Abstract
Objective To assess the frequency and predictors of regular monitoring of metabolic parameters as recommended by the American Diabetes Association (ADA)/American Psychiatric Association (APA) guidelines in adolescents receiving antipsychotics compared with an untreated comparison group in a primary care setting.
Method A retrospective cohort study was conducted using an electronic medical record database in the USA from January 2004 to July 2009. The exposure group consisted of adolescents with a first prescription for second-generation antipsychotics (SGAs). The comparison group, without antipsychotics, was matched (3:1) to the exposed. Baseline and follow-up metabolic measurements were assessed and patients were categorised as being regularly monitored based on recommendations by the ADA/APA guidelines. Multivariate logistic regression was conducted to assess the predictors of regular monitoring, adjusting for demographic characteristics, baseline medications and medical conditions.
Results The frequency of regular monitoring of body mass index (BMI), blood pressure, total cholesterol and fasting blood glucose, as recommended by ADA/APA guidelines among antipsychotic users (25, 55, 2.4 and 1.7%) was low but significantly higher compared with the matched comparison group (9.5, 37.4, 0.8 and 0.7%, respectively) (P < 0.05). Overall, antipsychotic treatment was associated with 1.5- to 4.26-fold increase in the likelihood of metabolic monitoring compared with the comparison group (P < 0.05). Other predictors included oral antidiabetic use for BMI monitoring and dyslipidaemia for blood pressure, total cholesterol and fasting blood glucose.
Conclusion The majority of adolescents on antipsychotics were under-monitored for BMI, lipids and glucose levels. Antipsychotic users with pre-existing and newly diagnosed metabolic conditions were more likely to be regularly monitored for metabolic parameters.
Keywords: adolescents, antipsychotics, guidelines, metabolic monitoring, primary care
Introduction
Second-generation antipsychotics (SGAs) are used as first-line treatment for psychotic and non-psychotic mental disorders.1–3 The national trend of prescriptions for antipsychotics in adolescents has increased sharply in the USA.4 The American Diabetes Association (ADA)/American Psychiatric Association (APA) guidelines, published in February 2004, recommend monitoring of metabolic parameters such as weight and body mass index (BMI), blood pressure, fasting plasma glucose and fasting lipid profile for all patients receiving antipsychotic treatment regardless of age.5 In 2004, the FDA required all manufacturers of SGAs (atypical antipsychotics), such as Clozaril® (clozapine), Risperdal® (risperidone), Zyprexa® (olanzapine), Seroquel® (quetiapine), Geodon® (ziprasidone, and Abilify® (aripiprazole) to add a new warning to the drugs' labels about the increased risk of hyperglycaemia and diabetes.6 In 2010, the FDA changed prescribing information for olanzapine, with recommendations that physicians consider the increased potential for weight gain, hyperlipidaemia and long-term risks in adolescents using the medication.7
Previous studies have assessed the monitoring patterns of metabolic parameters among adults on antipsychotics, which have been relatively low in commercially insured populations,8,9 but relatively better monitoring rates in veterans.10 Only one study assessed monitoring of metabolic parameters using Medicaid data. The authors of this study reported that lipid and glucose testing was performed in 13 and 32% of children and adolescents occurring 30 days before to 180 days after initiating SGA treatment.11 However, the results of this study may not be generalisable to the national population as the Medicaid population is generally sicker than the commercially insured population.11
Most importantly, none of the studies to date have assessed the monitoring of BMI and blood pressure in adolescents on antipsychotics. The ADA/APA guidelines suggest careful review of baseline metabolic parameters before or as soon as initiating antipsychotic treatment,5 but none of the studies have assessed baseline monitoring of metabolic parameters in adolescents. Also, very little information exists regarding the predictors of regular monitoring of metabolic parameters in adolescents treated with SGAs. A study by Morrato et al found that having an emergency department visit, multiple mental health comorbidities, office visit and hospitalisation were the strongest predictors of monitoring glucose and lipid testing among SGA-treated children.11 However, this study did not include an untreated comparison group and predictors of BMI or blood pressure monitoring were not assessed. It remains unknown whether demographic characteristics, insurance type, concomitant medication use and medical conditions in addition to antipsychotic use are associated with regular monitoring of metabolic parameters among adolescents.
