An ongoing dialogue on HIV/AIDS, infectious diseases,
January 22nd, 2019
Unanswerable Questions in Infectious Diseases — Treatment Duration in Endocarditis: 4 Weeks, 6 Weeks, Other?

National Library of Medicine
Time to get back to some tough clinical decisions. It’s been a while.
We’ve done The Abdominal Collection and Duration of Antimicrobial Therapy, Persistent MRSA Bacteremia, and The Positive Cultures for Candida in an ICU Patient.
However, that series of posts appeared here in early 2014, which means it’s been 5 years with no “Unanswerable Questions.”
Lest you think that all Unanswerable Questions have been answered — ha — here’s another one, inspired by a flurry of papers recently on a favorite topic, duration of antimicrobial therapy.
To summarize virtually all of them — shorter is usually as good, if not better, than longer. Nice editorial here.
Before we get to the case, an up-front apology that the question might seem simple. But I assure you, the answer is anything but straightforward. Bold prediction: there will be a substantial divergence in responses.
Additionally, while the question may not seem like such a big deal, there are many interested parties eagerly awaiting our answer — including our hospitalist, cardiology, and primary care colleagues, diverse insurance plans, homecare companies, outpatient pharmacies, and skilled nursing facilities. Most importantly, our patients and their families care a lot, too.
And even though the correct answer to “How long should I treat?” is “Long enough” (thank you, Bob Rubin!), this doesn’t change the fact that consulting services really want to know a precise answer. Ironic, huh?
Here’s the case, an amalgam of many seen over the years:
A 52 year old man is admitted with fever. On exam, he has a temperature of 101.5F and a loud systolic murmur. Two sets of admission blood cultures grow methicillin-sensitive Staph aureus. A cardiac ECHO shows a bicuspid aortic valve with a 0.5-cm vegetation.
He is treated with oxacillin; follow-up blood cultures are negative by hospital day 3, and fevers slowly resolve. A peripherally inserted central catheter (PICC) is placed on hospital day 5, and cefazolin replaces oxacillin in anticipation of discharge home on IV antibiotics.
The patient, the medical team, the homecare company, and your OPAT colleagues all await a specific “stop date” for the IV cefazolin.
The question:
If he has a clinically stable course (no further positive cultures, no recurrent fevers, and no metastatic sites of infection), how long would you treat him with intravenous antibiotics?
Before you answer, you’re welcome to look at the guidelines — this is an open-book blog, after all:
- American Heart Association
- European Society of Cardiology
- British Society for Antimicrobial Chemotherapy
Or, you could just wing it based on your clinical experience and judgment, or cite the POET study if you’re feeling cutting edge. Clock starts the day of the first negative blood culture.
And please defend your choice in the comments section, especially if you choose option 4!
(Apologies to Drs. Wald-Dicker and Spellberg for the Days-of-the-Week Units.)
January 13th, 2019
Are We ID Doctors Really So Unhappy Outside of Work?

Must be a rheumatologist or otolaryngologist. (Source: National Library of Congress.)
Medscape released their 2019 Physician Lifestyle & Happiness Report, and the results aren’t pretty for a certain cognitive specialty, one commonly abbreviated “ID.”
Out of 29 medical and surgical specialties, infectious diseases physicians ranked second to last when responding to a 7-point scale rating on their happiness. Only neurologists were gloomier than we were during their off-hours.
Rheumatologists and otolaryngologists finished first and second, smiling all the way.
The news was even worse for ID docs in “self-esteem,” where we ranked last, right behind oncology and internal medicine. Meanwhile, the plastic surgeons, urologists, and ophthalmologists scored highest on this measure.
Some of my ID colleagues have commented about possible methodologic issues with the survey, as the results don’t correlate with our own happy non-work lives.
For example, Dr. Dan McQuillen weighed in with this critique:
Much like crappy data in major salary surveys, ID docs were only 1% of this sample, roughly 150 of 15,069 docs surveyed. @IDSAInfo salary surveys w/ larger sample size have shown small sample size data to be inaccurate. Hard to take 150 responses and generalize to > 11K ID docs.
— Dan McQuillen (@McQHoya81) January 11, 2019
I have a query into Medscape to get more information about these concerns, and they’ve kindly agreed to get back to me. For example, I don’t understand how the 7-point happiness scale translated into the percentages shown in the linked figure.
More importantly, what were the demographics of these 150 ID respondents, and how do they compare with ID practitioners as a whole?
Methodology notwithstanding, it’s worth postulating at least a few reasons why indeed we might be somewhat less happy outside work than other doctors right now. So here goes:
1. The current political climate. It’s no secret that ID doctors skew strongly to the left when it comes to politics. I can’t be the only ID doctor who, on a certain Wednesday in early November 2016, went to work and encountered several colleagues and trainees literally in tears.
(Or maybe you were in tears yourself. It’s OK to admit that to other ID docs.)
