Diagnostic Microbiology and Infectious Disease xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Diagnostic Microbiology and Infectious Disease journal homepage: www.elsevier.com/locate/diagmicrobio
High bacterial titers in urine are predictive of abnormal postvoid residual urine in patients with urinary tract infection F. Caron a, K. Alexandre b, M. Pestel-Caron c, P. Chassagne d, P. Grise e, M. Etienne a,⁎ a
Infectious Diseases Department, Rouen University Hospital, GRAM (EA2656), Rouen, France Infectious Diseases Department, Rouen University Hospital, Rouen, France Microbiology Department, Rouen University Hospital, GRAM (EA2656), Rouen, France d Geriatrics Department, Rouen University Hospital, Rouen, France e Urology Department, Rouen University Hospital, Rouen, France b c
a r t i c l e
i n f o
Article history: Received 18 February 2015 Received in revised form 30 April 2015 Accepted 11 May 2015 Available online xxxx Keywords: Post-void residual urine Urine analysis Urinary tract infection
a b s t r a c t Urine bacterial titers (BTs) are influenced by bacterial and host factors. The impact of an abnormal postvoid residual (PVR) on BT in urine was investigated. A total of 103 inpatients with a urine growing Enterobacteriacae (≥102 CFU/mL) and a PVR measure were analyzed, mostly female (62%), elderly (mean age: 72 years), with urinary tract infection (25% of asymptomatic bacteriuria) due to Escherichia coli (85%). Fifty-two subjects (56%) had BT ≥106 CFU/mL; 48 (53%) had a PVR ≤100 mL, while 26 (25%) had a PVR N250 mL. PVR increased with BT, and a significant (P b 0.0001) threshold was reached for 106 CFU/mL: 100 mL mean PVR for patients with BT ≤105 CFU/mL versus 248 mL for patients with BT N105 CFU/mL. High PVR and BT were associated with complicated infections, concomitant bacteremia, and delayed apyrexia. Screening for patients with BT ≥106 CFU/mL is an easy way to identify patients at high risk for acute retention and voiding disorders. © 2015 Elsevier Inc. All rights reserved.
1. Introduction Urine culture is by far the most common microbiological analysis performed in daily practice worldwide, due to both the high incidence of urinary tract infection (UTI) or asymptomatic bacteriuria (AB) that deserve detection and treatment (e.g., during pregnancy or before an urological surgery) and to the high feasibility and acceptability of the method (Foxman, 2002; Manickam et al., 2013). Because of a common contamination of the sample by periurethral bacteria (even for midstream urines), colony count thresholds have been established to help interpret positive culture (Hooton et al., 2013). For decades, numerous studies have investigated the influence of several bacterial or host factors on such thresholds. It is usually admitted that thresholds are lower for bacterial species such as Escherichia coli, Proteus spp., and Staphylococcus saprophyticus that produce virulence factors favoring invasion of the urothelium (i.e., the so-called uropathogens), as compared to organisms less frequently responsible for UTI such as enterococci or Pseudomonas spp. (Aspevall et al., 2001). It is also admitted that thresholds for UTI are lower than for AB (Nicolle et al., 2005). Curiously, the influence of voiding dysfonction on bacterial titers (BTs) in urines of patients with UTI or AB has been theorized but has not been investigated in clinical studies (O'Grady and Cattell, 1966a,b). Conceptually, high postvoid residual (PVR) urine is at risk to ⁎ Corresponding author. Tel.: +33-2-32-88-87-39, +33-6-24-94-87-56 (cell); fax: +33-2-32-88-86-42. E-mail addresses: [email protected], [email protected] (M. Etienne).
favor microbial proliferation, while complete voiding should clear most bacteria from bladder (Sobel, 1987). Thus, a positive correlation between BT and PVR values could be expected, but this remains to be demonstrated. For this purpose, the primary aim of this study was to evaluate the correlation between BT and PVR among patients with positive urine culture. The secondary aim was to analyze the role of other factors that might interfere with both parameters (BT and PVR): gender, AB or UTI, risk factors for complication, and concomitant bacteremia. 2. Materials and methods 2.1. Patients This study was conducted retrospectively at the Rouen University Hospital (Normandy, France) from January 2013 to May 2014. Patients with an urine sample growing Enterobacteriacae (whatever the BT, and the presence or absence of leucocyturia) were first identified by the department of microbiology. Patients were then included if they had at least 1 PVR measure within 48 h after the urine sample was collected. The study was limited to the departments of infectious diseases, geriatrics, and urology where a PVR measure by portable ultrasound device is performed when urine retention is suspected. A diagnosis of UTI was retained for patients with clinical symptoms, BT ≥102 CFU/mL, and leukocyte count N10 4/mL. Patients were considered at risk for complicated UTI in case of functional or anatomical abnormality of the urinary tract, male gender, diabetes mellitus, or if they had healthcare-related infections. For patients with febrile UTI (T ≥ 38 °C), the delay of apyrexia
http://dx.doi.org/10.1016/j.diagmicrobio.2015.05.003 0732-8893/© 2015 Elsevier Inc. All rights reserved.
Please cite this article as: Caron F, et al, High bacterial titers in urine are predictive of abnormal postvoid residual urine in patients with urinary tract infection, Diagn Microbiol Infect Dis (2015), http://dx.doi.org/10.1016/j.diagmicrobio.2015.05.003
2
F. Caron et al. / Diagnostic Microbiology and Infectious Disease xxx (2015) xxx–xxx
was noticed; it was defined as the interval between the initiation of an effective antibiotic regimen and 2 consecutive temperature measures ≤37.3 °C in the morning and 37.8 °C in the afternoon. A diagnosis of asymptomatic bacteriuria was retained in asymptomatic patients with bacteriuria, whatever the leukocyturia. Healthcare-related bacteriuria were defined as those occurring 48 h after admission or within 1 month after any invasive urological procedure (Hooton et al., 2010). 2.2. BT measure The BT was determined using quantitative urinanalysis, which consisted of inoculating 10-μL urine onto a nonselective chromogenic agar plate (CPS3; bioMérieux, Marcy l'Etoile, France) using a calibrated loop. After 24 h of incubation at 37 °C in aerobic conditions, the number of colonies was counted and expressed as CFU/mL and quantified in a 102 to ≥107 range for AB and UTI, in order to investigate the correlation between BT and PVR. Patients with urine growing organisms other than Enterobacteriacae were excluded because detection of bacteria such as enterococci or streptococci in midstream urines is not necessarily predictive of their presence in bladder, as recently established among premenopausal women with acute uncomplicated cystitis (Hooton et al., 2013).