The purpose of the current study was to assess the frequency of regular monitoring of metabolic parameters as recommended by ADA/APA guidelines in adolescents receiving SGAs compared with an age- and gender-matched comparison group in a predominantly primary care setting. The secondary purpose of the study was to assess the predictors of regular monitoring of metabolic parameters among adolescents.
Methods
Data source
The General Electric (GE) Centricity electronic medical record (EMR) database was used for this study. The GE EMR database has data on ∼ 10 million patients. The EMR database is de-identified and HIPAA-compliant. It contains longitudinal clinical patient data including, but not limited to, demographic information, vital signs, laboratory orders and results, medication list entries and prescriptions, and diagnoses or problems. The EMR data are submitted voluntarily from physician groups across the country. A variety of practice types is represented ranging from solo practitioners to community clinics, to academic medical centres and large integrated delivery networks. Previous studies have used the GE EMR database in health outcomes studies.12–14
Study design
A retrospective cohort study design was used for this study. The retrospective cohort consisted of adolescents 12–19 years old. Within this cohort, adolescents with a prescription for antipsychotics were considered the exposed group. Monitoring of this group was compared with what the guidelines recommend, controlling for sporadic monitoring in the population by including the healthy comparison group without prescription for antipsychotics.
Study population
Patients were eligible for inclusion in this study if they were 12–19 years old and had at least one activity in the EMR database between January 2004 and July 2009. Patients in the exposure group had at least one prescription for any SGA, whereas the comparison group was eligible for inclusion if they had no prescription for any antipsychotic agent during the study period. The date of patients' first prescription for any SGA after a pre-index period of 180 days in the GE EMR database was denoted as the index date for the exposed group. Both the exposed and comparison groups had at least one documented physician visit ≥ 180 days before and 395 days after the index date (Figure 1). The antipsychotic drug prescribed on the index date was designated as the index drug. Patients were categorised as exposed to individual SGAs by their index drug. Among the exposed individuals, analysis was limited to patients who remained on the index antipsychotic drug throughout the follow-up period (monotherapy). Both the exposed and comparison groups were followed for a period of 395 days from the index date. Stratified random sampling with 3:1 matching was used and the comparison group was matched to the exposed group based on age, gender and month of visit to the index antipsychotic prescription date.
Figure 1.
Flowchart of patient selection
EMRs for eligible patients were examined for the presence of baseline and follow-up metabolic measurements such as BMI, blood pressure, total cholesterol and fasting blood glucose levels. Metabolic measurements that occurred within 15 days before or after the index date were classified as baseline measurements. Metabolic measurements occurring within the 16th day after the index date to the end of follow-up (index date + 395 days) were classified as follow-up measurements.
Demographic characteristics including age, gender, insurance type and region were assessed during the pre-index period. Baseline medication use of beta blockers, oral antidiabetic agents associated with weight gain (such as human insulin, sulphonylureas and thiazolidinediones) or loss (such as incretin mimetic agents and biguanides), antidepressants, anticonvulsants and corticosteroids that may influence metabolic monitoring was assessed during the pre-index period. Baseline medical conditions such as dyslipidaemia, hypertension, obesity, hypothyroidism, type 2 diabetes, bipolar disorder, schizophrenia, depression and other mental illness (such as psychoses, neurotic disorders and mental retardation) that may influence metabolic monitoring were identified, using ICD-9 codes, in both the preindex and follow-up periods. Newly diagnosed medical conditions were defined as those that were newly identified during the follow-up period for which there was no evidence in the pre-index period.
ADA/APA monitoring guidelines were used as a reference standard to categorise patients as being monitored regularly or irregularly for metabolic parameters.5 The guidelines state that including baseline measurement BMI should be monitored at least seven times, fasting glucose and blood pressure at least three times, and lipid levels at least twice during 1-year follow-up period. Based on these criteria, patients in the exposure and comparison groups were categorised as being regularly monitored if the number of measurements was ≥ 7 for BMI, ≥ 3 for blood pressure, ≥ 2 for total cholesterol and ≥ 3 for fasting blood glucose during the 395 days of the follow-up period.
