Response to antibiotics of women with symptoms of urinary tract infect
http://www.100md.com
《英国医生杂志》
1 Department of Public Health and General Practice, Christchurch School of Medicine and Health Sciences, PO Box 4345, Christchurch, New Zealand, 2 Department of Pathology, Christchurch School of Medicine and Health Sciences
Correspondence to: D Richards derelie.richards@chmeds.ac.nz
Objective To assess the effectiveness of antibiotic treatment of women with symptoms of urinary tract infection but negative urine dipstick testing.
Design Prospective, double blind, randomised, placebo controlled trial.
Setting Primary care, among a randomly selected group of general practitioners in Christchurch, New Zealand.
Participants 59 women aged 16-50 years presenting with a history of dysuria and frequency in whom a dipstick test of midstream urine was negative for both nitrites and leucocytes. Participants with complicated urinary tract infection were excluded.
Intervention Trimethoprim 300 mg daily for three days or placebo.
Main outcome measures Self reported diary of symptoms for seven days, recording the presence or absence of individual symptoms each day, followed by a structured telephone questionnaire after seven days. The main clinical outcome was resolution of dysuria at three and seven days and median time to resolution. Secondary outcomes were resolution of other symptoms.
Results The median time for resolution of dysuria was three days for trimethoprim compared with five days for placebo (P = 0.002). At day 3, five (24%) of patients in the treatment group had ongoing dysuria compared with 20 (74%) in the placebo group (P = 0.005). This difference persisted until day 7: two patients (10%) in the treatment group v 11 (41%) in the placebo group; P = 0.02). The number needed to treat was 4. The median duration of constitutional symptoms (feverishness, shivers) was reduced by four days.
Conclusions Although a negative dipstick test for leucocytes and nitrites accurately predicted absence of infection when standard microbiological definitions were used (negative predictive value 92%), it did not predict response to antibiotic treatment. Three days' treatment with trimethoprim significantly reduced dysuria in women whose urine dipstick test was negative. These results support the practice of empirical antibiotic use guided by symptoms. Balancing the competing interests of symptom relief and the minimisation of antibiotic use remains a dilemma—further research is needed to determine clinical predictors of response to antibiotics.
Infections of the urinary tract are extremely common, and the numbers of patients presenting to general practice with this condition represent considerable morbidity and workload.1 Because the symptoms of dysuria and frequency are unpleasant, doctors are under pressure from patients to provide relief.2 3 However, not all such symptoms are associated with growth of bacteria in a standard midstream urine specimen. More detailed microbiological investigation of women with apparently negative urine cultures on standard testing indicate that a proportion of them have low count bacteriuria.4 Urine testing with dipsticks that detect the presence of leucocytes and nitrites is commonly used in primary care to predict the subsequent diagnosis of urinary tract infection as determined by standard midstream urine culture and to guide the use of antibiotics. In one general practice study, the presence of leucocytes or nitrites in turbid urine had a positive predictive value of finding a pure growth on subsequent culture of around 66%.3 Conversely, a negative dipstick test for both leucocytes and nitrites has a negative predictive value of finding a pure growth on subsequent culture of 80-98.5%.2 3 In 2000 we carried out an epidemiological study to determine the prevalence of antibiotic resistance in bacteria causing uncomplicated urinary tract infections in the community and confirm the negative predictive value of the dipstick.5
Of 374 specimens collected, 96 (26%) were negative for both leucocytes and nitrites.5 Eight of these (8%) contained pure growth cultures above the standard conservative cut-off point of 100x106 colony forming units per litre.6 The negative predictive value of the dipstick test was 92%.
The approach to women with symptoms of uncomplicated urinary tract infection and positive urine dipstick results is to give empirical antibiotic treatment.7-9 Recommendations for the treatment of women with symptoms and negative dipstick results vary. Some suggest empirical treatment,1 10 but others do not.7 8 11 We carried out a pragmatic trial of antibiotic compared with placebo in women with symptoms of uncomplicated urinary tract infection and negative dipstick results.
Methods
We designed the study as a double blind randomised placebo controlled trial. The intervention was treatment with trimethoprim 300 mg daily for three days.
