I'd been spared from most snow shoveling due to mild winters the past few years. Fortunately, this week I discovered that my outdated acetaminophen still seems to work.
Artificial Intelligence for Pediatric Infectious Disease Diagnosis
Investigators at Baylor College of Medicine (bias alert: my alma mater) developed an interesting method to distinguish the covid-associated Multisystem Inflammatory Syndrome in Children (MIS-C) from endemic typhus, the latter enjoying a bit of a resurgence in south Texas. The rest of us don't have reason very often to consider these 2 entities together since endemic typhus (aka murine or flea-borne typhus) is pretty rare in the US beyond southern California, southern Texas, and Hawaii. However, the results from this preliminary study serve as a proof-of-concept model for other diseases. The methodology is very complex, to say the least. Briefly, they looked at electronic medical record data over a 2-year period for anyone being tested for Rickettsia typhi, the etiologic agent for endemic typhus, or having a rheumatology consult request for MIS-C. All of the typhus-positive patients and a subset of MIS-C patients were included in a preliminary dataset for the AI modeling. A large number of patient variables were tested in an iterative process to come up with a preliminary scoring system which was then validated on another MIS-C patient set. The final scoring system included 15 variables: age of patient, duration of fever, height of fever, highest heart rate, neutrophil to lymphocyte ratio, AST, ALT, sodium, troponin, BNP, fibrinogen, epidemiologic link to COVID-19 case, antecedent illness, conjunctivitis, and rash. The authors claimed their scoring system correctly classified all 220 patients in their training dataset (100% accuracy) and was 99% accurate in the 160-patient MIS-C cohort used as the validation dataset. Of course we still need further evidence that this works well in other settings and institutions. I hope the AI and machine learning process itself, so far published only as a preprint, can be applied to other clinical situations.
Another Tick-borne Virus
Speaking of vector-borne infections, now researchers in China report a novel virus from the Nairoviridae family was found to be a cause of febrile illness in 26 of 252 febrile patients tested in northeastern China. They named it XCV (Xue-Cheng Virus) after the geographic site. (IMHO, I wish we could get away from naming diseases according to a geographic site, it just encourages xenophobia.) They also found evidence that XCV caused cytopathic effect in vitro and was present in 3-6% of ticks in the area. They authors don't provide much clinical information about the cases, and a link to a supplementary appendix didn't contain any more details. Overall it does appear this is a new infectious pathogen.
If you never heard of nairoviruses, don't feel bad. Some infectious diseases physicians may be aware that Crimean-Congo hemorrhagic fever virus is the main human pathogen in this family.
Chronic Wasting Disease
I don't think I've ever mentioned CWD in these pages previously. It is a fatal neurodegenerative disease seen in cervids (deer, elk, moose, reindeer) and caused by an infectious prion. I've been watching reports about spread of CWD in the US and across the globe for many years. I mention it now because of the publication of a new report from CIDRAP concerning for potential for spillover into other species, including humans, similar to what has happened in humans with Creutzfeld-Jakob Disease, kuru, and bovine spongiform encephalopathy (mad cow disease). Here's the North American distribution of CWD.
CWD has been expanding in numbers and in geographic areas in the US the past 20 years, increasing the possibility of spillover into humans. The CWD prion can remain intact in the environment for years, and prions are not inactivated by cooking infected meat.
The report maps out several recommendations for monitoring this situation; the recommendations bullet list alone takes up 5 pages of the 102-page document. It's a good example of how to plan proactively, but of course implementation takes funding of public health endeavors.
Venison steak, anyone?
Good News in the Vaccine Department
Three quick comments on recent vaccine studies.
First, I was somewhat surprised to see a pretty good uptake of RSV prevention modalities for pregnant people and infants. In the 2023-24 RSV season in the Kaiser Permanente Northern California system, around 75% of at risk infants received protection either by maternal vaccination or nirsevimab administration after birth.
I'm hoping it is even better this season. Remember that Kaiser is a health system very well designed to deliver high rates of vaccinations and other preventive interventions to its subscribers. However, individual private medical practices can and should aim to achieve high rates of RSV prevention.
Another study provided more evidence that it is safe to administer 2-month vaccinations to hospitalized preterm infants. It was a randomized controlled trial of preterm (<33 weeks gestation) infants hospitalized at 3 NICUs and eligible for vaccination at 6 - 12 weeks of age. 223 babies were randomized to receive either 2-month vaccines or no vaccine and then monitored for 48 hours. The unvaccinated group of course could receive vaccines after this 2-day period. Although apnea was more common in the vaccinated group, it didn't appear to have any adverse effects.
The vaccines administered were PCV13, DTaP, HBV, IPV, and Hib.
The last vaccine study I'll mention is another estimate of flu vaccine effectiveness last year. Specifically, it looked at outpatients at least 8 months of age in 7 states that were part of a flu VE surveillance system. The target endpoint was outpatient visits with positive influenza testing. VE against any influenza illness was 41% (95% Confidence Interval [CI]: 32 to 49): 28% (95% CI: 13 to 40) against influenza A(H1N1)pdm09, 68% (95% CI: 59 to 76) against B/Victoria, and 30% (95% CI: 9 to 47) against A(H3N2). Protection was found in all age groups except for the 50-64-year-old group. Differences in age groups (and also related to influenza A subtypes) likely are related to imprinting, i.e. what flu strains people are exposed to early in life, but this study wasn't designed to answer this question.
WRIS
As we roll along in the winter respiratory infection season it's worth taking a look at where we are. CDC has some newer data for the "big 3" viruses in pictorial form. (See https://www.cdc.gov/flu-burden/php/data-vis/2024-2025.html, https://www.cdc.gov/rsv/php/surveillance/burden-estimates.html?ACSTrackingID=USCDC_2067-DM142871&ACSTrackingLabel=CDC%20Updates%20%7C%20New%20In-Season%20Estimates%20of%20COVID-19%20and%20RSV%20-%201%2F7%2F2024&deliveryName=USCDC_2067-DM142871, and https://www.cdc.gov/covid/php/surveillance/burden-estimates.html?ACSTrackingID=USCDC_2067-DM142871&ACSTrackingLabel=CDC%20Updates%20%7C%20New%20In-Season%20Estimates%20of%20COVID-19%20and%20RSV%20-%201%2F7%2F2024&deliveryName=USCDC_2067-DM142871.)
Note that the influenza numbers include 2 weeks longer than for RSV and covid. I did a quick calculation of death rates per illnesses for each, using the midpoint of the ranges: influenza 1.3 deaths/1000 illnesses, RSV 2.5, and covid 3.1. Of course these are all ballpark estimations but still show that infections with any of the big 3 are worth preventing.
Here's the most recent look from FluView, which captures respiratory illnesses other than influenza.
Still hoping that flu and RSV reach their peaks before covid ramps up.
Español en la Nieve
My shoveling activities were greatly enhanced by one of my new neighbors, a 3-year-old boy who just moved here from Colombia and only speaks Spanish. I'm forever embarrassed by the fact that the Spanish I've retained from my south Texas childhood can't be used in polite company, but my new friend was speaking his native tongue to me while we shoveled together. The upside if we have a heavy snow winter in Maryland is that I may develop new Spanish fluency!