Use and Discontinuation of Insulin Treatment Among Adults Aged 75 to 79 Years With Type 2 Diabetes | Clinical Pharmacy and Pharmacology | JAMA Internal Medicine | JAMA Network
[Skip to Navigation]
Figure 1.  Flowchart of Cohort Formation and Outcome Assessment
Flowchart of Cohort Formation and Outcome Assessment

T1DM indicates type 1 diabetes mellitus.

Figure 2.  Cohort Health Status Definition Compared With American Diabetes Association (ADA) Guideline Definition
Cohort Health Status Definition Compared With American Diabetes Association (ADA) Guideline Definition

ADL indicates activity of daily living; CHF, congestive heart failure; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; CVD, cardiovascular disease; HTN, hypertension; IADL, instrumental activity of daily living; MI, myocardial infarction.

Figure 3.  Insulin Use at Baseline and Insulin Use Discontinuation After Age 75 Years by Health Status and Hemoglobin A1c Category
Insulin Use at Baseline and Insulin Use Discontinuation After Age 75 Years by Health Status and Hemoglobin A1c Category

A, Bars are grouped into sets of 3 based on hemoglobin A1C (HbA1c) level measured at baseline. B, Bars are grouped into sets of 3 based on HbA1c level measured prior to most recent insulin dispensing. Health status was measured prior to censorship (ie, end of cohort, discontinuation, or death).

