General Medicine Update
Minnesota ACP
November 7, 2008
Steve Hillson
Hennepin County Medical Center
University of Minnesota
Objectives
• At the end of this session you should be able to:– Describe the main results of several
important reports from the past year– Decide how you want to change your
practice in the context of these findings
Disclosure
• I have no direct financial relationships with any commercial firm having any interest in any of the reports or topics I am about to discuss.
Process• Personally reviewed title of every original research
article from 10/01/07 till 10/22/08 in:– Annals of Internal Medicine– BMJ– JAMA– Lancet– New England Journal of Medicine
• Reviewed subspecialty updates, scattered other sources
• Personally reviewed abstract of every article with “interesting” title.
Process (cont’d)
• Selected “promising” articles by initial abstract review (about 100)
• Re-reviewed all abstracts, selecting about 60 with medium or high impact potential
• Solicited abstract reviews from colleagues to select subset of greatest importance
• Critically appraised final subset for presentation
Limitations on Process
• Personal idiosyncrasies• Incomplete survey of medical literature• No claim to comprehensive context for
assessing these articles• Very simplified presentation of complex
research• Final slide set available at
– www.paralleltext.net/ppt.html
In Pursuit of the Perfect A1C
• How intensely should we be controlling type 2 diabetes?
• 3 Important Articles– ACCORD, NEJM, June 2008
• Funded by NIH, CDC, with drugs contributed by many makers
– ADVANCE, NEJM, June 2008• Funded by maker of gliclazide
– UKPDS, NEJM, October 2008• Funded initially by UK government agencies, this follow-
up funded by drug makers
Purpose
• Assess tighter vs looser glycemic control in type 2 diabetes
• Previously limited information– None showing mortality or macrovascular
benefit in type 2 DM
• But extensive promulgation of the idea that lower is better
#1 - ACCORD
• Compare target A1C <6.0 to less tight (7-7.9) for cardiovascular outcomes
• Clinical Trial, unblinded– 10,000 US/Canadian patients with DM-2,
A1C≥7.5, and CV disease or risk factor– Any standard diabetes medications– More frequent visits and medication adjustments
for intensive therapy group– Followed 3.5 years for CV death, MI, CVA
#2 - ADVANCE
• Compare target A1C (<6.5) to less tight (local guideline) for vascular outcomes
• Clinical Trial, unblinded• 11,000 patients worldwide, type 2 diabetes,
age≥55, no insulin, and pre-existing vascular disease or a risk factor
• Gliclazide, plus frequent clinic visits and other drugs as needed, OR
• Usual care, with gliclazide excluded• Followed 5 years for vascular events
#3 - UKPDS 10-year follow-up
– Compare tight glycemic control (fasting glucose 108), to less tight (fasting glucose < 270) for macro- and microvascular outcomes
– Clinical Trial, unblinded– 4000 UK patients with new DM-2, age 25-65– Received one of several drug-based strategies OR– “Usual Care” with diet alone unless FPG>270– Treated 10 years, then followed additional 10
years on community standard care, for vascular outcomes
Findings - Achieved A1C
0
1
2
3
4
5
6
7
8
ACCORD ADVANCE UKPDS*
IntensiveStandard
Findings - Primary Outcomes
0
10
20
30
40
50
60
ACCORD ADVANCE UKPDS - All UKPDS -Metformin
IntensiveStandard
*
* *
Findings - Death
0
5
10
15
20
25
30
35
ACCORD ADVANCE UKPDS All UKPDSMetformin
IntensiveStandard
*
* *
Limitations• ACCORD used a lot of rosiglitazone
• Neither ACCORD nor ADVANCE achieved target A1C on most patients
• UKPDS “usual care” isn’t
Implications
• Target A1C of 6.5 or less is at best ambiguous for macrovascular disease, possibly dangerous– May depend on drug choice– Death (NNH of 100) trumps improved
