Tuesday, December 20, 2011

Aortic Stenosis - How Bad?

Motivation: If you listen closely enough, you will hear a murmur.  That has been my feeling about listening to murmurs over the aortic valve.  Almost everyone has a "soft ejection murmur."  The question, of course, is how tight is the aortic valve.  We have all heard cautionary tales about the lack of correlation of murmur intensity with aortic valve area.  During morning rounds, an attending suggested using the physical exam sign of "brachioradial delay" as a measure of severity of aortic stenosis.  How good is the sign?

For background, brachioradial delay is assessed by gently palpating the brachial and radial pulses simultaneously.  If there is a palpable delay between the brachial and radial pulses, then "brachioradial delay" is present.  From a pathophysiology viewpoint, the quicker the upstroke of the arterial pulse the quicker the pulse pressure wavefront spreads through the arterial tree.  In aortic stenosis, the upstroke of the arterial pulse is slower allowing some of the energy to be absorbed by circumferential stretch of the arteries.  Consequently, with a slow upstroke, there is a palpable delay in the spread of the pulse upstroke between proximal and distal arteries.

Paper: Leach, R.M. and McBrien, D.J. "Brachioradial delay: a new clinical indicator of the severity of aortic stenosis." Lancet (1990); 335: 1199-201.

Methods: Patients presenting at echocardiography clinic at Worthington Hospital (U.K.) were prospectively assessed by four clinicians about the presence of brachioradial delay.  Subjects were not preselected by any diagnosis.  The sign was counted positive if all four clinicians agreed.  Subsequently, for all the patients, the delay between the brachial and radial pulses was timed, and echocardiography was done in all patients.  In the paper, severe aortic stenosis was defined as aortic valve area of 0.5-0.75 sqcm.

Results:
Patient selection: Among the patients assessed, there were 33 patients with aortic stenosis and 25 patients with other heart diseases (9 heart failure, 4 MR, 3 post MI, 3 HOCM, 2 aortic regurg, 2 afib, 1 bicuspid aortic valve, 1 heart transplant).  Of the 33 patients with aortic stenosis, 17 had severe aortic stenosis (valve area between 0.5-0.75 sqcm).  The authors also reported on 27 controls without heart disease who were age and sex matched.

Control Subjects: For none of the 27 control subjects was a brachioradial delay palpable.  On measurement by pressure transducers,  the mean brachioradial delay was 24.3 msec below age 55 and 21.7 msec for those over age 75.  None of these delays were palpable.

Aortic Stenosis: In the 17 patients with "severe" aortic stenosis (valve area 0.5-0.75 sqcm), the brachioradial delay was palpable in all patients.  For 16 patients with mild aortic stenosis (valve area > 0.75 sqcm), brachioradial delay was palpable in only 6 out of 16 patients.

Other Heart Conditions: In 25 patients with other heart conditions, brachioradial delay was palpable in one patient with hypertrophic obstructive cardiomyopathy.  In no other patient was brachioradial delay palpable.

Discussion: This paper is remarkable in many ways.  I was surprised to see a novel physical exam sign for aortic stenosis described in the relatively recent past (1990).  Also surprising is the 100% sensitivity of this sign for aortic stenosis with valve area <0.75 sqcm.  The physical exam sign is also good because to determine a positive sign, you do not need to have many years of experience examining normal subjects.   The sign is a relative palpable delay between brachial and radial arteries.  Of note, the sensitivity of the sign falls off rapidly as the valve area increases above 0.75 sqcm (only 6/16 positive).

The paper, however, also has many weaknesses.  First of all, there are relatively few subjects examined.  Secondly, the specificity of the sign is very poorly established.  The group with "other" heart diseases was heterogeneous with few representatives of each type of cardiac pathology.  Finally, of note, what is called "severe" aortic stenosis (valve area <0.75 sqcm) would likely be termed critical aortic stenosis now (at least in the U.S.) while the mild aortic stenosis in the paper would be severe aortic stenosis at present (valve area 0.75 to 1 sqcm).  Despite these weakness, I think that the brachioradial delay is a good test to assess for critical AS in a patient with classic AS murmur.