Analysis
Tests of proportions were used to evaluate differences between the antipsychotic group and comparison group in baseline demographics, baseline medications, baseline medical conditions and frequency of baseline and regular monitoring of metabolic parameters. Wilcoxon rank sum test was used to compare the mean number of metabolic measurements between the antipsychotic group and comparison group. Separate multivariate logistic regression analyses were conducted to assess the likelihood of regular monitoring of BMI and blood pressure. The frequency of monitoring of lipids and fasting blood glucose was extremely low, since they are performed less frequently than blood pressure or BMI, for logistic regression modelling. To account for rare events, separate multivariate logistic regression for rare events was conducted to assess the likelihood of regular monitoring of total cholesterol and fasting blood glucose.15,16 Regression analyses were conducted controlling for the individual SGAs, demographic characteristics, baseline medications and baseline and newly diagnosed medical conditions thought to be related to monitoring of metabolic parameters. All analyses were performed using STATA Version 10.0 (StataCorp. 2007. Stata Statistical Software: Release 10. StataCorp LP: College Station, TX).
Results
A total of 7967 patients aged 12–19 years, with at least one prescription for a single type of SGA were identified in the GE EMR database between January 2004 and July 2009. Of these, the final sample consisted of 3038 patients with at least one documented physician visit 180 days before the index date and at least one documented physician visit 395 days after the index date. The comparison group consisted of 9114 patients randomly matched to the exposure group. See Figure 1 for details of patient selection.
Table 1 provides the comparison of demographics, baseline medication use and medical conditions among patients in the exposure group and the matched comparison group. Patients in the exposure group had a higher percentage of concurrent prescriptions for beta blockers, oral antidiabetic medications associated with weight loss, antidepressants and anticonvulsants (P < 0.01), whereas the comparison group had a higher proportion of oral antidiabetic medications associated with weight gain and corticosteroids (P < 0.01) during the study period. The exposed group had a significantly higher proportion of baseline medical conditions including obesity, type 2 diabetes and psychiatric conditions such as bipolar disorder, schizophrenia, depression and other mental illness compared with the comparison group (P < 0.05).
Table 1.
Demographics, medication use, and medical conditions among adolescents in the exposure and comparison groups, 2004 to 2009
| Exposure group (n = 3038) | Comparison group (n = 9114) | P | |||
|---|---|---|---|---|---|
| n | % | n | % | ||
| Age (Mean, SD) | 15.53 | 2.22 | 15.53 | 2.22 | 1.000 |
| Gender | |||||
| Males | 1647 | 54.21 | 4,941 | 54.21 | 1.000 |
| Females | 1391 | 45.79 | 4,173 | 45.79 | 1.000 |
| Race | |||||
| White | 1033 | 34.00 | 2,532 | 27.79 | < 0.001 |
| Black | 103 | 3.39 | 576 | 6.32 | < 0.001 |
| Hispanic | 39 | 1.28 | 273 | 3.00 | < 0.001 |
| Other | 35 | 1.15 | 140 | 1.54 | 0.124 |
| Unknown | 1828 | 60.17 | 5,591 | 61.36 | 0.251 |
| Region | |||||
| Northeast | 729 | 24.00 | 1,899 | 20.84 | < 0.001 |
| South | 1176 | 38.71 | 2,709 | 29.72 | < 0.001 |
| Midwest | 542 | 17.84 | 2,628 | 28.83 | < 0.001 |
| West | 591 | 19.45 | 1,878 | 20.61 | 0.172 |
| Insurance type | |||||
| Commercial | 1231 | 40.52 | 4,495 | 49.32 | < 0.001 |
| Medicare | 565 | 18.60 | 686 | 7.53 | < 0.001 |
| Medicaid | 48 | 1.58 | 19 | 0.21 | < 0.001 |
| Self-pay | 65 | 2.14 | 215 | 2.36 | 0.485 |
| Other/unknown | 1129 | 37.16 | 3,699 | 40.59 | 0.001 |
| Baseline medication use | |||||
| Beta blockers | 38 | 1.25 | 27 | 0.3 | < 0.001 |
| Antidiabetics weight gain | 12 | 0.39 | 69 | 0.76 | 0.034 |
| Antidiabetics weight loss | 21 | 0.69 | 29 | 0.32 | 0.005 |
| Antidepressants | 56 | 1.84 | 38 | 0.42 | < 0.001 |
| Anticonvulsants | 297 | 9.78 | 66 | 0.72 | < 0.001 |
| Corticosteroids | 77 | 2.53 | 363 | 3.98 | < 0.001 |
| Baseline conditions | |||||
| Dyslipidaemia | 22 | 0.72 | 61 | 0.67 | 0.751 |
| Hypertension | 12 | 0.39 | 28 | 0.31 | 0.465 |
| Obesity | 111 | 3.65 | 170 | 1.87 | < 0.001 |
| Hypothyroidism | 12 | 0.39 | 20 | 0.22 | 0.102 |
| Type 2 diabetes | 130 | 4.28 | 71 | 0.78 | < 0.001 |
| Bipolar disorder | 361 | 11.88 | 23 | 0.25 | < 0.001 |
| Schizophrenia | 25 | 0.82 | 2 | 0.02 | < 0.001 |
| Depression | 119 | 3.92 | 25 | 0.27 | < 0.001 |
| Other mental illnessa | 1168 | 38.45 | 669 | 7.34 | < 0.001 |
a Other mental illness was identified using ICD-9 codes 290 to 294 and 297 to 319.
