Effect of Splitting Simvastatin Tablets for Controlof Low-Density Lipoprotein Cholesterol

David Parra, PharmD, Nick P. Beckey, PharmD, Harsha S. Raval, PharmD,Kimberly R. Schnacky, PharmD, Vincent Calabrese, PharmD, Roy W. Coakley, MS,

and Robert C. Goodhope, MD, MBA, for the Veterans Integrated Service Network 8Pharmacy Benefits Management Utilization Committee

The efficacy, safety, and economics of a voluntaryconversion from whole simvastatin tablets to splittablets in 6 Veterans Affairs medical centers wereretrospectively evaluated in 3,787 patients who re-ceived a consistent daily dose (5 to 40 mg) of simva-statin in 1999. Baseline and final low-density lipo-protein cholesterol levels and average change frombaseline were not significantly different betweengroups (p >0.05), nor were the incidence of transam-inase increases (p >0.05) or measurements of patientcompliance (p � 0.07). Widespread implementationof this initiative resulted in a cost avoidance of >$1.2million in the 6 medical centers and $10.3 millionacross the Veterans Affairs medical system in 1999,with >$46 million avoided in 2003. �2005 by Ex-cerpta Medica Inc.

(Am J Cardiol 2005;95:1481–1483)

In the United States, because of the number ofavailable statins and their enormous costs, this class

of drugs often undergoes formulary limitations, guide-lines for use, and therapeutic conversions by medica-tion use or pharmacy and therapeutic committeesacross managed care organizations or governmenthealth care systems such as the Veterans Affairs (VA)system. These systems often have a preferred statinwith a substantially discounted contracted price. Con-tract price competitions induce therapeutic conver-sions from 1 agent to another formulary alternative orother efforts to decrease costs associated with theseagents. We analyzed 1 such economic initiative, thetherapeutic substitution of split for whole tablets of thesame dose of simvastatin, and its effects on low-density lipoprotein (LDL) cholesterol, transaminaselevels, and patient compliance.

• • •Formulary management across the VA system is

done both by individual VA medical centers and re-

gionally through Veterans Integrated Service Network(VISN) medication use committees with oversight atthe national level. Through a recommendation by theVISN 8 Pharmacy Benefits Management UtilizationCommittee and approval from the VISN 8 MedicationUse Committee, a voluntary tablet-splitting programfor simvastatin (the preferred statin on the VA formu-lary) was initiated at 6 VA medical centers and theirassociated outpatient clinics that service Florida,Puerto Rico, and southern Georgia. The recommenda-tion was based on the need to address the significantincreasing costs that are associated with statins and thepotential cost avoidance to be had in conjunction withthe wide therapeutic index of statins; in addition,previous statin trials found that using a larger dosetablet given every other day was safe and equallyeffective as the smaller dose given daily.1–3 Patientswere converted from whole to split simvastatin tabletsat the same total daily dose and issued a tablet splitterand instructions about the conversion. Patients whohad visual limitations or other disabilities (i.e., severearthritis of the hands) were exempted from the con-version as were patients whose health care provider orpharmacist deemed them unable to perform the tabletsplitting. Because formulary decisions happen on aregular basis, it is important to evaluate such pro-cesses to ensure that safety and efficacy parameters are notcompromised. This is particularly true with tabletsplitting, because several studies have indicated thattablet splitting may result in high incidences of weightvariation that may or may not be of clinical conse-quence depending on the drug being split.4–11 Subse-quently, we conducted a retrospective analysis to eval-uate any changes in efficacy or safety and to measurethe cost effectiveness of the initiative. The investiga-tors’ institutional review boards and research and de-velopment committees approved the study.

A search of the VISN 8 data mart identified 3,787patients who met the following criteria: (1) remainedon a consistent daily dose of simvastatin betweenJanuary 1, 1999 and December 31, 1999; (2) used adose of simvastatin that could be achieved by splittinga larger dose in half (5, 10, 20, or 40 mg); and (3) had�1 LDL cholesterol result before and after the switchto a split tablet or �2 LDL cholesterol results forpatients who did not split tablets.