within 10 minutes of voiding. The PVR value retained for the study was the mean of 3 consecutive measures during a single evaluation. When patients had several PVR measures by ultrasound device spanning several days, the PVR value considered for the study was the one performed at the time of urine sampling, or the closest one. Patients with PVR measured performed more than 48 h apart from urine sampling were excluded from the study. For patients who required bladder catheterization (either for an acute retention or for a chronic voiding dysfonction), the PVR was measured by catheterization. PVR values up to 100 mL were considered as normal for that population of adult inpatients with a majority of subjects older than 60 years (see below) (Gehrich et al., 2007; Grabe et al., 2014). Patients with internal or external urinary bypass, indwelling Foley catheter, ascitis, and pregnant women were excluded because PVR cannot reliably be measured by ultrasound portable devices in such situations. 2.4. Ethical issues In accordance with French regulations, no written informed consent was required because the data were obtained anonymously from medical files, with no participation of patients.
2.3. PVR measure
2.5. Statistics
For patients with natural micturition, PVR was assessed by ultrasound portable device (BladderScann®, BVI 9400; Verathon, USA)
Data were expressed as mean and SD. Continuous values were compared using Mann–Whitney test. To assess the accuracy of BT at log10
Fig. 1. Post-void residual urine (PVR) according to the bacterial titer (BT) in the urine sample of patients with urinary tract infection (UTI), asymptomatic bacteriuria (AB) or for the total series.
Please cite this article as: Caron F, et al, High bacterial titers in urine are predictive of abnormal postvoid residual urine in patients with urinary tract infection, Diagn Microbiol Infect Dis (2015), http://dx.doi.org/10.1016/j.diagmicrobio.2015.05.003
F. Caron et al. / Diagnostic Microbiology and Infectious Disease xxx (2015) xxx–xxx
3
if their P value was b0.05. Multivariate analysis was performed using R satistical software (version 3.1.2, http://www.r-project.org). Twosided statistical analysis was performed using GraphPad software (version Prism 5.0 Software; USA) with P value ≤0.05 considered significant. 3. Results
Fig. 2. Mean volume of post-void residual urine (PVR) in patients with urine bacterial titers b 106 CFU/mL or ≥ 106 CFU/mL in a series of patients with urinary tract infection (UTI) or asymptomatic bacteriuria (AB).
quantity 2 to 7 to predict abnormal PVR (i.e., N100 mL), the sensitivity, specificity, and odds ratio were calculated and summarized in a receiver operating characteristic (ROC) curve analysis whose area under the curve was measured. PVR and BT were compared using 2-way tables. Plots were generated to illustrate the relationship between BT in urine cultures and PVR measures, and a Spearman's correlation coefficient was then calculated to explore a linear correlation. Clinical characteristics were individually entered into the univariate model as dependent variables. All the variables that were associated with high PVR in the univariate analysis at a P b 0.05 level were included in a multivariate model. Multiple linear regression was used, and variables were retained
A total of 103 patients were included, with a 0.61 sex ratio (64 females, 62%). The median age was 72 years (range: 19–99), with 79% of patients older than 60 years. There was a majority of UTI (78 cases, 76%), among which the most common clinical manifestation was acute pyelonephritis (48 cases, 62%) followed by male febrile UTI (25 cases, 32%). Two thirds of the UTI (50 cases, 64%) were considered as complicated, mainly due to male gender (29 cases, 37%), diabetes mellitus (18 cases, 23%), and underlying urologic disorder (15 cases, 19%). Only 16 cases (16%) of bacteriuria were healthcare related (12% of the UTI and 28% of the AB). E. coli grew from the urine sample of 88 patients (85%). The remainder cases were due to Klebsiella spp. (n = 8), Enterobacter spp. (n = 5), or Proteus spp. (n = 2 ). Half of the patients (52 cases, 50%) had BT ≥10 6 CFU/mL. Among the 78 cases diagnosed UTI, only 5 (6%) had BT b10 3 CFU/mL. Among the 25 cases diagnosed AB, 16 (64%) had BT ≥105 CFU/mL. Blood cultures were performed for 61 patients with febrile UTI and isolated the same bacteria than in urines for 18 (30%) of them, attesting of a concomitant bacteremia. PVR was assessed by ultrasound device for 84 (82%) patients and by catheterization for 19 (18%) other patients. The mean PVR was 175 mL (range: 0–900); 47% had a PVR ≤100 mL, while 25% had a PVR ≥250 mL. BT increased with PVR in patients with UTI and AB as well (Fig. 1). Though the correlation in the increase of BT and PVR was not linear (Spearman r coefficient = 0.3), a significant BT threshold was observed for a 106 CFU/mL value (Figs. 1 and 2). Indeed, for the total cohort, patients with BT ≤10 5 CFU/mL had a mean PVR of 100 mL, while those with BT N10 5 CFU/mL had a mean PVR of 248 mL (P b 0.01). This 10 6 CFU/mL threshold was statistically valid for patients with UTI (P b 0.01), for those with AB (P = 0.01), as well as for the total cohort (Fig. 2). Among the 52 patients with BT ≥10 6 CFU/mL, 40 (77%) had an abnormal PVR (i.e., N100 mL). Conversely, among the 51 patients with a BT ≤10 5 CFU/mL, 35 (69%) had a normal PVR. The correlation in between BT and PVR was summarized (Fig. 3) in a fitted ROC curve analysis (area under the curve: 0.737, P = 0.05), showing that BT ≥106 CFU/mL was predictive of an abnormal PVR with a 71% sensitivity and a 74% specificity. In univariate analysis, BT ≥106 CFU/mL was significantly associated with PVR N100 mL (P b 0.05), with a 2.8 odds ratio as compared to a patient with a BT ≤10 5 CFU/mL. Of note, all of the patients with UTI and a low BT (10 2 or 103 CFU/mL) had a PVR b100 mL. In univariate analysis, higher BT was noticed in patients with UTI (2.10 6 CFU/mL) than in patients with AB (5.10 5 CFU/mL) (P = 0.01), as shown Table 1. However, a direct correlation between higher BT in urine and a more severe UTI was not demonstrated: BT did not significantly differ according to bacteremia, risk factors of complication (Table 1), and delay of apyrexia. The relation between higher PVR and more severe UTI was more consistent. In fact, PVR values were more commonly N100 mL in patients with bacteremia (11/18, 61%) than in patients without bacteremia (20/43, 45%) (P b 0.05), and mean PVR values were higher in patients with bacteremia (211 mL versus 131 mL, P b 0.01). PVR values were also more commonly N100 mL in patients with complicated UTI (33/50, 66%) than in those with uncomplicated UTI (9/19, 32%) (P = 0.01), and mean PVR was higher in patients with complicated UTI (210 mL versus 115 mL, P b 0.01). Symptomatic patients with PVR values N100 mL also had a trend to a longer mean delay of apyrexia (56 h) than patients with PVR ≤100 mL (44 h) (P = 0.6). Parameters significantly associated with PVR N100 mL in univariate analysis were included in a multivariate analysis conducted on patients with febrile UTI: BT ≥106 CFU/mL and risk factors of complication were independently associated with higher PVR (P b 0.01), whereas bacteremia was not (P = 0.06).