I’ve discussed this issue before, postulating that the “safety net” and inclusive ethos of ID doctors was more in line with one particular party than the other. Some even choose ID as a specialty because of these political leanings. Here again are the facts, which could very well contribute to a lower happiness score during the recent survey period:
I can count on one hand the number of ID/HIV docs I know who are openly Republicans, and still have fingers left. https://t.co/m2ewtxpcwd pic.twitter.com/90gEaIWISM
— Paul Sax (@PaulSaxMD) October 6, 2016
How about today, two-plus years later? Decent chance that the fraction who have registered red is even smaller.
2. Salaries, debts, and money concerns in general. During a time when certain government workers are sadly not being paid at all (see #1, above), it seems petty to complain about ID doctor salaries — which, in this latest comprehensive salary report, aren’t really that bad:
In general, full-time ID physicians in private practice (n = 366) reported higher incomes, with a median annual salary of $260 000, than respondents employed by hospitals, clinics, or academic medical centers (median salaries of $237 500 and $181 500, respectively).
But context is everything, and here are some important considerations. Students graduate from medical school with on average nearly $200,000 in medical school debt, a hefty sum to pay down with these ID salaries.
Furthermore, ID doctors are often paid less than hospitalists and primary care physicians — doctors who have spent less time training, and frequently work fewer hours, than ID doctors. And all of us cognitive clinicians can only dream of accumulating the RVUs (and hence revenues) of a plastic surgeon, urologist, or ophthalmologist.
Hey! Those are the docs who just happen to lead in the “self-esteem” metric! Hmmm.
3. We’re by nature big-time worriers. Obsessive to a fault, we ID doctors take the most detailed histories, frequently contact the outside labs, march down to the radiologists or ECHO room or microbiology lab to review primary data, write the longest (too long) notes, and still — we live in terror of missing something.
Could it be that this personality trait doesn’t translate into happiness? Maybe the sensation that our work is never done translates into non-work “worry hours” that cloud the responses to a happiness survey.
Imagine the thought bubble of some of us as we head home from work:
How can I be happy at home when that patient with Staph aureus on a urine culture may have bacteremia? Or that other patient with a positive IGRA could develop active tuberculosis when starting etanercept? Or that person being discharged on IV antibiotics won’t have close follow-up? Or that guy who missed his HIV follow-up appointment might have stopped ART? Or that patient getting blasted with immunosuppression for graft-versus-host disease might have an undiagnosed fungal infection? Or that person …
You get the idea.
4. Our work lives are so interesting, rewarding, and wonderful that everything else pales by comparison. This must be the explanation, right? For example, look what I accomplished at work just this last week — how could anything beat that?
Apologies for bragging, but I'm the proud owner of a new degree.
Cc'ing @NobelPrize , just in case this was the one missing piece. pic.twitter.com/q5Y83YKCu6— Paul Sax (@PaulSaxMD) January 9, 2019
So, what do you think? Is the Medscape survey valid?
January 6th, 2019
Rabies After Trip to India, Aortic Dissections with Quinolones, a Vaccine for Candida, Koala Bites, and More: A Welcome-to-2019 ID Link-o-Rama

From the Library of Congress, government shutdown notwithstanding.
As 2018 tips over into 2019, here are a bunch of ID- and HIV-related studies that, for one reason or another, haven’t made their way to this site yet — but still yearn for your attention:
- Cases of infective endocarditis have increased since release of the 2007 dental prophylaxis guidelines. Recall that those guidelines only recommended prophylaxis for patients at the highest risk of IE. The results of this population-based study suggest (but do not prove) that perhaps some moderate-risk patients should also receive antibiotic prophylaxis.
- A woman in Virginia died of rabies after being bitten by a puppy during a trip to India. Though rare, these horrific cases underscore the value of pre-travel counseling and immunizations for certain trips. And dogs (especially puppies) are cute and wonderful, but stay away from strays in countries that still have canine rabies!
- Per FDA, fluoroquinolone use is linked to aortic dissections. As the parade of adverse events for this drug class marches on, it’s worth remembering that they remain extremely useful antimicrobials, especially in patients with complex infections. For those interested in the history of these alerts, here’s a great graphic, courtesy Travis Jones, PharmD:
- Dolutegravir monotherapy lands with a thud. In the linked study, nearly 10% failed by week 48 (some of whom developed resistance), prompting cessation of the study. A second study done in patients treated during early infection suggests the monotherapy strategy may work in patients with a low viral reservoir, but 1) we don’t measure reservoir in clinical practice, and 2) what’s the point?
- Starting antiretroviral therapy right away works great — even in the most difficult-to-treat populations. In a group of 216 patients with high rates of mental illness, substance use disorder, and homelessness, an astounding 95.8% achieved an HIV viral load <200 at least once after receiving immediate ART. I can’t imagine this occurring without integrase inhibitor-based therapy.
- A 39-year-old man nearly died of influenza last year, and now he’s publicizing the importance of the flu vaccine. I have tremendous admiration for people who endure such extreme hardship and then try to make something good out of the experience.