Table 1.  Individual Characteristics by Insulin Use at Age 75 Yearsa
Individual Characteristics by Insulin Use at Age 75 Yearsa
Table 2.  Individual Characteristics by Insulin Use Discontinuation After Age 75 Yearsa
Individual Characteristics by Insulin Use Discontinuation After Age 75 Yearsa
1.
Centers for Disease Control and Prevention. United States Diabetes Surveillance System. https://gis.cdc.gov/grasp/diabetes/DiabetesAtlas.html. Accessed August 14, 2018.
2.
Gerstein  HC, Miller  ME, Byington  RP,  et al; Action to Control Cardiovascular Risk in Diabetes Study Group.  Effects of intensive glucose lowering in type 2 diabetes.  N Engl J Med. 2008;358(24):2545-2559. doi:10.1056/NEJMoa0802743PubMedGoogle ScholarCrossref
3.
Patel  A, MacMahon  S, Chalmers  J,  et al; ADVANCE Collaborative Group.  Intensive blood glucose control and vascular outcomes in patients with type 2 diabetes.  N Engl J Med. 2008;358(24):2560-2572. doi:10.1056/NEJMoa0802987PubMedGoogle ScholarCrossref
4.
Lipska  KJ, Krumholz  H, Soones  T, Lee  SJ.  Polypharmacy in the aging patient: a review of glycemic control in older adults with type 2 diabetes.  JAMA. 2016;315(10):1034-1045. doi:10.1001/jama.2016.0299PubMedGoogle ScholarCrossref
5.
Huang  ES, Laiteerapong  N, Liu  JY, John  PM, Moffet  HH, Karter  AJ.  Rates of complications and mortality in older patients with diabetes mellitus: the diabetes and aging study.  JAMA Intern Med. 2014;174(2):251-258. doi:10.1001/jamainternmed.2013.12956PubMedGoogle ScholarCrossref
6.
ORIGIN Trial Investigators.  Predictors of nonsevere and severe hypoglycemia during glucose-lowering treatment with insulin glargine or standard drugs in the ORIGIN trial.  Diabetes Care. 2015;38(1):22-28. doi:10.2337/dc14-1329PubMedGoogle ScholarCrossref
7.
Bruderer  SG, Bodmer  M, Jick  SS, Bader  G, Schlienger  RG, Meier  CR.  Incidence of and risk factors for severe hypoglycaemia in treated type 2 diabetes mellitus patients in the UK: a nested case-control analysis.  Diabetes Obes Metab. 2014;16(9):801-811. doi:10.1111/dom.12282PubMedGoogle ScholarCrossref
8.
American Diabetes Association.  11. Older Adults: Standards of Medical Care in Diabetes—2018.  Diabetes Care. 2018;41(suppl 1):S119-S125. doi:10.2337/dc18-S011PubMedGoogle ScholarCrossref
9.
Moreno  G, Mangione  CM, Kimbro  L, Vaisberg  E; American Geriatrics Society Expert Panel on Care of Older Adults with Diabetes Mellitus.  Guidelines abstracted from the American Geriatrics Society Guidelines for Improving the Care of Older Adults With Diabetes Mellitus: 2013 update.  J Am Geriatr Soc. 2013;61(11):2020-2026. doi:10.1111/jgs.12514PubMedGoogle ScholarCrossref
10.
Department of Veterans Affairs, Department of Defense. VA/DoD Clinical Practice Guideline for the Management of Type 2 Diabetes Mellitus in Primary Care. https://www.healthquality.va.gov/guidelines/CD/diabetes/VADoDDMCPGFinal508.pdf. Published April 2017. Accessed August 14, 2018.
11.
Qaseem  A, Wilt  TJ, Kansagara  D, Horwitch  C, Barry  MJ, Forciea  MA; Clinical Guidelines Committee of the American College of Physicians.  Hemoglobin A1c targets for glycemic control with pharmacologic therapy for nonpregnant adults with type 2 diabetes mellitus: a guidance statement update from the American College of Physicians.  Ann Intern Med. 2018;168(8):569-576. doi:10.7326/M17-0939PubMedGoogle ScholarCrossref
12.
Lipska  KJ, Yao  X, Herrin  J,  et al.  Trends in drug utilization, glycemic control, and rates of severe hypoglycemia, 2006-2013.  Diabetes Care. 2017;40(4):468-475. doi:10.2337/dc16-0985PubMedGoogle ScholarCrossref
13.
Pathak  RD, Schroeder  EB, Seaquist  ER,  et al.  Severe hypoglycemia requiring medical intervention in a large cohort of adults with diabetes receiving care in U.S. integrated health care delivery systems: 2005-2011.  Diabetes Care. 2016;39(3):363-370. doi:10.2337/dc15-0858PubMedGoogle ScholarCrossref
14.
Lee  SJ.  So much insulin, so much hypoglycemia.  JAMA Intern Med. 2014;174(5):686-688. doi:10.1001/jamainternmed.2013.13307PubMedGoogle ScholarCrossref
15.
Tseng  CL, Soroka  O, Maney  M, Aron  DC, Pogach  LM.  Assessing potential glycemic overtreatment in persons at hypoglycemic risk.  JAMA Intern Med. 2014;174(2):259-268. doi:10.1001/jamainternmed.2013.12963PubMedGoogle ScholarCrossref
16.
Lipska  KJ, Ross  JS, Miao  Y, Shah  ND, Lee  SJ, Steinman  MA.  Potential overtreatment of diabetes mellitus in older adults with tight glycemic control.  JAMA Intern Med. 2015;175(3):356-362. doi:10.1001/jamainternmed.2014.7345PubMedGoogle ScholarCrossref
17.
Arnold  SV, Lipska  KJ, Wang  J, Seman  L, Mehta  SN, Kosiborod  M.  Use of intensive glycemic management in older adults with diabetes mellitus.  J Am Geriatr Soc. 2018;66(6):1190-1194. doi:10.1111/jgs.15335PubMedGoogle ScholarCrossref
18.
Sussman  JB, Kerr  EA, Saini  SD,  et al.  Rates of deintensification of blood pressure and glycemic medication treatment based on levels of control and life expectancy in older patients with diabetes mellitus.  JAMA Intern Med. 2015;175(12):1942-1949. doi:10.1001/jamainternmed.2015.5110PubMedGoogle ScholarCrossref
19.
Thorpe  CT, Gellad  WF, Good  CB,  et al.  Tight glycemic control and use of hypoglycemic medications in older veterans with type 2 diabetes and comorbid dementia.  Diabetes Care. 2015;38(4):588-595.PubMedGoogle Scholar
20.
Karter  AJ, Schillinger  D, Adams  AS,  et al.  Elevated rates of diabetes in Pacific Islanders and Asian subgroups: the Diabetes Study of Northern California (DISTANCE).  Diabetes Care. 2013;36(3):574-579. doi:10.2337/dc12-0722PubMedGoogle ScholarCrossref
21.
Messer  LC, Laraia  BA, Kaufman  JS,  et al.  The development of a standardized neighborhood deprivation index.  J Urban Health. 2006;83(6):1041-1062. doi:10.1007/s11524-006-9094-xPubMedGoogle ScholarCrossref
22.
Stoddard  PJ, Laraia  BA, Warton  EM,  et al.  Neighborhood deprivation and change in BMI among adults with type 2 diabetes: the Diabetes Study of Northern California (DISTANCE).  Diabetes Care. 2013;36(5):1200-1208. doi:10.2337/dc11-1866PubMedGoogle ScholarCrossref
23.
Centers for Disease Control and Prevention. Defining adult overweight and obesity. https://www.cdc.gov/obesity/adult/defining.html. Updated April 11, 2017. Accessed August 7, 2018.
24.
Levey  AS, Stevens  LA, Schmid  CH,  et al.  A new equation to estimate glomerular filtration rate.  Ann Intern Med. 2009;150(9):604-612. doi:10.7326/0003-4819-150-9-200905050-00006PubMedGoogle ScholarCrossref
25.
Campbell  KH, Huang  ES, Dale  W,  et al.  Association between estimated GFR, health-related quality of life, and depression among older adults with diabetes: the Diabetes and Aging Study.  Am J Kidney Dis. 2013;62(3):541-548. doi:10.1053/j.ajkd.2013.03.039PubMedGoogle ScholarCrossref
26.
Parker  MM, Moffet  HH, Adams  A, Karter  AJ.  An algorithm to identify medication nonpersistence using electronic pharmacy databases.  J Am Med Inform Assoc. 2015;22(5):957-961. doi:10.1093/jamia/ocv054PubMedGoogle ScholarCrossref
27.
Karter  AJ, Moffet  HH, Liu  J,  et al.  Glycemic response to newly initiated diabetes therapies.  Am J Manag Care. 2007;13(11):598-606.PubMedGoogle Scholar
28.
Reeve  E, Shakib  S, Hendrix  I, Roberts  MS, Wiese  MD.  Review of deprescribing processes and development of an evidence-based, patient-centred deprescribing process.  Br J Clin Pharmacol. 2014;78(4):738-747. doi:10.1111/bcp.12386PubMedGoogle ScholarCrossref
29.
Scott  IA, Hilmer  SN, Reeve  E,  et al.  Reducing inappropriate polypharmacy: the process of deprescribing.  JAMA Intern Med. 2015;175(5):827-834. doi:10.1001/jamainternmed.2015.0324PubMedGoogle ScholarCrossref
30.
Reeve  E, Gnjidic  D, Long  J, Hilmer  S.  A systematic review of the emerging definition of ‘deprescribing’ with network analysis: implications for future research and clinical practice.  Br J Clin Pharmacol. 2015;80(6):1254-1268. doi:10.1111/bcp.12732PubMedGoogle ScholarCrossref
31.
Anderson  K, Stowasser  D, Freeman  C, Scott  I.  Prescriber barriers and enablers to minimising potentially inappropriate medications in adults: a systematic review and thematic synthesis.  BMJ Open. 2014;4(12):e006544. doi:10.1136/bmjopen-2014-006544PubMedGoogle ScholarCrossref
32.
Anderson  K, Foster  M, Freeman  C, Luetsch  K, Scott  I.  Negotiating “unmeasurable harm and benefit”: perspectives of general practitioners and consultant pharmacists on deprescribing in the primary care setting.  Qual Health Res. 2017;27(13):1936-1947. doi:10.1177/1049732316687732PubMedGoogle ScholarCrossref
33.
Wallis  KA, Andrews  A, Henderson  M.  Swimming against the tide: primary care physicians’ views on deprescribing in everyday practice.  Ann Fam Med. 2017;15(4):341-346. doi:10.1370/afm.2094PubMedGoogle ScholarCrossref
34.
Reeve  E, Wolff  JL, Skehan  M, Bayliss  EA, Hilmer  SN, Boyd  CM.  Assessment of attitudes toward deprescribing in older Medicare beneficiaries in the United States.  JAMA Intern Med. 2018;178(12):1673-1680. doi:10.1001/jamainternmed.2018.4720PubMedGoogle ScholarCrossref
35.
Thompson  W, Lundby  C, Graabaek  T,  et al.  Tools for deprescribing in frail older persons and those with limited life expectancy: a systematic review.  J Am Geriatr Soc. 2019;67(1):172-180. doi:10.1111/jgs.15616PubMedGoogle ScholarCrossref
36.
Munshi  MN, Slyne  C, Segal  AR, Saul  N, Lyons  C, Weinger  K.  Simplification of insulin regimen in older adults and risk of hypoglycemia.  JAMA Intern Med. 2016;176(7):1023-1025. doi:10.1001/jamainternmed.2016.2288PubMedGoogle ScholarCrossref
37.
Fonseca  VA.  Defining and characterizing the progression of type 2 diabetes.  Diabetes Care. 2009;32(suppl 2):S151-S156. doi:10.2337/dc09-S301PubMedGoogle ScholarCrossref
38.
Lang  L, Clifford  A, Wei  L,  et al.  Prevalence and determinants of undetected dementia in the community: a systematic literature review and a meta-analysis.  BMJ Open. 2017;7(2):e011146. doi:10.1136/bmjopen-2016-011146PubMedGoogle ScholarCrossref
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    Original Investigation
    September 23, 2019

