The general agreement in the literature is that oral iron supplementation does not cause a false-positive guaiac-based fecal occult blood test (GFOBT).
GFOBT is based on rapid oxidization of α-guaiaconic acid to “guaiacum blue”, with hemoglobin serving as a catalyst through a non-enzymatic or “pseudoperoxidase” action. Although in vitro Fe3+ may serve as an oxidizing agent, this reaction is possible only under acidic conditions not found in the stool (pH ≥ 6-7)1. Also, in the absence of a catalyst, Fe3+ alone would not be expected to cause rapid (within 30 seconds) conversion of α-guaiaconic acid to guaiacum blue1.
Although a number of earlier clinical studies reported false-positive GBFOBT because of oral iron supplementation, subsequent investigations have uniformly failed to confirm these findings2. Potential reasons for earlier false-positive GBFOBT results include false interpretation of the color change—eg, green instead of blue— particularly when the discoloration is weakly positive, and non-standardized method of stool collection with the possibility of stool sample contamination by toilet water.
Other fascinating facts: Did you know that guaiac plant extract was used for centuries for treatment of syphilis and that the earliest application of guaiac testing was in forensic medicine in 1800s?
If you liked this post, sign up under MENU and catch future pearls right into your inbox!
- McDonnell WM, Ryan JA, Seeger DM, Elta GH. Effect of iron on the guaiac reaction. Gastroenterology. 1989 Jan;96(1):74-8. https://www.ncbi.nlm.nih.gov/pubmed/2909440
- Anderson GD, Yellig TR, Krone RE. An investigation into the effects of oral iron supplementation on in vivo hemoccult stool testing. Am J Gastroenterol 1990;85:558-561. https://www.ncbi.nlm.nih.gov/pubmed/218661
Contributed by Brian Li, Medical Student, Harvard Medical School
Yes! That’s because any urinary tract infection (UTI) in men has the potential for prostatic involvement1 —-as high as 83% by one report. 2
To make the matters more confusing, patients with acute bacterial prostatitis (ABP) often present with symptoms just like those of UTI, such as urinary frequency, dysuria, malaise, fever, and myalgias. 3 In the elderly, atypical presentation is not uncommon (eg, confusion, incontinence, fall). 4 Under these circumstances, bacteriuria and pyuria may also be related to ABP and the prostate exam should be an important part of your evaluation.
Although the sensitivity of prostate tenderness on digital rectal exam varies widely for ABP (9%-100%), a painful exam should raise suspicion for ABP, and by itself may be an independent predictor for clinical and bacteriologic failure of therapy. 1 Along with tenderness, fluctuance of prostate, particularly in the setting of voiding difficulties and longer duration of symptoms, may also suggest the presence of prostatic abscess. 5,6
But be gentle when performing a prostate exam and don’t massage it because you could potentially cause bacteremia and worsening of sepsis! 1,7
Liked this post? Download the app on your smart phone and sign up below to catch future pearls right into your inbox, all for free!
- Etienne M, Chavanet P, Sibert L, et al. Acute bacterial prostatitis: heterogeneity in diagnostic criteria and management. Retrospective multicentric analysis of 371 patients diagnosed with acute prostatitis. BMC Infectious Diseases 2008;8:12. https://bmcinfectdis.biomedcentral.com/track/pdf/10.1186/1471-2334-8-12?site=bmcinfectdis.biomedcentral.com
- Ulleryd P, Zackrisson B, Aus G, et al. Prostatic involvement in men with febrile urinary tract infection as measured by serum prostate-specific antigen and transrectal ultrasonography. BJU Int 1999;84:470-4. http://onlinelibrary.wiley.com/doi/10.1046/j.1464-410x.1999.00164.x/pdf
- Krieger JN, Nyberg L, Nickel JC. NIH consensus definition and classification. JAMA 1999;282:236-37. http://jamanetwork.com/journals/jama/article-abstract/1030245
- Harper M, Fowlis. Management of urinary tract infections in men. Trends in Urology Gynaecology & Sexual Health. January/February 2007. http://onlinelibrary.wiley.com/doi/10.1002/tre.8/pdf
- Lee DS, Choe HS, Kim HY, et al. Acute bacterial prostatitis and abscess formation. BMC Urology 2016;16:38. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4936164/
- Oliveira P, Andrade JA, Porto HC, et al. Diagnosis and treatment of prostatic abscess. International Braz J Urol 2003;29: 30-34. http://www.scielo.br/pdf/ibju/v29n1/v29n1a06.pdf
- Lipsky BA, Byren I, Hoey CT. Treatment of bacterial prostatitis. Clin Infect Dis 2010; 50:1641-52. https://academic.oup.com/cid/article/50/12/1641/305217
There is no firm evidence either for or against the use of systemic anticoagulants (ACs) in patients with asymptomatic pulmonary embolism (PE) following hemodialysis vascular access declotting (HVAD).
