Monday, June 22, 2020

Lowering cholesterol? How, why?

This post was suggested by a reader, who had read enough about the coronavirus, and it is a topic that merits attention. Cardiovascular disease remains by far the number one killer in the developed world and is rapidly rising to that spot in the developing world as well, and our ability to prevent heart attacks is in large part related to lowering cholesterol.

The first point to note is that cholesterol is not arsenic – cholesterol is part of normal cells, and no one dies of cholesterol. It has, however, been known for many decades that an elevated level of cholesterol, along with high blood pressure and smoking, increases the risk of developing coronary disease. Norwegian researchers in 1939 described families with abnormally high levels of blood cholesterol and premature heart attacks. This led, beginning in the 1950s, to attempts to lower cholesterol with diet and medication. Most of the medications which were then available, which included nicotinic acid, fibrates and cholestyramine, were not very effective and/or were difficult for people to take for any time, and their use did not reduce the risk of heart attack by a large amount.

A breakthrough in cardiology occurred with the development of the class of drugs technically known as HMG-CoA reductase inhibitors and widely known as “statins.” The first in this class, lovastatin, was discovered in 1978 and first tested in humans in 1980. Multiple similar drugs followed, and we now have seven on the American market. They vary in potency, but all work in a very similar way. The first major trial of using a statin to prevent CV death was conducted in the west of Scotland, among men at high risk of heart attack. Participants were enrolled between 1989 and 1991 and the results of the trial were published in 1995. Those taking the active drug, which in this trial was pravastatin, had about 31% fewer heart attacks, and there was about a 9% reduction in total mortality (from 38.6% to 34.7%) over an extended 20-year follow-up. Multiple later trials, done in lower risk patients, all showed similar results: fewer heart attacks and a trend to lower overall deaths in statin-treated groups.

I do not own stock in any healthcare companies, so have no conflict of interest in stating that the statins as a class are one of the best medicines in our armamentarium. Most are now available generically and are dirt-cheap in the big picture of pharmaceutical products; they are well-tolerated by most people; and have very few bad side effects. There was a flurry of concern a few years back about them contributing to dementia, but this has been disproven by many large studies. They do seem to slightly hasten the onset of diabetes in people prone to it, but the benefit far outweighs the risk.

There is one fly in the ointment: many people develop muscle aches from taking a statin. Only a tiny number have muscle damage (which can be measured with a blood test), and in my opinion if you fall in that tiny number you should never take any statin, but 5-10% complain of muscle aching that goes away when you stop the drug. In some cases, taking a different statin can be done without the problem recurring, but some get the same reaction to every statin they try. What options are available if you are in this group?

The first step is to decide with your doctor if you truly need to have your cholesterol lowered. While statins seem to have a similar relative reduction in heart attack risk in all groups, there is a difference between relative and absolute risk. Let’s say that taking a moderate dose of a statin reduces your risk of having a heart attack in the next 10 years by 25%. If that risk is 40%, then reducing it by 25% means you have reduced your odds by 10%. If that risk is 4%, reducing it by 25% means your actual benefit is a 1% risk reduction – perhaps not worth it in your mind. There are on-line risk calculators that will help you calculate your personal risk.

If you feel that lowering your cholesterol is important and cannot take a statin, there are several choices available, but NONE of them has been studied for risk reduction when taken alone – all have been shown to lower your risk of heart attack when added to a maximally-tolerated dose of a statin.

Ezetimibe (Zetia) inhibits absorption of cholesterol from the intestine and reduces cholesterol by about 18% (compared to 25-60% reduction with a statin). There are two new injectable drugs, called PCSK9 inhibitors, that are very potent, but are extremely expensive and require self-injection every 2-4 weeks. Unless you have a very high risk, your health insurance company is likely to balk at paying for these. The newest entry is bempedoic acid (Nexletol), which also lowers cholesterol by about 18%, and it has also only been tested as an add-on.