    Use and Discontinuation of Insulin Treatment Among Adults Aged 75 to 79 Years With Type 2 Diabetes

    Author Affiliations
    • 1Division of Research,Kaiser Permanente of Northern California, Oakland
    • 2Department of Internal Medicine, Yale School of Medicine, New Haven, Connecticut
    • 3Section of General Internal Medicine, University of Chicago, Chicago, Illinois
    • 4Chicago Center for Diabetes Translation Research, University of Chicago, Chicago, Illinois
    JAMA Intern Med. 2019;179(12):1633-1641. doi:10.1001/jamainternmed.2019.3759
    Key Points

    Question  Is insulin treatment used less frequently and discontinued more often among older individuals with poor health compared with those in good health?

    Findings  In this cohort study of 21 531 adults, it was demonstrated that patients in poorer health were most likely to use insulin at age 75 years and that subsequent discontinuation of insulin use over a 4-year follow-up period was more common in healthier patients even after accounting for level of glycemic control.

    Meaning  Persistent insulin use among older adults with poor health is associated with increased risk for hypoglycemia and limited future health benefit; these results suggest a need to better align current practice with guidelines that support reducing treatment intensity as health status declines.

    Abstract

    Importance  Among older individuals with type 2 diabetes, those with poor health have greater risk and derive less benefit from tight glycemic control with insulin.

    Objective  To examine whether insulin treatment is used less frequently and discontinued more often among older individuals with poor health compared with those in good health.

    Design, Setting, and Participants  This longitudinal cohort study included 21 531 individuals with type 2 diabetes followed for up to 4 years starting at age 75 years. Electronic health record data from the Kaiser Permanente Northern California Diabetes Registry was collected to characterize insulin treatment and glycemic control over time. Data were collected from January 1, 2009, through December 31, 2017, and analyzed from February 2, 2018, through June 30, 2019.

    Exposures  Health status was defined as good (<2 comorbid conditions or 2 comorbidities but physically active), intermediate (>2 comorbidities or 2 comorbidities and no self-reported weekly exercise), or poor (having end-stage pulmonary, cardiac, or renal disease; diagnosis of dementia; or metastatic cancer).

    Main Outcomes and Measures  Insulin use prevalence at age 75 years and discontinuation among insulin users over the next 4 years (or 6 months prior to death if <4 years).

    Results  Of 21 531 patients, 10 396 (48.3%) were women, and the mean (SD) age was 75 (0) years. Nearly one-fifth of 75-year-olds (4076 [18.9%]) used insulin. Prevalence and adjusted risk ratios (aRRs) of insulin use at age 75 years were higher in individuals with poor health (29.4%; aRR, 2.03; 95% CI, 1.87-2.20; P < .01) and intermediate health (27.5%; aRR, 1.85; 95% CI, 1.74-1.97; P < .01) relative to good health (10.5% [reference]). One-third (1335 of 4076 [32.7%]) of insulin users at age 75 years discontinued insulin within 4 years of cohort entry (and at least 6 months prior to death). Likelihood of continued insulin use was higher among individuals in poor health (aRR, 1.47; 95% CI, 1.27-1.67; P < .01) and intermediate health (aRR, 1.16; 95% CI, 1.05-1.30; P < .01) compared with good health (reference). These same prevalence and discontinuation patterns were present in the subset with tight glycemic control (hemoglobin A1c <7.0%).

    Conclusions and Relevance  In older individuals with type 2 diabetes, insulin use was most prevalent among those in poor health, whereas subsequent insulin discontinuation after age 75 years was most likely in healthier patients. Changes are needed in current practice to better align with guidelines that recommend reducing treatment intensity as health status declines.

    Introduction

    In the United States, type 2 diabetes affects more than 20% of adults older than 75 years,1 yet there is little evidence to guide treatment decisions in older patients. Many of the landmark randomized clinical trials that established glycemic control targets for reducing the risk of microvascular and macrovascular complications excluded patients older than 75 years.2,3 Subsequent follow-up of trial participants and observational studies have demonstrated that the benefits conferred by tight glycemic control may not be realized for 5 to 9 years,4,5 a time period that may exceed many older adults’ life expectancies. Moreover, adults older than 75 years are at the greatest risk of hypoglycemia, particularly when insulin is used.6,7

    Professional societies have published guidelines based on clinical expertise and interpolation of the existing evidence. The American Diabetes Association (ADA), American Geriatrics Society, and the US Department of Veterans Affairs all recommend that healthier adults with longer life expectancies be treated to lower glycemic targets, whereas patients with poor health and shorter life expectancy to higher glycemic targets.8-10 The American College of Physicians’ newest guidelines11 do not specify hemoglobin A1c (HbA1c) targets for adults with limited life expectancy; however, the organization recommends that treatment intensity be limited by symptoms of hyperglycemia with general goals of reducing burden of treatment. The unifying theme across these disparate guidelines is the recommendation to individualize treatment in older adults to balance risks and benefits of treatment.

