Validity of urine dipstick test to assess eradication of urinary tract infection in persons with spinal cord injury

25 juin 2017

Auteurs : J.G. Previnaire, J.M. Soler, L. Chouaki, L. Pawlicki, M. Le Berre, E. Hode, P. Denys
Référence : Prog Urol, 2017, 7, 27, 424-430




 




Introduction


Bacteriuria, either symptomatic or not, is the most common medical complication in patients with spinal cord injury (SCI), both in the acute phase and throughout their lifetime [1, 2, 3]. Prevalence of lower urinary tract colonization (asymptomatic bacteriuria) has been reported between 50% and 90%, both at home and in hospital [4, 5, 6], while urinary tract infection (UTI - symptomatic bacteriuria) and urinary complications are the first cause of rehospitalisation in tetraplegic patients [7].


The diagnosis of symptomatic UTI is based on a combination of clinical signs and bacteriological criteria determined from urine cultures [8, 9, 10]. Symptoms include urinary, bladder or general signs [9, 11], while significant bacteriuria is dependent on the mode of voiding. However, cultures are costly, time consuming and require at least 24hours [12]. Over the last thirty years, the urinary dipstick test has become increasingly used as an alternative, as it is cheap and can be used as a bedside test, enabling a quick diagnosis. Studies have shown its utility in variety non-neurological disorders, particularly to rule out infections [12].


In patients with SCI, dipstick results for nitrite (NIT) and leukocyte esterase (LE) are part of the essential components of the UTI basic data set [11], and are widely used in rehabilitation settings [13]. However, there is a clear lack of scientific evidence to support their utility and there is still some confusion as to the validity of dipstick testing to confirm infection. In a recent review of the literature, Cameron et al. concluded that NIT and LE dipsticks are sensitive and specific for predicting a UTI [4]. The results are particularly robust when a positive NIT or LE is considered [6, 14]. The rate of missed infections and undertreatment is low, making it an excellent screening test [4]. However, the earlier study by Hoffman et al. found poor sensitivity (64%) and specificity (52%), suggesting that reliance on dipstick testing alone could lead to a high rate of overtreatment for UTI [6].


New regulations in France state that the effectiveness of antibiotic regimens, whether prescribed following culture results or clinical signs, should be assessed in the first days of the treatment, in order to control antimicrobial overuse and resistance [15].


The aim of this prospective study was to assess the clinical value of dipstick testing as an alternative to urine culture, to assess the effectiveness of antibiotic treatment, as well as its utility during routine testing and suspicion of UTI, in persons with SCI on IC or with an indwelling catheter.


Material and methods


Patients


From January to September 2013, urinalysis, urine culture and automated dipstick analysis (Clinitekâ„¢) were prospectively performed on all urine samples taken from all inpatients with SCI. Patients were included if they were on only IC, IC combined with spontaneous voiding, or indwelling catheters. Patients with a recent history of antibiotic treatment (within 2 weeks) or presenting with a suspicion of febrile UTI (temperature above 38°5C) were excluded.


The urine specimens were divided into three conditions for analysis:

routine: on admission, prior to hydrotherapy or urodynamic tests;
suspicion of a non-febrile UTI (temperature less than 38°5C);
on day 4 (ATB+3) of a 5-day course of antibiotics for significant bacteriuria with or without symptoms.


Early morning urine samples were collected whenever possible. Urine for culture was always collected as a clean-catch midstream technique for patients on IC, and from a new urine catheter whenever possible for those with indwelling catheters.


At the time the sample was collected, participants completed a specifically designed questionnaire with the nurse, in which they reported any clinical signs or symptoms. This included urinary signs such as cloudy or malodorous urine, bladder signs such as urinary incontinence, urgency or modification of bladder behavior (decreased bladder capacity), and general signs such as fever without other causes, increased spasticity, bouts of autonomic dysreflexia and sense of unease (suprapubic pain).


Urine samples were collected using standard procedures (clean or aseptic IC, when performed by the patients or the nurses, respectively) or aseptically through the catheter port and were immediately sent to the microbiology laboratory. Several samples could be included for the same patient.


Ethical approval was not required for this study since it involved routine treatment with no blood samples.


Laboratory methods


The samples were processed within 3hours of collection.


Dipstick analyses were performed first, with the remainder of the sample used for urine culture and urinalysis.


Clinitekâ„¢ was used for dipstick chemical urine analysis.