nephropathy/retinopathy (NNT of 70)
• Metformin, without a tight target A1C, is useful for survival in obese diabetics (NNT about 15 over 20 years)
• I will not seek extremely tight A1C• I will use still more metformin
Preventing the Clot
• There’s a new perioperative anticoagulant on the block - 2 studies– RECORD1, NEJM, June 2008– RECORD3, NEJM, June 2008
Purpose
• Compare rivaroxaban to enoxaparin for preventing post-op VTE– Total Hip Arthroplasty (RECORD 1)– Total Knee Arthroplasty (RECORD 3)
• Funded by makers of rivaroxaban– Orally administered, fixed dose factor Xa inhibitor– Reportedly out in January
• Related drugs– Argatroban - parenteral– Ximelagatran - oral, withdrawn due to liver toxicity– Dabigatran - oral, possibly out in 2010
Method• Clinical trials, blinded• 2500 (knee) and 4400 (hip) patients, age≥ 18
with no hepatic or renal disease• Given rivaroxaban 10 mg orally each day, OR• Enoxaparin 40 mg SC each day
– KNEE study: 10-14 total days– HIP study: 35 total days
• Followed 2-6 weeks for venographic DVT and symptomatic VTE or death
Findings - Detectable Venous Thromboembolism
02468
1012
14161820
Hip Knee Bleeding
RivaroxabanEnoxaparin
Limitations
• Symptomatic VTE was rare (about one-tenth of all VTE events)
• Industry-funded research has many opportunities to mislead
• Issue of spinal catheter management not clarified
Implications
• I’m usually a turtle, but…– I will start using perioperative rivaroxaban when it
is released• Easier for everyone• Question of pricing
– Not for frail or otherwise high-risk patients– Does not replace heparin– Watch for studies comparing it to chronic
coumadin for long term anticoagulation– Look for dabigatran
The Infected Respiratory Tract
• Two studies of antibiotics– BMJ, October 2008– JAMA, December 2007
Purpose
• Assess the value of antibiotics (and steroids) for common respiratory tract infections
• Many guidelines and some prior evidence– Largely recommend against antibiotics for
most conditions in absence of pneumonia– Acute bacterial sinusitis more equivocal
#1 - Antibiotics for common respiratory infections
• Historical cohort study• 1.1 million episodes of respiratory infection
(URI, “chest infection,” sore throat, otitis,) in UK
• Record assessed for antibiotic prescription• Followed 1 month for diagnosis-specific
complications (pneumonia, quinsy, mastoiditis)
• Funded by UK Department of Health
Findings - Complications of Respiratory Infections
0
0.51
1.52
2.53
3.54
4.5
URI SoreThroat
Otitis ChestInfection
TreatedUntreated
(Elderly Patients Only)
#2 - Antibiotics and topical steroids for maxillary sinusitis
• Clinical trial, blinded• 240 adults with < 4 weeks acute
bacterial sinusitis (purulent discharge, local pain, pus on exam), no diabetes
• Treated with amoxicillin, budesonide spray, both or neither
• Followed for clinical cure at 10 days• Funded by UK Department of Health
Findings - Resolution of Sinusitis
01020304050
60708090
100
Amoxicillin Budesonide Nothing
10 Day Cure
Limitations• The respiratory complication study was
not a trial– Many ways that treated and untreated
groups may have differed– Including getting diagnosis of complication
• The sinusitis study was small– Could have missed difference in serious
complications
Implications
• Despite limitations– Antibiotics don’t seem important for bacterial
sinusitis, otitis, sore throat, URI– BUT, may be quite useful for “Chest Infection”
• Acute bronchitis?• NNT 40 to prevent pneumonia
– I will try to use less antibiotic for sinusitis (even acute bacterial) and otitis
– I will try to distinguish “chest infection” in older patients and treat
How Do You See the Colon?