    None of these existing guidelines specifically address the use of insulin in older adults. While insulin may be necessary to achieve glycemic targets, it is also associated with dramatic increases in the risk of hypoglycemia in older adults—especially for those with multiple comorbidities.12,13 Prior studies have raised concern for potential insulin overtreatment in older adults with poor health status and suggest medication deprescribing may be warranted.14-19 However, these cross-sectional or short-term (follow-up, <1 year) analyses do not capture how insulin treatment may change over time.

    Studying the longitudinal relationship between health status and insulin therapy in a real-world context can help inform future interventions, practice guidelines, and policy recommendations to reduce overtreatment of older adults with type 2 diabetes. We investigated the prevalence and predictors of insulin use in a large cohort of 75-year-old patients and assessed insulin discontinuation during the subsequent 4 years. We tested the hypothesis that adults with poor health would be more likely to discontinue insulin over a 4-year follow-up period after age 75 years.

    Methods
    Study Design and Setting

    We conducted a longitudinal cohort study of individuals older than 75 years with type 2 diabetes who were members of Kaiser Permanente Northern California (KPNC), a large integrated health care delivery system that serves 4.2 million members. In this care system, members with type 2 diabetes older than 75 years are primarily managed by their primary care physicians. Individuals included in the study cohort were in the KPNC Diabetes Registry, turned 75 years of age between January 1, 2009, and December 31, 2013, and had no gaps 2 months or longer in health insurance coverage during the period from their 73rd birthday until their 79th birthday or death. The KPNC Diabetes Registry uses a validated algorithm that has been shown to be 99% sensitive and 99% specific in identifying members with diabetes.20 Individuals with type 1 diabetes were excluded based on International Classification of Diseases, Ninth Revision (ICD-9) codes. We evaluated clinical characteristics of each member in the 2 years prior to cohort entry (ie, ages 73 and 74 years) and assessed outcomes for each member up to 4 years (through December 31, 2017) following cohort entry (ie, ages 75 to 79 years) or until death (Figure 1).

    We compared individuals by insulin use at age 75 years (baseline cohort). In the subset of individuals who used insulin at baseline, we compared individuals who continued insulin vs those who discontinued insulin during the 4-year follow-up period (follow-up cohort). The KPNC Institutional Review Board approved the study and granted permission for a waiver of consent for study participants.

    Clinical Measures

    We collected standard demographic, diagnostic, medication, hospital use, and test result data directly from the electronic health record (EHR). Race and ethnicity information were collected by participant self-report and documented in the EHR. Because KPNC is an integrated health care delivery system with a single EHR, virtually all clinically relevant data (>98%) were available within the electronic record. Neighborhood deprivation index was computed using the American Community Survey data.21,22 Baseline HbA1c; body mass index, calculated as weight in kilograms divided by height in meters squared; and estimated glomerular filtration rate (eGFR) values were taken from the most recent results at the time of cohort entry. We categorized weight using body mass index categories based on Centers for Disease Control and Prevention definitions.23 Chronic kidney disease stage was calculated based on the estimated glomerular filtration rate using a validated approach.24,25 For the follow-up period, we report the HbA1c result preceding the last insulin prescription dispensed (ie, the last-measured HbA1c). Because the ADA-recommended individualized HbA1c targets for different populations based on health status do not have lower limits, we specified HbA1c categories of less than 7.0% (53 mmol/mol), 7.0% to 8.4% (53-68 mmol/mol), and 8.5% or greater (69 mmol/mol) to define 3 mutually exclusive ranges. We defined baseline use of noninsulin medications as those medications dispensed from an outpatient pharmacy within 6 months of age 75 years. We gathered 2 functional status measures from the EHR: self-reported prior-week exercise (dichotomized as none vs any and collected at KPNC in a standardized intake form as part of routine clinical care), and prescription for a walker.

    Health Status Category Definition

    We defined health status corresponding to categories proposed by the ADA as a treatment framework,8 which classifies older adults as having poor, intermediate, or relatively good health based on medical comorbidities, functional status measures, and cognitive impairment. A previous study applied this framework to categorize a sample of older adults from the National Health and Nutrition Examination Survey.16 For this investigation, we adapted a modified version of the ADA health status categories to a large, real-world patient population using available data from the EHR (Figure 2).

    We defined baseline health status categories at age 75 years using comorbidities (cardiovascular disease, stroke, diabetic retinopathy, chronic kidney disease, chronic obstructive pulmonary disease, and congestive heart failure), functional status (self-reported prior-week exercise and walker prescription), and indicators of end-stage disease (home oxygen use, metastatic cancer, diagnosis of dementia, or end-stage renal disease). We created 3 mutually exclusive health status groups that corresponded to ADA guidelines for individualizing HbA1c targets in older adults: good health (<2 comorbidities or 2 comorbidities with evidence of physical activity), intermediate health (>2 comorbidities or 2 comorbidities and no self-reported physical activity in the prior week or use of a walker), and poor health (any end-stage disease, regardless of number of comorbidities). In a validation analysis, we found that these categories aligned well with mortality rates (7.4% for good health, 21.4% for intermediate health, and 52.4% for poor health; P < .01) and hospitalization (31.7%, 55.0%, and 69.0%, respectively; P < .01) during the follow-up period.

    Insulin Use and Discontinuation

    We assessed insulin use at the time of cohort entry and for each 6-month period during the 4-year follow-up period. Prevalent insulin use at age 75 years was defined as having insulin dispensed from an outpatient pharmacy in both the first half and the second half of the 74th year of age. Among insulin users, the duration of insulin use at baseline was defined as the time between the first insulin prescription and 75th birthday. For our longitudinal analyses, we investigated insulin discontinuation over 4 years among the subset of 75-year-olds at cohort entry who were prevalent insulin users. Individuals were followed until the end of the cohort period (ie, age 79 years), insulin discontinuation, or death. Insulin discontinuation was defined as no insulin dispensed over 6 months. We chose a 6-month gap to allow for a grace period of roughly 3 months after the 3-month insulin supply ended (>95% of prescriptions were written for 100 days) based on a previously validated approach for detecting medication discontinuation using KPNC pharmacy data.26 There is a closed pharmacy system at KPNC, ensuring that nearly all prescriptions are detected.27

    We defined insulin persistence as no gaps greater than 6 months between insulin dispensings during the 4 years of cohort follow-up. Individuals who died within 6 months of their last insulin prescription were classified as insulin persistent, whereas those who died more than 6 months after their last insulin dispensing were classified as insulin discontinuers. In a sensitivity analysis using a 3-month gap in insulin dispensing instead of 6 months, the relationship between health status and insulin discontinuation was unchanged. Among insulin-persistent participants on both short-acting and long-acting insulin at baseline, we defined insulin regimen simplification as discontinuing short-acting insulin (eg, at least 6 months with no short-acting insulin dispensed) while maintaining the long-acting basal regimen.