Colony-forming units (cfu) were counted by the surface streak method using CPS agar (BioMérieux, Marcy l'Étoile, France). The minimal detection limit was 20 cfu/mL. Other culture plates were used as appropriate according to the results of Gram stain smear examination. Bacterial species were identified according to standard methods.


Definitions


Dipstick tests were defined as positive if either NIT or LE was positive.


For urine cultures, the criteria of the American Paraplegia Society were followed:

significant bacteriuria: a threshold of 102cfu/mL for patients on IC and any detectable concentration for specimens from indwelling catheters;
UTI: significant bacteriuria and at least one clinical sign or symptom;
asymptomatic bacteriuria: significant bacteriuria without any clinical signs.


To evaluate the effectiveness of the antibiotic treatment, the following definitions were used:

microbiological cure: urine cultures with negative results at ATB+3;
clinical cure: the absence of urinary symptoms.


Clinical signs were considered positive if there was at least one sign or symptom within one of the 3 subgroups (urinary, bladder or general signs).


The patients were categorized into two groups, according to their method of bladder drainage (IC and Foley indwelling catheter [FC]).


Statistical analysis


The results of the urine cultures were compared with the dipstick results for NIT and LE in each group of patients (IC and FC).


Validity of the screening test: sensitivity, specificity and predictive values (PV) were calculated for positive NIT tests, positive LE tests, both positive and negative NIT and LE tests, and either positive NIT or LE tests compared with urine culture results in each clinical condition (routine, suspicion of UTI and ATB+3).


Results


A total of 157 urine samples from 61 patients were included. Twenty patients had paraplegia, 41 tetraplegia, 14 were female and 47 were male. Their characteristics are shown in Table 1. Thirty-four patients were on IC (95 samples) and 27 on FC (62 samples). Seventy samples were routine, 51 for suspicion of UTI, and 36 on day 4 of antibiotic treatment for UTI (ATB+3). One to 12 samples were collected per patient.


Urine culture


In the routine condition (patients without symptoms), significant bacteriuria was found in 35 out of 36 samples in the IC group, and 29 out of 34 samples in the FC group. The prevalence of asymptomatic bacteriuria was 89% for the 56 patients. The prevalence was higher in the IC group (97%) than in the FC group (81%). In the suspicion of UTI condition, all samples had significant bacteriuria.


Significant bacteriuria was treated with antibiotics in 36 cases: 22 following suspicion of UTI and 14 before urodynamic testing or hydrotherapy. Antibiotic therapy followed antimicrobial susceptibility testing in 27 cases and was empirically started in 9.


Antibiotics were prescribed as single treatments in 33 cases (cefixime [12], Nor- or ciprofoxacine [1], trimethoprim/sulfamethoxazole [7], nitrofurantoin [3] and as dual therapies in 3 cases: amikacin+cefixime [1], gentamycin+ciprofloxacin [1] or fosfomycin trometamol [1]).


At ATB+3, microbiological cure was found in 27 out of 30 specimens in the IC group and 2 out of 6 specimens in the FC group.


Treatment failure was mainly due to reinfection with a change in the bacterial strain (6 specimens: 2 in the IC group [7%] and 4 in the FC group [67%]), while the same bacteria persisted in one specimen (one patient on IC). The causative bacteria were thus eradicated in 35 out of 36 specimens.


Leukocyturia≥103 was found in all 157 specimens, including the 29 specimens, which had microbiological cure.


Clinical signs


Suspicion of UTI (51 samples) was based on one, or a combination of urinary signs for 36 cases, bladder signs for 17 cases and general signs for 22 cases.


At ATB+3, clinical cure occurred in 20 out of the 22 patients with suspicion of UTI and bladder signs persisted in 2 patients on IC despite microbiological cure.


Dipsticks


The results and comparisons of dipstick and culture tests are presented in Table 2, Table 3, Table 4, Table 5. As can be seen in Table 3, the predictive value of the dipstick-screening test varied greatly depending on whether NIT was considered alone or in combination with LE.


It must be noted that in this study, a positive dipstick LE could not be used as a discriminant factor since the urinalyses showed leukocyturia in all samples. Thus, the discussion of dipstick findings is based on the NIT results. Furthermore, care must be taken in the interpretation of some results due to the small number of negative dipsticks in the routine and suspicion of UTI conditions.


