A gloved hand holding an empty orange prescription vial against a pale marble background

How Do I Switch From Compounded Semaglutide or Tirzepatide to Wegovy or Zepbound?

On November 6, 2026November 5, 2026 By Dr. Scott RennieIn Weight Loss & Metabolic HealthLeave a comment

The short answer: you switch by getting a new prescription for Wegovy, Zepbound, or one of the new pills, and your prescriber picks the starting dose. A compounded dose does not convert cleanly to a brand dose, so the number on your old vial is a clue rather than an instruction.

If you have been off for more than a couple of weeks, or nobody can tell you exactly what was in your vial, the safest default is to restart near the bottom of the label schedule and climb. That costs you some weeks. Guessing high can cost you days of vomiting.

Why your compounded supply is drying up

Pharmacies were allowed to copy semaglutide and tirzepatide while the brand products were in shortage. The shortages ended. FDA declared the tirzepatide shortage resolved on October 2, 2024, confirmed that on December 19, 2024, and declared the semaglutide shortage resolved on February 21, 2025. The enforcement grace periods for state-licensed pharmacies and outsourcing facilities then closed in steps through the spring of 2025 (1).

2026 has added pressure from the other direction. FDA’s April 1, 2026 update says plainly that tirzepatide and semaglutide are on neither the 503B bulks list nor the drug shortage list (1). On April 30 the agency proposed excluding semaglutide, tirzepatide, and liraglutide from that bulks list permanently, with a comment period that closed July 30 after an extension (2). And on March 3 it sent warning letters to 30 telehealth companies over marketing that implied their compounded products were the same as approved ones (3). As of early November 2026 there was no final decision on the bulks proposal, so treat that piece as pending.

A narrow lane survives. FDA will not treat a compounded product as essentially a copy when a prescriber determines and documents a change that makes a significant difference for an identified patient (1). It also says a semaglutide plus B12 product can still be judged a copy when both strengths fall within 10 percent of the commercial products (1). If your supplier is shipping less, relabeling, or closing its compounded program, this is the reason. It is a regulatory story and says nothing about you.

I covered the safety checklist and the history in Compounded Semaglutide and Tirzepatide: What to Know Now. This post is the practical follow-up for someone who has to make the move.

Why can’t I just convert my compounded dose?

Because the label on a compounded vial and the label on a pen are written in different languages, and sometimes describe different molecules.

Start with units. Compounded product usually arrives in a multi-dose vial that you draw up with a syringe. A poison control case series in the Journal of the American Pharmacists Association described three patients with adverse events after compounded semaglutide, two of them 10-fold dosing errors, and one who was dosing in milliliters and units instead of milligrams. One needed an antiemetic and intravenous fluids (4). A pen clicks to a fixed dose. A syringe asks you to do arithmetic.

Then the ingredient. FDA states that semaglutide sodium and semaglutide acetate are different active ingredients from the one in the approved drugs, and says it knows of no lawful basis for using those salts in compounding. It has also logged reports of patients getting more product per dose, dosing more often, or titrating faster than the label allows (5). As of May 31, 2026, the agency counted 990 adverse event reports tied to compounded semaglutide and more than 730 tied to compounded tirzepatide, and it says underreporting is likely (5).

So a compounded vial labeled “2.5 mg weekly” and a Zepbound 2.5 mg pen share a number. They do not share a guarantee.

Restart low, or match the dose?

No label tells you how to switch from a compounded product, because none of them contemplate it. What the labels do tell us is how to restart after a gap, and that is the closest evidence we have.

The Wegovy label says that if two or more consecutive weekly doses are missed, reinitiate dose escalation at a lower dosage to reduce the risk of gastrointestinal adverse reactions (6). The Zepbound label addresses only a single late dose (7), but Lilly’s medical information page describes what the SURMOUNT trials did: after one or two missed doses, treatment restarted at the same dose if it had been well tolerated, and after three or more, it restarted at 5 mg regardless of the prior dose (8). The Foundayo label sets its trigger at seven or more consecutive missed daily doses, again with reinitiation at a lower dosage (9).

Read together, gap length is the first variable. A short gap on a known, well-tolerated dose makes matching defensible. A longer gap pushes you down the schedule. The second variable is certainty. If the pharmacy is unknown, the concentration is unclear, or the patient cannot say how many units they drew, leaning low is the safer choice. My view is that the two errors are not symmetrical. Starting too low delays the result by a few weeks. Starting too high can mean days of vomiting and, rarely, a trip for fluids.

For scale, here is what low looks like. Wegovy starts at 0.25 mg and reaches 1 mg at week 9, with each step lasting four weeks. Zepbound starts at 2.5 mg and reaches 5 mg at week 5 (6,7). The Wegovy label also allows delaying escalation by four weeks if a dose is not tolerated (6).

There is no single number I use. If someone has been getting their medication through a reliable compounding pharmacy, that helps, because there is likely to have been no gap in dosing. If the pharmacy is unknown, or there is some question about how they have felt on the compounded drug, I adjust. Where I start depends on how the patient is doing, which pharmacy and which medication they were using, and whether they have been sick to their stomach or have seen no change in their hunger at all. If they were having side effects on the compounded product, such as nausea bad enough to need a lot of ondansetron, or severe diarrhea or constipation, but they want to try a brand-name medication to see whether that improves, I start them on a lower dose than they were taking. If they have been on a particular compounded dose for a long time and are not losing any weight, I might start higher than the label’s starting dose. It really depends.

What I would gather before the video visit

On a video visit I cannot pick up the vial, so the vial has to come to me as photographs. The useful list is short: the pharmacy label showing drug, concentration, and lot; the syringe type; how many units the patient draws; the date and size of the last injection; the highest dose reached; and every side effect along the way. Patients who bring that make the dose decision easy. Patients who bring “it was the blue one” make it a judgment call.

What I ask for is a photo of the current prescription. It has to show the name of the medication, the dose, the date it was prescribed, and the prescribing doctor. It also has to show any additives, because compounded versions sometimes include them, and I want to see exactly what has been in that vial.

What restarting will feel like

Expect the early weeks of a new start, even if you were on a high dose before. The labels build their escalation schedules specifically to reduce GI side effects (6,7), and tolerance you built on a product of uncertain strength is not a reliable credit. Nausea, early fullness, and constipation are the usual complaints. If a step is rough, hold there rather than push, which the Wegovy label explicitly permits (6).

Some patients feel like they are going backward when the dose drops. They are not. The goal is to land on a dose you can verify, then climb to the effect you want. Persistent vomiting, or an inability to keep fluids down, is a reason to call. My side effects post has the incidence numbers.

What will the brand version cost?

Less than many people expect, though prices move. As I write this, NovoCare lists the Wegovy pen at $199 a month for new patients on the first two fills of the 0.25 mg and 0.5 mg doses, then $349 a month for 0.25 through 2.4 mg, and the Wegovy pill from $149 a month, rising to $299 at the 9 mg and 25 mg strengths. Lilly lists Zepbound self-pay starting at $299 a month for the KwikPen or four vials, and the $449 price on higher doses, pen or vial, requires a refill within 45 days, otherwise it goes to $499 at 7.5 mg and $699 from 10 mg up (11). Novo’s offer runs through December 31, 2026 and it says pricing will be updated after that (10). Lilly’s page lists no end date.

Both routes need a prescription. Novo’s prices apply when it is filled through NovoCare Pharmacy or with its savings offer (10). If you have insurance, check coverage and prior authorization first, since a covered fill beats any cash price. My cash pay prices post tracks the full ladder, including Medicare and the other federal programs.

Pills as a way out

Needles and syringes are part of why some people used compounded vials in the first place, and now there are two oral choices. Wegovy tablets start at 1.5 mg daily for 30 days, then 4 mg, then 9 mg, then 25 mg (6). Foundayo starts at 0.8 mg, steps to 2.5 mg after at least 30 days, then 5.5 mg, with room to 9, 14.5, or 17.2 mg (9). The Wegovy label does include a rule for moving from the 2.4 mg injection to the 25 mg tablet. Nothing in either label covers moving from a compounded injection, so the same logic applies: start at the bottom unless there is a good reason not to (6,9). I compared the two pills in Foundayo vs. the Wegovy Pill.

For most people it comes down to two things, preference and coverage. If insurance will help pay for Foundayo or the Wegovy pill, that is probably the most important deciding factor. If the patient is paying cash, that changes the math. And if someone doesn’t like injections, a pill usually makes more sense, as long as it is affordable and it keeps them on the medication.

Gray-market sellers and “research use” vials

When a compounded supplier folds, the next search result is often worse. FDA has warned companies that sold semaglutide, tirzepatide, and related peptides falsely labeled “for research purposes” or “not for human consumption” while supplying dosing instructions to consumers. It also says it is aware of counterfeit Ozempic, and advises buying only from state-licensed pharmacies (5). A price far below the brand cash prices above is the tell. Use FDA’s BeSafeRx tools if you are unsure about a site.

The same caution covers telehealth ads. FDA’s position is that compounded drugs are not FDA-approved (3), so a site promising “the same as Wegovy” is making a claim the agency has already penalized.

If you are the patient

Do not stretch your last vial to bridge the gap, and do not mix products. Photograph the label now, write down your last dose date, and book the switching visit before you run out. Ask your prescriber which brand product fits your coverage and budget, and expect a lower starting dose than you were on.

If you are the clinician

Document the compounded product, the pharmacy, the last dose date, and your reasoning for the starting dose, since no label supports a direct conversion. Counsel on the GI risk of reinitiation (6,7). Remember that state-licensed pharmacies that are not outsourcing facilities are not required to report adverse events to FDA, so your MedWatch report may be the only one (5).

The Bottom Line

Moving from compounded semaglutide or tirzepatide to Wegovy, Zepbound, or an oral option means a new prescription and a starting dose chosen for you, because compounded doses do not convert cleanly to brand doses.

Gap length and certainty drive the choice. A short gap on a verified, well-tolerated dose can be matched. A longer gap or an uncertain vial points toward the label starting dose and a four-week-step climb. Costs for the brand products are lower than they were, and the pills give needle-averse patients a path. Avoid gray-market sellers entirely.

Sources

U.S. Food and Drug Administration. FDA clarifies policies for compounders as national GLP-1 supply begins to stabilize (current as of April 1, 2026). https://www.fda.gov/drugs/drug-alerts-and-statements/fda-clarifies-policies-compounders-national-glp-1-supply-begins-stabilize
U.S. Food and Drug Administration. FDA proposes to exclude semaglutide, tirzepatide, and liraglutide on 503B bulks list. April 30, 2026. https://www.fda.gov/news-events/press-announcements/fda-proposes-exclude-semaglutide-tirzepatide-and-liraglutide-503b-bulks-list
U.S. Food and Drug Administration. FDA warns 30 telehealth companies against illegal marketing of compounded GLP-1s. March 3, 2026. https://www.fda.gov/news-events/press-announcements/fda-warns-30-telehealth-companies-against-illegal-marketing-compounded-glp-1s
Lambson JE, Flegal SC, Johnson AR. Administration errors of compounded semaglutide reported to a poison control center: case series. J Am Pharm Assoc (2003). 2023;63(5):1643-1645. PMID 37392810. https://pubmed.ncbi.nlm.nih.gov/37392810/
U.S. Food and Drug Administration. FDA’s concerns with unapproved GLP-1 drugs used for weight loss (current as of October 1, 2026). https://www.fda.gov/drugs/drug-alerts-and-statements/fdas-concerns-unapproved-glp-1-drugs-used-weight-loss
Wegovy (semaglutide) injection and tablets, prescribing information (revised June 2026). DailyMed. https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=ee06186f-2aa3-4990-a760-757579d8f77b
Zepbound (tirzepatide) injection, prescribing information. DailyMed. https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=487cd7e7-434c-4925-99fa-aa80b1cc776b
Eli Lilly Medical Information. What should a patient do if they miss a dose of Zepbound (tirzepatide)? https://medical.lilly.com/us/products/answers/what-should-a-patient-do-if-they-miss-a-dose-of-zepbound-tirzepatide-209169
Foundayo (orforglipron) tablets, prescribing information. DailyMed. https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=8ac446c5-feba-474f-a103-23facb9b5c62
Novo Nordisk. NovoCare: Wegovy savings and self-pay options. https://www.novocare.com/wegovy/savings-card.html
Eli Lilly and Company. Zepbound savings and insurance options. https://zepbound.lilly.com/savings

A smartphone with a blank screen resting on a beige linen sofa cushion

What Should You Do If a Lab Result in Your Patient Portal Looks Critical?

On October 30, 2026October 8, 2026 By Dr. Scott RennieIn Healthcare System & TechnologyLeave a comment

If a result in your patient portal is marked critical, or you see a number that looks dangerously high or low and you feel unwell, do not wait for a callback. Call the clinician who ordered the test right now. If you also have chest pain, trouble breathing, confusion, fainting, or severe weakness, call 911 or go to the nearest emergency department.

If you feel fine and the result is only flagged high or low, it is usually not an emergency. A message or phone call during business hours is the right speed.

The hard part is telling those two situations apart when the result lands on your phone at 9 p.m. and nobody has explained it. So here is how it works, which numbers matter, and what to do.

Why results show up before anyone has called you

The 21st Century Cures Act information-blocking rules changed the default. The prohibition began applying on April 5, 2021, and from October 6, 2022 it covered the full electronic health information in the designated record set rather than a narrower list of data elements (1). Federal guidance does not tell health systems to push every result to you the moment it posts. It does say that once a patient asks for access, the response has to be timely, and in practice you may see a result at about the same time the ordering clinician does (2).