Patients were then removed from the analysis if thetime between the first LDL cholesterol measurement,or switch date, and the final LDL cholesterol measure-ment was �45 days. This left 1,331 patients who were

From the Sections of Clinical Pharmacy, Patient Support Service and Cardi-ology, Department of Veterans Affairs Medical Center, West PalmBeach, Florida; the Pharmacy Service Orlando Outpatient Clinic,Department of Veterans Affairs, Orlando, Florida; Pharmacy BenefitsManagement, Department of Veterans Affairs, Hines, Illinois; VISN 8Pharmacy Benefits Management, Department of Veterans Affairs, BayPines, Florida; and the Primary Care Tallahassee Outpatient Clinic,Department of Veterans Affairs, Tallahassee, Florida. Dr. Parra’s addressis: Patient Support Service and Section of Cardiology (119A), 7305 N.Military Trail, West Palm Beach, Florida 33410-6400. E-mail: david.parra@med.va.gov. Manuscript received September 24, 2004; re-vised manuscript received and accepted February 10, 2005.

1481©2005 by Excerpta Medica Inc. All rights reserved. 0002-9149/05/$–see front matterThe American Journal of Cardiology Vol. 95 June 15, 2005 doi:10.1016/j.amjcard.2005.02.019

converted to split tablets and 2,099 patients who werenot. Patients in each group were then stratified by thefollowing baseline LDL cholesterol groups: (1) 75 to100 mg/dl, (2) 101 to 130 mg/dl, (3) 131 to 160 mg/dl,or (4) 160 to 225 mg/dl. Within each baseline LDLcholesterol group, an equal number of patients wasrandomly selected for each simvastatin dose per day(5, 10, 20, or 40 mg). This resulted in 1,098 patients ineach group for analysis. For patients who split tablets,the most recent LDL cholesterol value before theconversion date served as the baseline LDL choles-terol value and the LDL cholesterol value closest toDecember 31, 1999 served as the final LDL choles-terol value. For patients in the unsplit group, the LDLcholesterol value closest to January 1, 1999 served asthe baseline LDL cholesterol value and the LDL cho-lesterol value closest to December 31, 1999 served asthe final LDL cholesterol value.

Primary end points of the study were the average

final LDL cholesterol values and theaverage change from baseline be-tween the split group and the wholetablet group. Secondary end pointsincluded comparison of total yearlysimvastatin costs between groups, in-cidence of transaminase increases�2 to 3 times the upper limit ofnormal, and assessment of compli-ance.

Average baseline LDL choles-terol, average final LDL cholesterol,average change in LDL cholesterol,patient compliance, and cost per yearwere compared with Student’s t test.The chi-square test was used to com-pare incidence of aspartate amino-transferase or alanine aminotransfer-ase levels �100 mg/dl. The sampleof 2,196 patients allowed the detec-tion of what we considered a clini-cally significant difference in LDLcholesterol of �5 mg/dl betweengroups with a power of 80% and an� value of 0.05. If the 95% confi-dence level of the difference betweengroups was �5 mg/dl, then tabletsplitting was judged to be noninfe-rior to the use of whole tablets.

Because equal numbers of pa-tients were randomly selected foreach dose of simvastatin, the averagedaily dose for each group was ex-actly the same (15 � 10 mg, p � 1).Mean baseline LDL cholesterol levelin the split group was 119 � 29mg/dl compared with 120 � 30mg/dl in the whole tablet group (p �0.728). The final LDL cholesterollevel after tablet splitting was 111 �30 mg/dl compared with 112 � 32mg/dl in the group that did not splittablets (p � 0.304); this comparison

is shown in Figure 1. Changes in LDL cholesterollevel from baseline for each group are shown in Figure2, with a mean change of �9 � 32 versus �8 � 33mg/dl for the split group and whole tablet groups,respectively (p � 0.503). The lower and upper bound-aries of the 95% confidence interval for the meandifference in LDL levels between groups (primary endpoint) were �3.5 and 0.8 mg/dl, indicating that tabletsplitting was noninferior to the use of whole tablets.The incidence of alanine aminotransferase increases�2 to 3 times the upper limit of normal was similarbetween groups (0.53% for the split group and 0.61%for the whole tablet group) and was not statisticallydifferent (p � 0.95), nor was the incidence of aspar-tate aminotransferase elevations (0.13% for the splitgroup and 0.28% for the whole tablet group; p �0.32).

Because the most reliable objective measurement ofadherence in large cohorts of patients has been reported

FIGURE 1. Final LDL cholesterol results (milligrams per deciliter).

FIGURE 2. Change in LDL cholesterol from baseline.

1482 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 JUNE 15, 2005

to be a failure to refill a prescription, we estimatedmedication compliance by measuring patient adherencewith refills over the study period.12 The quantity ofmedication dispensed in terms of days supplied and thenumber of actual days supplied (times between filling ofeach prescription) were used to calculate the proportionof days a patient had a statin available.13 The averageproportion of days a patient had a statin available was0.87 for the split group compared with 0.85 for the wholetablet group (p � 0.07).

