Available online at www.sciencedirect.com

General Hospital Psychiatry 34 (2012) 380–384

Prevalence of baseline lipid monitoring in patients prescribedsecond-generation antipsychotics during their index hospitalization:

a retrospective cohort study☆,☆☆

Walter Laundon, Pharm.D.a, Andrew J. Muzyk, Pharm.D.b, c,⁎,Jane P. Gagliardi, M.D., M.H.S.c,d, Eric J. Christopher, M.D.c,d,

Tracie Rothrock-Christian, Pharm.D.a, e, Wei Jiang, M.D.c,daDepartment of Pharmacy, Duke University Hospital, Durham, NC 27710, USA

bDepartment of Pharmacy Practice, Campbell University School of Pharmacy and Health Sciences, Buies Creek, NC 27506, USAcDepartment of Psychiatry and Behavioral Sciences, Duke University, PO Box 3089–Pharmacy, Durham, NC 27710, USA

dDepartment of Medicine, Duke University, Durham, NC 27710, USAeDepartment of Pharmacy Practice, UNC Eshelman School of Pharmacy, Chapel Hill, NC 27599, USA

Received 26 August 2011; accepted 20 March 2012

Abstract

Background: Second-generation antipsychotics have been found to increase a patient's risk of dyslipidemia. Despite consensus statementrecommendations for lipid monitoring, studies indicate that up to 90% of patients still do not have a baseline lipid panel prior to prescriptionof a second-generation antipsychotic.Methods: This study retrospectively examined the prevalence of baseline lipid monitoring in patients prescribed second-generationantipsychotics during their index psychiatric hospitalization at Duke University Hospital between July 1, 2005, and July 1, 2010.Results: Seventy patients were included in the study, with a mean age of 21.5±2.5 years. Of these patients, 22 (31.4%) had baseline lipidpanels drawn during hospitalization. Lipid monitoring was statistically more frequent in males than in females (P=.01). Although notstatistically significant, lipid monitoring was also more likely to occur among subjects who were African American (40%; P=.07) and withthe prescription of olanzapine (50%; P=.07). About half of baseline lipid panels demonstrated either a low high-density lipoprotein or hightriglycerides, indicating at least one risk factor for the metabolic syndrome.Conclusion: This study provides alarming evidence that, even in an academic setting with active discussions among psychiatrists regardingissues of metabolic risk and appropriate monitoring, adherence to American Psychiatric Association/American Diabetes Associationconsensus statement recommendations on rates of baseline lipid monitoring is disappointingly low in the absence of systems to encourage orautomate best practice.© 2012 Elsevier Inc. All rights reserved.

Keywords: Antipsychotics; Atypical antipsychotics; Second-generation antipsychotics; Lipid monitoring; Dyslipidemia; First episode; Schizophrenia;Bipolar; Psychosis

☆ Financial disclosure: No financial support was received for thispublication. The authors of this manuscript have no sources of financialsupport relevant to this publication.

☆☆ Additional disclosure: This manuscript (in part or in whole) has notbeen previously presented or published.

⁎ Corresponding author. Tel.: +1 919 681 3438.E-mail address: Andrew.Muzyk@duke.edu (A.J. Muzyk).

0163-8343/$ – see front matter © 2012 Elsevier Inc. All rights reserved.doi:10.1016/j.genhosppsych.2012.03.016

1. Introduction

Patients with mental illness are at increased risk formortality compared to the general population, with up to a30-year decreased life span as compared to patients withoutmental illness [1]. Two large public health studies examiningall-cause mortality reported that cardiovascular disease(CVD) was the single largest cause of death in patientswith schizophrenia [1,2]. Other studies have demonstrated

381W. Laundon et al. / General Hospital Psychiatry 34 (2012) 380–384

that, as compared to patients without mental illness, patientswith schizophrenia have a 1.5 to 2 times greater incidence ofCVD risk factors such as weight gain, hypertension, diabetesand dyslipidemia (a cluster of illnesses also known as themetabolic syndrome) [3]. Studies in patients with bipolardisorder have reported similar findings [4,5], although thereis substantially less research in this population. Dyslipidemiais one of the major causes of coronary heart disease (CHD), adisease under the rubric of CVD [6], which has also beenshown to be a major cause of mortality in patients withmental illness [7].

