diabetes guideline update: pushing SGLT-s inhibitors for CKD

The American Diabetes Association (ADA) just updated their diabetes "Standards of Medical Care" to incorporate results from the CREDENCE trial, which found that the sodium-glucose cotransporter-2 (SGLT2) inhibitor canagliflozin was associated with decreased risk for both renal and cardiovascular failure events in patients with type 2 diabetes (DM2) and chronic kidney disease (CKD). see  http://www.diabetes.org/newsroom/press-releases/2019/updates-standards-medical-care-diabetes.html , for their summary of the revisions, with links to the appropriate sections of the original document, conveniently with the updated materials highlighted in yellow.

For my review of the original 2019 full guidelines, seehttp://gmodestmedblogs.blogspot.com/2019/01/2019-diabetes-guidelines.html

Details of update (will focus on those with DM2):

--urinary albumin excretion (eg spot urinary albumin/creatinine ratio, UACR) should be measured at least once a year along with eGFR in all patients with DM2 [not much change here]

--for those with DM2 and CKD "consider use of an SGLT-2 inhibitor or GLP-1 receptor agonist to reduce risk of chronic kidney disease progression, cardiovascular events, or both"

--continue to monitor UACR in those with albuminuria treated with ACE/ARB to assess response to Rx and progression of CKD

--consider using SGLT2 in those when eGFR >30 and especially in those with albuminuria >300

the CREDENCE trial (see dm canagliflozin dec renal outcomes credence nejm2019 in dropbox, or DOI: 10.1056/NEJMoa1811744):

Details:

--4401 patients with DM2 and albuminuric CKD (eGFR 30-90, UACR 300-5000 mg/g), already on a renin-angiotensin system blocker, were randomized to canagliflozin 100mg daily vs placebo

    --patients from 690 sites in 34 countries

--median age 63, 34% female, 67% white/5% black/20% asian, 15% smoker, 97% hypertension, 15% heart failure, diabetes duration 16 yrs, cardiovasc dz 50%, amputation 5%, BMI 31, BP 140/78

--a1c 8.3%, eGFR 56, UACR 927 mg/g

--primary outcome: composite of end-stage renal disease (dialysis, transplant, or sustained eGFR<15), doubling of serum creatinine, or renal/cardiovasc deaths

--mediation adherence to treatment 84%

--the trial was stopped early (at 2.6 years) because of planned interim analysis of med benefit

Results:

--canagliflozin was associated with:

    --A1c decrease: 0.31 percentage points

    --BP decrease: 3.3/0.95 mmHg

    --weight decrease: 0.8 kg

    --UACR decrease (next checked at 6 months; decrease might have been much sooner??): about 350 mg/g (!!), then paralleled the placebo group

    --eGFR actually got worse in the first year (happened at next check at 3 weeks after starting canagliflozin and continued over next 3 checks until 12 mo), then got better with a less steep deterioration of renal function than placebo

--primary outcome:

    --30% lower with canagliflozin, HR 0.70 (0.59-0.82), p=0.00001; event rate of 43.2 vs 61.2 per 1000 patient-yrs

--renal-specific endpoints (ESRD, doubling of creat, death from renal causes):

    --34% lower, HR 0.66 (0.53-0.81), p<0.001

    --ESRD (by itself): 32% lower, HR 0.68 (0.54-0.86), p=0.002

--Cardiovascular death, MI, or stroke

    --20% lower, HR 0.80 (0.67-0.95), p=0.01

    --cardiovascular death (by itself): 22% lower, HR 0.78 (0.61-1.00), p=0.05

--hospitalization for heart failure

    -- 39% lower, HR 0.61 (0.47-0.80), p<0.001

--adverse events: no difference in amputation or fracture [BUT diabetic ketoacidosis occurred in 2.2 vs 0.2 per 1000 patient-years, though this was not mentioned in the abstract or tabulation of adverse effects, areas readers may look at if not reading the full article in detail]

--from their data, they would extrapolate that for 1000 patients treated with canagliflozin for 2.5 years, there would be:

    --47 fewer patients who had the primary outcome of ESRD, doubling of creatinine, or renal/cardiovasc deaths

    --36 fewer with composite renal outcomes of ESRD, doubling of creatinine, or renal deaths

    --24 fewer ESRD events

    --and 22 fewer hospitalizations for heart failure, and 25 fewer composite cardiovasc events (cardiovasc death, MI, or stroke)

Commentary:

--this trial does add to several others finding renal (and cardiovascular) protection from SGLT-2 inhibitors (the CREDENCE trial having sicker and more renally-compromised patients):

    -- for example, see the CANVAS trials: http://gmodestmedblogs.blogspot.com/2016/06/canagliflozin-decreases-macrovasc.html for eval/critique of these trials which had significant design problems but found a decrease in cardiovascular and renal endpoints, though twice the number of amputations and 71% increase in fractures. other studies have found benefit for empagliflozen, dapagliflozin (though also flawed trials, eg see http://gmodestmedblogs.blogspot.com/2015/12/empagliflozin-good-and-bad.html )

--so, the SGLT-2's do seem to confer renal protection, but is this a result of a direct renoprotective effect??

    --there was a difference in A1c levels in the canagliflozin group: 0.31 percentage points, and:

         -- hyperglycemia itself is associated with renal hyperfiltration/hypertension and renal dysfunction, as well as with increased albumin excretion (UACR). there are several postulated mechanisms by which hyperglycemia alone can lead to hyperfiltration (eg nitric oxide mebabolites), creating the common clinical scenario that patients with high blood sugars who have micro- or macroalbuminuria have much lower levels of albumin excretion just by lowering the blood sugar {see https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3258534/ ).  

        --and there seemed to be a pretty quick large decrease in UACR in this study, noted in the first 6 months of canagliflozin therapy (though might have happened much sooner, since UACR not checked earlier than 6 months), but no further decrease after that time, which seems to me to be consistent with the mechanism that SGLT-2's decrease blood sugar reasonably quickly and this by itself decreased albuminuria.  And many studies have documented that albuminuria is associated with decreased renal function, and that decreasing albuminuria, especially when at the high levels in the CREDENCE study, decreases the rate of further renal decline, both in diabetics and nondiabetics

    --also, there were other changes that might have been renoprotective:

        --the blood pressure decrease in those on canagliflozin, of 3.3/0.95 mmHg

        --the weight decrease of 0.8 kg may play a role, since perhaps the decreased adiposity was associated with decreased hyperinsulinemia (not measured in this study), and hyperinsulinemia is also associated with developing progressive renal disease (see https://www.ncbi.nlm.nih.gov/pubmed/15504934 )

    --and, though these individual differences in these above factors may seem to be small, it is often the case that the combined effects of all of them exceeds the sum of the incremental effects of each one (as well demonstrated with cardiovasc risk factors)

   --which all means, that to be sure it is the specifics of the SGLT-2's (which, by the way, can also cause acute kidney injury, and in this study did lead to a general deterioration of renal function initially), it would be really useful to have a study done where the placebo group received more antidiabetic meds that would equivalently decrease A1c, perhpas weight/blood pressure as well, but choosing meds that themselves do not seem to cause cardiac/renal effects (eg the hyperinsulinemia associated with insulin or sulfonylureas).

    --and, it really would be useful to have patient-specific data: did those patients who had the most dramatic effect on blood sugar have the best renal outcomes?? which might suggest that it was the lowering of blood sugars which was largely responsible (and the average A1c in this study was 8.3, not great control). maybe just more aggressive diabetic control is the answer???

--though it is notable that canagliflozin was effective on top of ACE/ARB, there was no mention of the doses of these meds (were they optimized??). other drugs also decrease albumin excretion, including statins and spironolactone, and nondihydropyridine calcium blockers. perhaps adding these to the diabetes regimen would help with albuminuria and decrease the rate of renal deterioation??

--the CREDENCE participants were pretty advanced diabetic patients with about 1 gram of albuminuria and CKD, so these results may not be applicable to less renal-impaired people (though the renoprotection, as noted above, extended to other SGLT2's)

--another concern is that this study was stopped early because of benefit. but, it may well be (as the authors suggest) that this might limit the power of some secondary outcomes and overestimate the effect size. but i would add: several of the adverse effects may take time, so there might be an important underestimation of the many documented and clinically very severe known effects of SGLT-2 inhibitors (eg, ketoacidosis despite lowish blood sugars, urosepsis, fractures/amputations)

-- and, we know that liraglutide is also renoprotective (also documentented with semaglutide and presumably for the other GLP-1 agonists), eg see http://gmodestmedblogs.blogspot.com/2017/09/liraglutide-also-decreases-adverse.html [of course, the renal protection of liraglutide and semaglutide  may alse be largely mediated in their studies through some of the same issues raised above for canagliflozin, ie, just lowering blood sugar in a way which is also cardioprotective]