Frequency of baseline monitoring in the exposed and comparison groups
Patients on quetiapine had higher rates of baseline monitoring of BMI, blood pressure, total cholesterol and fasting blood glucose than the comparison group (P < 0.01). Patients on risperidone and olanzapine had higher rates of baseline monitoring of BMI and blood pressure (P < 0.01), whereas only patients on risperidone had higher rates of fasting blood glucose monitoring compared with the comparison group (P < 0.01). Patients on aripiprazole had higher baseline monitoring for BMI, total cholesterol and fasting blood glucose (P < 0.05), and patients on ziprasidone had a higher frequency of monitoring for total cholesterol (P < 0.05) and fasting blood glucose compared with the comparison group.
Metabolic measurements in the exposed and comparison groups
Adolescents on individual antipsychotics had a lower mean number of metabolic measurements relative to those recommended by ADA/APA guidelines, except for blood pressure which was higher (Table 2). Adolescents on individual antipsychotics had a significantly higher mean number of BMI, blood pressure, total cholesterol and fasting blood glucose measurements compared with the group not prescribed SGAs (P < 0.01).
Table 2.
Mean number of metabolic measurements among adolescents on antipsychotics compared to comparison group within one year relative to ADA/APA recommendations
| BMI | Blood pressure | Total cholesterol | Fasting blood glucose | |
|---|---|---|---|---|
| ADA/APA guideline recommendations | ≥ 7 | ≥ 3 | ≥ 2 | ≥ 3 |
| Comparison group mean (SD) | 2.53 (3.28) | 2.54 (3.28) | 0.07 (0.33) | 0.14 (0.60) |
| Exposure group (all SGAs) mean (SD) | 4.74 (6.28) | 3.86 (3.90) | 0.17 (0.46) | 0.27 (0.75) |
| Aripiprazole mean (SD) | 5.12 (6.56) | 3.82 (4.13) | 0.26 (0.54) | 0.28 (0.71) |
| Olanzapine mean (SD) | 5.19 (8.25) | 4.18 (4.00) | 0.11 (0.36) | 0.27 (0.74) |
| Risperidone mean (SD) | 4.49 (5.52) | 3.33 (3.37) | 0.16 (0.44) | 0.26 (0.83) |
| Quetiapine mean (SD) | 4.64 (6.20) | 4.23 (4.06) | 0.14 (0.43) | 0.25 (0.68) |
| Ziprasidone mean (SD) | 4.65 (6.04) | 4.17 (4.20) | 0.26 (0.60) | 0.43 (0.90) |
* Wilcoxon rank sum test was used and P < 0.01 for comparison of number of BMI, blood pressure, total cholesterol and fasting blood glucose measurements among the exposure group compared with the comparison group and individual antipsychotic groups compared with the comparison group.
Frequency of regular monitoring among the exposed and comparison groups
There was evidence of adherence to monitoring for all metabolic parameters in the EMR for < 1% of patients on antipsychotics. Regular monitoring of all four metabolic parameters, as recommended by the ADA/APA guidelines, was extremely low and not significantly different between the exposed and comparison groups (P > 0.05), except for patients on aripiprazole in whom the frequency was slightly higher (P < 0.01). The frequency of regular monitoring of BMI, blood pressure, total cholesterol and fasting blood glucose was significantly higher among adolescents on aripiprazole, risperidone, quetiapine and ziprasidone compared with the comparison group (P < 0.05) (Table 3). Adolescents on olanzapine had a higher frequency of regular monitoring of fasting blood glucose, but the monitoring of total cholesterol was not significantly different compared with the comparison group (P = 0.478).
Table 3.
