I don't have kids at home, nor do I have different duties when the school year starts, but somehow I'm still sad to think that summer soon will come to a close.
It's not been the best of weeks for infectious diseases, highlighted by HHS cancelling $500 million worth of grants for mRNA vaccines based on completely faulty/bizarre reasoning by our HHS Secretary. Meanwhile, increases in emergency department visits and percent positivity of covid tests in many areas of the country suggest a covid bump is in process, just in time for start of the school year.
Scariest Journal Article in Recent Memory
Evidence is increasing that paper mills and other drivers of fraudulent scientific publications are a big problem. It ensnares all scientific fields; medicine and biology aren't immune. However, it's a slippery problem to define precisely given the newer advances in artificial intelligence. A professional field of "metascience" that examines scientific methodology and, increasingly, detects data falsification, has now become very important.
That scary article published last Monday comes from a group of metascientists at Northwestern University and the University of Sydney, Australia. I first heard about it through one of my favorite sources for tracking fraudulent publications, Retraction Watch. The article is scarier than [pick your scariest childhood movie ever, mine was the original 1958 version of The Blob starring a very young Steve McQueen] because of the massive scale of research fraud; it's going to be very difficult for some investigators to avoid referencing some of these fraudulent articles before they are retracted, if they ever are.
The authors of this study tried to get a better handle at this industrial-scale-sized research fraud whereby paper mills, for a price, will ensure an individual can publish multiple papers in scientific journals over a relatively short period of time, without ever lifting a finger to perform any research. This is far beyond the single unethical scientist fudging data or photographs to produce better results. I was fascinated to see how these authors approached the subject.
They performed various standard literature search techniques but focused on a few areas that lent themselves well to tracking falsifications. Some journals, such as PLOS ONE and journals in the Hindawi group, make it easy to access article content and metadata and also publish the names of the individual who edited (reviewed) the manuscript for publication. This allowed the metascientists to look for trends in individual editors and journals. Also, I learned about the Academic Research and Development Association (ARDA). The name sounds impressive, but actually it's an organization anyone can use to find journals or conferences that guarantee publication, generally without regard to content.
I won't delve deeply into the methodology, in part because it is very complicated and certainly not for the faint of heart. However, please note that PLOS ONE and Hindawi journals are thought to be of high quality overall and were used here because they were very transparent in their editorial processes. It makes it even scarier to see how quality journals can be hijacked. Here are a few things I learned.
The investigators discovered 22 editors (in red below) who accepted articles that were subsequently retracted significantly more frequently, as a percentage of total articles they reviewed, than would be expected by chance alone.
Keep in mind that the y axis is a log scale. Also, the number of articles handled by these flagged editors were on the high side, which may be a reflection of the fact that many journals allow authors to suggest reviewers (editors) to assess their manuscripts. Paper mills can funnel articles to editors they know will accept anything for publication.
Want more? I learned a new term, "journal hopping." Paper mills need to guarantee publication for those willing to pay, but over time a particular journal may become less desirable, e.g. because it loses indexing in key literature search engines, or just goes out of business. Paper mills need to keep ahead of these changes, ergo journal hopping. The metascientists found that journals listed by ARDA are deindexed at a much higher rate than those not listed. As an example, 13 of 39 (33.3%) of Scopus-indexed journals listed by ARDA in 2020 were deindexed later, versus 147 of 27,197 (0.5%) journals not indexed by ARDA over the same time period. They also listed some amazing examples of ARDA journal fraud, such as an article about roasting chestnuts published in a journal focusing on HIV/AIDS care and an article about malware detection that found a home in a special education journal.
The number of retracted articles in scientific journals has been increasing exponentially (note again the log scale on the y axis) recently, driven mostly by paper mills.
The article has many more scary examples that I won't elaborate on further. The authors point out that, in order to limit paper mill impact on scientific discourse, a lot of things need to change in how research fraud is detected, investigated, and sanctioned, especially taking care to remove such monitoring from those with potential conflicts of interest. The practice of financial rewarding reviewers for rapid publication and for increasing journal impact factors (a worthless metric that encourages fraud, IMHO) must end. Research institutions should not be left to investigate their own scientists. Also, if this is really such a wide-scale fraud enterprise, leaving fraud detection to a small group of (mostly) volunteer metascientists would be overwhelming. We need a redo of our systems, and quickly.
If you're looking for a bottom line to take away from this article, look no further than its title: "The entities enabling scientific fraud at scale are large, resilient, and growing rapidly."
When Did You Last Treat a Case of Bubonic Plague?
