Concordance of Near Infrared Spectroscopy With Pressure FlowStudies in Men With Lower Urinary Tract Symptoms

Doreen E. Chung,* Richard K. Lee, Steven A. Kaplan and Alexis E. TeFrom the Department of Urology, Weill Cornell Medical College, New York-Presbyterian Hospital, Cornell University, New York, New York

Abbreviations

and Acronyms

BOO � bladder outlet obstruction

I-PSS � International ProstateSymptom Score

LUTS � lower urinary tractsymptoms

MAD � Madersbacher et alclassification for bladder outletobstruction in men

NIRS � near infraredspectroscopy

Pdet � detrusor pressure

PSA � prostate specific antigen

PVR � post-void residual urine

Qmax � maximum flow rate

QoL � quality of life

ROS � Rosier et al classificationfor bladder outlet obstruction inmen

Submitted for publication April 14, 2010.Study received institutional review board ap-

proval.* Correspondence: Division of Urology, Univer-

sity of Chicago, Mount Sinai Hospital, 1500 S.California Ave., Department of Surgery-F944, Chi-cago, Illinois 60608 (telephone: 773-257-2639; FAX:773-257-6548; e-mail: doreen_chung@yahoo.ca).

Purpose: Near infrared spectroscopy is a technology with the potential to diag-nose bladder outlet obstruction noninvasively. We investigated the correlationbetween the near infrared spectroscopy algorithm classification of obstructed orunobstructed (based on near infrared spectroscopy pattern, maximum flow rateand post-void residual) and pressure flow studies in men with lower urinary tractsymptoms. We also assessed whether the near infrared spectroscopy patternitself, independent of the near infrared spectroscopy algorithm (which also in-cludes maximum flow rate and post-void residual urine), correlates with bladderoutlet obstruction.Materials and Methods: Subjects simultaneously underwent urodynamics andnear infrared spectroscopy evaluation. Bladder outlet obstruction classificationwas based on the International Continence Society nomogram. The near infraredspectroscopy algorithm was evaluated against the comparable noninvasive algo-rithms with receiver operating characteristic curves. For significance the Mann-Whitney U test was used with p �0.05 considered significant.Results: Of 42 subjects 33 (79%) had evaluable data. Area under the curve was0.484 for the near infrared spectroscopy algorithm relative to pressure flowdiagnosis. A downward near infrared spectroscopy pattern is associated withobstruction whereas an upward pattern is associated with nonobstruction. Of 26patients with urodynamic bladder outlet obstruction 9 (35%) had a downward, 4(15%) a flat and 13 (50%) an upward near infrared spectroscopy pattern. Of 7unobstructed cases 4 (57%) had a downward, 1 (14%) a flat and 2 (29%) anupward near infrared spectroscopy pattern.Conclusions: The near infrared spectroscopy algorithm had a 0.484 AUC fordiagnosing bladder outlet obstruction. The near infrared spectroscopy patterncomponent itself did not correlate strongly. The near infrared spectroscopy algo-rithm does not appear to provide substantial clinical usefulness in diagnosingbladder outlet obstruction in men with lower urinary tract symptoms. Morestudies are needed to further define and validate uses for near infrared spectros-copy in urology.

Key Words: spectroscopy, near-infrared; urodynamics; prostatic hyperplasia

2434 www.jurology.com

LOWER urinary tract symptoms inolder men may occur from BOO, de-trusor overactivity, impaired con-tractility or any combination of theseprocesses. Diagnosing BOO in thispopulation can be challenging. Stud-

ies show that men with urodynamic

0022-5347/10/1846-2434/0THE JOURNAL OF UROLOGY®

© 2010 by AMERICAN UROLOGICAL ASSOCIATION EDUCATION AND RES

obstruction benefit more from trans-urethral resection of the prostatethan those without obstruction, whichis a reason distinguishing BOO inmen is important.1,2

The gold standard for diagnosing

BOO is pressure flow studies. It was

Vol. 184, 2434-2439, December 2010Printed in U.S.A.