Saturday, December 10, 2011

DVT Prophylaxis - Surprising Truths

Motivation: Frankly, DVT prophylaxis seems like a lot of voodoo to me.  Most patients who become hospitalized have probably been sick in bed at home for some days, and many continue to be bed-bound after discharge.  During the few days of hospitalization, I often wonder what difference is made by daily stabs in the stomach with heparin.  Yet, there is increasing institutional push to remember to use heparin prophylaxis.  I was browsing for some evidence for DVT prophylaxis when I came across American College of Physician systematic review about this very topic published this year.  The results are surprising.

Paper: Qaseem, A., Chou, R., Humphrey, L., et. al. "Venous Thromboembolism Prophylaxis in Hospitalized Patients: A Clinical Practice Guideline from the American College of Physicians."  Ann. Intern Med. (2011) 155: 625-632.

Methods: Systematic review of published randomized trials from 1950 to 2011.  Primary outcome was total mortality 120 days after randomization.  Secondary outcomes were symptomatic DVT, all PE, fatal PE, all bleeding, and major bleeding.  The review separated analysis for patients with and without acute stroke.

Results:
Effect of heparin prophylaxis versus no prophylaxis:
Medical patients without stroke - The review found ten trials (total of 20,717 patients) of patients without stroke.  There was NO significant effect on mortality at 120 days (RR: 0.94, CI: 0.84-1.04).  Heparin was associated with reduced risk for PE (RR: 0.69, CI: 0.52-0.90) but also increased risk of any bleeding (RR: 1.34, CI: 1.08-1.66). The differences in major bleeding and symptomatic DVT were not significant. In summary, heparin use prevents 4 PE per 1000 treated but causes 9 events of any bleeding per 1000 treated.
Acute Stroke - Review found 8 trials (total of 15,405 patients).  Pooled results showed NO significant reduction in mortality, PE, or symptomatic DVT. Prophylaxis was associated with increased risk for major bleeding (RR: 1.66, CI: 1.20-2.28). Pooled data was, however, pretty heterogenous in findings (wide spread in data).  The largest randomized trial with acute stroke had 14,578 patients and found NO reduction in mortality or PE.  However, a significant reduction in recurrent ischemic stroke was detected (RR: 0.65, CI: 0.54-0.80) at the risk of increased risk of hemorrhagic stroke or serious extracranial hemorrhage (RR: 1.73, CI: 1.22-2.46).

Low-Molecular Weight Heparin versus Unfractionated Heparin:
Medical patients without acute stroke: NO statistical difference in mortality, PE, or major bleeding events.
Acute Stroke: NO statistical difference in mortality, PE, or bleeding events.

Compression Stockings versus no stockings:
Sparse data (three trials, 2518 patients) making separation of general medical patients from patients with acute stroke difficult.  Overall, compression stockings did NOT reduce mortality, symptomatic DVT, or PE.  Risk for lower extremity skin damage was significantly increased among patients wearing stockings (RR: 4.02, CI: 2.34-6.91) conferring risk of 39 events per 1000 treated.

Discussion: To me, the most surpirsing part of the review was that despite the large pooled cohort size (>20,000 patients), there was no mortality benefit with heparin prophylaxis four months post-randomization.  The reduction in PE risk by about 30% presumably involves decreased risk of small PE.  As expected, the cost of reducing PE is increased bleeding events.  Amazingly, acute stroke patients did not have similar PE risk reduction despite a large cohort size.  Heparin use in acute stroke is a different balance between reduction in recurrent stroke (reduction by about 35%) compared to increased risk of major bleeds, including intracranial bleeds.  The review also helped clarify a couple of other misconceptions that I had.  Overall, low molecular weight heparin is no better than unfractionated heparin other than once a day dosing versus three times daily dosing.  I had thought that LMWH was better than unfractionated heparin.  Also, compression stockings proved to be pretty useless.  There is no demonstrable benefit while incurring increased risk of lower extremity skin damage.  On balance, there appears to be increased harm.  The review did not cover pneumatic compression devices.

One caveat to keep in mind in these large systematic reviews is that by pooling data, a lot of granular differences among subgroups are lost.  For example, ICU patients may have different risks and benefits.  There are also number of other cohorts such as patients with cirrhosis or kidney disease who may benefit variably with heparin.  When evaluating 30,000 patients, only effects that are consistent across entire groups are apparent.  One way in which this paper changes my perspective on DVT prophylaxis is that I will be less hesitant to hold prophylaxis in patients at risk for bleeding (after all, not saving lives with heparin).  And, I will stop pushing for stockings - good for Christmas but probably not for preventing PE.