Well, for me, never. It isn't that common in the US, and I've never practiced near an endemic area. Treatment traditionally has been with parenteral antibiotic therapy, often with aminoglycosides, but quinolones do have activity against Yersinia pestis. Now we have a new study from Madagascar that included a lot of children and provides some reassurance for using oral quinolone therapy monotherapy for bubonic plague. It was an open label randomized non-inferiority trial comparing oral ciprofloxacin for 10 days to initial 3 days of parenteral aminoglycoside (streptomycin or gentamicin) followed by 7 days of oral ciprofloxacin. The results did show noninferiority for the primary day 11 endpoint of death, fever, development of secondary pneumonic plague, or receipt of alternative or additional treatment for plague up to and including day 11.
Secondary endpoints included reduction of bubo size of less than 25%, and the ciprofloxacin monotherapy group did have a higher rate of failure at 46.8% compared to 36% in the aminoglycoside-ciprofloxacin group. This did not meet noninferiority criteria, though the clinical significance of this endpoint is open to question.
Plague is on the list of bioterrorism agents, so I have always kept up with new developments in the field even though I've never seen, and probably never will see, a human case.
Neuropsychiatric Events With Oseltamivir Treatment of Flu
I've been puzzled by prior reports of serious neuropsychiatric side effects from oseltamivir during influenza treatment; most seemed to have come from other countries, and there wasn't much about this in the US, including in my own experience and that of my colleagues. Now we have reassurance that oseltamivir likely isn't to blame for these events. The study is a retrospective cohort of children 5 - 17 years of age enrolled in Tennessee Medicaid from July 1, 2016 to June 30, 2020. They ended up with almost 700,000 children enrolled. Oseltamivir treatment was associated with lower rates of neuropsychiatric events compared to no treatment. "The outcome definition includes both neurologic events (seizures, encephalitis, altered mental status, ataxia or movement disorders, vision changes, dizziness, headache, sleeping disorders) and psychiatric events (suicidal or self harm behaviors, mood disorders, psychosis or hallucination)."
The authors performed a large number of sensitivity analyses for various confounding variables, and the findings remained robust. Oseltamivir does have side effects, particularly minor GI disturbances, but the study design and available data did not include other potential side effects.
Management of Congenital CMV Infection: Start With Mom
CMV is the most common congenital infection, averaging about 1 in 200 newborns in the US and even higher in certain socioeconomic groups. Most are asymptomatic at birth but may develop hearing loss and neurodevelopmental disabilities later. Many experts have advocated routine screening of all newborns for congenital CMV infection, but these researchers in France provide early screening of pregnant people. Hot off the presses (August 9) is information on antibody avidity, amniotic fluid PCR sampling, and maternal treatment with valacyclovir starting in the first trimester of pregnancy.
Most primary care pediatric care providers probably don't remember that antibody avidity, the degree to which IgG antibody "sticks" or is bound to the antigen (CMV in this case) varies with stage of infection. As a general rule of thumb, a finding of high avidity antibody binding excludes recent (within 3-4 months) infection, while low affinity antibody binding cannot exclude recent infection.
Pregnant women seen at one French center from 2012 to 2024 all were offered CMV IgG and IgM testing in the first trimester of pregnancy. Those with positive IgG and positive or equivocal IgM were then tested for IgG antibody affinity. Fetal infection was defined as a positive CMV PCR on amniotic fluid, with amniocentesis performed starting at 17 weeks gestation. The study inclusion criteria were those who had positive IgG, positive or equivocal IgM, low IgG antibody avidity values, and a positive amniotic fluid CMV PCR. Starting in 2019 at this center, women thought to have primary CMV infection in the first trimester were offered valacyclovir prophylaxis while awaiting confirmation by amniotic fluid PCR. That's all pretty confusing, here's a study flow chart that might explain it better. (Liaison and Vidas are two different commercially available avidity assays.)
I'll show just one of many outcome figures, looking at vertical transmission of CMV by timing of maternal primary infection and whether the mother received antiviral treatment or not.
This study didn't really discuss its limitations, but one always needs to be careful about making sweeping conclusions from retrospective studies which have multiple confounding factors compared to prospective randomized controlled trials. These confounders are basically impossible to predict reliably. Also, CMV serologic assays aren't perfect, always tough to rely on these, but it's the best we can do. The study did add substantially to our understanding of reliability of the 2 avidity assays. The data from this study support early treatment of pregnant people with primary CMV infection in the first trimester to prevent fetal infection. Some states in the US now routinely screen newborns for CMV, and perhaps one day we'll see combination early screening of pregnant people plus of their newborns at birth in the hopes of decreasing congenital CMV infection and identifying infected infants sooner so that further evaluation can guide earlier treatment.
Missing John Prine
Some music lovers would have recognized the title of this week's post as part of the lyrics from a song, Summer's End, from the last album released before he died from covid in 2020. I can't multitask when listening to Prine, his lyrics are too thoughtful for me to let go as background noise.