A study comparing baseline demographic data ofschizophrenic subjects enrolled in the Clinical AntipsychoticTrials of Intervention Effectiveness trial to matched controlswithout mental illness from the National Health andNutrition Examination Survey III database found thatsignificantly more patients with schizophrenia met thecriteria for measures of atherogenic dyslipidemia andmetabolic syndrome [8]. Atherogenic dyslipidemia isassociated with low high-density lipoprotein (HDL) andelevated triglycerides, which are two of the five modifiablerisk factors for metabolic syndrome [6]. A similar studyusing the same patient population found significantly higher10-year risk estimates for CHD among patients withschizophrenia compared to matched controls [9]. Since themajority of subjects in these trials reported chronicpsychiatric illness and had a long duration of multipleantipsychotic treatments, these findings heighten concernsthat dyslipidemia and the metabolic syndrome attributable tomedication use may be contributing to the morbidity andmortality in this patient population.

The American Psychiatric Association (APA), in collab-oration with a number of other organizations, including theAmerican Diabetes Association (ADA), published a consen-sus statement addressing concerns about second-generationantipsychotic-inducedmetabolic syndrome [10]. Focusing ondyslipidemia, the consensus among experts was that second-generation antipsychotics can cause changes to lipidmetabolism, and the group proposed monitoring guidelinesfor lipids at baseline, 12-week intervals and 5-year intervalsfor patients prescribed these agents regardless of theindication. The Mt. Sinai Conference, an expert panelconvened to address adverse effects with antipsychotics,came to a similar consensus on the need for lipid monitoring[11]. Despite these recommendations, the prevalence of lipidmonitoring in patients prescribed second-generation antipsy-chotics remains consistently low at 10%, approximately thesame monitoring rate before and after the publication of theAPA/ADA consensus statement [12–15].

This study was conducted to determine the prevalencerate of baseline lipid monitoring in psychiatric inpatientsstarted on second-generation antipsychotics at an academicteaching hospital. Between 2002 and 2006, all inpatientpsychiatric services at our hospital were provided by duallytrained physicians in internal medicine and psychiatry, andour academic teaching hospital has developed a faculty and

training program in evidence-based psychiatry and internalmedicine/psychiatry. Even before the consensus statementsregarding lipid monitoring were published, journal clubs andacademic discussions frequently include issues of appropri-ate monitoring and metabolic risk with second-generationantipsychotics. We, therefore, anticipated that rates of lipidmonitoring in our institution would be higher than rates oflipid monitoring reported in studies conducted of patientsprescribed second-generation antipsychotics (whether newlystarted or continued) in commercial health plans andMedicaid databases [12–15].

2. Methods

Individuals were identified by performing a DrugUtilization Review search on second-generation antipsy-chotics. Patients were eligible for inclusion if they werebetween the ages of 18 and 25 years; were hospitalized onthe adult inpatient psychiatry 18-bed ward between July 1,2005, and July 1, 2010; were diagnosed with bipolar disorder[International Classification of Diseases, Ninth Revision(ICD-9) codes 296.0–296.9], schizophrenia (ICD-9 codes295.0–295.9) or psychosis not otherwise specified (NOS)(ICD-9 code 298); were given new prescription (e.g., duringindex hospitalization) for a second-generation antipsychoticand discharge prescription for a second-generation antipsy-chotic and had medical records available for review. “Firsttime use” was defined as having never received anyantipsychotic prior to the index hospitalization. Patientmedical records and chart notes were reviewed by twoauthors (W.L. and A.M.) to confirm that included subjectshad not been previously prescribed an antipsychotic.Lurasidone (Lutada) was not included in this study since itwas not approved by the Food and Drug Administration(FDA) until October 2010 [16].

Data collected from patients' medical records includedthe following: age, sex, ethnicity, admitting ICD code,length of hospital stay, comorbid illnesses, prescribedsecond-generation antipsychotic and the presence of abaseline lipid panel. If the patient had a baseline lipidpanel, the individual values were recorded as well as thenumber of days between the lipid panel and initiation of asecond-generation antipsychotic.