Dipstick testing for NIT compared with significant bacteriuria


In the routine condition (absence of symptoms), NIT tests had excellent specificity and positive PV of 1.00, with sensitivity of 0.63 and 0.66, but low negative PV. The first two results can be attributed to the absence of false positives in both groups (positive NIT with asymptomatic bacteriuria).


In the suspicion of UTI condition (presence of symptoms), the sensitivity of tests for NIT was higher in the IC than in the FC group (0.69 vs. 0.55). The positive PV was 1.00 and negative PV 0.00, for both groups. The latter two results can be attributed to the absence of negative urine cultures in both groups.


At ATB+3, the results of the tests for NIT differed significantly between the 2 groups. In the IC samples, there was a negative PV and sensitivity of 1.00, specificity of 0.85 but low positive PV of 0.43. In the FC samples, there was specificity of 1.00 with low negative PV of 0.33 and sensitivity of 0.00.


Bacteria


A single bacterium was found in 104 specimens and 2 bacteria were found in 17 specimens. Twelve different strains of bacteria were found (Table 6), of which the most common were Escherichia Coli (36%) and Klebsiella Pneumoniae (18%). Pseudomonas Aeruginosa was mostly present in the FC samples (18% in this group).


Discussion


The results of this study showed that dipstick testing for NIT had a high specificity and excellent negative PV to confirm microbiological cure (in patients on IC only), and limited clinical value to rule out significant bacteriuria (routine), or confirm UTI (presence of clinical signs). Urine culture at ATB+3 showed eradication of the causative germs in all but one patient, while reinfection (bacterial change) mostly occurred in the FC group. Urinalaysis showed persistence of leukocyturia in all patients, even in the case of microbiological cure.


The excellent positive PV (1.00) found in this study in the routine and suspicion of UTI conditions is, at first sight, clinically relevant. A reliably positive dipstick would be very useful to rule in symptomatic bacteriuria or UTI, order a urine culture and start empiric antibiotic treatment when necessary. However, asymptomatic bacteriuria is common in patients with SCI, with reported prevalences between 30% and 89% [4, 5]. In the study by Hoffman et al., on dipstick testing in patients with SCI living at home, asymptomatic bacteriuria was found in 81% of the samples [6]. In this study, we found significant bacteriuria in 91% of the samples and estimated the prevalence of asymptomatic bacteriuria at 89%. Thus, the clinical value of this test is clearly limited. Reliance on dipstick results could lead to high rates of overtreatment for UTI. Urine cultures are therefore mandatory to ensure appropriate treatment [6].


In patients with non-neurological disorders, dipstick testing has been shown to be valid, mainly to rule out infections [12], but surprisingly, studies have rarely compared the negative PV of dipstick testing with cultures. Two studies reported a good negative PV: from 0.91 to 0.98 (combined for LE, NIT and protein) [16], and 0.95 (combined for LE and NIT) [14], while another study later found a low negative PV of 0.35 for significant bacteriuria and a negative PV of 0.78 for positive UTI (combined for LE and NIT) [6]. Another study concluded that asymptomatic patients with SCI who present for annual evaluations should not undergo routine urine cultures if they have negative nitrite [17]. In our study, the negative PV of the dipstick test was very poor in the routine and suspicion of UTI conditions, but was excellent for ATB+3 in patients on IC. Moreover, its sensitivity was excellent (1.00). The dipstick test can therefore be safely used to rule out significant bacteriuria and confirm the effectiveness of antibiotic treatment.


Leukocytes


The presence of leukocytes in the urine, which reflects the inflammatory state of the bladder, might help distinguish between true infection and colonization [2]. In patients who do not have SCI, pyuria is an excellent indicator of infection: pyuria is almost always present (>96%) in symptomatic UTI, is often present (∼50%) in asymptomatic bacteriuria, and is rare (<1%) in patients with asymptomatic non-bacteriuria [1, 18].


In patients with SCI, leukocyturia is a frequent finding because the bladder wall is irritated by intermittent catheterization or indwelling catheters, even when there is no infection [8]. We further showed that leukocyturia was present in all samples, in the IC and FC groups alike, including those with microbiological cure at ATB+3. This highlights the low specificity of LE as a diagnostic test for UTI [1] and that NIT should be preferentially considered when carrying out dipstick testing. Quantitative criteria for assessing pyuria in persons with SCI have yet to be determined for urinalysis [2].