There are lots of prior blogs on the problems with SGLT-2 inhibitors:

-- http://gmodestmedblogs.blogspot.com/2016/06/canagliflozin-decreases-macrovasc.html for eval/critique of the CANVAS trials which had significant design problems but found a decrease in cardiovascular and renal endpoints though twice the number of amputations and 71% increase in fractures, and:

-- http://gmodestmedblogs.blogspot.com/2016/05/another-fda-alert-about.html for an FDA alert on amputations

-- http://gmodestmedblogs.blogspot.com/2018/09/sglt-2-inhibitors-and-increased-lower.html for a large-scale analysis showing the risk of amputations was high for all of the SGLT-2 inhibitors

-- http://gmodestmedblogs.blogspot.com/2019/05/sglt-2-inhibitors-and-fourniers-gangrene.html a newer analysis finding increases in Fournier’s gangrene with SGLT-2 inhibitors

-- http://gmodestmedblogs.blogspot.com/2015/05/sglt2-inhibitors-for-diabetes-may-cause.html for a blog on SGLT-2 inhibitors causing ketoacidosis, even with blood sugars <200 mg/dL

-- https://www.fda.gov/drugs/drug-safety-and-availability/fda-drug-safety-communication-fda-revises-labels-sglt2-inhibitors-diabetes-include-warnings-about for FDA alert on serious UTIs/urosepsis with SGLT-2’s

so, how should we put this new study on canagliflozin into context, given the updated ADA guidelines strongly pushing it???

--as with all studies, this one was also flawed. would be great to know the doses of the ACE/ARBs in these patients, the patient-specific changes in blood sugar vs change in albuminuria (as noted above, since a major factor improving renal function may just be lowering the blood sugar), whether they were on other meds that might affect albuminuria (statins, spironolactone, nondihydropyridine calcium channel blockers, etc).  And perhaps just putting patients on these meds might have equal renoprotection as the SGLT2's just by lowering urinary albumen levels???

--these SGLT2-'s may really be an advance in diabetes management, but i still go to GLP-1 agonists first for several reasons:

    --GLP-1's have very impressive decreases in A1c levels, even in chronically hyperglycemic diabetics who are unable to change much of their lifestyle issues.  i have seen many cases of double-digit A1cs falling to the 6.5-7.5 range.

    --GLP-1's actually replace a physiologic aberrancy associated with diabetes (the "incretin" effect), such that diabetics' innate GLP-1 secretion is dramatically reduced (about 40%) by a meal.  SGLT-2's are effectively loop diuretics which reduce glucose uptake (they do help with heart failure as a result, but the regular old torsemide/furosemide certainly do and are titratable to the degree of heart failure). And GLP-1's are also associated with renal and cardiac protection.  

    --GLP1's are quite good at decreasing appetite and weight (liraglutide has made it onto the list of weight-loss drugs). 

    --one downside of GLP-1's (vs SGLT2-s) is that they are injectable and mostly once a week. but there was an impressive new trial of the GLP-1 semaglutide as an oral agent (see dm glp1 oral semaglutide jama2019 in dropbox, or doi:10.1001/jama.2019.2942), comparing it to a DPP-4 inhibitor (sitagliptin) finding semaglutide was associated with both larger decreases in A1c by 0.5 percentage points and body wt by 2.5 kg  [i should add:  even now, i have been able to convince several injection-averse patients, who it turns out in fact were more opposed to insulin than the thought of an injection, to trial GLP-1's]. 

--it would be very helpful to have a real head-to-head comparison of an SGLT-2 to a GLP-1, with reasonable equivalence of achieved A1c’s/body weight/blood pressure, to really see what the best drug is [though I think this is very unlikely to happen: drug companies are very unlikely to participate in a trial which might show that their current blockbuster was actually inferior to another’s]

All of the studies on all of these meds (SGLT-2’s and GLP-1’s) are drug company sponsored and flawed, likely designed to improve the likelihood of a positive result for them. But, I really cannot understand the ADA's increased emphasis on using SGLT-2’s, especially since the GLP-1’s are so dramatically beneficial in terms of both large decreases in A1c levels and cardiorenal protective effects, are actually reinforcing a normal physiological mechanism for glucose homeostasis which happens to be diminished in diabetics, are associated with significant weight loss, and are bereft of many significant adverse effects (especially as compared to SGLT-2’s). so, to me, a more reasoned approach would be to prescribe GLP-1's, and if not tolerated, consider SGLT2's...

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:

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

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