☰ Menu

8th Postgraduate Course for Training in Reproductive Medicine and Reproductive Biology

Review of Two Rapid Screening Tests for
Asymptomatic Bacteriuria during Pregnancy

E.J. Abalos
Centro Rosarino de Estudios Perinatales (C.R.E.P.), Rosario, Argentina
Maternidad «Martin», Rosario, Argentina

Objectives: To determine the clinical effectiveness of nitrite and leukocyte esterase reagent strip tests as screening methods for predicting asymptomatic bacteriuria during pregnancy, confirmed by urinary culture.

Search strategy: A computerised Medline search was conducted to identify all relevant studies using the words «bacteriuria», «urinary tract infection», «asymptomatic», «pregnan*», «diagnosis», «screening» and «sensitivity», in different combination. For each article identified a search of the references and the references of the references was performed.

Selection criteria: Case series or cohorts studies that evaluate nitrite and/or leukocyte esterase reagent strip test as screening methods for predicting bacteriuria in asymptomatic pregnant women. Results of the tests were compared with an adequate gold standard (i.e. urinary culture). Data that could be extracted to perform 2x2 tables.

Data collection and analysis: Six relevant studies were included and the results were stratified in four groups: nitrite test, leukocyte esterase test, both tests positive and either one test or the other positive. The summary estimates of the likelihood ratio for positive and negative test results, with their 95% confidence intervals (CIs), were computed in each stratum according to a random-effects model. Sensitivity was combined across studies using a weighted mean and 95%CI was calculated.

Main results: Summary sensitivity was higher when either nitrite or leukocyte esterase test were positive (80.2%, 95%CI 46.1-100). All summary likelihood ratios for a positive test were high. When either nitrite or leukocyte esterase test results were considered, the summary likelihood ratio for a negative test was lower, but with a wide confidence interval (0.2, 95%CI 0.02-2.0). Furthermore, in this strata there was significant heterogeneity among the studies (chi-square: 23.48, df=1).

Conclusions: Although combined tests seem to be promising in detection of bacteriuria in asymptomatic pregnant women there is insufficient evidence for reassuring clinicians and patients that a negative result is a true negative one. Further trials need to be conducted with adequate sample size in order to assess the effectiveness of combined tests as screening methods for predicting asymptomatic bacteriuria during pregnancy.

Background

Urinary tract infection (UTI) is a relatively common medical complication of pregnancy and asymptomatic bacteriuria (defined as bacterial colonisation of the urinary tract without urinary tract symptomatology) is the most prevalent of these infections (1, 2). The prevalence of this condition in pregnancy is similar to that seen in non-pregnant women; therefore, pregnancy alone is not believed to be a predisposing factor in the development of asymptomatic bacteriuria (1, 3, 4). Many studies have shown the overall prevalence of bacteriuria in pregnancy ranges from 4% to 7% (5, 6, 7, 8, 9), although in certain sub-populations lower or higher rates were reported (10, 11). These differences could be explained by risk factors such as age, sexual activity, parity, socio-economic status or history of recurrent UTIs (12, 13). The natural history of bacteriuria in pregnancy has also been extensively evaluated. It has been shown that, when untreated, 20% to 40% of mothers with asymptomatic bacteriuria will develop acute pyelonephritis (2, 3, 9). In a systematic review of randomised controlled trials Smaill demonstrated by treating asymptomatic bacteriuria in pregnancy a substantial reduction in developing pyelonephritis (RR=0.25, 95%CI 0.19, 0.34) (14). Many complications of pregnancy have been attributed to UTIs during gestation, including preterm labour and delivery, low birth weight and growth retardation (15, 16), and the beneficial effect of treatment of pregnant women with asymptomatic bacteriuria has been established (14). Other complications of pregnancy associated with bacteriuria, such as anaemia, hypertension or long-term renal function impairment, are controversial.

A strategy of screening for asymptomatic bacteriuria is currently included in routine prenatal care. It is important to detect those patients that must be treated in order to avoid the development of symptomatic infection. In a large Swedish study, the optimal time of screening for bacteriuria during pregnancy was evaluated. The authors recommend screening all women early in pregnancy, particularly during the sixteenth gestational week (9). Although women with bacteriuria often have a history of UTIs, the history has low sensitivity to identify a group that should be screened (13).

The traditional reference test for bacterial UTI is the quantitative culture of urine (17). However, this method is relatively expensive, costing approximately $20.00 to $50.00 per sample (18, 19, 20). Additionally, this test is time consuming and labour intensive and needs the infrastructure of a microbiology laboratory and of qualified and trained staff. This situation, however desirable, is not available in many areas of the world, especially in regions where the consequences of preterm delivery, low birth weight, and all the complications of infectious diseases are associated with higher morbidity and mortality.