- A vaccine is under investigation for recurrent vulvovaginal candidiasis. This phase 2 study demonstrated safety and immunogenicity and hinted that it would be effective as well. For the record, there are currently zero vaccines for fungal infections.
- Here is a superb summary of the severe infrastructure problems fueling the national hepatitis A outbreaks. This disease is strongly linked to breakdowns in sanitation and hygiene, and not surprisingly have targeted the homeless and drug user populations.
- Complications are very high with outpatient parenteral antimicrobial therapy (OPAT) for severe Staph aureus infections. Roughly 33% experienced an adverse event, and 64% were readmitted at 90 days. Clearly a very tenuous patient population! And will Dr. Spellberg be right in his bold prediction?
- In North Carolina, infective endocarditis related to drug use has increased more than 12-fold over the past 10 years. As I’ve noted numerous times before, this dramatic rise in drug-use-related endovascular infections is one of the most profound changes in ID since starting practice in the early 1990s.
- Add wound botulism to the list of infectious complications of injection drug use. These cases in California involved black tar heroin, which is made mostly in Mexico. During transport inside car tires and other unsanitary locations, it may be exposed to soil containing Clostridium botulinum spores.
- Using both pneumocystis PCR and blood beta glucan improved the diagnosis of Pneumocystis pneumonia in cancer patients. These two tests complement each other — PCR is highly specific, but it may be too sensitive; a high value on beta glucan can help distinguish between colonization and true infection.
- Beta glucan is frequently elevated after abdominal surgery. Another study underscoring the basic problem with this test — if it’s ordered in the wrong context, there are innumerable false positives.
- Fusobacterium necrophorum is often implicated in recurrent peritonsillar abscess. This anaerobic gram-negative infection is more famously known for causing septic jugular vein thrombosis (Lemierre’s syndrome), but it clearly has a role in other suppurative complications. Question: Would a throat swab diagnostic test for F. necrophorum help identify the non-strep pharyngitis cases that would benefit from antibiotics?
- Cefiderocol versus imipenem for (what else?) complicated UTI — and cefiderocol comes out the winner. The winner, that is, based on microbiologic response. As a reminder, cefiderocol is a “siderophore” antibiotic (you can read here what this means), with activity against multi-drug resistant Pseudomonas, Stenotrophomonas, and Acinetobacter spp. Looking forward to seeing the studies in more difficult to treat patient populations, such as hospital-acquired pneumonia.
- Here’s a simple framework to guide antibiotic prescribing: 1) Is infection one that needs an antibiotic? 2) Have cultures or other studies been sent? 3) What empiric treatment should be started? 4) How should results alter therapy? 5) What duration is required? ID clinicians have internalized these thoughts, but no harm making them explicit.
- A woman acquired HIV that was highly resistant to all available integrase inhibitors. Transmission of integrase inhibitor resistant virus is very rare; these viruses have typically still retained susceptibility to dolutegravir and bictegravir. Not this one — the baseline isolate had three major integrase mutations, E138A, G140S, and Q148H. The critical information from the case report was the identification of the source patient, for whom the authors obtained detailed resistance information.
- For ID specialists, management of Staph aureus bacteremia is all over the map. Treatment of choice, management of persistent bacteremia, duration of therapy — plenty of variation! Example: For MSSA endocarditis, 32% chose cefazolin, 29% favored nafcillin, while 32% considered the two the same. Practice variation can represent differences in quality, but here I think they reflect lack of a clear, evidenced-based choice.
- For infected burn wounds, bacteriophages did worse than standard of care in reducing Pseudomonas aeruginosa bacterial burden. This approach to treating resistant infections is highly promising, but still in its infancy.
- Three patients suffered wound infections from Lonepinella koalarum after koala bear bites. No, I’ve never heard of that bug either — which, given the source, will never become a global threat — but the molecular investigation into these cases is quite the tour de force.
Have you written 2018 on a form or check yet?
Of course you have. Happy New Year!
And just 18 million views for this video (and counting):
January 2nd, 2019
How Did Our Medical Notes Become So Useless?
Among the many complaints about electronic medical records (EMRs), the death of the useful medical note ranks very high.
Notes are too long, too complex, and filled with unhelpful words. It’s often impossible to glean what the clinician thinks is going on, or what’s planned.
Ever get a note from an urgent care clinic on a patient who went there with a viral syndrome? Or a discharge summary? The note contains pages of indecipherable gobbledygook, ICD10 codes, irrelevant review of systems, stock phrases — the medical words are there, but where is the content? Give a click on this note for a particularly egregious example (all identifying information removed). Then come on back here. I’ll wait.*
(*Good chance these notes are faxed to your office, then scanned into your EMR’s “media manager,” or whatever your EMR calls it. My wife, a primary care pediatrician, calls this part of the EMR “the place where information goes to die.” Yep.)
It hasn’t always been this way. I’ve worked with EMRs of various sorts for decades. One of them, designed for outpatient care, had two ways to file notes — the clinician either dictated a narrative (for complex cases) or, more often, wrote a brief handwritten note in a 3-line section that was immediately typed in by clerical staff.