Participants
We invited women to participate who were aged between 16 and 50 and presenting with a history of dysuria and frequency to general practitioners from the randomly selected Christchurch sentinel network.5 Seventy five (91%) of the original 82 randomly selected general practitioners in the network were participating in a surveillance study, and 30 (40%) of these agreed to recruit patients for the randomised controlled trial. Exclusion criteria were a dipstick test that was positive for leucocytes or nitrites, complicated urinary tract infections, pregnancy, or known allergy to trimethoprim. We also excluded women with proved urinary tract infection or treatment for presumed urinary tract infection in the past month.
All the women who agreed to participate provided a midstream urine specimen that their general practitioner tested immediately with a standard urine dipstick and then sent for microbiological examination and culture. We randomly allocated patients whose dipstick test was negative for both leucocytes and nitrites to receive either three days of trimethoprim 300 mg (standard treatment) or placebo. To randomise the participants, the biostatistician chose SAS code, which used random block lengths with a maximum of 10 to generate a random sequence for the medication packs. This went to a dispensing pharmacist, together with a sequential alphanumerical code list. The pharmacist prepared identical placebo and active drug capsules and attached an alphanumerical code to the medication bottles that the patients received from their general practitioner. The pharmacist put the codes with the list identifying allocation to placebo or treatment in a sealed envelope. The study team and general practitioners enrolling patients had no access to the code list until the data collection was complete. Labels preprinted with the code were provided to attach to the patient's clinical notes as well as the initial urinary specimen form and questionnaire. General practitioners, investigators, and research nurses were all blind to allocation until the data collection was complete.
Data collection
Two accredited pathology laboratories (International Accreditation New Zealand) used standard techniques to test the urine specimens. A 0.001 ml loop was used to plate the specimens on to MacConkey agar and sheep blood agar. Plates were incubated at 37°C for 48 hours and colony counts recorded. We defined pyuria as 20 leucocytes or more per ml of urine on microscopy and clinically relevant bacteriuria as 105 organisms/ml of urine.
Participants completed a short written questionnaire at the surgery. This asked about demographic details, current symptoms, including, on the day the participant entered the trial, the presence or absence of dysuria, increased urinary frequency, low back pain, abdominal pain, appearance of blood in the urine, itching, and feeling "hot or shivery." Women were also asked about potential risk factors for infection, including use and type of contraceptive, recent sexual activity, past history of infection, and use of other preparations to alleviate symptoms. Participants received a seven day diary to record their symptoms and return by mail. They recorded the presence or absence of individual symptoms each day. We contacted all women after seven days and administered a structured telephone questionnaire, recording the presence or absence of each symptom. Where a symptom was absent and had been recorded as present on the day 1 questionnaire, we asked the woman to recall the day on which the symptom resolved. Adverse effects potentially related to medication and any other treatments taken were recorded. We asked participants to provide a second midstream urine specimen for microbiological examination and culture.
The main outcome was resolution of dysuria in the intervention and control groups at three and seven days, and median time to resolution. Secondary outcomes were resolution of other symptoms. We investigated predictors of response to treatment.
Statistical methods
The study was powered to detect clinically important differences between the two groups. With 30 participants in each arm, using 2 = 0.05, the study has 78% power to show a difference between groups if the "true" rates are such that symptoms will not resolve in only 10% of women taking antibiotics compared with 40% of women taking placebo. This sample size has 99% power to detect a difference between groups if the true rates are such that symptoms will not resolve in only 25% of women taking antibiotics compared with 75% of women taking placebo.
We used the median test to analyse days to resolution by symptom, and, where appropriate, we also calculated the number needed to treat. We included women whose symptoms had not resolved in this analysis, with censoring after day 7. We used the 2 test for contingency tables to analyse proportions. We used SAS, version 8.02 (SAS Institute, Cary, North Carolina, USA), to carry out all our analyses.
Results
The trial ran from November 2001 until November 2003. Sixty six patients who fitted the case definition for uncomplicated urinary tract infection with a negative dipstick result consented to participate in the study and were randomised. We decided that collecting accurate screening log data from a large group of general practitioners about every woman presenting with symptoms of urinary tract infection over an extended period was impractical. Using data from our 2000 study among the same group of general practitioners,5 we estimated that the number of dipstick negative patients in the eligible age range seen over a two year period by these 30 general practitioners would be 300 (five dipstick negative patients per year per general practitioner). The remainder were not recruited because the general practitioner did not remember to approach them about the study or did not have time to complete the informed consent process, or the patient did not consent to involvement. This gives an estimated response rate of 20% from the potentially eligible pool. Seven patients were excluded after randomisation (three in one arm and four in the other) and did not complete the study as they were outside the inclusion criteria for age. This left 26 patients in the treatment arm and 33 in the placebo arm (numbers are unbalanced as each general practice had some unused packs). Figure 1 shows the flow of participants through the trial. For the 12 participants who did not complete the symptom diary, we used the data from the telephone call on day 7. In this way we ensured that for all participants, complete data were available on the day when their symptom resolved.
