Preliminary Results of TORCH trial suggest a greater role for combo therapy in COPD

TORCH stands for "Towards a Revolution in COPD Health" and may well earn the yearly prize for the most self aggrandizing name for a trial . Here is a reference explaining some details of the trial and purported reasons for doing it.

Briefly, this was a 6200 patient trial involving 4 arms one of which was B.I.D.administration of an inhaler with salmeterol 50/fluticazone 500- the others were: placebo, LABA alone and ICS alone. COPD patients had to have a FEV1 less than 60 % to be eligible.

Here is a medical newspaper type writeup of the presentation of the results.

There was an all-cause mortality reduction of 17.5 % which was not statistically significant (p = 0.052). As usual, if one statistical parameter is not what you want, you can always do another which was done. In this case,the other one was the Cox proportional analysis by which the hazard ratio was 0.811 and this was statistically significant at p=0.03 and the COPD death reduction was statistically significant. So does the combo improve survival or not? It seems to depend on what statistical test you think better reflects reality.Within the temple of evidence based medicine, priests may differ as to what statistical incantations best produce the truth.

Life is messy and so often are the results of clinical trials. On balance it does look like the 50/500 combo of a ICS and a LABA seemed to be beneficial but we have not heard the end of the argument over "does it really decrease all-cause mortality". No doubt the statistical brain power that GSK can muster will put forth good argument on the pro side. Already we are hearing terms like "landmark" study" and comments like "new hope for COPD patients". The combo of an ICS and a LABA is already widely used in the treatment of COPD. In COPD, it is standard to use a LABA first and then add on an ICS for more severe disease with the opposite sequence being the rule in asthma treatment.

According to the 2006 GOLD criteria, one should recommend a long acting bronchodilator if the FEV1 is between 50% and > 30 %. This would mean that in the TORCH trial patients were given the combo (ICS + LABA) at an earlier stage than suggested by GOLD.In addition many patients with moderate COPD are given a long acting anticholinergic and later an ICS is added So,in a way the trial differs from usual clinical practice.GOLD 's position is that there is no definitive evidence indicating if a LABA is better or worse than an anticholinergic and with the continuing controversy over the safety of LABAs many doctors may well choose to begin with an anticholenergic particularly since there is now one that is long acting (once a day use), i.e. tiotropium (Spiriva).

My sense of it is that many physicians already would add Advair to Spriva as COPD worsens or for more severely impaired COPD patients and that would be my approach.I doubt the TORCH study will be much a "revolution" though some may add the combo earlier than Stage iii after reading TORCH.

Why can't the practice of medicine be more like a widget factory or running an airline?

Perhaps goaded on- in part by the urging of insurers and third party payers to save money- we are from time to time treated to a commentary by a giant or would be giant of industry or business to instruct the generally hapless medical profession about how to run a railroad.

When the extremely competent and accomplished Andy Grove on Intel editorialized on this topic in JAMA , Roy Poses of Health Care Renewal responded with a post worth reading and which more than suggested that Mr. Grove may not know much of the doctor business.

We are often admonished to emulate the principles embodied in the safety practices found in the airline industry and there is truth to be found there. A recent post by Aggravated DocSurg entitled "Fly me to the OR" is a gem -nothing unique there-which makes the point how little being a surgeon has to do with flying an airplane.

There is no substitute for experience

An editorial by Geoffrey Norman, Ph.D entitled "Building on Experience-The Development of Clinical Reasoning" is found in the November 23, 2006 issue of the New England Journal of Medicine.

I have written before on the importance of experience to gather up the particulars necessary to begin to be an expert. Dr. Norman's editorial is a pleasure to read because his views coincide with mine.

He points out that in regard to the nature of expert clinical reasoning there was-for a while- a school of thought that posited that it involved the acquisition of some general problem solving skill. Apparently this approach was wrong. He speaks of "content specificity" which means that success in problem solving was strongly related to have the right kind of content knowledge.

Basically experts have to know their subject matter. Here is a key quote from his editorial:

"The process of pattern recognition, so characteristic of an expert's approach, is a product of extensive experience with patients overlaid on a formal knowledge background."

He continues:

"...trying to teach or evaluate clinical problem solving or clinical reasoning skills is quixotic. Knowledge counts."

It is all about practice and experience and previously I wondered what the consequences are/will be of the time shortened internal medicine training program. Will the novice internists leave their training programs with the expertise needed to qualify as even a rookie expert?

Last year,an Annals of Internal Medicine article highlighted a case of TB that was very badly mishandled in a teaching hospital. The emphasis in the discussion-inappropriately in my view-was on a systems approach fix. What the problem was that the pattern recognition skills of the house officers and apparently the radiologists were seriously lacking.

Still more on the SMART trial and the Salpeter Annals article

I have written several times before on the SMART trial and the meta-analysis (MA) written by the Salpeter family team in the Annals of Internal Medicine and suggest that the former might serve as a teaching tool on how not to do a clinical trial and believe the latter could be instructive to those who wish to publish a flawed and opinion laden meta-analysis.