Based on the assumption that if a lipid panel was presentwithin the previous 6 months and available for review at thetime of initiation of an atypical antipsychotic, a repeat“baseline” might not be necessary during the indexhospitalization, the authors of this study reached consensusto define lipid panels performed within the 6 monthspreceding current hospitalization as well as those performedduring hospitalization as “baseline lipid monitoring.” Inaddition to the primary clinical question regarding overallrates of baseline monitoring, we sought to determine whetherthere were differential rates of monitoring based onadmitting diagnosis, sex, ethnicity and prescribed second-

able 2ates of baseline lipid monitoring by subject characteristics

haracteristic Number (%) of patientswith a lipid panel

exa

Male (n=39) 17 (43.6)Female (n=31) 5 (16.1)

382 W. Laundon et al. / General Hospital Psychiatry 34 (2012) 380–384

generation antipsychotic. Data analysis was mainly per-formed using descriptive statistics, but χ2 statistical analyseswere used for comparisons of categorical data.

This study was reviewed by the Duke University HealthSystem Investigational Review Board, which granted anexemption. No protected health information was collected ormaintained in study records.

iagnosisPsychosis NOS 10 (47.6)Bipolar disorder 11 (27.5)Schizophrenia 2 (22.2)thnicityAfrican American (n=40)b 16 (40.0)Caucasian (n=25) 5 (20.0)Asian (n=2) 0 (0)Not reported (n=3) 1 (33.3)econd-generation antipsychotic prescribedRisperidone (n=28) 9 (32.1)Olanzapine (n=16)c 8 (50.0)Aripiprazole (n=12) 3 (25.0)Quetiapine (n=10) 1 (10.0)Ziprasidone (n=2) 0 (0)Paliperidone (n=2) 1 (50.0)

a Comparison of males vs. females (χ2=6.04, P=.01).b Comparison of African American vs. other ethnicities (χ2=3.18,

=.07).c Comparison of olanzapine vs. other second-generation antipsychotics

2=3.32, P=.07).

3. Results

During the study period, 210 individuals hospitalized atour institution were prescribed a second-generation antipsy-chotic and were between the ages of 18 and 25 years. Ofthese 210 patients, 140 were excluded because of (1)previous prescription of a second-generation antipsychotic(n=106), (2) admission to a service other than the (n=30) or(3) insufficient documentation (n=4). Table 1 contains thecharacteristics of the 70 subjects whose data were analyzed.

Baseline lipid panels were obtained in 31.4% of subjects.All lipid panels were early morning and assumed to be donein the fasting state. Of those with lipid panels, 68.1% hadresults posted prior to the initiation of a second-generationantipsychotic, with a mean time from lipid panel to anti-psychotic initiation of 4.3±7.4 days. Lipid panels drawn aftersecond-generation antipsychotic initiation were completedwithin a mean of 2.0±2.1 days.

Median values were as follows: total cholesterol, 143mg/dl (range 68–218 mg/dl); triglycerides, 62 mg/dl (range,29–225 mg/dl); HDL, 41.5 mg/dl (range, 27–69 mg/dl) andlow-density lipoproteins, 76 mg/dl (range, 27–149 mg/dl).Half of the subjects qualified as having a low HDL level

Table 1Subject characteristics

Characteristic Number (%) ormean±S.D.

Number of subjects (first-time user of asecond-generation antipsychotics)

70

Male 39 (55.7)Age (years) at admission 21.5±2.5Ethnicity (% of total)African American 40 (57.1)Caucasian 25 (35.7)Asian 2 (2.9)Not reported 3 (4.3)

Admitting diagnosis (% of total)Bipolar 40 (57.1)Psychosis NOS 21 (30.0)Schizophrenia 9 (12.9)

Number (% of total) of second-generation antipsychotic prescriptionRisperidone 28 (40.0)Olanzapine 16 (22.9)Aripiprazole 12 (17.1)Quetiapine 10 (14.3)Ziprasidone 2 (2.9)Paliperidone 2 (2.8)Iloperidone, asenapine 0

Mean±S.D. length (days) of hospital stay 7.9±5.3

TR

C

S

D

E

S

P

(χ

(b40 mg/dl for males, b50 mg/dl for females). Thirty-fourpatients (47.8%) had elevated triglyceride levels (N150mg/dl). Approximately 5% of subjects had both a low HDLand elevated triglycerides.