Please cite this article as: Caron F, et al, High bacterial titers in urine are predictive of abnormal postvoid residual urine in patients with urinary tract infection, Diagn Microbiol Infect Dis (2015), http://dx.doi.org/10.1016/j.diagmicrobio.2015.05.003
4
F. Caron et al. / Diagnostic Microbiology and Infectious Disease xxx (2015) xxx–xxx
Fig. 3. Fitted ROC curve analysis of the accuracy of BT at log10 quantity 2 to 7 to predict abnormal PVR urine (i.e. N100 mL) in a series of patients with UTI or AB (area under fitted curve: 0.74, estimated SE: 0.05).
4. Discussion This work is, to our knowledge, the first study demonstrating a clear correlation between BTs in urine and PVR urine for bacteriuric patients. At first sight, it may seem surprising that such a basic finding had not been investigated yet. Two explanations can be postulated. First, most studies on bacterial thresholds have included women with acute uncomplicated cystitis (Hooton, 2012; Hummers-Pradier et al., 2005; Kunin et al., 1993; Stamm et al., 1982). In this specific population, the influence of abnormal voiding on BT could not be investigated since patients do not have any functionnal or anatomical disorder of the urinary tract. Second, until recently, it was very difficult to detect and, furthermore, to quantify in the daily practice an abnormal PVR, except for the rare patients with urine retention that require drainage. Ultrasound portable instruments (“bladder scan”) now allow a quick, accurate, and noninvasive measure of PVR. In our study, high BT was strongly and independantly correlated with uncomplete voiding and high PVR values, with a clear 106 CFU/mL cutoff: while three-quarters of the patients with UTI and a BT b106 CFU/mL had a complete or almost complete micturition (PVR ≤100 mL), threequarters of those with UTI at BT ≥106 CFU/mL had an abnormal amount of urine retained in the bladder after urination (PVR N 100 mL, up to 900 mL). The correlation between BT and PVR was not linear, as shown by the relatively low Spearman's coefficient of the corresponding curves, both during UTI and AB (1. 1). However, the 10 6 CFU/mL threshold offered the best compromise in terms of sensitivity (71%) and specificity (74%), as shown by the ROC curve. This finding strongly suggests that urine retention is correlated with microbial proliferation in the bladder. Though it cannot be excluded that bacterial colonization might inflame
the bladder mucosa or the prostate and, consequently, alter bladder voiding, it seems rather likely that uncomplete micturition favors bacterial proliferation and UTI. This hypothesis is reinforced by previous series having demonstrated that high PVR is a risk factor for UTI or AB (Carlson et al., 2001; Truzzi et al., 2008) and that the reduction of PVR reduces the incidence of UTI (Minardi et al., 2010). Moreover, in our study, like in others (Marschall et al., 2012), patients with higher PVR had a poorer outcome as was shown by significantly higher rates of bacteremia and by the trend to a longer delay of apyrexia. These findings might have 3 practical consequences in routine. First, in the short term, wherever mobile ultrasound is not available, patients with BT ≥10 6 CFU/mL should immediately be explored for acute retention or severe voiding dysfunction by urgent ultrasonography in order to implement precocious urine drainage and prevent bacteremia and severe manifestations. Second, among patients with AB that mostly do not deserve immediate antibiotic treatment, patients with high BT should be investigated for voiding dysfunction to prevent UTI. Third, at the lab, a precise count of BT in urines up to 10 6 CFU/mL has to be favored for inpatients. At the present time, numerous laboratories in many countries mention high BT as ≥105 CFU/mL but do not specify if they reach 106 CFU/mL, a practice totally founded until it is demonstrated that more precise results of BT have practical consequences in the management of patients, as suggested by the current series. This work has some limitations. First, only a relatively low number of patients was included, but this did not preclude the observation of highly significant differences for bacterial cut-off at 106 CFU/mL, as discussed above. Second, due to the retrospective design of the study, the real prevalence of abnormal PVR among inpatients with UTI or AB could not be determined. Indeed, the current study intended to include only patients who required PVR evaluation by ultrasound or by bladder catheterization for a suspected voiding dysfunction. As a consequence, half of the patients had an abnormal PVR (50% with a PVR N100 mL and 25% with a PVR N250 mL). Though one can assume that the real prevalence of abnormal micturition might be overestimated in the current study, yet the correlation between high BT and high PVR was ascertained in this population of inpatients at risk for voiding dysfunction. Whether the link between high BT and high PVRs remains in broader populations is an area of further study. 5. Conclusion In conclusion, screening for urine BTs ≥106 CFU/mL is an easy noninvasive and reliable method to identify inpatients with abnormal PVR that require immediate urine drainage and that should secondarily be investigated for underlying voiding disorders. Funding No funding was received for this study. Conflicts of interest All authors: no potential conflicts of interest.