Holding results back for clinician review is hard to defend as a blanket policy. The same federal FAQs say a routine several-day delay across a broad range of results does not qualify for the Preventing Harm exception (3). An individual clinician can make a specific determination that releasing one result would put someone’s life or physical safety at risk, and a patient can ask for a delay on their own results (3,4). Enforcement is no longer theoretical either. Provider disincentives tied to Medicare programs were finalized in July 2024, and HHS announced in September 2025 that information-blocking enforcement is a priority (5).

Patients, for what it is worth, mostly like this. In a four-center survey of 8,139 people, 95.7 percent wanted results in the portal immediately, even before a clinician reviewed them. But 16.5 percent of people with not-normal results reported being more worried afterward, against 5.0 percent of those with normal results (6). That worry is what this post is for.

What does “critical” actually mean?

A critical value is a lab result that suggests something imminently life-threatening. Federal lab regulations under CLIA require the laboratory to immediately alert the person who requested the test when a result indicates an imminently life-threatening condition, or a panic or alert value (7). The Joint Commission describes a critical value as one requiring immediate communication of results (8).

Notice who gets the call. The lab calls the ordering clinician, not you. One hospital laboratory described its rule as a technologist phoning the ordering practitioner within 30 minutes, with escalation if that person cannot be reached (8). Each lab builds its own list and its own cutoffs, and accreditors including the College of American Pathologists and the Joint Commission expect an alert system to exist (8). That obligation runs separately from whatever your portal shows you.

What the portal shows is a different matter. ONC’s safety guide for test result reporting recommends that critical values be flagged differently from ordinary abnormal values in the record clinicians use (9). Whether your portal does the same depends on the system. Do not count on a special color or the word “critical” appearing.

Flagged does not mean dangerous

Reference ranges are built from the central 95 percent of results in a healthy group. By definition, that labels 5 percent of healthy people as abnormal on any single test (10). Order a panel of 20 tests and about 64 percent of healthy people will have at least one result flagged (10). MedlinePlus puts it plainly: it is common for healthy people to sometimes have results outside the reference range (11).

So the red H or L next to your result is a prompt to look, not a verdict. A potassium of 5.6 mmol/L, for example, sits in the mild band (5.5 to 5.9) of the UK Kidney Association’s hyperkalaemia grading, and the same guideline says most mild or moderate cases found in the community can be managed without admission unless the person is acutely unwell (12). A sodium of 132 falls in the mild hyponatremia band, 130 to 135 (13). Both are real findings that deserve a conversation. Neither is a 911 call in a person who feels well.

One more trap. A hemolyzed blood sample can make potassium read falsely high, and that guideline recommends paired samples when this pseudo-hyperkalaemia is suspected (12). A surprising potassium in someone with no symptoms and normal kidney function is a reasonable one to repeat before anyone panics.

Which numbers can be truly urgent?

Labs call a wide range of analytes. Electrolytes, especially potassium, hemoglobin, high blood glucose and positive blood cultures were all on one laboratory survey’s list of commonly called values (8). I am not going to hand you a universal cutoff for each, because your lab sets its own. But the published guidelines give a sense of where the danger starts.

Potassium first. The UK Kidney Association grades severe hyperkalaemia as 6.5 mmol/L or higher and recommends urgent hospital assessment for anyone with that level found in the community (12). Sodium next. The European hyponatraemia guideline calls a serum sodium below 125 mmol/L profound, and notes that symptoms become more common below that level. Nausea, confusion and headache count as moderately severe symptoms. Vomiting, seizures, deep sleepiness and coma count as severe (13).

Glucose runs both directions. The American Diabetes Association defines level 2 hypoglycemia as below 54 mg/dL, and level 3 as an event with altered mental or physical status that needs another person’s help, whatever the glucose number says (14). On the high side, the ADA says to seek immediate care if you cannot keep fluids down, have ongoing vomiting, or have altered mental status (14). Those symptoms matter more than the number.

Hemoglobin is context-heavy. In the 2023 AABB guideline, the transfusion threshold for stable hospitalized adults is below 7 g/dL (15). That is a transfusion trigger and not a lab’s critical cutoff, but it tells you what kind of low is alarming. A slowly developed anemia and a bleed produce very different pictures at the same number.

Troponin deserves its own warning because patients see it more often now. Each assay has its own 99th percentile cutoff, reported in ng/L, and a value above it means myocardial injury. It is sensitive for a heart attack but not specific for one (16). If you see an elevated troponin in your portal and you have chest pressure, shortness of breath, or pain spreading to the arm or jaw, call 911 immediately (17). My view is that an unneeded trip to the emergency department is a much better outcome than a delayed one.

Some results always get a phone call from me: a critical blood sugar, a critical potassium, a critical kidney function result, or another electrolyte in the critical range. It is easier to reach someone during office hours, but I treat a critical result as urgent no matter what time it comes in. A critical number can occasionally be a mistake, too. A lab error, or a hemolyzed sample, where red cells broke open after the draw, can show an artificially high potassium. That is one reason earlier results help so much. For new patients I try to get their previous lab work, so I have something to compare against when I order new tests. It isn’t always possible.

What to do, in order

Start with how you feel, not with the number. Any chest pain, trouble breathing, fainting, new confusion, seizure, or severe weakness means 911, whatever the result says.

If you feel okay, check the reference range printed on your own report and read the lab’s comment, if there is one. Then call the office that ordered the test. Portal messages are the wrong tool for anything you think might be urgent. Pick up the phone.

After hours, call anyway and ask for the on-call clinician. If no one can be reached and the value is in the range described above (a potassium of 6.5 or higher, or a sodium under 125), go to an emergency department. A drawn glucose is old news by the time you read it, so for glucose, how you feel right now decides. Bring the result, your medication list, and the name of the ordering clinician.

Do not wait for the next business day with a severe value because you assume someone has already been called. The lab alert exists, but a phone number can be wrong and a message can sit in a queue.

When a virtual clinician ordered the test

Telehealth adds a wrinkle. The lab you walked into is probably near your home, while the clinician who ordered the test may be in another city or time zone. The ONC guide says the ordering clinician is responsible for follow-up of outpatient results unless that responsibility is delegated to someone else, and the delegation should be documented and accepted (9). That is the right principle. You want to know that someone actually holds it.

Ask which phone number the lab has on file for the ordering clinician, and whether that number is monitored when the clinician is not in a visit. Make sure the lab has your current cell number. And know where you are during a video visit, because a clinician who needs to send help has to know your physical address.

When a lab I ordered comes back critical, I try to reach the patient directly, especially after hours, because they need to know. They may know before I do, since the lab sometimes releases the result to them first. I also try to loop in their primary care physician if they have one. And a local emergency contact in the chart helps in the situations where the patient can’t be reached.

Ask before the blood draw

The single best question is “Who calls me if a result is critical, and from what number?” Ask it when the test is ordered, not when the portal pings you.

Follow with three more. Will I see results before you have read them? Who covers if you are unavailable? And if I would rather talk first, can the release wait? Federal guidance treats a patient’s own request to delay as likely not an interference, as long as the delay lasts no longer than necessary to meet that request (4). Some people genuinely prefer to hear a frightening result from a person first. That is a legitimate choice.

For patients, and for clinicians

If you are the patient

Know your own phone number on file, keep your ringer on after a blood draw, and decide ahead of time who you will call. Write the question “who calls me if it is critical” in your notes app before the next lab visit. Treat symptoms as the deciding factor.

If you are the clinician

Tell patients when you order that results may post before you have seen them, since the evidence shows abnormal results raise worry more often (6). Keep your critical-value contact information current with every lab you use, and make sure patients know who is covering when you are not. The SAFER guide asks organizations to have policies for timely patient notification of both normal and abnormal results, with the timeliness monitored (9).

When I order labs, I tell patients which tests are on the list. For my patients that usually means a fasting lipid profile, a comprehensive metabolic panel, a TSH, a hemoglobin A1c, and a high-sensitivity C-reactive protein, plus two special lipid tests: lipoprotein(a), which the 2026 ACC/AHA dyslipidemia guideline recommends measuring at least once in every adult (18), and apolipoprotein B.

The Bottom Line

A result marked critical, or a very high or low number paired with symptoms, means call the ordering clinician now, and call 911 for chest pain, trouble breathing, confusion, or fainting. A result that is merely flagged high or low is common even in healthy people and can usually wait for a conversation.

Cures Act rules mean you will often see results first, so ask before the blood draw who calls you if something is critical, from what number, and what happens when they cannot reach you. For telehealth patients, that answer matters even more.

Sources

Department of Health and Human Services. Information Blocking and the ONC Health IT Certification Program: Extension of Compliance Dates and Timeframes in Response to the COVID-19 Public Health Emergency. Federal Register, November 4, 2020. https://www.govinfo.gov/content/pkg/FR-2020-11-04/html/2020-24376.htm
ASTP/ONC. FAQ: Are actors (for example, health care providers) expected to release test results to patients through a patient portal or API as soon as the ordering clinician can see them? https://healthit.gov/faq/are-actors-example-health-care-providers-expected-release-test-results-patients-through-patient/
ASTP/ONC. FAQ: Would the Preventing Harm Exception cover a blanket several-day delay in the release of laboratory or other test results? https://healthit.gov/faq/would-preventing-harm-exception-cover-blanket-several-day-delay-release-laboratory-or-other/
ASTP/ONC. FAQ: Can an actor grant a patient’s request to delay the release of the patient’s test results? https://healthit.gov/faq/can-actor-grant-patients-request-delay-release-patients-test-results-eg-laboratory-or-image/
Liebler AJ, Sullivan S, Boyer MD, Doe K. Information Blocking Enforcement Enters a New Phase: What Digital Health Stakeholders Should Expect in 2026. Alston & Bird Health Care Advisory. February 23, 2026. https://www.alston.com/en/insights/publications/2026/02/information-blocking-enforcement-2026
Steitz BD, Turer RW, Lin CT, et al. Perspectives of patients about immediate access to test results through an online patient portal. JAMA Netw Open. 2023;6(3):e233572. https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2802672
42 CFR 493.1291. Test report (CLIA), paragraph (g). https://www.law.cornell.edu/cfr/text/42/493.1291
Association for Diagnostics & Laboratory Medicine. Critical result notification preferences. Clinical Laboratory News. September-October 2024. https://myadlm.org/cln/Articles/2024/SeptemberOctober/Critical-result-notification-preferences
ONC. SAFER Guides: Test Results Reporting and Follow-Up. August 2024 self-assessment. https://www.healthit.gov/sites/default/files/topiclanding/2025-01/8.%20Test%20Results%20Reporting%20Final.pdf
Doles N, Ye Mon M, Shaikh A, et al. Interpreting normal values and reference ranges for laboratory tests. J Am Board Fam Med. 2025;38(1):174-179. https://pmc.ncbi.nlm.nih.gov/articles/PMC12096374/
MedlinePlus. How to Understand Your Lab Results. Updated September 4, 2025. https://medlineplus.gov/lab-tests/how-to-understand-your-lab-results/
UK Kidney Association. Clinical Practice Guideline: Management of Hyperkalaemia in Adults. Updated July 2026. https://www.ukkidney.org/sites/default/files/documents/FINAL%20VERSION%20-%20UKKA%20CLINICAL%20PRACTICE%20GUIDELINE%20-%20MANAGEMENT%20OF%20HYPERKALAEMIA%20IN%20ADULTS%20-%20UPDATED%20JULY%202026_0.pdf
Spasovski G, Vanholder R, Allolio B, et al. Clinical practice guideline on diagnosis and treatment of hyponatraemia. Eur J Endocrinol. 2014;170(3):G1-G47. https://doi.org/10.1530/EJE-13-1020
American Diabetes Association Professional Practice Committee. 6. Glycemic goals, hypoglycemia, and hyperglycemic crises: Standards of Care in Diabetes-2026. Diabetes Care. 2026;49(Suppl 1):S132-S149. https://doi.org/10.2337/dc26-S006
Carson JL, Stanworth SJ, Guyatt G, et al. Red blood cell transfusion: 2023 AABB international guidelines. JAMA. 2023;330(19):1892-1902. https://doi.org/10.1001/jama.2023.12914
Sandoval Y, Apple FS, Mahler SA. High-sensitivity cardiac troponin and the 2021 guidelines for acute chest pain. American College of Cardiology. July 14, 2022. https://www.acc.org/latest-in-cardiology/ten-points-to-remember/2022/07/14/18/12/high-sensitivity-ctn-and-2021-chest-pain
Centers for Disease Control and Prevention. About Heart Attack Symptoms, Risk, and Recovery. October 24, 2024. https://www.cdc.gov/heart-disease/about/heart-attack.html
American College of Cardiology. Lower Sooner: How the 2026 Dyslipidemia Guideline Changes Practice. July 1, 2026. https://www.acc.org/latest-in-cardiology/articles/2026/07/01/01/prioritizing-health

A row of worn cast iron kettlebells on a gym floor

What Happens If You Stop Taking a GLP-1?

On October 16, 2026 By Dr. Scott RennieIn Weight Loss & Metabolic HealthLeave a comment

Somebody asks me this almost every week, usually some version of the same sentence. “How long do I have to stay on this?”

It is a fair question and it deserves a straight answer rather than a lecture. So here is what the data says happens, what I do about it, and the handful of situations where I think stopping is the right call.

