Antibiotic prescribing in mental health units across the Veterans' Health Administration: How much and how appropriate?

We evaluated antibiotic-prescribing across 111 mental health units in the Veterans' Health Administration. We found that accurate diagnosis of urinary tract infections is a major area for improvement. Because non-mental-health clinicians were involved in most antibiotic-prescribing decisions, stewardship interventions for mental health patients should have a broad target audience to be effective.

Clinical Impact of GLP-1 receptor agonists in veterans receiving basal-bolus insulin.

BACKGROUND: In 2017, an estimated 7.4 million Americans used insulin to treat diabetes. Insulin is proven to lower A1c but can result in hypoglycemia and weight gain. Combining insulin with a glucagon-like peptide-1 receptor agonist (GLP-1-RA) may provide additional blood glucose control while limiting undesirable effects including weight gain. OBJECTIVE: To characterize the clinical impact of adding a GLP-1-RA to a basal-bolus insulin regimen in patients with type 2 diabetes. METHODS: This retrospective observational study used national Veteran's Health Administration data to identify patients with an existing basal-bolus insulin regimen who initiated a GLP-1-RA between January 1, 2005 and December 31, 2017. A1c, insulin total daily dose (TDD), and weight were collected at GLP-1-RA initiation (baseline), 3-, 6-, and 12-month time points and then analyzed using an intent-to-treat approach with the last observation carried forward. Decreases in A1c ≥ 0.5% and weight ≥2 kg were deemed clinically significant. RESULTS: Among 7651 patients initiating GLP-1-RA therapy, mean A1c had a clinically significant decline at 3, 6, and 12 months by -0.5%, -0.7%, and -0.7%, respectively, from a mean baseline of 9%. Patients with lower baseline A1c levels did not have clinically significant changes in A1c, whereas patients with baseline A1c ≥9% had more clinically significant declines. Insulin TDD decreased by -32, -38, and -42 units/day at 3, 6, and 12 months, respectively, where the mean decrease in insulin TDD at 12 months was 79 units/day among patients who discontinued bolus insulin (52.3%) compared with a mean decrease of 2 units/day among those who continued bolus insulin. Mean weight reductions at 3, 6, and 12 months were -1.2, -2.3, and -2.9 kg, respectively, from a mean baseline of 120.6 kg. CONCLUSION: Combining a GLP-1-RA with basal-bolus insulin had a clinically significant improvement on blood glucose control, lowered insulin TDD, and reduced weight. These outcomes were achieved within 3 to 6 months following GLP-1-RA initiation and were maintained through 1 year.

Evaluation of the metformin initiation rate in veterans with newly identified type 2 diabetes.

BACKGROUND: An estimated 35 million individuals in the United States have diabetes. The American Diabetes Association recommends metformin as first-line pharmacologic treatment. The primary objective of this study was to evaluate the metformin initiation rate in veterans with recently identified type 2 diabetes. METHODS: Veterans with new onset type 2 diabetes were identified using National Veterans Health Administration Data. Retrospective information was obtained from those with a first A1C ≥ 6.5% (48 mmol/mol) between 2013 and 2018. Veterans with at least one additional A1C < 6.5% (48 mmol/mol) documented in the three years prior to the A1C diagnostic for diabetes were included in the analysis. RESULTS: A total of 144,180 veterans were included. Of those, 45,776 (31.7%) were started on metformin within one year of diabetes diagnosis. The median time to metformin initiation was 12 days and median time to initiation of any anti-hyperglycemic was 11 days. Approximately 16,000 veterans were referred for lifestyle interventions within 90 days. CONCLUSION: Metformin initiation occurred in fewer patients than expected given metformin is a generic, well-tolerated medication recommended as first-line pharmacologic treatment option regardless of A1C. Further studies are needed to assess the barriers of initiating metformin at time of diabetes diagnosis.

Risk of Acute Kidney Injury and Clostridioides difficile Infection With Piperacillin/Tazobactam, Cefepime, and Meropenem With or Without Vancomycin.

