SGLT-2 inhibitors: twice the risk of amputation; new dapagliflozin study

A large study compared serious adverse outcomes of SGLT-2 inhibitors vs GLP-1 agonists in diabetes, finding signficant increases in lower limb amputations and diabetic ketoacidosis (see dm sglt2 inhib lower extrem amputation bmj2018 in dropbox, or doi.org/10.1136/bmj.k4365 ).

Details:
--17,213 new users of SGLT2 inhibitors (dapagliflozin, 61%; empagliflozin, 38%; canagliflozin, 1%) were compared to  17,213 new users of glucagon-like peptide 1 (GLP1) receptor agonists in a Swedish and Danish registry analysis from 2013-2017, including all patients aged 35 and older given a first prescription for these agents
--61% men, mean age 61, comorbidities: 16% ischemic heart disease/8% ACS/6% cardiomyopathy/4% afib/10% psych disorders/16% obesity/<1% DKA/13% retinopathy/4% renal disease, 80% on metformin/35% DPP-4/29% long-acting insulin/69% ACE or ARB/35% platelet inhibitors/71% lipid-lowering drugs/24% NSAIDs
--propensity-score matched analysis, based on 66 covariates (sociodemographic, comorbidities, comediations, and healthcare utilization)

Results:
--SGLT2 inhibitors vs GLP1 receptor agonists:
    --lower limb amputation: more than 2-fold increased risk, incidence rate 2.7 v 1.1 per 1000 person years, HR 2.32 (1.37 to 3.91)
        --for toe or metatarsal amputation, HR 1.55 (0.87-2.77)
        --for major osteoporotic bone fracture, HR 1.05 (0.80-1.43)
    --diabetic ketoacidosis: more than 2-fold increased risk, incidence rate 1.3 v 0.6 per 1000 person years, HR 2.14 (1.01 to 4.52)
    --bone fracture: non-significant increase, incidence rate 15.4 v 13.9 per 1000 person years, HR 1.11 (0.93 to 1.33)
    --acute kidney injury: non-significant decrease, incidence rate 2.3 v 3.2 per 1000 person years, HR 0.69 (0.45 to 1.05)
    --serious urinary tract infection: non-significant decrease, incidence rate 5.4 v 6.0 per 1000 person years, HR 0.89 (0.67 to 1.19)
    --venous thromboembolism: non-significant increase, incidence rate 4.2 v 4.1 per 1000 person years, HR 0.99 (0.71 to 1.38)
    --acute pancreatitis: non-significant increase, incidence rate 1.3 v 1.2 per 1000 person years, HR 1.16 (0.64 to 2.12)
--HRs for the above by subgroup analyses (for country, sex, age group, history of cardiovascular disease) were similar when compared to the above primary analysis; HRs were also similar in subgroup analysis for diabetic ketoacidosis according to insulin status, and for lower limb amputation by history of PAD or lower limb amputation (though the event rates were higher in those with such histories)
--also, because of some baseline differences between the groups, the HRs were further adjusted for A1c, blood pressure, albuminuria, eGFR, BMI, and smoking status. no difference in outcomes

--review of the supplementary material showed that both lower limb amputations and DKA increased after one year of SGLT2 treatment and then remained parallel for the next 2 1/2 years

Commentary:
--the incentive for this study was that SGLT-2 inhibitors have been reported to have increased lower limb amputations, diabetic ketoacidosis, acute kidney injury, and serious urinary tract infections/urosepsis. also possibly cases of acute pancreatitis (though this has been tagged to GLP-1 inhibitors as well, subsequent studies not endorsing this finding)
--they chose to have the active comparator of GLP-1 agonists, since they both share the apparent benefit of decreasing cardiovascular events (an important consideration, since diabetics die mostly from cardiovascular disease, in the 80% range).
--of note, very few of these patients were on canagliflozin, the one with a box warning by the FDA as having increased leg and foot amputations; this has led many to avoid it and use empagliflozin or dapagliflozin (which do not have that warning). there was no breakdown (including in the supplementary materials) by the different SGLT2s in terms of adverse outcomes, but only 254 people were on canagliflozin, vs 10,454 on dapagliflozin and 6506 on empagliflozin
--the CANVAS studies did find an increased risk of bone fractures with canaglifglozin, which has been associated with decreased bone mineral density  (again, only 1% of the above patients were on that drug)
--a US retrospective cohort study based on a large commercial claims database found a significant increase in lower limb amputations wtih SGLT-2s when compared to sulfonylureas, metformin, and glitazones, but nonsignificant increases when compared to GLP-1 agonists or DPP-4 inhibitors: see http://gmodestmedblogs.blogspot.com/2018/09/sglt-2-inhibitors-and-increased-lower.html , which also references prior blogs suggesting that the original studies on empagliflozin and canagliflozin were seriously flawed.
 --limitations of this study: observational study so unable to draw clear causality just associations, overall this was a pretty healthy and fit group (quite low levels of obesity vs what we see in the US), and the % of patients on statins and platelet inhibitors was pretty low