While most authors agree on the need for early urine screening, the best screening test to be used remains to be determined. However, a number of quick inexpensive office-based tests have been developed to be used as screening tests for bacteriuria (17, 21). Since the last two decades their efficacy as screening method was evaluated (17, 21, 22). However, the target study population of these studies was heterogeneous, since some studies had evaluated children, men, elderly people, patients from general clinics or inpatients and outpatients referred to specialised hospitals. Moreover, there were differences between studies in including women with or without symptoms of UTI. Also, during pregnancy the same uncertainty at the time of evaluating the tests remains.

Screening for disease control or prevention is «the examination of asymptomatic people in order to classify them as likely or unlikely to have the disease that is the object of screening. People who appear likely to have the disease are investigated further to arrive at the final diagnosis (...) Screening simply calls attention to the likelihood of the disease before symptoms appear» (23). In this context, tests with high sensitivity should be preferred. The positive and negative predictive values also need to be assessed, in order to evaluate the risk on individual patients, given the test results. Because predictive values directly depend on prevalence of the disease, this cannot be combined directly across studies.

The objective of this study is to evaluate the effectiveness of screening asymptomatic pregnant women for bacteriuria by two methods, the nitrite and leukocyte esterase reagent trips. This will be evaluated by performing a systematic review of the available evidence.

Objectives

The objective of this review is to determine the clinical effectiveness of two rapid tests (nitrite and leukocyte esterase reagent strips) as screening methods for predicting asymptomatic bacteriuria, confirmed by urinary culture during pregnancy.

Criteria for considering studies for this review

  • Types of participants: Pregnant women presenting to their antenatal care visit without symptoms of UTI.
  • Types of intervention: A study was eligible for inclusion if it evaluates the effectiveness of nitrite or/and leukocyte esterase tests to detect bacteriuria in asymptomatic pregnant patients.
  • Types of outcome measures: A study was eligible for inclusion if results of a positive and of a negative test were available in the report, compared with an appropriate gold standard.
  • Types of studies: All studies that use one or both methods for screening asymptomatic pregnant women for bacteriuria, and compare their results with a gold standard.

Search strategy for identifications of studies

A computerised Medline search was conducted to identify all studies that compared the effectiveness of nitrite and / or leukocyte esterase tests as screening method in asymptomatic pregnant women published before September 1998.

The methods for identifying trials was as follows:
  • 1- BACTERIURIA
  • 2- URINARY TRACT INFECTION
  • 3- #1 or #2
  • 4- ASYMPTOMATIC
  • 5- #3 and #4
  • 6-PREGNAN*
  • 7- #5 and #6
  • 8- #3 and #6
  • 9- DIAGNOSIS
  • 10- DIAGNOSTIC METHODS
  • 11- SCREENING
  • 12- #9 or #10 or #11
  • 13- #7 and #12
  • 14- #8 and #12
  • 15- #13 or #14

An additional MEDLINE search using two different Internet search services Pub Med and Grateful Med was done using the words «bacteriuria» or «urinary tract infection» and «asymptomatic» and «pregnan*», in one hand, and «bacteriuria» or «urinary tract infection» and «pregnan*» and «diagnosis» or «screening» or «sensitivity», in the other. The first strategy retrieved 245 citations, the other two 316 and 1209, respectively. All the citations in the bibliography of identified reports were evaluated for eligible articles, and papers that review the tests either in a general population or in pregnant women were evaluated in order to identify potentially eligible studies.

For each article identified a search of the references and the references of the references was performed.

Of the 147 studies selected at first, 52 were review articles. 54 studies used nitrite or leukocyte esterase tests to screen patients of general population and 27 articles referred to other screening devices used in obstetric patients. Thus, all were rejected. From the sixteen remaining articles six were finally selected because of the eligibility criteria defined for this review. (see characteristics of included and excluded studies).

No contacts with the authors were performed.

Methods of review

Reviewer was not blinded to the authors or sources of the articles. The criteria used assessing the quality of the trials were the following:
  • an independent and blinded comparison with an adequate gold standard. (a positive culture showing a single bacterial growth equal or more than 100 000 colonies).
  • the sample of patients included an adequate spectrum of women in whom the test will be performed in clinical practice. (for the objectives of this review: asymptomatic pregnant women).
  • the results of the evaluated tests had any influence in the decision of performing the gold standard.
  • the methods tested had been described with enough details to permit their reproduction.