Letters to the editor of the Annals and now a more formal rejoinder has been published-all very critical of the MA and defending the use of a long acting beta agonists (LABA) in conjuction with a inhaled corticosteroid (ICS) in asthma. This article and a 2005 Cochrane review
both provide good data and analysis that lead to the conclusion that LABAs when used in conjunction with ICS in asthma leads to better control, fewer exacerbations and not only does not pose the risk claimed by Salpter but represents the standard of care for patients with more severe asthma.

Mark Twain or Bismark or someone supposedly said that there are two things you should never watch being made- a law and sausage and it has been suggested that MAs should be added to the list.Yet accepting the results of a MA without knowing how it was really made is an act of faith which supposedly we decry in the "age" of evidence based medicine.When your MA turns on the results of one large trial and that trial is seriously flawed the MA is worthless and potentially harmful which is what happened with the SMART trial and the Salpeter Meta-analysis.

Hopefully,the recent Annals article will clear the air. The above cited reference regarding sausages also makes the important point that MA s should include- along with experts in the methodology used- subject matter experts. Some MAs that I see seem to be written by authors who are excited about their meta-analytic skills and seem to believe they can analyze the forests so well that they need not bother to ask a tree expert for input.

Even the PPD may be replaced?

Nothing seems secure from progress or change or innovation.The Mantoux skin test has been around for over 60 years and is now being challenged by two commercially available tests that measure the blood levels of interferon-gamma release from sensitized T cells after stimulation by antigens fairly specific ( apparently there is some sharing of antigen with M. Kansasii) to Mycobacterium tuberculosis.

The QuantiFERON-TB gold Assay has been approved and in 2005 CDC recommended its use in all situations where PPD has been used. The results can be obtained quicker and the assay kits are said to have greater specificity and equivalent sensitivity.The blood must be received in the laboratory within 12 hours, which can be a drawback.The test is to be used just as the PPD is used.A good review of particulars is found here. CDC is recommending its use as a replacement not as an addition to the PPD. The in vitro assays should eliminate the problem of the booster phenomenon which in the past has lead to pseudo-mini-epidemics of TB in hospital personnel, described here, and the vagaries of interpretation of the PPD in a person who has received BCG.

The other assay is called T-Spot.TB.Here is a interesting comparison of how results from the two tests may differ.

Surgical versus nonoperative treatment of herniated lumbar disk-a randomized trial

The SPORT trial results are published in the November 22,2006 issue of JAMA. The Spine Patient Outcomes Research Trial was a 501 patient randomized trial involving 13 centers over a 4 1/2 year period. All patients had imaging confirmed lumbar intervertebral disk herniation with persistent symptoms and signs lasting for at least six week.Surgery was a diskectomy with at most a small portion of the superior facet being removed and a disc fragment removed and the nerve root decompressed.

It is generally agreed that the appropriate way to analyze results in a randomized trial is by a " intent-to-treat" analysis. However, in this case because there was such a large percentage of patients who crossed over to the other treatment arm and a significant amount of missing data that the intention-to-treat analysis was not informative about which approach was better and in the words of the authors "conclusions about the supriority or equivalence of the treatments are not warranted based on the intent-to-treat analysis alone". However, it did show small and non statistically significant advantages to the surgical approach for most measures with a statistical improvement in sciatica.Here is an excellent discussion of intent to treat analysis and the traps involving in attempts to consider the problems of lack of adherence and loss of data in a clinical trial from Dr. Gerald Dallal,a Yale epidemiologist .

When one analyzes the as-treated groups there was a definite advantage to surgery.But the validity of this conclusion is clouded by the concern about confounders as the two groups were no longer randomized. So it seems that we cannot know if surgery is better or not regardless of which analytic technique is used. If you do look at the groups as treated there are much larger effects in favor of surgery which did not disappear after correction for recognized covariates.

It seems that at the end of the day this fairly large, multi center multi year trial did not provide the answers to the questions for which the trial was designed. So what is next?

One of the two editorialists in the same issue of JAMA, Dr. David R. Flum, believes the only way to answer the still unanswered questions raised in SPORT is to have a randomized, placebo controlled trial. Placebo control in this sense means sham operations. The authors of SPORT ruled out sham operations because they believed subjecting a control group to general anesthesia with its attendant risks was not ethical. Flum disagrees. He says that sham procedures would be ethically justified on the "question of community exposure to an invasive, high risk procedure with associated risk ". I think he is saying that since large numbers of folks are"exposed" to the risk of diskectomy a sham controlled RCT would be justified to learn the answer to the question, " Is the procedure justified?" This, I think, is a type of public health style justification implying that diskectomy is some sort of risk that people are exposed to against their will as opposed to a decision made by the individual patient to undergo the procedure.

I realize that sham surgery RCTS have been beneficial in the past at times showing the lack of value of certain operative procedures, internal mammary ligation for angina for one, but the dramatic improvement ones sees when someone with severe pain, and objective weakness, improves immediately after surgery is hard to attribute to a placebo effect particularly when the surgeon sees a nerve root being compressed and relieves that compression.