Table 1 Mean BTs and PVR urine according to gender, clinical and biological manifestations in a series of 103 patients with UTI or AB. Mean BT value (CFU/mL) Male (n = 39) Female (n = 64) UTI (n = 78) AB (n = 25) Uncomplicated UTI (n = 28) Complicated UTI (n = 50) Febrile UTI (n = 61) With bacteremia (n = 43) Without bacteremia (n = 18)
106 106 106 105 106 106
P = 0.27
2.6 × 106 1.6 × 106
P = 0.26
1.6 1.9 2.2 7.5 1.2 2.2
× × × × × ×
P = 0.01 P = 0.32
Mean PVR value (mL) 191 165 176 173 115 210
P = 0.13
211 131
P b 0.05
P = 0.67 P b 0.01
References Aspevall O, Hallander H, Gant V, Kouri T. European guidelines for urinalysis: a collaborative document produced by European clinical microbiologists and clinical chemists under ECLM in collaboration with ESCMID. Clin Microbiol Infect, 7(4). Blackwell Science Ltd.; 2001. p. 173–8. Carlson KV, Romes S, Nitti VW. Dysfunctionnal voiding in women. J Urol 2001;165(1): 143–8. Foxman B. Epidemiology of urinary tract infections: incidence, morbidity, and economic costs. Am J Med 2002;113(Suppl. 1A):5S–13S. Gehrich A, Stany MP, Fischer JR, Buller J, Zahn CM. Establishing a mean postvoid residual volume in asymptomatic perimenopausal and postmenopausal women. Obstet Gynecol 2007;110(4):827–32. Grabe M, Bartoletti R, Bjerklund-Johansen TE. European Association of Urology Guidelines on Urological Infections; 2014.
Please cite this article as: Caron F, et al, High bacterial titers in urine are predictive of abnormal postvoid residual urine in patients with urinary tract infection, Diagn Microbiol Infect Dis (2015), http://dx.doi.org/10.1016/j.diagmicrobio.2015.05.003
F. Caron et al. / Diagnostic Microbiology and Infectious Disease xxx (2015) xxx–xxx Hooton TM. Clinical practice. Uncomplicated urinary tract infection. N Engl J Med 2012; 366(11):1028–37. Hooton TM, Bradley SF, Cardenas DD, Colgan R, Geerlings SE, Rice JC, et al. Diagnosis, Prevention, and Treatment of Catheter-Associated Urinary Tract Infection in Adults: 2009 International Clinical Practice Guidelines from the Infectious Diseases Society of America. Clin Infect Dis 2010;50(5):625–63. Hooton TM, Roberts PL, Cox ME, Stapleton AE. Voided midstream urine culture and acute cystitis in premenopausal women. N Engl J Med 2013;369(20):1883–91. Hummers-Pradier E, Ohse AM, Koch M, Heizmann WR, Kochen MM. Management of urinary tract infections in female general practice patients. Fam Pract 2005;22(1):71–7. Kunin CM, White LV, Hua TH. A reassessment of the importance of “low-count” bacteriuria in young women with acute urinary symptoms. Ann Intern Med 1993;119(6):454–60. Manickam K, Karlowsky JA, Adam H, Lagacé-Wiens PRS, Rendina A, Pang P, et al. CHROMagar Orientation medium reduces urine culture workload. J Clin Microbiol 2013;51(4):1179–83. Marschall J, Zhang L, Foxman B, Warren DK, Henderson JP. CDC Prevention Epicenters Program. Both host and pathogen factors predispose to Escherichia coli urinarysource bacteremia in hospitalized patients. Clinical Infectious Diseases, 54(12). Oxford University Press; 2012. p. 1692–8.
5
Minardi D, d'Anzeo G, Parri G, Polito M, Piergallina M, Asmar El Z, et al. The role of uroflowmetry biofeedback and biofeedback training of the pelvic floor muscles in the treatment of recurrent urinary tract infections in women with dysfunctional voiding: a randomized controlled prospective study. , 75(6). Elsevier Inc.; 2010. p. 1299–304. Nicolle LE, Bradley S, Colgan R, Rice JC, Schaeffer A, Hooton TM, et al. Infectious Diseases Society of America guidelines for the diagnosis and treatment of asymptomatic bacteriuria in adults. Clin Infect Dis 2005:643–54. O'Grady F, Cattell WR. Kinetics of urinary tract infection. I. Upper urinary tract. Br J Urol 1966a;38(2):149–55. O'Grady F, Cattell WR. Kinetics of urinary tract infection. II. The bladder. Br J Urol 1966b; 38(2):156–62. Sobel JD. Pathogenesis of urinary tract infections. Host defenses. Infect Dis Clin North Am 1987;1(4):751–72. Stamm WE, Counts GW, Running KR, Fihn S, Turck M, Holmes KK. Diagnosis of coliform infection in acutely dysuric women. N Engl J Med 1982;307(8): 463–8. Truzzi JCI, Almeida FMR, Nunes EC, Sadi MV. Residual urinary volume and urinary tract infection—when are they linked? J Urol 2008;180(1):182–5.
