diabetes: SGLT-2 inhibitors as first-line therapy??

 A recent cohort study found that first-line treatment of patients with type II diabetes with SGLT2 inhibitors was associated with a lower risk for hospitalization for heart failure and all-cause mortality (see dm SGLT2 as firstline rx AIM2022 in dropbox, or doi:10.7326/M21-4012)

 

Details:

-- claims-based data from two large US commercial and the Medicare databases from April 2013 to March 2020 (the Optum Clinformatics Data Mart and the IBM MarketScan, along with Medicare for those over 65yo), all individuals without any use of antidiabetic medications before cohort entry

-- mean age 62, 51% male, 50% from the south/15% from other regions, 70% white, 35% overweight or obese, 15% smokers

-- diabetic nephropathy 4%, neuropathy 7%, retinopathy 2%

-- comorbidities: MI 4%, stroke 7%, other ischemic heart disease 20%, heart failure 7%, peripheral arterial disease 6%, angina 5%, hyperlipidemia 72%, hypertension 73%, CKD 7%, COPD 9%

-- concomitant medications ACE/ARB 55%, antithrombotics 13%, beta blockers 33%, calcium blockers 34%, loop diuretics 10%, statins 50%, thiazides 13%

-- mean A1c 7.5%, mean eGFR 120, LDL 41, HDL 46, triglycerides 184

-- study analyzed the 8613 people who began first-line SGLT2 inhibitors for new onset diabetes versus 70,226 first-line metformin initiators

    -- the SGLT2 inhibitors included empagliflozin, canagliflozin, or dapagliflozin

    -- those on first-line SGLT2's: were younger, more often had diabetic nephropathy and neuropathy, were followed more closely (more office visits, more HbA1c tests), and had fewer recent hospitalizations

-- primary outcome: composite of hospitalizations for MI, hospitalization for ischemic or hemorrhagic stroke, or all-cause mortality, and a composite of hospitalization for heart failure (HHF) or all-cause mortality (HHF/mortality). 

-- secondary outcomes included the individual components of the primary outcomes and a composite of MI/stroke/HHF/mortality

-- safety outcomes: acute kidney disease, bone fractures, genital infections, severe hypoglycemia, severe UTIs, diabetic ketoacidosis, and lower limb amputations.

-- propensity score matching was used to mathematically equalize the different subgroups of patients

-- mean follow-up 12 months

 

Results:

-- achieved A1c levels in the two groups: 0.46 percentage points higher in those on SGLT-2s

-- MI/stroke/mortality: HR 0.96 (0.77-1.19), not statistically significant

-- HHF/mortality: 20% decrease with SGLT2, HR 0.80 (0.66-0.97)

-- HHF: 22% reduction, HR 0.78 (0.63-0.97)

-- MI: 30% decrease (0.48-1.00), borderline significant

-- no significant difference with metformin for stroke, mortality, and the composite of MI/stroke/HHF/mortality

-- genital infections were increased twofold, HR 2.19 (1.91-2.51); no other differences in safety concerns

--Sensitivity analyses: No difference between the specific SGLT2 inhibitors, no cardiovasc benefit in those started on DPP-4 inhibitors (as anticipated)

 

Commentary:

-- SGLT2 inhibitors do have a pretty consistent benefit in several studies for hospitalizations for heart failure in those with cardiovascular disease or at high cardiovascular risk, leading to FDA approval for heart failure

-- these medications have also been endorsed as a preferred second-line treatment after metformin by the American Diabetes Association

     -- the 2022 update of the Diabetes Care guidelines from the American Diabetes Association's Standards of Care for pharmacologic approaches to diabetes still supports metformin as first-line along with lifestyle modifications (see https://diabetesjournals.org/care/article/45/Supplement_1/S125/138908/9-Pharmacologic-Approaches-to-Glycemic-Treatment ). A few points from this guideline:

        -- they note that GLP-1 agonists added to metformin has the greatest reduction in A1c level (similar to insulin regiments)

        -- there is equipoise about GLP-1 versus SGLT2 as second-line therapy

        -- the combination of GLP-1 and SGLT2 does confirm more benefit than either alone, for those who need a third medication

-- the authors above note that the new European guidelines do support using SGLT2s in treatment naïve diabetics, but in my reviewing these guidelines, there is clearly equipoise between them and GLP-1 receptor agonists (see 2019 ESC Guidelines on diabetes, pre-diabetes, and cardiovascular diseases developed in collaboration with the EASD | European Heart Journal | Oxford Academic (oup.com)  

-- to me, there still remain questions about the specific cardiovascular benefit from SGLT2 inhibitors.