All criteria were rated as met, unmet or unclear. The following data were taken in consideration and were extracted from each study: number of patient enrolled, gestational age, socio-economic status, exclusions (for any cause).

Likelihood ratios for positive and negative tests results were computed in each study, along with their 95% confidence intervals (95%CI)(24). For each stratum, a X2 test of heterogeneity was computed before the summary likelihood ratio was calculated. A random-effects model was used to generate a summary estimate of the likelihood ratio. Sensitivities were calculated for each trial. Data were combined across studies using a weighted mean of these sensitivities. A formula which has been described for the estimation of precision in cluster sampling was used to calculate the 95% confidence interval of these summary sensitivities (25). The computation of this confidence interval takes into account the sum of the inter-study and intra-study variances.

Description of studies

All six studies were conducted in industrialised countries (5 in the United States and 1 in United Kingdom). The articles were published between 1984 and 1998. Prevalence of bacteriuria ranged from 2.3% to 9.3 %, and socio-economic status was low, middle and high across the studies. Patients were enrolled at their first antenatal visit. Gestational age, reported in three studies (26, 27, 30) ranged from 9 to 17 weeks. Samples of urine were collected at the clinic in all studies. The reagent strips used as screening test were CHEMISTRIP (Byodynamics BMC Division of Boehringer Mannheim Corp, Indianapolis, In) (27, 28, 29), MULTISTIX (Bayer Diagnostic, Basingstoke) (26, 30) and MICROSTIX (no more specifications) (31). Results of nitrite test were presented in all studies, leukocyte esterase activity was tested in four studies (26, 27, 28, 29), but in one of them results could not be extracted because they were presented in combination with nitrites, proteins and blood (26). Results of Leukocyte esterase activity and presence of nitrites were presented in four strata (Nitrite positive, Leukocyte esterase positive, both tests positive and either one or the other positive) in two studies (27, 28). Gold standard was defined as formal urine culture in all trials, and positive results were established as a growth of 100 000 Colony Forming Units/ml in five studies (26, 27, 28, 29, 31). The study Abbasi (30) only mentioned formal quantitative cultures.

Methodological qualities of included studies

Blinding between the assessment of tests and the evaluation of gold standard was performed in two studies (27, 30). In one paper (29), quantitative definition of the gold standard (number of CFU/ml to consider the culture as positive) was no mentioned.

All studies included asymptomatic pregnant women, but only one, reported the exclusion of symptomatic patients. (26). In two studies samples were excluded because of contamination (26, 31). Characteristics of population were not listed in some studies, i.e., gestational age was presented in three studies (26, 27, 30), and socio-economic status was described only in three other papers (27, 28, 30).

Results of the evaluated tests had not influenced the decision of performing the gold standard in any of the studies selected for this review (absolute exclusion criteria).

Explanation of the sample size calculation was no reported in any of these studies.

Results

A total of 3341 asymptomatic pregnant women who had a nitrite test were included in six studies (Table I). A summary likelihood ratio for a positive test result was 48.5 (95% CI 23.1-102.1). The summary likelihood ratio for a negative test result was 0.6 (95%CI 0.5-0.8). Three studies evaluated leukocyte esterase tests, involving 1862 patients (Table II). The summary likelihood ratios for a positive test result and for a negative test result were 13.2 (95%CI 6.7-26.0) and 0.4 (95%CI 0.1-2.1), respectively. Either leukocyte esterase or nitrite tests positive were evaluated in two studies (1797 patients) (Table III). The summary likelihood ratio for positive test result was 17.8 95%CI 13.4-23.6), and for negative test result was 0.21 (95%CI 0.1-2.0). In these two studies both were evaluated, a positive nitrite and leukocyte esterase test. As they reported no false positive results (100% specificity), likelihood ratios were not calculated because these results become unreliable. The summary sensitivity was 37.6% (95%CI 23.4-51.7) for nitrite tests, 61.8% (95%CI 21.2-100) for leukocyte esterase tests, 26.7% (95%CI 8.6-44.9) when both tests were positive, and 80.2% (95%CI 46.1-100) when either one test or the other was positive.

As prevalence of asymptomatic bacteriuria across the studies ranged from 2.3 to 9.3, summary predictive values were no estimated.

Summary of analyses

Table I. Characteristics and results of nitrite reagent strip tests performed in pregnant women asymptomatic for bacteriuria to assess the prediction of bacteriuria.