Both types of notes were infinitely more useful than today’s behemoths. The long dictated notes told a logical story, the short ones highlighted only the most relevant information. Example of the latter:
New painful rash on back. PE: vesicles in T10 dermatome on L, otherwise neg. Dx zoster, Rx Valacylovir 1 gm TID for 7d. Discussed possible complications, reasons to return or call for f/u.
That’s it! Today, this would be unimaginable. In a paper published in the Annals of Internal Medicine, three experts in EMR optimization compared the length of notes in the USA vs other countries:
In other countries, [a note] tends to be far briefer, containing only essential clinical information; it omits much of the compliance and reimbursement documentation that commonly bloats the American clinical note. In fact, across this same EHR, clinical notes in the United States are nearly 4 times longer on average than those in other countries
So how did we get here? What caused the note to shift from being the primary means of communicating medical information to this gargantuan beast? Three primary reasons:
1. Money. Some might call this “billing” or “regulatory” or “compliance,” but let’s call it by its root source — money. Based on quirks of our strange American healthcare system, certain words or phrases or diagnoses yield higher reimbursement than others. This hierarchy has nothing to do with delivering good patient care or communicating with other clinicians.
It’s not just individual words — entire sections of notes owe their very existence to maximizing revenue from clinical services. Dr. Mark Reid, author of the entertaining Medical Axioms, complained last week about being forced to include certain words in his notes.
He received this painful response from a Cardiology fellow, who recently had his notes reviewed by a “Cardiology Coder”:
I got this message in my epic inbox. As a FIT, it was hard to look at this in any positive way. What can I do to help? pic.twitter.com/7oJ34DOUdF
— Ali A. Azeem (@aliahsanazeem) December 29, 2018
Not only did poor Dr. Azeem include a Review of Systems to satisfy the Insatiable Billing Monster, but someone reviewed his Review of Symptoms to ensure he used the correct words! Could there be a better example of what’s wrong with medical documentation than this anecdote?*
(*And could you imagine having that reviewer’s job? Shudder.)
2. Copy/paste. Some EMRs have a feature where you can highlight only the original — not the copied or imported — content of a note. If you do this, you instantly understand why “ID consulted, awaiting input” appears several days after you’ve done your consult and have been communicating regularly with the primary team. They’re not ignoring your beautiful consult, they just haven’t gone back to update the text.
Other symptoms of copy/paste madness are the gobs of laboratory and radiology data appearing in every note, copied from the actual reports and then pasted into the “Results” section, or imported via macros (see #3 below).
How bad is the copy/paste phenomenon in medical documentation? Researchers at UCSF reviewed the source of text from medical notes over an 8-month period, and their findings were not pretty:
We analyzed 23, 630 notes written by 460 clinicians. In a typical note, 18% of the text was manually entered; 46%, copied; and 36%, imported.
3. Text expanders. Call them what you like — “smart text” or “auto text” or “templates” or “dot phrases” — but these tricks of the trade, once mastered, are simply irresistible to most of us, for better or worse. Dr. Grace Farris captures our ambivalent relationship with this strategy perfectly in the cartoon that led off this post.
It works like this — we enter a magic little short string of characters, press return, and voila! Everything from a complex (but commonly used) sentence to a full medical note appears on the screen. Take a bow and admire your work!
From a Reddit thread on this strategy:
.NICU: As a peds resident, I made this dot phrase. Took me about 2 full days of work to get it together, but I basically created it to pull all the info I needed to preround on a patient. All the numbers from all systems, weight change, ecmo/ventilator settings, even the number of desaturations they had overnight. Approximately 6-8 pages worth of data, arranged in order of systems that we would present on rounds.
Impressive! But do we clinicians really learn, or interpret, material that “autopopulates” a note? And how can these one-size-fits-all notes apply to the infinite diversity of patient care?
Full disclosure: After once receiving feedback that my notes didn’t have sufficient documentation about the time spent on counseling and patient education, how did I respond? By creating “.saxcounsel”, of course — which when typed, expanded into a thorough description of time spent on counseling and patient education! That will show them!
So is there any hope for the medical note?
I do think there is a way to improve them, at least a bit, that won’t require a complete overhaul of billing regulations — but that will be a topic for a different post. In the meantime, I very much welcome your suggestions in the comments!
December 16th, 2018
ID Doctors Are Lousy Golfers, and Time to Pick Your Favorite Cartoon Caption
Some might wonder how people who take care of patients, who deal with illness and suffering on a regular basis, can find humor in medicine.
Alternatively, one could take the opposite perspective — with so much misery around all the time, how could we survive without humor?
Clearly the folks at The BMJ are in the latter camp, as each year they bring us some high-quality chuckles with their annual Christmas issue.
