leader notes pioglitazone and rosglitazone

A Comparison of Lipid and GlycemicEffects of Pioglitazone and Rosiglitazonein Patients With Type 2 Diabetes andDyslipidemiaRONALD B. GOLDBERG, MD1

DAVID M. KENDALL, MD2

MARK A. DEEG, MD, PHD 3

JOHN B. BUSE, MD, PHD4

ANTHONY J. ZAGAR, MS5

JANE A. PINAIRE, PHD5

MENG H. TAN, MD5

MEHMOOD A. KHAN, MD6

ALFONSO T. PEREZ, MD7

SCOTT J. JACOBER, DO5

FOR THE GLAI STUDY INVESTIGATORS

OBJECTIVE Published reports suggest that pioglitazone and rosiglitazone have differenteffects on lipids in patients with type 2 diabetes. However, these previous studies were eitherretrospective chart reviews or clinical trials not rigorously controlled for concomitant glucose-and lipid-lowering therapies. This study examines the lipid and glycemic effects of pioglitazoneand rosiglitazone.

RESEARCH DESIGN AND METHODS We enrolled subjects with a diagnosis oftype 2 diabetes (treated with diet alone or oral monotherapy) and dyslipidemia (not treated withany lipid-lowering agents). After a 4-week placebo washout period, subjects randomly assignedto the pioglitazone arm (n 400) were treated with 30 mg once daily for 12 weeks followed by45 mg once daily for an additional 12 weeks, whereas subjects randomly assigned to rosiglita-zone (n 402) were treated with 4 mg once daily followed by 4 mg twice daily for the sameintervals.

RESULTS Triglyceride levels were reduced by 51.9 7.8 mg/dl with pioglitazone, butwere increased by 13.1 7.8 mg/dl with rosiglitazone (P 0.001 between treatments). Addi-tionally, the increase in HDL cholesterol was greater (5.2 0.5 vs. 2.4 0.5 mg/dl; P 0.001)and the increase in LDL cholesterol was less (12.3 1.6 vs. 21.3 1.6 mg/dl; P 0.001) forpioglitazone compared with rosiglitazone, respectively. LDL particle concentration was reducedwith pioglitazone and increased with rosiglitazone (P 0.001). LDL particle size increased morewith pioglitazone (P 0.005).

CONCLUSIONS Pioglitazone and rosiglitazone have significantly different effects onplasma lipids independent of glycemic control or concomitant lipid-lowering or other antihy-perglycemic therapy. Pioglitazone compared with rosiglitazone is associated with significantimprovements in triglycerides, HDL cholesterol, LDL particle concentration, and LDL particlesize.

Diabetes Care 28:15471554, 2005

Two core metabolic defects contrib-ute to the development of type 2 di-abetes: relative insulin insufficiencyand insulin resistance. Approximately92% of patients with type 2 diabetes dem-onstrate insulin resistance (1). Even in theabsence of overt hyperglycemia, insulinresistance is associated with a cluster ofabnormalities that increase the risk forcardiovascular disease (CVD), includingdyslipidemia, increased expression of in-flammatory markers, activation of proco-agulants, hemodynamic changes, andendothelial dysfunction (2,3).

The dyslipidemia associated with in-sulin resistance and type 2 diabetes ischaracterized by elevated triglyceridesand decreased HDL cholesterol (46).Although LDL cholesterol may not be el-evated in type 2 diabetes, an increase inthe proportion of small, dense, and po-tentially more atherogenic LDL choles-terol particles is observed (7). In additionto LDL cholesterol, elevated triglyceridelevels and reduced HDL cholesterol levelsare both risk factors for coronary heartdisease (CHD) (811). Compared withnondiabetic individuals, patients withtype 2 diabetes have a two- to fourfoldhigher risk of CVD, and dyslipidemia isan important contributor to the increasedrisk in this population (12).

By targeting insulin resistance, themembers of the thiazolidinedione class oforal antihyperglycemic medications pos-sess both a glucose-lowering effect andthe potential to alter lipid/lipoprotein me-tabolism. Two members of the thia-zolidinedione class are currently availablefor the treatment of type 2 diabetes: pio-glitazone hydrochloride (Actos; TakedaPharmaceuticals North America, Lincoln-shire, IL) and rosiglitazone maleate(Avandia; GlaxoSmithKline, ResearchTriangle Park, NC).

This study was conceived followingthe report of a nonrandomized clinicalcomparison of potential differences inlipid effects among thiazolidinediones

From the 1Division of Endocrinology, Metabolism, and Diabetes, University of Miami School of Medicine,Miami, Florida; the 2International Diabetes Center, Park Nicollet Institute, Minneapolis, Minnesota; the3Division of Endocrinology and Metabolism, Department of Veterans Affairs and the Indiana UniversitySchool of Medicine, Indianapolis, Indiana; the 4Divisions of Endocrinology and of General Medicine, Uni-versity of North Carolina School of Medicine, Chapel Hill, North Carolina; the 5Lilly Research Laboratories,Eli Lilly, Indianapolis, Indiana; 6Takeda Pharmaceuticals North America, Lincolnshire, Illinois; and the7Takeda Global Research and Development Center, Lincolnshire, Illinois.

Address correspondence and reprint requests to Scott J. Jacober, DO, Lilly Research Laboratories, ADivision of Eli Lilly, Lilly Corporate Center, DC 5116, Indianapolis, IN 46285. E-mail: [email protected].

Received for publication 10 February 2005 and accepted in revised form 31 March 2005.Abbreviations: CHD, coronary heart disease; CVD, cardiovascular disease; LOCF, last observation car-

ried forward; PAI-1, plasminogen activator inhibitor-1.A table elsewhere in this issue shows conventional and Syste`me International (SI) units and conversion

factors for many substances. 2005 by the American Diabetes Association.The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby

marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

C l i n i c a l C a r e / E d u c a t i o n / N u t r i t i o nO R I G I N A L A R T I C L E

DIABETES CARE, VOLUME 28, NUMBER 7, JULY 2005 1547

(13). Since that time, multiple reports(1422) have been published suggestingthat pioglitazone has differential effectson lipid parameters in patients with type2 diabetes when compared with rosiglita-zone. However, these previous studieswere either retrospective chart reviewstudies, clinical trials not rigorously con-trolled for concomitant glucose-loweringand lipid-lowering therapies, or system-atic reviews. With the primary objectiveto test the hypothesis that pioglitazonehas greater triglyceride-lowering effectsthan rosiglitazone, this study reports re-sults from the first multicenter, prospec-tive, randomized, double-blind, parallel-group comparison of maximally effectivemonotherapy doses of pioglitazone androsiglitazone in patients with type 2 dia-betes and dyslipidemia receiving no con-comitant glucose-lowering or lipid-lowering therapies.

RESEARCH DESIGN ANDMETHODS Subjects eligible forparticipation in this clinical trial were

men or women 35 years of age with adiagnosis of type 2 diabetes (based onWorld Health Organization criteria) withfasting triglyceride levels150 mg/dl and600 mg/dl and fasting LDL cholesterollevels130 mg/dl. Other inclusion crite-ria included fasting serum C-peptide lev-els1 ng/ml and HbA1c (A1C) values7and 11% if nave to previous oral anti-hyperglycemic therapy or A1C values7and9.5% if previously treated with oralantihyperglycemic monotherapy.

Subjects were excluded from partici-pation in this study for any of the follow-ing: treatment within 60 days of screeningwith insulin, systemic glucocorticoidtherapy, combination oral antihypergly-cemic therapy, any lipid-lowering agent,or any weight loss agent; known allergy toany thiazolidinedione; serum creatinine176.8 mol/dl (2.0 mg/dl) or 2dipstick proteinuria at screening; alanineaminotransferase or aspartate amino-transferase 1.5 times the upper limit ofnormal or significant clinical liver disease;hemoglobin 10.5 g/dl (females) or

11.5 g/dl (males) at screening; abnor-mal thyrotropin; functional New YorkHeart Association Cardiac Disease ClassIII or IV, history of CVD, or heart surgerywithin 6 months of screening; receivingrenal dialysis or having renal transplant;current therapy for malignancy otherthan basal cell or squamous cell skin can-cer; known history of HIV infection; signsor symptoms of drug or alcohol abuse;and any condition or situation precludingadherence to and completion of the pro-tocol. For female subjects, appropriatebirth control was required, and preg-nancy, breast-feeding, or the intent to be-come pregnant during the study periodprohibited participation.