First author Prevalence of Bacteriuria Nº of Women Sensitivity (95%CI) Likelihood ratio
for positive Test results
Likelihood ratio
for negative Test results
Tincello, 1998 5.4% 893 18.8 39.6 (12.6-124) 0.8 (0.7-0.9)
Bachman, 1993 2.3% 1047 45.8 156.3 (46.6-524.5) 0.5 (0.4-0.8)
Robertson, 1988 8.3% 750 43.4 42.8 (19.4-94.3) 0.6 (0.5-0.7)
Abbasi, 1985 4.6% 65 66.7 78.8 (4.5-1379.6) 0.3 (0.1-1.7)
Campos-Outcalt, 1985 4.7% 299 57.1 18.1 (8.2-39.6) 0.4 (0.2-0.8)
Archbald, 1984 9.3% 287 36.7 191.4 (11.6-3169.4) 0.6 (0.5-0.8)
Summary 37.6 (23.4-51.7) 48.5 (23.1-102.1) 0.6 (0.5-0.8)
X2 Heterogeneity 7.7 10.7 (df=5) 17.4 (df=5)

Table II. Characteristics and results of leukocyte esterase reagent strip test performed in pregnant women asymptomatic for bacteriuria to assess the prediction of bacteriuria.

First author Prevalence of Bacteriuria Nº of Women Sensitivity (95%CI) Likelihood ratio
for positive Test results
Likelihood ratio
for negative Test results
Bachman, 1993 2.3% 1047 16.7 5.9 (2.2-15.4) 0.9 (0.7-1.0)
Robertson, 1988 8.3% 750 77.4 18.9 (12.8-28.1) 0.2 (0.2-0.4)
Abbasi, 1985 4.6% 65 100 15.5 (6.0-40.0) 0.1 (0.01-1.8)
Summary 61.8 (21.2-100) 13.2 (6.7-26.0) 0.4 (0.1-2.2)
X2 Heterogeneity 11.7 4.9 (df=2) 73.0 (df=2)

Table III. Characteristics and results when either nitrite or leukocyte esterase react in urine samples of pregnant women asymptomatic for bacteriuria.

First author Prevalence of Bacteriuria Nº of Women Sensitivity (95%CI) Likelihood ratio
for positive Test results
Likelihood ratio
for negative Test results
Bachman, 1993 2.3% 1047 50.0 16 (9.4-27.0) 0.5 (0.4-0.8)
Robertson, 1988 8.3% 750 92.0 18.6 (13.3-26.0) 0.1 (0.02-2.0)
Summary 80.2 (46.1-100) 17.8 (13.4-23.6) 0.2 (0.02-2.0)
X2 Heterogeneity 3.8 0.2 (df=1) 23.5 (df=1)

Discussion

Asymptomatic bacteriuria during pregnancy has serious consequences. Routine screening and antimicrobial treatment of confirmed positive cases is recommended. Although there is no doubt that formal urine culture is the best method to confirm presence of bacteria in urine samples, this method is not feasible for many areas of the world, especially in developing countries. An effective, non expensive and easy to perform office screening method is needed in those rural areas where specialised laboratories are not available. Patients at risk of having bacteriuria could be referred to central Hospitals for further analysis to confirm the disease.

These tests have to be sensitive with high positive predictive value, in order to detect all patients that should be referred, without missing positive cases. The likelihood ratio is useful for the comparison of test performance and the computation of predictive values in populations with varying prevalence of the disease. As an example, in a population with a baseline risk of 6%, an asymptomatic woman with negative nitrite and leukocyte esterase test result (likelihood ratio 0.2) has a risk of 1,2% to have bacteriuria (24). That means that every 100 patients tested by the screening method, five will be detected and one will be missed. In a population with prevalence of bacteriuria of 15% and a negative test result, the calculated post-test probability is 3.4%, which is not completely reassuring. However these data has to be interpreted with extreme caution because ofits wide confidence intervals (for the test of the example: 95%CI 0.02-2.0).

Conclusion

Implications for practice

Multiple tests (i.e. combination of nitrite and leukocyte esterase when at least one test is positive) seem to be promising in detecting bacteriuria in asymptomatic pregnant women. However there is insufficient evidence reassuring that a negative result is a truly negative.

Implications for research

Further investigation has to be conducted with adequate sample size in order to assess the effectiveness of combined tests as screening methods for predicting asymptomatic bacteriuria during pregnancy.

Potential conflict of interest

Not known.

Acknowledgements

Dr. Olivier Irion, Dr. Michel Boulvain

Characteristics of included studies

Study Tincello, 1998 (26)

Methods: Prospective case series. No reports were done about blindness in the assessment of the test and the gold standard.