    Statistical Methods

    In bivariate analyses, we used χ2 tests, t tests, or nonparametric tests as appropriate. We assessed insulin use by the 3 mutually exclusive health status categories stratified by baseline HbA1c category. To examine the independent association of health status with prevalent insulin use at age 75 years, we created a multivariate log-binomial regression that adjusted for demographic variables (including gender, race, and neighborhood deprivation index) and measures of diabetes status (baseline HbA1c and diabetes duration).

    In the longitudinal analysis in the subset of individuals taking insulin at age 75 years, we examined insulin discontinuation by the 3 mutually exclusive health status categories with a multivariate log-binomial model. We included the same baseline variables used in the prevalent insulin use model, with the exception of updating HbA1c and health status categories to last measured values prior to censoring (ie, insulin discontinuation, death, or end of the follow-up period) and adding history of hypoglycemia defined as an inpatient or emergency department encounter with the primary diagnosis of hypoglycemia. In a sensitivity analysis excluding individuals with stage 4 or higher chronic kidney disease (who may have contraindications to alternative therapies such as metformin), the results did not notably change. All analyses were performed using SAS, 9.3 version (SAS Institute, Inc).

    Results
    Study Cohort

    Our final analytic cohort included 21 531 individuals. Mean (SD) follow-up time was 3.7 (0.9) years. The cohort was demographically diverse, including 10 396 (48.3%) women. Mean (SD) diabetes duration was 9.4 (6.0) years (Table 1). At baseline, 11 041 patients, roughly half of the cohort (51.3%) were classified as having good health, 8632 (40.1%) as having intermediate health, and 1858 (8.6%) as having poor health.

    Insulin Use at Age 75 Years

    Nearly 1 in 5 patients (4076; 18.9%) used insulin in the year prior to turning 75 years old, and the mean (SD) duration of insulin use was 7.9 (5.5) years. Insulin use was associated with ethnicity/race, longer diabetes duration, higher baseline HbA1c, and the presence of diabetes-related comorbidities (Table 1). Diagnosed dementia prevalence was not statistically significantly different between insulin users and nonusers. Insulin use was most common among those with poor health (547; 29.4%) compared with those in the intermediate health (2370; 27.5%) and good health groups (1159;10.5%) (P < .01). Differences in insulin use by health status persisted after stratifying by HbA1c category (Figure 3A), with smaller proportions of patients with good health prescribed insulin at each of the 3 HbA1c strata. In a multivariate model, patients with worse overall health status had increasingly greater likelihood of insulin use: intermediate health group adjusted risk ratio (aRR), 1.85 (95% CI, 1.74-1.97; P < .01) and poor health group aRR, 2.03 (95% CI, 1.87-2.20; P < .01), with good health as the reference (eTable 1 in the Supplement).

    Insulin Use Discontinuation After Age 75 Years

    One-third of patients (1335 of 4076 [32.7%]) using insulin at age 75 years discontinued insulin during the 4-year follow-up period, and insulin regimens were simplified in only 7.9% (321 of 4076) of patients. The mean (SD) time to discontinuation was 1.6 (1.2) years. In contrast with insulin use at age 75 years, prevalent microvascular and macrovascular diabetes-related complications were not associated with likelihood of insulin discontinuation after age 75 years. Insulin discontinuation was significantly more prevalent among patients with a last-measured HbA1c 7.0% or less (Table 2). Depression was associated with increased insulin discontinuation (288 of 1335 [21.6%] vs 514 of 2741 [18.8%]; P = .03).

    Insulin discontinuation during the follow-up period was greatest among patients with good health (306 of 787 [38.9%]), followed by intermediate health (778 of 2380 [32.7%]), and poor health (251 of 909 [27.6%]) (P < .01). By contrast, insulin regimen simplification was most common among those with poor health (99 of 909 [10.9%] vs 185 of 2380 [7.8%] for intermediate health and 37 of 787 [4.7%] for good health; P<.01). In a multivariate model, likelihood of continued insulin use was higher among individuals in poor health (aRR, 1.47; 95% CI, 1.27-1.67; P < .01) and intermediate health (aRR, 1.16; 1.05-1.30; P < .01) compared with good health (reference). These same prevalence and discontinuation patterns were present in the subset with tight glycemic control (HbA1c <7.0%) (Figure 3B). Diabetes duration less than 10 years, HbA1c less than 7.0%, and use of long-acting insulin (reference, combination) were independently associated with insulin discontinuation (eTable 2 in the Supplement).

    Discussion

    Existing guidelines recommend individualizing glycemic targets based on health status but do not make specific recommendations about insulin use. We studied the prevalence of insulin use and discontinuation among a cohort of 75-year-olds with type 2 diabetes to test the hypothesis that older adults with poor health would be less likely to use insulin and be more likely to discontinue insulin over time. We found that nearly 1 in 5 individuals were receiving insulin therapy at age 75 years. Insulin use was most prevalent among those in poor health, whereas subsequent insulin discontinuation after age 75 years was most likely in healthier patients, even after accounting for level of glycemic control.

    The results of this study suggest that neither prevalent insulin use nor subsequent insulin discontinuation among older patients is closely aligned with current recommendations to incorporate health status (in conjunction with life expectancy and patient preferences) when making treatment decisions. These patterns remained evident even when accounting for level of glycemic control. For example, we would expect to find less insulin discontinuation among relatively healthy patients with poor glycemic control (HbA1c ≥8.5%) relative to less healthy patients because these healthier patients are more likely to realize long-term clinical benefit with the tighter control that would be expected from continuing insulin therapy. However, Figure 3B demonstrates that discontinuation follows the opposite pattern: patients with poor health are least likely to discontinue insulin.