Take this recently published paper on golf habits among doctors, a leisure activity apparently quite unpopular among us ID physicians (my explanation below):
That's because we don't want to catch Lyme, anaplasmosis, ehrlichiosis, RMSF, Powassan, etc. https://t.co/ZSxMwro6mI
— Paul Sax (@PaulSaxMD) December 11, 2018
Note that there was also a strong inverse correlation between the percentage of golfers in a given specialty and their average golf handicap (whatever that is). I’m pretty sure this means that we stink at it.
Now it’s time to add to the holiday fun by offering you the chance to vote on your favorite caption from our last contest.
As always, we used a rigorously tested algorithm to come up with the finalists. This fully automated process harnesses the full computing power of the NEJM Journal Watch servers, slowing other publishing activities to a crawl.
First the cartoon (thank you again, Anne Sax), then your vote, please.
Which is the funniest cartoon caption?
- "Any recent travel?" (39%, 171 Votes)
- "I know what the Internet says, but Lyme disease is not endemic here." (28%, 125 Votes)
- "I think my water just broke." (20%, 90 Votes)
- "As a matter of fact, a lot of patients complain about inconvenient parking. Why do you ask?" (13%, 57 Votes)
Total Voters: 443
December 12th, 2018
Two Weeks of Attending on the ID Consult Service, with One-A-Day ID Learning Units
For those of us who don’t do inpatient medicine all the time, the “blocks” doing inpatient Infectious Diseases consults are a stark reminder of just how complex and challenging the case material can be.
Think about it — if a hospitalized patient has a straightforward ID problem, we are not getting involved. No one consults ID for cellulitis that rapidly improves, for community-acquired pneumonia responsive to antibiotics, or for the post-op infectious complication easily amenable to incision and drainage.
I’ve said it before — you know that randomized study of short-course antibiotic therapy for abdominal fluid collections? The one where the entry criteria included “adequate source control”?
We’re never consulted on those cases. Just these.
That’s why it was no surprise to read this recent paper, which showed that among 2.5 million patients in Canada, those seen by ID ranked second in complexity among all the sub-specialists. We trailed only nephrology — who, by taking care of all those people on dialysis, certainly have their hefty share of complex patients.
In order to provide some structure to this on-service experience, I try to find at least one item daily to for us to learn.
My “criteria” for inclusion:
- Has to be related to a case.
- Has to have a reference.
- Has to be interesting.
- Has to have no patient confidentiality issues.
Unlike previous rotations, this time I did it on the fly (so to speak) using Twitter. While some stay away from Twitter since it is (to certain individuals) an irresistible way to say something stupid, medical Twitter can also elicit fascinating responses and dialogue from an incredibly diverse group of clinicians. Thank you for that!
So here are the daily ID Learning Units from two weeks on service — a truly enjoyable time spent with an outstanding first-year fellow and a great second-year medical resident, someone I’m hoping will one day go into ID!
Day #1 On-Service ID Learning Unit: The only approved HIV regimen with no renal metabolism is DTG + RPV — hence no dose adjustment required in ESRD. For pts with viral suppression only. Pivotal study here: https://t.co/VRXb6puD6F
— Paul Sax (@PaulSaxMD) November 27, 2018
Day #2 On-Service ID Learning Unit: In vertebral osteomyelitis, ESR > 55 and CRP > 2.75 at week 4 of antibiotics associated with higher risk of treatment failure. Small study but consistent with clinical experience. https://t.co/HGEdqVAh5O
— Paul Sax (@PaulSaxMD) November 28, 2018
Day #3 On-Service ID Learning Unit: Dental work may contribute to the development of streptococcal PV endocarditis; most cases still NOT related. But how to manage future dental work in patients who appear to have "failed" prophylaxis? https://t.co/CYumfjUJ7q
— Paul Sax (@PaulSaxMD) November 29, 2018
Day #4 On-Service ID Learning Unit: Who to screen for latent TB?