Participants: 893 pregnant asymptomatic women in their first antenatal visit. Gestational age ranged from 12 to 14 weeks. Exclusion of symptomatic patients and those who were under antibiotic treatment for any reason (n= 35). Exclusion of contaminated samples (n=32). No further data about socio-economic status or characteristics of population were reported. The prevalence of bacteriuria was 5.4%.

Interventions: Midstream specimen of urine obtained in the clinic. The reagent strip test was Ames 8SG Multistix; Bayer Diagnostic, Basingstoke. Sampled were tested by midwifery staff (between 30-60 seconds). A positive test was defined when the strip showed any of the following: more than a trace of protein, more than a trace of blood, any positive result for nitrite, any positive result for leukocyte esterase. Authors present the results in two strata: Nitrite test (+ or -), and all determinations combined, defined as positive if any of them was positive. Formal urine cultures were defined as: positive if they showed more than 100 000 colony forming units (CFU)/ml of urine, contaminated if they showed a mixed culture of any density or a pure culture of less than 100.000 CFU/ml, and negative if there was no growth of bacteria. No specifications were done about time spent before performing cultures.

Outcomes: Positive or negative test. Sensitivity, specificity, positive and negative predictive values of the test in diagnosing bacteriuria.

Study Bachman, 1993 (27)

Methods: Prospective case series. Test results were interpreted without knowledge of other test results. This study contains a second part in where a cohort of patients was followed up by screening tests in each antenatal visit, but these results were not confronted with their respective gold standard (cultures were performed only with a positive test). For this reason only the first part of the study was considered in this review.

Participants: 1047 pregnant asymptomatic women in their initial prenatal care. Gestational age ranged from 7 to 17 weeks. No data reported about exclusions (either for the presence of symptoms or for contaminated cultures). They described the population as mainly white, older than 25 years, with more than 16 years of education and higher family incomes than the national average. The prevalence of bacteriuria was 2.3%.

Interventions: Clean-voided midstream urine specimens collected in the clinic. The reagent strip test was Chemistrip-9 (Biodynamics BMC Division of Boehringer Mannheim Corp, Indianapolis, Ind). The test was defined as positive if either leukocyte esterase activity or presence of nitrites or both were detected. All laboratory testing was performed by medical laboratory technicians and both, screening tests and cultures, were processed within 45 minutes from collection. Cultures were defined as positive if they showed 100 000 CFU/ml or more. A mixed culture with one predominant organism of more than 100 000 uropathogens/ml was considered as positive.

Outcomes: Positive or negative test. Sensitivity, specificity, positive and negative predictive values were presented in four strata: nitrite, leukocyte esterase, both test positive, and either one test or the other positive.

Study Robertson, 1988 (28)

Methods: Prospective case series. Blindness between the assessments of the results of tests and cultures was not mentioned. Specimen were obtained by random voiding.

Participants: 750 pregnant asymptomatic women at their initial appointment. Gestational age was not mentioned. No data reported about exclusions (presence of symptoms or contaminated cultures). Population was described as middle class with unlimited access to medical care. Prevalence of asypmtomatic bacteriuria was 8.3%.

Interventions: Clean-catch midstream urine specimens collected in the clinic. Results were read at 60 seconds using Chemstrip LN dipstick. The test was defined as positive with a positive result of any of its components (nitrite or leukocyte esterase portion). Cultures were defined as positive if two consecutive cultures showed at least 100 000 CFU/ml of a single uropathogen. Contaminated cultures (single or multiple pathogens with a colony counts less than 100 000 CFU/ml) were considered as negative. No specifications were done about time spent before performing cultures.

Outcomes: Positive or negative test. Sensitivity, specificity, positive and negative predictive values were presented in four strata: nitrite, leukocyte esterase, both test positive, and either one test or the other positive.

Study Abbasi, 1985 (29)

Methods: Prospective case series. Blindness between the screening methods and the gold standard was not mentioned.

Participants: 65 consecutive women at their initial obstetric visit. Gestational age and characteristics of the population were not reported. No data about exclusion were presented. Prevalence of asymptomatic bacteriuria was 4.6%.

Interventions: Clean-catch midstream urine specimens were collected at the clinic. Leukocyte esterase activity and presence of nitrites were determined by Chemstrip 9 (Byodynamics) at 60 seconds. Cultures were defined only as positive or negative.

Outcomes: Positive or negative test.

Study Campos-Outcalt, 1985 (30)

Methods: Prospective case series. Even the laboratory personnel that perform the cultures than the clinic aides testing for nitrites were unaware of the results of the other test.