Subjects were enrolled from the U.S.(78 sites), Puerto Rico (11 sites), Mexico(4 sites), and Colombia (7 sites). Con-ducted in accordance with the Declara-tion of Helsinki guidelines on goodclinical practice, this study was approvedby each investigators institutional ethicalreview board.

Screening for eligibility occurred atvisit 1 after written informed consent wasobtained. At visit 2, subjects were ran-domly assigned to one of the two treat-ment groups, although active study drugadministration was not initiated until 4weeks later (visit 3). Randomization oc-curred in a stratified fashion with fourstrata corresponding to previous oral an-tihyperglycemic treatment (previouslytreated or nave) and sex (male or female).Subjects discontinued any current oralantihyperglycemic therapy and receivedoral placebo therapy throughout the4-week, single-blind, lead-in period. Atvisit 3, subjects received either 30 mg pio-glitazone once daily or 4 mg rosiglitazoneonce daily for 12 weeks according to therandomization assigned at visit 2. Quali-fied personnel provided dietary counsel-ing on the American Heart Associationweight-maintaining Step I diet, and allsubjects were instructed to follow this dietthroughout the entire study. Clinic visitsoccurred every 4 weeks following visit 3through visit 6. At visit 6 and for the final12 weeks, the doses of pioglitazone androsiglitazone were increased to the maxi-mally effective doses (for monotherapy)of 45 mg once daily (23) or 4 mg twicedaily (24), respectively. Clinic visits oc-curred every 6 weeks (visits 7 and 8) forthe remainder of the 24-week total study.

Figure 1Patient flow through the study.

Effects of pioglitazone and rosiglitazone

1548 DIABETES CARE, VOLUME 28, NUMBER 7, JULY 2005

Analytical methodsThe following analyses were performedby Covance Central Laboratory Services(Indianapolis, IN): triglycerides, totalcholesterol, and plasma glucose in bloodsamples (following at least 10 h of fasting)using standard enzymatic methods; HDLand LDL cholesterol (Roche Diagnostics,Indianapolis, IN) by direct methods; free

fatty acid by the Wako enzymatic method(Wako Chemicals, Richmond, VA); apo-lipoprotein B by immunoassay (BeckmanIMMAGE Immunochemistry System,Beckman Instruments, Brea, CA); A1C bychromatography (Bio-Rad, Hercules,CA); total insulin by immunoassay (Ab-bott IMX Microparticle EIA, Abbott Lab-oratories, Abbott Park, IL); C-peptide by

radioimmunoassay (Adaltis Italia, Rome,Italy); highly sensitive C-reactive proteinby immunonephelometry (Dade Behring,Newark, DE); and plasminogen activatorinhibitor-1 (PAI-1) by immunoassay (As-serachrom PAI-1 Antigen EIA, Diagnos-tica Stago, France). LDL particle size andconcentration were measured using pro-ton nuclear magnetic resonance spectros-copy at LipoScience (Raleigh, NC).Surrogates of insulin resistance and -cellfunction were estimated by homeostasismodel assessments (25). Safety assess-ments included adverse events, bloodpressure and heart rate, hemoglobin andhematocrit, liver function, pedal edema,body weight, and hypoglycemic episodes.

Statistical methodsData are presented as means SE (or SDwhere indicated). Differences betweentreatments in demographics and baselinelevels (visit 3) for patients entering activedrug therapy were evaluated using a 2

test for categorical variables or an inde-pendent-groups t test for continuous vari-ables. Efficacy analyses were conductedon subjects providing a baseline measure-ment and at least one postbaseline mea-surement. The last-observation-carried-forward (LOCF) change from baselinelevel and LOCF actual value were ana-lyzed using a fixed-effects ANCOVA. TheANCOVA model was composed of termsfor strata, geographic region in which theinvestigative site was located (five regions:Mexico/Puerto-Rico/Colombia, Mid-Atlantic/Eastern, West/Midwest/Texas,South/Southeast, and West Coast/Hawaii), treatment, and baseline value.The change from baseline to the last ob-served value was of primary interest, andthe triglycerides change was the primaryefficacy variable. The visit-wise changesfrom baseline were also analyzed usingLOCF. LOCF percent change from base-line was also analyzed for the lipid vari-ables. Treatments were compared usingleast-square means (2627).

A mixed-model repeated-measuresanalysis was used to confirm the LOCFtriglyceride results. The model used wascomprised of terms for strata, geographicregion, treatment, visit, treatment visitinteraction, baseline value, and visit baseline level interaction. The covariancestructure was modeled using an unstruc-tured covariance matrix within each treat-ment. The Kenward-Roger degrees offreedom were used for the tests. Addition-

Table 1Characteristics and demographics at visit 3 of all randomly assigned subjects be-ginning active therapy

Pioglitazone Rosiglitazone P value

n 369 366Sex 0.824

Male 199 (53.9) 201 (54.9)Female 170 (46.1) 165 (45.1)

Age (years) 55.9 10.5 56.3 11.3 0.572Race 0.842

White 239 (64.8) 219 (59.8)Hispanic 105 (28.5) 118 (32.2)Asian 10 (2.7) 12 (3.3)African 9 (2.4) 10 (2.7)Other 6 (1.6) 7 (1.9)

Duration of diabetes (years) 3.9 4.4 4.0 4.6 0.847Body weight (kg) 93.7 20.6 92.5 21.0 0.450BMI (kg/m2) 33.7 12.9 32.6 6.6 0.122A1C (%) 7.6 1.2 7.5 1.2 0.230Fasting plasma glucose (mg/dl) 180.5 59.7 177.3 57.4 0.459Previous treatment* 280 (75.9) 274 (74.9) 0.797

Metformin 97/212 (45.8) 92/206 (44.7)Insulin secretagogues 97/212 (45.8) 95/206 (46.1)Thiazolidinediones 18/212 (8.5) 19/206 (9.2)

Fasting lipid profileTotal triglycerides (mg/dl) 258.5 159.4 239.5 132.6 0.079Total cholesterol (mg/dl) 193.4 31.3 193.7 33.7 0.882LDL cholesterol (mg/dl) 106.9 25.4 109.0 25.9 0.260HDL cholesterol (mg/dl) 38.8 10.0 39.7 10.3 0.211

Apolipoprotein B (g/l) 1.05 0.20 1.04 0.20 0.781Fasting free fatty acid (mEq/l) 0.64 0.28 0.62 0.29 0.579Fasting insulin (U/ml) 19.7 19.4 17.8 14.3 0.132Fasting C-peptide (ng/ml) 3.9 1.7 3.6 1.6 0.043HOMA

Insulin resistance 8.2 6.4 7.7 7.2 0.386-Cell function 83.6 125.2 71.1 69.7 0.095

Systolic blood pressure (mmHg) 128.4 15.9 129.1 16.4 0.584Diastolic blood pressure (mmHg) 78.7 9.3 78.5 8.9 0.765Preexisting CVD or previous MI 31 (8.4) 24 (6.6) 0.401Aspartate aminotransferase (units/l) 22.0 7.1 21.7 7.5 0.547Alanine aminotransferase (units/l) 27.0 10.7 25.6 10.9 0.063Creatine phosphokinase (units/l) 108.0 85.1 109.6 77.7 0.790Plasminogen activator inhibitor 1 (ng/ml) 64.2 43.4 59.6 40.4 0.143C-reactive protein (mg/l) 7.0 10.1 6.6 7.7 0.580

Data are n (%) or means SD. *Totals of 280 (pioglitazone) and 274 (rosiglitazone) patients reportedprevious antidiabetes therapy, but only 212 (pioglitazone) and 206 (rosiglitazone) patients reported the classof previous antidiabetes therapy. Includes short-acting secretagogues repaglinide and nateglinide. HOMA,homeostasis model assessment; MI, myocardial infarction.

Goldberg and Associates


ally, the change from baseline triglycer-ides was analyzed for the subset ofpatients who completed the study. SASversion 8.2 (SAS Institute, Cary, NC) wasused for all analyses. All tests were twosided, and results were considered statis-tically significant at P 0.05.

RESULTS Figure 1 summarizes pa-tient flow through the study. Similar num-bers of subjects completed therapy for thepioglitazone and rosiglitazone groups. Thedistribution of subjects among the variouswithdrawal categories was also similar be-tween treatment groups.

