Research articles of interest

Why most published research findings are false

PLoS Medicine, 2005

I have never seen so much controversy surrounding a published article as this one.

John Ioannidis used a statistical argument, not a patient trial. When studies are small, the effect being sought is small, many ideas are tested, and bias or flexible analysis creeps in, a “positive” result is often wrong. He argued that repeating the work is what sorts a real finding from a lucky one. The paper is about how research is done. It is not a finding about a particular disease.

Read the article

A child’s skull and ancient meat eating

PLoS ONE, 2012

Anthropologists described a skull fragment from a child about two years old, found at Olduvai Gorge in Tanzania and dated to about 1.5 million years ago. The bone showed a pattern they read as anemia, and they argued the child was short of nutrients such as vitamin B12 during weaning, when meat was scarce. From that one fossil they inferred that meat had already become important in the early human diet. It is a reading of one ancient bone, not a study of what a person should eat now.

Read the article

A ketogenic diet in mice

Schizophrenia Research, 2015

This was a laboratory study in mice, not a trial in people. Mice given a high-fat, very-low-carbohydrate diet showed fewer behaviors used as a stand-in for schizophrenia, and they also weighed less and had lower blood sugar than mice on ordinary chow. The researchers said the next step would be other animal models and, later, a possible human trial. A result in mice is not a treatment for schizophrenia.

Read the article

A restricted ketogenic diet and brain cancer

Epilepsy Research, 2012

This is a review, not a new patient trial. The authors argue that a calorie-restricted ketogenic diet lowers the glucose available to a malignant brain tumor. In mice, that diet slowed tumor growth. The human evidence they cite is case reports. Case reports cannot show that a diet should replace surgery, radiation, or chemotherapy.

Read the article

Ketogenic diets and the nervous system

Frontiers in Neuroscience, 2012

This review explains why a high-fat, low-carbohydrate diet has been used for drug-resistant epilepsy for about a century, and it walks through newer ideas about metabolism, adenosine, thinking, and pain. The established clinical use it describes is seizure control. The broader uses are presented as emerging research, much of it in animals.

Read the article

Fat quality in a short feeding trial

American Journal of Clinical Nutrition, 2016

Thirty-eight men with abdominal obesity were randomly assigned to a higher-carbohydrate pattern or a very-high-fat pattern, with about half of that fat saturated, from foods such as butter, cream, and cold-pressed oils. Both patterns used fresh, lightly processed foods and little added sugar, and calories were similar. In this short trial the high-fat pattern did not raise calculated cardiovascular risk. Belly, liver, and heart fat stores improved, as did triglycerides, insulin, and blood sugar. LDL cholesterol did not rise in a meaningful way, and HDL rose on the high-fat pattern. The study is small. It does not mean there is no limit on saturated fat.

Read the article

The Women’s Health Initiative diet trial

JAMA, 2006

Nearly 49,000 postmenopausal women were randomly assigned to their usual diet or to a low-fat pattern that also pushed vegetables, fruit, and grains. After about eight years, that pattern did not significantly lower coronary heart disease, stroke, or cardiovascular disease as a group. The diet change the women actually made was smaller than the one the trial had planned. This is a large trial of one dietary pattern. It is not a test of ketogenic eating, and it does not measure saturated fat by itself.

Read the article

Which fat replaces which

American Journal of Clinical Nutrition, 2009

This paper pools 11 cohort studies. People were not assigned a diet. The authors estimated what happened when saturated fat in the diet was mathematically replaced by other calories. Replacing some of it with polyunsaturated fat was linked with fewer coronary events. Replacing it with carbohydrate was not. An association in observational studies is not the same as a feeding trial.

Read the article

An opinion on salt advice

The New York Times, 2012

This is an opinion essay by Gary Taubes, not a study. He argues that the evidence behind a single low-salt target for everyone is weaker than public advice has implied, and that salt does not raise blood pressure the same way in every person. An essay can be worth reading. It is not a trial.

Read the article

A disputed meat guideline

Annals of Internal Medicine, 2019

A guideline panel reviewed earlier studies and issued a weak recommendation that adults continue their current intake of unprocessed red meat and processed meat. They judged the evidence that cutting back prevents cancer, heart disease, or diabetes to be low certainty, and the likely change in risk to be small. Other groups disagreed, in public. This is a panel’s judgment. It is not a new feeding experiment.

Read the article

A magazine feature on inflammation

Harvard Magazine, 2019

This is journalism, not a research paper. It describes the idea that a long-running, low-grade inflammatory response may connect chronic conditions that look unrelated, and that diet is one of the influences researchers are studying. It does not test a diet.