Why you want to stop matters more than the fact that you want to stop

The first thing I ask is why.

The answers cluster into three groups, and they lead to three completely different conversations. Some people have hit their goal and believe they can hold it from here on their own. Some are looking at the monthly cost and doing math. And some are uneasy about taking any medication for years on end, the same way people feel about a blood pressure pill or a statin.

None of those are unreasonable. But the plan for a patient who cannot afford the drug looks nothing like the plan for a patient who is worried about what it is doing to them over a decade, and lumping them together as “wants to stop” wastes the visit.

What the evidence actually shows

The honest version is that most of the weight comes back, and it comes back faster than people expect.

The best synthesis we have is a 2026 systematic review and meta-regression in eClinicalMedicine that pooled 48 studies (1). About 60 percent of the weight lost during treatment is regained within 52 weeks of stopping. The curve is steep early, slows, and flattens out around 60 weeks, with regain plateauing near 75 percent of what was lost. Read the other direction, roughly a quarter of the loss is still there.

The individual trials tell the same story. In the STEP 1 extension, participants had lost 17.3 percent of body weight on semaglutide. A year after the drug and the lifestyle support both stopped, they had regained 11.6 percentage points, leaving a net loss of 5.6 percent, and 48 percent were still at least 5 percent below where they started (2). In SURMOUNT-4, people who came off tirzepatide regained about 14 percent over the following year while those who stayed on lost another 6.7 percent, and 82 percent of the group that stopped gave back more than a quarter of their loss (3).

Here is the part that gets left out of the headlines, and the part I care about more than the scale. The metabolic improvements unwind first. Around 70 to 80 percent of the blood pressure improvement is gone within 12 weeks of stopping. About half of the HbA1c improvement is back by 8 to 12 weeks (1). The weight takes a year to come back. The blood pressure takes three months.

That ordering is the reason I do not treat stopping as a neutral event in someone with hypertension or diabetes.

It is physiology, not a character flaw

I want to be clear about what that regain represents, because patients are brutal with themselves about it.

These drugs work on appetite signaling. Take the drug away and the signaling returns to where it was, which means hunger returns to where it was. The body defends the higher weight. Regaining is the expected biological result of removing the treatment, not evidence that somebody got lazy in month four.

I say this out loud in visits because the shame around regain is what stops people from coming back and restarting when they need to.

Most people stop anyway

Worth knowing: stopping is the norm, not the exception.

In a cohort of 125,474 adults starting these medications, 46.5 percent of those with type 2 diabetes and 64.8 percent of those without had stopped within a year (4). An analysis of 60,000 insured patients presented at ENDO 2026 found about four in ten stopped in the first year, and nearly six in ten by two years (5).

But look at what happens next in that same data. Of the people who stopped, 41.5 percent restarted within a year and 58 percent within two (5). This is not a drug people take once and abandon. It is a drug people fall off and climb back onto, usually after the weight starts coming back.

One more number, and it is the one that changed how I think about this. An integrated health system that added structured pharmacist onboarding and real follow-up got its one-year discontinuation rate down to 3.9 percent (6). Compare that to the 46 to 65 percent above. Most of the stopping in this country is not a drug problem. It is an access, cost and support problem.

What I actually do: lower the dose rather than stop

When somebody has reached a healthy weight, my preference is not to stop the medication. It is to lower the dose to the amount that holds the weight steady instead of the amount that keeps driving it down.

I will say plainly that the evidence here is thin. There are no randomized trials of tapering or maintenance dosing after weight loss, only small observational reports (7). So this is clinical judgment rather than a protocol, and I tell patients that.

What I like about it is that it answers the three concerns at once. It costs less. It involves less drug for people who are uneasy about long-term exposure. And it keeps the appetite brake partly engaged, which is the thing that actually prevents the rebound.

When the reason is cost

This is the most common real reason, and it has the most workable solutions.

First, the cash-pay programs. Eli Lilly and Novo Nordisk both sell direct now, and the prices are meaningfully better than they used to be. For a lot of patients that alone closes the gap.

Second, insurance, which is more winnable than people assume. If a patient carries a weight-related condition, and most do, there is often a path to coverage. Diabetes, prediabetes, hypertension, high cholesterol, GERD, obstructive sleep apnea. Those are not footnotes. They are the clinical justification that gets a prior authorization approved. It is worth the paperwork.

Third, switch the drug rather than abandon the class. Semaglutide instead of tirzepatide. Liraglutide instead of either. Not identical, but real.

Fourth, Medicare, and this is the part most patients have not heard. Part D still cannot pay for a drug prescribed for weight loss alone, because that exclusion sits in the statute and the 2025 attempt to reinterpret it was never finalized. What changed is the Medicare GLP-1 Bridge, which went live July 1, 2026 and runs through the end of 2027. It puts Foundayo, Wegovy and the Zepbound KwikPen at $50 a month for eligible beneficiaries (8). The Zepbound vials and single-dose pens are not in it, which is catching people out. If a patient was told no before last summer, that answer is worth getting again.

And if none of that works, we move to what is affordable. Bupropion and naltrexone as generics. Metformin. These are not equivalent to a GLP-1 and I do not pretend otherwise. They are considerably better than nothing, which is the actual alternative on the table.

Protecting muscle

If weight is going to move in either direction, I want as much of it as possible to be fat.

I target protein intake around 1.5 grams per kilogram. I want resistance training at least twice a week, and 30 minutes of aerobic activity five days a week. That applies while losing, while maintaining, and especially after stopping.

The concern worth naming is that we do not know whether regained weight comes back in the same proportions it left. If fat returns faster than muscle, somebody can land at their old weight with worse body composition than they had before treatment. The data on that is genuinely unsettled, so I am not going to overstate it. The defense is the same either way: protein and resistance training.

When I bring up the long-term question, and why not sooner

This is where I part company with a lot of advice.

I do not open with “this is a lifelong medication.” Early on, we do not yet know whether the drug will even work for this person, and plenty of patients arrive with reservations already sitting in the back of their minds. Leading with a life sentence can talk somebody out of a treatment that might have changed their health, before anyone knows if it helps.

So I raise it as we approach maintenance, when there is something real to discuss: this is working, here is what holding onto it looks like.

When I support stopping

Not every stop is a mistake. I will back the decision in a patient who becomes pregnant, who cannot tolerate the side effects, who genuinely cannot afford it after we have worked the options, or who is not able to use it safely. That last one covers skipping the lab work, taking it in ways it was not prescribed, or developing disordered eating patterns such as bingeing and purging or restricting toward anorexia. It also covers losing more weight than is healthy and not sticking to the maintenance plan.

In those situations stopping is the correct medical decision, not a failure of resolve.

Restarting

People restart. It is common, and I treat it as a normal part of the course rather than a setback.

If somebody has been off for more than about a month, we go back to introductory dosing and step up again. Restarting at the dose you left off at is how people end up with a week of nausea they did not need to have.

For clinicians

Three things worth carrying into your own visits.

Ask the why before you negotiate the how, because cost, goal attainment and long-term safety worry need different responses.

Counsel on the cardiometabolic rebound, not only the weight. Your patient with hypertension will give back most of their blood pressure improvement inside three months, which is a faster and more clinically meaningful change than the scale (1).

And treat the discontinuation statistics as a system signal rather than a patient-adherence story. A 3.9 percent one-year discontinuation rate in a supported program, against 46 to 65 percent in the broader population, tells you where the problem actually lives (4,6).

The Bottom Line

Stopping a GLP-1 means most of the weight comes back, typically around 60 percent of it within a year, with the blood pressure and blood sugar improvements unwinding much faster than that.

That is physiology rather than failure. If you have reached a healthy weight, lowering the dose is usually a better move than stopping outright. If cost is the problem, there is more room to work than most people realize, between cash-pay pricing, prior authorization based on the conditions you already have, switching agents, and the Medicare GLP-1 Bridge if you are on Medicare. And if you do stop, protect your muscle, and come back without embarrassment if the weight starts climbing. Restarting is common and it works.

Sources

Budini B, Luo S, et al. Trajectory of weight regain after cessation of GLP-1 receptor agonists: a systematic review and nonlinear meta-regression. eClinicalMedicine. 2026. https://pmc.ncbi.nlm.nih.gov/articles/PMC13043475/
Wilding JPH, Batterham RL, Davies M, et al. Weight regain and cardiometabolic effects after withdrawal of semaglutide: The STEP 1 trial extension. Diabetes Obes Metab. 2022;24(8):1553-1564. https://pmc.ncbi.nlm.nih.gov/articles/PMC9542252/
Cardiometabolic Parameter Change by Weight Regain on Tirzepatide Withdrawal in Adults With Obesity: A Post Hoc Analysis of the SURMOUNT-4 Trial. JAMA Internal Medicine. https://jamanetwork.com/journals/jamainternalmedicine/fullarticle/2841273
Discontinuation and Reinitiation of Dual-Labeled GLP-1 Receptor Agonists Among US Adults With Overweight or Obesity. JAMA Network Open. https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2829779
Endocrine Society. More than half of those who stop GLP-1s restart within a year. ENDO 2026, presented June 14, 2026. https://www.endocrine.org/news-and-advocacy/news-room/2026/sontha-press-release-endo-2026
Health System Pharmacies Keep Patients on GLP-1 Therapy. American Journal of Managed Care. https://www.ajmc.com/view/health-system-pharmacies-keep-patients-on-glp-1-therapy
Scripps. Patients Maintained Weight Loss After Reducing GLP-1 Treatment Frequency. https://www.scripps.org/news_items/8102-scripps-patients-maintained-weight-loss-after-reducing-glp-1-treatment-frequency
Medicare GLP-1 Bridge: Information for Providers. Centers for Medicare & Medicaid Services. https://www.cms.gov/medicare/coverage/prescription-drug-coverage/medicare-glp-1-bridge/information-providers

A row of clear glass vaccine vials with a syringe resting against them on a grey background

Are You Still Protected From Measles? The 1963 to 1967 Vaccine Gap and What Adults Should Do

On October 8, 2026October 28, 2026 By Dr. Scott RennieIn Infectious DiseaseLeave a comment

Pennsylvania crossed a thousand measles cases this week. The state is at 1,078 across 41 counties, with 207 hospitalizations and five measles-associated deaths, and Lancaster County alone accounts for 401 of them. Nationally we are at roughly 3,880 cases. That is the highest annual total since 1991.

New York’s governor declared a disaster emergency on Monday so that paramedics, pharmacists and midwives can give vaccines. Florida’s health department held a hearing in Orlando on October 8 on dropping hepatitis B, chickenpox, Hib and pneumococcal from its school and daycare requirements. Measles is not among them. The MMR requirement is written into state statute, so only the legislature can touch it. Comment on the other four closed on October 19 and the rule has not been adopted.

Almost all of the coverage is about children, because that is where the cases are.

I have had one of these myself, by video. An unvaccinated child, high fever, body aches, and a rash whose appearance made me think measles before I thought of anything else. Looking like measles is not the same as being measles. Confirmation takes a swab for PCR and a blood sample for measles IgM, and neither of those travels down a camera, so that visit ended where it had to end, in a referral for testing in person.

That is the honest limit of the work I do. I could recognize it. I could not prove it.

The question I get by video is a different one, and it comes from adults. Some version of: am I still covered?

It is a better question than it sounds, because for a specific group of people the honest answer is no.

The 1963 to 1967 problem

There were two measles vaccines licensed in 1963. One was a live attenuated vaccine, which is the ancestor of the MMR we give now. The other was an inactivated vaccine, a killed-virus product that was pulled from the market in 1967 because it did not work well enough.

It did something worse than not working. Some people who got the killed vaccine and later met the real virus developed atypical measles syndrome, an unusually severe illness with high fever, an odd rash that starts on the extremities, and pneumonia.

So the guidance is specific. Anyone vaccinated before 1968 with killed measles vaccine, or with a vaccine of unknown type, should be revaccinated with at least one dose of live attenuated measles vaccine. If you are in a high-risk group, that becomes two doses.

The number of people affected is small. The CDC puts it at under five percent of adults. Immunize.org estimates fewer than a million Americans.

Here is the practical problem with a small number. Almost nobody born in that window knows which shot they got. The record, if it exists at all, is a line in a pediatrician’s file from sixty years ago, and that pediatrician has been retired for thirty of them. “Measles vaccine, 1965” does not tell you which product it was.

Which is why the rule says unknown type counts the same as known killed. If you cannot establish it was the live vaccine, you are treated as though it was not.

If you were born before 1957

You are presumed immune and you do not need anything.

That cutoff is not a guess about vaccines. It is a statement about how common measles was. Before the vaccine existed, essentially everyone caught it in childhood, so the assumption is that you already did, whether or not you remember it.

One exception, and it is the one that applies to me and to anyone else reading this who works in health care. For healthcare personnel, facilities should consider vaccinating staff born before 1957 who have no laboratory evidence of immunity and no laboratory confirmation of past disease. The presumption is good enough for the general public. It is not good enough for someone standing in a room with a febrile child.

Should you get a titer drawn?

Usually not, and this is where I disagree with how the question is often handled.

The CDC does not recommend routinely screening adults for measles immunity outside an outbreak area, and it does not recommend testing after vaccination to confirm it took. If you have documentation of two MMR doses, you are done. No blood draw adds anything.

