Diabetes Technology for the Endocrinologist,...

Post on 21-Mar-2019

216 views 0 download

Transcript of Diabetes Technology for the Endocrinologist,...

Diabetes Technology for the Endocrinologist, 2017

Irl B. Hirsch, MD University of Washington

Dualities • Research: Helmsley Charitable Trust, JDRF, ADA,

NIDDK, CDC • Consulting: Abbott, Roche, Intarcia, Adocia,

Valeritas, Big Foot

Raise Your Hand If In Your Practice…

• Every patient gets their meter downloaded • Every patient gets their pump downloaded • Every patient gets their CGM downloaded • No patient gets downloaded, but you encourage your

patients to download at home • If you have at least one computer designated for

downloading • If your downloads are uploaded into your EMR

Required (or at least desirable) Infrastructure for Your Office

• Person knowledgeable with insurance/payers, PAs, “verbiage” to efficiently gain approval

• Person (doesn’t have to be CDE) who can train patients; pros and cons of using all industry support

• Coming soon? Kiosks in the waiting room for patients to download their own technology

• Mechanisms (stickers) to ensure technologies do not get mixed up

• Dedicated computer(s) for downloading

What You Need (Minimal) • A program which allows downloading of various

meters/pumps/sensors • Clinipro® (Numedics.com), Glooko/Diasend,

Carelink®, Tidepool • Ideally, many of the native softwares are also

available • A better solution: immediate upload to “the cloud”

• Livongo®, Accucheck® Aviva Connect, Dexcom Clarity

Metrics in Diabetes • Meaningful Use (performance metrics) • “Glucometrics”: analysis of blood glucose data

• To better understand the glycemic fingerprint of each individual patient: A1C, mean, SD, CV, TIR, LBGI

What we always seem to be doing in diabetes, especially in an ACO environment: updating

the metrics (“work in progress”)

Standard Deviation Our clinically available measurement of glycemic

variability for both SMBG and CGM Many other statistical analysis are available but

correlation will be with CGM and outcomes, not SMBG (current studies and new consensus using CV)

Can determine both overall and time specific SD SMBG: I prefer a month of data for less potential

bias/outliers CGM: 2 weeks is fine

Calculation To Determine SD Target

SD X 2 < mean, may be difficult for some type 1 patients. Formulas only relevant for

mean BG between 120-180

SD X 3 < MEAN SMBG

CGM SD X 3 < MEAN

Better metrics: CV, TIR, TBR, TAR all to be correlated with outcomes

• Read “aggregate” mean/SD

Example: Richard

Example: Richard

♦ Read “aggregate” mean/SD ♦ Read frequency of testing

Example: Richard ♦Read “aggregate” mean/SD ♦Read frequency of testing ♦Review time-specific means/SDs

?

Other Advances in Home Blood Glucose Testing

• Business model: meters and strips are provided for free (e.g., large company such as Boeing, Ford, Amazon, etc.)

• Glucose is uploaded to cloud immediately when tested

• “Poor control” can be intercepted early by company HCPs

• If critical BG tested, patient is called or texted immediately by Livongo CDE

• Waiting for studies showing overall reduced cost to system with this model

Available Now (but not yet in US)

Small BG meter, size of

memory stick

Insulin pen

monitor

Mobile Application

Your office

Audible activation

Glucome.com

CASE 1: Christie • 32 y/o woman on 8 units glargine BID with pre-meal lispro 1:15, ISF 50 day,

60 at HS; jogs at 7am 5X’s/week; b’fast at 8a, lunch at 1p, dinner at 7p

1. Too much basal Too much prandial with downward trend

Poor prandial replacement Mean/SD 126/47; A1C 6.0%

As is often the case, the A1C doesn’t reflect all of the major challenges are patients are having

Prediction: CV Will Replace SD (for CGM) By Both Providers and Their Patients

• July, 2017: Dexcom Clarity introduced CV to their statistical home page

• Some understand SD, but how in the world to interpret CV?

• Recall: issue of glycemic variability is risk of added hypoglycemia, which brings us to

So What is the Goal CV?

Diabetes Care 2017;40:832-38

ONE Reason Why Downloading Is So Important: Understanding the Mean and Estimated A1C

This patient’s HbA1c is 8.2%. She has a glycation gap due to iron deficiency anemia

24 year-old woman, MDI, using Dexcom, 14 years T1D

What Alters A1C Hematologic conditions Anemia Accelerated erythrocyte turnover Thalassemia Sickle cell disease Reticulocytosis Hemolysis Physiologic States Aging Pregnancy Drugs/Medications Alcohol Opioids Vitamin C Vitamin E Aspirin Erythropoetin Dapsone Ribavirin

Disease States HIV infection Uremia Hyperbilirubinemia Dyslipidemia Cirrhosis Hypothyroidism* Medical Therapies Blood transfusion Hemodialysis Miscellaneous Glycation rate Protein turnover Race and ethnicity* Laboratory assay Glycemic Variability Smoking Mechanical heart valves Exogenous testosterone?