Frequency of regular monitoring of metabolic parameters as recommended by ADA/APA guidelines among adolescents in the exposure and comparison groups, 2004 to 2009
| Total | BMI | Blood pressure | Total cholesterol | Fasting blood glucose | |||||
|---|---|---|---|---|---|---|---|---|---|
| n | % | n | % | n | % | n | % | ||
| Comparison group | 9114 | 862 | 9.46 | 3411 | 37.43 | 69 | 0.76 | 65 | 0.71 |
| Exposure group (all SGAs) | 3038 | 765 | 25.18 | 1668 | 54.90 | 74 | 2.44 | 53 | 1.74 |
| Aripiprazole | 642 | 168 | 26.17 | 339 | 52.80 | 23 | 3.58 | 10 | 1.56 |
| Olanzapine | 262 | 61 | 23.28 | 163 | 62.21 | 3* | 1.15 | 6 | 2.29 |
| Risperidone | 931 | 231 | 24.81 | 456 | 48.98 | 18 | 1.93 | 13 | 1.40 |
| Quetiapine | 1090 | 272 | 24.95 | 648 | 59.45 | 23 | 2.11 | 19 | 1.74 |
| Ziprasidone | 113 | 33 | 29.20 | 62 | 54.87 | 7 | 6.19 | 5 | 4.42 |
* P > 0.05, test of proportions was used to compare proportion of patients with baseline metabolic parameters on individual antipsychotic group to the comparison group.
Predictors of BMI and blood pressure monitoring
Patients on risperidone had the highest odds ratio (OR) of 2.65 followed by ziprasidone (OR, 2.62; 95% confidence interval [CI], 1.68–4.09), olanzapine (OR, 2.60; 95% CI, 1.89–3.57), aripiprazole (OR, 2.46; 95% CI, 1.89–3.57) and quetiapine (OR, 2.35; 95% CI, 1.98–2.80) for regular monitoring of BMI compared with the comparison group (Table 4). Similar to BMI monitoring, being on antipsychotic treatment was significantly associated with a higher likelihood of regular monitoring of blood pressure, except for patients on ziprasidone. Although statistically significant OR values were observed for regular monitoring of blood pressure, the likelihood of regular monitoring between the antipsychotic agents was not significantly different from each other because the confidence intervals for individual antipsychotics overlapped. Compared with males, females were 1.52 and 1.44 times more likely to be regularly monitored for BMI and blood pressure, respectively. Patients on medications such as oral antidiabetic agents, corticosteroids or with medical conditions such as obesity, type 2 diabetes and other mental illness were more likely to be regularly monitored for BMI and blood pressure. Given usual clinical practice, one might assume that adolescents are monitored for BMI and blood pressure at every physician's office visit. According to the results, that may not necessarily be true for adolescents with dyslipidaemia. Patients with dyslipidaemia were highly likely to be monitored regularly for blood pressure, but this condition was not a significant predictor of monitoring BMI. Newly diagnosed medical conditions such as dyslipidaemia and type 2 diabetes significantly increased monitoring of BMI and blood pressure during the follow-up period. The proportion of case patients experiencing regular monitoring of BMI and blood pressure increased by 24 and 11% each year from 2004 to 2009, respectively, indicating an increase in the monitoring after the guidelines were published in 2004.
Table 4.