A titer is useful in one specific situation, which is when you genuinely do not know your history and the result would change what happens next. Positive measles IgG means immune. Negative or equivocal means treat as non-immune, unless you meet one of the other criteria such as birth before 1957.

But consider the arithmetic before you order it. An MMR dose costs less than the antibody test in most places, takes one visit instead of two, and carries no penalty if you turn out to have been immune all along. An extra dose in someone who is already protected does nothing at all. It does not boost, it does not harm, it simply has no target.

So for most adults who are uncertain, getting the shot is faster and cheaper than proving you do not need it.

The exception worth naming is pregnancy. MMR is a live vaccine and is not given during pregnancy, so a woman who is pregnant and uncertain about her status gets the titer instead, and the vaccine afterward if she needs it.

About the death count, since you will see it argued about

Pennsylvania reports five measles-associated deaths this year. The CDC’s national count is two, and the CDC’s database does not say which states those two are in.

That gap is real and it is worth understanding, because it is being read as evidence of a cover-up in one direction and of inflated numbers in the other.

The mechanism is duller than either. The CDC moved to counting measles deaths from National Center for Health Statistics mortality data, which is built from death certificates and lists an underlying cause, rather than from state health department reports. NCHS data is designed to measure long-term trends across a population. It was not built to adjudicate individual infectious disease cases in the middle of an outbreak, and it runs on a slower clock, which is the substance of what the outside experts objecting to the change have said.

Pennsylvania uses a different definition: a laboratory-confirmed measles infection and death within thirty days of symptom onset, absent an obvious unrelated cause such as a car crash.

Both definitions are defensible. They will produce different numbers during a fast-moving outbreak, and the federal one will lag.

There is also a political layer, which I will describe rather than referee. The CDC initially included Pennsylvania’s first two deaths and then removed them pending review, after the Health Secretary publicly questioned whether measles had caused them. That detail was first reported by a Substack citing an anonymous CDC employee, so I hold it more loosely than the rest of this. The count later returned to two after NCHS verified one death. Pennsylvania has since conditioned its request for federal Epi-Aid assistance on the CDC publicly recognizing the state’s full count.

What I would say clinically is narrow. Of Pennsylvania’s five, two were unvaccinated adults, aged 18 and 40, and two were infants too young to be vaccinated. That pattern is what measles mortality has always looked like. It kills the unprotected and the too-young-to-protect, and it does so through pneumonia and encephalitis rather than through the rash anyone remembers.

As the infectious disease physician Amesh Adalja put it, to the public a gap like this looks like either incompetence or concealment. Neither is true here, and the fact that neither is true is not doing much to help.

What I would actually do

If you have records showing two MMR doses, nothing. You are protected and the protection does not fade.

If you were born before 1957 and do not work in health care, nothing.

If you were vaccinated between 1963 and 1967, or you have no idea what you got and you were born after 1957, get one dose of MMR. Do not spend three weeks trying to find a sixty-year-old paper record first.

If you are pregnant and uncertain, ask for the titer, and plan the vaccine for after delivery.

And if you have an infant under twelve months in a household traveling to an outbreak area, that is a real conversation with your pediatrician about an early dose, which is a thing that exists and is not widely known.

The Bottom Line

Measles is at a thirty-five year high, and the people getting sick are overwhelmingly children who were never vaccinated. That is a different problem from the one most adults are asking me about.

For adults, the short version is that the vaccine works, it does not wear off, and the only real gap is a four-year window in the sixties when a product that did not work well was in use. If you fall in that window or you simply cannot establish what you got, a single MMR dose closes it, and it is cheaper and faster than proving you were fine all along.

Sources

Pennsylvania Department of Health Provides Statewide Update on Measles Outbreak as Cases Reach 1,078 in 41 Counties. Pennsylvania Department of Health, October 7, 2026. https://www.pa.gov/agencies/health/newsroom/pa-department-of-health-provides-statewide-update-on-measles-out9
Measles cases surpass 1,000 in Pennsylvania amid the state’s largest outbreak in decades. ABC News, October 2026. https://abcnews.com/Health/measles-cases-surpass-1000-pennsylvania-outbreak-continues-spread/story?id=137002024
Pennsylvania confirms fifth measles-associated death as US outbreak grows. Al Jazeera, September 30, 2026. https://www.aljazeera.com/news/2026/9/30/pennsylvania-confirms-fifth-measles-associated-death-as-us-outbreak-grows
New York governor declares measles emergency. CT Mirror, October 7, 2026. https://ctmirror.org/2026/10/07/new-york-measles-emergency-hochul
Vaccine change may bring largest measles outbreak ever. WGCU, October 7, 2026. https://wgcu.org/health/2026-10-07/vaccine-change-may-bring-largest-measles-outbreak-ever
Measles Cases and Outbreaks. Centers for Disease Control and Prevention. https://www.cdc.gov/measles/data-research/index.html
Measles Vaccine Recommendations. Centers for Disease Control and Prevention. https://www.cdc.gov/measles/hcp/vaccine-considerations/index.html
Questions About Measles. Centers for Disease Control and Prevention. https://www.cdc.gov/measles/about/questions.html
Ask the Experts: MMR (Measles, Mumps, and Rubella). Immunize.org. https://www.immunize.org/ask-experts/topic/mmr/
Measles Prevention: Recommendations of the Immunization Practices Advisory Committee (ACIP). MMWR. https://www.cdc.gov/mmwr/preview/mmwrhtml/00041753.htm
CDC’s new tracking method for measles deaths puzzles experts. CIDRAP, University of Minnesota. https://www.cidrap.umn.edu/measles/cdc-s-new-tracking-method-measles-deaths-puzzles-experts
Measles is spreading across the US and officials can’t agree on the death toll. The Conversation, 2026. https://theconversation.com/measles-is-spreading-across-the-us-and-officials-cant-agree-on-the-death-toll-293234
Changes to CDC Vaccine Recommendations in 2025 and 2026. Congressional Research Service, IN12684. https://www.congress.gov/crs-product/IN12684

A woman seen from behind at a pharmacy counter talking with a pharmacist

Why Can’t You Get an Estrogen Patch? The 2026 Estradiol Shortage, Explained

On October 5, 2026October 5, 2026 By Dr. Scott RennieIn Menopause & Women's HealthLeave a comment

Several of my patients are having a hard time filling their estradiol patches right now. Not one pharmacy being out for a week. Calling four places, getting a different strength than was prescribed, driving across town, running out before the next box arrives.

So let us go through the explanations people are offering, in order, and then what the evidence actually supports.

Hypothesis one: everyone started using them at once

This one is true, and the numbers are not subtle.

Estradiol patch prescriptions went from about 594,000 in June 2024 to 1.6 million in May 2026 (1). Among women aged 45 to 54, prescribing rose 184 percent, and by February 2026 roughly one in twenty women in that age group had an estrogen prescription (2). Patches now account for a large share of all estrogen prescriptions, and they were never the volume product the supply chain was built around.

The accelerant was regulatory. On November 10, 2025, the FDA announced it was removing the boxed warnings about cardiovascular disease, breast cancer and dementia from menopausal hormone therapy, with Commissioner Marty Makary saying the agency intended to “stop the fear machine steering women away from this life-changing, even lifesaving, treatment” (3). Those warnings had suppressed use for two decades. Pull them off, and a very large number of women who had been told no for years hear a different answer in the same month.

I wrote about that decision when it happened and I thought it was defensible. I did not expect it to empty the shelves.

Hypothesis two: something broke in manufacturing

This one is mostly false, and it is worth saying so clearly because it is the version that frightens people.

There is no recall behind this. No contamination finding, no plant shutdown, no safety problem with the patches that are reaching pharmacies. The FDA’s position, repeated through the summer, is that all the manufacturers are producing at full capacity (4).

What is true is that patches are hard to make more of quickly. A transdermal patch is a drug-in-adhesive matrix that has to deliver a precise dose through skin for three or four days. The production lines are specialized, and companies describe expansion timelines in years rather than months. Add thin generic margins, and nobody is rushing to build a new line for a product that might be a fad.

So: not a manufacturing failure. A manufacturing capacity ceiling, which is a different thing and gets solved on a much slower clock.

Hypothesis three: RFK Jr. and the administration did this

This is the one I keep seeing in comment sections, and the honest answer is more interesting than either the accusation or the denial.

There was no sabotage. Nobody restricted estrogen. If anything, this administration’s FDA pushed hormone therapy harder than any in twenty years, and the demand surge is a direct downstream consequence of that push. You cannot simultaneously argue that they suppressed access and that they promoted the drug into a shortage.

Where policy does have fingerprints is slower and less dramatic. HHS cut roughly 3,500 FDA positions in 2025 as part of a larger restructuring (5). The agency said reviewers and inspectors were protected. Reporting since has described reviewers absorbing administrative work and longer, less predictable timelines. None of that caused the shortage. All of it makes a nimble response harder, and a nimble response is exactly what a demand shock requires.

And there is a specific, consequential decision sitting right in the middle of this.

The part that actually matters: the FDA has not declared a shortage

Pharmacists have been listing estradiol patches as short since January 30, 2026. ASHP’s bulletin has grown to roughly 17 affected packages across five manufacturers, including Dotti, Lyllana, and the Noven, Sandoz, Viatris and Zydus generics (6).

The FDA has acknowledged that women are having trouble and says it is working with manufacturers to expand supply (7). It has not added the patches to its official shortage list.

That gap is not bureaucratic trivia. A formal shortage designation changes what other parties are allowed and incentivized to do. It can ease the path for additional manufacturers, it supports compounding in some circumstances, and it affects how insurers handle a prescription that has to be split across strengths or pharmacies. An Arizona gynecologist, Dr. Suzanne Hackenmiller, started a petition asking for the declaration, and Democratic members of the House have written to the agency asking the same (8).

So the fair summary of the policy question is this. The administration did not cause a shortage by restricting anything. It drove demand deliberately and then declined to use the one administrative tool that would most directly help the women it told to ask for treatment.

Two more explanations worth naming

Fragmentation makes it worse. When a patient cannot get one 0.1 mg patch, the workaround is often two 0.05 mg patches, or a different brand at a different pharmacy. Every one of those workarounds consumes more units of a constrained product and can generate a second copay. Scarcity plus substitution produces more scarcity.

Stockpiling is rational and corrosive. A woman who has been told to ration her patches will fill three months when she can get three months. I do not blame anybody for that. It also means supply reaches fewer people in any given week.

What to do if you are caught in this

First, do not simply stop. Vasomotor symptoms come back fast once estrogen is withdrawn, and the people who suffer most in a shortage are the ones who quietly go without and assume there was no alternative.

There are several, and most of them keep you on transdermal delivery, which matters if the reason you were on a patch in the first place was migraine with aura, a clot history, or anything else that argued against oral estrogen.

Gels and sprays are the closest substitute: EstroGel, Divigel, Evamist. Same route, same avoidance of first-pass liver metabolism, different daily routine. Oral estradiol is reasonable for many women and is widely available and cheap, though it is not the right swap for everyone. A different patch brand or a weekly patch instead of a twice-weekly one is often available when your usual one is not. If your symptoms are limited to vaginal dryness and painful sex, local vaginal estrogen is a different supply chain entirely and is not caught up in this.

Practical things that work: ask the pharmacy to check the strengths they do have rather than just saying no, call independent pharmacies because they order through different channels than the big chains, ask for a 90-day fill when you find stock, and have your clinician send the prescription with the substitution permitted so the pharmacist can act without a second call.

For clinicians

Write the alternative into the prescription up front. A patch prescription with no permitted substitution turns every stockout into a phone call, and in a shortage that is where the days get lost.

Know which of your patients actually need transdermal delivery rather than merely prefer it, because that determines whether oral is an acceptable bridge or a bad idea.

And counsel on the rebound explicitly. A patient who runs out and waits two weeks does not experience that as a supply problem. She experiences it as hormone therapy failing, and some of them do not come back.

The Bottom Line

The estrogen patch shortage is a demand shock, not a manufacturing failure and not a conspiracy.

The FDA removed a twenty-year-old warning, prescriptions nearly tripled in two years, and a small, low-margin, slow-to-expand corner of the generic industry could not absorb it. The policy criticism that holds up is not that anyone blocked access. It is that the agency drove the demand and has still not declared the shortage that would help unlock supply.

If you are affected, you have options, and most of them keep you on the same route of delivery. Call before you run out rather than after.

Sources

Estrogen patch shortage prompts doctors, lawmakers to seek FDA action. CBS News, September 3, 2026. https://www.cbsnews.com/news/estrogen-patch-shortage-fda-declaration-hormone-replacement-therapy/
Menopause Estrogen Patches Are in Short Supply. What Are the Alternatives? Healthline, April 15, 2026. https://www.healthline.com/health-news/menopause-estrogen-patch-shortage-after-fda-endorsement
HHS. HHS Advances Women’s Health, Removes Misleading FDA Warnings on Hormone Replacement Therapy. November 10, 2025. https://www.hhs.gov/press-room/hhs-advances-womens-health-removes-misleading-fda-warnings-hormone-replacement-therapy.html
Estrogen patches are in short supply as women seek menopause support. CNBC, June 26, 2026. https://www.cnbc.com/2026/06/26/estrogen-patches-are-in-short-supply-as-women-seek-menopause-support.html
HHS to cut 10,000 staff in major restructuring. BioPharma Dive. https://www.biopharmadive.com/news/hhs-layoffs-restructuring-kennedy-fda-cms-trump/743694/
ASHP Drug Shortage Detail: Estradiol Transdermal System. https://www.ashp.org/drug-shortages/current-shortages/drug-shortage-detail.aspx?id=1206
FDA acknowledges estrogen patch shortage, says it’s working to increase supply. Stateline, September 16, 2026. https://stateline.org/2026/09/16/fda-acknowledges-estrogen-patch-shortage-says-its-working-to-increase-supply/
FDA claims there’s no estrogen patch shortage as women struggle to get prescriptions filled. NBC News. https://www.nbcnews.com/health/womens-health/fda-estrogen-patch-shortage-hormone-therapy-alternatives-rcna344233
FDA removes menopause hormone therapy black box warnings. Harvard Health. https://www.health.harvard.edu/womens-health/fda-removes-menopause-hormone-therapy-black-box-warnings

A made bed below a bedroom window with autumn leaves visible outside in daylight

Menopause and Sleep: What Is Actually Waking You Up

On September 30, 2026September 30, 2026 By Dr. Scott RennieIn Menopause & Women's HealthLeave a comment

It is almost always three in the morning. You were asleep, and then you were not, and now you are doing arithmetic about how many hours are left before the alarm.

