You don't need to be signed in to read BMJ Blogs, but you can register here to receive updates about other BMJ products and services via our site.

Archive for June, 2016

Primary Care Corner with Geoffrey Modest MD: Statins Adverse Effects, Again

30 Jun, 16 | by EBM

By Dr. Geoffrey Modest

There are several significant adverse effects associated with statins noted in the literature and in our clinical practice. A blog a few years ago noted that there does not seem to be much difference in the incidence of myalgias between those on statins and controls, there is a 9% increased likelihood of diabetes, and there is a pretty high incidence of mild transaminase elevations (see http://blogs.bmj.com/ebm/2013/11/25/primary-care-corner-with-dr-geoffrey-modest-adverse-statin-effects )

  1. A recent blog looked at statins and myalgias, perhaps the most common patient complaint with statins, showing that a majority of people who think they are intolerant of even 3 different statins because of myalgias really do not have a true statin myopathy (a placebo-controlled study found that either they did not have more of a problem with atorvastatin vs placebo, or only had myalgias with the placebo!). See http://blogs.bmj.com/ebm/2016/04/29/primary-care-corner-with-geoffrey-modest-md-nonstatin-lipid-lowering-drugs-and-statin-myopathy/.

I should note that there is a relatively new entity found, an autoimmune myositis associated with statins: developing antibody against HMG-CoA reductase, and persistent symptoms after statin withdrawal. This needs aggressive therapy with prednisone or immunosuppressants — see Mammen AL. N Engl J Med 2016; 374: 664.

 

  1. Another oft-cited issue is statin-induced diabetes. This also is a tad murky, since:
  • There are studies suggesting that it is actually the low LDL itself and not the statin per se which is related to the diabetes (see blog http://blogs.bmj.com/ebm/2015/10/15/primary-care-corner-with-geoffrey-modest-md-low-ldl-and-diabetes-risk/ )
  • It is hard to find really granular data in the relevant studies on the association between statins and diabetes, bringing up a few issues:
    • The studies typically define the development of diabetes as we all do: arbitrarily as a HgbA1c of 6.5 (which is based actually on the approximate point where the incidence of diabetic retinopathy starts to increase), or by increased fasting glucose etc.
    • 70-80% of diabetics die from atherosclerotic cardiovasc disease (ASCVD), and even those with glucose intolerance (sub-6.5 A1c) have about a 2-fold increased risk of ASCVD. And statins work well in them in preventing ASCVD, as shown in several studies (same relative risk reduction as nondiabetics, higher absolute risk reduction)
    • This all makes it hard to determine the real clinical significance of “developing diabetes”. Are there really small changes in A1c levels (e.g., going from 6.4 to 6.5) causing us to lump these patients as “new diabetics”? Do these types of small differences really matter, since there seems to be a continuous relationship between A1c and ASCVD even in the pre-diabetic range? Similar argument for small changes in fasting glucose etc. A much more clinically significant outcome would be if there were more dramatic changes in A1c, both in the nondiabetic range (e.g. going from 5.0 to 6.4) or the diabetic range (e.g. 7.0 to 9.8), but we don’t know based on the data presented. For one review of the studies, see Preiss D. JAMA 2011: 305(24): 2556, which looked at older studies, mostly using fasting plasma glucose to diagnose diabetes, though the issue is the same.

 

  1. There have been a few reports suggesting that statins are associated with dementia. A recent large retrospective analysis did find an increase in reports of acute memory impairment in the first 30 days after initiating treatment, though this was found with all lipid lowering drugs (LLDs). The researchers utilized The Health Improvement Network, a database of anonymized patient records extracted from the general practitioners’ (GPs’) offices in the UK (see Strom BL. JAMA Intern Med 2015; 75: 1399). Prior studies have found mixed results; some finding improved memory [e.g. Jick H. Lancet 2000; 356(9242: 1627)] and others with no difference.

Details:

  • From 1987 through 2013, 482543 new statin users were compared to 482543 matched nonusers and 26484 users of nonstatin LLDs (bile-acid sequestrants, fibrates, niacin). Patients with a history of dementia /cognitive dysfunction or meds used for dementia were excluded, as well as those with traumatic brain injury or schizophrenia. A new statin user was defined as a person not on a statin for at least 1 year previously; similarly for nonstatin LLDs.
  • A case-crossover review of 68028 patients with incident acute memory loss looked at statin exposure in the period immediately before this diagnosis and the 3 earlier periods, to assess a statin association especially in patients who might have been intermittent statin users to see if they used statins more often in the immediate 30 days vs up to 300 days before.
  • The primary outcome: acute, reversible memory impairment.

Results:

  • Not so surprisingly, those on statins were older (63.4% >60yo vs 57.3% in nonusers), had more medical diagnoses with medical indications for LLDs/higher prevalence of comorbidities and meds. And (perhaps more surprisingly) those on nonstatin LLDs had more diabetes, hypercholesterolemia and cardiovascular disease than statin users, as well as more depression/anxiety. So the analyses were adjusted by propensity-score matching based on sex, age, and duration of enrollment.
  • First exposure to statins vs no LLD therapy: strong relationship with incident acute memory loss, with fully-adjusted OR 4.40 (3.01-6.41). This elevated risk decreased dramatically after the first 30-day period (they also looked at the 60-day reporting to see if there were some delay: again noting only an early association, decreasing monotonically thereafter).
  • Both atorvastatin and simvastatin showed the increase in the first 30 days, but there were fewer users of other statins in the studies (simvastatin being by far the most prescribed)
  • Overall, the association was strongest with the most lipophilic statins, and there was a clear dose-response curve, with the highest incidence of dementia with the highest doses of statins
  • But no statistically-significant difference between using statins vs nonstatin LLDs. Again, with nonstatin LLDs there was a notable increased association in the first 30 days [OR 3.60 (1.34-9.70)], decreasing monotonically thereafter.
  • The case-crossover analysis showed little association. 18.9% of patients with an episode of acute memory loss had received a statin in the prior 30 days. This was only slightly more when compared to other time periods up to 300 days prior.
  • There was a validation study of 100 patients with acute memory loss, done by direct outreach to the GPs. Most confirmed the acute memory loss, but only 38% had confirmed resolution of the memory loss. 43% had slowly progressive memory loss.
  • A rechallenge analysis found very few positive cases of memory loss (see table 10 in the supplement)

Commentary:

