Embed Size (px)
Citation preview
UNCERTAINTIES
Do men with lower urinary tract symptoms have anincreased risk of advanced prostate cancer?Jenny Østerø í Jákupsstovu research assistant 1, John Brodersen professor 1 2
1Centre of Research and Education in General Practice, Department of Public Health, University of Copenhagen, Denmark; 2Primary Health CareResearch Unit, Region Zealand, Denmark
What you need to know• It is uncertain if lower urinary tract symptoms are associated with an
increased risk of advanced prostate cancer• Do not routinely offer prostate specific antigen (PSA) testing in men
with only lower urinary tract symptoms and no risk factors for prostatecancer
• Explain the limitations and harms of a PSA test to the patient in termsof downstream procedures, and the possibility of being diagnosed withprostate cancer that might not cause him symptoms or shorten his life
A 72 year old man complains of increased frequency ofurination, particularly at night. He also describes a sense ofincomplete voiding and a poor stream. He is concerned aboutwhether he might have prostate cancer, and wants to be reassuredthat this is not the case.More than half of all men over 50 report problems withurinating.1
Grouped as lower urinary tract symptoms (LUTS), these signifyproblems with the storage and/or voiding of urine, and are oftencaused by benign prostatic enlargement.1 LUTS are habituallyconsidered a possible symptom of prostate cancer,2 andqualitative studies show that men with LUTS are worried abouthaving prostate cancer and consult a doctor to seekreassurance.3-5 Prostate cancer is one of the most diagnosedcancers in men worldwide. During the previous two to threedecades, prostate cancer incidence has greatly increased in highincome countries; however, at the same time the mortality ofprostate cancer has stayed stable or decreased slightly. Advancedprostate cancer is potentially fatal. In contrast, indolent prostatecancer is per definition harmless and does not cause morbidityor mortality. Moreover, indolent prostate cancer is the mostfrequent form of prostate cancer, eg, autopsy studies estimatethe mean prevalence of indolent prostate cancer to be 59% inmen >79 years.6 Collated with the high prevalence of LUTS inolder men, a connection between LUTS and indolent prostatecancer is nearly unavoidable. Most general practice clinicalguidelines (table 1) recommend doctors to consider the prostate
specific antigen (PSA) test in men with LUTS for prostate cancerdetection.7 9-11 However, it is uncertain if men with LUTS haveincreased risk of advanced or potentially fatal prostate cancerand therefore, if they should be offered the PSA test.The controversy of PSA screening is much discussed, and nowa consensus on not screening asymptomatic men is emerging(appendix 1).12 13 Data from the five Nordic countries suggestthat, despite a rise in the diagnosis of prostate cancer,particularly since the introduction of PSA testing, there has beenlittle change in death from the disease (fig 1).14 Patterns suchas these suggest overdiagnosis. Recognising the lack of benefitsin early prostate cancer detection,12 15 PSA testing in men withLUTS places them at risk of being diagnosed with indolentprostate cancer and being treated for a condition that would notresult in either morbidity or mortality.What is the evidence of uncertainty?
Search strategy and study selectionWe searched PubMed in September 2016 with the free-text and MESH-terms:“lower urinary tract symptoms” OR “LUTS” AND “prostatic neoplasm” OR“prostate cancer” OR “prostatic cancer” OR “cancer of the prostate.” Weidentified four16-19 observational studies that examine whether LUTS is predictiveof advanced and potentially fatal prostate cancer. We included studiesconducted in a general practice population or screening population, withtransparent identification of people with and without LUTS and a specifiedPSA level as indication for biopsy. We excluded studies conducted in aselected population, eg, referred patients, studies where the differentiation ofpeople with and without LUTS was not clear, and studies that had otherindications for biopsy, eg, abnormal digital rectal examination.