Somewhere along the way you decided this is menopause. It might be. But I have learned not to assume that, and the reason is that four or five different problems all arrive wearing the same costume.

The first thing I ask

What does not sleeping mean for you?

Trouble falling asleep is one thing. Waking up drenched at two in the morning is another. Waking at three for no reason you can name and lying there until five is a third. Those three answers send me down three different roads, and a woman who tells me only that she is not sleeping well has not yet told me the thing I need.

So we start there, before anyone says the word estrogen.

The one I do not want to miss

Here is what surprises patients: I do not automatically blame estrogen.

Menopause genuinely does fragment sleep. It also happens to be the stretch of life when obstructive sleep apnea stops being a mostly male problem. Prevalence climbs sharply after the transition, and in some series the majority of postmenopausal women have at least some degree of it (1). Surgical menopause, after both ovaries come out, roughly doubles the risk, and early menopause raises it as well (2).

The reason it gets missed is that everyone is looking for the wrong thing. The picture in most clinicians’ heads is a heavyset man who snores like a chainsaw and whose wife has watched him stop breathing. Women often do not present that way at all. They present with fatigue, morning headaches, trouble concentrating, waking to urinate, and insomnia (3). Every one of those gets filed under hormones, or stress, or getting older.

I use STOP-BANG. Not everyone does, and I know its performance in women is not what it is in men, but I would rather run a screen that is imperfect than rely on whether somebody’s husband has mentioned snoring.

Now the honest part, and it is a limitation of how I work rather than a clinical opinion. On the platform I currently practice through, I cannot order a home sleep apnea test. I would like that to change. What I can do is recognize the pattern, tell the patient plainly what I think is going on, and get her to someone who can test her. That is not as good as ordering it myself. It is considerably better than calling it menopause and moving on.

When the hot flashes really are the problem

Sometimes the story is simple. She wakes because she is soaked, three or four times a night, and has been for a year.

When vasomotor symptoms are driving the awakenings, treating the vasomotor symptoms usually fixes the sleep. That is the cleanest version of this problem and the most satisfying to treat. Hormone therapy works. Among the non-hormonal options, elinzanetant (Lynkuet) is worth knowing about: approved October 24, 2025, it blocks NK1 in addition to NK3, and the OASIS trials picked up an effect on sleep disruption rather than only on flash frequency (4). Fezolinetant, approved in 2023, works on NK3 alone.

If somebody’s nights are being destroyed specifically by the waking rather than the flashes themselves, the drug with a sleep signal behind it is the more logical place to start.

When insomnia has taken on a life of its own

The other pattern is the one where the hot flashes have settled down, or were never that bad, and she still cannot sleep. The insomnia has detached from whatever started it and become its own condition.

For that, cognitive behavioral therapy for insomnia is first-line. Not a suggestion, not an alternative to try if the pill does not work. It is what the American Academy of Sleep Medicine and the European Sleep Research Society both put ahead of medication for chronic insomnia, and that holds for menopausal women specifically (5).

I refer out for it. We have psychologists who work with us, which makes that referral considerably easier than it is in a lot of settings. I mention that because “CBT-I is first-line” is easy to write and genuinely hard to act on when there is nobody to send people to.

What I will prescribe, and what I will not

I am conservative with sleep medication, and more so after 60.

Before I write anything I want to know what is actually breaking the sleep. Night sweats. Sleep apnea. Restless legs. Alcohol. Pain. Another medication already on the list. Fix the cause and you often do not need the hypnotic at all.

When falling asleep is the main problem and medication is warranted, I may prescribe ramelteon 8 mg. It is not controlled, weight gain is not a typical adverse effect, and it has little dependence potential. The benefit is modest, and I say so upfront rather than letting somebody discover it themselves.

Melatonin can be reasonable when the timing of sleep is part of the problem. I do not present it as a strong treatment for chronic insomnia, and more is not better. Higher doses tend to buy morning grogginess and vivid dreams without buying much additional sleep. Magnesium comes up constantly. I am comfortable discussing a cautious trial when kidney function is normal, particularly if dietary intake is low, but the evidence in insomnia is weak and I will not pretend otherwise.

The list of things I do not do is longer. I do not start chronic benzodiazepines for sleep. I generally avoid zolpidem and the other Z-drugs in older women. Diphenhydramine and doxylamine, no. Quetiapine as a sleeping pill, no. Sedating antidepressants used purely as hypnotics, no. I avoid tricyclics here too, because anticholinergic effects and weight gain are a poor trade for a modest sleep benefit.

After 60 I want the treatment to improve the next day as well. Eight hours of sedation followed by fogginess, a bigger appetite, or a fall is not successful insomnia care.

For clinicians

Three things worth holding onto.

Ask what the sleep complaint actually is before reaching for a hormonal explanation. Sleep onset, vasomotor awakenings and maintenance insomnia are different problems and the history separates them in about ninety seconds.

Screen for apnea in this population, and do not wait for a snoring history to prompt you. STOP-BANG underperforms in women relative to men, which is an argument for screening more of them rather than fewer.

And when you tell a patient CBT-I is first-line, have somewhere to send her. The recommendation is only worth what the referral pathway behind it is worth.

The Bottom Line

Menopause can genuinely wreck your sleep, and it is not the only thing that does.

Start by working out what kind of not-sleeping you have. If night sweats are waking you, treating them usually fixes the nights. If the insomnia has become its own problem, cognitive behavioral therapy comes before a prescription. And if you are exhausted, foggy, waking to urinate and getting headaches in the morning, somebody should be thinking about sleep apnea rather than writing it all off to hormones.

That last one is the one I most want not to miss.

Sources

Advances in the Diagnosis and Treatment of Obstructive Sleep Apnea in Women. https://pmc.ncbi.nlm.nih.gov/articles/PMC12992857/
Type of Menopause, Age at Menopause, and Risk of Developing Obstructive Sleep Apnea in Postmenopausal Women. American Journal of Epidemiology. https://academic.oup.com/aje/article/187/7/1370/4819319
Sleep Apnea in Women Often Missed After Menopause. Medscape. https://www.medscape.com/viewarticle/sleep-apnea-women-often-missed-after-menopause-2026a1000cri
FDA Approves Elinzanetant as First Nonhormonal Therapy for Menopause Vasomotor Symptoms. Pharmacy Times, October 2025. https://www.pharmacytimes.com/view/fda-approves-elinzanetant-as-first-nonhormonal-therapy-for-menopause-vasomotor-symptoms
Insomnia in Postmenopausal Women: How to Approach and Treat It? https://pmc.ncbi.nlm.nih.gov/articles/PMC10816958/

Empty bench beside a wet winding path in a foggy leafless forest

2026-2027 Flu Shot: When to Get It, Which One, and Is Tamiflu Worth It?

On September 25, 2026September 25, 2026 By Dr. Scott RennieIn Infectious DiseaseLeave a comment

Last winter was rough. If you spent a week in January flat on the couch with a fever of 102 and a cough that hung on into February, you had plenty of company, and there is a decent chance the virus that did it was an H3N2 variant called subclade K.

This year’s flu shots are out, and every strain in them has changed. The federal advisory committee that normally tells everyone what to do is also, for practical purposes, not functioning. So I get a lot of the same three questions this time of year, and I am going to answer them here: when should I get it, which one should I get, and if I catch the flu anyway, is Tamiflu worth taking?

What last season looked like

CDC’s preliminary numbers for 2025-2026 are at least 32 million illnesses, 390,000 hospitalizations, and 24,000 deaths from flu (1). The agency classified the season as moderate for adults and high severity for children. By late May, 174 children had died of flu, and about 85 percent of those who were eligible and had a known vaccination status had not been fully vaccinated (1).

Subclade K was the reason. It surfaced in Australia in July 2025, rode the Southern Hemisphere winter north, and by January it accounted for roughly nine of every ten influenza A viruses sequenced in the United States. It had drifted away from the H3N2 strain in last year’s shot.

The shot still did something. CDC’s interim estimate put vaccine effectiveness against flu-related hospitalization at 30 percent in adults and 41 percent in children and teenagers, with outpatient protection at 22 to 34 percent in adults (2). Those are low numbers by flu vaccine standards. Thirty percent sounds unimpressive until you remember the denominator was 390,000 hospitalizations. I will take a 30 percent cut in my odds of being one of them.

What changed in this year’s shot

All three components were updated for 2026-2027, and the H3N2 piece was swapped specifically to cover subclade K (3). FDA’s advisory committee voted on the strains on March 12, 2026. The egg-based vaccines carry an A/Missouri/11/2025 H1N1 strain, an A/Darwin/1454/2025 H3N2 strain, and a B/Victoria strain; the cell-based and recombinant vaccines use closely matched versions (4).

One thing that trips people up at the pharmacy counter. If the pharmacist says they have four different flu shots, that means four types of vaccine, not four strains. Every flu shot in the U.S. this year covers the same three strains. What differs is how the vaccine is made and how strong a response it produces: standard-dose, high-dose (Fluzone High-Dose), adjuvanted (Fluad), recombinant (Flublok), and cell-based (Flucelvax), plus FluMist and the new mRNA shot. Until a couple of years ago most shots were quadrivalent, meaning they covered four strains. One B lineage, B/Yamagata, has not been confirmed anywhere in the world since March 2020, so FDA dropped it, and every U.S. flu vaccine has been trivalent since the 2024-2025 season.

A small aside, because patients ask. Those place names are where the reference virus was collected, not where the strain came from or where it is worst. “Darwin” means a lab in northern Australia isolated the sample that became the template. Nobody should cancel a trip to the Northern Territory over it.

For what it’s worth, Australia’s 2026 winter has been far quieter than 2025’s, which was the deadliest flu season there this century. Their experts are careful to say a mild Southern Hemisphere season does not reliably predict ours (5). I agree with them. I would not plan around it.

Who is actually making the recommendations this year

This part is messy, and I think you deserve a straight account of it.

In 2025, HHS removed the members of the Advisory Committee on Immunization Practices and appointed new ones. On March 16, 2026, a federal judge in Boston stayed those appointments and the votes the reconstituted committee had taken, finding they likely violated federal law (6). That ruling is on appeal. The practical result is that no functioning ACIP has voted on anything for this fall.

CDC’s response, issued September 1, was to state that the flu recommendations from the July 2025 immunization schedule remain in effect for 2026-2027 (7). In plain terms: annual flu vaccination for everyone 6 months and older, the same as it has been for years. On September 2, the American Academy of Family Physicians, the American Academy of Pediatrics, the American College of Obstetricians and Gynecologists, and the Infectious Diseases Society of America published their own respiratory-season guidance, and on flu they all say the same thing (8).

Will your insurance cover it? For flu, almost certainly. Medicare Part B covers the flu shot by statute, and AHIP, the trade group for health insurers, has said its members will keep covering the vaccines that were recommended before the committee was replaced. Flu is the one vaccine with decades of annual guidance behind it, so it is the least disrupted of the three fall shots. COVID and RSV are a different and murkier story, and I covered COVID separately.

September or October? My answer depends on your age

CDC’s guidance for most people is September or October (7). I would push that a little depending on who you are.

If you are under 65 and healthy, get it whenever it is convenient between now and the end of October. There is no advantage in waiting, and the most common way people miss their flu shot is by planning to get it “in a few weeks.”

If you are 65 or older, I would aim for October. Protection from the flu shot fades over the season, and it fades faster in older adults, which is why CDC specifically says to avoid July and August vaccination in this group unless there is no other realistic chance to get it (7). Flu in the U.S. usually peaks somewhere between December and February, so a shot in mid-October puts your best protection squarely over those months, while one in early September has already started to fade by the time the worst of it arrives. Do not let “wait for October” turn into “forgot until January,” though. A September shot beats no shot by a wide margin.

Children between 6 months and 8 years who have never had a flu shot, or have had only one dose ever, need two doses four weeks apart. Start those now so the second dose lands before the end of October.

If you are pregnant, you can get the shot in any trimester (7). A shot in the third trimester also protects the baby for the first months of life, before infants are old enough to be vaccinated themselves.

If you are 65 or older, ask for the stronger shot

I encourage everyone to get a flu shot every year, and I push hardest with older adults and anyone medically fragile. Adults 65 and over should preferentially get one of three vaccines: high-dose (Fluzone High-Dose), recombinant (Flublok), or adjuvanted (Fluad) (7). All three produce a stronger immune response than the standard shot in older adults, whose immune systems respond less to a standard dose. Pharmacies stock them. Ask by name.