  • It would not be surprising if any LLD caused a change in neurologic functioning, since cholesterol is such an important component of neuronal membranes and changes could conceivably affect depolarization/neural transmission.
  • Several earlier studies (pre-statin) did find small increases in violence/suicides/accidents in those on LLDs. Monkey experiments did show more aggression with lipid lowering. But subsequent lipid-lowering studies in humans did not confirm this finding.
  • But, on the other hand, b-amyloid plaques incorporate cholesterol, and statins could conceivably interrupt this process. And there might be some additional benefit from their anti-oxidant and anti-inflammatory effects.
  • A few caveats of this study:
    • It was a really large, but still retrospective study. There was not a systematic approach to assessing acute memory change, but instead they relied on patient report and physician documentation. (Though the validation study found 76 of 86 patients had confirmed diagnosis, the actual incidence of reversible acute dementia was validated in only 38% of them)
    • There are likely biases: confounding by indication (patients on LLDs had many different baseline characteristics, and there may be unknown confounders there), detection bias (patients on LLDs probably saw their GPs more often, so more chances of reporting the memory change)
    • There was no systematic attempt to look at LLDs and chronic/slowly progressive memory changes.
  • It is at least a little reassuring that the acute effects on memory were short-term (within the first 30 days) with no evidence that they persist/happen later
  • So, a few issues going forward, assuming that there may be some adverse effect of statins on memory in some individuals:
    • We do not have a clear idea of what the goal LDL should be, and the new ACC/AHA guidelines of 2013 dismiss this issue (as those of you who have followed these blogs for a bit know, I do not agree with several of the recommendations of their guideline, including their dismissal of LDL goal)
    • For example, if I have a patient who has an indication for a statin (perhaps known atherosclerotic disease) but has an LDL of 96, should I really put them on a “high-intensity” statin and lower their LDL to 38? Or should I/the patient be happy with a lower intensity statin getting their LDL down to 60? Is there any cardiovascular benefit to such a low LDL? And, are there more adverse effects which could be associated with such a low LDL, such as diabetes, or memory issues etc. (at least there seems to be dose-response for the adverse effects)?
    • Are there differences between different statins – and, specifically between those most lipophilic (e.g. lovastatin, simvastatin or atorvastatin), which presumably cross the blood-brain barrier more readily, vs the more hydrophilic ones (e.g. pravastatin, and to a lesser degree, rosuvastatin). We do know clinically that the lipophilic ones are more associated with some CNS effects, such as insomnia or nightmares.

 

But, to me, perhaps the most important of the common statin adverse effects is that patients on statins are really happy with their lipid improvement (as are we), but a large % stop doing the lifestyle changes that they may have started after they hear their lipids are good (i.e., stopping the diet, exercise, wt loss). I think it is really important that we keep pushing on the lifestyle changes, since so many more bad health outcomes are related to lifestyle issues that are not addressed by statins (including diabetes, cognitive impairment, arthritis, cancer…….)

Primary Care Corner with Geoffrey Modest MD: Tai Chi for Knee OA; Mindfulness for Chronic Pain

29 Jun, 16 | by EBM

By Dr. Geoffrey Modest

  1. A recent studyfound that Tai Chi was at least as good, and sometimes better, than physical therapy (PT) for patients with painful knee osteoarthritis, OA (see doi:10.7326/M15-2143). There have been some earlier studies finding efficacy of Tai Chi for knee osteoarthritis, rheumatoid arthritis, and fibromyalgia, by decreasing pain and improving physical and psychological health. The current study compared Tai Chi with PT.

Details:

  • 204 people with symptomatic knee OA
  • Mean age 60, 70% women,  53% white/35% black, BMI 33, duration of knee pain 8 years, mostly moderate radiologic OA (Kellgren-Lawrence grade 2 in 38%, 3 in 37%), 50% hypertensive, 20% diabetic, mean WOMAC pain score (Western Ontario and McMaster Universities Osteoarthritis Index) 253 (range 0-500)
  • Interventions (patients allowed to continue meds, including acetaminophen and NSAIDs):
    • Tai Chi: 60 minute sessions 2x/week for 12 weeks. Explanation of mind-body exercise theory and procedures. Patients instructed to do home Tai Chiat least 20 min/d (videotaped with feedback throughout the study). At end of 12 weeks, patients asked to continue at home for the duration of the study
    • PT: 30 minute sessions 2x/week for 6 weeks. Individual assessment and targeted regimens. Exercise at home. At end of 6 weeks, patient asked to continue with 30-minute sessions 4 x/week for 6 weeks. [i.e., shorter intervention than Tai Chi, but this is a standard PT regimen]
  • Results (with 52 week follow-up)
    • Overall attendance: 74% for Tai Chi and 81% for PT
    • Clinical outcomes (WOMAC pain, physical function and stiffness scores; patient global assessment score; Beck depression inventory;  SF-36,  a health survey; arthritis self-efficacy score; and both the 6-minute and 20-minute walk scores): patients in each group showed improvement over time, including at 52 weeks, well after the active interventions. But, comparing the interventions: Tai Chiwas better than PT for essentially every outcome and at weeks 12, 24 and 52. However, the difference was statistically significant only for the physical component of SF-36 and Beck depression inventory.
    • Use of NSAIDs and analgesics: also generally less with Tai Chi, but not reach statistical significance.

Commentary:

  • This was perhaps a somewhat unexpected finding since the focus of PT is so different from Tai Chi. PT largely involves stretching and strengthening exercises and some local therapies, leading to improved quadriceps dynamics in particular, developing increased support for the knee and decreasing the load on the joint itself (at least that is my understanding. Studies have shown that quadriceps weakness correlates with the degree of knee pain). Tai Chicombines meditation, slow and gentle movements, deep diaphragmatic breathing and relaxation (i.e., physical as well as psychosocial/emotional/behavioral elements).
  • In this study on Tai Chi, it is impressive that the results remained pretty consistent at 12, 24, 52 weeks. My guess is that the Tai Chi group did continue their home-based exercises after the formal study stopped (data not in article), but either way, that suggests that the benefits are durable (and perhaps Tai Chi really can be incorporated into one’s life long-term)

 

  1. A complementary article appeared near the same time in JAMA, stressing a role for mindfulness meditation in pain management (seedoi:10.1001/jama.2016.4875). Briefly, mindfulness meditation involves an increasing awareness of body sensations (e.g. breathing), techniques to promote mindful practice (yoga, meditation), learning how to understand and change how we react to stress, understanding the relationship between stress and pain, and viewing the reactions to stress without judgment. This JAMA Perspective highlights some pretty impressive studies:
  • A randomized controlled trial showing that mindfulness-based stress reduction (MSBR)was comparable to cognitive behavioral therapy (CBT) in reducing chronic low back pain, finding that there was no difference between MBSR and cognitive behavioral therapy (CBT), both with about a 45% reduction in pain (vs 25% with usual care). For my review of full article, see prior blog: http://blogs.bmj.com/ebm/2016/04/07/primary-care-corner-with-geoffrey-modest-md-low-back-pain-improves-with-stress-reduction-mindfulness-and-cognitive-behavioral-therapy/
  • Another RCT involved 282 older adults with chronic low back pain, also finding that those with 8 weeks of mindfulness meditation followed by 6 monthly sessions showing that 45% of the patients experienced >30% reduction in pain vs 25% of the patients in the control group.
  • Using functional magnetic resonance imaging of volunteers exposed to a noxious stimulus, those who practiced mindfulness meditation had a 57% decrease in how unpleasant the stimulus felt and a 40% decreased rating of pain intensity vs control (and those who paid attention just to breathing did not have these benefits). They found that meditation was associated with more activation of the orbitofrontal cortex (OFC), an area of the brain which “controls how people put into context what they sense in the environment”. Subjects commented that they did in fact feel the pain but were able to “let it go” and not dwell on it. The meditation also led to less activation of the thalamus, which serves as the pain gateway from the spinal cord to the brain, and activation of the anterior cingulate cortex (ACC), involved in cognitive control and emotional regulation. And, interestingly enough, though there are plenty of opioid receptors in the OFC and ACC, mindfulness did not affect these receptors (naloxone had no effect).