Read the article

Non-drug care for pain

Explore, 2018

This white paper reviews non-drug options for pain, including spinal manipulation, and places them inside a broader plan of care. It was written while opioid prescribing was under heavy scrutiny. It summarizes earlier evidence and states a position. It is not a new experiment.

Read the article

A guideline for low back pain

Journal of Manipulative and Physiological Therapeutics, 2018

This Canadian chiropractic guideline recommends spinal manipulative therapy as one option for recent and for persistent low back pain, together with other conservative care such as exercise. A guideline gathers earlier studies and turns them into advice. It is not a new trial, and it does not mean every person with back pain should be adjusted.

Read the article

Melatonin, the uterus, and the placenta

International Journal of Molecular Sciences, 2019

This is a review of laboratory and earlier research on ways melatonin might support the uterus and placenta, including antioxidant and anti-inflammatory actions. It is not a trial of melatonin as a pregnancy treatment. Do not start a supplement in pregnancy because of a review.

Read the article

Melatonin across female reproduction

Frontiers in Endocrinology, 2020

This review gathers research on melatonin in the ovary, in pregnancy, and around menopause. It is a map of a large literature. It is not a study that assigned women to take melatonin, and it is not a dosing guide.

Read the article

Melatonin around the time of birth

International Journal of Molecular Sciences, 2019

This review looks at animal studies and early human research on melatonin given around birth, with attention to heart and nervous-system outcomes in the offspring. The authors are describing a research direction. They are not reporting a completed trial that shows babies should be given melatonin.

Read the article

Measured melatonin in infants and mothers

International Journal of Molecular Sciences, 2019

Researchers measured melatonin in preterm infants, term infants, and their mothers. The paper is about how those levels compare. It does not test a melatonin supplement.

Read the article

Melatonin and the corpus luteum

Journal of Endocrinological Investigation, 2019

This laboratory study looked at human corpus luteum tissue, the temporary gland left on the ovary after ovulation. The authors report that melatonin can act directly on that tissue. It is not a trial of melatonin for fertility.

Read the article

One probiotic strain and colic

Alimentary Pharmacology & Therapeutics, 2019

Eighty infants with colic were split between drops of one named strain, Bifidobacterium animalis subspecies lactis BB-12, and a placebo, once a day for 28 days. Crying fell by more than half in 80 percent of the probiotic group and in 32.5 percent of the placebo group. Sleep and stools also moved in a favorable direction, and the gut bacteria made more butyrate. The authors said the result belongs to this strain. It does not transfer to a different probiotic.

Read the article

Rare gut microbes and rheumatoid arthritis

Genome Medicine, 2016

Researchers compared gut bacteria in people with rheumatoid arthritis, their relatives, and people without the disease. Groups of bacteria that are usually rare, including Collinsella, were expanded in the people with arthritis. That is an association. It does not prove those bacteria caused the arthritis, and it is not a treatment.

Read the article

A news report of two arthritis studies

Mayo Clinic news release, 2016

This news story describes the patient study of rare gut microbes and a separate mouse experiment. In the mice, a human gut bacterium, Prevotella histicola, reduced the frequency and severity of arthritis-like disease. The physicians quoted in the story said human prevention trials had not been done. A news headline that says bacteria can prevent rheumatoid arthritis is ahead of the evidence.

Read the article

Probiotics, strain by strain

Clinical Gastroenterology and Hepatology, 2012

This is a review written for gastroenterologists. The useful evidence is tied to a particular strain and a particular gut problem. A label that only says “probiotic” does not tell you which studies apply. The paper is a guide to the literature. It is not a new trial.

Read the article

A small probiotic pilot in schizophrenia

Brain, Behavior, and Immunity, 2017

Fifty-six adults with schizophrenia took either a placebo or a probiotic containing Lactobacillus rhamnosus and Bifidobacterium animalis. In the men, antibodies to Candida yeast fell much more on the probiotic. Bowel complaints and the change in symptoms such as delusions depended on whether those yeast antibodies were high at the start. The sample of women was too uneven to judge. The authors said the study was too small to recommend the probiotic widely. It is not a treatment for schizophrenia.

Read the article

A proposed name, mapranosis

PLOS Pathogens, 2017

Two researchers proposed the word “mapranosis” for a process in which amyloid proteins made by gut microbes might misfold proteins in the nervous system and add to inflammation. Their earlier work was in rats and worms. This paper names an idea and asks for more research. It is not a treatment for Parkinson’s disease, Alzheimer’s disease, or ALS.