There were no statistically significantdifferences between the treatment groupsin respect to demographics, baselinecharacteristics, and laboratory measure-ments with the exception of fasting C-peptide levels, which were lower insubjects randomly assigned to rosiglita-zone (Table 1).

Figure 2 shows the 24-week timecourse for fasting triglycerides, HDL cho-lesterol, non-HDL cholesterol, LDL cho-lesterol, and A1C. Baseline and end pointvalues for each are found in Table 2.

By the first posttherapy visit, triglyc-eride levels were significantly decreasedwith pioglitazone and significantly in-creased with rosiglitazone compared withbaseline. Differences in triglyceride levelsbetween treatment groups were signifi-cant at every time point (P 0.001) (Fig.2A). By week 24, triglyceride levels in pio-glitazone-treated subjects were signifi-cantly reduced by 51.9 7.8 mg/dl(12.0 3.0%, median of 19.8%) belowbaseline (Table 2), whereas triglyceridelevels in rosiglitazone-treated subjectswere elevated by 13.1 7.8 mg/dl(14.9 3.1%, median of 3.5%) abovebaseline (the percent change representsthe mean of the individual percentchanges from baseline). The 95% CI(based on the ANCOVA model) for themean change from baseline to the last ob-served value was 67.2 to 36.6 mg/dlin the pioglitazone group and 2.2 to28.5 mg/dl in the rosiglitazone group.The mixed-model repeated-measures re-sults for change from baseline triglyceridelevel (data not shown) as well as the re-sults from an analysis of study completers(data not shown) were similar to theLOCF results.

Both pioglitazone and rosiglitazoneincreased HDL cholesterol over time, butmean changes from baseline to end pointwere significantly greater with pioglita-zone compared with rosiglitazone, re-spectively: 5.2 0.5 mg/dl (14.9 1.2%) versus 2.4 0.5 mg/dl (7.8 1.2%) (P 0.001) (Table 2). The differ-ences in HDL cholesterol levels betweentreatment groups were significant at everytime point (P 0.001) (Fig. 2B).

Non-HDL cholesterol levels re-mained relatively constant in pioglita-zone-treated subjects but were markedlyincreased with rosiglitazone therapy overthe treatment period, such that the differ-ences between treatment groups were sig-nificant at every time point (P 0.001)

Figure 2Comparison of fasting triglyceride (A), HDL cholesterol (B), non-HDL cholesterol(C), LDL cholesterol (D), and mean A1C (E) levels observed during 24 weeks of therapy withpioglitazone (PIO) and rosiglitazone (ROSI). LDL cholesterol levels were directly measured (notcalculated). Patients were randomly assigned to either 12 weeks of 30 mg/day pioglitazone fol-lowed by 12 weeks of 45 mg/day pioglitazone () or 12 weeks of 4 mg/day rosiglitazone oncedaily followed by 12 weeks of 4 mg rosiglitazone twice daily (f). Vertical bars represent SE. *P0.001 between treatment groups;P 0.05 between treatment groups.



(Fig. 2C). By week 24, non-HDL choles-terol levels in pioglitazone-treated sub-jects were 3.6 1.9 mg/dl (3.8 1.3%)above baseline, whereas levels in rosigli-tazone-treated subjects increased 25.7 2.0 mg/dl (18.6 1.3%) above baseline(P 0.001 between treatments) (Table2).

Both pioglitazone and rosiglitazonealso increased LDL cholesterol over time,but mean changes from baseline to endpoint were significantly less with pioglita-zone compared with rosiglitazone, re-spectively: 12.3 1.6 mg/dl (15.7 1.9%) versus 21.3 1.6 mg/dl (23.3 1.9%) (P 0.001) (Table 2). The differ-ences in LDL cholesterol levels betweentreatment groups were significant at everytime point (P 0.001) except at weeks 4and 12 (Fig. 2D). Lastly, apolipoprotein Bwas unchanged in the pioglitazone groupbut significantly increased in the rosigli-tazone group (Table 2).

Figure 3 compares the effects of pio-glitazone and rosiglitazone on LDL parti-cle concentration and particle size at endpoint. These agents had opposing effectson particle concentration: a significant re-duction was observed with pioglitazone,whereas rosiglitazone therapy resulted ina significant increase. Both agents in-creased particle size, but the increaseobserved with pioglitazone was signifi-cantly greater than that observed withrosiglitazone.

Both agents significantly improvedglycemic control. Although statistical dif-ferences between treatment groups wereobserved for A1C values between weeks 4and 12, these differences are not clinicallysignificant (Fig. 2E). Furthermore, therewas no difference between agents with re-spect to A1C or fasting plasma glucosechanges at end point (Table 2). No signif-icant differences were observed betweenagents for changes in free fatty acid levels,PAI-1, C-reactive protein, and indices ofinsulin secretion and sensitivity (Table 2).

Mean body weight changes frombaseline were similar for both pioglita-zone (2.0 0.2 kg) and rosiglitazone(1.6 0.2 kg) (P 0.164). Additionally,no differences between agents were ob-served with regard to liver function tests(alanine aminotransferase and aspartateaminotransferase), creatine phosphoki-nase, blood pressure and heart rate, he-moglobin and hematocrit, hypoglycemicepisodes, or adverse events includingedema and congestive heart failure.

Table 2Effects of pioglitazone and rosiglitazone on outcome measures for subjects with atleast one postbaseline measurement

Pioglitazone Rosiglitazone P value

n 363 356Triglyceride (mg/dl)

Baseline 257.8 8.2 235.3 6.6Change from baseline 51.9 7.8* 13.1 7.8 0.001Percent change from baseline 12.0 3.0* 14.9 3.1* 0.001

HDL cholesterol (mg/dl)Baseline 38.8 0.5 39.8 0.6Change from baseline 5.2 0.5* 2.4 0.5* 0.001Percent change from Baseline 14.9 1.2* 7.8 1.2* 0.001

Non-HDL cholesterol (mg/dl)Baseline 154.8 1.6 153.6 1.6Change from baseline 3.6 1.9 25.7 2.0* 0.001Percent change from baseline 3.8 1.3* 18.6 1.3* 0.001

LDL cholesterol (mg/dl)Baseline 107.1 1.3 109.1 1.4Change from baseline 12.3 1.6* 21.3 1.6* 0.001Percent change from baseline 15.7 1.9* 23.3 1.9* 0.002

Total cholesterol (mg/dl)Baseline 193.6 1.6 193.4 1.8Change from baseline 8.8 1.9* 28.2 1.9* 0.001Percent change from baseline 5.7 1.0* 15.9 1.0* 0.001

Total-to-HDL cholesterol ratioBaseline 5.3 0.1 5.1 0.1Change from baseline 0.3 0.1* 0.7 0.1* 0.001

Apolipoprotein B (g/l)Baseline 1.05 0.01 1.04 0.01Change from baseline 0.00 0.01 0.11 0.01* 0.001

Free fatty acid (mEq/l)Baseline 0.64 0.01 0.62 0.02Change from baseline 0.11 0.02* 0.12 0.02* 0.681

A1C (%)Baseline 7.6 0.1 7.5 0.1Change from baseline 0.7 0.1* 0.6 0.1* 0.129

Fasting plasma glucose (mg/dl)Baseline 180.6 3.1 176.5 3.0Change from baseline 33.2 2.2* 36.6 2.2* 0.233

Fasting insulin (U/ml)Baseline 19.7 1.0 17.9 0.8Change from baseline 4.5 0.5* 4.6 0.5* 0.918

Fasting C-peptide (ng/ml)Baseline 3.9 0.1 3.7 0.1Change from baseline 0.7 0.1* 0.7 0.1* 0.652

HOMA of insulin resistanceBaseline 8.2 0.3 7.8 0.4Change from baseline 2.8 0.2* 3.0 0.2* 0.449

HOMA -cellsBaseline 83.8 6.6 71.8 3.7Change from baseline 8.0 3.5* 6.7 3.6 0.780

PAI-1 (ng/ml)Baseline 62.8 2.3 60.0 2.3Change from baseline 10.4 2.0* 11.7 2.0* 0.623

C-reactive protein (mg/l)Baseline 7.0 0.6 6.6 0.4Change from baseline 2.0 0.3* 2.5 0.3* 0.288

Data are means SE. *P 0.05 vs. baseline; change from baseline and percent change from baseline areleast-square means adjusted for baseline level. LDL cholesterol levels were directly measured (not calcu-lated). HOMA, homeostasis model assessment.