However, despite the common occurrence of asymptomatic PE following HVAD procedures (~40%), symptomatic PE—at times fatal—has also been reported in these patients1,2.
In the absence of hard data and any contraindications, anticoagulation can be justified in our patient for the following reasons:
- Asymptomatic segmental PE is commonly treated as symptomatic PE irrespective of setting2,3
- Hemodialysis patients are often considered hypercoagulable due to a variety of factors eg, platelet activation due to extracorporeal circulation, anti-cardiolipin antibody, lupus anticoagulant, decreased protein C or S activity, and/or reduced anti-thrombin III activity4-7
- Overall, chronic dialysis patients have higher incidence of PE compared to the general population8
- There is no evidence that asymptomatic PE following HVAD has a more benign course compared to that in other settings
- Untreated PE may be associated with repeated latent thrombosis or progression of thrombosis in the pulmonary artery5
- Calderon K, Jhaveri KD, Mossey R. Pulmonary embolism following thrombolysis of dialysis access: Is anticoagulation really necessary? Semin Dial 2010:23:522-25. https://www.ncbi.nlm.nih.gov/pubmed/21039878
- Sadjadi SA, Sharif-Hassanabadi M. Fatal pulmonary embolism after hemodialysis vascular access declotting. Am J Case Rep 2014;15:172-75. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4004792/pdf/amjcaserep-15-172.pdf
- Chiu V, O’Connell C. Management of the incidental pulmonary embolism. AJR 2017;208:485-88. http://www.ajronline.org/doi/pdf/10.2214/AJR.16.17201
- Kearon C, Akl EA, Ornelas J, et al. Antithrombotic therapy for VTE disease: Chest guideline and expert panel report. CHEST 2016;149:315-52. http://journal.chestnet.org/article/S0012-3692(15)00335-9/fulltext
- Yamasaki K, Haruyama N, Taniguchi M, et al. Subacute pulmonary embolism in a hemodialysis patient, successfully treated with surgical thrombectomy. CEN Case Rep 2016;5:74-77 https://link.springer.com/article/10.1007/s13730-015-0195-9
- Nampoory MR, Das KC, Johny KV, et al. Hypercoagulability, a serious problem in patients with ESRD on maintenance hemodialysis, and its correction after kidney transplantation. Am J Kidney Dis 2003;42:797-805. https://www.ncbi.nlm.nih.gov/pubmed/14520631
- O’Shea SI, Lawson JH, Reddan D, et al. Hypercoagulable states and antithrombotic strategies in recurrent vascular access site thrombosis. J Vasc Surg 2003;38: 541-48. http://www.jvascsurg.org/article/S0741-5214(03)00321-5/pdf
- Tveit DP, Hypolite IO, Hshieh P, et al. Chronic dialysis patients have high risk for pulmonary embolism. Am J Kidney Dis 2002;39:1011-17. https://www.ncbi.nlm.nih.gov/pubmed/11979344
Although traditionally we think of epistaxis as a potential sign of hypertension, particularly when severe, whether hypertension causes epistaxis is unclear and even the association of these 2 conditions has been challenged in recent years.
A 2014 systematic review found that although the majority of studies reported an association between these 2 conditions, many did not include a control group, were of poor methodological quality and did not adjust for confounding variables such as age, sex, and anticoagulation1. Indeed, a larger study that controlled for many potential confounding factors failed to confirm such an association2. A small prospective study also found no correlation between the severity of hypertension and epistaxis3.