EARCH, INC. DOI:10.1016/j.juro.2010.08.027

NEAR INFRARED SPECTROSCOPY FOR DIAGNOSIS OF BLADDER OUTLET OBSTRUCTION 2435

stated at the 5th International Consultation on Be-nign Prostatic Hyperplasia that pressure flow stud-ies remain the only modality to establish BOO withacceptable reproducibility and reliability.3 Urody-namics also remain the only available study foridentifying the etiology of LUTS in men such asobstruction and detrusor overactivity.4

A disadvantage of pressure flow studies is inva-siveness as they require bladder and rectal cathe-terization. Several noninvasive methods for diag-nosing BOO have been studied including Dopplerresistive index, bladder wall thickness and Dopplerultrasound urodynamics, but none have been shownto reliably predict BOO.5 NIRS is an advancing tech-nology that may have the potential to study detrusorfunction noninvasively.

NIRS uses light in the near infrared spectrum tomonitor changes in the concentration of chro-mophores (oxyhemoglobin and deoxyhemoglobin) intissue. Current applications include pulse oximetryand cerebral oxygenation monitoring.6,7 Near infra-red photons are absorbed largely by chromophores,and less so by water, fat and other proteins.8 TheBeer-Lambert law is the physical and mathematicalbasis of NIRS applications. It states that light pass-ing through a solution of a colored compound (chro-mophore) is absorbed by that compound, resulting inreduced intensity of emerging light.9

Macnab and Stothers described the developmentof an instrument to interrogate the bladder detrusor(Tetra™ NIRS).10 They hypothesized that NIRSmay be able to evaluate the etiology of voiding dys-function if it is related to oxidative metabolism orhemodynamics. They also theorized that detrusorpressure and chromophore concentration may be di-rectly related.11 To assess BOO an algorithm to clas-sify men with LUTS as being obstructed or unob-structed was developed based on the 3 componentsof data from NIRS (pattern of slope of changes inchromophore concentration), Qmax and PVR. Initialstudies showed greater than 80% concordance withpressure flow studies.12 According to the NIRS algo-rithm a downward chromophore concentration sloperelates to a higher probability of obstruction and anupward slope relates to a higher probability of non-obstruction.

Blood flow and oxidative metabolism are factorspertinent to NIRS that influence total and oxyhemo-globin concentration changes. They are assumed tobe the basis for the relationship between BOO andchanges in chromophore concentration,10,12 and the-oretically contribute to the upward and downwardslope of chromophore concentration change whilevoiding in unobstructed and obstructed cases. Datafrom animal studies demonstrate that filling de-creases bladder dome blood flow and that emptying

elicits reactive hyperemia.13 Muscle contraction de-

pends on oxygen consumption, decreasing total andoxyhemoglobin concentrations.14 In BOO less reac-tive hyperemia may occur, potentially leading to thedownward slope of chromophore concentrationchange.

The primary study objective was to determine ifthe NIRS algorithm classification (based on NIRSpattern, free flow Qmax and PVR) correlated withthe classification of obstructed or unobstructedbased on pressure flow studies in men with LUTS.The secondary objective was to compare the predic-tive value of the NIRS algorithm to that of 2 othernoninvasive algorithms for BOO.15,16 A post hocanalysis was performed to assess whether the NIRSpattern itself, independent of the NIRS algorithm(which also includes Qmax and PVR), correlateswith pressure flow diagnosis.

MATERIALS AND METHODS

After institutional review board approval men with LUTSreferred for urodynamics were simultaneously evaluatedwith NIRS. All patients had serum PSA measured, com-pleted the I-PSS including the QoL item and underwenttransrectal ultrasound to determine prostate volume. Ex-clusion criteria were urinary retention, prior radical pros-tatectomy, neurological disease or voiding less than 100ml on uroflowmetry. To be able to compare the NIRSalgorithm diagnosis of obstructed or unobstructed withpressure flow diagnosis, patients with equivocal BOO onurodynamics were excluded from study.