Fig 1 Flow of participants through the trial
The treatment and placebo groups were similar in demographic characteristics (table 1). Among all women who reported dysuria on day 1, the median time for resolution of dysuria was three days in the trimethoprim group and five days in the placebo group (P = 0.002; table 2). At day 3, 24% of patients in the treatment group who had dysuria on day 1, still had dysuria compared with 74% of the placebo group (P = 0.005; table 3). This difference was still present at day 7; only 10% of the treatment group who reported dysuria on day 1 were experiencing dysuria, compared with 41% of the placebo group (P = 0.02). Less than a third of patients in each group experienced constitutional symptoms of feverishness or shivers. However, among those who had symptoms, the median time from clinical presentation and treatment initiation to resolution of feverishness or shivers was two days in the trimethoprim arm, compared with six days in the placebo arm (P = 0.02). At day 3, 46% of patients in the placebo group were still feeling hot and shivery compared with none of the treatment group (P = 0.04). The difference between groups did not reach significance at day 7 (0% v 36%, P = 0.09). The median time to resolution or proportion experiencing symptoms at day 3 and 7 for urinary frequency, abdominal pain, itch, or low back pain did not differ.
Table 1 Patients' characteristics at baseline. Values are numbers (percentages) of patients unless otherwise indicated
Table 2 Days to resolution by symptom
Table 3 Patients with symptom at baseline still experiencing symptom after day 3 and day 7
No patient or illness characteristic predicted response to treatment at day 3. Twenty six participants had 20 leucocytes per ml in their urine on microscopy, 13 in each arm, and this did not predict response to treatment. In the placebo arm, 6/13 (46%) of those who reported no dysuria at day 3 and 7/20 (35%) who still reported dysuria had pyuria at trial entry. In the treatment arm, 8/19 (42%) of those who reported no dysuria at day three and 5/7 (71%) who still reported dysuria had pyuria at trial entry. Five women had microbiological evidence of bacterial infection on midstream urine testing when we used standard criteria of 20 leucocytes/ml of urine and pure growth of 105 organisms/ml of a uropathogen.6 Four of these grew Escherichia coli and one Klebsiella pneumoniae. Three were in the treatment arm and two in the placebo arm. The negative predictive value of the dipstick in this study was therefore 92%. Six women had low count bacteriuria, three in each arm, and this did not predict response to treatment.
We obtained follow-up urine specimens from 42 participants (71%). Two participants had clinically significant bacteriuria (E coli in both cases), both of whom were in the placebo arm, and 13 had 20 leucocytes/ml in their urine on microscopy.
We saw few adverse effects in either arm. Six patients (18%) in the placebo arm and three (12%) in the treatment arm reported minor symptoms (nausea, sore mouth, itching skin, sedation after taking medication, mouth ulceration, and thrush). Nine patients in each group (treatment group 29%, placebo group 35%) had used other preparations. These included cranberry juice, urinary alkalinisers, herbal preparations, and homeopathic remedies.
Discussion
Brooks D. The management of suspected urinary tract infection in general practice. Br J Gen Pract 1990;40: 399-402.
Verest LF, Van Esch WM, Van Ree JW, Stobberingh EE. Management of acute uncomplicated urinary tract infections in general practice in the south of the Netherlands. BrJ Gen Pract 2000;50: 309-10.
Ditchburn RK, Ditchburn JS. A study of microscopical and chemical tests for the rapid diagnosis of urinary tract infections in general practice. Br J Gen Pract 1990;40: 406-8.
Stamm WE, Wagner KF, Amsel R, Alexander R, Turck M, Counts GW, et al. Causes of the acute urethral syndrome in women. N Engl J Med 1980;303: 409-15.