The current evidence is limited to observational studies primarilyconducted in PSA screening populations and suggests that menwith self-reported LUTS are not at increased risk of havingadvanced or potentially fatal prostate cancer compared withmen without LUTS (see table 2 and appendix 1).These studies are large and population based but they vary indesign. Three16-18 of four studies were conducted in PSAscreening populations. The Norwegian19 and Swedish17 studiesused longitudinal follow-up, and the other two studies16 18 used
Correspondence to [email protected]
For personal use only: See rights and reprints http://www.bmj.com/permissions Subscribe: http://www.bmj.com/subscribe
BMJ 2018;361:k1202 doi: 10.1136/bmj.k1202 (Published 3 May 2018) Page 1 of 6
Practice
PRACTICE
a cross-sectional design, although they were based on PSAscreened cohorts. The tools used to record symptoms aredifferent and there is a potential for bias. Three16 18 19 of the fourstudies used a validated self-reported questionnaire to measureLUTS. In two studies,18 19 asymptomatic men were asked tocomplete a questionnaire on lower urinary tract symptoms, andin the other two studies16 17 men with raised PSA levels wereasked about symptoms. The outcomes varied in terms of severityof prostate cancer and the time period. Only one study19 recordedprostate cancer incidence and mortality nine years later.Three16-18 of four studies offered biopsy when PSA was ≥3ng/ml.9 Hence, the true association between LUTS and prostatecancer now or in the future remains unknown, and thediscovered association can only be applied for men with a raisedPSA level. An American population based cohort study hasshown that increased diagnostic intensity among men whopresent with LUTS is possible for the association between LUTSand prostate cancer observed in some studies.20
Is ongoing research likely to providerelevant evidence?We searched clinical trial registers (clinicaltrials.gov,controlled-trials.com, who.int/trialsearch) in September 2017for observational studies of associations between LUTS andcurrent or future prostate cancer diagnosis. We did not identifyany relevant studies examining whether LUTS are predictiveof advanced or potentially fatal prostate cancer or studiesoffering PSA testing in men with LUTS for cancer detection.
What should we do in light of theuncertainty?There is no consensus regarding PSA testing in men with LUTS,and most guidelines recommend shared decision making (table1).7 9-11 In addition to cancer detection, urologists request thePSA test in men with LUTS to assess the prostate volume andpredict prostate growth and clinical progression.21 Someguidelines support the use of clinical nomograms in selectedpatients to determine the risk of high grade prostate cancer.10
Otherwise, current guidelines recommend a PSA test if thepatient’s LUTS are suggestive of bladder outlet obstruction,7 ifthe man is over 40 with unexplained symptoms such as lowerback pain, bone pain, or weight loss suggestive of metastaticprostate cancer,10 or there is a family history of prostate cancer.11
We have not identified evidence that these recommendationsimprove the prognosis in men with LUTS.There is insufficient evidence to recommend PSA testing inmen with only LUTS and no other known risk factors forprostate cancer.Explore why your patient is concerned about prostate cancer,and assist him in making an informed choice regarding PSAtesting. Ask for age, family history of prostate cancer, andethnicity to assess risk for prostate cancer.Offer to perform a physical examination including a digitalrectal examination that can help detect prostate enlargement orcancer. Offer a urine dipstick test to detect blood, glucose,protein, leucocytes, and nitrites. You might ask the patient tocomplete a urinary frequency volume chart to assess the severityand perhaps help pinpoint the type of the problem andfurthermore aiding the decision making regarding medicaltreatment.7
Discuss with your patient the lack of direct evidence that LUTSsymptoms in isolation suggest prostate cancer. Explain the
possible limitations and harms of a PSA test such as downstreamprocedures, and the possibility of being overdiagnosed withprostate cancer and related treatment. When communicating theuncertainty about testing, using terms like “abnormal indolentcells in the prostate gland” rather than “localised prostate cancer”might reassure the patient and reduce patient preferences foraggressive management responses to a low risk condition.22
Recommendations for further research• A prognostic intervention study to evaluate if men with LUTS are
overdiagnosed with prostate cancer if they have a PSA test, and howthe PSA test affects their prognosis. Men with LUTS in general practicewould be randomly selected to receive PSA testing or not. Both groupswould be offered usual care including a digital rectal examination, urinaryfrequency volume chart, urine flow rate examination, andpharmacological or surgical treatment for their LUTS in discussion withtheir healthcare provider. Primary outcomes include prostate cancer:localised or disseminated, prostate cancer mortality and all-causemortality. Secondary outcomes include quality of life, morbidity, andtreatment.