There is also something new this year. On August 5, 2026, FDA approved mFlusiva, Moderna’s mRNA flu vaccine and the first of its kind, for adults 50 and older (9). In a phase 3 trial of about 40,700 adults, it was about 27 percent more effective than a standard-dose vaccine at preventing PCR-confirmed flu illness. Approval for ages 50 to 64 is standard. For 65 and older it is accelerated approval, which means Moderna still owes FDA a confirmatory study.

I like it. It is what I have chosen for myself this year, and for the members of my family old enough to get it. One fair caveat: the trial compared it against the standard-dose shot, not against high-dose or adjuvanted, so for people over 65 nobody has shown yet that it beats those. For someone 50 to 64 who wants more than the standard shot, it joins Flublok, which is approved for adults of any age, as a good choice. Coverage may lag, because the usual trigger for insurer coverage of a new vaccine is an ACIP vote, and ACIP is not voting.

Honestly, I am fine with whichever version you are comfortable with. The medical groups put it well: take an enhanced vaccine if one is available, but do not delay vaccination hunting for one (8). The best flu shot is the one you actually get.

FluMist, now at home

FluMist, the nasal spray vaccine, is approved for ages 2 through 49. Since last season, AstraZeneca has shipped it directly to homes through its FluMist Home program, and an adult can give it to themselves or to a child in the household (10). Last season it was free with most commercial insurance plus a shipping fee of about $9, and it was not available in every state. I have not been able to confirm this year’s state list, so check the program site before you plan on it.

One catch. It is a live attenuated vaccine, so it is off the table in pregnancy and for people who are significantly immunocompromised or who live with someone severely immunocompromised. For a household of healthy kids who dread needles, it is a genuinely good option.

If you get the flu anyway: the 48-hour window

This is where telemedicine is the right tool for the job.

Flu antivirals work best when started within 48 hours of the first symptom (11). That window closes fast, because a lot of people spend the first day telling themselves it is only a cold and the second day on hold trying to get an appointment somewhere, and by the time they are seen the 48 hours are gone. A video visit collapses that. When flu is circulating in your area, sudden fever, body aches, and a dry cough are enough to diagnose it clinically without a swab, and the prescription goes to your pharmacy the same hour. A home combination COVID and flu test helps if you have one, but it is not required.

When I see someone for flu symptoms by video, the first thing I ask is whether they have had a positive flu test. Then whether they have been around someone they know had the flu recently. And whether they have had a flu shot this year. If the story fits, I prescribe an antiviral.

Now the honest part, about oseltamivir (Tamiflu). For an otherwise healthy adult, the benefit is modest. Real, but modest. The Cochrane review of the drug’s full clinical study reports found it shortened symptoms by about 17 hours, from roughly 7 days to 6.3 (12). A 2024 meta-analysis in JAMA Internal Medicine found no reduction in hospitalization among outpatients treated with oseltamivir, including older and higher-risk patients, and a clear increase in nausea and vomiting (13). CDC and IDSA still recommend treatment for anyone at higher risk of complications, even after 48 hours, and I do treat those patients past the 48-hour mark. The high-risk group includes adults 65 and older, pregnancy, asthma, COPD, diabetes, heart disease, and a BMI of 40 or above.

For a healthy 30-year-old on day one, I think the fair thing is to lay out those numbers and let the patient decide. Some people will gladly take a day off a week of misery. Others would rather skip the nausea. Both reasonable. When an antiviral makes sense, I prescribe either oseltamivir or baloxavir.

Baloxavir (Xofluza) is the other option: a single dose instead of ten pills over five days, approved for ages 5 and up. In the CENTERSTONE trial, published in the New England Journal of Medicine in 2025, a single dose of baloxavir taken within 48 hours cut household transmission of flu from 13.4 percent to 9.5 percent, a 32 percent reduction in the odds of passing it on (14). That is the first antiviral shown to reduce spread. If you live with a newborn, an elderly parent, or someone on chemotherapy, that result matters. Two caveats. CDC does not recommend baloxavir during pregnancy or breastfeeding, and it is often not covered by insurance the way generic oseltamivir is. Some of my patients prefer it anyway, so before I send it to the pharmacy I let them know there may be an out-of-pocket cost.

For patients

Get the shot. If you are under 65, get it now. If you are 65 or older or have serious medical problems, do not skip it; aim for October, and ask about the high-dose, recombinant, adjuvanted, or mRNA versions. If you are between 50 and 64 and want more than the standard shot, ask your pharmacy about Flublok or mFlusiva, and check whether your plan covers the one you pick.

If you get sick, do not wait to see whether it turns into something. Start a video visit on the first day of symptoms. If you are in a higher-risk group, treatment is worth starting even after 48 hours.

Two things mean you should skip the video visit and be seen in person: trouble breathing, or a fever that improves for a day or two and then comes back with a worse cough. That second pattern can be a bacterial pneumonia following the flu. It is the one I worry about most.

For clinicians

Three practical points for this season. First, with no functioning ACIP, standing orders in many systems reference the July 2025 schedule, and CDC’s September 1 interim guidance is the document to cite if anyone asks what authority you are vaccinating under (7). Second, for your patients 65 and older, the preferential-product recommendation still stands, and mFlusiva was compared against standard dose only, so there is no evidence yet that it outperforms Fluad or high-dose in that group. Third, CENTERSTONE gives you a concrete reason to reach for baloxavir when the index patient lives with someone at high risk, provided the patient is 5 or older, not pregnant, and not severely immunocompromised (14).

On oseltamivir for low-risk outpatients, I think shared decision-making is the honest approach, and the Hanula meta-analysis is worth reading before you tell a healthy 30-year-old it will keep them out of the hospital (13). The evidence for that claim in low-risk patients is weak.

The Bottom Line

The 2026-2027 flu shot has been rebuilt around the H3N2 strain that made last winter so hard, and it is covered and available now. Get it by the end of October. If you are 65 or older, the high-dose, recombinant, adjuvanted, and new mRNA shots are all good choices, and October is the sweet spot for timing. Pick the one you are comfortable with and do not delay hunting for a particular brand. And if you do get the flu, the first 48 hours matter. A video visit on day one is the fastest way to use them.

Sources

Centers for Disease Control and Prevention. Weekly US Influenza Surveillance Report: Key Updates for Week 20, ending May 23, 2026. https://www.cdc.gov/fluview/surveillance/2026-week-20.html
Interim Estimates of 2025–26 Seasonal Influenza Vaccine Effectiveness, United States, September 2025–February 2026. MMWR Morb Mortal Wkly Rep. 2026;75(9):116-123. https://pmc.ncbi.nlm.nih.gov/articles/PMC12981535/
Centers for Disease Control and Prevention. 2026-2027 Flu Season. Updated September 2, 2026. https://www.cdc.gov/flu/season/2026-2027.html
U.S. Food and Drug Administration. Influenza Vaccine Composition for the 2026-2027 U.S. Influenza Season. https://www.fda.gov/vaccines-blood-biologics/vaccines/influenza-vaccine-composition-2026-2027-us-influenza-season
Doherty Institute. How severe is this flu season shaping up to be? https://www.doherty.edu.au/articles/how-severe-is-this-flu-season-shaping-up-to-be/
American College of Physicians. Federal Judge Blocks Immunization Schedule Changes, Stays ACIP Member Appointments. March 2026. https://www.acponline.org/acp-newsroom/federal-judge-blocks-immunization-schedule-changes-stays-acip-member-appointments
Centers for Disease Control and Prevention. Interim Clinical Considerations for the Use of Seasonal Influenza Vaccines in the United States. September 1, 2026. https://www.cdc.gov/flu/hcp/vax-summary/seasonal-influenza-vaccines.html
Healthline. U.S. Medical Groups Issue Flu, COVID Vaccine Guidance. September 2026. https://www.healthline.com/health-news/medical-groups-issue-independent-flu-covid-rsv-vaccine-guidance
BioPharma Dive. FDA approves Moderna’s mRNA flu vaccine. August 2026. https://www.biopharmadive.com/news/moderna-fda-approve-mflusiva-seasonal-influenza/826864/
AstraZeneca. FluMist, the nation’s only nasal spray flu vaccine, now available for home delivery. 2025. https://www.astrazeneca-us.com/media/press-releases/2025/FLUMIST-the-nations-only-nasal-spray-flu-vaccine-now-available-for-home-delivery.html
Centers for Disease Control and Prevention. Influenza Antiviral Drug Baloxavir Marboxil. https://www.cdc.gov/flu/treatment/baloxavir-marboxil.html
Jefferson T, Jones MA, Doshi P, et al. Neuraminidase inhibitors for preventing and treating influenza in adults and children. Cochrane Database Syst Rev. 2014;(4):CD008965. https://www.cochranelibrary.com/cdsr/doi/10.1002/14651858.CD008965.pub4/full
Hanula R, Bortolussi-Courval É, Mendel A, et al. Evaluation of Oseltamivir Used to Prevent Hospitalization in Outpatients With Influenza: A Systematic Review and Meta-Analysis. JAMA Intern Med. 2024;184(1):18-27. https://jamanetwork.com/journals/jamainternalmedicine/fullarticle/2805976
Monto AS, Kuhlbusch K, Bernasconi C, et al. Efficacy of Baloxavir Treatment in Preventing Transmission of Influenza. N Engl J Med. 2025;392(16):1582-1593. https://www.nejm.org/doi/full/10.1056/NEJMoa2413156

A wooden bedside table with a closed notebook, reading glasses, and a mug beside an unmade linen bed in soft morning window light

Should Women Take Testosterone After Menopause?

On September 11, 2026September 18, 2026 By Dr. Scott RennieIn Menopause & Women's HealthLeave a comment

Update, September 18, 2026: What Came Out of the FDA Workshop

The workshop happened yesterday. Nothing was approved, and approval was never on the table. What changed was the tone. Admiral Brian Christine, the Assistant Secretary for Health, opened by calling the lack of an FDA-approved testosterone product for women “unacceptable” when Australia has one. The FDA’s own position was more measured: the approval standard is the same for every drug, data from other countries will be judged case by case, and the agency says it stands ready to review an application (10).

That last point matters most. Nothing changes for patients until a company actually files. Products made for women are already approved in the UK, Australia, New Zealand and South Africa, so the data exists. The question is whether anyone brings it here.

The science presented matched what I wrote below. Dr. Rajita Patil of UCLA’s menopause program said the evidence supports testosterone for postmenopausal women with low sexual desire and is insufficient for mood, cognition or general well-being. She also cautioned against oral testosterone because of what it does to cholesterol (10).

My own practice hasn’t changed. Until there’s an approved product I can prescribe, I refer patients out. Public comments are open until October 19, 2026.

On September 17, 2026, the FDA holds its first public workshop on testosterone use in menopausal women. It is a hybrid session, in person at the agency’s White Oak campus and online at the same time, filed under docket FDA-2026-N-5479, with the public comment period open through October 19 (1). Right now, no testosterone product is FDA-approved for use in women in the United States. Every prescription written for a woman is off-label, borrowed from a product built and dosed for men. When I think a patient would benefit, my role right now is to refer her to someone who can prescribe it, not to write it myself.

That gap has been widening for a decade, and the FDA is only now catching up to it.

Why the Agency Is Paying Attention Now

Prescribing tells the story better than any regulatory notice could. Testosterone prescriptions to women in the US rose roughly 2.6-fold between 2016 and 2025, from 50.0 to 130.8 per 100,000 eligible women (2). In 2025 alone, 90,482 prescriptions went out, up 58.7 percent from the year before. About three in five, 62.2 percent, went to women between 45 and 64. More than half, 51.2 percent, already carried at least one documented cardiometabolic risk factor (2).

Rapid off-label growth in a population that already carries meaningful cardiovascular risk, for a drug whose long-term safety in women is still not established: that combination is what put a federal workshop on the calendar.

The Evidence That Actually Holds Up

Low sexual desire is where testosterone has real trial support. A 2019 systematic review and meta-analysis by Islam and colleagues in The Lancet Diabetes & Endocrinology pooled 36 randomized controlled trials and 8,480 women (3). In postmenopausal women, testosterone significantly improved desire, arousal, orgasm, and sexual pleasure, and reduced the distress that comes with low libido. The Global Consensus Position Statement, released the same year, reached the same conclusion from a different angle: the only evidence-based indication for testosterone therapy in women is hypoactive sexual desire disorder, confirmed only after a genuine biopsychosocial assessment rules out the other usual explanations, a strained relationship, untreated depression, a medication doing this to her (4).

I put it this way when the subject came up this week:

“Low desire is where the evidence is strongest, and that alone is reason not to dismiss it. A lot of women also just feel better on it. We cannot fully explain why yet, but at a physiologic dose with honest counseling about what is proven, that is a reasonable place to stand.”

What the Data Still Cannot Explain

The same Islam review found no measurable effect of testosterone on general wellbeing, cognition, body composition, or musculoskeletal outcomes. No real change in bone density. No change in lean mass or strength. The consensus statement says it plainly: no effect on general wellbeing has been demonstrated in trials to date.

That sits awkwardly next to what a lot of women report once they are actually on it: more energy, sharper focus, a sense of being back in their own body. That reported experience deserves to be taken seriously rather than waved off, and I am not going to claim an energy or cognition benefit I cannot point to a trial for either. The evidence has not caught up to the experience yet.