Richards DA, Toop LJ, Chambers ST, Sutherland MG, Harris BH, Ikram RB, et al. Antibiotic resistance in uncomplicated urinary tract infection: problems with interpreting cumulative resistance rates from local community laboratories. N Z Med J 2002;115: 12-4.
Kunin CM. Bacteriuria, pyuria, proteinuria, hematuria and pneumaturia. In: Urinary tract infections. Detection, prevention and management. 5th ed. Baltimore: Williams and Wilkins, 1997: 28-9.
Infection POEMs. Patient orientated evidence that matters. National antibiotic campaign: information for primary care prescribers. Dunedin: Best Practice Advocacy Centre, 2003: 1-20.
Hummers-Pradier E, Kochen M. Urinary tract infections in adult general practice patients. Br J Gen Pract 2002;52: 752-61.
Office on Women's Health US Department of Health and Human Services. Managing acute uncomplicated cystitis in women in the era of antibiotic resistance. Clin Courier 2003;21: 11.
Fenwick EAL, Briggs AH, Hawke CI. Management of urinary tract infection in general practice: a cost-effectiveness analysis. Br J Gen Pract 2000;50: 635-9.
Car J, Sheikh A. Recurrent urinary tract infection in women. BMJ 2003;327: 1204.
Foxman B, Barlow R, D'Arcy H, Gillespie B, Sobel JD. Urinary tract infection: self-reported incidence and associated costs. Ann Epidemiol 2000;10: 509-15.
Stamm WE, Running K, McKevitt M, Counts GW, Turck M, Holmes KK. Treatment of the acute urethral syndrome. N Eng J Med 1981;304: 956-8.
Stamm WE, Counts GW, Running KR, Fihn S, Turck M, Holmes KK. Diagnosis of coliform infection in acutely dysuric women. N Eng J Med 1982;307: 463-8.
Kunin CM, VanArsdale White L, Hua Hua T. A reassessment of the importance of "low-count" bacteriuria in young women with acute urinary symptoms. Ann Intern Med 1993;119: 454-60.
Moore T, Hira NR, Stirland RM. Differential urethrovesical urinary cell-count. Lancet 1965;1: 626-7.
Department of Health. UK Antimicrobial Resistance Strategy and Action Plan. London: Department of Health, 2000:1-25.(Dee Richards, senior lecturer1, Les Toop)
Correspondence to: D Richards derelie.richards@chmeds.ac.nz
Objective To assess the effectiveness of antibiotic treatment of women with symptoms of urinary tract infection but negative urine dipstick testing.
Design Prospective, double blind, randomised, placebo controlled trial.
Setting Primary care, among a randomly selected group of general practitioners in Christchurch, New Zealand.
Participants 59 women aged 16-50 years presenting with a history of dysuria and frequency in whom a dipstick test of midstream urine was negative for both nitrites and leucocytes. Participants with complicated urinary tract infection were excluded.
Intervention Trimethoprim 300 mg daily for three days or placebo.
Main outcome measures Self reported diary of symptoms for seven days, recording the presence or absence of individual symptoms each day, followed by a structured telephone questionnaire after seven days. The main clinical outcome was resolution of dysuria at three and seven days and median time to resolution. Secondary outcomes were resolution of other symptoms.
Results The median time for resolution of dysuria was three days for trimethoprim compared with five days for placebo (P = 0.002). At day 3, five (24%) of patients in the treatment group had ongoing dysuria compared with 20 (74%) in the placebo group (P = 0.005). This difference persisted until day 7: two patients (10%) in the treatment group v 11 (41%) in the placebo group; P = 0.02). The number needed to treat was 4. The median duration of constitutional symptoms (feverishness, shivers) was reduced by four days.
Conclusions Although a negative dipstick test for leucocytes and nitrites accurately predicted absence of infection when standard microbiological definitions were used (negative predictive value 92%), it did not predict response to antibiotic treatment. Three days' treatment with trimethoprim significantly reduced dysuria in women whose urine dipstick test was negative. These results support the practice of empirical antibiotic use guided by symptoms. Balancing the competing interests of symptom relief and the minimisation of antibiotic use remains a dilemma—further research is needed to determine clinical predictors of response to antibiotics.