    -- There is no question that there is a decrease in new onset heart failure and hospitalizations for heart failure in those on an SGLT2. This is not so surprising, since the SGLT2s function as loop diuretics. Though studies are lacking, it seems quite likely that if people were put on low-dose standard loop diuretics (e.g. torsemide 5 mg), they would also have a decrease in heart failure episodes. And, the old loop diuretics have been around for ages, have well-known long-term adverse effects that can easily be monitored/fixed, and are titratable to the patient’s heart failure needs. There is certainly an array of adverse effects from SGLT2s, quite a long list, though the really severe adverse effects that we have found to date are pretty uncommon.

    -- the SGLT2 benefit for atherosclerotic events is a lot less clear. The study cited most commonly is the empagliflozin one (see http://gmodestmedblogs.blogspot.com/2015/12/empagliflozin-good-and-bad.html), where the researchers did not disaggregate cardiovascular deaths related to heart failure. In fact, their table S5, buried in the appendix, notes that there were 137 cardiovascular deaths in the placebo group (5.9%) versus 172 with empaglioflozin (3.7%). But further breakdown shows:

    -- sudden-death: placebo 1.6% versus empagliflozin 1.1%

    -- worsening heart failure: 0.8% versus 0.2%

    -- acute MI: 0.5% versus 0.3%

    -- stroke: 0.5% versus 0.3%

    -- cardiogenic shock: 0.1% for both

    -- “other cardiovascular death”, defined as “fatal cases that were not assessable due to lack of information and were presumed to be cardiovascular deaths”: 2.4% versus 1.6%

        -- BUT this latter amorphous category comprised 42% of all of the deaths, far overwhelming all of the more specific cardiovascular death categories assessed. so, who knows what this all means???

    -- and, in order to try to get comparable groups, the researchers had to increase the use of other diabetic meds in the control group. Unfortunately, many more people were on insulin or sulfonylureas, which in several studies increase cardiovascular disease. so, was it benefit of empagliflozin in that group or harm in the control group getting more insulin/sulfonylureas?? 

        -- for more info on insulin and cardiovascular events, see http://gmodestmedblogs.blogspot.com/2016/08/insulin-vs-glp-1-agonists-for-patients.html , or http://gmodestmedblogs.blogspot.com/2018/03/loosening-a1c-goal-is-low-a1c-really.html

-- there was a brief report suggesting that the studies on canagliflozin, empagliflozin, and dapagliflozin all showed a trend to benefit for MI in those with CVD but not stroke or amputations, though the authors' quick calculations suggested that putting all three studies together revealed an 11% MI benefit. But this report lacks any details of their methodology in adding together 3 disparate trials, so hard to assess their results more rigorously (for their report see dm dec MI SGLT2 review cardiovascdiabet2019 in dropbox, or doi.org/10.1186/s12933-019-0812-6 ). 

-- GLP-1 agonists, on the other hand, seem to have clearer benefits than SGLT-2s, as noted in http://gmodestmedblogs.blogspot.com/2020/02/diabetes-glp1-agonists-vs-sglt2.html , which has a detailed summary of the strengths and weaknesses of these two medication classes

 

-- one concern I have is that it seems that there were some regular overstatements in this article, including overstating that SGLT2s are the preferred second line drug by the American Diabetes Association, as well as attributing significant atherosclerotic benefit to the SGLT2 inhibitors. To me this smacks of drug company bias (all of the studies were drug company supported) with some embellishment of the benefits of SGLT2s (especially with their primo study on empagliflozin, which really seems to have uninterpretable results regarding atherosclerotic disease)