The NIRS device emits 3 different lasers in the nearinfrared spectrum by a multifilament optical, through tis-sue to a photodiode receiver. It is attached to the patientby a suprapubic adhesive patch. Macnab and Stothershave described NIRS technology and the methodology forvoiding function analysis.12

NIRS data were collected during free uroflowmetry aswell as during pressure flow studies. Baseline NIRS datawere collected for 30 seconds before patients were givenpermission to void. Patients were catheterized for PVRbefore urodynamics. Testing was performed according toInternational Continence Society guidelines. Pressureflow classification of BOO was based on the InternationalContinence Society nomogram.17

The NIRS algorithm classification of BOO is based onthe pattern of chromophore concentration (upward, flat,downward) during voiding, Qmax (ml per minute) andPVR (ml).12 Upward NIRS pattern, Qmax 12 ml per sec-ond or greater and PVR less than 80 ml are each assigneda point for nonobstruction. Downward NIRS pattern,Qmax less than 10 and PVR 120 ml or greater are eachassigned a point for obstruction. The category, obstructedor unobstructed, with the greater number of points deter-mines classification. For an equivalent score the algorithmfavors a diagnosis of obstruction but the final decision ismade by the physician based on clinical information. Freeflow NIRS algorithm classification was compared to pres-

sure flow diagnosis. The slope of the chromophore concen-

78

NEAR INFRARED SPECTROSCOPY FOR DIAGNOSIS OF BLADDER OUTLET OBSTRUCTION2436

tration from NIRS was recorded and compared to thepressure flow diagnosis of BOO.

To further investigate the ability of the NIRS algorithmto diagnose BOO we compared the NIRS algorithm clas-sification from the free flow phase to 2 comparable nonin-vasive classification systems for BOO. Madersbacher et aldeveloped nomograms for BOO using Qmax, prostate vol-ume and PVR.15 Using this nomogram we classified pa-tients with a greater than 60% chance of obstruction asobstructed. The second BOO classification system towhich we compared the NIRS algorithm was the clinicalprostate score of Rosier et al.16 This system combines freeQmax, prostate size, PVR and voided urine volume toderive a score. Using data from 871 men with LUTS thesensitivity of this score, with 11 or higher classified asobstructed, was 80.7%, and the specificity was 53.1%.

During pressure flow studies NIRS data were collectedsimultaneously with pressure flow parameters. NIRS al-gorithm and pressure flow diagnoses from this phase werecompared. For tests of significance the Mann-Whitney Utest was used and p �0.05 was considered significant.ROC curves were used to compare noninvasive classifica-tion systems for BOO. The statistical software used wasSPSS® 15.0.

RESULTS

A total of 42 patients were enrolled in the study and33 (79%) had evaluable data. Three patients wereexcluded from analysis for prior radical prostatec-tomy and 4 voided an inadequate volume at uroflow-metry. Of 2 patients with equivocal BOO on urody-namics 1 was excluded from analysis due tocommunication error with urodynamics instru-ments.

Mean patient age was 67 � 14 years. BaselineI-PSS was 19 � 6 and mean QoL score was 4 � 1. Atotal of 26 (79%) patients were obstructed on pres-sure flow studies and 7 (21%) were unobstructed.Baseline data, free flow parameters and pressure

Table 1. Baseline, free flow and pressure flow parameters

Mean � SD All Pts M

I-PSS 19 � 7QoL 4 � 1Prostate vol (ml) 48 � 39PSA (ng/ml) 2.0 � 2.3Free flow study:

Qmax (ml/sec) 10.0 � 5.5Voided vol (ml) 197 � 134PVR (ml) 161 � 188

Pressure flow study:Qmax (ml/sec) 6.7 � 4.4Voided vol (ml) 179 � 131PVR (ml) 287 � 209Pdet at Qmax (cm H2O) 72 � 31Max Pdet 80 � 32BOO index 64 � 36

flow parameters are listed in table 1. Patients with

obstruction had a larger prostate (53 � 42 vs 28 � 23ml, p � 0.048) and higher PSA (2.3 � 2.4 vs 0.6 � 0.8ng/ml, p � 0.022) than those unobstructed. MeanPVR after free flow was 161 � 188 ml and wassimilar between the groups.