Infections of the urinary tract are extremely common, and the numbers of patients presenting to general practice with this condition represent considerable morbidity and workload.1 Because the symptoms of dysuria and frequency are unpleasant, doctors are under pressure from patients to provide relief.2 3 However, not all such symptoms are associated with growth of bacteria in a standard midstream urine specimen. More detailed microbiological investigation of women with apparently negative urine cultures on standard testing indicate that a proportion of them have low count bacteriuria.4 Urine testing with dipsticks that detect the presence of leucocytes and nitrites is commonly used in primary care to predict the subsequent diagnosis of urinary tract infection as determined by standard midstream urine culture and to guide the use of antibiotics. In one general practice study, the presence of leucocytes or nitrites in turbid urine had a positive predictive value of finding a pure growth on subsequent culture of around 66%.3 Conversely, a negative dipstick test for both leucocytes and nitrites has a negative predictive value of finding a pure growth on subsequent culture of 80-98.5%.2 3 In 2000 we carried out an epidemiological study to determine the prevalence of antibiotic resistance in bacteria causing uncomplicated urinary tract infections in the community and confirm the negative predictive value of the dipstick.5
Of 374 specimens collected, 96 (26%) were negative for both leucocytes and nitrites.5 Eight of these (8%) contained pure growth cultures above the standard conservative cut-off point of 100x106 colony forming units per litre.6 The negative predictive value of the dipstick test was 92%.
The approach to women with symptoms of uncomplicated urinary tract infection and positive urine dipstick results is to give empirical antibiotic treatment.7-9 Recommendations for the treatment of women with symptoms and negative dipstick results vary. Some suggest empirical treatment,1 10 but others do not.7 8 11 We carried out a pragmatic trial of antibiotic compared with placebo in women with symptoms of uncomplicated urinary tract infection and negative dipstick results.
Methods
We designed the study as a double blind randomised placebo controlled trial. The intervention was treatment with trimethoprim 300 mg daily for three days.
Participants
We invited women to participate who were aged between 16 and 50 and presenting with a history of dysuria and frequency to general practitioners from the randomly selected Christchurch sentinel network.5 Seventy five (91%) of the original 82 randomly selected general practitioners in the network were participating in a surveillance study, and 30 (40%) of these agreed to recruit patients for the randomised controlled trial. Exclusion criteria were a dipstick test that was positive for leucocytes or nitrites, complicated urinary tract infections, pregnancy, or known allergy to trimethoprim. We also excluded women with proved urinary tract infection or treatment for presumed urinary tract infection in the past month.
All the women who agreed to participate provided a midstream urine specimen that their general practitioner tested immediately with a standard urine dipstick and then sent for microbiological examination and culture. We randomly allocated patients whose dipstick test was negative for both leucocytes and nitrites to receive either three days of trimethoprim 300 mg (standard treatment) or placebo. To randomise the participants, the biostatistician chose SAS code, which used random block lengths with a maximum of 10 to generate a random sequence for the medication packs. This went to a dispensing pharmacist, together with a sequential alphanumerical code list. The pharmacist prepared identical placebo and active drug capsules and attached an alphanumerical code to the medication bottles that the patients received from their general practitioner. The pharmacist put the codes with the list identifying allocation to placebo or treatment in a sealed envelope. The study team and general practitioners enrolling patients had no access to the code list until the data collection was complete. Labels preprinted with the code were provided to attach to the patient's clinical notes as well as the initial urinary specimen form and questionnaire. General practitioners, investigators, and research nurses were all blind to allocation until the data collection was complete.
Data collection
Two accredited pathology laboratories (International Accreditation New Zealand) used standard techniques to test the urine specimens. A 0.001 ml loop was used to plate the specimens on to MacConkey agar and sheep blood agar. Plates were incubated at 37°C for 48 hours and colony counts recorded. We defined pyuria as 20 leucocytes or more per ml of urine on microscopy and clinically relevant bacteriuria as 105 organisms/ml of urine.
Participants completed a short written questionnaire at the surgery. This asked about demographic details, current symptoms, including, on the day the participant entered the trial, the presence or absence of dysuria, increased urinary frequency, low back pain, abdominal pain, appearance of blood in the urine, itching, and feeling "hot or shivery." Women were also asked about potential risk factors for infection, including use and type of contraceptive, recent sexual activity, past history of infection, and use of other preparations to alleviate symptoms. Participants received a seven day diary to record their symptoms and return by mail. They recorded the presence or absence of individual symptoms each day. We contacted all women after seven days and administered a structured telephone questionnaire, recording the presence or absence of each symptom. Where a symptom was absent and had been recorded as present on the day 1 questionnaire, we asked the woman to recall the day on which the symptom resolved. Adverse effects potentially related to medication and any other treatments taken were recorded. We asked participants to provide a second midstream urine specimen for microbiological examination and culture.