 

Overall Commentary:

  • These articles reinforce the intimate connection between pain perception and one’s psychosocial state.
  • There seems to be a shift in thinking about chronic vs acute pain, with argument that the issue with chronic pain involves different/more extensive central involvement (hyperalgesia, changes in functional neuroimaging, more somatic symptoms such as fatigue, memory problems, insomnia, mood disorders), which supports the use of different CNS-directed treatments (SNRIs, anticonvulsants) — for more info see Phillips K. Best Pract Res Clin Rheumatol 2011; 25: 141.
  • Those with chronic pain often have increased response to peripheral stimuli (hyperalgesia/allodynia), rate pain as more severe, and those with chronic widespread pain often have specific focal triggers, such as myofascial trigger points, ligamentous trigger points, or osteoarthritis of the spine or joints. And these focal triggers can lead to/perpetuate the chronic pain (see Staud R. Best Pract Res Clin Rheumatol 2011; 25: 155.) This ties together the complex interaction between peripheral triggers/local changes (e.g. increased lactic acid production, cytokines) and the central interpretation of that pain, including the sensation of chronic widespread pain. (It is really common for a patient with a particularly painful local site to have much more diffuse bodily pain. in my experience, I have sometimes been able to treat the triggering source with injections, leading to a generalized decrease in pain overall).
  • Also, it is pretty clear that stress itself may be a bad actor: stress leads to muscle tightness (perhaps part of the fight/flight response and readiness to act). But chronic stress leads to chronic muscle tightness and pain transmitted largely through the spinothalamic tract pain fibers to the thalamus/ACC/etc. and then to the cerebral primary somatic sensory cortex. As noted above, several of these processing stations alter their pain response by meditation.
  • In terms of the peripheral musculoskeletal effects of chronic stress, it seems to me that there are certain areas of muscle tightness that are more common with chronic stress, such as at the occipital insertion of the trapezius, diffusely in the muscles around the cranium/tension headaches, costochondral areas in the sternum, lower back, and several of the “trigger points” of fibromyalgia.)
  • To me, the above studies suggest a couple of things:
    • Decreasing the functional impact of stressors, whether through mindfulness meditation, Tai Chi or CBT, can decrease the direct effects of stress on the muscles (and there is the argument that many people with chronic pain “catastrophize” it: thinking negative thoughts about how pain will affect their function, which might expectantly exacerbate pain sensation. these psychological techniques directly affect how stress is perceived and handled)
    • And, there are impressive data that there are also meditation effects on how various parts of the brain itself (thalamus, anterior cingulate cortex, etc.) fundamentally respond to painful stimuli
  • The CDC stresses that the preferred therapy for pain management is maximizing nonpharmacologic and nonopioid pharmacologic pain management (see http://blogs.bmj.com/ebm/2016/03/25/primary-care-corner-with-geoffrey-modest-md-new-cdc-guidelines-for-opiate-prescribing/ for review)
  • So, whether the beneficial effect of mindful meditation, CBT or Tai Chi is through moderating the perception of stress or the changing the way pain is handled centrally, or both (hard to separate), it seems to me to be increasingly clear that there are some important and perhaps fundamental differences in how patients with chronic pain experience their pain, and that we should be more aggressively pursuing a more global approach, including mindfulness meditation, Tai Chi, or CBT as a really important nonpharmacologic component to helping people with chronic pain. this approach coincides with the CDC focus on maximizing nonpharmacologic adjunctive therapies as a way to avoid opiates or at least minimize their use. In this context, it is intriguing that the above alterations in central pain pathways by mindfulness meditation do not seem to be entirely related to opiate receptors, both suggesting that either endogenous or exogenous opiates may not be necessary for effective chronic pain management and that there may be room for development of other, nonopiate meds as part of chronic pain treatment…

Primary Care Corner with Geoffrey Modest MD: More Superbugs

28 Jun, 16 | by EBM

By Dr. Geoffrey Modest

Following on the last blogs on colistin-resistant E coli (see http://blogs.bmj.com/ebm/2016/06/21/primary-care-corner-with-geoffrey-modest-md-e-coli-superbug-is-spreading/ and http://blogs.bmj.com/ebm/2016/06/23/primary-care-corner-with-geoffrey-modest-md-response-and-further-comments-on-e-coli-superbug-is-spreading/), Paul Susman sent me the link http://www.coha.org/super-bacteria-in-rio-de-janeiros-olympic-arenas/ on the increasing spread of  Klebsiella pneumoniae carbapenemase (KPC)-producing bacteria, which has been around for many years. A 2011 article (Arnold RS, South Med J 2011; 104: 45) noted at that time that this difficult-to-treat organism, associated with significant morbidity and mortality, had already spread from the northeastern US to most of the world. The CDC commented that 50% of patients infected with this organism will die from it. The current link references the upcoming Olympics in Brazil, noting:

  • KPC is found off the beaches in Rio, where rowing, canoeing and swimming events will take place (i.e., the swimmers, sailors and rowers will be exposed). The source is likely sewage contamination of the waterways
  • 10 samples were taken from five different beaches (Copacabana, Ipanema, Leblon, Botafogo, and Flamengo) and all tested positive for KPC
    • Copababana is the site of the triathlon and open swimming competitions, tested positive in 10% of the samples
    • Botafogo, where sailing and canoeing events are to be held, had 100% of samples positive
    • Ipanema and Leblon, popular tourist beaches, had 50 and 60% of samples positive
    • Flamengo beach, where sailing competitions will be held, 90% tested positive
  • Another study found that Guanabara Bay was contaminated, likely because waste from thousands of households and hospitals that dump into the streams and rivers that empty there. They note that “at least 50% of the untreated sewage from Rio de Janeiro is dumped in to Guanabara Bay.
  • KPC has been found there since 2010. The authorities promised the water would be cleaned. But alas….
  • Another article on Brazil highlights the high rates of methicillin-resistant Staph, vancomycin resistant Enterococci, b-lactamase resistant Klebsiella and E coli, and KPC (see Rossi F. Clinical Infectious Diseases 2011; 9: 1138). They attribute the resistance to: overuse of superantibiotics (e.g. colistin in the ICUs), high total consumption of antibiotics, availability of over-the-counter antibiotics, inadequate dosing of antibiotics, and poor adherence.
  • Also Brazil is the world’s largest beef exporter with the world’s largest commercial cattle herd. The government does regulate antibiotic use (as opposed to the US), though antimicrobial resistance has been found in Brazilian cattle

Commentary:

  • I guess there is more concern about the Olympics than Zika…
  • I don’t mean to single out Brazil by this example. But this example yet again reinforces the big picture of potential outbreaks of untreatable infections caused by antibiotic resistance in common microbes. This is a worldwide phenomenon and really does need to be approached in a coherent worldwide manner

Primary Care Corner with Geoffrey Modest MD: USPSTF Colorectal Cancer Screening Guidelines

27 Jun, 16 | by EBM

By Dr. Geoffrey Modest

The USPSTF just published their 2016 guidelines for colorectal cancer screening (see doi:10.1001/jama.2016.5989).