Finally, do not forget lifestyle changes: a better diet, weight loss, moderate aerobic exercise. If you are a smoker, quitting will do more to reduce your heart attack risk than any pill.

Prescription for Bankruptcy. Buy the book on Amazon

Saturday, June 6, 2020

Health care: black and white

The Covid-19 pandemic has brought into sharp relief the failings of the U.S. health care system to give adequate priority to public health: our woeful performance in testing, contact tracing and availability of personal protective equipment. It has also highlighted the marked social disparities in how we deliver health care.

Across the United States, the death rate from Covid-19 per 100,000 among non-Hispanic whites is 22.7, among Asian-Americans 24.3, among Hispanics 24.9 and among African Americans 54.6. The strikingly higher death rate reflects many factors. Blacks are more likely to live in crowded housing, making “social distancing” more difficult; they are more likely to have to depend on crowded public transportation; they have a higher incidence of such medical conditions as hypertension and diabetes that increase mortality; and they are more likely to work in such “essential” jobs as meat packing, nursing homes and home health care and grocery stores, and less likely to be able to work from home.

The racial disparity in health care exposed by the pandemic is not new. Overall deaths per 100,000 in the U.S. in 2018 were 725 for non-Hispanic whites and 852 for African Americans. The maternal death rate in this country has been (or should have been) an embarrassment for a long time. While the death rate per 100,000 live births in most of the western world ranges from 5 to 8 and has been falling, in the U.S, during the decade 2007-2016, it rose from 15 to 17. Here too there was a marked racial disparity. The maternal death rate among white women averaged 12.7/100,000 in the most recent figures, while it was 40.8 among black women. Poverty alone is not the explanation; among women with at least a college degree, the maternal death rate of black mothers was 5.2 times that of whites.

Too often, minority populations live in the least safe housing, are less likely to have health insurance and have less access to health care.

As detailed in Prescription for Bankruptcy, some 80% of a nation’s health, as measured by factors such as our life expectancy at birth, is determined by social factors rather than the traditional health care delivery system. To lead healthy lives, people need safe housing, access to healthy food, green spaces, educational opportunity, social support and, most important of all, a living wage.

Whether through development of a vaccine or the gradual development of effective treatment and herd immunity, the pandemic will end. The racial disparity in health and health care will not unless we make a conscious decision to make societal change.

Prescription for Bankruptcy. Buy the book on Amazon

Tuesday, June 2, 2020

What do Covid-19 tests tell us?

To understand what various tests for Covid-19 mean, you must first accept one important fact: medical tests are not perfect. Even the most accurate tests we have at our disposal may say that someone has a disease when they do not (a “false positive”) or that they do not have the disease when they do (“false negative”). All tests must be interpreted together with the clinical findings and the likelihood of someone having the disease in question.

Let’s look at how Covid-19 is diagnosed. The “gold standard” is detection of the virus’ RNA in specimens from the respiratory tract of a sick person. These tests go by the highfalutin label of “reverse transcriptase-polymerase chain reaction” or RT-PCR, and this describes a technique by which the presence of viral RNA is measured and quantified. This technique can be used on any specimen, but the earliest and still most common specimen has been a swab taken from the very back of the nasal passage. It has also been used on sputum (spit), throat culture, stool samples and even semen.

In most people who develop symptoms of Covid-19, viral RNA becomes detectable in the nasopharyngeal swab as early as day 1 of symptoms and in over 98% within the first week. Detection falls off after week 3, though positive tests have been reported as late as six weeks after symptoms begin. Tests for Covid-19 RNA in the stool and semen have also been found positive long after patients recover. It appears that late tests often do NOT indicate that the person is still contagious, and that they may simply be shedding decayed viral remnants.

The RT-PCR is highly accurate when positive. False negative tests do occur because of technical problems: the specimen may have not been taken properly or not transported properly to the lab, or there may have been a problem with the way the test was conducted.