• Two studies of CT Colonography– NEJM, October 2007
• Funding not reported,investigators receive money from makers of the colonography processing software
– NEJM, September 2008• Funded by National Cancer Institute and
American College of Radiology
Purpose• Determine whether a relatively non-invasive colonic
imaging technique can approach the ability of colonoscopy to detect pre- and early malignancies
• Colonoscopy never proven to reduce colon cancer mortality, but almost certainly does (FOBT does)
• Colonoscopy is expensive, inconvenient, and not completely safe– 1-3/1,000 have serious consequences, usually
associated with biopsies
• CT Colonography uses similar prep, insufflation, plus fluid tagging
#1 - CT Colonography for advanced neoplasia
– Cohort study, sort of– 6300 adults with no bowel disorder
• Half had enrolled in a CT colonography screening program (why?), with colonoscopy follow-up for selected findings
• Half were getting ordinary colonoscopic screening
– Assessed number and pathology of lesions found
– No follow-up
#2 - Accuracy of CT colonography
• “Test of a Test”• 2600 adults over 50, asymptomatic,
referred for ordinary colonoscopic screening– First received CT colonography– Follwed by immediate colonoscopy
• Assessed concordance for important polyps
Findings - Cohort Study
0
0.5
1
1.5
2
2.5
3
3.5
Advanced Adenomas Cancers
CTScope
*
Findings - Sensitivity Study
• CT detected– 90% of advanced lesions ≥ 1 cm– 65% of advanced lesions ≥ 5 mm
• CT incorrectly called abnormalities in 14% of subjects
Limitations
• First study had no direct comparison of CT to scope in the same patient– Why the excess of cancers in colonography?
• In both studies, CT found extracolonic stuff in majority of patients– Mostly trivial, often requiring further assessment
• In practice, unlikely to get immediate colonoscopy after positive CT– Requires repeat preps, other inconvenience
Implications
• CT Colonography still not ready for prime time– Difficult prep– Lots of follow-up colonoscopies– Lots of irrelevant findings
• I won’t be doing it• Fecal Occult Blood for my patients who
don’t want colonoscopy
After the Fall
• Prevention after a hip fracture
• NEJM, November 2007
• KW Lyles et al.
Purpose
• Determine whether annual infusion of zoledronic acid reduces subsequent fracture after hip fracture repair
• Inconclusive prior evidence about bisphsphonates following hip fracture
• Funded by the maker of zoledronic acid
Method
• Clinical Trial, blinded• 2100 adults with recent “minimal trauma” hip
fracture, previously ambulatory, no kidney disease, and refusing oral bisphosphonate
• Received Calcium and Vitamin D, plus– 5 mg IV zoledronic acid or placebo infusion
annually
• Followed 2 years for new clinical fractures and survival
Findings
0
2
4
6
8
10
12
14
Hip Fx Vertebral Fx Any Fx Death
Zoledronic AcidPlacebo
Limitations
• Mortality benefit unexpected and unexplained
• Industry-funded research has many opportunities for misleading reports
Implications
• Bisphoshonates reduce subsequent fractures and possibly mortality following hip fracture repair– NNT for another hip fx = 70 over 2 years– NNT for death = 27 (!)
• If oral bisphosphonates aren’t an option, zoledronic acid can be given IV yearly– Alendronate $100/month– Zoledronic acid $1200/year
Is the Blockade Working?
• Perioperative beta blockers
• The Lancet, May 2008
• The POISE study group– Funded by governments of Canada,
Australia and Spain, with some support from maker of the study drug
Purpose
• Reassess perioperative beta-blockade for preventing cardiac complications after non-cardiac surgery
• Several prior studies indicate improved post-operative cardiac outcomes with beta-blockade
• “Standard of care” for higher risk patients for at least 5 years– Some doubts due to study limitations and some
conflicting results
Method
• Clinical trial, blinded• 8300 adults worldwide, age ≥ 45, either existing
major vascular disease or at least 3 risk factors– Age>70, TIA, DM, CRF (2.0), CHF history, emergent
or high-risk surgery
• Received metoprolol, starting 4 hours pre-op, or placebo– Held for P<45 or SBP < 100
• Followed 1 month for major vascular outcomes and death
Findings
0
1
2
3
4
5
6
7
CompositeEndoint
MI Stroke Death
MetoprololPlacebo
Limitations
• Beta-blocker started immediately pre-op
• Drug held only for “consistent” severe bradycardia or hypotension
• Excluded patients whose physicians had planned to beta-block
Implications
• Perioperative beta-blockade, at least as done in this study, may be dangerous
• I’m limiting my use– Only beta-block if otherwise indicated– Only with plenty of advance time for slow up-
titration (a month!)– Not in higher stroke risk setting
• (Sad sigh…)
All you need is…Salt?
• Saline or bicarbonate for preventing contrast nephropathy
• JAMA, September 2008
• SS Brar et al.