Background:

  • Colorectal cancer is the second leading cause of cancer death in the US
  • In 2016, estimates are that 134K people will be diagnosed with the disease, and 49K will die from it.
  • Most frequent age range for diagnosis is 65-74, median age of death is 68
  • About 1/3 of adults never get screened

Recommendations, for asymptomatic adults at average colorectal cancer risk:

  • Screen from age 50-77 (grade A recommendation)
  • Screen from age 76-85 (grade C recommendation), as an individual decision, taking into account the patient’s overall health and prior screening history. specifically focusing on:
    • Adults in this age group who have never been screened (higher likelihood of abnormality)
    • Those healthy enough to undergo treatment if cancer is detected and do not have comorbid conditions that would significantly limit their life expectancy
    • They conclude “with moderate certainty that the net benefit of screening in adults aged 76-85 who have been previously screened in small”
  • The benefit of screening >85yo is likely to be very small, since the time between detection/treatment and the actual mortality benefit can be substantial
  • 7 screening options:
    • Guaiac fecal occult blood test (gFOBT) every year
    • Fecal immunochemical test (FIT) every year
    • Multitargeted stool DNA test (FIT-DNA) every 1 or 3 years (this test combines a FIT with testing for altered DNA biomarkers, has higher sensitivity than FIT but more false positives and leads to more colonoscopies, with their attendant risks). The manufacturer recommends every 3 years
    • Colonoscopy every 10 years
    • CT colonography every 5 years
    • Flex sigmoidoscopy every 5 years
    • Flex sig every 10 years, plus FIT every year
  • They do not prioritize the order, noting that all of these tests have benefit, and that the main goal is screening (with the hope that multiple different options will increase the screening rate)
  • The harms of screening increase with age, going from small in those 50-74, to small-to-moderate in those >75, especially for colonoscopy
  • They provide a bar graph of benefits and harms, noting
    • Life-years gained per 1000 screened: in the 260-270 range for colonoscopy and FIT-DNAevery year; 240-250 for FIT, gFOBT, CT colonoscopy, and flex sig/ FIT; in the 220 range for flex sig and for FIT-DNA every 3 years
    • For colorectal cancer deaths averted per 1000 people screened: 23-24 for colonoscopy, flex sig/FIT and FIT-DNA every year; 22 for FIT, gFOBT, CT colonoscopy; 20 for flex sig or FIT-DNA every 3 years.
    • For harms per 1000 screened: lowest (9-11) for flex sig/FIT, FIT-DNA every 3 years, FIT, gFOBT, CT colonography, and flex sig/FIT; 12 for FIT-DNA every year and highest (15) for colonoscopy
    • For lifetime burden of colonoscopies per 1000 screened: under 1820 for FIT-DNA every 3 years , FIT, CT colonography and flex sig; 2200-2300 for gFOBT and flex sig/FIT; 2662 for FIT-DNA every year; 4069 for colonoscopy.

Commentary:

  • I agree with them that the most important thing is doing the screening, since the benefits really are not that different by which test is done, and that offering noninvasive testing is clearly attractive to many patients (and may increase the screening rate). I would certainly add that any person getting a non-colonoscopic screen should understand and agree to pursuing colonoscopy if positive
  • Note that the above numbers are from different studies in different populations and are not based on head-to-head comparisons of different screening strategies, so the actual numbers above (risks/benefits) are not really strictly comparable. But they do probably provide a reasonable ballpark estimate
  • For the recently released Canadian guidelines, see  http://blogs.bmj.com/ebm/2016/03/04/primary-care-corner-with-geoffrey-modest-md-colorectal-screening-guidelines-from-canada/ , which is pretty dismissive of colonoscopy screening, are not in favor of screening those >74 yo, and actually mostly support using gFOBT or FIT every 2 years or flex sig every 10 years. This blog includes my additional concerns about colonoscopy, where the risk does increase significantly with age, and argues that the lesions that benefit from screening are basically the ones the sigmoidoscope reaches.
  • I am concerned about using CT colonoscopy (radiation exposure, competence/accuracy in different settings esp if not done regularly, pick up of incidental non-colon findings which lead to unnecessary testing) and the FIT-DNA testing. For the latter, what does one do if the FIT-DNA is positive but the colonoscopy is negative? Is it a false positive, or is there a small cancer lurking in the crypts? Should there be a follow up colonoscopy in 1 year. What if that is negative — keep repeating it every 1-2 years??? Seems to me to be a black hole, and we should avoid that event horizon.
  • They do not suggest a different strategy for Black or Alaska Native individuals (who have higher incidence and mortality than the general population) because there are no empirical data on the effectiveness of different screening strategies for these populations. And there are some studies showing that equal treatment seems to produce equal outcomes, suggesting that the issue may be more of access than actual risk.
  • So, my own approach is evolving, though I am much more strongly leaning towards FIT testing alone or in combination with flex sig, since these options seem to maximize the benefit and cut in half the number of colonoscopies with their associated intensive prep, conscious sedation, and increased risk of perforation (all worse in older people), and with small apparent benefit.

Primary Care Corner with Geoffrey Modest MD: Carvedilol, Best Drug For Portal Hypertension?

24 Jun, 16 | by EBM

By Dr. Geoffrey Modest

A recent review looked at the varying efficacies of different nonselective b-blockers in patients with portal hypertension in cirrhosis (see Li T. BMJ Open 2016; 6: e010902), finding that carvedilol may be more effective than propranolol or nebivolol and as effective as the combo of nadalol plus isosorbide mononitrate.