The other type of test about which you will read is the antibody test. When we are exposed to an infectious agent (or a vaccine), our immune system gears up to fight it off by producing proteins called antibodies that both directly fight the infectious agent and call in various cells to do the same. It takes time for antibodies to be produced, so they are unlikely to be found very early in an illness. In the case of Covid-19, some antibodies have been found as early as 4 days into symptoms, but it is generally in week 2 or 3 that they are reliably found. First to appear are a group called IgM antibodies, that fall off by week 6, while IgG antibodies appear slightly later but persist, for months or years.

Antibodies are commonly measured in blood samples and provide proof of prior exposure to an infectious agent. In many, but not all, cases, presence of antibodies indicates at least some immunity to the agent against which they are targeted. The degree to which the presence of antibodies means immunity to the disease varies a lot. If you have measles antibodies, you are safe, but that is less true for whooping cough or for the coronaviruses that cause the common cold. We simply do not yet know if antibodies against Covid-19 will protect one from re-infection. It is clearly premature to use these tests to offer “immune passports.”

The other major problem with current Covid-19 antibody tests is that many of the tests now available are far from reliable. Many were rushed to market with limited testing and have a high number of both false positives and false negatives. Remember that Covid-19 is one of a group of coronaviruses, many of which cause the common cold, and a poorly designed antibody test may simply be detecting a prior exposure to one of these other viruses.

Bottom line: until we have more consistently reliable antibody tests, and until we know if the presence of antibodies is a reliable way to guarantee immunity to re-infection, I would not be in a hurry to get tested.

Prescription for Bankruptcy. Buy the book on Amazon

Wednesday, May 27, 2020

Musings on masks

Let me start with the punch line and then work backwards.

You do not wear a face mask to protect yourself, though it may offer some modest protection. You wear one to protect your fellow human beings. Refusing to wear a mask does not indicate bravery and toughness but rather a lack of concern for others.

How does the coronavirus spread? It spreads almost entirely when viral particles are expelled from the respiratory tract of an infected person and inhaled by another person. Huge numbers of viral particles are spread when an infected person coughs or sneezes, large numbers when they sing or shout, and fewer but enough to spread with normal talking or even breathing.

Who can be a virus spreader? Alas, it is not only people who are obviously ill. There are still gaps in our knowledge, and it is hard to give exact numbers, but all experts agree, backed up by increasing data, that many people who harbor coronavirus and can spread it have minor or no symptoms. The most recent study I found, published in a British medical journal, found that 81% of passengers on a cruise ship who tested positive had no symptoms when they were tested. Moreover, even those who develop symptoms are most infectious just before symptoms develop. Lack of symptoms seems to be more common in women than men, and in younger adults. Roughly half the Covid-19 patients in China appeared to have been infected before the patient from whom they caught it knew they were sick.

How much do masks protect the wearer? Probably not a lot. Only N95 masks, which remain in short supply and are needed by healthcare workers, filter out most viral particles, which are very small and can pass through the fabric of cloth or disposable surgical masks. They do offer modest protection against moist droplets which carry viral particles from a sneeze or cough.

The masks do cut down on spread of the droplets generated by speaking, coughing or sneezing. The key point is not that masks do not block every virus particle, but that they filter out many. Every virus particle that does not escape into the air is a virus that will not be inhaled by others or fall on surfaces and perhaps be picked up on the hands of others.

Just as wearing a seat belt does not confer immortality if you get in a car accident, but does cut down on deaths, wearing a face mask will not prevent all spread of coronavirus but will reduce it. Wearing a mask is an easy, inexpensive and harmless way to show that you care about others.


Prescription for Bankruptcy. Buy the book on Amazon

Friday, May 22, 2020

Surrogate or real?