Please cite this article as: Caron F, et al, High bacterial titers in urine are predictive of abnormal postvoid residual urine in patients with urinary tract infection, Diagn Microbiol Infect Dis (2015), http://dx.doi.org/10.1016/j.diagmicrobio.2015.05.003
Contents lists available at ScienceDirect
Diagnostic Microbiology and Infectious Disease journal homepage: www.elsevier.com/locate/diagmicrobio
High bacterial titers in urine are predictive of abnormal postvoid residual urine in patients with urinary tract infection F. Caron a, K. Alexandre b, M. Pestel-Caron c, P. Chassagne d, P. Grise e, M. Etienne a,⁎ a
Infectious Diseases Department, Rouen University Hospital, GRAM (EA2656), Rouen, France Infectious Diseases Department, Rouen University Hospital, Rouen, France Microbiology Department, Rouen University Hospital, GRAM (EA2656), Rouen, France d Geriatrics Department, Rouen University Hospital, Rouen, France e Urology Department, Rouen University Hospital, Rouen, France b c
a r t i c l e
i n f o
Article history: Received 18 February 2015 Received in revised form 30 April 2015 Accepted 11 May 2015 Available online xxxx Keywords: Post-void residual urine Urine analysis Urinary tract infection
a b s t r a c t Urine bacterial titers (BTs) are influenced by bacterial and host factors. The impact of an abnormal postvoid residual (PVR) on BT in urine was investigated. A total of 103 inpatients with a urine growing Enterobacteriacae (≥102 CFU/mL) and a PVR measure were analyzed, mostly female (62%), elderly (mean age: 72 years), with urinary tract infection (25% of asymptomatic bacteriuria) due to Escherichia coli (85%). Fifty-two subjects (56%) had BT ≥106 CFU/mL; 48 (53%) had a PVR ≤100 mL, while 26 (25%) had a PVR N250 mL. PVR increased with BT, and a significant (P b 0.0001) threshold was reached for 106 CFU/mL: 100 mL mean PVR for patients with BT ≤105 CFU/mL versus 248 mL for patients with BT N105 CFU/mL. High PVR and BT were associated with complicated infections, concomitant bacteremia, and delayed apyrexia. Screening for patients with BT ≥106 CFU/mL is an easy way to identify patients at high risk for acute retention and voiding disorders. © 2015 Elsevier Inc. All rights reserved.
1. Introduction Urine culture is by far the most common microbiological analysis performed in daily practice worldwide, due to both the high incidence of urinary tract infection (UTI) or asymptomatic bacteriuria (AB) that deserve detection and treatment (e.g., during pregnancy or before an urological surgery) and to the high feasibility and acceptability of the method (Foxman, 2002; Manickam et al., 2013). Because of a common contamination of the sample by periurethral bacteria (even for midstream urines), colony count thresholds have been established to help interpret positive culture (Hooton et al., 2013). For decades, numerous studies have investigated the influence of several bacterial or host factors on such thresholds. It is usually admitted that thresholds are lower for bacterial species such as Escherichia coli, Proteus spp., and Staphylococcus saprophyticus that produce virulence factors favoring invasion of the urothelium (i.e., the so-called uropathogens), as compared to organisms less frequently responsible for UTI such as enterococci or Pseudomonas spp. (Aspevall et al., 2001). It is also admitted that thresholds for UTI are lower than for AB (Nicolle et al., 2005). Curiously, the influence of voiding dysfonction on bacterial titers (BTs) in urines of patients with UTI or AB has been theorized but has not been investigated in clinical studies (O'Grady and Cattell, 1966a,b). Conceptually, high postvoid residual (PVR) urine is at risk to ⁎ Corresponding author. Tel.: +33-2-32-88-87-39, +33-6-24-94-87-56 (cell); fax: +33-2-32-88-86-42. E-mail addresses: [email protected], [email protected] (M. Etienne).
favor microbial proliferation, while complete voiding should clear most bacteria from bladder (Sobel, 1987). Thus, a positive correlation between BT and PVR values could be expected, but this remains to be demonstrated. For this purpose, the primary aim of this study was to evaluate the correlation between BT and PVR among patients with positive urine culture. The secondary aim was to analyze the role of other factors that might interfere with both parameters (BT and PVR): gender, AB or UTI, risk factors for complication, and concomitant bacteremia. 2. Materials and methods 2.1. Patients This study was conducted retrospectively at the Rouen University Hospital (Normandy, France) from January 2013 to May 2014. Patients with an urine sample growing Enterobacteriacae (whatever the BT, and the presence or absence of leucocyturia) were first identified by the department of microbiology. Patients were then included if they had at least 1 PVR measure within 48 h after the urine sample was collected. The study was limited to the departments of infectious diseases, geriatrics, and urology where a PVR measure by portable ultrasound device is performed when urine retention is suspected. A diagnosis of UTI was retained for patients with clinical symptoms, BT ≥102 CFU/mL, and leukocyte count N10 4/mL. Patients were considered at risk for complicated UTI in case of functional or anatomical abnormality of the urinary tract, male gender, diabetes mellitus, or if they had healthcare-related infections. For patients with febrile UTI (T ≥ 38 °C), the delay of apyrexia
http://dx.doi.org/10.1016/j.diagmicrobio.2015.05.003 0732-8893/© 2015 Elsevier Inc. All rights reserved.
Please cite this article as: Caron F, et al, High bacterial titers in urine are predictive of abnormal postvoid residual urine in patients with urinary tract infection, Diagn Microbiol Infect Dis (2015), http://dx.doi.org/10.1016/j.diagmicrobio.2015.05.003
2
F. Caron et al. / Diagnostic Microbiology and Infectious Disease xxx (2015) xxx–xxx
was noticed; it was defined as the interval between the initiation of an effective antibiotic regimen and 2 consecutive temperature measures ≤37.3 °C in the morning and 37.8 °C in the afternoon. A diagnosis of asymptomatic bacteriuria was retained in asymptomatic patients with bacteriuria, whatever the leukocyturia. Healthcare-related bacteriuria were defined as those occurring 48 h after admission or within 1 month after any invasive urological procedure (Hooton et al., 2010). 2.2. BT measure The BT was determined using quantitative urinanalysis, which consisted of inoculating 10-μL urine onto a nonselective chromogenic agar plate (CPS3; bioMérieux, Marcy l'Etoile, France) using a calibrated loop. After 24 h of incubation at 37 °C in aerobic conditions, the number of colonies was counted and expressed as CFU/mL and quantified in a 102 to ≥107 range for AB and UTI, in order to investigate the correlation between BT and PVR. Patients with urine growing organisms other than Enterobacteriacae were excluded because detection of bacteria such as enterococci or streptococci in midstream urines is not necessarily predictive of their presence in bladder, as recently established among premenopausal women with acute uncomplicated cystitis (Hooton et al., 2013).