Subject characteristics and rates of lipid panel ascer-tainment are listed in Table 2. There were no statisticallysignificant differences in rates of lipid monitoring bydiagnosis. Of 21 subjects diagnosed with psychosis NOS,10 patients (47.6%) had baseline lipid panels ascertained.Of 40 subjects with bipolar disorder, 11 (27.5%) had lipidpanels at baseline. Only two (22.2%) of nine subjects withschizophrenia had baseline lipid panels at initiation ofsecond-generation antipsychotic. Lipid panels during indexhospitalization were statistically more prevalent amongmale subjects compared to their female counterparts(χ2=6.04, P=.01). There was a trend toward a differencebased on ethnicity and a difference based on prescriptionof olanzapine.

4. Discussion

This study is the first to examine the prevalence ofbaseline lipid monitoring in patients prescribed second-generation antipsychotics during their index hospitalizationand with varying psychiatric diagnoses. Overall, the rate oflipid monitoring was low. Almost 70% of psychiatricinpatients started for the first time on atypical antipsychoticshad no baseline lipid panel. Among those who did havebaseline lipid monitoring, 50% had at least one risk factor for

383W. Laundon et al. / General Hospital Psychiatry 34 (2012) 380–384

atherogenic dyslipidemia, with either a low HDL or elevatedtriglycerides. Ascertainment and treatment of modifiablemetabolic risk factors are indicated regardless of whetherincreased cardiovascular morbidity and mortality are related topreexisting baseline characteristics or are the subsequenteffects of antipsychotic medications prescribed to patientswith psychiatric disorders. Our results highlight a need forgreater vigilance in lipid monitoring and, more importantly,given that the culture in our institution is influenced by astrong CVD program and has included frequent discussions ofthe data, the ADA/APA and Mt. Sinai guidelines and theimportance of baseline lipid monitoring in patients prescribedsecond-generation antipsychotics, a need for systems andprograms to assure appropriate baseline lipid monitoring.

Compared with similar studies that found an approxi-mately 10% rate of baseline lipid monitoring in patientsprescribed second-generation antipsychotics [12–15], ourrates of lipid monitoring are relatively high at 31.4%. Thisdifference may be best accounted for by the setting in whichthe data were collected. The subjects were hospitalized in anacademic teaching hospital where recognition of CVD riskwith second-generation antipsychotics should be highercompared to patient data collected from commercial healthplans and Medicaid databases. Additionally, we found thatmales experienced a higher prevalence of lipid monitoring,whereas other studies have not found sex-based differences[13,14]. Lastly, subjects in this study were younger and first-time users of second-generation antipsychotics compared toolder and chronic users in other studies [12–15].

A number of findings warrant further discussion. Subjectsdiagnosed with bipolar disorder had a lipid monitoring rateof 27.5%. There are few studies that examine the prevalenceof lipid monitoring in patients with bipolar disorder; themajority of studies examining the risk of metabolicsyndrome are on patients diagnosed with schizophrenia.Two studies found bipolar patients to be at a high risk formetabolic syndrome but may be underevaluated for thiscondition [4,5]. Both studies conclude that health policiesimplemented for metabolic syndrome screening need toassess this risk in bipolar patients. The use of second-generation antipsychotics in patients with bipolar disorder israpidly increasing [17]. As majority of second-generationantipsychotics are now FDA approved for bipolar disorder,prescribers must be aware of the potential for adversemetabolic effects despite indication, and systems shouldinstitute safeguards to ensure appropriate monitoring.

An interesting finding is the statistically lower rates oflipid monitoring in females compared to males in this study.This finding is concerning since studies examining theprevalence of metabolic and cardiac risks find that femaleshave a greater risk compared not only to matched controls[8,9] but also to male patients with schizophrenia [8]. Twoother additional findings are intriguing, although it ispossible that our study was underpowered to detect astatistically significant difference on these comparisons.First, the differences in rates of lipid monitoring based on

ethnicity may have implications for future study and,possibly, interventions, particularly for African Americanssince they may be at an increased risk for dyslipidemia andmetabolic syndrome [18]. Additionally, there was anonsignificantly higher rate of baseline lipid ascertainmentamong subjects prescribed olanzapine vs. other second-generation antipsychotics. Over the last decade, olanzapinehas been cited as the worst offender of the routinelyprescribed second-generation antipsychotics, and this find-ing makes sense; however, other second-generation anti-psychotics also carry risks of metabolic derangements,weight gain and CHD [19–21]. Quetiapine was also foundto increase these measures [20,21]. In this study, lipidmonitoring of subjects prescribed quetiapine was the lowestof all the agents.