The main outcome was resolution of dysuria in the intervention and control groups at three and seven days, and median time to resolution. Secondary outcomes were resolution of other symptoms. We investigated predictors of response to treatment.
Statistical methods
The study was powered to detect clinically important differences between the two groups. With 30 participants in each arm, using 2 = 0.05, the study has 78% power to show a difference between groups if the "true" rates are such that symptoms will not resolve in only 10% of women taking antibiotics compared with 40% of women taking placebo. This sample size has 99% power to detect a difference between groups if the true rates are such that symptoms will not resolve in only 25% of women taking antibiotics compared with 75% of women taking placebo.
We used the median test to analyse days to resolution by symptom, and, where appropriate, we also calculated the number needed to treat. We included women whose symptoms had not resolved in this analysis, with censoring after day 7. We used the 2 test for contingency tables to analyse proportions. We used SAS, version 8.02 (SAS Institute, Cary, North Carolina, USA), to carry out all our analyses.
Results
The trial ran from November 2001 until November 2003. Sixty six patients who fitted the case definition for uncomplicated urinary tract infection with a negative dipstick result consented to participate in the study and were randomised. We decided that collecting accurate screening log data from a large group of general practitioners about every woman presenting with symptoms of urinary tract infection over an extended period was impractical. Using data from our 2000 study among the same group of general practitioners,5 we estimated that the number of dipstick negative patients in the eligible age range seen over a two year period by these 30 general practitioners would be 300 (five dipstick negative patients per year per general practitioner). The remainder were not recruited because the general practitioner did not remember to approach them about the study or did not have time to complete the informed consent process, or the patient did not consent to involvement. This gives an estimated response rate of 20% from the potentially eligible pool. Seven patients were excluded after randomisation (three in one arm and four in the other) and did not complete the study as they were outside the inclusion criteria for age. This left 26 patients in the treatment arm and 33 in the placebo arm (numbers are unbalanced as each general practice had some unused packs). Figure 1 shows the flow of participants through the trial. For the 12 participants who did not complete the symptom diary, we used the data from the telephone call on day 7. In this way we ensured that for all participants, complete data were available on the day when their symptom resolved.
Fig 1 Flow of participants through the trial
The treatment and placebo groups were similar in demographic characteristics (table 1). Among all women who reported dysuria on day 1, the median time for resolution of dysuria was three days in the trimethoprim group and five days in the placebo group (P = 0.002; table 2). At day 3, 24% of patients in the treatment group who had dysuria on day 1, still had dysuria compared with 74% of the placebo group (P = 0.005; table 3). This difference was still present at day 7; only 10% of the treatment group who reported dysuria on day 1 were experiencing dysuria, compared with 41% of the placebo group (P = 0.02). Less than a third of patients in each group experienced constitutional symptoms of feverishness or shivers. However, among those who had symptoms, the median time from clinical presentation and treatment initiation to resolution of feverishness or shivers was two days in the trimethoprim arm, compared with six days in the placebo arm (P = 0.02). At day 3, 46% of patients in the placebo group were still feeling hot and shivery compared with none of the treatment group (P = 0.04). The difference between groups did not reach significance at day 7 (0% v 36%, P = 0.09). The median time to resolution or proportion experiencing symptoms at day 3 and 7 for urinary frequency, abdominal pain, itch, or low back pain did not differ.