CONCLUSIONS This prospec-tive, randomized, multicenter, double-blind clinical trial demonstrates thatpioglitazone and rosiglitazone exert dif-ferent effects on plasma lipids. Pioglita-zone is associated with significantimprovements versus rosiglitazone in tri-glyceride, HDL cholesterol, non-HDLcholesterol, and LDL particle concentra-tions and LDL particle size, despite simi-lar effects on glycemic control andsurrogate measures of insulin resistance.

The different lipid responses to max-imal monotherapy doses of pioglitazoneand rosiglitazone observed in this studyare consistent with results from prior, lesswell-controlled comparison studies andthe large, randomized, multicenter place-bo-controlled trials for pioglitazone(23,28,29). Additionally, a meta-analysisof these studies (18) demonstrates results

very similar to those of our study. Themechanism(s) by which these agents ex-ert differential effects on the lipid profileare not clearly understood, and studiesare underway to elucidate these mecha-nisms.

The effects of lipids on cardiovasculardisease are well known. At increased con-centrations, LDL cholesterol, total choles-terol, and triglycerides, and HDLcholesterol at decreased concentrations,are known to be risk factors for CVD inthe general population (8 11) and insubjects with type 2 diabetes (34). Al-though lowering LDL cholesterol is theprimary target according to both the Na-tional Cholesterol Education ProgramAdult Treatment Panel III (35,36) andAmerican Diabetes Association guidelines(12), raising HDL cholesterol is a second-ary target, the benefit of which was dem-

onstrated by the Veterans Affairs High-Dens i ty Lipoprote in Choles tero lIntervention Trial (VA-HIT) study (37),in which increasing HDL cholesterol andlowering triglycerides with the fibrategemfibrozil decreased cardiovascularevents by 24%. In subjects with hypertri-glyceridemia (200 mg/dl), loweringnon-HDL cholesterol to levels 130mg/dl is recommended for high-risk sub-jects by the National Cholesterol Educa-tion Program Adult Treatment Panel III(36). In our study, rosiglitazone raisednon-HDL cholesterol levels and pioglita-zone did not.

The dyslipidemia of diabetes is usu-ally characterized by a combination of in-creased triglyceride and decreased HDLcholesterol levels and, most often, near-normal LDL cholesterol concentrations(38). However, insulin resistance with orwithout hyperglycemia is associated withqualitative changes in the composition ofLDL particles shown to be associated withgreater risk for atherosclerosis and cardio-vascular disease (39). These changes inthe LDL particles include a decrease inparticle size and a greater density of eachparticle concomitant with a relative de-crease in the cholesterol content of eachparticle (40). It is generally accepted thatthe increase in triglyceride levels in type 2diabetes is in part responsible for theseatherogenic changes in the LDL profile(5,7,11,41,42).

In the current study, pioglitazone androsiglitazone differed significantly withopposing effects on triglycerides. Al-though both agents increased HDL cho-lesterol, pioglitazone increased HDLcholesterol more. Furthermore, pioglita-zone caused a shift from small, dense LDLparticles to larger, more buoyant LDL par-ticles. This change in the size of the par-ticles was accompanied by a decrease inparticle concentration (particle number),an effect not observed with rosiglitazone.Rosiglitazone was associated with an in-crease in LDL particle size as well as anincrease in triglyceride levels and LDLparticle numbers. These observed differ-ences in the lipid effects between piogli-tazone and rosiglitazone raise severalmechanistic questions. First, the shift to-ward larger LDL particles observed withrosiglitazone therapy cannot be due to ef-fects on triglyceride levels. Secondly, theincrease in LDL by pioglitazone can onlybe explained by increases in particle size,as the measured number of LDL particles

Figure 3Comparison of mean LDL particle concentration (A) and particle size (B) at baselineand end point (24 weeks) for patients treated with pioglitazone (PIO) and rosiglitazone (ROSI).Vertical bars represent SE.



was slightly reduced. The greater increasein LDL observed with rosiglitazone is theresult of both increases in particle size andparticle number. These observations sug-gest that previous speculation (43,44)that differences in lipid effects betweenpioglitazone and rosiglitazone might bedue to differential effects on particle sizeare no longer tenable.

Finally, in light of the recent guide-lines, subjects with diabetes often are pre-scribed lipid-lowering medication, mostfrequently a statin. However, subjects inthis study were not treated with suchtherapy. This study was purposely de-signed to eliminate the confounding in-fluence of concomitant glucose-loweringand lipid-lowering medications to allowfor a clear assessment of the glycemic andlipid effects of the two thiazolidinediones.Simvastatin, when added to either piogli-tazone or rosiglitazone, produced similarmean lipid changes from baseline withboth these thiazolidinediones, suggestingthat differences between these agentswould be preserved with concomitantstatin therapy (45). A current, ongoingstudy named COMPLEMENT has beendesigned to assess these parameters insubjects with diabetes who are simulta-neously taking a statin (46).

In summary, the current study dem-onstrates that pioglitazone and rosiglita-zone differ in their effects on triglyceride,HDL cholesterol, non-HDL cholesterol,and LDL cholesterol particle concentra-tions and particle size. These differenceswere observed despite the finding thatthese agents produced similar improve-ments in many of the nonlipid CHD/CVDrisk factors associated with insulin resis-tance and type 2 diabetes (A1C, fastingplasma glucose, fasting insulin levels, ho-meostasis model assessment of insulin re-sistance, PAI-1, and C-reactive protein).Whether these differences in lipid effectstranslate into differences for the risk ofCVD is not clear. Although no trials di-rectly comparing the effects of pioglita-zone and rosiglitazone on CVD outcomesare underway, multiple ongoing trials areevaluating CVD event reduction with ei-ther pioglitazone or rosiglitazone (Pro-spective Pioglitazone Clinical Trial inMacrovascular Events [PROactive] [47],Rosiglitazone Evaluated for Cardiac Out-comes and Regulation of glycemia in Di-abetes [RECORD] [48], and the Rationalefor the Bypass Angioplasty Revasculariza-tion Investigation 2 Diabetes [BARI 2D]

trial [49]). These studies should provideinsight into the cardiovascular benefits ofthe two drugs.

Acknowledgments This study was jointlyfunded by Eli Lilly and Takeda Pharmaceuti-cals North America.

We acknowledge the efforts of Lisa Martin,Sandra Althouse, Melvin Prince, MD, and Ja-net Tobian, MD, PhD, without whom thisproject could not have reached completion.

References1. Haffner SM, DAgostino R Jr, Mykkanen

L, Tracy R, Howard B, Rewers M, Selby J,Savage PJ, Saad MF: Insulin sensitivity insubjects with type 2 diabetes: relationshipto cardiovascular risk factors: the InsulinResistance Atherosclerosis Study.DiabetesCare 22:562568, 1999

2. Nesto RW: Correlation between cardio-vascular disease and diabetes mellitus:current concepts.Am JMed 116:11S22S,2004

3. Hsueh WA, Lyon CJ, Quinones MJ: Insu-lin resistance and the endothelium. Am JMed 117:109117, 2004

4. Siegel RD, Cupples A, Schaefer EJ, WilsonPW: Lipoproteins, apolipoproteins, andlow-density lipoprotein size among dia-betics in the Framingham offspring study.Metabolism 45:12671272, 1996

5. Ginsberg HN: Insulin resistance and car-diovascular disease. J Clin Invest 106:453458, 2000

6. Krauss RM: Lipids and lipoproteins in pa-tients with type 2 diabetes. Diabetes Care27:14961504, 2004

7. Garvey WT, Kwon S, Zheng D, Shaugh-nessy S, Wallace P, Hutto A, Pugh K, Jen-kins AJ, Klein RL, Liao Y: Effects of insulinresistance and type 2 diabetes on lipopro-tein subclass particle size and concentra-tion determined by nuclear magneticresonance. Diabetes 52:453462, 2003

8. Gordon DJ, Probstfield JL, Garrison RJ,Neaton JD, Castelli WP, Knoke JD, JacobsDR Jr, Bangdiwala S, Tyroler HA: High-density lipoprotein cholesterol and car-diovascular disease: four prospectiveAmerican studies. Circulation 79:815,1989