---------------------------------------------------
 --there was a recent study presented at the Am Heart Assn meetings finding that 17,160 diabetic patients at risk for ASCVD (41% known ASCVD, 59% multiple risk factors) randomized to dapagliflozin vs placebo and followed 4.2 years had no difference in major adverse cardiovascular events, but did have a lower rate of cardiovascular death or hospitalization (4.9% vs 5.8%, 17% decrease, HR 0.83 (0.73-0.95), p=0.005), though this really just represented a lower rate of hospitalization for heart failure (27% decrease, HR 0.73 (0.61-0.88). see dm dapagliflozin dec cardiac events nejm2018 in dropbox, or DOI: 10.1056/NEJMoa1812389 ). there was also a 24% decrease in renal events (with >40% decrease in eGFR) on dapagliflozin, but DKA was more common with dapagliflozin (0.3% vs 0.1, p=0.02); and more genital infections or other adverse effects leading to dapagliflozin discontinuation (0.9% vs 0.1%, p<0.001). notably, there was no difference in amputation or fractures.
    --it should be noted that this was a placebo-controlled study, the patients on dapagliflozin had a lower A1c (0.42%) throughout the trial, 10% received GLP-1 agonists, and 5% open-label SGLT-2 antagonists. also those on dapagliflozin had a 1.8 kg decrease in weight, and 2.7/0.7 mmHg decrease in BP; these differences might explain some part of the observed cardiovascular benefit from dapagliflozin
    --though dapagliflozin decreased only the rate of hospitalization for heart failure, empagliflozin did so as well, but also decreased cardiac death rates. And canagliflozin also decreased heart failure hospitalizations, though no other specific cardiac outcome was significantly improved
--it is always difficult to reconcile the different outcomes of different studies. it is pretty striking that canagliflozin was associated with amputations in one study, another did not with empagliflozin, this study new dapagliflozin study found no association, but the above Swedish/Danish study did (at least, for the 99+% on dapagliflozin and empagliflozin). This brings up another important issue: if it were true that the different SGLT2’s may have differing adverse effect profiles, that has profound clinical implications: we should not be routinely prescribing any new cousin of an SGLT2 being marketed until it has been around several years to see the specific adverse events for this newcomer before prescribing it more generally…

--------------------------------------------------

so, this Swedish/Danish study above continues to add concerns about the SGLT-2 inhibitors, especially the amputations and DKA (though the recent study on dapagliflozin did also find an issue with genital infections, as with other SGLT-2 studies but not found in the Swedish/Danish study). the most consistent cardiovascular finding for the SGLT-2 inhibitors seems to be decreasing hospitalization for heart failure. and these were in studies comparing the SGLT-2's to placebo, with some adjustment of the other meds (in the empagliflozin study, glitazones were added in the placebo groups which could increase heart failure).  i do think that the big issue for me is diabetes drugs which decrease cardiovascular disease, given that this is what diabetics die from. in this case, i think that the most useful comparison is in fact between SGLT-2's and GLP-1 agonists, which both seem to have some cardiovascular benefit. and, to me, the GLP-1's win, hands-down. mostly because they have been around a long time, do not appear to have significant adverse events, and seem to be very potent in decreasing A1c levels (as per my several prior blogs on this).

geoff

If you would like to be on the regular email list for upcoming blogs, please contact me at gmodest@uphams.org

to get access to blogs since 8/15/17:

1. go to http://gmodestmedblogs.blogspot.com/ to see them in reverse chronological order

2. click on 3 parallel lines top left, if you want to see blogs by category, then click on "labels" and choose a category​

3. or you can just type in a name in the search box and get all the blogs with that name in them

to access older blogs from the BMJ website, from October 2013 until 8/15/17: go to http://blogs.bmj.com/bmjebmspotlight/category/archive/ 

please feel free to circulate this to others. also, if you send me their emails, i can add them to the list