BACKGROUND: Empiric antimicrobial therapy for healthcare-acquired infections often includes vancomycin plus an anti-pseudomonal beta-lactam (AP-BL). These agents vary in risk for adverse events, including acute kidney injury (AKI) and Clostrioides difficile infection (CDI). Studies have only examined these risks separately; thus, our objective was to evaluate AKI and CDI risks simultaneously with AP-BL in the same patient cohort. METHODS: This retrospective cohort study included 789 200 Veterans Health Administration medical admissions from 1 July 2010 through 30 June 2016. The antimicrobials examined were vancomycin, cefepime, piperacillin/tazobactam, and meropenem. Cox proportional hazards regression was used to contrast risks for AKI and CDI across individual target antimicrobials and vancomycin combination therapies, including adjustment for known confounders. RESULTS: With respect to the base rate of AKI among patients who did not receive a target antibiotic (4.6%), the adjusted hazards ratios for piperacillin/tazobactam, cefepime, and meropenem were 1.50 (95% CI: 1.43-1.54), 1.00 (.95-1.05), 0.92 (.83-1.01), respectively. Co-administration of vancomycin increased AKI rates (data not shown). Similarly, against the base rate of CDI (0.7%), these ratios were 1.21 (1.07-1.36), 1.89 (1.62-2.20), and 1.99 (1.55-2.56), respectively. Addition of vancomycin had minimal impact on CDI rates (data not shown). CONCLUSIONS: Piperacillin/tazobactam increased AKI risk, which was exacerbated by concurrent vancomycin. Cefepime and meropenem increased CDI risk relative to piperacillin/tazobactam. Clinicians should consider the risks and benefits of AP-BL when selecting empiric regimens. Further well-designed studies evaluating the global risks of AP-BL and patient specific characteristics that can guide empiric selection are needed.

Contrasting qSOFA and SIRS Criteria for Early Sepsis Identification in a Veteran Population.

The quick Sequential Organ Failure Assessment lacks sensitivity to be an effective replacement for the Systemic Inflammatory Response Syndrome criteria for sepsis screening.

Identifying Opportunities to Improve Guideline-Concordant Antibiotic Prescribing in Veterans with Acute Respiratory Infections or Cystitis.

We investigated the frequency and determinants of guideline-discordant antibiotic prescribing in outpatients with respiratory infections or cystitis. Antibiotic prescribing was guideline discordant in 60% of patients. The most common reason for discordance was prescribing an antibiotic when not indicated. In a multivariate analysis, physicians in training had the highest likelihood of guideline-concordant antibiotic prescribing. Infect Control Hosp Epidemiol 2017;38:724-728.

Effect of selected clinical trial publication on adjunctive nonstatin medication prescribing in the Veterans Health Administration system.

PURPOSE: The question of whether publication of selected clinical trials is temporally followed by changes in prescribing of adjunctive lipid-lowering medications was evaluated. METHODS: In this retrospective preanalysis and postanalysis, Veterans Health Administration (VHA) patients 18 years or older who received a new or renewed order for any lipid-lowering medication between April 2, 2004, and September 2, 2014, were included. This period was chosen based on the publication dates of three trials investigating the efficacy of nonstatin medications: Simvastatin with or without Ezetimibe in Familial Hypercholesterolemia (ENHANCE, April 3, 2008), Effects of Combination Lipid Therapy in Type 2 Diabetes Mellitus (ACCORD Lipid, March 14, 2010), and Niacin in Patients with Low HDL Cholesterol Levels Receiving Intensive Statin Therapy (AIM-HIGH, December 15, 2011). Annual prescribing rates for ezetimibe, fibrates, and niacin were analyzed for 4 years before and after the ENHANCE, ACCORD, and AIM-HIGH trial publication dates, respectively (3 years for niacin in AIM-HIGH) and reported as percent of patients in the cohort. RESULTS: Among patients receiving lipid-lowering medications, relatively low overall prescribing rates were observed for all three target medications. Prescribing rates for each medication decreased after its respective trial publication, with ezetimibe having the greatest change. CONCLUSION: Prescribing of fibrates, niacin, and ezetimibe in the VHA system decreased after the publication of landmark trials assessing their addition to a statin, consistent with the recommendations in the 2013 American College of Cardiology/American Heart Association (ACC/AHA) guideline, which did not encourage routine use of adjunctive therapies to lower the risk of cardiovascular disease.

Adding glucose-lowering agents delays insulin initiation and prolongs hyperglycemia.