9. Scandinavian Simvastatin Survival StudyGroup: Randomised trial of cholesterollowering in 4444 patients with coronaryheart disease: the Scandinavian Simvasta-tin Survival Study (4S). Lancet 344:13831389, 1994

10. Shepherd J, Cobbe SM, Ford I, Isles CG,Lorimer AR, MacFarlane PW, McKillopJH, Packard CJ: Prevention of coronaryheart disease with pravastatin in men withhypercholesterolemia: West of Scotland

Coronary Prevention Study Group.N Engl J Med 333:13011307, 1995

11. Ginsberg HN: Hypertriglyceridemia: newinsights and new approaches to pharma-cologic therapy. Am J Cardiol 87:11741180, 2001

12. American Diabetes Association: Dyslipi-demia management in adults with diabe-tes (Position Statement). Diabetes Care 27(Suppl. 1):S68S71, 2004

13. King AB: A comparison in a clinical set-ting of the efficacy and side effects of threethiazolidinediones (Letter). Diabetes Care23:557, 2000

14. Boyle PJ, King AB, Olansky L, MarchettiA, Lau H, Magar R, Martin J: Effects ofpioglitazone and rosiglitazone on bloodlipid levels and glycemic control in pa-tients with type 2 diabetes mellitus: a ret-rospective review of randomly selectedmedical records. Clin Ther 24:378396,2002

15. Khan MA, St Peter JV, Xoe JL: A prospective,randomized comparison of the metaboliceffects of pioglitazone or rosiglitazone in pa-tients with type 2 diabetes who were previ-ously treated with troglitazone. DiabetesCare 25:708711, 2002

16. LaCivita KA, Villarreal G: Differences inlipid profiles of patients given rosiglita-zone followed by pioglitazone. Curr MedRes Opin 18:363370, 2002

17. Olansky L, Marchetti A, Lau H: Multicenterretrospective assessment of thiazolidinedi-one monotherapy and combination therapyin patients with type 2 diabetes: compara-tive subgroup analyses of glycemic controland blood lipid levels. Clin Ther 25:B64B80, 2003

18. van Wijk JPH, de Koning EJP, MartensEP, Rabelink TJ: Thiazolidinediones andblood lipids in type 2 diabetes. Arterio-scler Thromb Vasc Biol 23:17441749,2003

19. Buse JB, Tan MH, Prince MJ, Erickson PP:The effects of oral anti-hyperglycaemicmedications on serum lipid profiles in pa-tients with type 2 diabetes. Diabetes ObesMetab 6:133156, 2004

20. Derosa G, Cicero AFG, Gaddi A, RagonesiPD, Fogari E, Bertone G, Ciccarelli L, Pic-cinni MN: Metabolic effects of pioglita-zone and rosiglitazone in patients withdiabetes and metabolic syndrome treatedwith glimepiride: a twelve-month, multi-center, double-blind, randomized, con-trolled, parallel-group trial. Clin Ther 26:744754, 2004

21. Gegick CG Altheimer MD: Thiazo-lidinediones: comparison of long-termeffects on glycemic control and cardiovas-cular risk factors. Curr Med Res Opin 20:919930, 2004

22. Peters Harmel AL, Kendall DM, Buse JB,Boyle PJ, Marchetti A, Lau H: Impact ofadjunctive thiazolidinedione therapy on



blood lipid levels and glycemic control inpatients with type 2 diabetes. Curr MedRes Opin 20:215223, 2004

23. Aronoff S, Rosenblatt S, Braithwaite S,Egan JW, Mathisen AL, Schneider RL:Pioglitazone hydrochloride monotherapyimproves glycemic control in the treat-ment of patients with type 2 diabetes. Di-abetes Care 23:16051611, 2000

24. Phillips LS, Grunberger G, Miller E, Pat-wardhan R, Rappaport EB, Salzman A:Once- and twice-daily dosing with rosigli-tazone improves glycemic control in pa-tients with type 2 diabetes. Diabetes Care24:308315, 2001

25. Matthews DR, Hosker JB, Rudenski AS,Naylor BA, Treacher DF, Turner RC: Ho-meostasis model assessment: insulin re-sistance and beta cell function fromfasting plasma glucose and insulin con-centrations in man. Diabetologia 28:412419, 1985

26. Searle SR: Linear Models for UnbalancedData. New York, John Wiley & Sons,1987

27. Gallo PP: Practical issues in linear modelsanalyses in multicenter clinical trials. Bio-pharm Rep 6:29, 1998

28. Rosenblatt S, Miskin B, Glazer NB, PrinceMJ, Robertson KE: The impact of piogli-tazone on glycemic control and athero-genic dyslipidemia in patients with type 2diabetes, mellitus. Coron Artery Dis 12:413423, 2001

29. Herz M, Johns D, Reviriego J, GrossmanLD, Godin C, Duran S, Hawkins F, Loch-nan H, Escobar-Jimenez F, Hardin PA,Konkoy CS, Tan MH: A randomized, dou-ble-blinded, placebo-controlled, clinicaltrial of the effects of pioglitazone on gly-cemic control and dyslipidemia in oralantihyperglycemic medication-nave pa-tients with type 2 diabetes mellitus. ClinTher 25:10741095, 2003

30. Patel J, Anderson RJ, Rappaport EB: Ros-iglitazone monotherapy improves glycae-mic control in patients with type 2diabetes: a twelve-week, randomized,placebo-controlled study. Diabetes ObesMetab 1:165172, 1999

31. Nolan JJ, Jones NP, Patwardhan R, Dea-con LF: Rosiglitazone taken once dailyprovides effective glycaemic control inpatients with type 2 diabetes mellitus.Diabet Med 17:287294, 2000

32. Raskin P, Rappaport EB, Cole ST, Yan Y,Patwardhan R, Freed MI: Rosiglitazoneshort-term monotherapy lowers fastingand post-prandial glucose in patients withtype II diabetes. Diabetologia 43:278284, 2000

33. Lebovitz HE, Dole JF, Patwardhan R, Rap-paport EB, Freed MI: Rosiglitazonemonotherapy is effective in patients withtype 2 diabetes. J Clin Endocrinol Metab86:280288, 2001

34. Turner RC, Millns H, Neil HA, StrattonIM, Manley SE, Matthews DR, HolmanRR: Risk factors for coronary artery dis-ease in non-insulin dependent diabetesmellitus: United Kingdom Prospective Di-abetes Study (UKPDS: 23). BMJ 316:823828, 1998

35. Expert Panel on Detection, Evaluation,and Treatment of High Blood Cholesterolin Adults (Adult Treatment Panel III): Ex-ecutive Summary of the Third Report ofthe National Cholesterol Education Pro-gram (NCEP). JAMA 285:24862497,2001

36. Grundy SM, Cleeman JI, Bairey Merz CN,Brewer HB Jr, Clark LT, HunninghakeDB, Pasternak RC, Smith SC Jr, Stone NJ:A summary of implications of recent clin-ical trials for the national cholesterol ed-ucation program adult treatment panel IIIguidelines. Arterioscler Thromb Vasc Biol24:13291330, 2004

37. Rubins HB, Robins SJ, Collins D, Fye CL,Anderson JW, Elam MB, Faas FH, LinaresE, Schaefer EJ, Schectman G, Wilt TJ,Wittes J: Gemfibrozil for the secondaryprevention of coronary heart disease inmen with low HDL-cholesterol. N EnglJ Med 341:410418, 1999

38. Zarich SW: Treating the diabetic patient:appropriate care for glycemic control andcardiovascular disease risk factors. RevCardiovasc Med 4 (Suppl. 6):S19S28,2003

39. Lamarche B, Tchernof A, Mauriege P,Cantin B, Dagenais GR, Lupien PJ, Des-pres JP: Fasting insulin and apolipopro-tein B levels and low-density lipoproteinparticle size as risk factors for ischemicheart disease. JAMA 279:19551961,1998

40. Reaven GM, Chen YD, Jeppesen J, Ma-heux P, Krauss RM: Insulin resistance andhyperinsulinemia in individuals with

small, dense low density lipoprotein par-ticles. J Clin Invest 92:141146, 1993

41. Lahdenpera S, Syvanne M, Kahri J, Taski-nen MR: Regulation of low-density li-poprotein particle size distribution inNIDDM and coronary disease: impor-tance of serum triglycerides. Diabetologia39:453461, 1996