    Also shown in Figure 3B is the higher prevalence of insulin discontinuation among patients with good health and tight glycemic control (HbA1c <7.0%) relative to patients in intermediate or poor health. Here, the clinical question relates to insulin discontinuation, sometimes described as deprescribing in the recent literature.28-30 Deprescribing potentially harmful medicines, such as insulin, when the risks outweigh the benefits represents a novel and potentially robust strategy for reducing adverse events and improving quality of individualized care in older patients. The observed pattern of insulin discontinuation in the present study runs contrary to what we would expect to find based on ADA and other guideline recommendations that suggest relaxed glycemic control in adults with poor health status.

    Clinicians have reported barriers to deprescribing related to the lack of evidence to guide decisions, lack of time for informed shared decision-making conversations, and concerns that patients may feel like their care is being diminished.31-33 Recent nationally weighted survey data of Medicare beneficiaries, in contrast, indicate that two-thirds of older patients wanted to reduce the total number of medicines they were taking, and most (92%) would be willing to stop a medicine if recommended by their provider.34 Although this survey was not specific to insulin, the results indicate that the opportunity for safe deprescribing exists. Given the well-documented and severe clinical consequences of iatrogenic hypoglycemia in older patients on insulin, the results of the present study suggest that efforts to define and implement insulin deprescribing guidelines in high-risk patients will likely be applicable to a substantial proportion of older patients with tight glycemic control despite poor health status and limited life expectancy.

    Existing medication deprescribing guidelines provide frameworks for prescribers to contemplate deintensification but do not necessarily provide practical recommendations to implement this process into everyday practice. A recent review of medication deintensification tools noted that only 4 of 15 published guidelines were medication specific, 1 of 15 pertained to antihyperglycemic medicines, and none had high or moderate quality evidence supporting them.35 To date a small, single-arm trial of 65 patients testing the efficacy of an insulin deintensification algorithm has been performed,36 and more trials are needed to provide clinicians with practical tools and protocols to reduce the use of high-risk, low-benefit medications.

    In contrast with insulin discontinuation, we found that insulin regimen simplification (eg, from long-acting and short-acting insulin use to long-acting alone) was relatively uncommon (prevalence, 7.9%) but was more frequent in patients with poor health. This finding underscores the clinical utility of insulin regimen simplification to reduce hypoglycemia risk in certain situations when clinicians or patients are less willing to discontinue insulin. For example, patients with long-standing diabetes may become insulin-dependent owing to progressive beta-cell dysfunction.37 Indeed, we found that characteristics indicative of reduced beta-cell function, such as longer diabetes duration and use of both long-acting and short-acting insulin, predicted insulin persistence. Patients with renal insufficiency may have contraindications to noninsulin medications, and their HbA1c may be artificially low, which makes clinical decisions about insulin use more challenging. In these situations, insulin regimen simplification may allow clinicians to prioritize practical aspects of diabetes management while also reducing risk of iatrogenic hypoglycemia.

    Limitations

    The results of this study must be interpreted in the context of the study design. First, the observational design precludes any definitive inference about causality. However, the cohort had few exclusion criteria and likely represents the general population better than that of a clinical trial. Second, we studied an insured population in an integrated health system, which may limit generalizability, but using the KPNC closed pharmacy system allowed us to capture near-complete insulin prescribing information. Third, without being present with patients, we are unable to evaluate the discussion (or lack thereof) that informed the decision to continue or discontinue insulin therapy in this older, higher-risk patient population. Rather, we provide a population-level perspective of the scope of insulin use in different older patient subgroups. Further work is now needed that can inform system-level efforts to guide safer and more standardized insulin continuation, discontinuation, and simplification frameworks for older patients. Fourth, because we measured insulin dispensing rather than insulin ordering, we were unable to determine whether insulin discontinuation was because of the clinician (ie, stopped prescribing insulin) or patient (ie, stopped picking up prescriptions). This measure has the advantage of capturing true discontinuation but requires further research to better understand the role of clinician vs patient in the discontinuation process. Fifth, despite robust pharmacy data, we were unable to examine insulin dose reductions because doses are not reliably captured in prescription information in the pharmacy data. Finally, the health status classification scheme used EHR data and was susceptible to underrepresentation of medical comorbidities such as dementia, a condition often underdiagnosed.38 Nonetheless, this approach was empirically validated by the strong association of worse health status with death and hospitalizations.

    Conclusions

    As the population with type 2 diabetes continues to age, there is a growing need for evidence-based treatment strategies related specifically to the use of insulin for these older patients. We found that the older adults in poorest health were most likely to use insulin and that subsequent insulin discontinuation was most common among healthier individuals. The substantial and persistent insulin use among older adults with a high risk of hypoglycemia and limited future benefit suggests that more work is needed to develop systems-based approaches that support guideline-concordant insulin use in people older than 75 years.

    Back to top
    Article Information

    Accepted for Publication: July 7, 2019.

    Corresponding Author: Jonathan Z. Weiner, MD, MPH, Division of Research, Kaiser Permanente of Northern California, 2000 Broadway, Oakland, CA 94612 (jonathan.z.weiner@kp.org).

    Published Online: September 23, 2019. doi:10.1001/jamainternmed.2019.3759

    Author Contributions: Dr Weiner had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

    Study concept and design: Weiner, Gopalan, Karter, Grant.

    Acquisition, analysis, or interpretation of data: Weiner, Gopalan, Mishra, Lipska, Huang, Laiteerapong, Grant.

    Drafting of the manuscript: Weiner.

    Critical revision of the manuscript for important intellectual content: All authors.

    Statistical analysis: Weiner, Mishra.

    Obtained funding: Gopalan, Grant.

    Administrative, technical, or material support: Karter.

    Study supervision: Huang, Grant.

    Conflict of Interest Disclosures: Dr Weiner reports receiving support from the Division of Research Delivery Science Fellowship Program. Dr Lipska reports receiving grants from the National Institutes of Health (NIH) and support from the Centers for Medicare and Medicaid Services to develop and evaluate publicly reported quality measures. Dr Huang reports receiving grants from NIH. Dr Laiteerapong reports receiving grants from the NIH National Institute of Diabetes and Digestive and Kidney Diseases and the American Diabetes Association. Dr Karter reports receiving grants from NIH. Dr Grant reports receiving grants from the NIH National Institute of Diabetes and Digestive and Kidney Diseases. No other disclosures were reported.