OBJECTIVES: Nine classes of glucose-lowering agents (GLAs) are available for patients with type 2 diabetes prior to starting insulin. This study's aim was to determine if the number of GLAs is associated with a difference in glycated hemoglobin (A1C) at insulin initiation in the US Department of Veterans Affairs Health Care System (VAHCS). STUDY DESIGN: Retrospective cohort study. METHODS: A retrospective cohort with national Veterans Health Administration data was created. Veterans with type 2 diabetes and first insulin prescription filled in the VAHCS between January 1, 2009, and August 28, 2013, were identified. Included veterans refilled insulin within the first year, had an A1C > 7% (53 mmol/mol) at least 60 days prior to insulin initiation, and received a GLA within 6 months prior to insulin. Veterans were grouped into 4 cohorts according to the number of GLAs used. RESULTS: A total of 90,497 veterans with type 2 diabetes met inclusion criteria. Insulin was initiated at a mean A1C of 9.9% (85 mmol/mol). The mean A1Cs prior to insulin for 1, 2, 3, or > 3 GLAs were 10.3% (89 mmol/mol), 9.9% (85 mmol/mol), 9.6% (81 mmol/mol), and 9.6% (81 mmol/mol), respectively. Months to insulin increased with the number of GLA trials and prolonged the time veterans were exposed to A1C > 8% (64 mmol/mol). CONCLUSIONS: Multiple glucose-lowering drug classes are associated with a numerical, but not a clinical, difference in A1C at insulin initiation in the closed formulary of the VAHCS.

The impact of posttraumatic stress disorder on blood pressure and heart rate in a veteran population.

Hyperarousal is a hallmark of posttraumatic stress disorder (PTSD). PTSD has been associated with increased blood pressure (BP) and heart rate (HR) in veteran populations. We retrospectively identified male patients consulted to outpatient psychiatry at the Iowa City Veterans Affairs Healthcare System. Patients were divided into PTSD (n = 88) and non-PTSD (n = 98) groups. All PTSD patients and a subset of non-PTSD patients had documented blast exposure during service. The study investigated whether patients with PTSD had higher systolic blood pressure (SBP), diastolic blood pressure (DBP), and heart rate (HR) than patients without PTSD. The effect of trauma exposure on BP was also examined. Mean SBP (133.8 vs. 122.3 mm Hg; p < .001), DBP (87.6 vs. 78.6 mm Hg; p < .001), and HR (78.9 vs. 73.1 bpm; p < .001) were all significantly higher in the PTSD group. Trauma-exposed patients without PTSD had significantly higher BP than nonexposed patients. The prevalence of hypertension (HTN) was 34.1% (diagnosed and undiagnosed) among PTSD patients. Patients with PTSD had higher BP and HR compared to patients without PTSD. Trauma exposure may increase BP in this population. These findings will increase awareness about the cardiovascular implications of PTSD.

Glycemic control and hypoglycemia in Veterans Health Administration patients converted from glyburide to glipizide.

BACKGROUND: In 2009, the Veterans Health Administration (VHA) released a national bulletin regarding the risk of hypoglycemia associated with the use of glyburide in elderly patients with renal dysfunction. Providers were encouraged to avoid glyburide and use glipizide in patients with a calculated creatinine clearance (CrCl) of less than 50 mL per minute. Since this initiative, many veterans were converted by their providers from glyburide to glipizide regardless of renal impairment. OBJECTIVES: To (a) identify whether hemoglobin A1c remained equivalent in patients converted from glyburide to glipizide, (b) evaluate the prevalence of hypoglycemia during treatment with glyburide or glipizide, (c) compare change in glycemic control for renally impaired versus nonimpaired patients, and (d) analyze dosage conversion ratios selected by providers and measures of patient follow-up after conversion including time until A1c measurement and number of glipizide dose titrations. METHODS: This was a single-center, retrospective analysis of veterans converted from glyburide to glipizide from January 1, 2008, through May 31, 2010, who had documented A1c values concurrent with glyburide and glipizide use. A 2-sided equivalence analysis was used for the primary outcome. Equivalence was defined as a change in mean A1c of ± 0.2. Hypoglycemia was defined as blood glucose of less than 70 mg per dL, symptoms of hypoglycemia, or hypoglycemia that led to a fall, loss of consciousness, emergency room visit, hospitalization, or death. The pre- to post-conversion change in rates of hypoglycemia was tested for significance using a McNemar's test. RESULTS: In the 141 (99.3% male, 53.9% CrCl less than 50 mL per minute, mean age = 74.0 years) patients meeting inclusion criteria between 2008-2010, the average change in A1c (+ 0.34) was nonequivalent after conversion from glyburide to glipizide (7.08% vs. 7.42%, respectively). Hypoglycemia occurred more frequently during treatment with glyburide than glipizide (31.2% vs. 12.8%, respectively, P less than 0.001). Mean dose conversion ratios were consistent with VHA recommendations (1 mg per day glyburide = 1.26- 1.55 mg per day glipizide). CONCLUSIONS: Conversion from glyburide to glipizide was associated with an increase in A1c, but the incidence of hypoglycemia was reduced. Results of this study are consistent with the recommendation of the American Diabetes Association and European Association for the Study of Diabetes to use second-generation sulfonylureas other than glyburide. Patients converted to glipizide should be monitored closely to adjust therapy as appropriate to maintain glycemic control.