Even when an association between hypertension and epistaxis has been found, it is unclear how much of the stress of bleeding itself and white coat syndrome may affect the readings1. However, an interesting 2017 study found masked hypertension (normal blood pressure in office, abnormal on ambulatory measurements) in 33.3% of patients with epistaxis with night time blood pressures that were significantly higher among patients with epistaxis4.
So the data is all over the place! It makes sense that long standing hypertension through its effects on blood vessels such as atherosclerosis and endothelium dysfunction may set the stage for epistaxis1,5, particularly in our ever-aging population on anticoagulants. But whether hypertension by itself is enough to cause epistaxis is likely to be debated for years to come.
- Kikidis D, Tsioufis K, Papanikolaou V, et al. Is epistaxis associated with arterial hypertension? A systematic review of the literature 2014;271:237-243. https://www.ncbi.nlm.nih.gov/pubmed/23539411
- Fuchs FD, Moreira LB, Pires CP, et al. Absence of association between hypertension and epistaxis: a population-based study. Blood Press 12:145-48. http://www.tandfonline.com/doi/abs/10.1080/08037050310001750
- Knopfholz J, Lima-Junior E, Précoma-Neto D, et al. Association between epistaxis and hypertension: A one year follow-up after an index episode of nose bleeding in hypertensive patients. Internat J Cardiol 2009;134:e107-e109. https://www.ncbi.nlm.nih.gov/pubmed/18499285
- Acar B, Yavuz B, Yildiz E, et al. A possible cause of epistaxis: increased masked hypertension prevalence in patients with epistaxis. Braz J Otorhinolaryngol 2017;83:45-49. http://www.scielo.br/pdf/bjorl/v83n1/1808-8694-bjorl-83-01-0045.pdf
- Celik T, Iyisoy A, Yuksel UC, et al. A new evidence of end-organ damage in the patients with arterial hypertension: epistaxis? Internat J Cardiol 2008;141:105-107. https://www.ncbi.nlm.nih.gov/pubmed/19138805
Even in this age of high-tech medicine, physical exam is still a great starting point for assessing the severity of aortic stenosis (AS) even if you are not a skilled cardiologist like most.
Start out by listening over the right clavicle. If you don’t hear a systolic murmur, you can be pretty confident that your patient doesn’t have moderate to severe AS (>98% sensitivity, LR 0.10)1.
If you hear a systolic murmur look for combination of findings that may increase the likelihood of moderate to severe AS: slow carotid artery upstroke, reduced carotid artery volume, maximal murmur intensity at the second right intercostal space, and reduced intensity of the second heart sound. The presence of 3 or 4 of these signs increases the likelihood of moderate to severe AS (LR 40), with less than 3 not helping much1.
When considered individually, many of the signs we often attribute to significant AS2 may not be as helpful in part because most of us are not skilled cardiologists and over the years the cause of AS has changed from primarily rheumatic heart disease-related to that advancing age and valve degeneration3.
So it may not be surprising that murmur intensity (eg grade 3/6 or above) may have a poor sensitivity and is an unreliable predictor of the severity of AS when patients with left ventricular failure are also studied3. Remember also that the absence of the 2nd sound may not distinguish between moderate and severe AS4.
- Etchells E, Glenns V, Shadowitz S, et al. A bedside clinical prediction rule for detecting moderate or severe aortic stenosis. J Gen Intern Med 1998;13:699-704. https://link.springer.com/article/10.1046/j.1525-1497.1998.00207.x
- Etchells EE, Bell C. Robb KV. Does this patient have an abnormal systolic murmur? JAMA 1997;277:564-71. https://www.ncbi.nlm.nih.gov/pubmed/10376577
- Das P, Pocock C, Chambers J. The patient with a systolic murmur: severe aortic stenosis may be missed during cardiovascular examination. Q J Med 2000;93:685-8. https://oup.silverchair-cdn.com/oup/backfile/Content_public/Journal/qjmed/93/10/10.1093_qjmed_93.10.685/1/930685.pdf?Expires=1500852139&Signature=TwyO6Z4fUfbPc1yiA~2xZC7jOjed0juH604DshdvRYy~VqeNQ57Sv1yE-LNsImthgQogkawMruBPdXn6PvVCVmdvXxE9QsMzQYhZ13JqXDTQhRiPBcsIBKDdROr~xbz0gp0nv-zEmjCp1M8-CXjrlVnjVtwJ6q2nIPTRW5h-CUOnDAmf8vCeJHRi2M9Dt3a4vGALDJQPaETvxKDfoADamBDtZHzzoCIH3OyXT3–jHRtv9AJI2uHlzN79Vzkh~oIrR-rI5mkHle3Yz0R3qIBY0l4P3PssMng~v-IXMNKS~Ghjav8YFTigHN23aEA5yUYllsC7hR25L6h9PA0SZP3QA__&Key-Pair-Id=APKAIUCZBIA4LVPAVW3Q
- Aronow WS, Kronzon I. Prevalence and severity of valvular aortic stenosis determined by Doppler echocardiography and its association with echocardiographic and electrocardiographic left ventricular hypertrophy and physical signs of aortic stenosis in elderly patients. Am J Cardiol 1991;67:776-7. https://www.ncbi.nlm.nih.gov/pubmed/1826070
Several lines of evidence suggests that even in the absence of atrial fibrillation, patients with aortic stenosis (AS) may be at higher risk of a cerebrovascular accident (CVA).