1) High likelihood of infection (e.g., from highly endemic region) OR
2) High likelihood of of progression to TB disease if infected (e.g., anti-TNF Rx)
Also, IGRA preferred over TST. https://t.co/0FIfbR2pic— Paul Sax (@PaulSaxMD) November 30, 2018
Day #5 On-Service ID Learning Unit: The always-confusing EBV serologies–which are unfortunately called "antigens" (i.e., antigen antibodies). https://t.co/GKwhnGQjhR pic.twitter.com/uIvhyWgtXk
— Paul Sax (@PaulSaxMD) December 1, 2018
Day #6 On-Service ID Learning Unit: Linezolid vs vancomycin in MRSA pneumonia. Controversial clinical trial at the time–but critiques notwithstanding, linezolid at least as good as (and possibly better than) vanco for Staph aureus pneumonia.https://t.co/Lpr2Dzw950
— Paul Sax (@PaulSaxMD) December 2, 2018
Day #7 On-Service ID Learning Unit: The oral antibiotic options for penicillin susceptible streptococci in the POET trial of partial oral therapy for endocarditis. Not easy given pill burden, dosing frequency, drug interactions, etc. https://t.co/1LjCCwCARJ pic.twitter.com/qOFIis9SOo
— Paul Sax (@PaulSaxMD) December 3, 2018
Day #8 On-Service ID Learning Unit: Fever of Unknown Origin or Fever of Too Many Origins? Still the best depiction of the challenge of ICU ID consults. "Frequently, the treatment approach is like playing Whac-A-Mole." https://t.co/uaCw6IKIeo
— Paul Sax (@PaulSaxMD) December 4, 2018
Day #9 On-Service ID Learning Unit: Diabetes is a consistent risk factor for invasive gp B strep infection in adults. A representative recent study:https://t.co/YgLZblL7kW
— Paul Sax (@PaulSaxMD) December 5, 2018
Day #10 On-Service ID Learning Unit: C diff is a common cause of "unexplained" leukocytosis in hospitalized patient. Still a valid observation — but if you just ask about C diff symptoms, it doesn't remain unexplained for very long! https://t.co/34Wtji8n5h
— Paul Sax (@PaulSaxMD) December 6, 2018
Day #11 On-Service ID Learning Unit: Multiplex PCR for diagnosis of CNS infections may yield false + (as in this case report) and false – (esp HSV and cryptococcus) results. Tests increasingly used; larger validation studies warranted! (h/t @iddocjen) https://t.co/A9HWULtqMa
— Paul Sax (@PaulSaxMD) December 7, 2018
Day #12 On-Service ID Learning Unit: Can you stop PCP prophylaxis with viral suppression and CD4 < 200? In this cohort study, for pts with CD4 100-200 on ART, the incidence of PCP was zero after stopping prophylaxis. (Which is what I do.) https://t.co/Z1Tlgin92W
— Paul Sax (@PaulSaxMD) December 8, 2018
Day #13 On-Service ID Learning Unit: In several studies (including this recent one), RSV causes as much morbidity among adults as influenza — esp in those with concomitant cardiopulmonary dz or immunosuppression. A very underappreciated pathogen! https://t.co/M4r8BHHYv0
— Paul Sax (@PaulSaxMD) December 9, 2018
Day #14 (and final, for now) On-Service ID Learning Unit: Is that pesky low-level viremia (50-200) in our HIV patients adherent to ART of any clinical consequence? Depends! In this study, don't worry! [THREAD] https://t.co/aSNiEGWwp6 pic.twitter.com/OkIaclhW5J
— Paul Sax (@PaulSaxMD) December 10, 2018
December 2nd, 2018
As A Strategy for HIV Prevention, Disabling the CCR5 Gene in Embryos Implanted in HIV-Negative Mothers Makes Zero Sense

The CRISPR Way to Identify Proteins Essential to HIV-1 Infection. From N Engl J Med 2017; 376:1290-1291.
One of the great joys of being an ID/HIV specialist is looking back at how far we’ve come in HIV prevention and treatment since the beginning of the epidemic.
Here are a bunch of things we know about HIV prevention, listed roughly in order of when we learned them — and forgive me if this is an oversimplification for this sophisticated readership:
- Condoms work very well in preventing HIV transmission.
- Taking a brief course of HIV therapy soon after exposure reduces the risk of HIV acquisition.
- Babies born to HIV-positive mothers do not contract HIV if the moms take suppressive HIV therapy.
- Male circumcision reduces the risk of these men acquiring HIV.
- People do not contract HIV from their HIV-positive partners if the person with HIV takes suppressive HIV therapy.
- People taking pre-exposure prophylaxis markedly reduce their risk for HIV acquisition.
You’ll note that nowhere on this list is anything about preventing HIV in babies born to women who don’t have the virus to begin with — because the babies are not at risk, even if the mother’s male sexual partner has HIV.
Just typing that sentence felt a little strange, it’s so obvious. However, it seems that He Jiankui may not understand this basic fact.
He’s the scientist who startled the world by releasing news that he and his research team had used CRISPR–Cas9 genome-editing to alter the embryos of two babies. The editing disabled the CCR5 gene, which means the babies lack a key co-receptor that HIV uses to infect cells.
However, as noted at 1:53 in the above-linked video, it’s the father who has HIV. Indeed, reports indicate that eight serodiscordant couples have participated in his studies — all with the fathers having HIV.
I’ll let others with far greater knowledge of genetics, embryology, and medical ethics comment on just how reckless this experiment was — here’s a good take (there have been many).
But from an HIV prevention perspective, it’s easy to judge — it makes zero sense to do this since the babies aren’t at risk of getting HIV to begin with. For them and their families, the genome-editing was all risk and no reward.
Let’s hope the mothers understood this before they agreed to participate in this disturbing experiment.
November 25th, 2018
Does Experiencing Childhood Illness Make Someone Stronger? How One Person Turned Adversity into Remarkable Success
Many people growing up with chronic illness become resilient.
Whether it’s Crohn’s Disease, or cystic fibrosis, or diabetes, or the sequelae of an accident, or whatever condition they have, they impressively live their life just like the rest of us — occasionally regressing or slowing down only during a flare of the illness.