• A large prognostic observational cohort study including men in generalpractice with baseline assessment of LUTS and other factors potentiallypredictive of prostate cancer. The cohort should be prospectivelyfollowed and prostate cancer related outcomes measured. This studymight reveal predictors of advanced prostate cancer and potentiallyhelp develop a clinical nomogram to aid decision making in men withLUTS.
• Research for new biomarkers capable of distinguishing between lowgrade and high grade prostate cancer
What patients need to know• There is no evidence that problems with urinating suggest advanced
prostate cancer• Your doctor will perform a digital rectal examination and might discuss
prostate specific antigen (PSA) testing if you have a family history ofprostate cancer in addition to lower urinary tract symptoms (LUTS)
• Be aware of the risk of overdiagnosis and overtreatment of an indolentprostate cancer if you have a PSA test with LUTS as a single symptomwith no other risk factors
Education into practice• How do you discuss lower urinary tract symptoms with a patient who
is concerned about the risk for prostate cancer?• To whom do you consider offering a PSA test for prostate cancer? Does
this article offer you ideas for change?• How do you explain the benefits and harms of testing for prostate
cancer? Are there ways you could alter this?• If you were to make a leaflet at your practice for men with LUTS, what
information about prostate cancer testing would you include?
How patients were involved in the creation of this articleA man who had a PSA test as part of medical examination kindly agreed toreview this article. He shared his concern regarding the high rate of indolentprostate cancer in older men, but was also concerned about overdiagnosisand said a diagnostic label could cause fear in patients. He suggested thatdoctors consider the personal context of the patient in the diagnostic processand stressed the need for patient information tools, especially information onpotential harms. We have summarised the current evidence and ways in whichdoctors might provide this crucial information to patients.
We would like to extend our sincerest thanks and appreciation to Iona Heath forher critical comments and linguistic editing.
We have read and understood The BMJ policy on declaration of interests anddeclare that we have no competing interests.
Provenance and peer review: Commissioned, based on an idea from the author;externally peer reviewed.
1 Vedanayagam M, Kumar A, Madaan S. Lower urinary tract symptoms in an older man.BMJ 2017;357:j1493. 10.1136/bmj.j1493 28420635
For personal use only: See rights and reprints http://www.bmj.com/permissions Subscribe: http://www.bmj.com/subscribe
BMJ 2018;361:k1202 doi: 10.1136/bmj.k1202 (Published 3 May 2018) Page 2 of 6
PRACTICE
2 NHS prostate cancer symptoms. Available from: http://www.nhs.uk/Conditions/Cancer-of-the-prostate/Pages/Symptoms.aspx.
3 Lammers HA, van Wijnhoven R, Teunissen TA, Harmsen S, Lagro-Janssen AL. Why domen suffering from LUTS seek primary medical care? A qualitative study. J Eval ClinPract 2015;21:931-6. 10.1111/jep.12407 26111045
4 Brown CT, O’Flynn E, Van Der Meulen J, Newman S, Mundy AR, Emberton M. The fearof prostate cancer in men with lower urinary tract symptoms: should symptomatic menbe screened?BJU Int 2003;91:30-2. 10.1046/j.1464-410X.2003.04013.x 12614245
5 Wong MC, Goggins WB, Wang HH, etal . Global Incidence and Mortality for ProstateCancer: Analysis of Temporal Patterns and Trends in 36 Countries. Eur Urol2016;70:862-74. 10.1016/j.eururo.2016.05.043 27289567
6 Bell KJ, Del Mar C, Wright G, Dickinson J, Glasziou P. Prevalence of incidental prostatecancer: A systematic review of autopsy studies. Int J Cancer 2015;137:1749-57.10.1002/ijc.29538 25821151
7 National Institute for Health and Care Excellence guidance. Lower urinary tract symptomsin men: management https://www.nice.org.uk/guidance/cg97/chapter/1-Recommendations.