    Funding/Support: Internal funding was awarded by Kaiser Permanente of Northern California.

    Role of the Funder/Sponsor: The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; or decision to submit the manuscript for publication.

    References
    1.
    Centers for Disease Control and Prevention. United States Diabetes Surveillance System. https://gis.cdc.gov/grasp/diabetes/DiabetesAtlas.html. Accessed August 14, 2018.
    2.
    Gerstein  HC, Miller  ME, Byington  RP,  et al; Action to Control Cardiovascular Risk in Diabetes Study Group.  Effects of intensive glucose lowering in type 2 diabetes.  N Engl J Med. 2008;358(24):2545-2559. doi:10.1056/NEJMoa0802743PubMedGoogle ScholarCrossref
    3.
    Patel  A, MacMahon  S, Chalmers  J,  et al; ADVANCE Collaborative Group.  Intensive blood glucose control and vascular outcomes in patients with type 2 diabetes.  N Engl J Med. 2008;358(24):2560-2572. doi:10.1056/NEJMoa0802987PubMedGoogle ScholarCrossref
    4.
    Lipska  KJ, Krumholz  H, Soones  T, Lee  SJ.  Polypharmacy in the aging patient: a review of glycemic control in older adults with type 2 diabetes.  JAMA. 2016;315(10):1034-1045. doi:10.1001/jama.2016.0299PubMedGoogle ScholarCrossref
    5.
    Huang  ES, Laiteerapong  N, Liu  JY, John  PM, Moffet  HH, Karter  AJ.  Rates of complications and mortality in older patients with diabetes mellitus: the diabetes and aging study.  JAMA Intern Med. 2014;174(2):251-258. doi:10.1001/jamainternmed.2013.12956PubMedGoogle ScholarCrossref
    6.
    ORIGIN Trial Investigators.  Predictors of nonsevere and severe hypoglycemia during glucose-lowering treatment with insulin glargine or standard drugs in the ORIGIN trial.  Diabetes Care. 2015;38(1):22-28. doi:10.2337/dc14-1329PubMedGoogle ScholarCrossref
    7.
    Bruderer  SG, Bodmer  M, Jick  SS, Bader  G, Schlienger  RG, Meier  CR.  Incidence of and risk factors for severe hypoglycaemia in treated type 2 diabetes mellitus patients in the UK: a nested case-control analysis.  Diabetes Obes Metab. 2014;16(9):801-811. doi:10.1111/dom.12282PubMedGoogle ScholarCrossref
    8.
    American Diabetes Association.  11. Older Adults: Standards of Medical Care in Diabetes—2018.  Diabetes Care. 2018;41(suppl 1):S119-S125. doi:10.2337/dc18-S011PubMedGoogle ScholarCrossref
    9.
    Moreno  G, Mangione  CM, Kimbro  L, Vaisberg  E; American Geriatrics Society Expert Panel on Care of Older Adults with Diabetes Mellitus.  Guidelines abstracted from the American Geriatrics Society Guidelines for Improving the Care of Older Adults With Diabetes Mellitus: 2013 update.  J Am Geriatr Soc. 2013;61(11):2020-2026. doi:10.1111/jgs.12514PubMedGoogle ScholarCrossref
    10.
    Department of Veterans Affairs, Department of Defense. VA/DoD Clinical Practice Guideline for the Management of Type 2 Diabetes Mellitus in Primary Care. https://www.healthquality.va.gov/guidelines/CD/diabetes/VADoDDMCPGFinal508.pdf. Published April 2017. Accessed August 14, 2018.
    11.
    Qaseem  A, Wilt  TJ, Kansagara  D, Horwitch  C, Barry  MJ, Forciea  MA; Clinical Guidelines Committee of the American College of Physicians.  Hemoglobin A1c targets for glycemic control with pharmacologic therapy for nonpregnant adults with type 2 diabetes mellitus: a guidance statement update from the American College of Physicians.  Ann Intern Med. 2018;168(8):569-576. doi:10.7326/M17-0939PubMedGoogle ScholarCrossref
    12.
    Lipska  KJ, Yao  X, Herrin  J,  et al.  Trends in drug utilization, glycemic control, and rates of severe hypoglycemia, 2006-2013.  Diabetes Care. 2017;40(4):468-475. doi:10.2337/dc16-0985PubMedGoogle ScholarCrossref
    13.
    Pathak  RD, Schroeder  EB, Seaquist  ER,  et al.  Severe hypoglycemia requiring medical intervention in a large cohort of adults with diabetes receiving care in U.S. integrated health care delivery systems: 2005-2011.  Diabetes Care. 2016;39(3):363-370. doi:10.2337/dc15-0858PubMedGoogle ScholarCrossref
    14.
    Lee  SJ.  So much insulin, so much hypoglycemia.  JAMA Intern Med. 2014;174(5):686-688. doi:10.1001/jamainternmed.2013.13307PubMedGoogle ScholarCrossref
    15.
    Tseng  CL, Soroka  O, Maney  M, Aron  DC, Pogach  LM.  Assessing potential glycemic overtreatment in persons at hypoglycemic risk.  JAMA Intern Med. 2014;174(2):259-268. doi:10.1001/jamainternmed.2013.12963PubMedGoogle ScholarCrossref
    16.
    Lipska  KJ, Ross  JS, Miao  Y, Shah  ND, Lee  SJ, Steinman  MA.  Potential overtreatment of diabetes mellitus in older adults with tight glycemic control.  JAMA Intern Med. 2015;175(3):356-362. doi:10.1001/jamainternmed.2014.7345PubMedGoogle ScholarCrossref
    17.
    Arnold  SV, Lipska  KJ, Wang  J, Seman  L, Mehta  SN, Kosiborod  M.  Use of intensive glycemic management in older adults with diabetes mellitus.  J Am Geriatr Soc. 2018;66(6):1190-1194. doi:10.1111/jgs.15335PubMedGoogle ScholarCrossref
    18.
    Sussman  JB, Kerr  EA, Saini  SD,  et al.  Rates of deintensification of blood pressure and glycemic medication treatment based on levels of control and life expectancy in older patients with diabetes mellitus.  JAMA Intern Med. 2015;175(12):1942-1949. doi:10.1001/jamainternmed.2015.5110PubMedGoogle ScholarCrossref