Prevalence of achievement of A1c, blood pressure, and cholesterol (ABC) goal in veterans with diabetes.

BACKGROUND: The "ABCs of Diabetes" are defined as hemoglobin A1c < 7.0%, blood pressure < 130/80 millimeters mercury (mm Hg), and lowdensity lipoprotein cholesterol (LDL-C) < 100 milligrams per deciliter (mg per dL). Assessments of 3-part goal attainment of A1c, blood pressure, and cholesterol have been reported using data from the National Health and Nutrition Examination Survey (NHANES) for several time periods (e.g., 1988-1994, 1999-2000, 1999-2002, and 2003-2004), Look Action for Health in Diabetes (Look AHEAD, 2001-2004), and community-based endocrinology practice (CBEP, 2000-2004). In 2002, an unpublished analysis of data from 2001-2002 at the Iowa City Veterans Affairs (ICVA) Medical Center found less than 50% of patients met each of the 3 individual goals. In the 5 years following the 2001-2002 assessment, the care for veterans with diabetes at the ICVA was enhanced to include (a) an increased number of diabetes classes and clinics, (b) implementation of the diabetes Care Coordination/Home Telehealth (CCHT) program, and (c) clinical reminders for diabetes performance measures that were added to the electronic medical record (EMR). OBJECTIVES: To (a) describe the prevalence of veterans meeting the ABC goals of diabetes in 1 VA medical center; (b) differentiate the proportion of diabetes patients who met the individual targets for A1c, blood pressure, and LDL-C and compare the results for 2008 through September 2009 with the earlier data from this facility (2001-2002); and (c) examine results reported previously in the literature for NHANES, Look AHEAD, and CBEP data sources. METHODS: Single-center, retrospective analysis of veterans at the ICVA for dates of service from January 1, 2008, through September 30, 2009, who (a) filled at least 1 prescription for an antidiabetic medication and (b) had each of the 3 biomarker values recorded in the EMR for A1c, blood pressure, and LDL-C after the antidiabetic prescription fill date. RESULTS: Of the 5,426 (97.6% male) patients meeting inclusion criteria in 2008-2009, 17.3% (n = 936) achieved the 3-part ABC goal. In this managed care setting, achievement of the 3-part ABC goal surpassed the proportions reported in previous studies in NHANES data (5.2% in 1988-1994, 7.3% in 1999-2000, 7.0% in 1999-2002, 13.2% in 2003-2004), and 10.1% in Look AHEAD 2001-2004, but fell short of the 22.0% reported in CBEP 2000- 2004. When compared with the 2001-2002 results at ICVA, the proportion of patients achieving the individual A1c goal in 2008-2009 increased by 10.8 percentage points (from 43.2% to 54.0%), 12.6 percentage points for blood pressure (from 29.2% to 41.8%), and 17.1 percentage points for LDL-C (from 49.5% to 66.6%, P < 0.001) for the 3 individual comparisons. CONCLUSIONS: The proportion of patients achieving each of the 3 goals for A1c, blood pressure, and LDL-C improved significantly in 2008-2009 compared with the 2001-2002 assessment in this medical center, following implementation of yearly clinical reminders for diabetes care, enhanced patient education, and other program changes that included home-based telephone monitoring with diabetes case management for some patients. Achievement of the 3-part ABC goal in 2008-2009 (17.3%) surpassed 5 assessments reported in the literature but was lower than the CBEP (2000- 2004) performance (22.0%).

Inappropriate prescribing predicts adverse drug events in older adults.