42. Taskinen MR: Diabetic dyslipidemia. Ath-eroscler Suppl 3:4751, 2002

43. Freed MI, Ratner R, Marcovina SM, Krei-der MM, Biswas N, Cohen BR, BrunzellJD, Rosiglitazone Study 108 investigators:Effects of rosiglitazone alone and in com-bination with atorvastatin on the meta-bolic abnormalities in type 2 diabetesmellitus. Am J Cardiol 90:947952, 2002

44. Bell DS: Dyslipidemia in type 2 diabetesand the effect of thiazolidinediones. Endo-crinologist 13:496504, 2003

45. Lewin AJ, Kipnes MS, Meneghini LF,Plotkin DJ, Perevozskaya IT, Shah S, Mac-cubbin DL, Mitchel YB, Tobert JA, Simva-statin/Thiazolidinedione Study Group.Effects of simvastatin on the lipid profileand attainment of low-density lipoproteincholesterol goals when added to thiazo-lidinedione therapy in patients with type2 diabetes mellitus: a multicenter, ran-domized, double-blind, placebo-con-trolled trial. Clin Ther 26:379389, 2004

46. CARE Foundation: CARE today [articleonline], 2004. Available from http://www.aspirus.org/heart_institute/newsletters/CaRETodayWinter04.pdf. Accessed 30April 2005

47. Charbonnel B, Dormandy J, Erdmann E,Massi-Benedetti M, Skene A: The pro-spective pioglitazone clinical trial in mac-rovascular events (PROactive): canpioglitazone reduce cardiovascular eventsin diabetes? Study design and baselinecharacteristics of 5238 patients. DiabetesCare 27:16471653, 2004

48. CME on diabetes [article online], 2004.Available from http://www.cmeondiabetes.com/pub/shifting.the.paradigm.from.stepwise.to.early.combination.therapy.php.Accessed 8 August 2004

49. Sobel BE, Frye R, Detre KM: Burgeoningdilemmas in the management of diabetesand cardiovascular disease. Rationale forthe Bypass Angioplasty RevascularizationInvestigation 2 Diabetes (BARI 2D) trial.Circulation 107:636642, 2003



DIABETESReps will be touting new evidence that pioglitazone (Actos)

improves lipids better than rosiglitazone (Avandia).Both glitazones reduce A1C equally.But pioglitazone seems better at lowering triglycerides and

increasing HDL. Pioglitazone LOWERS triglycerides about 9% to 12%... rosiglitazone can INCREASE triglycerides up to 15%.

Pioglitazone also raises HDL about 12% to 19%...compared to8% to 19% for rosiglitazone.

Keep this in perspective. Most diabetes patients take statinsor other lipid-lowering drugs, which may negate any differences.

Pio and rosi differ in cost. Pio is about $192 for a months supplyof the 45 mg tabs...rosi is about $175 for a month of the 8 mg tabs.

Keep in mind that lipid levels are only SURROGATE MARKERS...theydont show if either glitazone is better for long-term OUTCOMES.

(For more on this topic, see Detail-Document #210802 at www.prescribersletter.com.)

Primary Reference Goldberg RB, Kendall DM, Deeg MA, et al. A comparison of lipid and glycemic effects of pioglitazone and rosiglitazone in patients with type 2 diabetes and dyslipidemia. Diabetes Care 2005;28:1547-54.

Overview of current therapy

1. How commonly are pioglitazone and rosiglitazone used in Type 2 diabetes?

Both pioglitazone (Actos) and rosiglitazone (Avandia) are Top 100 Drugs in the U.S. Over 9 million prescriptions in 2004.

2. What are the differences between these agents?

There are NO significant differences in glucose-lowering; both lower hemoglobin A1C by 0.5% 1%. There are NO outcomes studies for diabetes-related endpointscardiovascular diseaserenal diseaseor

microvascular events such as retinopathy. There is NOT much difference in adverse effectsedemaelevation of liver function tests (LFTs). Preliminary evidence suggests that there MAY be a difference in serum lipid effects.

3. What is known about the effects of glitazones on lipids?

Tend to increase HDL cholesterol. Variable to neutral effects on LDL cholesterol. Big difference is rosiglitazone tends to increase triglycerides where pioglitazone is more likely to reduce

triglycerides. This study evaluates the effects of rosiglitazone and pioglitazone on triglycerides, HDL, LDL, and LDL

particle composition and size.

Analysis of new study

4. What type of study was this and how were the patients selected?

A randomized, double-blind trial of rosiglitazone versus pioglitazone. Subjects with Type 2 diabetes were screened from the U.S., Mexico, Puerto Rico, and Columbia. Eligible patients were > 35 years of age, triglycerides between 150 and 600 mg/dL, a C-peptide level >1

ng/mL, and an A1C between 7% 11% (if nave to oral hypoglycemic medications) or between 7% 9.5% (if previously treated with oral hypoglycemic monotherapy).

LEADER NOTES LEADER NOTES

The following succinct analysis appeared in Prescribers Letter.

Discussion Questions

- 2 - PL Journal Club August 2005 PL Journal Club August 2005 - 11 - Welcome to PL Journal Club PL Journal Club gives you insights and guides you to the discoveries that Prescribers Letter researchers and editors uncover. Each month we analyze many new studies and help you discover the answers to the hard questions. What are the real advantages and disadvantages of new therapies? How do they compare with other options? What do prescribers and pharmacists need to know? We look beyond the headlines and promotional materials to interpret the clinical studies and data. Sometimes the marketing spin doesnt stand up to scrutiny. Sometimes studies do not really prove what they are reported to prove. PL Journal Club helps guide you to the truth and how to apply new ndings to patient care.

PL Journal Club builds on Prescribers Letter to provide you with background for your own journal club discussions. We wont bring up every possible question, but you canin your own journal club. If a question comes up, go to www.prescribersletter.com to nd more background. As a PL Journal Club participant, you get access to all of Prescribers Letter and you get access to the special PL Journal Club online tools. Youll be able to communicate with other PL Journal Club participants in many other locations. Plus, youre able to send questions and comments to the author of the published study that is discussed in PL Journal Club. And all PL Journal Club par-ticipants will benet from seeing the responses from the study author. To see this, go to www.prescribersletter.com, click PL Journal Club, then click DISCUSSION FORUM. Feel free to call or email us with suggestions or if we can be of assistance... 209-472-2240 or [email protected].

InstructionsGo to www.prescribersletter.com to get materials and use the PL Journal Club tools. PL Journal Club leaders can use the tools to invite participantssend reminders...get attendance rosters...download interpretations and analyses of studies...etc. Participants can use the tools to get discussion materials...get advance notice of discussion questions...reference citations...etc. Youll also get great background materi-als, including Prescribers Letter and Prescribers Letter Detail-Documents. PL Journal Club functions like a typical journal club meet-ing, except that it is pre-organized for you and you get the expert analysis of important new studies done by the large Prescribers Letter research and editorial staff. Let the questions serve as a springboard for your discussions. Let your own patient cases shape the discussion. Each month PL Journal Club gives you two topics that are also covered in Prescribers Letter. Your group might discuss only one or both topics. Youll also nd a library of previous PL Journal Clubs online for your use. And the discussion forum lets you share your thoughts, observations, questions, etc with other PL Journal Club participants online.