The findings from this study become even more importantwhen viewed in the context of studies examining lipidmetabolism changes in patients with first-episode psychosisprescribed second-generation antipsychotics. Two studies,one of 8-week duration [22] and one of 12-week duration[23], both reported statistically significant changes in lipidprofiles from baseline to study end, underscoring the fact thatchanges in lipid metabolism can occur rapidly in patientsexperiencing a first episode. A 1-year follow-up studyshowed statistically significant persistence in lipid metabo-lism changes [24]. Even more alarming is the fact that studieson drug-naive patients experiencing a first-episode psychosishad similar baseline lipid measures compared to healthycontrols, which strongly indicates an association betweensecond-generation antipsychotics and dyslipidemia [25,26].As a whole, this study, along with other first-episode studies,emphasizes the need for improved baseline monitoring as apreventive strategy to minimize future risk.

This study has several limitations. It is a retrospectivecohort with a small sample size, and prescribers chosemedications based on clinical impressions that may haveincluded risk assessment with respect to the metabolicsyndrome. The small sample size of this study may haveaffected the results of our statistical comparisons. Effort wasmade to identify only those patients who were newly startedon second-generation antipsychotics. Although all medicalrecords were reviewed by two study authors to ensure thatsubjects met the inclusion criteria, it is possible that patientsmay have been erroneously included or excluded from thisstudy. We did not have access to patient weights, which werenot documented in the electronic portion of the medicalrecord during the study period, so we are unable to commenton any association between weight or body mass index andprobability of lipid monitoring. Given the age range we setfor this study, we may have excluded older patients whowere first-time users of second-generation antipsychotics.There was no comparison group, such as a preguidelinegroup used in other studies, to measure the prevalence oflipid monitoring before and after the publication of the 2004consensus statement on this topic, although we did find ahigher rate of baseline lipid measurement than previously

384 W. Laundon et al. / General Hospital Psychiatry 34 (2012) 380–384

published studies demonstrating consistent rates of about10% both before and after the guidelines were published.

5. Conclusion

Although a higher rate of lipid monitoring was found inour study compared to other similar studies, the prevalenceof baseline lipid monitoring is low despite generalknowledge among psychiatric providers in our institutionregarding consensus statement guidelines for lipid monitor-ing. Further interventions to improve lipid monitoring inpatients newly prescribed atypical antipsychotics are neces-sary and overdue.

References

[1] Colton CW, Manderscheid RW. Congruencies in increased mortalityrates, years of potential life lost, and causes of death among publicmental health clients in eight states. Prev Chronic Dis 2006;3:1–13.

[2] Osby U, Correia N, Brandt L, Ekbom A, Sparen P. Mortality andcauses of death in schizophrenia in Stockholm county, Sweden.Schizophr Res 2000;45:21–8.

[3] Newcomer JW, Hennekens CH. Severe mental illness and risk ofcardiovascular disease. JAMA 2007;298:1794–6.

[4] Fiedorowicz JG, Palagummi NM, Forman-Hoffman VL, Miller DD,Haynes WG. Elevated prevalence of obesity, metabolic syndrome, andcardiovascular risk factors in bipolar disorder. Ann Clin Psychiatry2008;20:131–7.

[5] Sicras A, Rejas J, Navarro R, Serrat J, Blanca M. Metabolicsyndrome in bipolar disorder: a cross-sectional assessment of ahealth management organization database. Bipolar Disord 2008;10:607–16.

[6] Expert Panel on Detection, Evaluation, and Treatment of High BloodCholesterol in Adults. Executive summary of the third report of theNational Cholesterol Education Program (NCEP) expert panel ondetection, evaluation, and treatment of high blood cholesterol in adults(Adult Treatment Panel III). JAMA 2001;285:2486–97.

[7] Newcomer JW. Comparing the safety and efficacy of atypicalantipsychotics in psychiatric patients with comorbid medical illnesses.J Clin Psychiatry 2009;70(Suppl 3):30–6.