BACKGROUND: Explicit measures of potentially inappropriate prescribing, such as the Beers criteria, have been associated with risk for adverse drug events (ADEs). However, no such link has been established for actual inappropriate prescribing using implicit measures. OBJECTIVE: To determine whether an implicit measure of inappropriate prescribing can predict ADE risk. METHODS: Patients were veterans aged 65 years and older who were seen in primary care clinics and participated in a randomized controlled trial of a pharmacist-physician collaborative intervention. Inappropriate prescribing was determined at baseline, using the 2003 Beers criteria as an explicit measure and the Medication Appropriateness Index (MAI) as an implicit measure. A modified MAI scoring approach was designed to target ADE risk and was used in addition to standard scoring. ADEs that occurred during the 3 months following baseline were assessed by patient interview and plausibility verification by blinded pharmacist review. Logistic regression analysis was used to determine whether inappropriate prescribing predicted risk for an ADE, controlling for potential confounding factors. RESULTS: Of 236 patients, 34 (14.4%) experienced an ADE. Inappropriate prescribing was common at baseline, with 48.7% of patients receiving a Beers criteria drug and 98.7% of patients having an inappropriate rating on at least 1 MAI criterion. Modified MAI scoring, but not other measures of inappropriate prescribing, significantly predicted ADE risk. For every unit increase in modified MAI score (3.1 +/- 3.5; mean +/- SD), the adjusted 3-month odds of an ADE increased 13% (OR 1.13; 95% CI 1.02 to 1.26). For example, patients with a modified MAI score of 3, near the precise mean score of 3.1, were at a nearly 40% greater risk for an ADE compared with patients with a score of zero. CONCLUSIONS: Implicit measurement of actual inappropriate prescribing predicted ADE risk, an important clinical outcome. This finding helps confirm the validity of prior studies that have relied on explicit measures to link potentially inappropriate prescribing to adverse health outcomes.

Preemptive dose reduction of warfarin in patients initiating metronidazole.

BACKGROUND: The goal of this study was to determine if preemptive dose reduction (PDR) of warfarin is effective in maintaining therapeutic anticoagulation in patients initiating metronidazole. METHODS: This is a retrospective, single-center, cohort study in a pharmacist-managed anticoagulation clinic of a university affiliated Veteran's Affairs (VA) Medical Center. Subjects were anticoagulation patients initiating metronidazole between 1 January 2002 and 30 March 2009. At the time of metronidazole initiation, patients were managed with PDR of warfarin or no dose reduction. The primary outcome was the average change in International Normalized Ratio (INR) between patients that received PDR vs. those that did not. RESULTS: In total, 20 patients met inclusion criteria with seven patients receiving PDR at the time of initiation of metronidazole, whereas 13 did not. Patients managed with PDR and those that were not were similar in age (mean±SD 69.4±12.9 years vs. 72.1±9.9 years, p=0.61), mean baseline INR before metronidazole (2.58±0.49 vs. 2.57±0.66, p=0.98), and mean time to follow-up after initiation of metronidazole (5.6±2.9 days vs. 7.0±3.7 days, p=0.40), respectively. The primary outcome was statistically significant with a mean difference in INR of 1.28 (p=0.01) between patients manag-ed with PDR vs. those that were not. The mean preemptive warfarin dose reduction was 34.6%±13.4% which resulted in no significant increase in INR (p=0.61). Secondary outcomes including INR values >4.0 (0% vs. 46%, p=0.05), the average number of warfarin doses omitted (0.43±0.79 vs. 1.15±1.27, p=0.17), use of phytonadione or fresh frozen plasma, and rates of bleeding events were not significantly different be-tween groups. No thromboembolic events occurred during the 30 days following metronidazole therapy. CONCLUSIONS: In patients determined to be appropriate candidates for PDR, a 30%-35% reduction in mean daily warfarin dose was effective in maintaining therapeutic anticoagulation in patients started on concomitant metronidazole.

Celecoxib-induced erythema multiforme with glyburide cross-reactivity.

Erythema multiforme is an acute inflammatory skin reaction that often is caused by drugs, especially sulfonamides and their derivatives. Celecoxib, a cyclooxygenase-2 inhibitor, is a sulfonamide derivative commonly prescribed to treat arthritis in patients who cannot tolerate or who have a contraindication for taking traditional nonsteroidal antiinflammatory agents. A 57-year-old man with a previously undocumented sulfa allergy experienced an allergic skin reaction and had difficulty breathing secondary to throat swelling. His condition was believed to be erythema multiforme associated with the introduction of celecoxib into his drug regimen. His drug therapy was discontinued, but a subsequent reaction occurred when the sulfonamide derivative glyburide was reintroduced. It is important for clinicians to obtain a careful history and perform a thorough medical evaluation in all patients receiving sulfonamides and their derivatives, as a potentially life-threatening allergic reaction may be prevented.