In a typical diabetes practice, 14%-25% of A1C measurements are misleading

But even with all of these limitations, HbA1c is actually a more crude test than this

Estimated Mean Glucose: Current Study vs. ADAG

HbA1c

Current Study N=598

(mg/dL) mean (95% CI)

ADAG N= 507 (mg/dL)

mean (95% CI) 6% 132 (93-171) 126 (100-152) 7% 155 (116-194) 154 (123-185) 8% 178 (140-217) 183 (147-217) 9% 201 (163-240) 212 (170-249)

10% 224 (186-263) 240 (193-282)

CASE 2

• 45 y/o Ethiopean man moved to Seattle to work as an executive for a coffee company (we don’t have Dunkin’ Donuts’ in Seattle)

• T2DM X 10 years, on insulin X8 years • BMI 36; A1C 8.7% • Has been on SAP for 4 years

SEPT 2014:MEAN 197; SD 75; A1C = 8.7%

Good basal

60 g CHO

BC: 17U

44 g CHO

BC: 11.5 U

Stayed high: last bolus not enough

No bolus: too much basal?

Insulin given with food

Insulin given with food

With All of This Technology, His Major Problem is Easy (or Should Be) to Fix

• Insulin needs to be given before one eats! • Why is this so difficult for so many patients?

OCT 2016: Mean = 156; SD = 41; A1C = 7.1%

So What Happened To Our Patient’s Diabetes Over the Next Two Years?

CSII: What Does The Clinical Endocrinologist Need To Know in 2017?

Features of CSII: Evolution Over the Years

• Many basal (alternating, temp) and bolus (extended, 2-component) options but to me, the most important ones both grossly under-utilized by patient and provider:

• Downloading-both for patients and providers. • Bolus calculator: when used appropriately is a

tremendous tool!

The Problem with Bolus Calculators

• They are not “smart”!

• If the glucose is trending up or down, more or less insulin will be required

Estimate Details: Bolus Calculator

Est. total 4.5 U

Food intake 36 g

(meter) BG 210

Food 3 U

Correction 2 U

IOB 0.5 U

Smart? Is There Any Reason To Care? Why We Should Be Excited About Smart

(Pumps) • Integrating pump with a sensor and a computer

could potentially make the insulin delivery smarter

CASE 3: 20 y/o T1D, Down Syndrome T1D X 11 years, CSII and CGM. Last SH 5 years ago; HbA1c 7.6%

Hypoglycemic seizure at 1:30am. Why?

High basals in the evening do help to “cushion” dinner but usually result in hypoglycemia if not snacking. Timing of nocturnal basal change is important!

CASE 4-Other Downloading Options

• 44 y/o woman, T1D X 22 years, using Omnipod and Dexcom

• Frustrated with downloading choices • Decided to try Tidepool (Tidepool.org) • Last HbA1c 6.8%

CGM SMBG

Toggle cursor: 29 g carb, suggested bolus 2.4 u, delivered 1.7 u

Ouch!

Blip

Case 8a: Twin Sister! Dexcom, No Pump

Highest average after lunch

Most variability after breakfast and bedtime

Move the cursor

BAD DAY

GOOD DAY

CASE 5 • 60 y/o woman, 41 years T1D using Animas Vibe • Past history significant for PDR s/p PRP, DAN with

gastroparesis • After many years of no diagnosis, found in 2010 to have stiff-

person’s syndrome • HbA1c prior to SPS Dx usually in the 7-8% range, after Dx

most in 9-11% range. • 75-80% of total daily insulin is basal insulin • What’s going on?

h

Case 5: AGP from 60 y/o Woman with SPS Feb 2016, HbA1c 10.1%

CASE 5: The Answer to the Problem

(but you need the download!) No bolus

Why Are We So Focused on CSII?

The majority of type 1’s use MDI and this is still the gold standard for severely insulin deficient type 2’s

Important Point • In the US, about 30% of our type 1 patients use CSII

(60% in the T1D Exchange) and most agree the majority of adult type 1 patients receive their care by non-endocrinologists (one recent estimate was that 2/3 of T1D is only seen by primary care)

• The majority of patients use MDI-will closed loop systems increase CSII use for those who are cared for by both endocrinologists and non-specialists?

• DIAMOND and GOLD studies (JAMA, 2017): outcomes improved with CGM and MDI

What About the 70% of T1D Who Use MDI? • Companion Medical “InPen” system • Approved by FDA 8/16, to be launched 2017 (?) • Will track prandial insulin doses (cartridge pens)

and send to paired app via Bluetooth • App also includes a bolus calculator (with real-time

IOB) • Many other companies working on these blue-

tooth enabled insulin pens

What does it look like?

Wait a Minute!

What about our growing number of patients who can’t afford this technology, and in fact

can’t even afford their insulin?

NPH Insulin in 2017 • 45 y/o man who still had insulin lispro and a few CGMs from

last year. Deductable is $4000 and can’t afford list price of glargine (let alone degludec), so he simply used NPH instead. His A1C is 6.9%

NPH isn’t so bad if you know how to use it!

Conclusion • Technology for MDI is finally improving • Downloading should be part of the “vital signs” for

every patient using a meter, pump, or sensor • The downloading software is improving, is connected

to the cloud, and should assist us in helping our patients

• How the endocrinologist can efficiently utilize all of this technology in our current system requires further research