Read the article

Microbes rewrite mouse chemistry

Nature, 2020

This was a chemistry map in mice, four with ordinary microbes and four raised germ-free. As much as 70 percent of the molecules in the gut differed, and about 20 percent differed even in distant organs such as the brain. Bacteria were attaching extra amino acids to bile acids. Similar modified bile acids turned up in some human samples, more often in infants and in people with inflammatory bowel disease or cystic fibrosis. The paper shows that microbes change the body’s molecules. It does not offer a treatment.

Read the article

Bile acids as immune signals

Nature, 2019

Two mouse studies found that gut microbes modify bile acids, and that those modified acids shift the balance between inflammatory T cells and calming regulatory T cells in the intestine. Mice that lacked the signal were more prone to colitis. The same bile-acid byproducts can be found in human stool, including stool from people with inflammatory bowel disease. These are mechanism studies in mice. They are not a treatment.

Read the article

Exosomes as a gene carrier

Nature Biomedical Engineering, 2019

Researchers used an electrical pulse to load donated human cells with a synthetic gene, then collected the tiny parcels those cells released. In mice with glioma, parcels carrying a tumor-suppressor gene were more likely to reach the brain tumor, slow its growth, and lengthen survival than the comparison injections. This is a proof of concept in mice. It is not a treatment available in the office.

Read the article

Aspirin for healthy older adults

New England Journal of Medicine, 2018

In the ASPREE trial, older adults without known cardiovascular disease took 100 milligrams of aspirin or a placebo every day. Aspirin did not lengthen survival free of disability, and major bleeding was more common. This was primary prevention in people who were already older and otherwise healthy. It does not mean a person should stop aspirin that was prescribed after a heart attack or stroke.

Read the article

Viruses and human protein change

eLife, 2016

This is a computer analysis of genomes. Proteins that interact with viruses show more signs of adaptive change than other proteins. The authors estimated that viruses account for a large share of those adaptive changes in humans. It is an argument about evolution. It is not a finding about a present infection.

Read the article

Shared genes in viruses and cells

News report of a 2017 genome comparison

Researchers compared protein structures in viruses with those in bacteria, archaea, and more complex organisms. They found a large overlap and read it as a long history of gene exchange between viruses and cells. This is evolutionary bookkeeping. It is not a finding about an illness in a patient.

Read the article

Transcription factors in clusters

eLife, 2017

Using yeast and a microscope that can see single molecules, researchers found that the proteins which switch genes on and off work in clusters of about seven to ten, roughly 30 nanometers across, rather than as lone molecules. They think the cluster size lets the proteins move quickly along packed DNA. This is basic cell biology in yeast. It is not a treatment.

Read the article

Enhancer RNAs and a gene-activating enzyme

Cell, 2017

Most of the genome does not code for protein. This laboratory study found that RNAs copied from enhancer regions bind an enzyme called CBP and increase a chemical mark that helps a nearby gene turn on. Many DNA changes linked with cancer sit in those enhancer regions rather than in the protein recipe itself. The paper explains a mechanism. It does not test a therapy.

Read the article

Droplets that gather broken DNA

Nature Communications, 2020

In yeast, repair proteins form liquid-like droplets, and short filaments inside the nucleus help those droplets fuse into a repair center. Damaged DNA is brought there along longer filaments. The older picture, that a broken strand simply drifts until it is fixed, does not match what they filmed. This is basic cell biology. It is not a treatment.

Read the article

Berberine in a computer model

Frontiers in Pharmacology, 2019

This paper reviews earlier laboratory reports and then uses a computer model to predict that berberine might touch several molecular targets related to more than one cancer type. A prediction of that kind is not evidence that berberine treats cancer in a person.

Read the article

Cannabidiol in laboratory cancer research

International Journal of Molecular Sciences, 2019

This is a review of laboratory research on cannabidiol, the non-intoxicating compound from cannabis. The authors describe many possible molecular targets seen in cells and in animals. A review of those experiments does not show that cannabidiol treats cancer in a person.

Read the article

Selection on a chemical mark, not only on DNA letters

Cell, 2020

The news headline calls this a “living fossil.” The organism is Cryptococcus neoformans, a yeast that can infect people with weakened immune systems. The paper is not about treating that infection. It shows that a chemical annotation on the yeast’s DNA, methylation, has persisted for tens of millions of years after the enzyme that normally adds new marks was lost. The authors argue that natural selection has been preserving those marks, probably because they silence jumping genes. That is an evolutionary claim. It is not medical advice.

Read the article