What a Reasonable Dose Looks Like

The goal is a physiologic premenopausal testosterone level. The dose I would want a patient on is about one-tenth of the standard male dose of a regulated 1 percent testosterone gel, roughly 5 mg applied to the skin daily, off-label since nothing is approved for women. Pellets and high-dose injections push levels into supraphysiologic territory that a properly dosed gel does not reach. Compounded products are a different problem: no consistent evidence backs their dosing, efficacy, or safety, whatever the label promises (4).

Monitoring is not complicated. Check a baseline total testosterone level, recheck at three to six weeks to confirm the dose landed where it should, and stop by six months if there is no symptomatic benefit. Formulation matters for one specific reason: oral testosterone raises LDL cholesterol, while a transdermal patch or gel leaves the lipid panel essentially untouched. Side effects run mild and cosmetic more often than dangerous, acne and unwanted hair growth chief among them. The same 2019 review also recorded an overall increase in weight with testosterone treatment. Worth knowing if weight is already part of the conversation.

Where Estrogen Usually Fits First

Estrogen treats hot flashes, night sweats, disrupted sleep, and vaginal dryness, none of which testosterone has been shown to fix, and it has a much longer safety record behind it. When I treat patients, estrogen usually comes before testosterone.

Without a need for contraception, the regimen I reach for is transdermal 17-beta-estradiol 0.025 to 0.05 mg per day, paired with micronized progesterone 100 mg nightly, continuously, if she still has a uterus. With a need for contraception, I use continuous ethinyl estradiol 20 mcg with levonorgestrel 100 mcg daily instead. About a quarter of the women I see on video visits arrive already started on hormone therapy from somewhere else, so a fair amount of my job is auditing a regimen I did not write.

Flibanserin, sold as Addyi, is worth knowing about too. On December 15, 2025, the FDA expanded its approval to cover postmenopausal women under 65 with HSDD; before that it had only been approved for premenopausal women, since 2015 (5). It is a separate, FDA-approved option aimed specifically at low desire rather than at hormone levels.

Getting Any of This Through a Screen

Testosterone is a Schedule III controlled substance. Under the DEA and HHS telemedicine flexibilities extended through December 31, 2026, a DEA-registered practitioner can currently prescribe it after a video visit alone, with no prior in-person exam, and generally without a separate DEA registration in the patient’s state, as long as both states’ laws allow it (6, 7). That is temporary: it rests on an extension already renewed four times, and the permanent replacement, a special-registration system that would keep Schedule III drugs like this one in an easier tier than Schedule II, has been in White House OMB review since August 25, 2026 with no final version yet (8, 9).

So federal law is not what stops me right now. Some telemedicine platforms do let a DEA-registered clinician prescribe controlled substances after a video visit. The platforms I work through don’t give me that ability at the moment, so when I think a patient would genuinely benefit, I refer her to an in-person clinician who can prescribe it. If that changes, or the DEA’s pending rule settles things for good, I may be able to write it myself.

For Patients

Before a video visit about this, bring your current medications and any hormone therapy already written down. Know what has changed for you, specifically, rather than a general “I don’t feel like myself.” And say what you want treatment to do: “I want my desire back” and “I want more energy” point toward different conversations, and only one has real trial data behind it.

One red flag worth knowing: pellets. The consensus statement is direct about this. Any testosterone preparation that produces supraphysiologic levels, pellets and high-dose injections included, is not recommended, in part because there is no way to adjust or remove the dose once a pellet is implanted. If a clinic is pitching pellets alongside language about restoring youthful vitality, that is worth questioning.

For Colleagues

A total testosterone level should never be the test that diagnoses HSDD. It does not correlate reliably with symptoms; this is a clinical diagnosis built on history and distress, other causes ruled out along the way. There is no data supporting use in premenopausal women, so keep the indication where the evidence lives, and screen for cardiometabolic risk before starting anything, given how much of the current prescribing population already carries it (2).

The Bottom Line

Low desire is where the evidence for testosterone is strongest in postmenopausal women, and that alone is reason not to dismiss it. Many women report feeling better on it, and nobody can fully explain why yet. If I were prescribing, the dose would be physiologic: about one-tenth of the standard male dose of a regulated 1 percent gel, roughly 5 mg on the skin daily, and estrogen usually comes first. Right now I am not the one writing that prescription. Testosterone is a Schedule III controlled substance, and the platforms I work through don’t currently let me prescribe it, so when a patient would benefit, I refer her to an in-person clinician who can. That could change once the DEA finalizes its pending rule or my platforms add the capability. The comment window on the FDA’s workshop closes October 19, 2026.

Sources

1. FDA. Public Meeting: Testosterone Use in Menopausal Women. September 17, 2026. Docket FDA-2026-N-5479. https://www.fda.gov/consumers/womens-health-events/fda-public-meeting-testosterone-use-menopausal-women-09172026

2. Avivi I, Stuenkel CA, Sampath-Kumar R, Ben-Yehuda O. Accelerating Testosterone Prescribing for U.S. Women: Implications for Cardiovascular Safety. JACC Adv. 2026;5(6 Pt 1):102724. PMID 42095794. https://pmc.ncbi.nlm.nih.gov/articles/PMC13309321/

3. Islam RM, Bell RJ, Green S, Page MJ, Davis SR. Safety and efficacy of testosterone for women: a systematic review and meta-analysis of randomised controlled trial data. Lancet Diabetes Endocrinol. 2019;7(10):754-766. PMID 31353194. https://pubmed.ncbi.nlm.nih.gov/31353194/

4. Davis SR, Baber R, Panay N, et al. Global Consensus Position Statement on the Use of Testosterone Therapy for Women. J Clin Endocrinol Metab. 2019;104(10):4660-4666. PMID 31498871. https://pmc.ncbi.nlm.nih.gov/articles/PMC6821450/

5. Historic First in Women’s Sexual Health: FDA Grants Approval for Addyi (flibanserin) in Postmenopausal Women. BioSpace, December 15, 2025. https://www.biospace.com/press-releases/historic-first-in-womens-sexual-health-fda-grants-approval-for-addyi-flibanserin-in-postmenopausal-women

6. Fourth Temporary Extension of COVID-19 Telemedicine Flexibilities for Prescription of Controlled Substances. Federal Register, December 31, 2025. https://www.federalregister.gov/documents/2025/12/31/2025-24123/fourth-temporary-extension-of-covid-19-telemedicine-flexibilities-for-prescription-of-controlled

7. Drug Enforcement Administration, Diversion Control Division. Telemedicine: State Registration Q&A (EO-DEA192, DEA-DC-044). https://www.deadiversion.usdoj.gov/GDP/(DEA-DC-044)(EO-DEA192)_Telemed_State_Reg_QA_(Final).pdf

8. Special Registrations for Telemedicine and Limited State Telemedicine Registrations. Federal Register, January 17, 2025. https://www.federalregister.gov/documents/2025/01/17/2025-01099/special-registrations-for-telemedicine-and-limited-state-telemedicine-registrations

9. DEA Prescribing of Controlled Substances via Telemedicine. Alliance for Connected Care. https://connectwithcare.org/dea-prescribing-of-controlled-substances/

10. Mahatole S, Ananthan P. US regulators urge more research on testosterone therapy for women. Reuters (via GV Wire), September 17, 2026. https://gvwire.com/2026/09/17/us-regulators-urge-more-research-on-testosterone-therapy-for-women/

An amber prescription bottle tipped over on a pale wooden table with pink and white antibiotic capsules spilling out, beside a folded paper card, in bright window light

You Probably Are Not Allergic to Penicillin

On August 31, 2026 By Dr. Scott RennieIn Infectious DiseaseLeave a comment

About one in ten Americans carries a penicillin allergy in their chart (1). When those people are actually tested, the large majority are not allergic.

That gap is one of the most consequential unforced errors in medicine, and almost nobody goes back and checks.

Where the label usually comes from

Most penicillin allergy labels were applied in childhood, by someone who is no longer involved, based on something that happened once.

The classic story is a rash during a course of amoxicillin at age four. Here is the problem with that story. Viral illnesses cause rashes. Children with viral illnesses get prescribed antibiotics. When a rash appears on day three, the antibiotic is the thing everyone is looking at, so the antibiotic gets the blame and the label goes in the chart. It never comes out.

The other common origins are a family member’s allergy transferred onto the patient, and side effects filed as allergy. Nausea is not an allergy. Diarrhea is not an allergy. A yeast infection is not an allergy. They are all reasons to dislike a drug and none of them are reasons to avoid it for life.

Even genuine allergy fades. IgE-mediated penicillin sensitivity wanes over time, and a substantial proportion of people who really were allergic at twenty are not allergic at fifty.

The label is not free

This is the part I want to land, because people assume that avoiding penicillin is the cautious choice and therefore the safe one.

It is not. Blumenthal and colleagues followed 64,141 adults carrying a penicillin allergy label against 237,258 matched comparators for a mean of six years (2). The people with the label had an adjusted hazard ratio of 1.69 for MRSA and 1.26 for Clostridioides difficile.

The mechanism is not mysterious, and the authors show it. Those patients received far more of the alternatives: macrolides at over four times the rate, clindamycin at nearly four times, fluoroquinolones at twice (2). Those drugs do more collateral damage to the gut and select harder for resistance. The excess MRSA and C. difficile were mediated by that substitution.

So the label does not sit there harmlessly. It quietly buys worse antibiotics for the rest of a person’s life.

There is a score for this, and it runs on history alone

The thing that changed my thinking here is that you do not need a skin test to identify most of the people who are low risk. You need four questions.

PEN-FAST was derived and validated by Trubiano and colleagues (3). It scores like this:

Five years or fewer since the reaction: 2 points. Anaphylaxis or angioedema, or a severe cutaneous adverse reaction: 2 points. Treatment required for the episode: 1 point.

A score under 3 means low risk. In the derivation cohort, only 17 of 460 low-risk patients tested positive, giving a negative predictive value of 96.3 percent (3).

Read the criteria again and notice what is not in them. No skin prick. No specific IgE. No blood work at all. Every input is something the patient tells you.

That is why this belongs in a conversation rather than a laboratory, and it is why I think most clinicians could be doing more of this than they are.

What I actually do with this on a video visit

I ask everyone about allergies before I prescribe anything. That part is unremarkable. What is worth describing is what happens next, because it is where most of these labels give themselves away.

When somebody tells me they are allergic to amoxicillin, I ask what the reaction was.

Two answers come back more than any others. The first is that they do not really remember, because they were a baby and their parents told them. The second is some version of “I think I am allergic because my mom is.”

Neither of those is an allergy history. The second one is not even about the patient.

Here is where I part company with the enthusiastic version of this argument. A person who is sick today and needs an antibiotic today is not in a good position to test a theory about their own immune system. If there is any real concern about anaphylaxis in what I am hearing, I do not push it. I pick something else and move on, and I am comfortable with that.

I also do not send people off for an oral challenge as a matter of course. That is a real procedure with a real indication, and routing every vague childhood label into an allergy clinic is not a good use of anyone’s afternoon.

What I do instead is tell them the label is worth reopening properly, and that the time to do it is when they are well rather than in the middle of an infection. It belongs with whoever manages their ongoing care, in a visit where nobody is waiting on a prescription.

The distinction matters. The label deserves to be examined. The examination should not happen while someone is sick and I am trying to treat them.

What a real reaction looks like, and what has changed about treating it

None of the above applies to someone who has had a genuine severe reaction. Anaphylaxis, angioedema, or a severe cutaneous adverse reaction such as Stevens-Johnson syndrome or DRESS means the label stays. Those people are not candidates for de-labeling and should not be challenged.

Worth knowing what has changed on the treatment side. For decades the only option for anaphylaxis outside hospital was an intramuscular auto-injector. There is now an FDA-approved epinephrine nasal spray, neffy, the first intranasal adrenaline product to reach the market (4). Pharmacokinetic studies show plasma concentrations comparable to or exceeding intramuscular injection, and, counterintuitively, nasal congestion does not appear to impair absorption (5).

I mention it because needle avoidance is a real reason people leave their epinephrine at home, and a device you will actually carry beats a better device you will not. If someone genuinely needs to carry epinephrine and the needle is the barrier, that conversation is now worth having.

It does not change the rule that epinephrine is the treatment and antihistamines are not.

For patients

If your penicillin allergy dates from childhood and you cannot remember what happened, it is worth reopening. Ask specifically to be assessed rather than mentioning it in passing, because in passing it stays in the chart.

Before that conversation, work out three things: roughly when it happened, what you or your parents remember seeing, and whether anyone treated it. That is most of the assessment.

Raise it when you are well. A visit for an infection is the wrong moment, because whoever is treating you needs to choose an antibiotic today and will reasonably play it safe. Bring it to a routine appointment instead, when nobody is waiting on a prescription.

If you have taken amoxicillin, Augmentin or any similar antibiotic since the original reaction and were fine, say so early. It is the single most useful thing you can tell your clinician.

And if your reaction involved breathing trouble, swelling of the face or throat, or a blistering rash, the label is real and it stays.

For colleagues

PEN-FAST costs four questions and identifies low-risk labels at the point of care with a negative predictive value of 96.3 percent (3). We are not short of the tool. We are short of anyone taking ownership of the conversation.

Frame it as a safety intervention rather than a convenience one, because it is. The hazard ratios in the Blumenthal cohort are for MRSA and C. difficile, not for inconvenience (2).