Details:

  • 12 RCTs were evaluated, though mostly not-so-great quality

Results:

  • 7 trials (379 patients, about 2/3 male, mean age around 50, most with alcoholic cirrhosis but some hepatitis B or C) compared carvedilol vs propranolol for hemodynamic outcomes: hepatic venous pressure(HVPG) reduction, hemodynamic response rate, post-treatment arterial pressure (mean arterial pressure, MAP). Most follow-up of only up to 6 weeks. Finding:
    • Carvedilol was associated with a greater (%) HVPG reduction within 6 months (mean difference -8.49 (-12.36 to -4.63) and improvement in hemodynamic response rate, both being consistently found in each of the trials. There was a nonsignificant decrease in MAP in almost all of the trials
  • 3 trials compared carvedilol vs endoscopic variceal band ligation (EVL) for clinical outcomes: all-cause mortality, bleeding-related mortality, upper GI bleeding.
    • 1 trial was secondary prevention, 64 patients, 26 months follow-up, Child-Pugh class B. No significant difference in clinical outcomes, though very wide confidence intervals
    • 2 trials of primary prevention, 341 patients, 13-26 month median follow-up, mostly Child-Pugh class A. Also no significant difference in clinical outcomes
  • 1 trial compared carvedilol vs nadalol plus isosorbide mononitrate, 121 patients, 30 month follow-up:
    • No significant difference in mortality or bleeding. carvedilol group had fewer adverse events (5/61 vs 23/61)
  • 1 trial compared carvedilol vs nebivolol, 20 patients, 14 day follow-up:
    • Carvedilol had greater reduction in HVPG, mean difference of -10.9 percentage of HVPG reduction

Commentary:

  • Acute variceal bleeding has 6 week mortality of 10-20%, depending on Child-Pugh class; and a 60% 1-year rate of recurrent variceal bleeds
  • HVPG is the strongest predictor of bleeds: if >10 mmHg, it predicts varices; if >12 mmHg, it is associated with high rate of bleeds; if >20 mmHg, there is a high mortality.
  • Risk of bleeding decreases if HVPG decreases to <12 mmHg or by 20% from baseline
  • Nonselective b-blockers work by both decreasing cardiac output (b1) and constricting splanchnic vessels (b2); but only 40% of patients reach therapeutic levels
  • Carvedilol (blocking both a1 and b1/b2) also decreases intrahepatic resistance. Prior reports have found that >50% of propranolol non-responders do respond to carvedilol, though there are concerns about systemic hypotension and renal failure (and perhaps increased mortality) especially in people with refractory ascites.
  • The above trials were pretty short-term. It is somewhat reassuring that in the carvedilol studies looking at hemodynamic outcomes, there was not much difference acutely and at 6 months, suggesting that the benefit was, and might continue to be, sustained
  • Studies suggest that there is not much additional effect of higher doses of carvedilol in terms of HVPG (i.e., it may be reasonable to start with 3.125 mg bid, then increase as tolerated to 6.25 mg bid)
  • There is not a lot of data comparing carvedilol to EVL, but the limited data above suggest relative equivalence, and the quality of evidence is low. Other studies have found that EVL is superior to nonselective b-blockers.

What does this all mean?

  • Interestingly (and anecdotally), I happened to see a patient today who was not tolerant of nadolol (made her feel weak, trouble exercising/walking). And, having seen this review, I started her on carvedilol 3.25mg bid.  We’ll see….
  • There are concerns about EVL, since it requires endoscopy and a procedure, and about 50% need retreatment within one year. The data comparing EVL and nonselective b-blockers is a bit unclear: a review (Cheung J. Alimentary Pharmacology and Therapeutics 2009; 30: 577), which included only propranolol and nadolol as meds, found pretty poor quality of evidence (significantly, no documentation of the number of patients actually achieving target heart rate). Their comparison noted that EVL was better than b-blockers if the mean dose was <80mg. But given other studies finding lots of people not achieving optimal b-blocker dose, it is hard to be sure that EVL really is better than even the old b-blockers.
  • So, I am a bit concerned about the reports of more hypotension with carvedilol, since many of my patients with cirrhosis have pretty low blood pressures to begin with, and the reported concerns of possible increased acute kidney injury/mortality in patients with advanced decompensated liver disease/refractory ascites is a tad worrisome. Since the above data are mostly based on not-so-great quality studies with small numbers of patients, I think it is important to have bigger, better, and longer comparative studies before using carvedilol as first-line. so, my approach will be to continue using nadolol or propranolol initially, but as in the case I saw today, to try carvedilol gingerly if the other b-blockers are not tolerated/target heart rate not achieved (reduction of 25% or down to <55-60 bpm)

Primary Care Corner with Geoffrey Modest MD: Response and Further Comments On: E. coli Superbug is Spreading

23 Jun, 16 | by EBM

By Dr. Geoffrey Modest

An article was sent by Burak Alsan, from an interview with his wife Marcie Alsan (see http://blogs.plos.org/publichealth/2016/05/31/to-fight-superbugs-fight-poverty). She stresses the connection between socioeconomic disparities and infectious diseases, specifically noting that “out of pocket payments were the most significant correlate of antimicrobial resistance across countries” and that “the entire correlation was driven by countries that had in place a policy by which copayments were imposed in the public sector”. In particular, she found that of 47 countries, out-of-pocket health expenditures were the only factor significantly associated with antibiotic resistance, controlling for socioeconomic and environmental factors (e.g. sanitation, animal husbandry, and poverty) and structural health-care features (e.g. physician density, hospital bed density, total health expenditures). In particular, a “ten point increase in percentage of health expenditures that were out-of-pocket was associated with a 3.2 percentage point increase in resistant isolates”. For details, see Alsan, M. Lancet Infect Dis 2015; 15: 1203.

Commentary

  • I think this is really important. I mentioned out-of-pocket expenses in my rantings on one of the fundamental problems with our health care system: we do not place primary care at its center. Providing easy access to primary care is not only much cheaper to the system but undoubtedly derives better outcomes, with more coordinated, less interventive care that prioritizes a more holistic approach to the broad biopsychosocial aspects of the patient and focuses on the therapeutic benefits of a strong provider-patient relationship. All of this means having free and easy access to primary care (decreasing obstacles to access, such as copayments by patients), as well as reorienting the incentives in the system to promote the training and job satisfaction of primary care providers
  • And, as mentioned before, I have seen way too many patients with treatable conditions (e.g., cellulitis, hypertension…), unable to pay their copays for meds, then hospitalized with serious conditions (sepsis, stroke…). A huge human as well as monetary cost…
  • In the case of antibiotic resistance, there was an article in the Boston Globe recently finding further spread of colistin-resistant E coli, noting a few pretty scary things: this “superbug” was found in another pig in the US, but concerning enough “each of the three US cases (2 in pigs, and the one woman from Pennsylvania) involve different strains of E. coli. The latest animal case suggests the gene is spreading through multiple routes here.” (See https://www.bostonglobe.com/news/nation/2016/06/14/superbug-found-second-pig-sample/OPUjpJ5aMA7ontXAjMW4xL/story.html ). Seems like a pretty urgent thing to tackle….

See http://blogs.bmj.com/ebm/2016/06/21/primary-care-corner-with-geoffrey-modest-md-e-coli-superbug-is-spreading/ for the original blog

Primary Care Corner with Geoffrey Modest MD: Liraglutide Decreases Cardiovascular Events

22 Jun, 16 | by EBM

By Dr. Geoffrey Modest

There was a recent article from a paper at the Am Diabetes Assn annual meeting finding decreased cardiovascular events in patients on the glucagon-like peptide 1 (GLP-1) agonist liraglutide (see DOI: 10.1056/NEJMoa1603827).