On pressure flow studies patients with obstruc-tion had significantly lower Qmax (5.7 � 3.7 vs10.4 � 5.0 ml per second, p � 0.007), higher Pdet atQmax (83 � 24 vs 33 � 13 cm H2O, p � 0.000),higher maximum Pdet (90 � 29 vs 43 � 12 cm H2O,p � 0.000) and higher BOO index (78 � 28 vs 16 �10, p � 0.000). Mean voided volume and PVR werenot significantly different between the groups.Figure 1 illustrates ROC curves for the NIRS algo-rithm, MAD and ROS. The AUC was 0.484 for theNIRS algorithm classification, 0.632 for MAD and0.703 for ROS in relation to pressure flow classifica-tion.

Next we analyzed the correlation of the NIRSpattern itself (independent of Qmax and PVR) toBOO. Free flow NIRS patterns are shown in table 2.

SD Obstructed Mean � SD Unobstructed p Value

� 7 18 � 7 0.774� 1 4 � 1 0.450� 42 28 � 23 0.048� 2.4 0.6 � 0.8 0.025

� 5.2 12.4 � 6.6 0.217� 92 299 � 230 0.067� 206 157 � 113 0.596

� 3.7 10.4 � 5.0 0.007� 104 283 � 171 0.042� 213 270 � 209 0.982� 24 33 � 13 0.000� 29 43 � 12 0.000� 28 16 � 10 0.000

1.00.80.60.40.20.0

1 - Specificity

1.0

0.8

0.6

0.4

0.2

0.0

Sens

itivi

ty

Reference LineRosObstrMarlObstrFNIRSObst

Source of the Curve

Diagonal segments are produced by ties.

Free Flow NIRS Algorithm Classifica�on

Madersbacher et al.’s Classifica�on

Rosier et al.’s Classifica�on

Reference Line

Figure 1. ROC curves of NIRS algorithm classification, MAD andROS in relation to diagnosis of obstructed or unobstructed

ean �

195

532.3

9.3173162

5.71502928390

based on pressure flow studies.

NEAR INFRARED SPECTROSCOPY FOR DIAGNOSIS OF BLADDER OUTLET OBSTRUCTION 2437

To review, according to the NIRS algorithm a down-ward NIRS pattern relates to a higher probability ofobstruction and an upward pattern relates to a higherprobability of nonobstruction. Of 26 patients ob-structed on pressure flow studies 9 (35%) had a down-ward, 4 (15%) a flat and 13 (50%) an upward NIRSpattern. Of 7 unobstructed cases 4 (57%) had a down-ward, 1 (14%) had a flat and 2 (29%) had an upwardNIRS pattern.

NIRS data were also collected during pressureflow studies. The goal of NIRS is to provide a non-invasive means to diagnose BOO. However, wewanted to examine the correlation between theNIRS algorithm classification and the pressure flowclassification of BOO using data collected simulta-neously during urodynamics. The ROC curve of theNIRS algorithm classification in relation to pressureflow classification is shown in figure 2. The AUC forthe NIRS algorithm during pressure flow studies is0.548. Table 2 shows NIRS data from urodynamics.