PL Journal Club Contributing Editors: Rex W. Force, Pharm.D., FCCP, BCPS Mark Graber, M.D. Brian C. Harrington, M.D., MPH, FAAFPProfessor, Department of Family Medicine Research Director, Past Program Director,Associate Professor, Department of Pharmacy Division of Emergency Medicine Darnall Family Medicine Residency ProgramPractice Associate Professor of Clinical Fort Hood, TXDirector of Research, Grants and Information Surgery and Family Medicine Current Teaching FacultySystems Univ. of Iowa Hospitals and Southern Colorado Family Medicine Residency Family Medicine Clinical Research Center Clinics Program Idaho State Univ. Assistant Clinical Professor Univ. of Colorado Health Sciences Center

Prescribers Letter Editorial Staff: Editor: Jeff M. Jellin, Pharm.D.; Senior Associate Editor: Karen Davidson, Pharm.D.; Assistant Edi-tors: Kim Palacioz, Pharm.D., Tanveer Khan, Pharm.D., Joseph A. Woelfel, Ph.D., FASCP, R.Ph., Melissa Murn, PA-C, Pharm.D., Kayla Dotson, Pharm.D., Crystal Amos, B.Sc.Pharm., ACPR; Assistant Editor and Director of Continuing Education: Tony Martin, Pharm.D., MBA; Associate Clinical Editor: Mark Graber, M.D.; Editorial Project Manager: Timothy Swaim; Editorial Liaison: Karen Wilson; Editorial Assistant: Tyler Nagata; Database Coordinator: Stephanie Feilzer-Pate; Drug Information Consultants: Neeta OMara, Pharm.D., BCPS, Jill E. Allen, Pharm.D., BCPS; Research and Review: Kwabena O.M. Adubofour, M.D., FACP, Thomas Barringer, M.D., FAAFP, Robert T. Browne, M.D., FAAFP, Matthew Cline, M.D., FAAFP, Sandra Harley Counts, Pharm.D., Rex Force, Pharm.D., BCPS, Joseph Guglielmo, Pharm.D., Marcia Isakari, M.D., Sheela Kapre, M.D., Adam Kaye, Pharm.D., FASCP, William Kehoe, Pharm.D., MA, FCCP, BCPS, Ted Lee, M.D., Stanley Leong, Pharm.D., Guey Mark, M.D., John Morozumi, M.D., Kay Niegel, R.Ph., Mike Pastrick, R.Ph., Allen Shaughnessy, Pharm.D., Tom Simpson, Pharm.D., FCSHP, Sharm Steadman, Pharm.D.

Prescribers Letter Editorial Advisory Board: Jan Basile, M.D., Professor of Medicine, Medical Univ. of South Carolina; Robert E. Bick-erton, M.D., FACP, Private Practice; Kevin Brown, M.D., FACOG, Private Practice; David B. Brushwood, J.D., Law Professor, University of Florida; Narinder Duggal, M.D., BSc (Pharm) CDE, CGP, FRCPC, Univ. of Washington; Col John D. Grabenstein, R.Ph., Ph.D., Deputy Director for Clinical Operations, US Army Medical Command; Mark A. Graber, M.D., Research Director, Div of Emergency Medicine, Associate Professor of Clinical Surgery and Family Medicine, University of Iowa Hospitals and Clinics; Col. Jefferson Harman, Jr., M.D., Director, Travis Air Force Base Family Practice Residency Program; Arthur Hull Hayes, Jr., M.D., President, MediScience Associates, Inc., former Commissioner of the FDA; B. Mark Hess, M.D., FACP, Wright State University; Eric Jackson, Pharm.D., BCPS, Asylum Hill Fam-ily Practice Center University of Connecticut; Peter Jacobsen, Ph.D., D.D.S., Professor of Oral Medicine, University of the Pacic, School of Dentistry; Evan L. Lipkis, M.D., Medical Director, Advanced Center for Total Health Care, Glenbrook Hospital; Stephen L. McKernan, BS.Pharm., N.D., D.O., FAAFP, Director, Conroe Family Practice Residency Program; Darius Noori, Pharm.D., M.D.; Robert W. Patter-son, M.D., FABFP, FAAFP, Family Practice; David E. Rosenthal, M.D., FACP, Abington Memorial Hospital; David Stadtner, M.D., Private Practice; Kathryn Taubert, Ph.D., American Heart Assoc National Center; Walter Way, M.D., University of Cal at San Francisco; C. Wayne Weart, Pharm.D., BCPS, Professor and Chair, Medical University of South Carolina; Charles, F. Weiss, M.D., Clinical Professor of Pediat-rics, University of South Florida; Craig Williams, Pharm.D., Purdue University.

DISCLOSURE: Therapeutic Research Center/Prescribers Letter does not receive any commercial support and does not accept any advertising. It is completely independent and is supported entirely by subscriptions. It is our policy that all editors refrain from having any ties with any or-ganization that could inuence the editorial objectivity of Prescribers Letter or PL Journal Club. Prescribers Letter and PL Journal Club focus on delivering completely objective, unbiased drug information and advice for the benet of subscribers.

jbottemillerDiabetes Care 2005;28:1547-54

jbottemiller


References

- 10 - PL Journal Club August 2005 PL Journal Club August 2005 - 3 -

Exclusion criteria included elevated serum creatinine, elevated liver enzymes, recent cardiovascular disease, combination oral hypoglycemic agents, and treatment for lipid disorders.

5. How are study groups defined?

4,410 patients screened, 802 patients randomized, 735 received study drug. Patients were an average age of 55, 54% were male, 60% 65% were white, 30% were Hispanic, BMI

of 33, an A1C of 7.5%, triglycerides of 250 mg/dL, LDL of 108 mg/dL, and < 10 % had pre-existingcardiovascular disease.

Study groups were well matched at baseline, indicating effective randomization procedures. Patients received either rosiglitazone 4 mg daily or pioglitazone 30 mg daily for 12 weeks. Doses were increased to rosiglitazone 4 mg twice daily or pioglitazone 45 mg daily for the final 12 weeks. Serum lipids were measured every 4 weeks.

6. What are the strengths and weaknesses of this trial?

Randomized, double-blind trial. Relatively small study with enough power to detect differences in endpoints. Evaluated changes in laboratory parameters, NOT clinical outcomes. Follow-up rates are so-so with about 20% of patients dropping out over the 24-week study period.

7. How were the study subjects not representative of real world diabetic patients?

Exclusion criteria significantly limit the generalizability of the study to the real world. Criteria included:o NO other diabetes medications or statins most patients with diabetes do take these.o any condition or situation precluding adherence to...the protocol. nonadherence occurs in reality

and would result in smaller lipid changes.

8. How did the drugs work in this trial?

At 24 weeks, both drugs reduced hemoglobin A1C by similar amounts (about 0.7 %). Rosiglitazone increased HDL by an average of 2.4 mg/dL, increased triglycerides by 13.1 mg/dL, and

increased LDL by 21.3 mg/dL. Pioglitazone increased HDL by an average of 5.2 mg/dL, decreased triglycerides by 51.9 mg/dL, and

increased LD by 12.3 mg/dL. LDL particles were larger and less buoyant with pioglitazone.

9. Were there any significant adverse affects in either group?

NO differences in weight gain, elevation of LFTs, hypoglycemia, edema, or heart failure.

10. Were there any other design issues that might influence the results?

A MAJOR limitation is that the results do NOT necessarily translate to an outcome we care about.Altering an intermediate outcome (LDL or blood pressure) does NOT always translate to a reduction in cardiovascular disease. For example, in the ALLHAT trial, doxazosin lowered blood pressure as well as chlorthalidone, but doxazosin was associated with new heart failure at an alarming rate.

Screen-failure rate was quite high only 18% of those screened were randomized. This calls into question the external validity.

Glycemic control was modest at best (A1C of ~7.5% on glitazone monotherapy). Withdrawal rate was quite high nearly 20%. Sponsored by the makers of pioglitazone (Takeda).

DIABETES

Chren M. Interactions between physicians and drug company representatives. Am J Med 1999;107:182-3.

Goldberg RB, Kendall DM, Deeg MA, et al. A comparison of lipid and glycemic effects of pioglitazone and rosiglitazone in patients with type 2 diabetes and dyslipidemia. Diabetes Care 2005;28:1547-54.

Grandage KK, Slawson DC, Shaughnessy AF. When less is more: a practical approach to searching for evidence-based answers. J Med Libr Assoc 2002;90:298-304.

Shaughnessy AF, Slawson DC, Bennet JH. Teaching information mastery: evaluating information provided by pharmaceutical representatives. Fam Med 1995;27:581-5.

Slawson DC, Shaughnessy AF. Obtaining useful information from expert based sources. BMJ 1997;314:947-9.

Slawson DC, Shaughnessy AF, Bennett JH. Becoming a medical information master: feeling good about not knowing everything. J Fam Pract 1994;38:505-13.

Watkins RS, Kimberly J Jr. What residents dont know about physician-pharmaceutical industry interactions. Acad Med 2004;79:432-7.

Additional Prescribers Letter Resources available at www.prescribersletter.com

Does Actos Improve Lipids Better than Avandia? Pharmacist's Letter/Prescriber's Letter 2005;21:210802.

Worksheet to Evaluate Drug Studies. Pharmacist's Letter/Prescribers Letter 2005;21:210610.