The dictionary defines surrogate as “a person appointed to act for another,” and to the general public probably the most common use is in terms such as surrogate parent

In medicine, it is common to have surrogate decision-makers, such as the parents of a minor child, or the spouse or child of a person who is incapable of making their own decisions. Such surrogate decision makers have all-too-often been put in the difficult position during the COVID-19 pandemic of making dread decisions such as whether to have a loved one put on a ventilator. It is very common for an elderly patient to undergo intensive care when they have asked this not be done, and researchers have also found that many surrogates’ wishes are not followed by the medical staff. It is always best, and Covid-19 has just underlined this, for a older person (or a person of any age with serious medical problems) to be sure their wishes regarding the trade-off between possibly life-sustaining care and comfort be explicit, in writing, and discussed with family.

I want to focus, however, on another form of surrogate – the use of “surrogate end points” to approve a new drug.

The gold standard for proving that a new treatment works, or works better than an existing one, is the controlled clinical trial. A large group of patients are randomly given Drug X or Drug Y, or, if there is no existing useful treatment, Drug X or a placebo. They are then followed until a pre-specified outcome occurs and the results of the two groups compared. As regular readers of these posts know, I am much more impressed with a trial that uses death as the outcome being measured. For many of the events reported, the definition is often fuzzy. The research group frequently establishes a committee to “adjudicate” whether an event occurs. You do not need a committee to decide if a patient is alive or dead.

From a researcher’s perspective, the problem with using death as the outcome is that to establish that one treatment is better than another may take a large sample followed over a long time unless the disease is rapidly lethal. Large trials over a long time are costly to perform.

Enter the “surrogate.” The trial of a cancer drug may use “shrinking of the tumor by 50% on CT scan” or “reduction of a biomarker (blood test)” as the outcome studied. While these MAY correlate well with outcomes of more value to the patient, such as lifespan or quality of life, they may not. More and more cancer drugs are now being approved by the FDA based on surrogate outcomes. A recent study published in JAMA Internal Medicine found only a very weak correlation between these end points and overall survival. While a drug approved on this basis is supposed to have further tests done that do track overall survival, in a large percentage these studies are never done. Since these new drugs are often very expensive and have many major side-effects, it appears that patients are often exposed to potential harm with minimal potential meaningful benefit.

So, when the latest “wonder drug” is suggested, be sure to ask just HOW wonderful it really is. The reported results may be “statistically significant,” but are they clinically significant? Can anyone tell you this treatment will make you feel better or live longer, or will there just be a meaningless "surrogate" improvement?

Prescription for Bankruptcy. Buy the book on Amazon

Thursday, April 30, 2020

A cure for COVID-19?

We are all desperate for a “cure” for the coronavirus illness, COVID-19. Since even the most optimistic accept that a vaccine is at least a year away (and that would be unprecedently fast), a drug that could offer a cure is the next great hope. On Wednesday, the press reported a statement by Dr. Anthony Fauci about the drug remdesivir from Gilead Science. The headlines touted “a cure,” a “game-changer,” and Gilead’s share price soared.

Do we really have an “effective treatment” against coronavirus? No, not really. Here is what Dr. Fauci said in a much-hyped White House announcement: “The data shows that remdesivir has a clear-cut significant positive effect in diminishing the time to recovery . . . The mortality rate trended towards being better in the sense of less deaths in the remdesivir group – 8 percent versus 11 percent in the placebo group. It has not reached statistical significance, but the data needs to be further analyzed.” When a result is described as not having reached statistical significance, this means that the differences could be due to chance alone. Flipping a coin three times, you will not be shocked to get heads three times, and it does not mean the coin is rigged. Similarly, small differences in response to different treatments will often show up without meaning there is a real difference, but simply by chance.

As to the “positive effect” cited by Dr. Fauci, he was describing a reduction in the mean duration of hospitalization from 15 days to 11 days. That is good, but hardly “an effective treatment against coronavirus” or a “drug that can block the virus” as the headlines blared.

Dr. Fauci’s announcement was unusual because it was made before the results of the trial have been published or subjected to peer review. To his credit, Dr. Fauci described the results as “preliminary.” He also tried to tamp down expectations arising from his announcement by saying that the data was a “proof of concept” rather than an effective treatment, similar to early results of drugs against AIDS. That important qualification was missing from many headlines. We later heard from Dr. Fauci that he made the announcement because he “feared it would come out in leaks.” While perhaps understandable, hardly the best reason or way to present science.