This and other trials have shown that after 2 years there may be little difference in the outcomes in the surgery and non surgical groups but when you have unrelenting neuropathic pain-for which typically usual analgesics work poorly-the option of quick relief is something I would opt for. In the long run the two approaches seem very similar in terms of outcome but remember what John Maynard Keynes said about the long run.

Do ABIM and ACP differ regarding P4P?

The American College of Physicians (ACP) has supported P4P. An organization,the American Board of Internal Medicine,frequently sees eye to eye with ACP. Apparently this is not the case in regard to P4P according to a recent letter to the editor published in the November 2006 issue of "ACP Observer". Dr. Christine Cassel is president and CEO of ABIM and says the following in her letter:

"ABIM does not 'support' pay for performance."

"... pay for performance is a strategy that has yet to prove itself as a mechanism to improve quality of care"

The apparent link from ABIM to P4P may lie in the fact that physicians who are taking part in the ABIM's " Maintenance of Certification" program can receive credit from many health plans and that credit could be linked to a P4P arrangement. So it seems like ABIM is sort of working with insurance plans on activities tied to "incentive" programs but do not support P4P and consider it a unproven way to improve quality. I can see why Dr. Cassel saw fit to write a letter to explain the situation because some would consider their arrangement with health plans "support".

Do thiazide diuretics cause "benign" diabetes?

That thiazides precipitate "benign"diabetes is, in part, the argument made in the recent issue of the Archives of Internal Medicine in an editorial accompanying still another analysis of data from the ALLHAT trial.This study was a post-hoc analysis with a followup time of five years in which three groups were compared-those on a diuretic, those on a calcium channel blocker ( CCB) and those on an ace inhibitor (ACEi).

Those taking a CCB or a ACEi were statistically less likely to develop diabetes than those receiving a thiazide. Strong advocates of ALLHAT's preference for thiazides have put forth what I consider to be a strange argument that I think goes something like this; Yes, thiazides cause more diabetes but it doesn't seem to alter the outcomes. This they say because the various post hoc analyses fail to show a mortality excess. But the time frame of these studies is short-five years in the current Archives article with an average time of followup of 3 years- and are post hoc, sub group analyses which in the catechism of evidence based medicine are not very high on the evidentiary pecking order. The absence of proof of an effect is not the same as proof that the effect does not occur.

Although the editorialist repeatedly admonishes the readers with references to adhere to the principles of evidence based medicine I think it takes quite a leap of faith- not reliance on evidence, which so far is inadequate- to accept the notion that drug induced diabetes is harmless and somehow the patient with diabetes precipitated by thiazides is immune to the ravages of micro and macro vascular disease. While it is possible that the elevated blood sugar in the thiazide treated patients does not represent the disease that we designate as type 2 diabetes the burden of proof lies with the moving party i.e. the one saying this type of "diabetes" carries no cardiovascular or renal risk. The short period of follow up is a major weakness in the study particularly in regard to end stage renal disease but even CV effects may occur only after prolonged periods of elevated blood sugar as was the case in the Diabetes Control and Complications Trial.Further, nearly half of the study group did not have fasting glucose levels measured.

The patient with the metabolic syndrome which we think is driven by high insulin levels secondary to resistance to insulin in various tissues may also be the person with decreasing numbers of pancreatic beta cells and is already close to having elevated fasting blood glucose levels and have his glucose pushed up a bit by thiazides. It is hard to believe his risk of CV disease over the long run is not going to not be elevated. The deleterious processes of type 2 diabetes are at work for years before the fasting blood glucose becomes elevated.I have felt uneasy about prescribing to the diabetes-waiting-to- happen- patient a drug that is well recognized to increase the risk of diabetes.

In spite of the valid criticisms of ALLHAT's original design and its lack of correspondence to real life treatment of HBP we continue to see more re-analyses of the data that was flawed to begin with.

And the number one reason to oppose P4P- It is unethical

Dr. Edmund Blum, an internist from Brooklyn makes the argument that pay for performance (P4P) involves a "irresolvable conflict " with the ethical standards of the medical profession.

His persuasive arguments can be found in the November 6,2006 issue of American Medical News (subscription required) in their "Professional Issue" section.

He says that P4P rests on 3 flawed premises or fallacies the most important of which is that P4P is consistent with medical ethics. He argues that it is not. (The other 2 fallacies are:P4P rests on a valid statistical foundation and P4P will improve the safety and quality of patient care)

I quote;

"[medical] standards derive from a core of fiduciary responsibility, in which one person, the patient, depends on the superior knowledge and skills of another, the physician, and places complete confidence in that person in regard to a particular transaction-in this case, medical care."

"The fiduciary is held to a higher standard of legal and moral conduct and trust than a stranger or a business person...[This] obligates the physician to do his or her best for the patient regardless of reward.The duty goes beyond the 'due care' standard or tort law to a higher level of loyalty and commitment that is not contingent or rewards or penalties."