Table 1 Patients' characteristics at baseline. Values are numbers (percentages) of patients unless otherwise indicated
Table 2 Days to resolution by symptom
Table 3 Patients with symptom at baseline still experiencing symptom after day 3 and day 7
No patient or illness characteristic predicted response to treatment at day 3. Twenty six participants had 20 leucocytes per ml in their urine on microscopy, 13 in each arm, and this did not predict response to treatment. In the placebo arm, 6/13 (46%) of those who reported no dysuria at day 3 and 7/20 (35%) who still reported dysuria had pyuria at trial entry. In the treatment arm, 8/19 (42%) of those who reported no dysuria at day three and 5/7 (71%) who still reported dysuria had pyuria at trial entry. Five women had microbiological evidence of bacterial infection on midstream urine testing when we used standard criteria of 20 leucocytes/ml of urine and pure growth of 105 organisms/ml of a uropathogen.6 Four of these grew Escherichia coli and one Klebsiella pneumoniae. Three were in the treatment arm and two in the placebo arm. The negative predictive value of the dipstick in this study was therefore 92%. Six women had low count bacteriuria, three in each arm, and this did not predict response to treatment.
We obtained follow-up urine specimens from 42 participants (71%). Two participants had clinically significant bacteriuria (E coli in both cases), both of whom were in the placebo arm, and 13 had 20 leucocytes/ml in their urine on microscopy.
We saw few adverse effects in either arm. Six patients (18%) in the placebo arm and three (12%) in the treatment arm reported minor symptoms (nausea, sore mouth, itching skin, sedation after taking medication, mouth ulceration, and thrush). Nine patients in each group (treatment group 29%, placebo group 35%) had used other preparations. These included cranberry juice, urinary alkalinisers, herbal preparations, and homeopathic remedies.
Discussion
Brooks D. The management of suspected urinary tract infection in general practice. Br J Gen Pract 1990;40: 399-402.
Verest LF, Van Esch WM, Van Ree JW, Stobberingh EE. Management of acute uncomplicated urinary tract infections in general practice in the south of the Netherlands. BrJ Gen Pract 2000;50: 309-10.
Ditchburn RK, Ditchburn JS. A study of microscopical and chemical tests for the rapid diagnosis of urinary tract infections in general practice. Br J Gen Pract 1990;40: 406-8.
Stamm WE, Wagner KF, Amsel R, Alexander R, Turck M, Counts GW, et al. Causes of the acute urethral syndrome in women. N Engl J Med 1980;303: 409-15.
Richards DA, Toop LJ, Chambers ST, Sutherland MG, Harris BH, Ikram RB, et al. Antibiotic resistance in uncomplicated urinary tract infection: problems with interpreting cumulative resistance rates from local community laboratories. N Z Med J 2002;115: 12-4.
Kunin CM. Bacteriuria, pyuria, proteinuria, hematuria and pneumaturia. In: Urinary tract infections. Detection, prevention and management. 5th ed. Baltimore: Williams and Wilkins, 1997: 28-9.
Infection POEMs. Patient orientated evidence that matters. National antibiotic campaign: information for primary care prescribers. Dunedin: Best Practice Advocacy Centre, 2003: 1-20.
Hummers-Pradier E, Kochen M. Urinary tract infections in adult general practice patients. Br J Gen Pract 2002;52: 752-61.
Office on Women's Health US Department of Health and Human Services. Managing acute uncomplicated cystitis in women in the era of antibiotic resistance. Clin Courier 2003;21: 11.
Fenwick EAL, Briggs AH, Hawke CI. Management of urinary tract infection in general practice: a cost-effectiveness analysis. Br J Gen Pract 2000;50: 635-9.
Car J, Sheikh A. Recurrent urinary tract infection in women. BMJ 2003;327: 1204.
Foxman B, Barlow R, D'Arcy H, Gillespie B, Sobel JD. Urinary tract infection: self-reported incidence and associated costs. Ann Epidemiol 2000;10: 509-15.
Stamm WE, Running K, McKevitt M, Counts GW, Turck M, Holmes KK. Treatment of the acute urethral syndrome. N Eng J Med 1981;304: 956-8.
Stamm WE, Counts GW, Running KR, Fihn S, Turck M, Holmes KK. Diagnosis of coliform infection in acutely dysuric women. N Eng J Med 1982;307: 463-8.
Kunin CM, VanArsdale White L, Hua Hua T. A reassessment of the importance of "low-count" bacteriuria in young women with acute urinary symptoms. Ann Intern Med 1993;119: 454-60.
Moore T, Hira NR, Stirland RM. Differential urethrovesical urinary cell-count. Lancet 1965;1: 626-7.
Department of Health. UK Antimicrobial Resistance Strategy and Action Plan. London: Department of Health, 2000:1-25.(Dee Richards, senior lecturer1, Les Toop)