And ask the question that costs nothing: has this person taken a penicillin since. A meaningful share of these labels fall over on that answer alone.

The Bottom Line

Roughly ten percent of people carry a penicillin allergy label and most of them are not allergic, usually because a childhood viral rash got blamed on an antibiotic. The label is not harmless. Carrying it is associated with a 69 percent higher risk of MRSA and a 26 percent higher risk of C. difficile, driven by the broader antibiotics used instead. Four questions, scored as PEN-FAST, identify most low-risk labels without any testing at all, which makes this one of the few things in medicine that is fixed by a conversation. If your reaction was anaphylaxis, angioedema or a blistering rash, the label stays and you keep your epinephrine, which now comes as a nasal spray as well as an injector.

Sources

1. Shenoy ES, Macy E, Rowe T, Blumenthal KG. Evaluation and management of penicillin allergy: a review. JAMA. 2019;321(2):188-199. PMID 30644987. https://pubmed.ncbi.nlm.nih.gov/30644987/

2. Blumenthal KG, Lu N, Zhang Y, Li Y, Walensky RP, Choi HK. Risk of meticillin resistant Staphylococcus aureus and Clostridium difficile in patients with a documented penicillin allergy: population based matched cohort study. BMJ. 2018;361:k2400. PMID 29950489. https://pubmed.ncbi.nlm.nih.gov/29950489/

3. Trubiano JA, Vogrin S, Chua KYL, et al. Development and validation of a penicillin allergy clinical decision rule. JAMA Intern Med. 2020;180(5):745-752. PMID 32176248. https://pubmed.ncbi.nlm.nih.gov/32176248/

4. Hernandez-Trujillo V, Tachdjian R, et al. Successful administration of neffy (epinephrine nasal spray) by patients and caregivers. Ann Allergy Asthma Immunol. 2026. PMID 42476303. https://pubmed.ncbi.nlm.nih.gov/42476303/

5. Takahashi K, Yanagida N. Intranasal adrenaline: a new treatment for anaphylaxis. Pediatr Allergy Immunol. 2026;37(3):e70316. PMID 41796069. https://pubmed.ncbi.nlm.nih.gov/41796069/

Two empty medical vials and syringe beside an illustrated human brain

Does the Shingles Vaccine Lower Dementia Risk? What the Evidence Shows

On August 30, 2026August 31, 2026 By Dr. Scott RennieIn Infectious DiseaseLeave a comment

I wrote about shingles here in 2012. Back then the vaccine question was simple and slightly boring: there was one shot, it worked reasonably well, and most people over 60 were told to consider it.

The question I get now is different. People do not ask me whether the shingles vaccine prevents shingles. They ask whether it protects the brain.

That is a fair question, the evidence behind it is more interesting than most vaccine headlines, and it has a catch in it that almost nobody reporting on it mentions.

Start with the boring part, because it is the part that matters

Shingles is the chickenpox virus waking up. It sat in a nerve root for decades and then came back out along that nerve, which is why the rash arrives in a band on one side and stops at the midline.

The rash heals. The nerve pain is the problem. Postherpetic neuralgia is burning, electric pain in the same distribution that can outlast the rash by months, occasionally years, and it is genuinely difficult to treat once it settles in. That, not the rash, is the reason to care about this.

The current vaccine is Shingrix, given as two doses. In the ZOE-50 trial across more than 15,000 adults aged 50 and over, efficacy against shingles was 97.2 percent (1). In ZOE-70, in adults 70 and older, it was 89.8 percent, and in the pooled analysis of everyone over 70 across both trials, efficacy against postherpetic neuralgia specifically was 88.8 percent (2).

Those are very good numbers. Vaccines rarely perform like that in the age group where you most want them to work.

It is a reactogenic shot. Roughly one in six people feels properly unwell for a day or two after a dose, sore arm, tired, achy, sometimes feverish. I tell people that in advance, because a person who was warned takes the second dose and a person who was ambushed does not.

The dementia finding, and the catch

Here is where it gets interesting.

In 2025, a group at Stanford published a study in Nature that used an accident of Welsh policy (3). When Wales rolled out the shingles vaccine, eligibility was set by exact date of birth. Anyone born before 2 September 1933 was ineligible, permanently. Anyone born on or after that date was eligible.

Think about what that creates. A person born on 1 September 1933 and a person born on 3 September 1933 are, on average, identical in every way that matters. Same generation, same health, same everything. The only systematic difference between them is that one could get the vaccine and one could not. Uptake went from 0.01 percent in the group one week too old to 47.2 percent in the group one week younger.

That is as close to a randomized trial as you will ever get out of routine health records without running one.

Over seven years, the vaccinated group had a 3.5 percentage point lower probability of a new dementia diagnosis, which works out to a 20 percent relative reduction (3). The effect was stronger in women. The authors then reproduced it in a separate population using death certificates rather than diagnoses, which guards against the finding being an artifact of who happens to get diagnosed.

Now the catch.

Wales was using Zostavax, the old live-attenuated vaccine. That product was withdrawn from the US market in November 2020. It is not what you would be offered today, and the study does not tell you what Shingrix does.

A separate group looked at exactly that, using the rapid switchover from live to recombinant vaccine as its own natural experiment, and found the recombinant vaccine associated with a 17 percent increase in dementia-free time, about 164 extra days without a diagnosis among those who went on to be diagnosed, over six years (4). Again larger in women. I will note, because they note it, that the senior author consults for the manufacturer, and that GSK had no involvement in the study and did not know about it until after acceptance.

What I actually think about it

I think the signal is real and I do not think it should be the reason you get the shot.

Those are compatible positions. The dementia work is observational in design even when it is cleverly built, the mechanism is unproven, and no randomized trial has tested it. Something could still be wrong with it. Meanwhile the case for Shingrix rests on two large randomized controlled trials showing it prevents a disease that causes months of nerve pain, and that case was already sufficient on its own.

So my framing is this. Get it because postherpetic neuralgia is miserable and largely preventable. If the dementia finding holds up, you will have gotten that for free.

What the finding does change is urgency. I used to be relaxed when someone in their early fifties wanted to put it off a few years. I am less relaxed now, because if there is a neurological benefit it presumably accrues with time, and there is no advantage to waiting.

The part that belongs to my day job

Most of my practice is metabolic, and shingles turns out to sit closer to that than it looks.

A meta-analysis of sixteen studies put the risk of shingles in people with diabetes at a pooled relative risk of 1.38 compared with the general population (5). A separate population cohort found that people with diabetes had roughly 1.45 times the risk of postherpetic neuralgia, and that when they got it, it tended to be more severe and more persistent (6).

So the group most likely to get shingles, and most likely to be left with lasting nerve pain afterward, overlaps heavily with the group I see for weight and glucose.

There is a second reason this matters in diabetes that has nothing to do with the vaccine. Postherpetic neuralgia and diabetic peripheral neuropathy feel similar, they get treated with the same drugs, and I have seen the second diagnosis absorb the first. If burning pain shows up in a band, on one side, in a person who also has stocking-glove neuropathy, that is shingles until proven otherwise, whatever the chart already says.

Age 50 is also, not coincidentally, right in the middle of the menopause transition, which is the other half of what I do. If a woman is already in front of me talking about hormone therapy and bone density, the shingles vaccine is a reasonable thing to raise in the same visit, and often nobody has.

If you already have it, the clock is the thing

Antivirals for shingles work best started within 72 hours of the rash appearing. After that the benefit falls off steeply.

That single fact is the strongest argument I know for handling suspected shingles by video. A one-sided painful band of blistering rash with a burning prodrome is one of the more recognizable things in medicine, and a decent photograph plus a history usually settles it. The bottleneck has never been diagnostic difficulty. It has been how long it takes to be seen, and a visit you can get this afternoon beats a better visit on Thursday.

Two exceptions I do not manage remotely. A rash near or in the eye, particularly on the tip of the nose, needs same-day ophthalmology because of the risk to the cornea. Shingles in someone significantly immunocompromised, or a disseminated rash crossing the midline, goes in person.

For patients

If you are 50 or over, you are eligible for Shingrix regardless of whether you remember having chickenpox, and regardless of whether you had the old vaccine years ago.

If you are immunocompromised, the recommendation starts at 19 (7).

Expect to feel rough for a day after each dose, and get the second one anyway. One dose is not the regimen.

If a painful one-sided rash appears, get seen inside three days. Do not wait to see whether it spreads.

For colleagues

The dementia data is worth knowing and worth describing accurately. The strongest causal design was done on a product no longer sold in the United States, and patients who have read a headline usually have not been told that.

Diabetes belongs on the list of reasons to push the vaccine actively rather than mention it in passing.

And check the vaccine history of the metabolic patients. Nobody owns this conversation, so it tends not to happen. I have started treating it as part of the same visit rather than something to refer out, and it costs about ninety seconds.

The Bottom Line

Shingrix prevents shingles about 97 percent of the time in adults over 50, and it prevents postherpetic neuralgia, which is the thing actually worth avoiding. A natural experiment in Wales found 20 percent less dementia over seven years in vaccinated adults, and a second study suggests the current recombinant vaccine carries a similar signal, but the strongest of that evidence came from a live vaccine the United States no longer uses. Get the vaccine for the nerve pain it reliably prevents rather than the dementia it might. If you have diabetes, your risk of both shingles and lasting nerve pain afterward is meaningfully higher, which moves this up the list. And if a one-sided painful rash turns up, the antiviral window is 72 hours.

Sources

1. Lal H, Cunningham AL, Godeaux O, et al. Efficacy of an adjuvanted herpes zoster subunit vaccine in older adults. N Engl J Med. 2015;372(22):2087-2096. PMID 25916341. https://pubmed.ncbi.nlm.nih.gov/25916341/

2. Cunningham AL, Lal H, Kovac M, et al. Efficacy of the herpes zoster subunit vaccine in adults 70 years of age or older. N Engl J Med. 2016;375(11):1019-1032. PMID 27626517. https://pubmed.ncbi.nlm.nih.gov/27626517/

3. Eyting M, Xie M, Michalik F, et al. A natural experiment on the effect of herpes zoster vaccination on dementia. Nature. 2025;641(8062):438-446. PMID 40175543. https://pubmed.ncbi.nlm.nih.gov/40175543/

4. Taquet M, Dercon Q, Todd JA, Harrison PJ. The recombinant shingles vaccine is associated with lower risk of dementia. Nat Med. 2024;30(10):2777-2781. PMID 39053634. https://pubmed.ncbi.nlm.nih.gov/39053634/

5. Huang CT, Lee CY, Sung HY, et al. Association between diabetes mellitus and the risk of herpes zoster: a systematic review and meta-analysis. J Clin Endocrinol Metab. 2022;107(2):586-597. PMID 34536279. https://pubmed.ncbi.nlm.nih.gov/34536279/

6. Wen SY, Ou-Yang C, Hsu CY, et al. Impact of type 1 versus type 2 diabetes on developing herpes zoster and post-herpetic neuralgia: a population-based cohort study. Acta Derm Venereol. 2023;103:adv9400. PMID 37787418. https://pubmed.ncbi.nlm.nih.gov/37787418/

7. Anderson TC, Masters NB, Guo A, et al. Use of recombinant zoster vaccine in immunocompromised adults aged 19 years and older: recommendations of the Advisory Committee on Immunization Practices. MMWR Morb Mortal Wkly Rep. 2022;71(3):80-84. PMID 35051134. https://pubmed.ncbi.nlm.nih.gov/35051134/

Why your compounded supply is drying up

Why can’t I just convert my compounded dose?

Restart low, or match the dose?

What I would gather before the video visit

What restarting will feel like

What will the brand version cost?

Pills as a way out

Gray-market sellers and “research use” vials

If you are the patient

If you are the clinician

The Bottom Line

Related Reading

Sources

Why results show up before anyone has called you

What does “critical” actually mean?

Flagged does not mean dangerous

Which numbers can be truly urgent?

What to do, in order

When a virtual clinician ordered the test

Ask before the blood draw

For patients, and for clinicians

If you are the patient

If you are the clinician

The Bottom Line

Related Reading

Sources

Why you want to stop matters more than the fact that you want to stop

What the evidence actually shows

It is physiology, not a character flaw

Most people stop anyway

What I actually do: lower the dose rather than stop

When the reason is cost

Protecting muscle

When I bring up the long-term question, and why not sooner

When I support stopping

Restarting

For clinicians

The Bottom Line

Related Reading

Sources

The 1963 to 1967 problem

If you were born before 1957

Should you get a titer drawn?

About the death count, since you will see it argued about

What I would actually do

The Bottom Line

Related Reading

Sources

Hypothesis one: everyone started using them at once

Hypothesis two: something broke in manufacturing

Hypothesis three: RFK Jr. and the administration did this

The part that actually matters: the FDA has not declared a shortage

Two more explanations worth naming

What to do if you are caught in this

For clinicians

The Bottom Line

Related Reading

Sources

The first thing I ask

The one I do not want to miss

When the hot flashes really are the problem

When insomnia has taken on a life of its own

What I will prescribe, and what I will not

For clinicians

The Bottom Line

Related Reading

Sources

What last season looked like

What changed in this year’s shot

Who is actually making the recommendations this year

September or October? My answer depends on your age

If you are 65 or older, ask for the stronger shot

FluMist, now at home

If you get the flu anyway: the 48-hour window

For patients

For clinicians

The Bottom Line

Related Reading

Sources

Update, September 18, 2026: What Came Out of the FDA Workshop