The NIRS pattern during urodynamics was in-cluded for 23 subjects. In 5 cases data were excludedfor inadequate voided volume, and in 5 there wasmiscommunication between urodynamics equipmentand the NIRS instrument. Of 18 patients with obstruc-tion 11 (61%) had a downward, 2 (11%) a flat and 5(28%) an upward NIRS pattern. Of 5 unobstructedcases 3 (60%) had a downward NIRS pattern duringpressure flow studies and 2 (40%) had an upwardpattern.

DISCUSSION

The NIRS algorithm is comparable to other algo-rithms combining multiple noninvasive parametersto predict BOO.16,18 Madersbacher et al studied 253patients with LUTS suggestive of BOO to determineif obstruction could be predicted by clinical evalua-tion.15 They observed that prostate volume, Qmax,PVR and voided volume correlated significantly withdegree of BOO rated according to linear passiveurethral resistance (linPURR). Nomograms were es-tablished by multiple logistic regression analysis for

Table 2. Free flow and pressure flow NIRS patterns

No.Obstructed (%)

No.Unobstructed (%) Totals (%)

Free flow NIRS (33):Downward 9 (35) 4 (57) 13 (39)Flat 4 (15) 1 (14) 5 (15)Upward 13 (50) 2 (29) 15 (45)

Pressure flow NIRS (23):Downward 11 (61) 3 (60) 14 (61)Flat 2 (11) 0 (0) 2 (9)Upward 5 (28) 2 (40) 7 (30)

the probability of BOO in patients with LUTS based

on Qmax, PVR and prostate volume. Using thesenomograms the probability of BOO is high in pa-tients with a Qmax less than 10 ml per second and aPVR greater than 50 ml. For prostate size greaterthan 50 ml the probability is 97% to 99%.

Rosier et al compared the results of free uroflow-metry and transrectal ultrasound to urodynamic di-agnosis of BOO in 871 consecutive men.16 Logisticregression was used to assign a relative weight toeach of free Qmax, voided volume, prostate volumeand PVR to construct a clinical score for obstruction,called the clinical prostate score. The sensitivity forthis score to detect obstruction when the lower limitof the clinical prostate score was 11 was 80.7% andspecificity was 11.8%.

In this study the noninvasive NIRS algorithm,based on free Qmax, PVR and NIRS pattern, had anAUC of 0.484 in relation to pressure flow studies inclassifying men with LUTS as obstructed or unob-structed. In this group the NIRS algorithm classifica-tion was inferior to MAD (AUC 0.632) and ROS (AUC0.703). ROS may have had a higher correlation withpressure flow studies due to a greater number of fac-tors combined. Like the NIRS algorithm MAD used acombination of 3 factors. If prostate volume has ahigher correlation with BOO than the NIRS pattern,this could explain the higher comparative correlationof MAD with pressure flow studies.

These results differ from those from the valida-tion study of the NIRS algorithm by Macnab andStothers.12 In their study males 19 years old or older(with no specific exclusion criteria) underwent freeuroflowmetry and pressure flow studies simulta-neously with NIRS. All patients with an equivocaldiagnosis of BOO based on pressure flow studies

1.00.80.60.40.20.01 - Specificity

1.0

0.8

0.6

0.4

0.2

0.0

Sens

itivi

ty

Reference LineUNIRSObstFNIRSObst

Source of the Curve

Diagonal segments are produced by ties.

Free-Flow NIRS Algorithm Classifica�on

Pressure-Flow NIRS Algorithm Classifica�on

Source of the Curve

Reference Line

Figure 2. ROC curves of NIRS algorithm classification of ob-structed or unobstructed in relation to pressure flow diagnosis,from data captured during pressure flow study and free flow

study.

NEAR INFRARED SPECTROSCOPY FOR DIAGNOSIS OF BLADDER OUTLET OBSTRUCTION2438

were assigned a diagnosis of obstructed or unob-structed based on other clinical data. Of 28 men withpressure flow BOO 24 (86%) had BOO with theNIRS algorithm. Of 27 unobstructed cases 24 (89%)were correctly classified by the NIRS algorithm.