Adjusted odds of regular monitoring of BMI, blood pressure, total cholesterol, and fasting blood glucose among exposed adolescents compared with the comparison group, 2004 to 2009
| BMI | Blood pressure | Total cholesterol | Fasting blood glucose | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| OR | 95% CI | P | OR | 95% CI | P | OR | 95% CI | P | OR | 95% CI | P | |||||
| Age | 0.97 | 0.94 | 0.99 | 0.015 | 1.07 | 1.05 | 1.09 | 0.000 | 1.04 | 0.96 | 1.12 | 0.305 | 1.04 | 0.95 | 1.15 | 0.360 |
| Females (ref = Males) | 1.52 | 1.36 | 1.70 | 0.000 | 1.44 | 1.33 | 1.55 | 0.000 | 0.63 | 0.43 | 0.92 | 0.016 | 1.27 | 0.85 | 1.89 | 0.239 |
| Region (ref = Northeast) | ||||||||||||||||
| Southeast | 1.47 | 1.27 | 1.71 | 0.000 | 0.86 | 0.78 | 0.96 | 0.005 | 0.74 | 0.48 | 1.13 | 0.161 | 0.95 | 0.56 | 1.62 | 0.855 |
| Midwest | 0.95 | 0.80 | 1.12 | 0.513 | 0.98 | 0.88 | 1.10 | 0.780 | 0.65 | 0.39 | 1.07 | 0.090 | 0.69 | 0.38 | 1.25 | 0.223 |
| West | 0.68 | 0.57 | 0.82 | 0.000 | 0.89 | 0.79 | 1.00 | 0.042 | 0.68 | 0.40 | 1.16 | 0.154 | 1.04 | 0.61 | 1.78 | 0.875 |
| Insurance Type (ref = Commercial) | ||||||||||||||||
| Medicare | 1.16 | 0.98 | 1.38 | 0.078 | 1.06 | 0.93 | 1.21 | 0.386 | 0.77 | 0.45 | 1.32 | 0.337 | 0.86 | 0.41 | 1.80 | 0.686 |
| Medicaid | 1.59 | 0.89 | 2.85 | 0.119 | 1.65 | 0.97 | 2.81 | 0.065 | 1.71 | 0.52 | 5.65 | 0.381 | 2.72 | 0.77 | 9.63 | 0.122 |
| Self-pay | 1.19 | 0.83 | 1.70 | 0.349 | 0.99 | 0.77 | 1.28 | 0.967 | 0.48 | 0.07 | 3.58 | 0.477 | 0.60 | 0.08 | 4.41 | 0.615 |
| Other/unknown | 0.81 | 0.72 | 0.92 | 0.001 | 0.91 | 0.84 | 0.99 | 0.021 | 0.77 | 0.52 | 1.13 | 0.181 | 1.08 | 0.71 | 1.64 | 0.722 |
| Treatment group (ref = Comparison group) | ||||||||||||||||
| Aripiprazole | 2.46 | 2.00 | 3.04 | 0.000 | 1.53 | 1.28 | 1.82 | 0.000 | 4.21 | 2.50 | 7.08 | 0.000 | 1.53 | 0.72 | 3.24 | 0.271 |
| Olanzapine | 2.60 | 1.89 | 3.57 | 0.000 | 2.22 | 1.70 | 2.89 | 0.000 | 1.65 | 0.49 | 5.54 | 0.420 | 2.96 | 1.21 | 7.25 | 0.018 |
| Risperidone | 2.65 | 2.21 | 3.19 | 0.000 | 1.51 | 1.31 | 1.76 | 0.000 | 2.48 | 1.45 | 4.26 | 0.001 | 1.82 | 0.94 | 3.53 | 0.077 |
| Quetiapine | 2.35 | 1.98 | 2.80 | 0.000 | 1.88 | 1.63 | 2.16 | 0.000 | 2.52 | 1.50 | 4.25 | 0.001 | 1.78 | 0.99 | 3.17 | 0.052 |
| Ziprasidone | 2.62 | 1.68 | 4.09 | 0.000 | 1.41 | 0.95 | 2.09 | 0.085 | 7.34 | 3.11 | 17.34 | 0.000 | 4.26 | 1.33 | 13.65 | 0.015 |
| Medications | ||||||||||||||||
| Beta Blockers | 1.55 | 0.85 | 2.81 | 0.149 | 1.45 | 0.86 | 2.43 | 0.163 | 3.07 | 0.87 | 10.79 | 0.08 | 2.18 | 0.40 | 11.79 | 0.365 |
| OAD Weight Gain | 3.65 | 2.19 | 6.08 | 0.000 | 2.88 | 1.79 | 4.63 | 0.000 | 3.68 | 1.00 | 13.48 | 0.049 | 4.44 | 1.17 | 16.88 | 0.029 |
| OAD Weight Loss | 2.99 | 1.57 | 5.71 | 0.001 | 1.48 | 0.80 | 2.74 | 0.214 | 4.60 | 1.17 | 18.09 | 0.029 | 3.72 | 0.98 | 14.07 | 0.053 |
| Antidepressants | 1.57 | 0.97 | 2.55 | 0.068 | 1.31 | 0.85 | 2.01 | 0.223 | 1.49 | 0.34 | 6.50 | 0.597 | 1.64 | 0.34 | 7.86 | 0.539 |
| Anticonvulsants | 1.15 | 0.87 | 1.50 | 0.324 | 1.10 | 0.88 | 1.37 | 0.424 | 1.62 | 0.83 | 3.15 | 0.158 | 1.25 | 0.55 | 2.86 | 0.600 |
| Corticosteroids | 1.40 | 1.07 | 1.85 | 0.016 | 1.34 | 1.10 | 1.64 | 0.004 | 1.21 | 0.48 | 3.02 | 0.686 | 1.35 | 0.53 | 3.46 | 0.533 |
| Baseline conditions | ||||||||||||||||