But then there’s an extreme version of this resilience — people born with or acquiring severe illness as a child, and somehow not just surviving, but thriving.
Their medical problems are just a small bump in the road as they go from one success to another, each achievement more remarkable given what they’ve had to go through to make this success happen.
The easy interpretation is that all the hardships they’ve endured make them stronger — but it must go beyond that, since most of us mere mortals would respond in no such way.
Dr. Eric Winer, who runs the breast cancer program at Dana-Farber Cancer Institute, is an example of this remarkable group. He kindly agreed to tell his story for a podcast on Open Forum Infectious Disease.
(It’s also available on iTunes, and Overcast; we’re working on Stitcher.)
He talks about growing up with hemophilia, and HIV, and hepatitis C, and how this influenced his career and family life. (Quick answer — both hardly at all, and more than most could possibly understand. That’s a theme here.)
The medical and personal details are fascinating, and not just to ID and hematology-oncology doctors.
Has he been extraordinarily lucky, or terribly unlucky?
The answer, again, is both — listen and find out.
November 18th, 2018
HIV and HCV Treatment, Shorter Antibiotic Courses, Malaria-Sniffing Dogs, and Other ID and HIV Reasons to Be Grateful, 2018 Edition
As noted here before, I’m a big fan of Thanksgiving, a great excuse to get together with family and friends, and to eat a gargantuan amount of food.*
(*On this last point, non-U.S. citizens will wonder how this differs from any other time we “Americans” get together. Indeed, we are the Land of Giant Portions. Ever eat at a Cheesecake Factory? Yeesh.)
But the primary purpose of Thanksgiving is expressing gratitude, a very healthy impulse that makes everyone happier. And since this is an ID blog, here are bunch of things ID and HIV specialists can be thankful for over the past year:
- HIV treatment continues to improve. In the past year, we’ve seen approvals of bictegravir/FTC/TAF, darunavir/c/FTC/TAF, doravirine and doravirine/3TC/TDF, and ibalizumab. If that’s not enough, generic 3TC/TDF and EFV/3TC/TDF also gained FDA approval. Treatment now is so good with bictegravir and dolutegravir-based regimens that essentially 100% of patients taking these simple, well-tolerated regimens achieve viral suppression, and the IAS-USA Guidelines responded by making them the preferred treatments. Starting HIV treatment on the same day of diagnosis has moved from an interesting idea to a practical reality (especially with bictegravir/FTC/TAF and darunavir/c/FTC/TAF). Finally, the remarkably good results of the two-drug dolutegravir plus 3TC regimen in the GEMINI studies suggest that therapy can be even safer — and cheaper — than it is now.
- When it comes to the duration of antibiotic therapy, less is almost always better than more. A barrage of studies have shown that almost regardless of the condition, a shorter course of antibiotic treatment is the right choice. Shorter is better than longer both in the hospital and in the outpatient office. The take-home message from these data should be that clinicians can use their clinical judgment to decide how long to treat someone, not relying on some arbitrarily chosen treatment course — and yes, I mean even a treatment course defined by an ID doctor.
- It is becoming increasingly clear that oral antibiotics can substitute for intravenous in almost every situation. In the much-discussed POET study, stable patients with endocarditis who completed treatment with oral therapy did just as well as those who completed treatment with IVs. (And here’s a big thank you to Dr. Grace Farris for allowing us to use her wonderful journal club cartoon.) In another study, linezolid — price dropping rapidly — was a fine option for completing treatment in uncomplicated Staph aureus bacteremia. OVIVA study, where are you?
- Hepatitis C treatment has become staggeringly simple. I’m going to estimate that 99.156% (approximately) of treatment-naive patients could be successfully treated with either 1) sofosbuvir/velpatasvir, one pill daily for 12 weeks or, 2) glecaprevir/pibrentasvir, three pills daily for 8 weeks (sometimes 12). Go ahead and abbreviate SOF/VEL and GLE/PIB, if you find saying them a mouthful. Both are pan-genotypic, safe, well tolerated, and incredibly effective, and the price has dropped dramatically since the crazy days of 2014. In other words, these are the “must know” HCV treatments; the rest are optional. Here are a couple of slides I’ve made to describe a few of the medical issues in the choice between them — it’s a very short lecture — since most of the time, either one will do.
- Undetectable = Untransmittable, even in the highest-risk patients. The first publication of the PARTNER study of “condomless sex” among HIV serodiscordant couples didn’t have enough MSM couples to make a confident statement about the risk in this population. No longer — in a follow-up presentation this summer in Amsterdam, 783 MSM couples contributed data including 1596 couple-years of follow-up, and 76,991 (impressive precision!) condomless sexual encounters — and there were still zero transmissions. Giving this U=U message to our patients never gets old.
- Treatment of African trypanosomiasis will very soon be safer, easier, and more effective. Current treatment for this life-threatening parasitic infection — better known as “African sleeping sickness” — involves a complicated infusions with the drug eflornithine (administered in a hospital) along with nifurtimox. By contrast, fexinidazole (just approved) is given as one pill daily for 10 days and is both more effective and less toxic. Thanks to several European countries, the Bill and Melinda Gates Foundation, Doctors Without Borders, and other donors for funding the study that led to its approval, which was coordinated by the Drugs for Neglected Diseases Initiative.