The idea of P4P involves an assumption that "the fiduciary relationship is insufficient motivation for the physicians to do their best."

To accept P4P is to accept the notion that physicians have not already been obligated to do their best for the patient and to place patient welfare above financial rewards and that they have to be giving a tip or a bribe to do their job. Dr. Faith Fitzgerald was on target when she said

" We must not servilely accept gratuities for doing our duty."

Forty years ago,I began the transformation from a lay person to a physician. Part of what was branded in to my limbic cortex in that years long process was the responsibility physicians have for their patients, a responsibility to do what is right for the patient,a responsibility to place their welfare above personal financial concerns. The acceptance of P4P is so antithecal to that tradition that I cannot believe some professional organizations of physicians are supporting it. It seems to me that support and advocacy for P4P is equivalent to saying the ethics and culture of physicians are not adequate and to provide good clinical care it is necessary for third parties to proscribe behavior and reward and sanction accordingly. To sanction such thinking, in the words of Dr. Blum, is to "push us farther down the slippery slope to deprofessionalization".

Physician group sues insurance company for defamation

According to an article in the November 1, 2006 of Internal medicine News (www.internalmedicine.com) six Washington state physicians and the Washington State Medical Association have filed a suit against Regence Blue Shield. This insurance company notified some 500 physicians that they did not meet the "standards" and were dropped from the network. But they went further- they then contacted 8,000 of those physicians' current patients and informed them that the docs did not meet quality standards basically claiming they did not practice quality medicine.

Attacks on professional integrity or on a professional person's integrity in some jurisdictions may be considered "per se" defamation. I had written before about the Washington state situation and wondered then if there might be a case of legal action and it looks like there may be and I applaud the Washington docs for taking action doing what they can to protect their rights and level the playing field .Other state medical associations have had some success in battling the big insurers, witness the Texas Medical Association action against several large HMOs under RICO.

The lawsuit claims deceptive trade practices,breach of contract and defamation.

Dr. Gail Wilensky (Ph.D) was also quoted in the article as she continues to play the strings of the support-P4P effort and supported the position of the insurance company saying in part "All data has errors but that doesn't mean the suggested conclusions are faulty "

Dr. Wilensky is often described in news articles as a senior fellow at Project HOPE and a member of the Institute of Medicine panel [onP4P]. Often not mentioned is the fact she is on the board of Unitedhealth Group,holds about $800,000 worth of UNH stock or that she cashed in 1.3 million dollars in UNH options in 2005. (information from Yahoo finance as of 11/11/06).

Annals Internal Medicine article"Former CEO Aetna recommends P4P

Dr. John W. Rowe, who is not without impressive academic credentials, recently retired as CEO of Aetna recently authored a five page article in the American college of physician's' journal advocating pay for performance. ACP's position on this is fairly well known and readers will not be surprised to see an article advocating its implementation. However, it does seem a bit audacious to have the former CEO of a major health insurer to pontificate on the "moral basis for physicians...to support efforts to control costs,improve quality of care and participate in pay-for-insurance initiatives."

Some internists members of the ACP might be puzzled as to how an insurance company executive
becomes qualified to lecture them on morality particularly when that insurance company- as well as others- was taken to court by numerous medical societies charging it with various illegal practices designed to seriously curtail payments to physicians and limit patients access to care.

There is, of course, no love lost between practicing physicians and insurance companies in general ,but at least in my experience, Aetna 's reputation in dealing with doctors is near the bottom. The editors of the Annals seem out of touch with real life practicing internists when they choose a former CEO of one of the least liked insurance companies to promote pay for performance.

NEJM "Perspective"Nov.2,2006:Misleading soundbite for P4P- "Shift from autonomy to accountability"

Dr. Elliot S. Fisher is the author for the NEJM piece "Paying for Performance-Risks and Recommendations" (NEJM 355:18 1845). Dr. Fisher has published many articles documenting the "remarkable variation in performance" by the players in ambulatory and hospital based care so we should not be surprised that he favors doing something about it-namely pay for performance (P4P).

Amazingly, after listing some of the major concerns about P4P he dispenses with them by simply saying that the concerns were discussed in a report by the Institute of Medicine (IOM) but that the IOM committee,of which he is a member" then "strongly recommended moving forward with pay for performance." This is a curous argument, indeed, that lists serious problems with the proposal, offers little reason to accept it (he does say payers are "demanding accountability") and then strongly recommends it. The arguments against P4P that he briefly covers are:

1.Concerns about the underlying goal. He says physicians fear that cost control will be the only focus.

2.Are the [quality] measures adequate? He says in part [medical care] "often requires a careful balancing of risk, benefits and patients' preferences, not rigid adherence to clinical guidelines."