Clinical signs


Signs and symptoms are included in the UTI basic data set [11], however clinical manifestations of UTI are often atypical and nonspecific in patients with SCI and may involve only aggravation of preexisting neurological symptoms, while some very suggestive symptoms such as fever may have a non-infectious origin [19].


Urinary signs are usually considered to have the highest accuracy for the diagnosis of UTI [2, 20]. The number of clinical signs is undoubtedly a major diagnostic factor [2]. In the case of microbiological cure at ATB+3, we showed that urinary signs and most other clinical signs always disappeared, except for the bladder signs that persisted in 2 cases, suggesting that these signs are less specific.


Mid-treatment assessment of ATB effectiveness


Reassessment of clinical and microbiological effectiveness of antibiotic regimens should be part of good clinical practice [15]. Confirmation of eradication of germs at mid-treatment is necessary to ensure ATBs are correctly used. There are limited data on the mid-treatment effect of antibiotics in UTI. In a study of catheter-free patients with SCI undergoing 7-14 day courses of antibiotics, Waites et al. found mid-treatment urine sterility rates of 54% [21] while Stannard et al. found that 97% of patients had negative cultures on day 3 of treatment in a study of 38 hospitalized patients with SCI treated with ciprofloxacin for 5 days [22]. In the present study, eradication of the causative germ was achieved in 35/36 (97%) samples and microbiological cure in 29/36 (81%) samples by day 4. This heterogeneity in rates of cure might be related to differences in the effectiveness of the antibiotics used or variations in microbial characteristics and virulence factors of the infecting organisms. Treatment failure was due to reinfection (6/7), mainly in the FC group. In patients on indwelling catheterization the risk of UTI is 4-times greater than in those on other bladder drainage methods (odds ratio 4.04) [23].


The optimal duration of ATB treatment for UTI is still a matter of debate. The prompt clinical and microbiological response to treatment observed in this study (day 4) supports the prescription of short courses of antibiotics (5 to 7 days), which potential benefits could include a reduction in adverse effects, antimicrobial resistance and cost [24].


Conclusion


This study showed that in the specific population of patients with SCI with a high prevalence of bacteriuria, dipstick testing has a high positive PV but does not help the decision making process for UTI and does not avoid urine cultures.


Dipstick testing has an excellent negative PV and is thus useful to confirm a prompt response to antibiotic treatment, especially in patients with slight bacteriuria (patients on IC and most probably those who are catheter-free).


Future research should assess the usefulness of systematic mid-treatment urine tests (either with dipstick or culture) to help discriminate responders from non-responders to antibiotics and to determine optimal treatment duration.


Disclosure of interest


The authors declare that they have no competing interest.



Acknowledgements


We are grateful to Dr Vincke Bernard for statistical assistance, and to Johanna Robertson for revision of the English




Table 1 - Patient characteristics.
  Intermittent catheter  Foley catheter 
Tetraplegic
Paraplegic 
14
20 
27
Female/male  7/27  7/20 
AIS A/B/C/D  22/5/3/4  20/5/0/2 
Time since lesion onset (years)  6.0±9.3
(0.1-42.3) 
3.4±5.8
(0.1-24.3) 
CIC/reflex voiding+CIC
Indwelling/suprapubic catheter 
31/3
−
26/1 
Urine samples (n 95  62 



Légende :
AIS: ASIA Injury Scale; CIC: clean intermittent catheterization.



Table 2 - Results of urine cultures and dipstick testing in each condition.
  Intermittent catheter 
Foley catheter 
  Urine culture 
Dipstick NIT 
Urine culture 
Dipstick NIT 
 
Routine  35  22  14  29  19  15 
Suspicion UTI  29  20  22  12  10 
ATB+ 27  23 
Total  67  28  49  46  55  31  31 



Légende :
NIT: nitrite; UTI: urinary tract infection; ATB+3: fourth day of antibiotic treatment.