| Dyslipidaemia | 1.12 | 0.60 | 2.08 | 0.731 | 1.60 | 1.01 | 2.52 | 0.044 | 9.78 | 4.15 | 23.03 | 0.000 | 7.83 | 2.95 | 20.78 | 0.000 |
| Hypertension | 1.89 | 0.86 | 4.14 | 0.115 | 1.43 | 0.72 | 2.85 | 0.306 | 0.65 | 0.03 | 15.12 | 0.787 | ||||
| Obesity | 1.57 | 1.16 | 2.12 | 0.003 | 1.53 | 1.19 | 1.98 | 0.001 | 0.61 | 0.21 | 1.80 | 0.373 | 0.70 | 0.18 | 12.76 | 0.611 |
| Hypothyroidism | 1.38 | 0.57 | 3.32 | 0.471 | 0.95 | 0.45 | 1.97 | 0.885 | 1.67 | 0.27 | 10.24 | 0.578 | 1.72 | 0.19 | 15.73 | 0.630 |
| Schizophrenia | 0.93 | 0.36 | 2.41 | 0.884 | 0.94 | 0.43 | 2.07 | 0.881 | ||||||||
| Bipolar Disorder | 0.98 | 0.76 | 1.27 | 0.896 | 1.15 | 0.91 | 1.44 | 0.239 | 1.29 | 0.64 | 12.57 | 0.477 | 1.13 | 0.49 | 12.63 | 0.774 |
| Depression | 1.00 | 0.66 | 1.51 | 0.997 | 1.38 | 0.96 | 1.97 | 0.079 | 0.56 | 0.13 | 12.46 | 0.439 | 0.68 | 0.10 | 14.57 | 0.693 |
| Type 2 diabetes | 1.84 | 1.32 | 2.55 | 0.000 | 1.36 | 1.00 | 1.85 | 0.048 | 0.90 | 0.31 | 12.63 | 0.852 | 1.03 | 0.37 | 12.86 | 0.949 |
| Mental illnessa | 1.52 | 1.32 | 1.75 | 0.000 | 1.52 | 1.36 | 1.70 | 0.000 | 0.98 | 0.64 | 11.50 | 0.931 | 1.42 | 0.90 | 12.25 | 0.133 |
| Incident conditions | ||||||||||||||||
| Dyslipidaemia | 2.16 | 1.38 | 3.39 | 0.001 | 3.17 | 2.02 | 4.98 | 0.000 | 11.85 | 6.09 | 23.05 | 0.000 | 7.81 | 3.62 | 16.86 | 0.000 |
| Type 2 diabetes | 2.91 | 2.09 | 4.04 | 0.000 | 2.75 | 1.95 | 3.87 | 0.000 | 1.78 | 0.79 | 14.01 | 0.167 | 4.05 | 1.92 | 18.52 | 0.000 |
| Year of index date | 1.24 | 1.19 | 1.29 | 0.000 | 1.11 | 1.08 | 1.14 | 0.000 | 1.13 | 0.99 | 11.29 | 0.076 | 1.31 | 1.14 | 11.50 | 0.000 |
OR, odds ratio; 95% CI, 95% confidence interval.
a Other mental illness was identified using ICD-9 codes 290 to 294 and 297 to 319. The logistic regression analysis was adjusted for age, gender, insurance type, individual antipsychotic medications, baseline medications, baseline medical conditions, incident medical conditions and year of index date.
Predictors of lipid and glucose monitoring
Antipsychotic treatment was significantly associated with regular monitoring of total cholesterol and fasting blood glucose among adolescents. The likelihood of being regularly monitored for total cholesterol was 2.48 (95% CI, 1.45–4.26) for risperidone, 2.52 (95% CI, 1.50–4.25) for quetiapine, 4.21 (95% CI, 2.50–7.08) for aripiprazole, and 7.34 (95% CI, 3.11–17.34) for ziprasidone compared with the comparison group (Table 4). Patients on olanzapine were three times more likely to be regularly monitored for fasting blood glucose compared with the comparison group. Also, patients on ziprasidone were four times more likely to be regularly monitored for fasting blood glucose. Adolescents on oral antidiabetic agents were associated with increased monitoring of total cholesterol and fasting blood glucose. Baseline diagnosis of dyslipidaemia and newly diagnosed dyslipidaemia were associated with increased monitoring of both total cholesterol and fasting blood glucose, but only adolescents with newly diagnosed type 2 diabetes were associated with increased monitoring of fasting blood glucose. The proportion of case patients experiencing regular monitoring of total cholesterol remained unchanged during the study period but regular monitoring of fasting blood glucose increased by 31% every year from 2004 to 2009.