In contrast, we excluded patients with equivocalBOO on pressure flow studies, prior radical prostatec-tomy, neurological disease or in urinary retention for amore homogenous population. In addition, we did notreclassify patients with equivocal BOO as being ob-structed or unobstructed. It is unclear from studies byMacnab and Stothers whether they used the NIRSalgorithm with data from free uroflowmetry or frompressure flow studies to validate the algorithm.10,12 Inthis study there was a higher correlation between theNIRS algorithm and pressure flow studies using datafrom the pressure flow study than from free uroflow-metry (AUC 0.548 vs 0.484, respectively).

Recently Stothers et al used classification andregression tree analysis with NIRS parameters toclassify 46 men with LUTS as being obstructed orunobstructed. 11 They found a misclassification rateof only 4% and a specificity of 88%.With linear dis-criminant analysis of concentrations of hemoglobinand oxyhemoglobin at Qmax, plotted against Pdet atQmax, the misclassification rate was 16%. Thisstudy had a sample size similar to ours, and used adifferent NIRS parameter from only the pressureflow phase. Absolute concentrations of hemoglobinand oxyhemoglobin may be better measures of ob-struction than concentration changes. However, evenwhen using the same NIRS algorithm as in thisstudy Macnab and Stothers had a much higher cor-relation with pressure flow diagnosis.10,12 An expla-nation for the higher accuracy of NIRS in both ofthese studies could be the NIRS equipment withcomparatively higher sensitivity.

For the NIRS algorithm the downward slope ofchange of chromophore concentration relates to ahigher probability of obstruction and the upwardslope relates to a higher probability of nonobstruc-tion. In this study this trend was not clearly seen.On free flow NIRS 35% of obstructed cases had adownward slope while 50% had an upward slope. Ofunobstructed cases 57% had a downward slope and29% had an upward slope. Using NIRS data frompressure flow studies 60% of unobstructed and 61%

of obstructed cases had a downward slope, whereas

REFERENCES

Long-term predictive role of urodynamics: an options: highlights from

40% of unobstructed and 28% of obstructed caseshad an upward slope.

That chromophore concentration changes are as-sociated with BOO is based on the hypothesis thatvascular changes are related to BOO. A reason thatslope of change in chromophore concentration (NIRSdata) did not strongly correlate is that this hypoth-esis may not be true. Another explanation is thatanother aspect of NIRS data may correspond betterwith obstruction in men with LUTS. In the valida-tion study by Macnab and Stothers the NIRS pat-tern of patients was not included.12 The high con-cordance rate of the NIRS algorithm with pressureflow diagnosis in their subjects may have been but-tressed by low Qmax and high PVR in obstructedand conversely in unobstructed cases.

Study limitations are sample size and lack of con-trol group without LUTS. Strengthening the study,on correlation between the NIRS algorithm classifi-cation and pressure flow classification of BOO inmen, is the clinical comparison to other noninvasiveclassification systems.

Even if information from NIRS cannot predictBOO it may, by assessing the hemodynamic andmetabolic component of voiding, provide insight intothe etiology and pathophysiology of LUTS that can-not be learned from pressure flow studies. Thiswould not be surprising since many of the agentsused to treat LUTS such as �-blockers, 5�-reductaseinhibitors and phosphodiesterase inhibitors have asignificant impact on vascular factors.

CONCLUSIONS

In this study the NIRS algorithm had an AUC of0.484 in relation to pressure flow studies when di-agnosing men with LUTS as obstructed or unob-structed. The NIRS pattern component itself did notcorrelate strongly with being obstructed or unob-structed based on pressure flow studies. The NIRSalgorithm does not currently provide substantialclinical usefulness in diagnosing BOO in men withLUTS. More studies are needed to determine theclinical value of NIRS in urology as well as to de-velop the specific NIRS parameters that will helpprovide insight into the voiding and storage phases

of micturition.