- Preventive therapy for latent TB is getting shorter. In a large randomized study, a 4-month rifampin regimen was just as effective as 9 months of INH, and was both less toxic and more likely to be completed. In people with HIV, 1 month of daily rifapentine/INH prevented TB as well as 9 months of daily INH. The latter study hasn’t yet been published — and included some “high risk” patients who lacked positive skin testing or IGRA studies — but one is hopeful that the favorable results would translate to people without HIV as well.
- There continue to be ID-related studies that test the diagnostic skills of dogs. I guess they can’t reliably sniff out C. diff after all, but they did pretty well identifying malaria from the socks of children. (Yes, that’s what they were asked to do. Dogs will sniff anything.) I don’t actually think any of these dog diagnostic strategies will ever become standard-of-care (alas), but the videos and pictures released with each study sure are cute.
What are you grateful for this Thanksgiving?
November 12th, 2018
Sharing Radiology Images Across EMRs Is Frustratingly Terrible — and It Doesn’t Have to Be This Way
In the United States, any person who has tried getting their own (or their patient’s) radiology images from another hospital or practice will find this brief anecdote painful:
https://twitter.com/sarahkliff/status/1059518100821983233
Here are several obvious reasons why the CD-ROM — briefly the darling of large-data transfer — is a truly terrible way to share radiology images in 2018:
- They require physical transfer. Remember the term “snail mail”? Do people still say that?
- They are slow. When you bring a CD-ROM down to your friendly radiologist to review the scans, also bring a good book — you will be waiting awhile for the images to load. Zzzzz …
- There’s no universal software to read them. Watching even the most tech-savvy radiologist trying to extract images from these disks is proof enough that this is a horribly outdated technology.
- The blank disks are disappearing. When was the last time you purchased a “spool” of these things? Back when Napster was a thing?
- The drives are disappearing from computers. They’ve been gone from most laptops for years. Desktop computers, especially the mass-market, small-form ones used in hospitals, often lack them as well.
- Hospitals spend significant time and money transferring images from CD-ROMs into their EMRs. It works like this: you walk the disk down to wherever the uploading machine is located. You fill out some forms. You hand the disk over. It goes into queue with other disks. Later — hard to predict exactly when, could be later that day, or tomorrow, or next week, but never during your patient’s office visit — you can view the images in your patient’s medical record. However, sometimes (and this has happened more times than I can count), the disk is unreadable, or doesn’t even have images at all — only the radiology report, and not the actual images. Gak.
- No one knows whether CD-ROM disks should be spelled “disk” or “disc.” Discuss among yourselves — I’m going with “disk.”
Non-clinicians might wonder, what’s the big deal if you can get the radiology report? Isn’t that what “Care Everywhere©™” does?
Sure, having the report is better than nothing. But in complex cases, and when making difficult diagnostic or therapeutic decisions, it is always better to review the actual images — preferably with a radiologist specially trained in the involved anatomic region.*
(*ID doctors do a lot of this kind of thing. This probably explains why getting ID consults on complicated cases is associated with better outcomes. And it definitely contributes to why we’re typically bottom-feeders in an RVU, procedure-based world. How do you bill for time spent chasing down images and reviewing them with radiology?)
Of course it doesn’t have to be this CD-ROM way. If there ever were an irrefutable argument for the benefits of digitizing medical information, the switch from hard copy “films” to digital images would have to be right near the top. Think of how far we’ve come from the days of searching for x-ray films that, not surprisingly, would disappear in direct proportion to how interesting the case, or how sick the patient.
That’s why the current CD-ROM strategy is so frustrating. Never mind that a faster and more reliable technology (the USB flash drive) has been available for years. Though cheap and ubiquitous, and better than CD-ROMs, USB flash drives would also require physical transfer.
The solution, of course, is to put the images on the web — which is apparently what many non-U.S. hospitals have been doing for years:
We were consulted on a patient transferred from a hospital in Thailand. We needed to review his MRIs. The doctors there kindly emailed us a link and, bingo, we could see them all (along with his whole medical record).
This was at least 5 years ago. https://t.co/I3ARLxEtO9
— Paul Sax (@PaulSaxMD) November 9, 2018
So here’s what I recommend we do, starting now:
- Patients scheduled for imaging fill out a form while they are waiting asking if they want their images available for review by the clinicians caring for them.
- If the answer is Yes — and I imagine it would be for all but the most paranoid individuals — then after the scan is done, they are provided a secure link. It can be communicated by email, text, a post-procedure print out, or all of the above.
- In order to make the link work, they need to click on it and verify that it can be accessed by others.
- They then get to choose the variety of ways others can access it — secure password? Two-step verification?
- The patients can then share the link with whomever they like.
There, wasn’t that easy?