A population-based study from Mayo Clinic examining the risk of cerebrovascular events (primarliy ischemic strokes or transient ischemic attacks) among patients with valvular heart disease reported severe aortic stenosis (study defined as mean pressure gradient >30 mm Hg) as a predictor of cerebrovascular event, independent from atrial fibrillation or age and at rates similar to those of mitral stenosis 1. Similarly, the simvastatin and ezetimibe in AS study involving patients with mild-to-moderate AS not prescribed oral anticoagulation found that CHA2DS2-VASc was a major predictor of stroke, independent of atrial fibrillation or aortic valve replacement2. Thromboembolic CVA has also been reported in the setting of calcified bicuspid aortic valve with moderate-severe AS3.
Potential factors increasing the risk of CVA in AS include calcium embolization from the valve and increased microthrombus formation on the valve which may be present in 53% of stenotic aortic valves3,4. In addition, severe AS may be a marker for other conditions that increase risk of CVA, such as atherosclerosis or prothrombotic tendencies1.
- Petty GW, Khandheria BK, Whisant JP. Predictors of cerebrovascular event and death among patients with valvular heart disease: A population-based study. Stroke 2000;31:2628-2635. http://stroke.ahajournals.org/content/strokeaha/31/11/2628.full.pdf
- Greve AM, Dalsgaard M, Bang CN, et al. Stroke in patients with aortic stenosis: The simvastatin and ezetimibe in aortic stenosis study. Stroke 2014;45:1939-1946. http://stroke.ahajournals.org/content/strokeaha/45/7/1939.full.pdf
- Mahajan N, Khetarpal V, Alfonso L. Stroke secondary to calcific bicuspid aortic valve: Case report and literature review. J Cardiol 2009;54:158-61. https://www.ncbi.nlm.nih.gov/pubmed/19632538
- Pleet AB, Massey EW, Vengrow ME. TIA, stroke, and the bicuspid aortic valve. Neurology 1981;31:1540-2. https://www.ncbi.nlm.nih.gov/pubmed/7198207
Cirrhotic patients with upper gastrointestinal bleed (UGIB) are at high risk of bacterial infections: 22% during the first 48 h after admission, 35-66% within 2 weeks of initial bleeding1. Antibiotic prophylaxis has been shown to reduce short term mortality, bacterial infections, early rebleeding and volume of blood transfused1-4.
But what is the exact connection between UGIB and bacterial infections in cirrhosis? One hypothesis is that UGIB sets up the host for bacterial infection via translocation (eg, due to hypovolemia), procedures necessary in the management of bleeding (eg endoscopy, sclerotherapy, IV access), and aspiration pneumonia. More intriguing is the reverse hypothesis—that is the bacterial infection serves as a trigger for UGIB. Several lines of evidence support this view1,2.
- Cirrhotic patients admitted for non-UGIB-related conditions may be 4x more likely to develop UGIB during their hospitalization in the presence of bacterial infection on admission4
- Infections predispose to early variceal rebleeding
- Infection/endotoxemia increase portal pressure, and impair liver function and coagulation
- Commonly cited risk factors for variceal bleeding (eg, hepatic venous pressure gradient, liver function, size of varices) do not readily explain why bleeding occurs unpredictably and why despite daily increases in portal pressure (eg, following daily meals and exercises), UGIB is relatively infrequent.