Details:

  • 9340 patients with type 2 diabetes and high cardiovascular risk, randomized to liraglutide vs placebo, with mean follow-up 3.8 yrs. Drug-company sponsored study begun in 2010, in 410 sites in 32 countries
  • 65% male, mean age 64, duration of diabetes 13 yrs, 35% North America/30% Europe/8% Asia, A1c 8.7, BMI 33, BP 136/77, 18% with heart failure, 81% prior MI, 16% prior stroke/TIA, 25% chronic kidney disease CKD with eGFR<60, 35% with normal renal function.
  • 76% were on lipid lowering meds, with 72% of them on statins; 67% on antiplatelet drugs of which 64% on aspirin; 93% on BP meds, of whom 57% on b-blocker, 84% on ACE/ARB.
  • Inclusion criteria — all patients needed initial A1c>7.0% (either treated or untreated), and:
    • Were over 50yo, and at least one cardiovascular problem (coronary heart dz, cerebrovasc dz, peripheral arterial dz, chronic kidney dz at least stage 3, or chronic heart failure NYHA class II or III)
    • Or, were >60yo with at least one cardiovasc risk factor (microalbuminuria or proteinuria, hypertension and LVH, LV systolic or diastolic dysfunction, ankle-brachial index of <0.9)

Results:

  • A1C was 0.40 percentage points lower in those on liraglutide, though more placebo patients were on metformin, sulfonylureas, TZDs (?which one), glinides, SGLT-2 inhibitors, or the array of insulin preparations
  • Liraglutide was also associated with significantly more weight loss (2.3kg), lower systolic blood pressure (1.2 mmHg), but higher diastolic (0.6 mmHg), and higher pulse (3.0 beats/min)
  • Primary composite endpoint (first occurrence of death from cardiovasc causes, nonfatal MI, or nonfatal stroke): 608 of 4668 patients on liraglutide (13.0%) vs 694 of 4672 on placebo (14.9%), a13% decrease [HR 0.87 (0.78-0.97), p<0.001 for noninferiority and p=0.01 for superiority]. Overall NNT to prevent one event in 3 years was 66
    • 219 patients died from cardiovascular causes (4.7%) vs 278 on placebo (6.0%), a 22% decrease [HR 0.78 (0.66-0.93),  p=0.007]
    • 381 patients died from all causes (8.2%) vs 447 on placebo (9.6%), a 15% decrease [HR 0.85 (0.74-0.97),  p=0.02]; NNT to prevent one death from any cause in 3 years was 98
    • Rates of nonfatal MI, nonfatal stroke and hospitalization for heart failure were nonsignificantly lower with liraglutide
  • Review of the graphs show that a clear separation (improvement with liraglutide) was seen after about 12-18 months, with perhaps some splaying of the curves over time
  • Subgroup analysis revealed: pretty consistent though usually nonsignificant benefit of liraglutide independent of the subgroup, though
    • Clear benefit in the 7598 patients in the >50 yo and documented cardiovasc disease group but a trend to doing less well in the >60 yo with only cardiovasc risk factors
    • Benefit in those with chronic kidney disease was most evident in those with eGFR <60 (too few in the <30 group to be really evaluable)
    • Clearer benefit in those with BMI>30 and those with A1c>8.3 and with duration of diabetes <11 years
  • Microvascular outcomes: liraglutide associated with decreased nephropathy events (1.5 vs 1.99/100 patient-yrs, a 22% reduction,p=0.003). Nephropathy events were the composite of new onset macroalbuminuria, doubling of creatinine and eGFR <45, need for continuous renal-replacement therapy, or death from renal disease
  • Adverse events:
    • Liraglutide assoc with more GI events leading to discontinuation of the med
      • More severe hypoglycemia in the placebo group (153 vs 114 events)
      • More GI events with liraglutide (significantly more acute cholecystitis in 36 vs 21; more nausea, vomiting, diarrhea, abdominal pain, decreased appetite
    • The incidence of pancreatitis (found in some prior GLP-1 studies) was nonsignificantly lower with liraglutide
    • No difference in neoplasms, though specifically: more pancreatic (13 vs 5, nonsignificant, and a few recent large analyses have not find increased pancreatic cancer), same numbers for melanoma, but significantly less prostate cancer (26 vs 47 cases in the placebo group)

Commentary:

  • One reassuring aspect of the study was that it was a large study of a diverse group of patients with long-standing diabetes (mean 13 years, which is a concern because the GLP-1 agonists require a functioning pancreas to secrete insulin, and there has been lingering concern in my mind that patients with such longstanding diabetes might not have much b-cell reserve, though I should note that those with longer-standing diabetes did a bit less well)
  • The decrease in cardiovascular events is pretty impressive, especially since it seems that patients were mostly on pretty appropriate cardiac meds
  • Some question about generalizability: the benefit was really confined to younger patients who had definite cardiovascular disease. Would that still be the case with longer-term follow-up?? (after all, the likelihood of decreasing atherosclerotic events is likely to be higher in those with more advanced disease, both because of a higher absolute risk in that group and a greater likelihood of an event sooner than later). My guess is that longer follow-up would show benefit in those with just a cardiovasc risk factor, since 80% of diabetics overall die from carvdiovac causes. Also, as per many prior blogs, I am concerned that short-term studies will tend to understate adverse effects, which may take longer to manifest themselves.
  • As per prior blogs, I am very impressed clinically with the GLP-1 agonists, because of their often dramatic effects on glucose control, their relative lack of hypoglycemia, their tendency to decrease weight and not cause hyperinsulinemia, their apparent ability to supply a very specific peptide which is part of normal glucose control but deficient in diabetics, and their overall acceptability by patients. Seehttp://blogs.bmj.com/ebm/category/diabetes/ for my litany of blogs on them and other meds
  • And my real concerns about most of the newer meds is that their approval is not based on real clinical outcomes, just A1c effects. So, this study reinforces that these drugs are clinically beneficial. By the way, the FDA just strengthened their warnings about SGLT-2 inhibitors because of increasing numbers of cases of acute renal injury, though they did not include empagliflozin. For my critique of the original empagliflozin study finding potential cardiovascular benefit, see http://blogs.bmj.com/ebm/2015/12/21/primary-care-corner-with-geoffrey-modest-md-empaglifozin-the-good-and-the-bad/ )

Primary Care Corner with Geoffrey Modest MD: E. Coli Superbug is Spreading

21 Jun, 16 | by EBM

By Dr. Geoffrey Modest

Not-so-shockingly, the US government reported the first case of a patient in the US with colistin-resistant E. coli, in a 49 yo Pennsylvania woman with a urinary tract infection. They found a plasmid-mediated piece of DNA which passed along the mcr-1 gene, conferring colistin resistance (see http://www.reuters.com/article/us-health-superbug-idUSKCN0YH2KT . For the case study, see doi:10.1128/AAC.01103-16). This is precisely the issue reported last year in China [see http://blogs.bmj.com/ebm/2015/12/10/primary-care-corner-with-geoffrey-modest-md-troubling-microbiome-changes . This blog looked at the initial report from China, noting that China is the world’s largest producer of poultry and pig products and one of the largest veterinary users or colistin (aka polymixin E), finding this resistant bug in 16 samples from various sites: sputum, urine, ascitic fluid, bile…]. Since the Chinese finding, the  Walter Reed National Military Medical Center has been testing all extended-spectrum b-lactamase-producing E. coli for colistin-resistance.