Saturday, December 3, 2011

Valves of the Right Stuff

Motivation: Morning rounds with a cardiologist is like going from room to room and trying to be a human echo machines.  Is there a valvular problem? The safe answer is almost always yes.  You may not be able to hear it, but the attending can.  Although human beings have refined their ears to distinguish harsh, blowing, and soft transvalvular flow sounds, the ability to fix valvular disease has lagged behind.  Artifical valves have major problems.  Mechanical valves require life long anti-coagulation, and dead tissue valves have finite lifespan.  A surgeon recently showed me the third way: growing valves out of your own tissue.  Sounds like fiction, but the patient did just fine after the tricuspid valve repair operation.  Here is a case series from Italy that demonstrates the magic.

Paper: Quarti, A. et. al. "Preliminary experience in the use of an extracellular matrix to repair congenital heart diseases." Interactive Cardiovascular and Thoracic Surgery (2011article in press.

Methods: Between 2009 and 2011, 26 patients with congenital heart defect underwent surgery using an extracellular matrix patch (manufactured by CorMatrix).  The matrix was constructed from porcine small intestinal submucosa with expectation that native heart tissue would grow over the matrix.  The immediate post-operative course was reported.

Results:
Patients: Surgery was performed in 26 patients (mean age 6.4 years, range of 8 days to 32 years):
  • 10 for pulmonary patch arterioplasty
  • 9 for valve leaflet repair (including tricuspid, pulmonary, aortic, and mitral valves)
  • 4 for ascending aortic patch aortoplasty
  • 3 for aortic arch reconstruction
  • 1 for right ventricular outflow tract obstruction
Results: There were no deaths in the peri-operative period.  At follow-up (mean 13.2 months, range 4-25 months), there was no evidence of thrombosis, disruption, shrinkage, leakage, or patch calcification on echo and CXR. 

For the patients with valve repair, follow-up echo only showed trivial to mild regurgitation.  No valve repair required reoperation.  Intial followup of the valve repair cohort (mean of 12.5 months) did not show evidence of progression of regurgitation. 

For cases in which extracellular matrix was used as vascular patch (18 cases), two had pulmonary artery stenosis distal to the patch.  One required operative repair.

Discussion: This paper may be part of the intial foundations of a marked change in valve repair surgery.  If the results eventually hold up in larger study, many of the long-term complications of valve repair surgery such as anti-coagulation, thrombogenic risk, and risk of endocarditis will be significantly decreased.  At present, however, this paper probably generates more questions than answers.  First of all, the follow-up is short, and questions about long-term viability of the ECM repair remain.  Also, most of the subjects were children, and older adults may repopulate the matrix at a different rate than children making extension of the results to adults problematic.  This paper probably lays the groundwork for larger randomized trials.  Finally, the paper does not mention the funding source.  In subsequent trials, it would be preferable if CorMatrix is not part of the trials.

Sunday, November 20, 2011

Regrowing the Heart

Motivation: With the recent AHA conference, a flurry of new findings has been released.  Although there have been many negative findings, there was one recent trial in particular that caught my attention.  Wouldn't it be nice if we could somehow regrow heart cells in patients with heart failure? People have tried it in the past and failed.  But, there was a recent report of success!

Paper:   Bolli, R. et. al. "Cardiac stem cells in patients with ischaemic cardiomyopathy (SCIPIO): initial results of a randomised phase 1 trial" Lancet (2011) epub.

Methods: The full trial consisted of an early safety analysis and then a randomized portion.  Given low recruitment so far, the reported results combine the results of the whole experience.  The trial included patients of (1) less than 75 who were (2) undergoing CABG with (3) LVEF <= 40% and (4) history of previous MI.  Initial screening occurred at time of CABG during which the right atrial appendage was harvested and cardiac stem cells bearing cell surface marker c-kit were isolated.  At a mean of 113 days after CABG, the cardiac stem cells were infused into the infarcted territory using a balloon catheter.

Results:
Subjects: In total, 16 patients received stem cells while there were seven patients in the control group.  The two groups were overall similar in terms of age, gender, and comorbidities though the number of patients in each group is obviously very small.

Treatment Effect: At four month follow-up, the left ventricular ejection fraction (LVEF) increased from mean of 30.3% to 38.5% (p = 0.001).  By contrast, in the controls, the LVEF remained unchanged from mean of 30.1% to 30.2%.  The authors had injected the stem cells into the infarcted territory of previous MI.  In the infused territory, the regional wall motion score improved significantly (1.97 to 1.78, p = 0.007), with lower scores indicating less wall motion defect.  The regional wall motion score did not change in the control group.