Participants: 299 asymptomatic pregnant women recruited in their initial prenatal examination. The mean estimated gestational age was 17 weeks. Population as described as low-income, 45% Hispanic and 13% black. No data reported about exclusions (symptomatic patients or contaminated cultures). The prevalence of bacteriuria was 4.7%.

Interventions: Midstream urine samples provided at the clinic were tested for nitrite with N-Multistix and results were recorded as either positive or negative. No further explanations about the methods were reported. The presence of organisms exceeding 100 000 CFU/ml was considered a positive culture. No more description about cultures was reported. They performed also a request about previous history of UTI, and combined the results with the nitrite tests. For the purpose of this study, those data had not been taken under consideration in this review.

Outcomes: Positive or negative test. Sensitivity, specificity, positive and negative predictive values of the test in diagnosing bacteriuria.

Study Archbald, 1984 (31)

Methods: Prospective case series. Blinding, regarding the assessment of the test results and the cultures was not mentioned.

Participants: 324 asymptomatic pregnant women registered for their first antenatal visit. Gestational age or any other characteristic of the population was not reported. Exclusion of symptomatic patients was not mentioned. 11.4% of the samples (n=37) were excluded from the analysis because of contamination. The prevalence of asymptomatic bacteriuria was 9.3%.

Interventions: Clean-catch randomly voided specimens were collected without specifications if those samples were collected in the clinic or at home. Tests were performed with Microstix-3 to search for nitrites within 3 hours of collection, and results were registered as positive or negative. Positive cultures were defined as having a colony count of 100 000 or more. Contaminated cultures were defined as having a growth of three or more organisms. No information was available on the time interval before cultures were performed.

Outcomes: Positive or negative test. Sensitivity, specificity, positive and negative predictive values of the test in diagnosing bacteriuria.

Characteristics of excluded studies

Study Hagay, 1996 (32)

This study included symptomatic and asymptomatic patients coming from both, inpatent and outpatient clinics and the results were not presented separately. Furthermore, nitrite and leukocyte esterase tests were not performed on all the specimens, because priority was given to another test (Uriscreen) if the sample volume was not sufficient to perform all tests. There was no information on independence and blindness between the evaluation of the results of the reagent strips and the gold standard.

Study Maillard, 1994 (33)

This study asymptomatic pregnant women at 30 weeks of gestation were enrolled and screened for bacteriuria by nitrite and leukocyte esterase reagent strips. During the study, patients performed tests at home, one in the interval between antenatal visits and another in the morning of their clinic appointment. Only after 5 tests had been performed by the patient, the results were compared with the gold standard. Also, a different test was performed during each antenatal visit and samples were only cultured in case of a positive test result. Cultures were defined as positive with a growth of 10 000 or more organisms.

Study Etherington, 1993 (34)

This study included symptomatic and asymptomatic pregnant patients («...both, for routine bacteriological investigation and where clinically indicated»). No information about number of patients of each group was available and the outcomes were not presented separately. The reagent strips were read in two different ways, visually using the reference code of the bottle during the first three months and photometrically in the last two months. The results were combined when the outcomes were assessed. Blinding for evaluating the strips and the gold standard was not mentioned..

Study Lorentzon, 1990 (35)

This study included symptomatic and asymptomatic pregnant patients without stratification in the assessment of both groups. There was no total independence between the results of the test and the gold standard: for some samples (but not in all) a second gold standard method(another culture for confirmation) was performed, according to the results of the test. Blinding of the personnel that evaluate the strips and the cultures was not mentioned. No data was available on the population characteristics.

Study McNeeley, 1987 (36)

This study analysed a subgroup of pregnant women without data if they were asymptomatic or not. Furthermore, urine samples were treated with a preservative (glycerin-boric acid-sodium formate preservative) before performing both cultures and tests. Blinding between assessment of strips and cultures was not mentioned. The gold standard was defined in a different way (10 000 or more CFU/ml of two or fewer probable pathogens for which antimicrobial therapy was indicated).

Study Soisson, 1885 (37)

This study is a retrospective review of medical records. The charts of 1062 obstetric patients were reviewed. The results of the urinalyses were compared to those of the subsequent urine cultures. No data were available about the methods used or about the women enrolled (eg if they were pregnant or postpartum patients)

Study Marquette, 1985 (38)

This study included symptomatic and asymptomatic patients without specification of the prevalence in each group or description of the characteristics of the patients. Another gold standard test(presence of leukocyturia by microscopy) was used as confirmatory test.