This current finding and the blog (which goes into detail on the China case) basically make the following points:

  • Antibiotic resistance is a huge and growing problem: >2 million illnesses blamed on antibiotic resistance in the US with 23,000 annual deaths
  • Potential for spread of this E. coli, resistant to the last-line big gun colistin, is real, and highlighted by this case in the US 6 months after the report in China. The concern is that the ability for E. coli to elaborate plasmid-mediated resistance makes it more spreadable than through the prior identified mechanism of chromosomal changes
  • Part of the issue is us guys: more than 1/2 of patients admitted to the hospital get an antibiotic, but 30-50% of the time it is considered unnecessary. Blogs below look at the overuse of antibiotics in the outpatient setting as well.
  • The biggest part of the problem is use/remarkable overuse of antibiotics in agriculture (as in the colistin case), leading to lots of resistant bugs (see blogs below).
  • And there is real concern, as per the CDC, that we may be entering a post-effective antibiotic era
  • Confounding the issue, is the lack of aggressive antibiotic development by drug companies: even expensive antibiotics (the $100-200/pill ones) are typically given for short, and therefore less-profitable, courses of treatment. The $$ is in long-term, preferably lifelong need for meds.  Some of the big drug companies who mostly foreswore against antibiotic drug development have “signed a declaration calling for new incentives from governments to support investment in development of medicines to fight drug-resistant superbugs”. Perhaps the hepatitis c drug model will develop: hugely overpriced drugs, which even for only a few months can develop huge profits…???

See http://blogs.bmj.com/ebm/category/antimicrobial-resistance/ for a litany of articles/critiques on antibiotic overprescribing, resistance, and efforts to combat resistance, including a US White House report from last year

Primary Care Corner with Geoffrey Modest MD: Continue Aspirin After Lower GI Bleeds?

17 Jun, 16 | by EBM

By Dr. Geoffrey Modest

A rather common clinical conundrum in patients with/at high risk for cardiovascular disease is whether to reinstate aspirin after a GI bleed. A recent retrospective study from Hong Kong suggested that the benefit outweighs the risk for lower GI bleeds (see doi.10.1053/j.gastro.2016.04.013). They reviewed their data on patients with documented lower GI bleed (melena or hematochezia and absence of upper GI source) from 2001-2008.

Details:

  • 295 patients who were on aspirin at the time of the lower GI bleed, followed up to 5 years
  • 93% were considered to be of “high cardiovascular risk”, using the Antithrombotic Trialists’ Collaborative definition: history of unstable angina, acute MI, prior MI, stroke, or TIA
  • Non-users were defined as taking aspirin <20% of the follow-up period (n=121; 87% actually took aspirin <10% of the follow-up period); users were those with cumulative use >50% of that period (n=174; 84% took aspirin >75% of the follow-up period). Aspirin use was determined by prescription patterns.
  • All used <=160mg aspirin, and 88% used only 80 mg/d
  • Non-users were older (76.7 vs 73.1 years, p=0.003), fewer smokers (26.4% vs 42.0%, p=0.006), and more needed transfusion of >= 2 units (54.5% vs 39.7%)
  • Predefined covariables at baseline: age, sex, alcohol consumption, smoking, severity of comorbidities, history of GI bleeding (upper and lower), blood transfusion, meds (anticoagulants, steroids, non-aspirin antiplatelet drugs) within the 30 days prior to index bleed.
  • Outcomes assessed: recurrent lower GI bleed, serious cardiovascular events (nonfatal MI, nonfatal stroke, death from vascular cause), and deaths from other causes
  • Results, comparing non-users to users:
    • Lower GI bleeding recurred in 18.9% of those on aspirin vs 6.9% of non-users (SHR 2.76; p=0.011) [SHR is subdistribution hazard ratios, which from my understanding is the probability of an event due to aspirin at a moment in time, comparing cause-specific cumulative incidences of different effects]
      • Overt bleeding in 6.6% of non-users (n=8) vs 17.8% of users (n=31). both groups were transfused a median of 2 units of blood
      • Occult bleeding in 1.7% (n=2) of non-users vs 6.9% of users (n=12)
    • Serious cardiovascular events occurred in 22.8% of those on aspirin vs 36.5% of non-users (SHR 0.59; p=0.019)
    • 2% of aspirin users died from other causes vs 26.7% of non-users (SHR 0.33; p=0.001): 42 patients overall died, including 22 from sepsis, 10 from cancer, and 6 from renal failure
    • Multivariable analysis: aspirin use was an independent predictor of rebleeding but protected  against major cardiovascular events and deaths

Commentary:

  • A large number of people are on aspirin, up to 50% of men >40 yo. The major GI toxicity is upper GI, which can be decreased by concomitant use of a proton pump inhibitor (though that has its concerns: see http://blogs.bmj.com/ebm/2016/02/25/primary-care-corner-with-geoffrey-modest-md-ppi-use-and-dementia/ which comments on a couple of articles on potential PPI-associated dementia and includes references to PPI-associated microbiome changes, MIs in those without prior history of heart disease, chronic kidney disease, pneumonia and a variety of GI infections, decreased bone density, etc.). But there is somewhat more concern about the lower GI bleeding: several observational studies have documented an increase in aspirin-associated lower GI bleeding (where PPIs are unlikely to be protective, so continued aspirin use is more dangerous). And those hospitalized for lower GI bleeding actually have a higher mortality than those with upper GI bleed.
  • It is hard to draw clear conclusions from a retrospective study, given that the nonusers were older (which perhaps explains their increase in noncardiovascular deaths), though fewer were smokers. So, the multivariate analysis may have not fully compensated for differences in all risk factors/biases.
  • But, I would add a few points:
    • This article highlights the very likely conclusion that those with a lower GI bleed but at high cardiovascular risk are more likely to get benefit over harm by continuing aspirin use. The issue with the more common adverse event of aspirin-associated upper GI bleeding is a bit easier given the potential of adding acid suppression therapy to minimize recurrent risk
    • I think it is also important to stress that aspirin may have real benefits in preventing cancer (see below). The data on this has been accumulating for many decades (I remember seeing data on aspirin or NSAIDs decreasing colonic adenomas/cancer dates back to the 1970s, with the support of a really large number of suggestive epidemiologic and animal studies), though potential cancer prevention was only recently incorporated into mainstream recommendations (see USPSTF guideline link below).

http://blogs.bmj.com/ebm/2015/09/09/primary-care-corner-with-geoffrey-modest-md-low-dose-aspirin-and-colon-cancer-risk/ looks at the data on colon cancer prevention by low dose aspirin with links to other blogs on ovarian, prostate and other cancers.

http://www.uspreventiveservicestaskforce.org/Page/Document/UpdateSummaryFinal/aspirin-to-prevent-cardiovascular-disease-and-cancer?ds=1&s=aspirin is a link to the USPSTF 2016 aspirin recommendations, which highlights using low dose aspirin in the primary prevention of both cardiovascular disease and colorectal cancer (i.e., not just for cardiovascular protection)

Primary Care Corner with Geoffrey Modest MD: Hep C Treatment and Regression of Cirrhosis

16 Jun, 16 | by EBM

By Dr. Geoffrey Modest

A recent study provides up to 12-13 year follow-up showing long-term improvement in important clinical outcomes (liver cancer, decompensation and death) in patients treated successfully for hepatitis C (see doi.org/10.1053/j.gastro.2016.03.036). The value of this study, using PEG interferon/ribavirin, is that it has long-term results which I think are applicable to our newer and much better therapies.