Contrast the NIH-sponsored trial with the results of a trial done in China and published on-line on April 29 in The Lancet, a prestigious medical journal. These researchers randomly assigned 237 patients with severe COVID-19 (all with low oxygen levels and pneumonia seen on CT) to remdesivir or placebo, They found that remdesivir use was not associated with a difference in time to clinical improvement, though there was trend to faster improvement with the drug in patients who were started on treatment earlier. They also found that treated patients had more “serious” adverse effects. There was no difference in death rates between the two groups (and those who read these posts regularly know I focus on that - a number that is neither subjective nor able to be fudged).

Why the difference between these two studies? Probably because the effects were so small that it was similar to a coin flip, but perhaps because one or the other trial had flaws. Since the Chinese study was peer-reviewed, it has inherently more credibility. Whichever proves to be more correct, the drug may (or may not) be helpful but is hardly “a cure.”

Prescription for Bankruptcy. Buy the book on Amazon

Saturday, April 18, 2020

The fox guarding the henhouse?

A tale of greed, bad science and poor decision-making by the FDA which is unfortunately not unique.

Pre-term birth, delivery before 37 weeks gestation, is a serious problem. It puts the baby at risk of disability or even death. Women who have had one “premie” are eager to do anything to avoid one in subsequent pregnancies, as are their doctors. The best estimates are that women who have had one premature delivery have a 30% risk of having one on a subsequent pregnancy.

Short of putting women on prolonged bedrest, there has been little available to reduce the rate of premature delivery. A “meta-analysis” of previously published studies done in 1990 suggested that while the hormone hydroxyprogesterone did not prevent miscarriage (for which it had been used), it appeared to protect against preterm birth. In 2003, a trial was conducted that claimed to show that injecting pregnant women weekly with hydroxyprogesterone in mid-pregnancy reduced the number of premature births. The treated group had 36% premature deliveries compared to 55% in the placebo group. There were many critics of the study, who pointed out that the placebo group had a much higher rate of premature birth than would have been expected, and those in the treated group had a similar but higher rate than expected if nothing were done. It was also pointed out that the pregnant uterus was “awash” in progesterone and so there was little biologic reason to expect the injections to do much.

Despite these reservations, many obstetricians began prescribing this treatment. Compounding pharmacies began making the injections and selling them for $15/shot. Enter stage right Adeza Biomedical, which used the 2003 trial to support their application for FDA approval of their branded hydroxyprogesterone for the prevention of premature birth. Many of the FDA’s own scientists pointed out the flaws of the trial, and when the branded product (Makena) was approved, it was with the stipulation that further trials be conducted. This trial, which enrolled over three times as many women, and reported results in March 2019, showed NO benefit from use of Makena.

The FDA convened another “expert panel” to decide if the drug should be kept on the market. The panel voted unanimously that the newer trial did not verify the benefit of Makena, and 13 to 3 that the combined evidence from the original and the newer trial did not provide evidence of any substantial benefit. When the vote came on whether to leave it on the market, the vote was much closer: 9 to 7 in favor of removing it from the market. The FDA decided to leave it on the market.

“Follow the money” has clarified many apparently confusing stories. While the product was sold for $15/shot by compounding pharmacies, Makena was marketed at an initial price of $1440/shot! As generics have entered the market, the list price has fallen, but is still $848/shot for a product that was profitable at $15. This obscene profit margin has allowed its current manufacturer, AMAG Pharmaceuticals, to generously support the American College of Obstetricians and Gynecologists and the Society for Maternal-Fetal Medicine, which supplied many of the expert panelists. Neither society has yet reflected the negative trial in their guidelines.

My take home is that all FDA panelists should be truly disinterested. Societies accepting money from drug manufacturers and researchers getting support from manufacturers should be automatically excluded from advisory panels.

Prescription for Bankruptcy. Buy the book on Amazon