Study Lenke, 1981 (39)

In this study pregnant women at high risk of having UTIs were included. All of them had been hospitalised previously and treated for pyelonephritis during the current pregnancy. It was not mentioned if any symptoms were present during the time of the study.and more than one urine sample was taken from each patient without further specification. The prevalence of UTI in this population was 12,3%, but a high rate of contaminated cultures (11,6%) was reported. Blinding between the evaluation of the strips and the cultures was not reported.

Study Czerwinski, 1971 (40)

This study did not mention if the samples were taken from women with urinary symptoms or not.. But the main reason to exclude this study was that the used screening method was a modified Griess test (a laboratory technique), instead of reagent strips.

Study Finnerty, 1968 (41)

This study used a modified Griess test as a screening method to detect nitrites in urine of asymptomatic pregnant women. Also, another gold standard was defined as confirmatory method.

References

  1. Millar LK, Cox SM: Urinary tract infections complicating pregnancy. Infect Dis Clin North Am., 1997 Mar; 11(1): 13-26.
  2. Patterson TF, Andriole VT: Detection, significance and therapy of bacteriuria in pregnancy. Infect Dis Clin North Am., 1997 Sept; 11(3): 593-608.
  3. Andrews W, Cox S, Gilstrap LC: Urinary tract infections in pregnancy. Int Urogynecol J, 1990; 1: 155-163.
  4. Bengtsson C, Bengtsson U, Lincoln K: Bacteriuria in a population sample of women. Acta Med Scand, 1980; 208: 417-423.
  5. Kincaid-Smith P, Bullen M, Mills J, Fussell U, Huston N: The reliability of screening test for bacteriuria in pregnancy. Lancet, 1964 Jul; i: 61-62
  6. Campbell-Brown M, McFayden IR, Seal DV, Stephenson ML: Is screening for bacteriuria in pregnancy worth while? Br Med J, 1987 Jun; 294: 1579-1582.
  7. Lawson DH, Millier AWF: Screening for bacteriuria in pregnancy. Lancet, 1971 Jan; i: 9-11.
  8. Dempsey C, Harrison RF, Moloney A, Darling M Walshe J: Characteristics of bacteriuria in a homogeneous maternity hospital population. Eur J Obstet Gynecol Reprod Biol, 1992; 44(3): 189-193.
  9. Stenqvist K, Dahlen-Nilsson I, Lidin-Janson G, Lincoln K, Oden A, Rignell S, Svanborg-Eden C: Bacteriuria in pregnancy. Frequency and risk of acquisition. Am J Epidemiol, 1989; 129: 372-379.
  10. De la Rosa M, Rojas A, Garcia V, Herruzo A, Moreno I: Asymptomatic bacteriuria and piuria during pregnancy. Enferm Infecc Microbiol Clin, 1994; 12(2): 79-81.
  11. Versi E, Chia P, Griffiths DJ, Harlow BL: Bacteriuria in pregnancy: a comparison of Bangladeshi and Caucasian women. Int Urogynecol J Pelvic Floor Dysfunct, 1997; 8(1): 8-12
  12. Lucas MJ, Cunningham FG: Urinary infection in pregnancy. Clin Obstet Gynecol, 1993; 36: 855.
  13. Chng PK, Hall MH: Antenatal prediction of urinary tract infection in pregnancy. Br J Obstet Gynaecol, 1982 Jan; 89: 8-11.
  14. Smaill F: Antibiotic versus no treatment for asymptomatic bacteriuria in pregnancy. Pregnancy and Childbirth Module of the Cochrane Database of Systematic Reviews. Updated 1997.
  15. Romero R, Oyarzun E, Mazor M, Sirtori M, Hobbins JC, Bracken M: Meta-analysis of the relationship between asymptomatic bacteriuria and preterm delivery / low birth weight. Obstet Gynecol, 1989; 73: 576-582.
  16. Meis PJ, Michielutte R, Peters TJ, Wells HB, Sands E, Coles EC, Johns KA: Factors associated with preterm birth in Cardiff, Wales. II. Indicated and spontaneous preterm birth. Am J Obstet Gynecol, 1995; 173: 597-602.
  17. Kellogg JA, Manzella JP, Shaffer SN, Schwartz BB: Clinical relevance of culture versus screens for the detection of microbial pathogens in urine specimens. Am J Med, 1987; 83: 739-745.
  18. Mac Millan MC, Grimes DA: The limited usefulness of urine and blood cultures in treating pyelonephritis in pregnancy. Obstet Gynecol, 1991; 78: 745-748.
  19. Wadland WC, Plante DA: Screening for asymptomatic bacteriuria in pregnancy. A decision and cost Analysis. J Fam Pract, 1989; 29(4): 372-376.
  20. Rouse DJ, Andrews WW, Goldenberg RL, Owen J: Screening and treatment of asymptomatic bacteriuria of pregnancy to prevent pyelonephritis: A cost-effectiveness and cost-benefit analysis. Obstet Gynecol, 1995; 86(1): 119-123.
  21. Hurlbut III TA, Littenberg B, et al: The diagnostic accuracy of rapid dipstick tests to predict urinary tract infection. Am J Clin Pathol, 1991; 96:582-588.
  22. Pollock H: Laboratory techniques for detection of urinary tract infection and assessment of value. Am J Med, 1983; 75 (Suppl 1B): 79-84.
  23. Morrison AS: Screening. Modern Epidemiology, Chp.25: 499-518
  24. Sackett DL, Haynes RB, Guyatt GH, Tugwell P: Clinical Epidemiology: a basic science for clinical medicine. 2nd ed. Toronto: Little, Brown, 1991.
  25. Donner A, Klar N: Methods for comparing event rates in intervention studies when the unit of allocation is a cluster. Am J Epidemiol, 1994; 140: 279-289.