3.Is implementation feasible? He acknowleges that for small office practices "costs will be high"

4.Could there be unintended consequences? Such as avoiding sick or challenging patients. (You think there might be a problem if physicians start avoiding sick patients)

Fisher then points out the funding problem. He proposes that the increased funds to "reward" (aka bride" docs for doing their job would be derived from cuts in the CMS programs so that " some providers would see little or no increase in fees."

Fisher then tells the readers that the IOM committee recognized that the evidence underlying P4P is weak and that unintended effects are possible and therefore the federal government was advised to have an effective monitoring and evaluation system in place to recognize potential harms and correct them. This would be a first- a government system that includes some sort of super-system to monitor itself and make mid-course corrections as it goes.

He closes with the claim that the "shift from autonomy to accountability" seems inevitable. This misleading and fradulent semantic ploy is reminiscent of the bogus term "managed care". Of course, physicians are already accountable-to their patients as well as to medical boards of examiners just as care was managed by physicians long before third party payers used the managed care mantra to cut costs.

Thoracic imaging for lung cancer screening-here we go again

In the October 26,2006 issue of the New England Journal of Medicine we find a clinical study and an editorial about lung cancer screening using CT which should rekindle the decades old arguments about this topic. ( Survival of Patients with stage I Lung Cancer detected on CT Screening, NEJM 355;17, p. 1763). So high profile are NEJM articles with pickup and amplification by the news media that I will not be surprised when smokers and possibly folks who worked around asbestos and possibly other lung carcinogens will be asking their physicians for chest CTs. A brief review of the current study is found here.

A convincing advocate of the value of imaging screening for early lung cancer has been Dr. Gary M. Strauss. Some of his views can be found here. When we talk about screening we have to talk about the difference between survival rate and mortality rate.The latter is defined as the total number of deaths from the disease in question divided by total number tested. Survival rate is defined as numbers of survivors after some time period divided by total number diagnosed with cancer. If the screening technique were to detect significant numbers of indolent cancers then the survival rate might appear to be improved after the institution of the screening test while the mortality rate might be unchanged. Prostate cancer screening with periodic PSA measurements is sometimes accused of being an example of that. After spending a few decades in the pulmonary disease business I was not impressed with the large number of indolent or clinically insignificant lung cancers.Every pulmonary doctor remembers the occasional case of cured lung cancer that happened to fortunately be detected by a chest xray done for whatever reason. The coin lesions (less than 3 cm by definition) have a much higher cure rate than lung cancers as a whole. All of that leads to the intuitive appeal l (or maybe just hope) that if we could come up with a way to catch lung cancer early the current rather dismal survival rate of lung cancer would improve.

Conventional wisdom contains the nugget that in regard to screening one should use the cause specific mortality as a measure of efficacy not the survival over a given time period.. Strauss has taken the opposite view. More of his thoughts can be found here and here and here.

This brings us to the current study,I-ELCAP aka The International EarlyLung cancer Action Program. It is survival rates that are emphasized in this study (so the issue of lead time bias has to be raised) and the numbers seem impressive. The study is very large with over 31,000 asymptomatic persons at risk of lung cancer being screening with low dose spiral CT and then evaluated with a detailed protocol that utilized followup CTs, PET scans and skinny needle biopsy. They report a "estimated 10-year survival of 88% in the subgroup with clinical stage I lung cancer"

A great deal has happened since the early chest x-ray lung cancer screening projects. We have spiral CT, PET scans and skinny needle biopsies.Perhaps we can now detect lungs cancer early enough (that is small enough?) to remove them while they is still time. Before I reviewed the article I had assumed they were talking about non-small cell cancers (NSCLC) since the small cell variety seems to be another animal entirely. However, no mention is made on survival for each cell type or any indication that they were managed differently and there were 7 small cell cancers detected on the annual screening.Were they resected also? Is it possible that we can actually detect and remove small cell lung cancers that are so early they have not spread? In fact, there are some data indicating long term survival for small cell lung cancers treated with resection.

"Marketing strategies masquerading as Evidence Based Medicine"

A commentary appears in the October 19,2006 issue of the New England Journal of Medicine regarding allegations about the length to which a drug company will go in influencing patient care by misleading physicians and improperly manipulating the creation of treatment guidelines. The commentary in the Perspective section is entitled " Surviving Sepsis-Practice Guidelines, Marketing Campaigns, and Eli Lilly".The authors are all from the Critical Care Medicine Department at NIH. A summary of their article can be found on the October 19 post on Health Care Renewal.

Xigris (drotrecogin, aka recombinant human activated protein C,rhPAC) is the drug, Lilly is the drug company and the condition for which treatment guidelines were said to be manipulated is sepsis. I have been concerned for some time about the degree to which Big Pharma have used ( and mis- used) the concepts of evidence based medicine to promote various medications but if the commentary accurately reflects reality this example rivals or even exceeds the gabapentin story.