- Thalheimer U, Triantos CK, Samonakis DN, et al. Infection, coagulation, and variceal bleeding in cirrhosis. Gut 2005;54:556-63. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1774431
- Goulis J. Bacterial infection in the pathogenesis of variceal bleeding. Is there any role for antibiotic prophylaxis in the cirrhotic patient. Ann Gastroenterol 2001;14:205-11. http://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=4&ved=0ahUKEwjNh-rhlpLVAhXGdD4KHSurANcQFgg4MAM&url=http%3A%2F%2Fwww.annalsgastro.gr%2Findex.php%2Fannalsgastro%2Farticle%2Fdownload%2F80%2F71&usg=AFQjCNHJfAyYAjuNXpwsWGrVuyuxxgJYKg
- Soares-Weiser K, Brezis, Tur-Kaspa R, et al. Antibiotic prophylaxis of bacterial infections in cirrhotic inpatients: a meta-analysis of randomized controlled trials. Scand J Gastroenterol 2003;38:193-200. http://www.tandfonline.com/doi/abs/10.1080/00365520310000690
- Anastasioua J, Williams R. When to use antibiotics in the cirrhotic patient? The evidence base. Ann Gastroenterol. 2013; 26(2): 128–131. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3959942
- Benavides J, Fernandez N, Colombato L, et al. Further evidence linking bacterial infection and upper G.I. bleeding in cirrhosis. Results from a large multicentric prospective survey in Argentina. J Hepatol 2003;38 (suppl 2):A176. http://www.journal-of-hepatology.eu/article/S0168-8278(03)80592-5/abstract
Don’t forget to sign up under menu to get future pearls right into your mailbox!
Mitral regurgitation (MR) murmur can radiate to several places on the chest wall as well as the spine and….ready for this…. the top of the head!
Classically, MR is thought to have 4 patterns of radiation1,2.
- Axilla and the inferior angle of the left scapula (typical)
- Left sternal border, base of the heart and into the neck
- Cervical and lumbar spine (down to sacrum)
- Right of the sternum (associated with a “giant left atrium”)
Less well-known and perhaps most intriguing is the radiation of MR to the top of the head. Original reports involved patients who often had ruptured chordae tendineae due to subacute bacterial endocarditis and/or rheumatic heart disease2.
It was posited that “the flail portion of the mitral valve folds back into the left atrial cavity forming a hood which deflects the regurgitant stream against the atrial wall”. In the setting of a flail anterior leaflet, if the jet stream is sufficiently high energy and comes in contact with the spine, the murmur may be transmitted by bone conduction to the top of the skull2!
I suggest you explain to your patient what you are doing before you auscultate the top of their heads!
Liked this post? Sign up under MENU and catch future pearls right into your mailbox!
- Chatterjee K. Physical examination. In Topol EJ, ed. Textbook of cardiovascular medicine, 2007, pp 193-224. Lippincott, Williams &Wilkins. Philadelphia. https://books.google.com/books?id=35zSLWyEWbcC&pg=PA219&lpg=PA219&dq=top+of+the+head+mitral+regurgitation+murmur&source=bl&ots=56Erim4eNM&sig=82TrOiU52ojmhVBMG7G2jMULxVo&hl=en&sa=X&ved=0ahUKEwit8_WnmorVAhVGyj4KHcwUC_8Q6AEIRzAF#v=onepage&q=top%20of%20the%20head%20mitral%20regurgitation%20murmur&f=false
- Merendino KA, Hessel EA. The “murmur on top of the head” in acquired mitral insufficiency: Pathological and clinical significance. JAMA 1967;199:892-896. http://jamanetwork.com/journals/jama/article-abstract/663746
Not extremely well! Although a recent multicenter prospective study in adult hospitalized patients reported that the median procalcitonin (PCT) concentration was significantly lower for community-acquired pneumonia (CAP) caused by viral pathogens ( 0.09 u/ml vs atypical bacteria [0.2 ug/ml] and typical bacteria [2.5 ug/ml]), PCT was <0.1 ug/ml and <0.25 ug/ml in 12.4% and 23.1% of typical bacterial cases, respectively1.