A cost analysis was conducted using the fiscal year1999 average acquisition cost of a 30-day supply ofsplit simvastatin tablets ($16.07) and unsplit tablets($25.73) derived from the VISN 8 Pharmacy BenefitsManagement Database. The total yearly cost of sim-vastatin (including the cost of tablet splitters, i.e.,$1.05 per splitter) for the 1,098 patients in the splitgroup was $211,738 compared with $339,018 for the1,098 patients in the whole tablet group, a cost avoid-ance of $126,127 (p �0.0001).

Throughout VISN 8 in fiscal year 1999, there were44,522 patients on simvastatin who received 375,17030-day prescription equivalents for simvastatin. Of these30-day prescriptions, 96.4% were for a dose that couldbe achieved by splitting a larger dose in half (5, 10, 20,or 40 mg), representing a potential cost avoidance of$3.45 million. Actual cost avoidance in VISN 8 for fiscalyear 1999 was $1.26 million secondary to incompleteimplementation of the program (only 36.8% of prescrip-tions that could be converted to a larger dose and split inhalf to achieve the same daily dose actually were). How-ever, in fiscal year 2003, after widespread implementa-tion of this cost-avoidance initiative, 98.3% of the1,309,786 30-day prescription equivalents of simvastatinthat could be achieved by splitting a larger dose in halfwere, resulting in a cost avoidance of $5.83 million forthat year. Across the VA system in fiscal year 2003,�$46.5 million was saved by splitting tablets of simva-statin.

• • •Together with 2 previous studies, findings from our

study confirm not only the cost avoidance associatedwith splitting tablets of statins but also the safety andefficacy of such an approach.14,15 Future studies areneeded to evaluate this intervention with other medi-cations that also possess favorable dose-response re-lations and therapeutic windows and to evaluate pa-tient and provider acceptability and patient adherenceto tablet splitting.

1. Rindone JP, Achacoso R, Bledsoe R. Effect of lovastatin administered everyother day on serum low-density lipoprotein cholesterol �160 mg/dl. Am JCardiol 1995;76:312–313.2. Dennis VC, Britton ML, Sirmans SM, Letassy NA, Freeman DA. The use ofalternate-day lovastatin in hypercholesterolemic men. Ann Pharmacother 1997;31:708–712.3. Rindone JP, Hiller D, Arriola G. A comparison of fluvastatin 40 mg every otherday versus 20 mg every day in patients with hypercholesterolemia. Pharmaco-therapy 1998;18:836–839.4. Flynn K, Alligood E, Adams E. Tablet splitting. VA TAP Short Rep2000;3:1–7.5. Rosenberg JM, Nathan JP, Plakogiannis F. Weight variability of pharmacist-dispensed split tablets. J Am Pharm Assoc 2002;42:200–205.6. Teng J, Song CK, Willliams RL, Polli JE. Lack of medication dose uniformityin commonly split tablets. J Am Pharm Assoc 2002;42:195–199.7. Stimpel M, Kuffer B, Groth H, Vetter W. Breaking tablets in half. Lancet1984;1:1299.8. Gupta P, Gupta K. Broken tablets: does the sum of the parts equal the whole?Am J Hosp Pharm 1988;45:1498.9. Sedrati M, Arnaud P, Fontan JE, Brion F. Splitting tablets in half. Am J HospPharm 1994;51:548–552.10. McDevitt JT, Gurst AH, Chen Y. Accuracy of tablet splitting. Pharmaco-therapy 1998;18:193–197.11. Peek BT, Al-Achi A, Coombs SJ. Accuracy of tablet splitting by elderlypatients. Research letter. JAMA 2002;288:451–452.12. Jackevicius CA, Mamdani M, Tu JV. Adherence with statin therapy in elderlypatients with and without acute coronary syndromes. JAMA 2002;288:462–467.13. Benner JS, Glynn RJ, Mogun H, Neumann PJ, Weinstein MC, Avorn J.Long-term persistence in use of statin therapy in elderly patients. JAMA 2002;288:455–461.14. Duncan MC, Castle SS, Streetman DS. Effect of tablet splitting on serumcholesterol concentrations. Ann Pharmacother 2002;36:205–209.15. Gee M, Hasson N. Effects of tablet-splitting program in patients takingHMG-CoA reduction inhibitors: analysis of clinical effects, patient satisfaction,compliance and cost avoidance. J Manag Care Pharm 2002;6:453–458.

BRIEF REPORTS 1483