was run on June 19, 2019. An Excel-based data extraction sheet was developed and piloted on five randomly selected studies. One reviewer extracted data from included studies and the second author validated the extracted data. No disagreement was encountered in data extraction. Information was extracted from each included study on-

Characteristics of participants – age, presence of symptoms, cohorts

Details of intervention, name, automated or manual, intended use, comparator

Outcomes and results of tests conducted; cutoff points used

Study design, sample size, prospective nature.

The risk of bias in individual studies was assessed by looking at their sample size and prospective nature.

Sensitivity, specificity, likelihood ratios, or receiver operating characteristic were considered the principal summary measures. Sensitivity and specificity above 80% were considered as valid. Likelihood ratios were converted to sensitivity and specificity. Among the studies describing the similar intervention, the highest diagnostic accuracy was considered. Studies were reviewed for explicit mention of terms rapid, cost-effective/affordable, or user-friendly, and the description of index test was reviewed for determining these criteria.

Screening only one database and only published articles introduced publication bias. No additional analyses were performed.

RESULTS

Seventy-eight records were retrieved from the electronic database, while 25 studies[13141516171819202122232425262728293031323334353637] were retained for qualitative synthesis [Figure 1]. The researchers had studied 11 individual tests and 7 combinations of various tests [Figure 1]. The report characteristics and study characteristics are described in Table 1.

Figure 1

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram

Table 1

Characteristics of enrolled studies (n=25)

Report characteristics	n
Range of year of publication (n)
1970-2019	25
Language
English	24
French	1
Publication status
Published	25
Article type
Original Research Article	25
Study characteristics	n
Participants
Population studied
Pregnant women	14
Nonpregnant women	1
Children	3
Urinary incontinence patients	1
Patients posted for urology procedures	1
No specific population	5
Symptomatology of study population
Symptomatic patients	8
Asymptomatic subjects	15
Both	2
Interventions
Intended use of intervention
Diagnostic test	3
Screening test	22
Automation status of intervention
Automated test	6
Manual test	19
Contribution to evidencebased prescription
Identification of common uropathogens studied	2
Identification of only one bacterium (Escherichia coli) studied	1
Antibiotic susceptibility providing tests studied	2
Comparator (urine culture)
Cutoff for labeling positive (CFU/ml)
105	20
104	3
103	1
103-106	1
Study design (diagnostic accuracy study)
Sequence of testing
Simultaneous testing of reference and index test	23
Consecutive testing of index test after reference test (with leftover sample)	2
Prospective nature
Prospective studies	24
Retrospective study	1

CFU: Colony Forming Units

In one study, diagnostic accuracy was separately described for males and females, while in two of the studies, separate cohorts of asymptomatic and symptomatic patients were evaluated. Since the review is concerned with asymptomatic women, the diagnostic accuracy in asymptomatic patients and female patients were considered. Out of 15 tests studied in 25 papers [Table 2], all were rapid in nature, but only seven were valid, namely, genetic signature identification CAPTURE assay, microfluidics-based direct single-cell imaging test, biosensor assay and phenotypic antimicrobial susceptibility test, chromogenic Limulus amoebocyte lysate assay, Gram staining, urine particle analyser UF-1000i, and flow cytometry-based leukocyte count (UF 100). Out of these tests, review of required infrastructure as described in the papers revealed that Gram staining and chromogenic Limulus amoebocyte lysate assay were cost-effective for use in large population. While the study on chromogenic Limulus amoebocyte lysate assay explicitly mentioned that it was user-friendly, the procedure to conduct Gram stain was deemed to be user-friendly by the authors.