This means that we could potentially miss about a quarter of CAP cases due to typical bacterial causes if we use the <0.25 ug/ml threshold (<0.20 is ug/ml has been used to exclude sepsis2). For these reasons and based on the results from another study3, no threshold for PCT can reliably distinguish bacterial from viral etiologies of CAP4. Clinical context is essential in interpreting PCT levels! Also go to a related pearl on this site5.
Can PCT distinguish Legionella from other atypical bacterial causes of CAP (eg, caused by Mycoplasma or Chlamydophila)? The answer is “maybe”! Legionella was associated with higher PCT levels compared to Mycoplasma and Chlamydophila in one study1, but not in another3.
- Self WH, Balk RA, Grijalva CG, et al. Procalcitonin as a marker of etiology in adults hospitalized with community-acquired pneumonia. Clin Infect Dis 2017;65:183-90. https://www.ncbi.nlm.nih.gov/pubmed/28407054
- Meisner M. Update on procalcitonin measurements. Ann Lab Med 2014;34:263-73.
- Krüger S, Ewig S, Papassotiriou J, et al. Inflammatory parameters predict etiologic patterns but do not allow for individual prediction of etiology in patients with CAP-Results from the German competence network CAPNETZ. Resp Res 2009;10:65. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2714042/pdf/1465-9921-10-65.pdf
- Bergin SP, Tsalik EL. Procalcitonin: the right answer but to which question? Clin Infect Dis 2017; 65:191-93. https://academic.oup.com/cid/article-abstract/65/2/191/3605416/Procalcitonin-The-Right-Answer-but-to-Which?redirectedFrom=fulltext
Two things to ask before you order procalcitonin (PCT): 1. Will it impact patient management?; and 2. If so, will the result be available in a timely manner ie, within hours not days?
Whatever the result, PCT should always be interpreted in the context of the patient’s illness and other objective data. Not surprisingly then, as a “screening” test, PCT may be more useful in patients with low pre-test likelihood of having bacterial infection, not dissimilar to the use of D-dimer in patients with low pre-test probability of pulmonary embolism1.
Several potential clinical uses of this biomarker have emerged in recent years, including:1,2
- Helping decide when to initiate antibiotics in patients with upper acute respiratory tract infections and bronchitis. A normal or low PCT supports viral infection.
- Helping decide when to discontinue antibiotics (ie, when PCT normalizes) in community-acquired or ventilator-associated pneumonia.
- Helping monitor patient progress with an expected drop in PCT of about 50% per day (half-life ~ 24 hrs) with effective therapy.
- PCT may be unremarkable in about a third of patients with bacteremia (especially due to less virulent bacteria, including many gram-positives)3.
- PCT levels are lowered by high-flux membrane hemodialysis, so check a baseline level before, not after, hemodialysis4.
- Lastly, despite its higher specificity for bacterial infections compared to other biomarkers such as C-reactive protein, PCT may be elevated in a variety of non-infectious conditions, including pancreatitis, burns, pulmonary edema or aspiration, mesenteric infarction (ischemic bowel), cardiogenic shock, and hypotension during surgery2.
- Schuetz P, Muller B, Chirst-Crain M, et al. Procalcitonin to initiate or discontinue antibiotics in acute respiratory tract infections (review). Evid-Based Child Health (A Cochrane Review Journal) 2013;8:4;1297-137. http://onlinelibrary.wiley.com/doi/10.1002/ebch.1927/pdf
- Gilbert GN. Use of plasma procalcitonin levels as an adjunct to clinical microbiology. J Clin Microbiol 2010;48:2325-29. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2897488/pdf/0655-10.pdf
- Yan ST, Sun LC, Jia HB. Procalcitonin levels in bloodstream infections caused by different sources and species of bacteria. Am J Emerg Med 2017;35:779-83. https://www.ncbi.nlm.nih.gov/m/pubmed/27979420/#fft
- Grace E, Turner RM. Use of procalcitonin in patients with various degrees of chronic kidney disease including renal replacement therapy. Clin Infect Dis 2014;59:1761-7. https://www.ncbi.nlm.nih.gov/pubmed/25228701