within 10 minutes of voiding. The PVR value retained for the study was the mean of 3 consecutive measures during a single evaluation. When patients had several PVR measures by ultrasound device spanning several days, the PVR value considered for the study was the one performed at the time of urine sampling, or the closest one. Patients with PVR measured performed more than 48 h apart from urine sampling were excluded from the study. For patients who required bladder catheterization (either for an acute retention or for a chronic voiding dysfonction), the PVR was measured by catheterization. PVR values up to 100 mL were considered as normal for that population of adult inpatients with a majority of subjects older than 60 years (see below) (Gehrich et al., 2007; Grabe et al., 2014). Patients with internal or external urinary bypass, indwelling Foley catheter, ascitis, and pregnant women were excluded because PVR cannot reliably be measured by ultrasound portable devices in such situations. 2.4. Ethical issues In accordance with French regulations, no written informed consent was required because the data were obtained anonymously from medical files, with no participation of patients.
2.3. PVR measure
2.5. Statistics
For patients with natural micturition, PVR was assessed by ultrasound portable device (BladderScann®, BVI 9400; Verathon, USA)
Data were expressed as mean and SD. Continuous values were compared using Mann–Whitney test. To assess the accuracy of BT at log10
Fig. 1. Post-void residual urine (PVR) according to the bacterial titer (BT) in the urine sample of patients with urinary tract infection (UTI), asymptomatic bacteriuria (AB) or for the total series.
Please cite this article as: Caron F, et al, High bacterial titers in urine are predictive of abnormal postvoid residual urine in patients with urinary tract infection, Diagn Microbiol Infect Dis (2015), http://dx.doi.org/10.1016/j.diagmicrobio.2015.05.003
F. Caron et al. / Diagnostic Microbiology and Infectious Disease xxx (2015) xxx–xxx
3
if their P value was b0.05. Multivariate analysis was performed using R satistical software (version 3.1.2, http://www.r-project.org). Twosided statistical analysis was performed using GraphPad software (version Prism 5.0 Software; USA) with P value ≤0.05 considered significant. 3. Results
Fig. 2. Mean volume of post-void residual urine (PVR) in patients with urine bacterial titers b 106 CFU/mL or ≥ 106 CFU/mL in a series of patients with urinary tract infection (UTI) or asymptomatic bacteriuria (AB).
quantity 2 to 7 to predict abnormal PVR (i.e., N100 mL), the sensitivity, specificity, and odds ratio were calculated and summarized in a receiver operating characteristic (ROC) curve analysis whose area under the curve was measured. PVR and BT were compared using 2-way tables. Plots were generated to illustrate the relationship between BT in urine cultures and PVR measures, and a Spearman's correlation coefficient was then calculated to explore a linear correlation. Clinical characteristics were individually entered into the univariate model as dependent variables. All the variables that were associated with high PVR in the univariate analysis at a P b 0.05 level were included in a multivariate model. Multiple linear regression was used, and variables were retained
A total of 103 patients were included, with a 0.61 sex ratio (64 females, 62%). The median age was 72 years (range: 19–99), with 79% of patients older than 60 years. There was a majority of UTI (78 cases, 76%), among which the most common clinical manifestation was acute pyelonephritis (48 cases, 62%) followed by male febrile UTI (25 cases, 32%). Two thirds of the UTI (50 cases, 64%) were considered as complicated, mainly due to male gender (29 cases, 37%), diabetes mellitus (18 cases, 23%), and underlying urologic disorder (15 cases, 19%). Only 16 cases (16%) of bacteriuria were healthcare related (12% of the UTI and 28% of the AB). E. coli grew from the urine sample of 88 patients (85%). The remainder cases were due to Klebsiella spp. (n = 8), Enterobacter spp. (n = 5), or Proteus spp. (n = 2 ). Half of the patients (52 cases, 50%) had BT ≥10 6 CFU/mL. Among the 78 cases diagnosed UTI, only 5 (6%) had BT b10 3 CFU/mL. Among the 25 cases diagnosed AB, 16 (64%) had BT ≥105 CFU/mL. Blood cultures were performed for 61 patients with febrile UTI and isolated the same bacteria than in urines for 18 (30%) of them, attesting of a concomitant bacteremia. PVR was assessed by ultrasound device for 84 (82%) patients and by catheterization for 19 (18%) other patients. The mean PVR was 175 mL (range: 0–900); 47% had a PVR ≤100 mL, while 25% had a PVR ≥250 mL. BT increased with PVR in patients with UTI and AB as well (Fig. 1). Though the correlation in the increase of BT and PVR was not linear (Spearman r coefficient = 0.3), a significant BT threshold was observed for a 106 CFU/mL value (Figs. 1 and 2). Indeed, for the total cohort, patients with BT ≤10 5 CFU/mL had a mean PVR of 100 mL, while those with BT N10 5 CFU/mL had a mean PVR of 248 mL (P b 0.01). This 10 6 CFU/mL threshold was statistically valid for patients with UTI (P b 0.01), for those with AB (P = 0.01), as well as for the total cohort (Fig. 2). Among the 52 patients with BT ≥10 6 CFU/mL, 40 (77%) had an abnormal PVR (i.e., N100 mL). Conversely, among the 51 patients with a BT ≤10 5 CFU/mL, 35 (69%) had a normal PVR. The correlation in between BT and PVR was summarized (Fig. 3) in a fitted ROC curve analysis (area under the curve: 0.737, P = 0.05), showing that BT ≥106 CFU/mL was predictive of an abnormal PVR with a 71% sensitivity and a 74% specificity. In univariate analysis, BT ≥106 CFU/mL was significantly associated with PVR N100 mL (P b 0.05), with a 2.8 odds ratio as compared to a patient with a BT ≤10 5 CFU/mL. Of note, all of the patients with UTI and a low BT (10 2 or 103 CFU/mL) had a PVR b100 mL. In univariate analysis, higher BT was noticed in patients with UTI (2.10 6 CFU/mL) than in patients with AB (5.10 5 CFU/mL) (P = 0.01), as shown Table 1. However, a direct correlation between higher BT in urine and a more severe UTI was not demonstrated: BT did not significantly differ according to bacteremia, risk factors of complication (Table 1), and delay of apyrexia. The relation between higher PVR and more severe UTI was more consistent. In fact, PVR values were more commonly N100 mL in patients with bacteremia (11/18, 61%) than in patients without bacteremia (20/43, 45%) (P b 0.05), and mean PVR values were higher in patients with bacteremia (211 mL versus 131 mL, P b 0.01). PVR values were also more commonly N100 mL in patients with complicated UTI (33/50, 66%) than in those with uncomplicated UTI (9/19, 32%) (P = 0.01), and mean PVR was higher in patients with complicated UTI (210 mL versus 115 mL, P b 0.01). Symptomatic patients with PVR values N100 mL also had a trend to a longer mean delay of apyrexia (56 h) than patients with PVR ≤100 mL (44 h) (P = 0.6). Parameters significantly associated with PVR N100 mL in univariate analysis were included in a multivariate analysis conducted on patients with febrile UTI: BT ≥106 CFU/mL and risk factors of complication were independently associated with higher PVR (P b 0.01), whereas bacteremia was not (P = 0.06).