Liver Function Test Scheduling. Pharmacist's Letter/Prescribers Letter 2004;20:200903.

Avandia (Rosiglitazone) Approved For Use With Insulin. Pharmacist's Letter/Prescribers Letter 2003;19:190402.

Relationship Between Industry And Research And Professional Education. Pharmacists Letter/Prescriber's Letter2001;8:171025.

WOMENS HEALTH

Abraham GE. Nutritional factors in the etiology of the premenstrual tension syndromes. J Reprod Med 1983;28:446-64.

Bertone-Johnson ER, Hankinson SE, Bendich A, et al. Calcium and vitamin D intake and risk of incident premenstrual syndrome. ArchIntern Med 2005;165:1246-52.

Clinical management guidelines for obstetricians-gynecologists: premenstrual syndrome. ACOG practice bulletin. No. 15. Washington, D.C.: American College of Obstetricians and Gynecologists, April 2000.

Dell DL. Premenstrual syndrome, premenstrual dysphoric disorder, and premenstrual exacerbation of another disorder. Clin Obstet Gynecol 2004;47:568-75.

Fletcher RH, Fletcher SW, Wagner EH. Clinical epidemiology: The essentials. 3rd ed. Baltimore: Williams & Wilkins,1996.

Friis RH, Sellers TA., Moffitt HL. Epidemiology for public health practice. 3rd ed. Gaithersburg, Maryland: Aspen Publishers, Inc., 2004.

Grady-Weliky TA. Clinical practice. Premenstrual dysphoric disorder. N Engl J Med 2003;348:433-8.

National Osteoporosis Foundation. Prevention: Calcium & Vitamin D. http://www.nof.org/prevention/calcium.htm. (Accessed 7/26/05).

Thys-Jacobs S, Starkey P, Bernstein D, Tian J. Calcium carbonate and the premenstrual syndrome: effects on premenstrual and menstrual symptoms. Premenstrual Syndrome Study Group. Am J Obstet Gynecol 1998;179:444-52.

Additional Prescribers Letter Resources available at www.prescribersletter.com

Calcium and Other Supplements for PMS. Pharmacist's Letter/Prescribers Letter 2005;21:210813.

Calcium and Vitamin D for Reducing Fracture Risk. Pharmacist's Letter/Prescriber's Letter 2005;21:210609.

Calcium and Colorectal Cancer. Pharmacist's Letter/Prescriber's Letter 2005;21:210656.

PMS Guidelines. Pharmacist's Letter/Prescribers Letter 2000;16:160803.

Calcium Salts. Pharmacist's Letter/Prescribers Letter 2000;16:160313.



11. Were the results expressed in terms we care about and can use?

Yes. Treatment resulted in similar reductions in hyperglycemia, but very different changes in good and bad cholesterol. However, whether this actually affects cardiovascular outcomes is yet to be determined.

How should the new findings change current therapy?

12. Are the patients studied similar to those you see?

Discuss your patient population. Study population:

o Generally, patients were young (age 55).o Recently diagnosed with diabetes (



o Results focused on RELATIVE change, rather than the ABSOLUTE changes in triglycerides. ABSOLUTE changes were relatively small.

A valid study may not be particularly RELEVANT. The best information is patient oriented evidence that matters (POEMs). Many studies look at

intermediate endpoints, disease oriented evidence (DOE). For example, small changes in lipid levels may not have a significant clinical benefit. A change in heart attack rates would be more useful.

Points that affect this studys relevance:o Looked at intermediate endpoints, NOT clinical outcomes. o Is not real world. Subjects differed in several ways from typical diabetes patients.

The validity and relevance of this study are low. This study does NOT dictate a practice preference of pioglitazone over rosiglitazone for patients with diabetes or high triglycerides.

16. What attributes should you consider when evaluating if one drug is better than another?

Shaughnessy and Slawson describe the STEP approach:Safety long-term or serious side effectsTolerability best to compare drop out rates, not relative incidence of side effectsEffectiveness may be hard to get good comparative data from drug representativesPrice direct + indirect (extra monitoring costs; extra visits, etc.)

17. What serious adverse reactions can occur in patients taking pioglitazone or rosiglitazone?

Heart failure. Both can cause fluid retention and edema, which can lead to or exacerbate heart failure, especially when combined with insulin.

Hepatotoxicity. Association was more prevalent with Rezulin (troglitazone), which eventually led to its withdrawal in 2000. Newer recommendations on LFT monitoring suggest doing so before starting and thenperiodically. Some experts interpret periodically to mean every 3 to 6 months.

18. What if the company rep states that Actos is the most prescribed thiazolidinedione among endocrinologists? Does this convince you to prefer Actos?

Prescription volume does NOT necessarily indicate it is the BEST choice. A top ranking may indicate an aggressive marketing program. Sometimes companies will sell their product at a competitively low price in order to gain market share and

thus claim it is the number one prescription item in its class.

19. How much do pioglitazone and rosiglitazone cost?

Approximate retail price for a month supply, as used in the study:pioglitazone 45 mg once a day = $165 $192rosiglitazone 8 mg daily (#60 of 4 mg or #30 of 8 mg) = $166 $175

20. How can providers interact with pharmaceutical representatives in a way that is mutually respectful and helps the representative be a positive resource for information?

Explain that you need information that is VALID and RELEVANT. Emphasize the need for POEMs, not DOEs; limit pre-clinical data. Ask them to provide information that conforms to the STEP approach.

For more content, and to see comments from other PL Journal Clubs on this topic go to www.prescribersletter.com - click on PL Journal Club - click on DISCUSSION FORUM.Youll be able to pose your questions to the study author, Dr. Goldberg.

9. Were there any design issues that might influence results?

The MAJOR limitation of case-control studies is that biases are not well controlled for.o For example, women who consume large amounts of calcium might have other health habits that

reduce their risk of developing PMS. o They may have a different personality type or outlook on life that results in better coping mechanisms

and responses to stresses.o Without randomization, we cannot be certain that women who consume large amounts of calcium or

vitamin D might not be different in some way from those who consume lower amounts of these nutrients.

Nurses may not be representative of ALL women; therefore, generalizability may be suspect. Recall biases can be a significant problem; that is, are subjects adequately recalling their health history? Other limitations:

o PMS may be difficult to diagnose, making correct identification of PMS cases complicated. o Small proportion of patients reported using calcium supplements, making it difficult to identify the

potential benefit of high-dose calcium supplementation in preventing PMS development.

10. Were the results expressed in terms we care about? Is the intervention feasible?

Yes. Young women are frequently NOT receiving enough calcium and vitamin D in their diet. This study demonstrates an association between high levels of calcium intake and subsequent development

of PMS. Its reasonable to recommend these nutrients in a larger subset of women. Calcium and vitamin D intake has been associated with a reduction in the risk of osteoporosis and some

cancers. It is extremely unlikely that these products will cause significant adverse effects. Since the cost of these products is relatively low, it makes sense to recommend an increased intake.

How should the new findings change current therapy?

11. Are the patients studied similar to those you see in clinic?

Discuss your patient population. Study population:

o A relatively homogeneous population of nurses; this may limit external validity.

12. Do the results force you to change your practice? How?

Maybe. The results are not earth shattering enough to mandate an immediate change. Results support potential additional benefits of calcium and vitamin D supplementation. Its reasonable to

use the results from this trial as additional supporting evidence in recommending calcium and vitamin D supplementation to younger patients. Again, the downside is relatively minimal.

Apply the new findings to the following case

TZ is a 22 year old female, who presents for her annual well woman exam. She is generally healthy, exercises regularly, and is non-obese. Her mother takes calcium pills for bone health and tells TZ that calcium also used to help her severe PMS symptoms. TZ denies PMS symptoms, but asks if she should be taking calcium too.

13. Do you think TZ should pay attention to her calcium intake?

Yes. Calcium intake is critical for bone strength. This is an important period for TZs bone growth. Women benefit from reaching the highest bone mass possible, which usually peaks during their mid-30s. This study suggests a role for calcium and vitamin D in preventing PMS. People should be getting adequate daily calcium intake with their diets. Many people do NOT. Poor dietary habits have been associated with PMS. If TZs dietary intake of calcium appears inadequate, and given her mothers history of PMS,

recommending dietary supplementation may be prudent.

jbottemillerwww.prescribersletter.com