Cardiac MRI: Seven patients treated with cardiac stem cells underwent cardiac MRI to assess mean infarct weight.  In the seven patients, the mean infarct weight decreased by 7.8 g (standard error, 1.7 g) at four months.  A reduction in infarct size was also noted.

Functional Data:  In the stem cell treated patient, the NYHA functional class decreased from mean of 2.19 to 1.63 four months after infusion.  The NYHA class of control patients stayed identical. 

Adverse effects: No serious adverse effects occurred more frequently in the treatment group including MI, arrhythmia, new tumor, or stroke.

Discussion: The trial is pretty exciting because it shows that a one-time procedure of stem cell infusion can generate long-lasting benefits including ejection fraction increase and functional benefit in terms NYHA heart failure severity.  This trial, however, must be taken in the context of a proof-of-concept safety analysis trial.  The number of patients in each group is pretty small.  Also, the entire trial was not randomized, and there may have been non-obvious differences in the two cohorts.  Despite media hype about this trial, both of these are strong limitations in extending stem cell infusions to routine practice just now.  Nonetheless, this trial provides hope that rather than just thinking about slowing down the pace of heart failure progression, we may actually in the future think about reversing heart failure.  Perhaps, in the future, standard therapy for STEMI will be PCI with stem cell infusion.  There is much more to come from this field!

Sunday, November 6, 2011

Parkinson's Disease Prevention

Motivation: Lately, I have been meeting too many people with Parkinson's Disease - one of those random streaks of fate.  Controllable at first, the disease is pretty disabling as it progresses.  I was wondering about causes of this "idiopathic" disease and whether there are protective factors like say eating a salmon each day.  During this search, I came across some rumors that calcium channel blockers may be protective against Parkinson's Disease.  And, the data?

Paper: Ritz, B. et. al. "L-Type Calcium Channel Blockers and Parkinson Disease in Denmark." Ann. Neurol. (2010) 67: 600-606.

Methods: A case-control study in Denmark using the Denmark National Health Service registry that covers medical service for all citizens.  Authors identified initial diagnoses of Parkinson's Disease (PD) between 2001-2006.  Cases were identified either by first coded diagnosis of PD or first prescription of PD medication.  Controls were chosen after matching for birth year and sex.  Patients with any type of dementia or cerebrovascular disease before diagnosis of PD were excluded (along with matched controls).  Also excluded were patients with PD who were not prescribed any PD drugs.

Results:
Subjects: In total, the authors identified 1931 cases of PD with 9651 matched controls.  Five years prior to the index date of PD, the general health of cases and matched controls as measured by the Charson index was similar. 

Non-dihydropyridine Calcium Channel Blockers: Authors looked at prescription drug use two years prior to diagnosis of PD to account for the pre-clinical phase and risk factors for developing PD.  For non-dihydropyridine calcium channel blockers (verapamil and diltiazem), no association with PD was found. 

Dihydropyridine CCB: For dihydropyridine calcium channel blockers (such as amlodipine or felodipine), authors separated analysis for amlodipine, which has low central nervous system penetrance, and other members of this class, which have higher CNS availability.  For amlodipine, no protective effects were found.  For other members of dihydropyridine CCB, use was associated with protective effects: odds ratio of 0.70, adjusted confidence interval (0.52-0.94).  Of note, there were 55 patients with PD and 368 controls using medications of this class.

Discussion: The paper shows that use of non-hydropyridine calcium channel blockers (excepting amlodipine) was associated with about 30% decreased risk of PD diagnosis.  While this paper is a case-control study, the results have some biological plausability.  Non-CNS penetrant drugs did not show any effect while CNS penetrant calcium channel blockers had some protective effects.  It is plausible that blocking calcium influx into oxidatively stressed cells may provide some degree of protection.  On the other hand, the association observed may be indirect indicators of some other condition that is protective.  In the study, not much information was provided about details of comorbidities and how comorbidities were different between cases and controls.  This study is suggestive but far from definitive in using calcium channel blockers in the clinic.  As usual, more rigorous study is needed.