Please cite this article as: Caron F, et al, High bacterial titers in urine are predictive of abnormal postvoid residual urine in patients with urinary tract infection, Diagn Microbiol Infect Dis (2015), http://dx.doi.org/10.1016/j.diagmicrobio.2015.05.003
4
F. Caron et al. / Diagnostic Microbiology and Infectious Disease xxx (2015) xxx–xxx
Fig. 3. Fitted ROC curve analysis of the accuracy of BT at log10 quantity 2 to 7 to predict abnormal PVR urine (i.e. N100 mL) in a series of patients with UTI or AB (area under fitted curve: 0.74, estimated SE: 0.05).
4. Discussion This work is, to our knowledge, the first study demonstrating a clear correlation between BTs in urine and PVR urine for bacteriuric patients. At first sight, it may seem surprising that such a basic finding had not been investigated yet. Two explanations can be postulated. First, most studies on bacterial thresholds have included women with acute uncomplicated cystitis (Hooton, 2012; Hummers-Pradier et al., 2005; Kunin et al., 1993; Stamm et al., 1982). In this specific population, the influence of abnormal voiding on BT could not be investigated since patients do not have any functionnal or anatomical disorder of the urinary tract. Second, until recently, it was very difficult to detect and, furthermore, to quantify in the daily practice an abnormal PVR, except for the rare patients with urine retention that require drainage. Ultrasound portable instruments (“bladder scan”) now allow a quick, accurate, and noninvasive measure of PVR. In our study, high BT was strongly and independantly correlated with uncomplete voiding and high PVR values, with a clear 106 CFU/mL cutoff: while three-quarters of the patients with UTI and a BT b106 CFU/mL had a complete or almost complete micturition (PVR ≤100 mL), threequarters of those with UTI at BT ≥106 CFU/mL had an abnormal amount of urine retained in the bladder after urination (PVR N 100 mL, up to 900 mL). The correlation between BT and PVR was not linear, as shown by the relatively low Spearman's coefficient of the corresponding curves, both during UTI and AB (1. 1). However, the 10 6 CFU/mL threshold offered the best compromise in terms of sensitivity (71%) and specificity (74%), as shown by the ROC curve. This finding strongly suggests that urine retention is correlated with microbial proliferation in the bladder. Though it cannot be excluded that bacterial colonization might inflame
the bladder mucosa or the prostate and, consequently, alter bladder voiding, it seems rather likely that uncomplete micturition favors bacterial proliferation and UTI. This hypothesis is reinforced by previous series having demonstrated that high PVR is a risk factor for UTI or AB (Carlson et al., 2001; Truzzi et al., 2008) and that the reduction of PVR reduces the incidence of UTI (Minardi et al., 2010). Moreover, in our study, like in others (Marschall et al., 2012), patients with higher PVR had a poorer outcome as was shown by significantly higher rates of bacteremia and by the trend to a longer delay of apyrexia. These findings might have 3 practical consequences in routine. First, in the short term, wherever mobile ultrasound is not available, patients with BT ≥10 6 CFU/mL should immediately be explored for acute retention or severe voiding dysfunction by urgent ultrasonography in order to implement precocious urine drainage and prevent bacteremia and severe manifestations. Second, among patients with AB that mostly do not deserve immediate antibiotic treatment, patients with high BT should be investigated for voiding dysfunction to prevent UTI. Third, at the lab, a precise count of BT in urines up to 10 6 CFU/mL has to be favored for inpatients. At the present time, numerous laboratories in many countries mention high BT as ≥105 CFU/mL but do not specify if they reach 106 CFU/mL, a practice totally founded until it is demonstrated that more precise results of BT have practical consequences in the management of patients, as suggested by the current series. This work has some limitations. First, only a relatively low number of patients was included, but this did not preclude the observation of highly significant differences for bacterial cut-off at 106 CFU/mL, as discussed above. Second, due to the retrospective design of the study, the real prevalence of abnormal PVR among inpatients with UTI or AB could not be determined. Indeed, the current study intended to include only patients who required PVR evaluation by ultrasound or by bladder catheterization for a suspected voiding dysfunction. As a consequence, half of the patients had an abnormal PVR (50% with a PVR N100 mL and 25% with a PVR N250 mL). Though one can assume that the real prevalence of abnormal micturition might be overestimated in the current study, yet the correlation between high BT and high PVR was ascertained in this population of inpatients at risk for voiding dysfunction. Whether the link between high BT and high PVRs remains in broader populations is an area of further study. 5. Conclusion In conclusion, screening for urine BTs ≥106 CFU/mL is an easy noninvasive and reliable method to identify inpatients with abnormal PVR that require immediate urine drainage and that should secondarily be investigated for underlying voiding disorders. Funding No funding was received for this study. Conflicts of interest All authors: no potential conflicts of interest.