8 American Urological Association. Management of benign prostatic hyperplasia (BPH).http://www.auanet.org/guidelines/benign-prostatic-hyperplasia-(2010-reviewed-and-validity-confirmed-2014) - x2512.
9 DSAM Klinisk vejledning for almen praksis: Udredning og behandling af nedreurinvejssymptomer hos mænd og kvinder http://www.dsam.dk/files/9/urinvejssymptomer_rettet.pdf.
10 Young SM, Bansal P, Vella ET, Finelli A, Levitt C, Loblaw AProstate Cancer ReferralWorking Group. Guideline for referral of patients with suspected prostate cancer by familyphysicians and other primary care providers. Can Fam Physician 2015;61:33-9.25756141
11 Prostate Cancer Management and Referral Guidance New Zealand. 2015http://www.health.govt.nz/system/files/documents/publications/prostate-cancer-management-referral-guidance_sept15-c.pdf.
12 Ilic D, Neuberger MM, Djulbegovic M, Dahm P. Screening for prostate cancer. CochraneDatabase Syst Rev 2013;1:CD004720.23440794
13 Lin K, Croswell JM, Koenig H, Lam C, Maltz A. Prostate-specific antigen-based screeningfor prostate cancer: An evidence update for the US preventive services task force. Rockville
(MD): Agency for Healthcare Research and Quality (US). A US preventive services taskforce evidence syntheses, formerly systematic evidence reviews. 2011.
14 NORDCAN project. http://www-dep.iarc.fr/NORDCAN/English/frame.asp.15 Wilt TJ, Jones KM, Barry MJ, etal . Follow-up of prostatectomy versus observation for
early prostate cancer. N Engl J Med 2017;377:132-42. 10.1056/NEJMoa1615869 2870084416 Collin SM, Metcalfe C, Donovan J, etal . Associations of lower urinary tract symptoms
with prostate-specific antigen levels, and screen-detected localized and advanced prostatecancer: a case-control study nested within the UK population-based ProtecT (Prostatetesting for cancer and Treatment) study. BJU Int 2008;102:1400-6.18540932
17 Frånlund M, Carlsson S, Stranne J, Aus G, Hugosson J. The absence of voiding symptomsin men with a prostate-specific antigen (PSA) concentration of ≥3.0 ng/mL is anindependent risk factor for prostate cancer: results from the Gothenburg RandomizedScreening Trial. BJU Int 2012;110:638-43. 10.1111/j.1464-410X.2012.10962.x 22540895
18 Matsubara A, Yasumoto H, Teishima J, etal . Lower urinary tract symptoms and risk ofprostate cancer in Japanese men. Int J Urol 2006;13:1098-102.10.1111/j.1442-2042.2006.01504.x 16903936
19 Martin RM, Vatten L, Gunnell D, Romundstad P, Nilsen TI. Lower urinary tract symptomsand risk of prostate cancer: the HUNT 2 Cohort, Norway. Int J Cancer 2008;123:1924-8.10.1002/ijc.23713 18661522
20 Weight CJ, Kim SP, Jacobson DJ, etal . The effect of benign lower urinary tract symptomson subsequent prostate cancer testing and diagnosis. Eur Urol 2013;63:1021-7.10.1016/j.eururo.2012.12.060 23313032
21 European Association of Urology. Treatment of Non-neurogenic Male LUTS http://uroweb.org/guideline/treatment-of-non-neurogenic-male-luts/ - 4.