Table 2

PICOS characteristics of studies related to various tests

Study	Study population	Interventional test (intended use in study)	Comparator (cut-off used in study)	Study design characteristics	Outcome (sensitivity and specificity)
O’Keefe et al., 2019[15]	Not specified Symptomatic	Genetic signature identification CAPTURE assay Diagnostic test	Urine culture 105 CFU/ml	Prospective diagnostic accuracy study Sample size-40 Simultaneous testing	Sensitivity90.0% Specificity100.0%
Baltekin et al., 2017[17]	Not specified Symptomatic	Microfluidics-based direct single-cell imaging test Diagnostic test	Urine culture 104 CFU/ml	Prospective diagnostic accuracy study Sample size-49 Consecutive testing	Sensitivity86.3 to 100% Specificity85.8 to 100%
Duong et al., 2016[18]	Febrile children Symptomatic	Flow cytometrybased leukocyte count (100i) Screening test	Urine culture 105 CFU/ml	Prospective diagnostic accuracy study Sample size-1247 Simultaneous testing	Sensitivity99.5% Specificity80.6%
Moshaver et al., 2016[19]	Not specified Symptomatic	Flow cytometrybased bacterial count Screening test	Urine culture 103 to 106 CFU/ml	Prospective diagnostic accuracy study Sample size-209 Simultaneous testing	Sensitivity99.0% Specificity58.0%
Altobelli et al., 2017[21]	Not specified Symptomatic	Biosensor assay and phenotypic antimicrobial susceptibility test Diagnostic test	Urine culture 105 CFU/ml	Prospective diagnostic accuracy study Sample size-84 Consecutive testing	Sensitivity98.5% Specificity96.6%
Dai et al., 2014[22]	Nonpregnant women Symptomatic	Urine particle analyser UF1000i Screening test	Urine culture 105 CFU/ml	Prospective diagnostic accuracy study Sample size-272 Simultaneous testing	Sensitivity99.0% Specificity82.0%
Waisman et al., 1999[24]	Children Symptomatic	Uriscreen, dipstick and microscopy Screening test	Urine culture 105 CFU/ml	Prospective diagnostic accuracy study Sample size-121 Simultaneous testing	SensitivityUriscreen 100% Dipstick 97.1% Microscopy 88.6% Specificity Uriscreen 68.6% Dipstick 82.5% Microscopy 88.4%
Aigere et al., 2013[26]	Pregnant women Asymptomatic	Enhanced urinanalysis Screening test	Urine culture 105 CFU/ml	Prospective diagnostic accuracy study Sample size-150 Simultaneous testing	Sensitivity57.1% Specificity98.6%
Thakre et al., 2012[27]	Pregnant women Asymptomatic	Griess nitrite and microscopy Screening test	Urine culture 105 CFU/ml	Prospective diagnostic accuracy study Sample size-300 Simultaneous testing	SensitivityGriess nitrite79.3% Microscopy 72.4% Specificity Griess nitrite97.1% Microscopy 94.5%
Ullah et al., 2012[28]	Pregnant women Asymptomatic	Gram staining and leukocyte esterase with nitrite Screening test	Urine culture 105 CFU/ml	Prospective diagnostic accuracy study Sample size-600 Simultaneous testing	SensitivityGram staining 91.7% Leukocyte esterase with nitrite 25% Specificity Gram staining 97.2% Leukocyte esterase with nitrite 99.7%
Greeff et al., 2002[31]	Pregnant women Both asymptomatic and symptomatic	Uricult trio Screening test	Urine culture 105 CFU/ml	Prospective diagnostic accuracy study Sample size-not specified Simultaneous testing	Sensitivity81.0% Specificity55.0%
Chongsomchai et al., 1999[33]	Pregnant women Asymptomatic	Urine analysis Screening test	Urine culture 105 CFU/ml	Prospective diagnostic accuracy study Sample size-774 Simultaneous testing	Sensitivity18.4% Specificity97.2%
Robertson and Duff 1988[36]	Pregnant women Asymptomatic	Nitrite with leukocyte esterase Screening test	Urine culture 105 CFU/ml	Prospective diagnostic accuracy study Sample size-750 Simultaneous testing	Sensitivity of Combination 92.0% Specificity of Combination 95.0%
Nachum et al., 1986[37]	Pregnant women Asymptomatic	Chromogenic Limulus amoebocyte lysate assay Screening test	Urine culture 105 CFU/ml	Prospective diagnostic accuracy study Sample size-1039 Simultaneous testing	Sensitivity88.7% Specificity98.7%

CFU: Colony Forming Units, PICOS: Population, Intervention, Comparison, Outcome, Study Design

None of the 15 tests was equipped to provide both identification of common uropathogens as well as susceptibility to various antibiotics for evidence-based prescription at the time of the first contact itself. Biosensor assay identified Enterobacteriaceae but gave susceptibility to only ciprofloxacin. CAPTURE assay identified common uropathogens but did not provide antibiotic susceptibility. Microfluidics-based test identified only E. coli and provided susceptibility to only ciprofloxacin. Gram staining could differentiate Gram-positive bacteria from Gram-negative bacteria. While the rest of the 11 tests were used in the role of triage tests for initial diagnosis of UTI to reduce the burden of performing urine culture on every sample.

No combination of tests proved to be valid unequivocally. Nitrite and leukocyte esterase together showed validity in two studies but failed to show validity in five studies.

DISCUSSION

Many hospitals in low-resource settings, lack the resources to screen every pregnant woman using the urine culture, and sensitivity test. As a result, they either do not screen them at all or screen them with tests having low diagnostic accuracy.

In our review of the literature, none of the alternative tests tried in the past had all the characteristics of a suitable test for screening asymptomatic bacteriuria, namely, rapid, valid, user-friendly, affordable, and providing identification and antibiotic sensitivity. Most of the tests just provided an initial triage of UTI so as to reduce the need of running urine culture for every sample. A positive woman on screening had to still undergo urine culture and wait for 2 days for initiation of appropriate antibiotic therapy, which in reality never happens. The women fail to come back to the hospital on the 2nd or 3rd day, when the culture and sensitivity reports come, due to many factors like lack of conveyance, cost of transportation, lack of compliance, and discomfort of repeated visits. Hence, they fail to avail adequate treatment with desirable and effective antimicrobials.

This reveals a state of affairs where 40% women[11] do not turn up for antenatal checkups, among the rest, majority do not get an early diagnosis and initiation of appropriate antibiotics. This leaves an unmet need in adequately screening the pregnant women during early pregnancy for asymptomatic bacteriuria, contributing to adverse outcomes of pregnancy.

Similarly, in cases of symptomatic pregnant women, treating physicians are unable to timely find out the causative organism and susceptibility. They are forced to mostly rely on their clinical judgment and prescribe wide-spectrum antibiotics empirically. Nonevidence-based irrational and unnecessary drug prescription of antibiotics has raised the rate of development of antimicrobial resistance for several commonly used broad-spectrum antibiotics.[38]

Despite the limitations like not being an extensive search in various databases and unpublished articles, the review provided an insight into why pregnant women are not being screened routinely and how it is contributing to a rise in cases of prematurity.

According to United States Preventive Services Task Force (USPTF), it is urgently needed to develop a test, which can screen for asymptomatic bacteriuria and is less labor-intensive at a lower cost too.[9] The authors also conclude that for making this testing possible in all health-care settings, further research is needed to provide an affordable, easy, rapid, and valid test for identification of bacteria and antibiotic sensitivity. Failing to do so will leave asymptomatic bacteriuria undiagnosed and will cause adverse outcomes of pregnancy.

Financial support and sponsorship

This study was funded by GYTI-SRISTI grant of Biotechnology Industry Research Assistance Council (BIRAC), Department of Biotechnology (DBT), Government of India.

Conflicts of interest

There are no conflicts of interest.