-- a couple of interesting (to me) sidelines about metformin 

    -- a part of metformin's mechanism of action may be related to creating a healthier gut microbiome. A study of diabetic mice found that those on metformin improved their glycemic profile (as expected), but there was a significant increase in the gut bacterium Akkermansia. In a parallel experiment, just increasing this bacterium in the gut (in the absence of metformin) also enhanced glucose tolerance and decreased adipose tissue inflammation, suggesting that an additional mechanism of action for metformin may be through its effect on the microbiome (see http://gmodestmedblogs.blogspot.com/2014/10/heart-failure-microbiome.html ). And, it may well be that the long-term effects of a healthier microbiome may be an added benefit above and beyond metformin's diabetic control: eg specifically for the array of bad outcomes related to chronic inflammation, such as heart disease, cancer, CKD, NAFLD, autoimmune disease, depression,... (see dm metformin microbiome GUT 2014 in dropbox, or doi.org/10.1136/gutjnl-2013-305370)

-- also, a recent article from the large VA Hospital database found that metformin use in diabetics with non-alcoholic fatty liver disease (as opposed to sulfonylureas or insulin) was associated with a 20% decreased risk of hepatocellular carcinoma [of note, diabetes is a risk factor for increased likelihood of HCC in those with NAFLD, and NAFLD is really common in diabetic patients], and this was independent of the degree of glycemic control (see nafld metformin decd hcc hepatology2021 in dropbox or DOI: 10.1002/hep.32244 )

Limitations:

-- this was an observational retrospective study, limiting its ability to determine causality. For example, were those treated initially with an SGLT-2 inhibitor different than those treated initially with metformin, the usual first-line medication? Were they intolerant of metformin, and this reflected a subgroup of patients who are in other ways different from metformin-tolerant patients, and this difference might have altered their results?

    -- as an observational study, one cannot draw causal conclusions, only associations

-- this was a short-term follow-up, which may have decreased the timeframe necessary to observe serious adverse effects

    -- there were also some differences in the length of follow-up, 10.7 months for SGLT2s and 12.2 months for metformin, and this 1.5 year difference might have affected the numbers of adverse events found

-- the group given SGLT2 inhibitors were younger and may have been very different from those on the standard metformin therapy. they were also likely healthier (fewer prior hospitalizations) so might have been at lower baseline risk for MI, etc. 

    -- we also have no information about nonpharmacologic therapies such as diet or exercise, which might have influenced the results. and these might have been different in this younger cohort. there was also no general socioeconomic data

-- propensity matching is a mathematical device to improve the equivalences of the comparators in a study. However, this statistical manipulation is not necessarily an accurate one: an article found that tramadol for OA was associated with increased mortality, but an associated editorial impressively challenged the results (see http://gmodestmedblogs.blogspot.com/2020/03/tramadol-fo-oa-inc-mortalityprobs-with.html )

So, i think there are significant holes in the studies/logic of substituting SGLT2s for metformin as first-line therapy, and that the benefits of metformin may well exceed the specific diabetic indication (eg healthy changes in the microbiome, decreases in inflammation and likely its associated bad outcomes, and decreasing hepatocellular carcinoma in those with diabetes who are prone to more frequent and aggressive non-alcoholic fatty liver disease). and, metformin is safe and well-known from its 7 decades of use... (to access many blogs on metformin, including its more liberal use in CKD and overall negligible association with lactic acidosis, see http://gmodestmedblogs.blogspot.com/search?q=metformin )

my bottom line: i will continue to use metformin as the primary med for diabetes, and will continue to consider a GLP-1 as my preferred second choice, with the addition of SGLT-2 as the third. To the extent possible, i think it makes sense to try to avoid insulin/sulfonylureas since there is pretty strong suggestion in many studies that cardiovascular disease may be increased with these agents. Both of these meds often lead to hyperinsulinemia/insulin resistance and its array of potential adverse health effects that can occur with these meds

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 all of the blogs (2 options):

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 click on the magnifying glass on top right, then  type in a name in the search box and get all the blogs with that name in them

 

or: go to https://www.bucommunitymedicine.org/ , a website from the Community Medicine section at Boston Medical Center.  This site does have a very searchable and accessible list of my blogs (though there have been a few that did not upload over the last year or two). but overall it is much easier to view blogs and displays more at a time.

 

 

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