22 Nickel B, Barratt A, Copp T, Moynihan R, McCaffery K. Words do matter: a systematicreview on how different terminology for the same condition influences managementpreferences. BMJ Open 2017;7:e014129. 10.1136/bmjopen-2016-014129 28698318
Published by the BMJ Publishing Group Limited. For permission to use (where not alreadygranted under a licence) please go to http://group.bmj.com/group/rights-licensing/permissions
For personal use only: See rights and reprints http://www.bmj.com/permissions Subscribe: http://www.bmj.com/subscribe
BMJ 2018;361:k1202 doi: 10.1136/bmj.k1202 (Published 3 May 2018) Page 3 of 6
PRACTICE
Tables
Table 1| Guidelines for PSA testing in men with LUTS
Recommendations regarding PSA testingCountry/guideline
Men with LUTS should be offered information, advice, and time to decide if they wish to have PSA testing if: - their symptoms are suggestive of bladder outlet obstruction secondary to benign prostate enlargement,or - the prostate feels abnormal on digital rectal examination, or - they are concerned about prostate cancer
UK/National Institute for Health and Care Excellence7
At initial evaluation for benign prostatic enlargement/LUTS, PSA is recommended in selected patients withat least a 10 year life expectancy for whom knowledge of the presence of prostate cancer would changemanagement, or patients for whom the PSA measurement might change the management of voiding symptoms
USA/American Urological Association8
All men with urinary problems should be tested for PSADenmark/Danish College of General Practitioners9
For a man presenting with lower urinary tract symptoms, perform a digital rectal examination, and discussthe benefits and harms of PSA testing
Canada/The College of Family Physicians ofCanada10
In men aged 50-70 (or over 40 if they have a family history of prostate cancer) who have prostate relatedconcerns, discuss the benefits and harms of PSA testing and obtain an informed consent before testing
New Zealand/Ministry of Health11
For personal use only: See rights and reprints http://www.bmj.com/permissions Subscribe: http://www.bmj.com/subscribe
BMJ 2018;361:k1202 doi: 10.1136/bmj.k1202 (Published 3 May 2018) Page 4 of 6
PRACTICE
Table 2| PICO summary of evidence
CommentsOutcomeComparisonInterventionPopulationStudy designStudy
Assessmentof LUTS waslimited by theuse of astudy-specificquestionnaire
Prostate cancer rate was higher amongasymptomatic men (378/1230=31% ofasymptomatic men vs 255/1123=23%of men with voiding symptoms,P<0.001). There was no statisticallysignificant difference in the proportion
Men who reportedvoiding symptomswere compared withmen without reportedvoiding symptoms
The attendants were invitedfor biennial measurement ofPSA from 1995 to 2010. Allmen with a PSA concentrationof ≥3.0 ng/mL were offeredfurther clinical assessment,
In 1995 20 000 menborn between 1930and 1944 wererandomly selectedfrom the Swedishpopulation register and
Prognosticcohort study(randomisedcontrolled trial)
Frånlund etal(Swedishstudy)
of locally advanced tumours betweenincluding digital rectal10 000 men werethe groups (36/378=9.5% ofexamination and trans-rectalrandomly allocated toasymptomatic men vs 29/255=11.4% ofultrasonography (TRUS)a PSA-screeningmen with voiding symptoms. P=0.473).guided biopsies of the prostate
(2590/7625 attendees).All men who accepted furtherevaluation were asked to
group and 10 000 mento a control group In the multivariate logistic regression
analysis, the presence of voidingsymptoms (odds risk 0.78, 95%confidence interval 0.63-0.98) wascomplete a questionnaire withindependently associated with lowerone question on obstructive
voiding symptoms (2353 men) odds of detecting prostate cancer inbiopsies