Table 3 - Predictive values of urine dipstick analysis against urine culture in the routine condition.
    Sensitivity  Specificity  Positive PV  Negative PV 
NIT  IC  0.63 (0.79-0.47)  1.00 (1.00-1.00)  1.00 (1.00-1.00)  0.07 (from 0.21 to −0.06) 
  FC  0.66 (0.83-0.48)  1.00 (1.00-1.00)  1.00 (1.00-1.00)  0.33 (0.57-0.09) 
LE  IC  0.43 (0.59-0.26)  0.00 (0.00-0.00)  0.94 (1.06-0.82)  0.00 (0.00-0.00) 
  FC  0.76 (0.91-0.60)  0.80 (1.15-0.45)  0.96 (1.04-0.87)  0.36 (0.65-0.08) 
NIT or LE positive  IC  0.80 (0.93- 0.67)  0.00 (0.00-0.00)  0.97 (1.03-0.90)  0.00 (0.00-0.00) 
  FC  0.90 (1.01-0.79)  0.80 (1.15- 0.45)  0.96 (1.03-0.89)  0.57 (0.94-0.20) 
NIT and LE positive  IC  0.26 (0.40-0.11)  1.00 (1.00-1.00)  1.00 (1.00-1.00)  0.11 (from 0.11 to −0.03) 
  FC  0.52 (0.70-0.34)  1.00 (1.00-1.00)  1.00 (1.00-1.00)  0.26 (0.46-0.07) 
NIT and LE negative  IC  0.20 (from 0.33 to −0.07)  1.00 (1.00-1.00)  1.00 (1.00-1.00)  0.03 (from 0.10 to −0.03) 
  FC  0.10 (from 0.21 to −0.01)  0.20 (from 0.55 to −0.15)  0.43 (0.80-0.06)  0.04 (from 0.11 to −0.03) 



Légende :
Values are given with confidence intervals in brackets. IC: intermittent catheterization; FC: Foley catheter.



Table 4 - Predictive values of urine dipstick analysis against urine culture in suspicion of UTI.
    Sensitivity  Specificity  Positive PV  Negative PV 
NIT  IC  0.69 (0.86-0.52)  1.00 (1.00-1.00)  0.00 (0.00-0.00) 
  FC  0.55 (0.75-0.34)  1.00 (1.00-1.00)  0.00 (0.00-0.00) 



Légende :
Values are given with confidence intervals in brackets. IC: intermittent catheterization; FC: Foley catheter.



Table 5 - Predictive values of urine dipstick analysis against urine culture for ATB+3.
    Sensitivity  Specificity  Positive PV  Negative PV 
NIT  IC  1.00 (1.00-1.00)  0.85 (0.99-0.72)  0.43 (0.80-0.06)  1.00 (1.00-1.00) 
  FC  0.00 (0.00-0.00)  1.00 (1.00-1.00)  0.33 (from 0.71 to −0.04) 



Légende :
ATB+3: fourth day of antibiotic treatment. Values are given with interval confidences in brackets. IC: intermittent catheterization; FC: Foley catheter.



Table 6 - Strains of bacteria found in the urine samples.
Bacteria  IC  FC  Total 
Escherichia Coli   30  20  50 
Klebsiella Pneumoniae   18  25 
Pseudomonas Aeruginosa   12  13 
Enterobacter Aerogenes   11 
Enterococcus Faecalis   10 
Staphylococcus Aureus  
Proteus Mirabilis  
Enterobacter Cloacae  
Corynebacterium Spp.   
Serratia Marcescens  
Staphylococcus Coagulase Negative    
Acinetobacter Baumanii    
Total  71  67  138 



Légende :
IC: intermittent catheterization; FC: Foley catheter.


