Predictors of Duloxetine versus Other Treatments among Veterans with Diabetic Peripheral Neuropathic Pain: A Retrospective Study

Authors

  • Yingnan Zhao MS,

    1. School of Public Health and Tropical Medicine, Tulane University, New Orleans, Louisiana
    2. Southeast Louisiana Veterans Health Care System, New Orleans, Louisiana
    Search for more papers by this author
  • Jinan Liu MPH,

    1. School of Public Health and Tropical Medicine, Tulane University, New Orleans, Louisiana
    2. Southeast Louisiana Veterans Health Care System, New Orleans, Louisiana
    Search for more papers by this author
  • Yang Zhao PhD,

    1. Global Health Outcomes, Eli Lilly and Company, Lilly Corporate Center, Indianapolis, Indiana
    Search for more papers by this author
  • Tina Thethi MD,

    1. School of Public Health and Tropical Medicine, Tulane University, New Orleans, Louisiana
    2. School of Medicine, Tulane University, New Orleans, Louisiana, U.S.A.
    Search for more papers by this author
  • Vivian Fonseca MD,

    1. School of Medicine, Tulane University, New Orleans, Louisiana, U.S.A.
    Search for more papers by this author
  • Lizheng Shi PhD

    1. School of Public Health and Tropical Medicine, Tulane University, New Orleans, Louisiana
    2. Southeast Louisiana Veterans Health Care System, New Orleans, Louisiana
    3. School of Medicine, Tulane University, New Orleans, Louisiana, U.S.A.
    Search for more papers by this author

  • Disclosures: Shi, Zhao YN, and Liu are WOC employees/investigators for the Southeast Louisiana Veterans Health Care System. Shi has received research support (to Tulane) from American Diabetes Association, Eli Lilly and Company, Takeda, and Blue Cross Blue Shield of Louisiana. Zhao Y is an employee of Eli Lilly and Company and owns the company stocks. Zhao YN drafted the article. Liu, a certified advanced programmer for SAS®9, primarily performed data analysis. All authors participated in data interpretation. Shi, Liu, Zhao Y, Thethi, and Fonseca provided comments on the article. All authors contributed to the revisions of the article.

Address correspondence and reprint requests to: Lizheng Shi, PhD, Assistant Professor, Department of Health Systems Management, School of Public Health and Tropical Medicine, Tulane University, 1440 Canal Street, New Orleans, LA 70112, U.S.A. E-mail: lshi1@tulane.edu.

Abstract

Objective:  This study used medical and pharmacy records from the Veterans Affairs (VA) health system to explore the predictors of duloxetine versus other treatments for patients with diabetic peripheral neuropathic pain (DPNP).

Methods:  The electronic medical and pharmacy records from January 2004 to December 2008 were requested from the Veterans Integrated Service Network 16 data warehouse. All select patients received either duloxetine or other treatments [tricyclic antidepressants (TCAs), venlafaxine, gabapentin, and pregabalin] over the study period, with the first dispense date of the index agent as the index date. All patients must have 1+ prior DPNP diagnosis (ICD-9-CM: 250.6x or 357.2), but no diagnoses of prior depression (ICD-9-CM: 296.2, 296.3, 300.4, 309.1, or 311.0), fibromyalgia (ICD-9-CM: 729.1), or neuralgia (ICD-9-CM: 729.2). Logistic regression was used to examine the predictors of receiving duloxetine versus other treatments, controlling for demographics, comorbidities, prior pain level, prior use of other medications, and opioid use.

Results:  The analytical sample included 2,694 patients (duloxetine cohort, n = 216; other-treatment cohort, n = 2,478). Prior uses of gabapentin (odds ratio [OR] = 13.66, 95% confidence interval [CI]: 9.70–19.24), TCAs (OR = 5.40, 95% CI: 3.73–7.82), or venlafaxine (OR = 3.67, 95% CI: 1.67–8.06) were strong predictors of duloxetine. Other comorbidities associated with duloxetine were anxiety (OR = 2.08, 95% CI: 1.40–3.08), cerebrovascular disease (OR = 1.44, 95% CI: 1.01–2.07), and substance abuse (OR = 2.11, 95% CI: 1.10–4.03). Prior opioid users were 1.47 (95% CI: 1.02–2.12) times as likely to receive duloxetine as those without prior opioid use. Patients with self-reported severe pain were 1.66 (95% CI: 1.11–2.50) times as likely to receive duloxetine as those with no pain reported.

Conclusion:  DPNP patients in the VA healthcare system with prior other treatment use, select comorbid conditions, prior substance abuse, prior opioid use, and higher pain level were more likely to receive duloxetine.

Ancillary