Prostate cancer was detected on biopsyin 29% (34/117) of asymptomatic menand in 28% (17/61) of symptomatic men.The difference was not statisticallysignificant (P value IPSS 0-7 vs IPSS8-35 0.2873 (χ2) in screened men and
Men with an IPSSscore of 0-7 wereclassified as beingasymptomatic(n=2574) and menwith an IPSS score of
The participants completedthe IPSS questionnaire(International ProstateSymptom Score) as ameasure of the severity ofLUTS when the PSA test was
3919 Men aged 50-79years, whoparticipated in aPSA-screeningprogramme in Japanbetween 2002 and2004
Cross-sectionalstudy
Matsubaraet al(Japanesestudy)
0.9010 (χ2) in biopsied men). These8-35 as beingdone. Disregarding the IPSSfindings were consistent across agesymptomaticscore, all participants with agroups. Multivariate logistic regression(n=4448). The cancerPSA value above 4.0 ng/mlanalysis for the biopsied men showeddetection rateswere invited to a TRUS guided
biopsy that IPSS score was not predictive ofbetween these twoprostate cancer (P-value 0.9812, 95%confidence interval (0.932-1.062)
groups werecompared
Unknown howcancer wasdiagnosed
Compared with men with no symptoms,men with moderate symptoms had twicethe risk of having prostate cancer (HR2.01). The association between LUTSseverity and prostate cancer wasconfined to localised prostate cancer
Men with differentIPSS levels werecompared with theasymptomatic men(IPSS=0) who servedas a reference group
The severity of LUTS wasclassified according to theIPSS score. Nine years laterthe prostate cancer incidenceand the prostate cancermortality in the cohort were
21159 Men aged 20years or older from theNord-TrøndelagHealth Study (HUNT2) in Norway who hadcompleted the IPSS
Prognosticcohort study
Martin et al(Norwegianstudy)
(hazard ratio 4.618 (95% confidenceidentified in the Norwegianquestionnaire frominterval 2.23-9.54) for IPSS 20-35).Cancer Registry and the1995 to 2007 wereThere was no association betweenNorwegian Cause of death
Registryidentified and followedup LUTS severity and advanced prostate
cancer incidence (hazard ratio 0.51(95% confidence interval 0.15-1.75%)for IPSS 20-35). There was noassociation between LUTS severity andprostate cancer mortality (hazard ratio0.73 (confidence interval 0.23-2.26) forIPSS 20-35)
LUTS were negatively associated withprostate cancer (odds ratio of 0.44 (95%confidence interval 0.22-0.83) fornocturia, 0.74 (95% confidence interval0.63-0.87) for leakage and 0.83 (95%confidence interval 0.73-0.94) for
Logistic regressionmodels were used toestimate odds ratiofor the associationbetween urinarysymptoms and
Men with a PSA value of ≥3.0ng/ml were invited forTRUS-guided biopsy. At thetime of the PSA screening, theparticipants completed theshort form version of the
A sub-population of 65871 men (aged 50-69years) in 2007 fromthe ProtecT (Prostatetesting for cancer andTreatment) study in
Cross-sectionalstudy
Collin et al(UK study)
hesitancy (occasionally/ sometimes vsprostate cancerInternational ContinenceUK who had all beenscreened with PSA never)).The association of LUTS with(localised orSociety Male Questionnaire
(ICSmaleSF) prostate cancer did not differ in theadvanced) in menwho had a biopsy proportion of locally advanced tumours
between the groups of symptomatic andasymptomatic men
For personal use only: See rights and reprints http://www.bmj.com/permissions Subscribe: http://www.bmj.com/subscribe
BMJ 2018;361:k1202 doi: 10.1136/bmj.k1202 (Published 3 May 2018) Page 5 of 6
PRACTICE
Figure
Fig 1 Prostate cancer incidence and mortality in Nordic countries 1953-201314
For personal use only: See rights and reprints http://www.bmj.com/permissions Subscribe: http://www.bmj.com/subscribe
BMJ 2018;361:k1202 doi: 10.1136/bmj.k1202 (Published 3 May 2018) Page 6 of 6
PRACTICE