This NEJM article and an earlier article in the Annals of Internal Medicine dealing with gabapentin and the disreputable techniques used to market it ( I posted about it here ) should be mandatory reading for medical students and included in the reading lists of med school courses on EBM and how EBM can be kidnapped and exploited . If the situation is, in fact, as described in the NEJM commentary, egregious is not strong enough an adjective for that type behavior. But "If" is the operative word. For a thoughtful counterpoint to the NEJM commentary to to the Oct 22, 2006 post by Dr. RW. He reminds us that criticism of guidelines or for that matter any treatment should not be based simply on the fact that a drug company may have used various techniues to promote it but should be based on logic and reference to solid evidence and in his analysis of the NEJM article the authors from the NIH fall short in that category.

No one is shocked by the fact that businesses not infrequently promote their goods or services by emphasizing the advantages and minimizing the downside of their products. However, if and when that spin puts patients well being or lives at risk I believe we move past "mere" unethical behavior. There is little room for error in the treatment of sepsis patients and underplaying or ignoring the risk of hemorrhage from activated protein C (Xigris) could well lead to patient fatalities. (The authors assert that the risk of bleeding was not properly noted in the promotional material generated for Xigris). To withhold information about serious side effects from a medication and promote its use by deceptive means included sponsoring guidelines may well move past negligence as well.

There is more .Go to the October 20,2006 post from Health Care Renewal for expression of concern about the role another drug company, Amgen, played in the development of guidelines for uses of Epogen in renal failure.

I do not know if and/or to what degree the experts who authored the two guidelines mentioned above were actually improperly influenced by the drug companies' activities and/or recommended treatments for which the evidence was insufficient but the perception that guidelines might be "usurped ...for commercial purposes" has to make physicians even more skeptical of guidelines in general particularly with such a commentary in a high impact journal. Further, when you consider the uses to which guidelines are put,including quality audits,pay for performance,arguments in legal proceedings it becomes even more important that we know how the sausages are made. And as the NEJM article asserts Xigris does seem to have found its way into performance standards even though I believe it is fair to say that the jury is still out ( or should be) regarding its efficacy and safety.

Basis for "Treat to goal" for cholesterol is questioned by Annals Internal Medicine review

A thought provoking-and in a way troubling review- in the October 3, 2006 issue of the Annals of Internal Medicine ( "Lack of Evidence for recommended Low-density Lipoprotein treatment Target: A solvable Problem" by R.A. et al Haywood) was highlighted on October 17,2006 by DB's Medical Rants and by MEDPUNDIT. It questions the evidentiary basis of NCEP's 2004 "treat to goal" set of recommendations. Haywood's article is instructive because of the thoughtful analysis of the data linking cholesterol level and response to statin therapy and outcome and troublesome because it raises doubt about the validity of adherence to the NCEP recommendations of treating to goal. I 'll have to admit that I accepted those targets as a reasonable thing to do leading to at times increasing the statin dose and sometimes adding ezetimibe. In 2004, the NCEP expert panel recommended that physicians treat patients at what they designate at "very high risk" for coronary events to achieve a LDL cholesterol of less than 70 and for those patients judged to be at " high risk" a value of less than 100.

The review's major point is this:

High quality data is lacking to provide basis for the recommendation to titrate lipid lowering therapy to LDL targets or to prove that such therapy is superior to simply prescribing doses of statin drugs used in the clinical trials for patients at high and very high cardiovascular risk.

It should be noted that the authors are quick to point out that they are not saying that there is strong evidence against the current recommendations.

The reply from Dr.Scott Grundy and the NCEP folks likely will be interesting. (I assume they have to reply to this). As thoughtful as this article is you have to wonder how much of an impact it will have.The cardiologists and a number of primary care doctors seemed to have accepted the lower goals rather widely and that concept may be a hard one to get back into the barn. At this point I do not know if we should try to or not.

More bad news regarding second generation antipsychotics.

The October 12, 2006 issue of The New England Journal of Medicine published an article regarding the use of the second generation antipsychotic drugs (SGAs) in the management of aggression and agitation and psychotic behavior in dementia. These drugs are widely used for this application although it is not approved (by the FDA) for that use and in fact there is a "black box" warning regarding increased risk of death in older patients with dementia.

Three drugs were compared with placebo in 421 patients in a multi center study; 1) olanzapine (Zyprexa), 2) risperidone (Risperdal) and 3)quetipine (Seroquel).

This trial differed from the typical efficacy-safety RCT done by drug companies as it looked at a "real clinical life" end point of time of discontinuing the medication because of any reason. The other primary outcome was the number of patients who had a minimal improvement a clinical behavior scale.

The authors concluded that the three drugs were more effective than placebo but the incidence of side effects limited their use. As seems to be more and more the case in clinical trials, there are so many comparisons made and often with rather arcane statistical tools it is difficult to know what to conclude. For example, Zyprexa was significantly better than placebo with the "Cox model" but not when compared with placebo with the "Hockberg adjustment" for multiple comparisons.Apparently this adjustment is an alternative to the Bonferroni technique to decrease the number of "false positives" when multiple comparisons are made. But how do you decide which technique to use-in this case the resultant answers seems 180 degrees apart.