References to studies included in this review

  1. Tincello DG, Richmond DH: Evaluation of reagent strips in detecting asymptomatic bacteriuria in early pregnancy: prospective case series. Br Med J, 1998; 316: 435-437.
  2. Bachman JW, Heise RH, Naessens JM, Timmerman MG: A study of various tests to detect asymptomatic urinary tract infections in an obstetric population. JAMA, 1993; 270(16): 1971-1974.
  3. Robertson AW, Duff P: The nitrite and leukocyte esterase tests for the evaluation of asymptomatic bacteriuria in obstetric patients. Obstet Gynecol, 1988; 71 (6): 878-881.
  4. Abbasi IA, Hess LW, Johnson TR, Mc Fadden E, Chernow B: Leukocyte esterase activity in the rapid detection of urinary tract and lower genital tract infections in obstetric patients. Am J Perinatol, 1985; 2(4): 311-313.
  5. Campos-Outcalt DE, Corta PJ: Screening for asymptomatic bacteriuria in pregnancy. J Fam Pract, 1985; 20(6): 589-591.
  6. Archbald FJ, Verma U, Tejani NA: Screening for asymptomatic bacteriuria with Microstix. J Reprod Med, 1984; 29(4): 272-274.

References to studies excluded in this review

  1. Hagay Z, Levy R, Miskin A, Milman D, Sharabi H, Insler V: Uriscreen, a rapid enzymatic urine screening test: useful predictor of significant bacteriuria in pregnancy. Obstet Gynecol, 1996; 87 (3): 410-413.
  2. Maillard F, Breart G, Uzan S: Evaluation of urinary rapid strip-tests for screening of urinary infection in pregnant women. J Gynecol Obstet Biol Reprod (Paris), 1994; 23(8):909-913.
  3. Etherington IJ, James DK: Reagent strip testing of antenatal urine specimens for infection. Br J Obstet Gynaecol, 1993; 100: 806-808.
  4. Lorentzon S, Hovelius B, Miörner H, Tendler M, Aberg A: The diagnosis of bacteriuria during pregnancy. Scand J Prim Health Care, 1990; 8: 81-83.
  5. McNeeley SG, Baselski VS, Ryan GM: An evaluation of two rapid bacteriuria screening procedures. Obstet Gynecol, 1987; 69(4): 550-553.
  6. Soisson AP, Waston WJ, Benson WL Read JA: Value of a screening urinalysis in pregnancy. J Reprod Med, 1985; 30(8): 588-590.
  7. Marquette GP, Dillard T, Bietla S, Niebyl J: The validity of leukocyte esterase reagent test strip in detecting significant leukocyturia. Am J Obstet Gynecol, 1985; 153: 888-889.
  8. Lenke R, Van Dorsten JP: The efficacy of the nitrite test and microscopic urinalysis in predicting urine cultures results. Am J Obstet Gynecol, 1981; 140: 427-429.
  9. Czerwinki AW, Wilkerson RG, Merrill JA, Braden B, Colmore JP: Further evaluation of the Griess test to detect significant bacteriuria. Am J Obstet Gynecol, 1971; 110 (5): 677-681.
  10. Finnerty Jr FA, Johnson AC: A simplified accurate method for detecting bacteriuria. Am J Obstet Gynecol, 1968; 101(2): 238-242.