Thread by @JJheart_doc

A tweetorial.

Here's an intriguing analysis just published in Diabetes Care by the great folks at @AmDiabetesAssn and @ADA_Journals. We examined concentrations of insulin-like growth factor binding protein-7 (IGFBP7) at BL and 1Y in the CANVAS study.

care.diabetesjournals.org/content/early/…

1/

Why IGFBP7? Good question.

Though IGFBP7 "looks like" an IGF binder, it is lousy at this job. But IGFBP7 has other roles: it is a cell cycle arrest biomarker in the 'senescence associated secretory phenotype'.

When cells are exposed to IGFBP7 in excess, fibrosis follows.

2/

We previously found IGFBP7 was associated with cardiac structural and functional abnormalities and outcome in patients with heart failure, but we were interested to see how it would act in patients with #T2D, such as those in CANVAS.

Why?

3/

IGFBP7 is higher in obesity, in those with #T2D, and has been shown in animal models to also be expressed in the proximal tubules of an animal model of diabetic kidney disease (#DKD). In urine IGFBP7 predicts AKI, but no data about blood measurement for DKD prediction exist.

4/

We therefore sought to examine whether baseline or 1Y IGFBP7 predicted renal outcomes in #CANVAS and whether the biomarker predicted response to #canagliflozin. Since the #CREDENCE Study also looked at the composite of CV death/renal composite, we also looked at that.

5/

Endpoints included:

Sustained eGFR reduction
Progression to macroalbuminuria
First progression of albuminuria
CV death
Renal composite (40% decrease in eGFR, RRT, or renal death)
Renal composite/macroalbuminuria
CV death/renal composite

6/

The study population included 3577 with baseline samples for IGFBP7. The study participants with biomarker samples were 62.5 ± 7.8 years of age, had T2D with a baseline eGFR of 77.5 ± 18.5 mL/min/1.73 m2 and UACR of 9.5 ± 36.6 mg/g.

7/

BL IGFBP7 was similar between those allocated to receive cana (97.6 ng/mL) or placebo (96.7 ng/mL). In models adjusted for log-transformed baseline IGFBP7, age, HbA1c, eGFR, UACR, and treatment assignment, cana did not lower IGFBP7 (β = –0.0044 [–0.0201, 0.0113]; P = 0.58).

8/

But here's where it gets very, very interesting.

In fully-adjusted models, IGFBP7 independently predicted renal and renal/CV outcomes.

Baseline was a strong predictor, but in each case the 1Y measurement added even more.

How much more?

9/

In my 20 years of doing this, I have never seen a continuous log-transformed variable with an adjusted hazard ratio of 15.7--that's a log-transformed result. Wow.

To see if baseline to 1Y IGFBP7 values predicted outcomes we then examined each by biomarker/treatment.

10/

In most cases, we found no baseline biomarker-by-treatment interaction. So, whether the IGFBP7 was high or not, outcome improvement from cana was similar. Still, with higher absolute risk comes greater absolute benefit (figure).

11/

One exception was the fact those with higher IGFBP7 concentrations appeared to specifically benefit from cana with respect to the endpoint of first progression of micro- or macroalbuminuria where interaction between baseline IGFBP7 and canagliflozin effects was observed.

12/

Curious, right? Consistent with a link to DKD, those with baseline IGFBP7 ≥96.5 ng/mL had greater impact of cana on progression of albuminuria (44.7% vs 35.7%; HR, 0.64) than in those with lower IGFBP7 concentrations (30.9% vs 32.2%; HR, 0.95; Pinteraction = 0.003).

13/

Shown another way, across baseline quartiles of IGFBP7, larger effects of canagliflozin vs placebo were confirmed above the median split (Figure 2; Pinteraction = 0.03).

14/

Though urinary IGFBP7 has been linked to AKI, these are the first to show blood measurement of IGFBP7 as a biomarker to predict progression of CHRONIC kidney disease. The data also raise the question as to whether IGFBP7 might be predict who might benefit most from SGLT2i.

15/

More data to come on the links to IGFBP7 and cardiac outcomes in CANVAS, as well as more data in other datasets with DKD/cardiac outcomes.

FIN