References



Cardenas D.D., Hooton T.M. Urinary tract infection in persons with spinal cord injury Arch Phys Med 1995 ;  76 (3) : 272-280 [cross-ref]
Ronco E., Denys P., Bernede-Bauduin C., Laffont I., Martel P., Salomon J., et al. Diagnostic criteria of urinary tract infection in male patients with spinal cord injury Neurorehabil Neural Repair 2011 ;  25 (4) : 351-358 [cross-ref]
Wyndaele J.J. Intermittent catheterization: which is the optimal technique? Spinal cord 2002 ;  40 (9) : 432-437 [cross-ref]
Cameron A.P., Rodriguez G.M., Schomer K.G. Systematic review of urological followup after spinal cord injury J Urol 2012 ;  187 (2) : 391-397 [cross-ref]
Colgan R., Nicolle L.E., McGlone A., Hooton T.M. Asymptomatic bacteriuria in adults Am Fam Physician 2006 ;  74 (6) : 985-990
Hoffman J.M., Wadhwani R., Kelly E., Dixit B., Cardenas D.D. Nitrite and leukocyte dipstick testing for urinary tract infection in individuals with spinal cord injury J Spinal Cord Med 2004 ;  27 (2) : 128-132 [cross-ref]
Klotz R., Joseph P.A., Ravaud J.F., Wiart L., Barat M., Tetrafigap G. The Tetrafigap Survey on the long-term outcome of tetraplegic spinal cord injured persons: part III. Medical complications and associated factors Spinal Cord 2002 ;  40 (9) : 457-467 [cross-ref]
Gribble M.J., Puterman M.L., McCallum N.M. Pyuria: its relationship to bacteriuria in spinal cord injured patients on intermittent catheterization Arch Phys Med Rehabil 1989 ;  70 (5) : 376-379
NIDRR National Institute on Disability and rehabilitation research Consensus Statement (NIDRR). The prevention and management of urinary tract infections among people with spinal cord injuries J Am Par Soc 1992 ;  33 (15) : 194-204
Vigil H.R., Hickling D.R. Urinary tract infection in the neurogenic bladder Transl Androl Urol 2016 ;  5 (1) : 72-87
Goetz L.L., Cardenas D.D., Kennelly M., Bonne Lee B.S., Linsenmeyer T., Moser C., et al. International spinal cord injury urinary tract infection basic data set Spinal cord 2013 ;  51 (9) : 700-704 [cross-ref]
Deville W.L., Yzermans J.C., van Duijn N.P., Bezemer P.D., van der Windt D.A., Bouter L.M. The urine dipstick test useful to rule out infections. A meta-analysis of the accuracy BMC Urol 2004 ;  4 : 4 [cross-ref]
Pannek J. Treatment of urinary tract infection in persons with spinal cord injury: guidelines, evidence and clinical practice. A questionnaire-based survey and review of the literature J Spinal Cord Med 2011 ;  34 (1) : 11-15 [cross-ref]
Tuel S.M., Meythaler J.M., Cross L.L., McLaughlin S. Cost-effective screening by nursing staff for urinary tract infection in the spinal cord injured patient Am J Phys Med Rehabil 1990 ;  69 (3) : 128-131 [cross-ref]
HAS Stratégie d'antibiothérapie et prévention des résistances bactériennes en établissement de santé  :  (2008). 
Gutman S.I., Solomon R.R. The clinical significance of dipstick-negative, culture-positive urines in a veterans population Am J Clin Pathol 1987 ;  88 (2) : 204-209 [cross-ref]
Jayawardena V., Midha M. Significance of bacteriuria in neurogenic bladder J Spinal Cord Med 2004 ;  27 (2) : 102-105 [cross-ref]
Stamm W.E. Measurement of pyuria and its relation to bacteriuria Am J Med 1983 ;  75 (1B) : 53-58 [cross-ref]
Linsenmeyer T.A., Oakley A. Accuracy of individuals with spinal cord injury at predicting urinary tract infections based on their symptoms J Spinal Cord Med 2003 ;  26 (4) : 352-357 [cross-ref]
Massa L.M., Hoffman J.M., Cardenas D.D. Validity, accuracy and predictive value of urinary tract infection signs and symptoms in individuals with spinal cord injury on intermittent catheterization J Spinal Cord Med 2009 ;  32 (5) : 568-573 [cross-ref]
Waites K.B., Canupp K.C., DeVivo M.J. Eradication of urinary tract infection following spinal cord injury Paraplegia 1993 ;  31 (10) : 645-652 [cross-ref]
Stannard A.J., Sharples S.J., Norman P.M., Tillotson G.S. Ciprofloxacin therapy of urinary tract infections in paraplegic and tetraplegic patients: a bacteriological assessment J Antimicrob Chemother 1990 ;  26 (Suppl. F) : 13-18 [cross-ref]
Esclarin De Ruz A., Garcia Leoni E., Herruzo Cabrera R. Epidemiology and risk factors for urinary tract infection in patients with spinal cord injury J Urol 2000 ;  164 (4) : 1285-1289 [cross-ref]
Dow G., Rao P., Harding G., Brunka J., Kennedy J., Alfa M., et al. A prospective, randomized trial of 3 or 14 days of ciprofloxacin treatment for acute urinary tract infection in patients with spinal cord injury Clin Infect Dis 2004 ;  39 (5) : 658-664 [cross-ref]






© 2017 
Elsevier Masson SAS. Tous droits réservés.