Purpose
• Reassess whether bicarbonate infusion reduces contrast nephropathy
• Prior evidence that contrast nephropathy is common, around 25% of high-risk patients
• A few prior reports showed reduced nephropathy with pre-procedure bicarbonate hydration
• Funded by Kaiser Permanente
Method
• Clinical Trial, unblinded• 350 adults having non-emergent cardiac
catheterization, with GFR ≤ 60 and at least 1 of:– DM, CHF, HBP, Age > 75– Received either Sodium Bicarbonate, 150 meq in
1 liter D5, OR Normal Saline.• 3 ml/kg/hour for 1 hour pre-procedure, then 1.5
ml/kg/hour during and 4 hours after
– Followed 4 days for 25% fall in GFR
Findings
0
2
4
6
8
10
12
14
16
25% Fall in GFR 0.5 Cr Rise
BicarbonateSaline
Limitations
• Relatively small study
• Only coronary angiography patients
• Relatively good baseline GFR
Implications
• Bicarbonate might not be necessary for renal protection from contrast dye– Saline hydration probably acceptable
substitute
• However– Bicarbonate is not hard or apparently
dangerous to use– Should certainly use some form of
hydration
Staying Off the Sauce
• Baclofen to maintain alcohol abstinence
• The Lancet, December 2007
Purpose
• Assess whether baclofen can help achieve and maintain alcohol abstinence in cirrhotic alcoholics
• Growing interest in several drugs to help prevent alcohol craving and relapse– Naltrexone, acamprosate, topiramate
• Limited information, particularly in cirrhotic patients
• Funded by Italian government
Method
• Clinical Trial, blinded• 84 adults, age 18-75, with alcoholic cirrhosis,
at least 14 (women) to 21 (men) weekly drinks, and no other major system disease
• Admitted, given baclofen 5-10 mg tid, for 12 weeks, or placebo– Also frequent visits with counseling
• Followed 4 months for self- and family-reported abstinence – Dropouts assumed to be relapsed
Findings - Abstinence from Alcohol
0
10
20
30
40
50
60
70
80
Total Abstinence
BaclofenPlacebo
Limitations
• Small study
• Many dropouts, assumed relapsed– But similar results if assumed abstinent
• Duration only 3 months
• Used in context of additional support for abstinence
Implications
• I will try using it– High gain, low risk (NNT 2.5)– Avoid in renal dysfunction, epilepsy– Attempt to provide broader treatment
context
• But I’m not pushing this
Also Noted
• N-3 Polyunsaturated fatty acid supplementation may reduce 3-year mortality in CHF, NNT=60.– Lancet, 10/4/2008
• Telling smokers their “lung age,” derived from FEV1, may improve quit rates, NNT=14– BMJ, 3/6/2008
• Arthroscopic debridement and lavage does not help the osteoarthritic knee more than medicine and PT– NEJM, 9/11/2008
More “Also Noted”• In new type 2 (Irish) DM on oral treatment, home
glucose monitoring did not improve A1C but did worsen depression and anxiety– BMJ, 4/17/2008
• Low-dose risperidone may improve response in depression refractory to monotherapy, NNT=7 (Industry funded)– AnnIntMed 11/6/2007
• The US Preventive Services Task Force still does not recommend prostate cancer screening, and recommends against it after age 75– AnnIntMed, 8/5/2008
And Last -
• Coffee might decrease cardiovascular and overall mortality
• At 6 cups per day, over 25 years:– Men were 20% less likely to die– Women were 17% less likely to die– Independent of caffeine
• WARNING: Brought to you by the Nurses’ Health Study– Remember HRT?
• AnnIntMed, 6/17/2008
Summary
• Reconsider the A1C goal, use more metformin
• Oral thrombin inhibitors for perioperative DVT prophylaxis look promising
• Avoid antibiotics for most non-pneumonia respiratory infections; “chest infection” in the elderly may be an exception
• CT Colonography is pretty good, not yet ready• Bisphosphonates may be important after hip
fracture
Summary, cont’d
• Perioperative beta-blockade looks more risky than helpful
• Saline may be as good as bicarbonate for IV dye renal protection
• Baclofen may help alcohol abstinence in cirrhotics
• Coffee?
Remember:
• Before acting on anything you heard here, you may wish to study the original research, and discuss with colleagues or domain experts