    19.
    Thorpe  CT, Gellad  WF, Good  CB,  et al.  Tight glycemic control and use of hypoglycemic medications in older veterans with type 2 diabetes and comorbid dementia.  Diabetes Care. 2015;38(4):588-595.PubMedGoogle Scholar
    20.
    Karter  AJ, Schillinger  D, Adams  AS,  et al.  Elevated rates of diabetes in Pacific Islanders and Asian subgroups: the Diabetes Study of Northern California (DISTANCE).  Diabetes Care. 2013;36(3):574-579. doi:10.2337/dc12-0722PubMedGoogle ScholarCrossref
    21.
    Messer  LC, Laraia  BA, Kaufman  JS,  et al.  The development of a standardized neighborhood deprivation index.  J Urban Health. 2006;83(6):1041-1062. doi:10.1007/s11524-006-9094-xPubMedGoogle ScholarCrossref
    22.
    Stoddard  PJ, Laraia  BA, Warton  EM,  et al.  Neighborhood deprivation and change in BMI among adults with type 2 diabetes: the Diabetes Study of Northern California (DISTANCE).  Diabetes Care. 2013;36(5):1200-1208. doi:10.2337/dc11-1866PubMedGoogle ScholarCrossref
    23.
    Centers for Disease Control and Prevention. Defining adult overweight and obesity. https://www.cdc.gov/obesity/adult/defining.html. Updated April 11, 2017. Accessed August 7, 2018.
    24.
    Levey  AS, Stevens  LA, Schmid  CH,  et al.  A new equation to estimate glomerular filtration rate.  Ann Intern Med. 2009;150(9):604-612. doi:10.7326/0003-4819-150-9-200905050-00006PubMedGoogle ScholarCrossref
    25.
    Campbell  KH, Huang  ES, Dale  W,  et al.  Association between estimated GFR, health-related quality of life, and depression among older adults with diabetes: the Diabetes and Aging Study.  Am J Kidney Dis. 2013;62(3):541-548. doi:10.1053/j.ajkd.2013.03.039PubMedGoogle ScholarCrossref
    26.
    Parker  MM, Moffet  HH, Adams  A, Karter  AJ.  An algorithm to identify medication nonpersistence using electronic pharmacy databases.  J Am Med Inform Assoc. 2015;22(5):957-961. doi:10.1093/jamia/ocv054PubMedGoogle ScholarCrossref
    27.
    Karter  AJ, Moffet  HH, Liu  J,  et al.  Glycemic response to newly initiated diabetes therapies.  Am J Manag Care. 2007;13(11):598-606.PubMedGoogle Scholar
    28.
    Reeve  E, Shakib  S, Hendrix  I, Roberts  MS, Wiese  MD.  Review of deprescribing processes and development of an evidence-based, patient-centred deprescribing process.  Br J Clin Pharmacol. 2014;78(4):738-747. doi:10.1111/bcp.12386PubMedGoogle ScholarCrossref
    29.
    Scott  IA, Hilmer  SN, Reeve  E,  et al.  Reducing inappropriate polypharmacy: the process of deprescribing.  JAMA Intern Med. 2015;175(5):827-834. doi:10.1001/jamainternmed.2015.0324PubMedGoogle ScholarCrossref
    30.
    Reeve  E, Gnjidic  D, Long  J, Hilmer  S.  A systematic review of the emerging definition of ‘deprescribing’ with network analysis: implications for future research and clinical practice.  Br J Clin Pharmacol. 2015;80(6):1254-1268. doi:10.1111/bcp.12732PubMedGoogle ScholarCrossref
    31.
    Anderson  K, Stowasser  D, Freeman  C, Scott  I.  Prescriber barriers and enablers to minimising potentially inappropriate medications in adults: a systematic review and thematic synthesis.  BMJ Open. 2014;4(12):e006544. doi:10.1136/bmjopen-2014-006544PubMedGoogle ScholarCrossref
    32.
    Anderson  K, Foster  M, Freeman  C, Luetsch  K, Scott  I.  Negotiating “unmeasurable harm and benefit”: perspectives of general practitioners and consultant pharmacists on deprescribing in the primary care setting.  Qual Health Res. 2017;27(13):1936-1947. doi:10.1177/1049732316687732PubMedGoogle ScholarCrossref
    33.
    Wallis  KA, Andrews  A, Henderson  M.  Swimming against the tide: primary care physicians’ views on deprescribing in everyday practice.  Ann Fam Med. 2017;15(4):341-346. doi:10.1370/afm.2094PubMedGoogle ScholarCrossref
    34.
    Reeve  E, Wolff  JL, Skehan  M, Bayliss  EA, Hilmer  SN, Boyd  CM.  Assessment of attitudes toward deprescribing in older Medicare beneficiaries in the United States.  JAMA Intern Med. 2018;178(12):1673-1680. doi:10.1001/jamainternmed.2018.4720PubMedGoogle ScholarCrossref
    35.
    Thompson  W, Lundby  C, Graabaek  T,  et al.  Tools for deprescribing in frail older persons and those with limited life expectancy: a systematic review.  J Am Geriatr Soc. 2019;67(1):172-180. doi:10.1111/jgs.15616PubMedGoogle ScholarCrossref
    36.
    Munshi  MN, Slyne  C, Segal  AR, Saul  N, Lyons  C, Weinger  K.  Simplification of insulin regimen in older adults and risk of hypoglycemia.  JAMA Intern Med. 2016;176(7):1023-1025. doi:10.1001/jamainternmed.2016.2288PubMedGoogle ScholarCrossref
    37.
    Fonseca  VA.  Defining and characterizing the progression of type 2 diabetes.  Diabetes Care. 2009;32(suppl 2):S151-S156. doi:10.2337/dc09-S301PubMedGoogle ScholarCrossref
    38.
    Lang  L, Clifford  A, Wei  L,  et al.  Prevalence and determinants of undetected dementia in the community: a systematic literature review and a meta-analysis.  BMJ Open. 2017;7(2):e011146. doi:10.1136/bmjopen-2016-011146PubMedGoogle ScholarCrossref
    ×