Thread by @TimAndersonMD

🚨New in @JAMANetworkOpen🚨

We examined clinical outcomes of intensifying hospitalized older adults' #diabetes medications at discharge...

🧓+💉+💊+ 🏥--->🏡

Link: ja.ma/3DYuVnI

This will be a thread...

For those who just want the punchline:

🧐Deeper Dive

Patients👵👴 often have blood sugar fluctuations in hospital, due to:
➖Acute illness
➖Home meds held
➖New meds
➖Stress
➖Diet change

We have guidelines for managing DM during hospitalization for other conditions, but no data on intensifying meds at discharge

In prior work, we found med changes are common, even among patients unlikely to benefit

Link: jamanetwork.com/journals/jaman…

So…what happens to hospitalized older adults with diabetes (DM) who are prescribed more intensified DM meds at discharge⁉️

Seems an important question for:
👉hospitalists (#JHMChat)
👉PCPs (@primarycarechat)
👉 endocrinologists
👉👉👉PATIENTS!

Research Context

Examined the national VA health system because we can reliably track medication use & clinical data like blood glucose & HbA1c.

(Downside of VA is generalizability esp for gender)

Focused on older adults 👴👵🧓 = highest risk of hospitalization & med harms.

Focused on hospitalizations for common medical/cardiac conditions (not DKA/HHS)

Study period = 2011-2016 (pre- SGLT2/GLP1 era)

Exposure = discharge with a new or higher dose diabetes med 💉💊

* Caveat, it near impossible to study insulin dose changes, as they don’t result in new prescriptions, so we studied only patients not taking insulin on admission, but included those receiving new insulin

Co-primary outcomes
1. Severe hypoglycemia
2. Svere hyperglycemia;

Severe = leading to ED or hospitalization
Time-frame = 30 days & 365 days

Secondary
1.Death
2.Persistence to intensified meds
3.HbA1c at 1 year

Nerdy stats…
🧮 Propensity score matching approach to adjust for measured confounding

🧮 Fine and Gray proportional subdistribution hazards models for outcomes

Top-line results

Older adults receiving diabetes intensifications at discharge had:
-Higher risk of severe hypoglycemia‼️
-No difference in hyperglycemia
-Both outcomes were quite rare

Second-line results
-Decreased mortality at 30 but not 365 days (More not this shortly)
-No difference in readmissions

No difference in HbA1c change between groups at 1 year

Persistence to intensifications was worse for intensified medications than continued medications

What about the mortality benefit?
- Prob confounding, as RCTs do not show a 30 day mortality benefit from DM meds
- We suspect driven by a group of patients with high risk of short-term mortality who (appropriately) were not intensified (e.g. patients headed to hospice)

- We excluded those receiving hospice, but some may decline hospice services (spent a year trying to track this down & remain😕)

- Fortunately, this odd finding should not strongly bias other outcomes given Fine-Gray models

So what does this all mean clinically?

My take: Most of the time, DM intensifications can wait until after discharge & patient recovery from acute illness. Inpatient docs should communicate concerns with patients & PCPs for f/u

When health stabilizes a decision to intensify meds may be safer, more patient-centered, & more likely to stick.

Old but true adage: Treat the patient not the number

If all this sounds familiar, our team has been studying outcomes of med changes at hospital discharge for a bit…

Why? This is a largely evidence-free scenario which inpatient docs & residents deal with every day!

Want to read more? See our prior work on blood pressure!
1. bmj.com/content/362/bm…

2. link.springer.com/article/10.100…

3. acpjournals.org/doi/10.7326/M2…

4. jamanetwork.com/journals/jamai…

Closing out with 🙏 to our team - @MikeSteinman Alexandra Lee @Bochizzle @seijlee @ShaniHerzig

🙏 @NIH & NIA

🙏 to @RozalinaMD & Patrick O’Conner for a thoughtful commentary on the paper, published today as well in @JAMANetworkOpen

ja.ma/3E5aaqE

/END

Unroll @ThreadReaderApp