1. Ball AJ and Smith PJ: Urodynamic factors inrelation to outcome of prostatectomy. Urology1986; 28: 256.

2. Jensen KM, Jørgensen TB and Mogensen P:

8-year follow-up of prostatic surgery for lowerurinary tract symptoms. Br J Urol 1996; 78: 213.

3. Roehrborn CG: Focus on lower urinary tract symp-toms: nomenclature, diagnosis, and treatment

the 5th International

Consultation on Benign Prostatic HyperplasiaJune 25–27, 2000, Paris, France. Rev Urol 2001;3: 139.

4. Hyman MJ, Groutz A and Blaivas JG: Detrusor

instability in men: correlation of lower urinary

NEAR INFRARED SPECTROSCOPY FOR DIAGNOSIS OF BLADDER OUTLET OBSTRUCTION 2439

tract symptoms with urodynamic findings. J Urol2001; 166: 550.

5. Belal M and Abrams P: Noninvasive methods ofdiagnosing bladder outlet obstruction in men.Part 1: nonurodynamic approach. J Urol 2006;176: 22.

6. Fauchère JC, Schulz G, Haensse D et al: Near-infrared spectroscopy measurements of cerebraloxygenation in newborns during immediate post-natal adaptation. J Pediatr 2010; 156: 372.

7. Terborg C, Gröschel K, Petrovitch A et al: Nonin-vasive assessment of cerebral perfusion and ox-ygenation in acute ischemic stroke by near-infra-red spectroscopy. Eur Neurol 2009; 62: 338.

8. Calderon-Arnulphi M, Alaraj A and Slavin KV:Near infrared technology in neuroscience: past,present and future. Neurol Res 2009; 31: 605.

9. Gosain A, Rabkin J, Reymond JP et al: Tissueoxygen tension and other indicators of blood lossor organ perfusion during graded hemorrhage.

Surgery 1991; 109: 523.

10. Macnab AJ and Stothers L: Development of anear-infrared spectroscopy instrument for appli-cations in urology. Can J Urol 2008; 15: 4233.

11. Stothers L, Guevara R and Macnab A: Classifica-tion of male lower urinary tract symptoms usingmathematical modelling and a regression treealgorithm of noninvasive near-infrared spectros-copy parameters. Eur Urol 2010; 57: 327.

12. Macnab AJ and Stothers L: Near-infrared spec-troscopy: validation of bladder-outlet obstructionassessment using non-invasive parameters. CanJ Urol 2008; 15: 4241.

13. Kozlowski R, Siroky MB, Krane RJ et al: Regula-tion of blood flow and microcirculation resistancein rabbit bladder. J Urol 2002; 168: 1608.

14. Ward KR, Ivatury RR, Barbee RW et al: Nearinfrared spectroscopy for evaluation of thetrauma patient: a technology review. Resuscita-tion 2006; 68: 27.

15. Madersbacher S, Klingler HC, Djavan B et al: Is

obstruction predictable by clinical evaluation in

patients with lower urinary tract symptoms? Br JUrol 1997; 80: 72.

16. Rosier PF, de Wildt MJ, Wijkstra H et al: Clinicaldiagnosis of bladder outlet obstruction in patientswith benign prostatic enlargement and lower uri-nary tract symptoms: development and urody-namic validation of a clinical prostate score forthe objective diagnosis of bladder outlet obstruc-tion. J Urol 1996; 155: 1649.

17. Griffiths D, Höfner K, van Mastrigt R et al: Stan-dardization of terminology of lower urinary tractfunction: pressure-flow studies of voiding, ure-thral resistance, and urethral obstruction. Neur-ourol Urodyn 1997; 16: 1.

18. van Venrooij GE and Boon TA: The value ofsymptom score, quality of life score, maximalurinary flow rate, residual volume and prostatesize for the diagnosis of obstructive benign pros-tatic hyperplasia: a urodynamic analysis. J Urol

1996; 155: 2014.