Monday, October 31, 2011

Prolonged QTc and Torsades de Pointes

Motivation: In the past few months, the most common EKG abnormality I have observed has been "nonspecific" ST-T wave changes.  Second to that has been mildly prolonged QT interval.  I often don't quite know what to do with a QT interval of 470 msec.  Does this mean that the patient does not get ondansetron? Or, should this QT interval lead to the reflex of K>4 and Mg>2?  The feared consequence of prolonged QTc is Torsades de Pointes.  What is the association between the degree of QTc prolongation and Torsades de Pointes?  The issue is complicated further by the fact that there is no good consensus on the definition of "normal." Conventionally, upper limit of normal for QTc is 450 msec for men and 470 msec for women.

This question turned out to be a lot harder to answer than anticipated.  There are not great prospective studies, but there have been reviews which have compiled cases of Torsades de Pointes and associated QTc intervals.

Paper:  Bednar, M.M. et. al. "The QT Interval" Progress in Cardiovascular Diseases. 43 (2001): 1-45 (supplement)

Methods: Data collected from 202 reports of Torsades de Pointes and prolonged QT interval.  Corrected QT interval was by Bazett formula.

Results:

QTc (ms)               TdP Cases (% of total cases)
<500                     9  (7.8)
500-549               13 (11.2)
550-599               24 (20.7)
600-649               36 (31.0)
650-699               21 (18.1)
>700                    13 (11.2)

Discussion: The results illustrate that while Torsades de Pointes from prolonged QTc is uncommon with QTc less than 500, the risk is not zero with mildly prolonged QTc (<500 msec).  The most commonly associated QTc interval with TdP is between 600-649 msec.  The patients who experienced QTc in the lowest QTc interval (<500 msec) may have had some inherited propensity for ventricular ectopy that was enhanced with mildly prolonged QTc.  Another explanation could be that the Bazett formula probably did not adequately correct the QT interval.  The Bazett formula overcorrects at faster heart rates and undercorrects at low heart rates.  Interestingly, Bazett proposed the formula after only examining 39 healthy patients.  An alternative way to calculate QTc is using an empirically derived formula from the Framingham study which used a much larger number of subjects (Pubmed ID: 1519533).  Anyway, in the future, I will think a little bit about risk of TdP in patients with QTc<500 and a lot more when QTc>500.

Wednesday, October 19, 2011

Eating during Times of Obstruction

Motivation: I used to believe in the maxim, "For SBO, keep NPO" meaning that for patient with small bowel obstruction, eating was forbidden. In fact, for decompression, the stomach should be suctioned out with an NG tube. Last month, while caring for patients with bowel obstruction, I came across papers that challenged this hegemony. They suggested that laxatives might even be beneficial in some cases of partial SBO. I asked around. Nobody uses laxatives in SBO. But, should we be changing out ideas?

Paper: Chen, S-C. et. al. "Specific oral medications decrease the need for surgery in adhesive partial small-bowel obstruction." Surgery (2006) 139: 312-316.

Methods: A randomized controlled trial in Taiwan comparing standard vs. novel treatment in patients with partial adhesive small bowel obstruction, defined by (1) history of intra-abdominal operation, (2) clinical signs and symptoms of SBO, and (3) passage of contrast to colon within 24 hours of administration. Standard treatment consisted of IV hydration, NG tube decompression, and NPO. Novel treatment was IV hydration, NG tube decompression, and oral solution containing magnesium oxide (laxative), Lactobacillus acidophilus (digestant), and simethicone (defoaming agent). The primary outcome tracked was success of non-operative management.

Results:
Patients: Total of 236 patients were randomized. Both groups were similar in terms of age, gender, and presenting symptoms (abdominal pain, distension, constipation, vomiting).

Comparison of treatments: Non-operative management success rates were less with standard approach (77%) compared to treatment with oral therapy (90%, p<0.01). In other words, more patients kept NPO required surgery. The complication and recurrence rates were not different between the two treatment arms.

Discussion: This randomized study challenges the traditional assumption that bowel rest is the best treatment for any type of small bowel obstruction. One important flaw in the study is that while the attending surgeon was blinded to allocation, the rest of the staff was not blinded. This may have introduced some bias into the decision making process. Otherwise, the study contradicts the main fear that giving PO during obstruction leads to excess complications. In fact, giving the oral regimen significantly improved chances of non-operative management success. Importantly, this study only examined partial SBO from adhesions (the most common cause of SBO). It is unclear whether diseases like Crohn's have different benefits from bowel rest. The next time I see partial SBO from adhesions, I will try to convince my attending to give this regimen a try!