Although the headline news- sound bites about this article may claim these drugs were useless, that characterization seems too simple.They can help control the symptoms of interest but often have to be discontinued because of side effects . Even the authors seems a bit ambivalent in their comments about the results;

"...our findings suggest that there is no large clinical benefit of treatment with atypical antipsychotic medications as compared with placebo."

They also say:

"Although the atypical antipsychotic drugs were more effective than placebo, adverse effects limited their overall effectiveness."

My take on all of this is these drugs may help a bit in the control of agitation and aggression in dementia patients but in a significant number of patients side effects lead to their discontinuation. Certaintly the exuberant enthusiasm driven in no small measure by drug company hype is waning. These drugs are not nearly as good as the efforts to promote them suggested. It would have been interesting and perhaps instructive for Haldol to have been included in the drugs that were compared in this study as for years it has been the stand by drug in difficult situations with dementia patients with aggressive behavior.

Answering services should not make it hard to talk to physician

A recent article in the September/October issue of the Journal of the American Board of Family Practice by David Hildebrandt called attention to an issue with some answering services techniques that serve to prevent patients from contacting their physicians. By simply being asked "Is this an emergency?", many contacts with the physician are eliminated. Patients often call their doctor because they do not know if the issue is serious enough to be considered an emergency or not. This "filtering" technique does not serve the interests of the patient.Procedural barriers limiting contact with doctors cannot be in the physicians' s best interests either.

The survey admittedly was small, only 35 physicians offices were contacted and of those 14 used answering services and 9 of those asked the patients to decide if their call should be fowarded to the doctor. The small sample size precludes robust conclusions about how widespread the practice might be. An larger earlier study by the same lead author involved 91 primary care offices and in 55 the answering services "forced" the patients to decide if it was an emergency requiring a call back from the physician. Clearly ,this is not a good practice but I have encountered worse. I have attempted to contact physicians after hours, and sometimes on Friday afternoon and been unable to contact them at all or anyone providing coverage. The answering machine-not even a human- informs the caller what the office hours are and that if they have an emergency they should go to the nearest emergency room.

Another approach is the nurse telephone triage. While this is better I have some uneasy feeling about this as well. When you get old and cranky you tend to think if things are not done like you did them they are off base. When I was in private practice all calls were referred by the answering service to either the patient's physician or the on call doctor.

Dr. Bruce Bagley,the medical director for quality at the American Academy of Family Physicians, is quoted in the American Medical News story about the article:

"You want the highest level clinical person determining what's an emergency, not a person at an answering service who knows nothing from nothing."

Is Merck gearing up for "son of Vioxx"?

The editorial commentary by David Graham in the October 4,2006 issue of the Journal of the American Medical Association ("COX-2 Inhibitors,other NSAIDs and cardiovascular Risk,The seduction of Common Sense" vol. 296,no 13 p 1653) is very critical in regard to both Merck and the FDA. If Dr. Graham's analysis is correct Merck is already cooking the books to get a Vioxx like drug approved by the FDA. Admittedly, this is Dr. Graham's analysis of the pre-approval activities of Merck and their version of those activities is likely to differ significantly from his comments and for those we will have to wait a while.

Here is what he said- Merck has announced they will proceed to get approval for etoricoxib,a new COX-2 inhibitor. They will rely , in part, on the results of the MEDAL trial in which etoricoxib was compared with diclofenac and found to demonstrate that the cardiovascular event risk was the same with either medication using a "noninferiority study" design, which according to Graham, is "especially poor" at identifying risk between drugs. Further, the comparator drug was diclofenac which apparently has been associated with an increased risk of c-v events. By inference, he argues etoricoxib must also increase c-v risk.

Graham then says;

"This veiled and misleading ambiguity has much in common with the stratagems used by VIGOR and APPROVe,where the true results were opposite to those claimed and promoted."

There are two articles regarding COX-2 inhibitors in the same issue, one of which demonstrates an increased risk of cardiovascular events with diclofenac.One is a meta-analysis of randomized clinical trials (RCTs) dealing with renal effects and cardiac rhythm problems and the other a systematic analysis of observational studies. It is the latter which provides the following:

The relative risks (RR) of cardiovascular events is elevated with lower and higher doses of celecoxib-1.33 for the 25 mg/day dose and 1.64 for the greater than 25 mg /day dose.

The highest RR reported is with diclofenac at 1.50 while no increased risk is noted with naproxen and ibuprofen but idomethacin's RR is increased at 1.36.(Ref.Cardiovascular Risk and Inhibition of Cyclooxygenase,McGettigen P and Henry,D JAMA vol 296,no 13,p 1633)

The authors make an important comment :

"Typically, in pharmacoepidemiological studies there is reluctance to accept as causal RR estimates much below 2" This is because this type of study is subject to various biases but, predictably, the authors still believe the demonstrated association "are real". (When do authors not believe what they find is true?)