[8] McEvoy JP, Meyer JM, Goff DC, et al. Prevalence of the metabolicsyndrome in patients with schizophrenia: baseline results from theClinical Antipsychotic Trials of Intervention Effectiveness (CATIE)schizophrenia trial and comparison with nation estimates fromNHANES III. Schizophr Res 2005;80:19–32.

[9] Goff DC, Sullivan LM, McEvoy JP, et al. A comparison of ten-yearcardiac risk estimates in schizophrenia patients from the CATIE studyand matched controls. Schizophr Res 2005;80:45–53.

[10] American Diabetes Association, American Psychiatric Association,American Association of Clinical Endocrinologists, North American

Associates for the Study of Obesity. Consensus developmentconference on antipsychotic drugs and obesity and diabetes. DiabetesCare 2004;27:596–601.

[11] Marder SR, Essock SM, Miller AL, et al. Physical healthmonitoring of patients with schizophrenia. Am J Psychiatry 2004;161:1334–49.

[12] Morrato EH, Newcomer JW, Kamat S, Baser O, Harnett J, Cuffel B.Metabolic screening after the American Diabetes Association'sconsensus statement on antipsychotic drugs and diabetes. DiabetesCare 2009;32:1037–42.

[13] Morrato EH, Newcomer JW, Allen RR, Valuck RJ. Prevalence ofbaseline serum glucose and lipid testing in users of second-generationantipsychotic drugs: a retrospective, population-based study ofMedicaid claims data. J Clin Psychiatry 2008;69:316–22.

[14] Haupt DW, Rosenblatt LC, Kim E, Baker RA, Whitehead R,Newcomer JW. Prevalence and predictors of lipid and glucosemonitoring in commercially insured patients treated with second-generation antipsychotic agents. Am J Psychiatry 2009;166:345–53.

[15] Morrato EH, Druss B, Hartung DM, et al. Metabolic testing rates in 3state Medicaid programs after FDA warning and ADA/APArecommendations for second-generation antipsychotic drugs. ArchGen Psychiatry 2010;67:17–24.

[16] Latuda (Lurasidone HCL) [package insert]. Marlborough (MA):Sunovion Pharmaceuticals Inc.; 2010.

[17] Domino ME, Swartz MS. Who are new users of antipsychoticmedications? Psychiatr Serv 2008;59:507–14.

[18] Scneiderhan ME, Batscha CL, Rosen C. Assessment of a point-of-caremetabolic risk screening program in outpatients receiving antipsy-chotic agents. Pharmacotherapy 2009;29(8):975–87.

[19] Lieberman JA, Stroup TS, McEvoy JP, et al. Effectiveness ofantipsychotic drugs in patients with chronic schizophrenia. N Engl JMed 2005;353:1209–23.

[20] Daumit GL, Goff DC, Meyer JM. Antipsychotic effects on estimated10-year coronary heart disease risk in the CATIE schizophrenia study.Schizophr Res 2008;105:175–87.

[21] Meyer JM, Davis VG, McEvoy JP. Impact of antipsychotic treatmenton nonfasting triglycerides in the CATIE schizophrenia trial phase 1.Schizophr Res 2008;103:104–9.

[22] Wu RR, Zhao JP, Liu ZN, et al. Effects of typical and atypicalantipsychotics on glucose-insulin homeostasis and lipid metabolismin first-episode schizophrenia. Psychopharmacology 2006;186:572–8.

[23] Perez-Iglesias R, Crespo-Facorro B, Amado JA, et al. A 12-weekrandomized clinical trial to evaluate metabolic changes in drug-naïve,first-episode psychosis patients treated with haloperidol, olanzapine, orrisperidone. J Clin Psychiatry 2007;68:1733–40.

[24] Perez-Iglesias R, Mata I, Pelayo-Teran JM, et al. Glucose and lipiddisturbances after 1 year of antipsychotic treatment in a drug-naïvepopulation. Schizophr Res 2009;107:115–21.

[25] Sengupta S, Parrilla-Escobar MA, Klink R, et al. Are metabolic indicesdifferent between drug-naïve first-episode psychosis patients andhealthy controls? Schizophr Res 2008;102:329–36.

[26] Verma SK, SubramaniamM, Liew A, Poon LY. Metabolic risk factorsin drug-naïve patients with first-episode psychosis. J Clin Psychiatry2009;70:997–1000.