Table 1 Mean BTs and PVR urine according to gender, clinical and biological manifestations in a series of 103 patients with UTI or AB. Mean BT value (CFU/mL) Male (n = 39) Female (n = 64) UTI (n = 78) AB (n = 25) Uncomplicated UTI (n = 28) Complicated UTI (n = 50) Febrile UTI (n = 61) With bacteremia (n = 43) Without bacteremia (n = 18)
106 106 106 105 106 106
P = 0.27
2.6 × 106 1.6 × 106
P = 0.26
1.6 1.9 2.2 7.5 1.2 2.2
× × × × × ×
P = 0.01 P = 0.32
Mean PVR value (mL) 191 165 176 173 115 210
P = 0.13
211 131
P b 0.05
P = 0.67 P b 0.01
References Aspevall O, Hallander H, Gant V, Kouri T. European guidelines for urinalysis: a collaborative document produced by European clinical microbiologists and clinical chemists under ECLM in collaboration with ESCMID. Clin Microbiol Infect, 7(4). Blackwell Science Ltd.; 2001. p. 173–8. Carlson KV, Romes S, Nitti VW. Dysfunctionnal voiding in women. J Urol 2001;165(1): 143–8. Foxman B. Epidemiology of urinary tract infections: incidence, morbidity, and economic costs. Am J Med 2002;113(Suppl. 1A):5S–13S. Gehrich A, Stany MP, Fischer JR, Buller J, Zahn CM. Establishing a mean postvoid residual volume in asymptomatic perimenopausal and postmenopausal women. Obstet Gynecol 2007;110(4):827–32. Grabe M, Bartoletti R, Bjerklund-Johansen TE. European Association of Urology Guidelines on Urological Infections; 2014.
Please cite this article as: Caron F, et al, High bacterial titers in urine are predictive of abnormal postvoid residual urine in patients with urinary tract infection, Diagn Microbiol Infect Dis (2015), http://dx.doi.org/10.1016/j.diagmicrobio.2015.05.003
F. Caron et al. / Diagnostic Microbiology and Infectious Disease xxx (2015) xxx–xxx Hooton TM. Clinical practice. Uncomplicated urinary tract infection. N Engl J Med 2012; 366(11):1028–37. Hooton TM, Bradley SF, Cardenas DD, Colgan R, Geerlings SE, Rice JC, et al. Diagnosis, Prevention, and Treatment of Catheter-Associated Urinary Tract Infection in Adults: 2009 International Clinical Practice Guidelines from the Infectious Diseases Society of America. Clin Infect Dis 2010;50(5):625–63. Hooton TM, Roberts PL, Cox ME, Stapleton AE. Voided midstream urine culture and acute cystitis in premenopausal women. N Engl J Med 2013;369(20):1883–91. Hummers-Pradier E, Ohse AM, Koch M, Heizmann WR, Kochen MM. Management of urinary tract infections in female general practice patients. Fam Pract 2005;22(1):71–7. Kunin CM, White LV, Hua TH. A reassessment of the importance of “low-count” bacteriuria in young women with acute urinary symptoms. Ann Intern Med 1993;119(6):454–60. Manickam K, Karlowsky JA, Adam H, Lagacé-Wiens PRS, Rendina A, Pang P, et al. CHROMagar Orientation medium reduces urine culture workload. J Clin Microbiol 2013;51(4):1179–83. Marschall J, Zhang L, Foxman B, Warren DK, Henderson JP. CDC Prevention Epicenters Program. Both host and pathogen factors predispose to Escherichia coli urinarysource bacteremia in hospitalized patients. Clinical Infectious Diseases, 54(12). Oxford University Press; 2012. p. 1692–8.
5
Minardi D, d'Anzeo G, Parri G, Polito M, Piergallina M, Asmar El Z, et al. The role of uroflowmetry biofeedback and biofeedback training of the pelvic floor muscles in the treatment of recurrent urinary tract infections in women with dysfunctional voiding: a randomized controlled prospective study. , 75(6). Elsevier Inc.; 2010. p. 1299–304. Nicolle LE, Bradley S, Colgan R, Rice JC, Schaeffer A, Hooton TM, et al. Infectious Diseases Society of America guidelines for the diagnosis and treatment of asymptomatic bacteriuria in adults. Clin Infect Dis 2005:643–54. O'Grady F, Cattell WR. Kinetics of urinary tract infection. I. Upper urinary tract. Br J Urol 1966a;38(2):149–55. O'Grady F, Cattell WR. Kinetics of urinary tract infection. II. The bladder. Br J Urol 1966b; 38(2):156–62. Sobel JD. Pathogenesis of urinary tract infections. Host defenses. Infect Dis Clin North Am 1987;1(4):751–72. Stamm WE, Counts GW, Running KR, Fihn S, Turck M, Holmes KK. Diagnosis of coliform infection in acutely dysuric women. N Engl J Med 1982;307(8): 463–8. Truzzi JCI, Almeida FMR, Nunes EC, Sadi MV. Residual urinary volume and urinary tract infection—when are they linked? J Urol 2008;180(1):182–5.
Please cite this article as: Caron F, et al, High bacterial titers in urine are predictive of abnormal postvoid residual urine in patients with urinary tract infection, Diagn Microbiol Infect Dis (2015), http://dx.doi.org/10.1016/j.diagmicrobio.2015.05.003