Here is a related quote form David Sackett et al in their second edition of "Evidence Based Medicine" (Churchhill Livingstone reprinted 2001, p.163)

"How Big should relative risk and odds ratios be before we should be impressed by them? ....We might not want to label an odds ratio from a case-control study as impressive unless it is greater than 4...in cohort studies..we might regard a relative risk of greater than 3 as convincing for more serious adverse events"

This systematic analysis utilized data from both case control and cohort studies so somewhere between 3 and 4 might be the threshold for concern according to the Canadian gurus of Evidence based medicine and these were all under 2.

Even though one can argue about causality and relative risk level this issue of JAMA will do little to encourage the use of COX-2 drugs. Graham's suggestion of using naproxen (or ibuprofen) plus a proton pump inhibitor (PPI) as an alternative makes sense and is what I was recommending for the last several years.

Second generation antipsychotics-efficacy versus effectiveness

In the 1960s, the phenothiazines changed the face of psychiatry when chlorpromazine was shown to be effective treatment for schizophrenia. These first generation antipsychotics (FGAs) are associated with major side effects-namely acute extrapyramidal symptoms and tardive dyskinesia. So when the first of the second generation antipsychotics(SGAs),clozapine,was approved by the FDA and it seemed to be less likely to cause these very troublesome side effects another new era in psychiatric therapeutics seemed to be launched. Other SGAs were developed and approved and were widely accepted and generally believed to not only be more efficacious regarding the so-called negative symptoms of schizophrenia but safer and capable of inducing a better quality of life.The evidence that clozapine did in fact produce superior results in symptom reduction in patients resistant to other drugs is convincing: the question seems to be are the other drugs (five have been approved in the last ten years) in the second generation category also superior.

In 2005 and 2006, two clinical trials (CATIE and CUtLASS1) were published which have raised considerable doubt about the alleged superiority of these SGAs. (Thanks to PHARMAGOSSIP for the reference.)

Dr. Jeffery Lieberman from Columbia Psychiatry Department published an excellent commentary on this issue and his article is available on line full text from the AMA site (go to "Newsroom" and then to "Publications" for the comments found in the October 2006 issue of the Archives of General Psychiatry.Also full text free links to the Studies are found in his reference list)

The issue of the relative value of the FGAs and the SGAs is important per se.Dr. Lieberman's comments not only address that but also the broader issue which can be stated as follows:

How can the following happen-A medication is approved by the FDA based on Phase 2 and Phase 3 RCTs, becomes widely accepted largely replacing the older drugs with that application and then following more Scrutiny and analysis is found to not be any better than the drug(s) it replaced?

He suggests two reasons"

1.The traditional efficacy-effectiveness gap. Things do not always work out the same in the helter-skelter world of clinical medicine as they do in the sometimes cherry picked world of randomized clinical trials. The second Gaultian axiom of evidence based medicine is that "Treatments do not work as well in the clinical practice as they do in randomized trials and they cause more problems".(The first axiom is "The basic fact of clinical trials is that everyone does not respond the same to a particular treatment and almost no one has the average effect")

2.Claims of a drug's superiority were "greatly exaggerated". Drug company hype and overt and sometimes covert promotional activities certainly play a role as does (and this is my contribution to that reason) the sincere desire of physicians to be able to have better tools and be better able to treat their patients. In other words, docs yearn for better drugs and sometimes overlook the weakness of the evidence that is presented.

There will be much more written about whether FGA or SGA are better or safer and neither Dr. Lieberman nor I claim to have the final word.But if there is a lesson here I think it is that because of the efficacy-effectiveness gap a RCT (or even several RCTs) is/are just the beginning of the process of deciding what to do for a given patient; it is not the determining factor and we often cannot really judge the value of a given treatment until there is enough real world clinical experience to see how it really works.

The virtual doctor's lounge-the successor to the now defunct real lounge

I written before about the demise of the doctor's lounge in the hospital-the previous site of free coffee, curb stone consultations, physician networking and the sharing of common shared interests and experiences and a chance to vent about whatever. My regular 4 or 5 readers will not be surprised that I blame this on managed care.

In a way, medical blogs have become a surrogate for this experience for physicians many of whom have significantly less face time with other physicians and according to one surgeon blogger even telephone time has become less common as more and more PAs and NPs are delegated the role of calling the consultant.

Today, for example, in this virtual lounge I learned useful information about neuropathic pain from Doctor RW and was reminded by DB about the importance of time in the context of taking it to explain prescription medication to your patients.Memories of 40 years ago were recharged by Dr. Schwab in his comments about the medical drama of a surgeon opening a chest in the emergency room. Medpundit gives me a chuckle when she relates what British soldiers in Iraq think of the British NHS when they claim they would rather get shot in the head and get to go to a great US military hospital or receive a less serious wound and end up in the NHS. As usual the pulmonary docs at their site present fascinating cases.

The medical web does provide some of what we had at the lounge (we can certainty vent 24-7) and in some ways much more in terms of connectivity with information but I think how great it would be if had both.Sitting down over coffee and discussing a difficult case with a respected colleague is something many of us miss.