Details:

  • 444 patients with hepatitis C (HCV) infection and compensated cirrhosis were treated with PEG interferon/ribavirin (PEG- IFN/RBV) in Italy, from 2001-2009, with mean of 7.6 years of follow-up
  • Mean age 58, 62% male, BMI 27, ALT 150, bili 1, albumin 4, 27% with diabetes, 83% genotype 1, 26% IL28B SNP CC/74% either CT or TT (those with the CC genotype have a stronger immune response to HCV than the non-CC genotypes and historically have been more likely to be cured of HCV infection)
  • They looked at 2 groups of cirrhotic patients: those without (n=218) and those with esophageal varices (EVs) (n=226) on baseline endoscopy
  • Routine ultrasounds were scheduled every 6 months, and EGD every 3 years for those without baseline EV and every 2 years in those with EV. Those who developed larger EVs were put on b-blockers or got banding
  • Results (and, feel free to skip over the mass of statistics):
    • SVR was achieved in 67 patients (30.7%) of those without EV and 41 (18.1%) in those with EV (p=0.003); and more often in those with higher ALT, albumin, non genotype 1, and those with IL28B  SNP CC polymorphism [all pretty much as expected with this therapy]
    • Risk of developing/progression of EV:
      • In those without baseline EV: those achieving SVR had 77% less likelihood of developing EV [HR 0.23 (0.11-0.48), p<0.001], and those developing EV had small ones. The rate was 2.1%/yr for those with SVR and 9.1% without. These curves continued to diverge over the 12-13 years of follow-up
      • In those with baseline EV: no significant difference in developing more EVs. 22% had varices regress, 44% were stable, and 34% had progression
    • SVR was associated with 75% lower risk of developing hepatocellular carcinoma (HCC): [HR 0.25 (0.12-0.55), p<0.001]. In those without EV the HCC rate was 0.7%/yr vs 2.9%/yr if no SVR; in those with EV, the respective rates were 0.9% vs 3.6%/yr. and the severity of the HCC was much less in those who developed SVR. The HCC incidence curves for those with and without EV diverged over time.
    • For liver decompensation (LD): those without EV who developed SVR, none developed LD, though 1.7%/yr did if no SVR. In those with EV with SVR, 2.3%/yr developed LD vs 4.9%/yr if no SVR.
    • For overall survival: 2 of 67 SVR patients without EV died (both from HCC), vs 27 of 151 without SVR (18 from HCC and 9 from LD): 0.4%/yr vs 2.1%/yr (p=0.0035).  in those with EV, 1.8%/yr died in the SVR group and 4.6%/yr without SVR (p=0.005)
  • For those with EV at baseline (vs those without), independent of SVR, were at higher risk for
    • Liver decompensation:  [HR 2.82 (1.73-4.59), p<0.001]
    • Death:  [HR 1.77 (1.12-2.80), p=0.015]
  • In those without baseline EV
    • HCC: 2.9% with SVR died from HCC vs 11.9% in those who did not develop SVR (p=0.03)
    • Liver decompensation: in none with SVR but 7.1% without (p=0.009)
  • In those with baseline EV:
    • HCC: 2% with SVR died from HCC vs 18.4% in those without (p=0.003)
    • Liver decompensation was not statistically significant, though trend to improvement with SVR (12.1% vs 25.4%,  p=0.15)

Commentary

  • There were some limits inherent in this study
    • They excluded patients with any history of decompensated cirrhosis, or those with more advanced EV (moderate to large varices) or on prophylactic b-blockers. Also those with co-infection with hepatitis B or HIV, or prior to getting cirrhosis.
    • They only looked at patients treated with PEG-IFN/RBV
    • There might be inherent biases in this outcome analysis of this prospective review, including a response  bias (i.e., did those patients with SVR by PEG-IFN/RBV differ from those who did not respond because of some unmeasured difference in the aggressiveness of the virus, comorbidities, or intrinsic issues of their immunologic response)
  • BUT, there are, I think, some reasonable and probably generalizable conclusions
    • This study offers us the ability to look at really long-term conclusions about important clinical events after SVR. Although possible, it seems likely that the meds used to achieve the SVR are not particularly important to outcomes years later (and, the effect of any selection bias introduced by those responding to the old meds is unlikely to be significant anyway with the new direct acting antivirals, DAAs, since they pretty uniformly lead to SVR). So, the application of these results to the new era of DAAs seems pretty reasonable
    • It is quite remarkable how well patients did who had initial cirrhosis without EV: basically they do not develop liver decompensation, and only rarely develop HCC. And, they have a very low risk of mortality from liver disease
    • And, for those with EV at the time of treatment, effective treatment significantly reduced the risk of HCC, liver decompensation, and mortality, though there was still a risk of EV progression (though the risk of progression from small EV is still less: those with SVR had progression at an annual rate of 5.9% vs 9.0% in those who did not achieve SVR)
    • But, even in those without SVR, the presence of EVs portended a poorer prognosis, also reinforcing the benefit of earlier treatment
    • For the future, it will be important in this era of DAA therapies that we look at the long-term efficacy of successful hep C treatment in those excluded in the above long-term study, including those with both less advanced liver disease, those with more advanced cirrhosis (e.g. with history of liver decompensation), and in those with co-infection with hep B and HIV

So, based on this:

  • It seems appropriate to aggressively pursue hep C treatment at an earlier stage, and specifically prior to developing significant portal hypertension/esophageal varices. And I would add that the continued risk of HCC in those with cirrhosis even without EVs and the need for future HCC monitoring supports treating HCV prior to developing cirrhosis (which some insurance companies seem reluctant to do).
  • But this study still confirms that even those with more severe cirrhosis/earlier stages of portal hypertension derive major benefit from effective HCV treatment (i.e., still important to treat HCV even in later stages of cirrhosis/portal hypertension)
  • It will be important to develop a granular database in this age of DAA to guide future monitoring. For example, in HCV infection, given the predictable progression of liver disease to fibrosis prior to HCC, are there clear markers where SVR will make it unnecessary to do continued routine surveillance for HCC? For example, is there a stage of liver disease, perhaps some degree of cirrhosis regression or other markers after SVR, when we do not have to continue doing serial RUQ ultrasounds? Or in the group with EVs, continuing to do follow-up EGDs?

Latest from Evidence-Based Medicine

Latest from EBM