Discussion
The GE EMR database used in this study is a predominantly primary care physician network database with a third of data representing care delivered by specialists. Therefore, any trends observed in the data may be largely representative of primary care physicians' practice. Primary care physicians are the first line of defence in diagnosing and treating psychiatric conditions. Many adolescents receive treatment from primary care physicians because of the dearth of child and adolescent psychiatrists. There is growing public health concern regarding the metabolic effects of these drugs in the adolescent population17,18 and regular monitoring of metabolic parameters is recommended.5 Overall, 55% of the adolescents on antipsychotics in this study were being regularly monitored for blood pressure, 25% for BMI and ∼ 2% for lipids and glucose, as recommended by ADA/APA guidelines in this study. The frequency of regular monitoring observed in this study was higher for adolescents on antipsychotics than the untreated comparison group. However, the incremental difference among the exposed and unexposed groups was extremely low. Specifically, the monitoring of lipids and glucose was < 1% higher among adolescents with a prescription for antipsychotics compared with the untreated cohort. Such low numbers are shocking given the FDA requirement of adding a warning to drug labels stating the increased risk of hyperglycaemia or diabetes associated with antipsychotics. The mean metabolic measurements and frequency of regular monitoring among adolescents on individual antipsychotics did not appear to vary drastically even though olanzapine is associated with the highest risk of diabetes. Most importantly, < 1% of exposed patients experienced monitoring of all four metabolic parameters, as suggested by the guidelines, which was not significantly different from the untreated comparison group. The reasons for low monitoring remain unknown and future studies should ascertain the challenges associated with low monitoring among adolescents.
The frequency of regular monitoring of lipids and glucose was extremely low compared with other published studies.11,19 The frequency might be lower because this study's definition of regular monitoring was stringent compared with previous studies. We defined regular monitoring of metabolic parameters as recommended by ADA/APA guidelines within 1 year of antipsychotic prescription, whereas previous studies assessed monitoring of at least one lipid or glucose including the baseline measurement within 6 months.
Antipsychotic treatment in adolescents was associated with increased monitoring of metabolic parameters. However, no differences were observed in the likelihood of regular monitoring of metabolic parameters across individual SGAs with the exception of fasting blood glucose. Adolescents on olanzapine or ziprasidone were being monitored regularly for fasting blood glucose, while no differences were observed in the likelihood of regular monitoring among other SGAs compared with the comparison group. Even though antipsychotic treatment was associated with regular monitoring, the frequency of regular monitoring of metabolic parameters, as recommended by ADA/APA guidelines, among adolescents on antipsychotics was low.
The strongest predictors of regular monitoring of BMI were oral antidiabetic use and new diagnosis of type 2 diabetes. Both of these were stronger predictors of BMI monitoring than antipsychotic use. Adolescents with diagnosis of type 2 diabetes are encouraged to lose weight along with healthy eating and regular monitoring of blood glucose levels. Also, adolescents with type 2 diabetes may be visiting their primary care physician more often than others, resulting in an increased monitoring of BMI compared with others. Pre-existing and new diagnosis of dyslipidaemia was found to be the strongest predictors of metabolic monitoring of blood pressure, total cholesterol and fasting blood glucose with higher OR values than antipsychotic use. This finding is consistent with the results of the study by Haupt et al conducted in adults.8
Although antipsychotic use was not the strongest predictor of metabolic monitoring, adolescents on antipsychotics with pre-existing conditions such as diabetes and dyslipidaemia had higher likelihood of metabolic monitoring compared with the comparison group with pre-existing conditions. Adolescents on antipsychotics with pre-existing chronic conditions may be monitored more regularly to manage the inherent chronic condition which may worsen because of the increased risk of cardiometabolic adverse effects linked to antipsychotic treatment. Baseline diagnosis of obesity and mental illness associated with increased metabolic monitoring was consistent with previously published studies in children and adults.8,11,19
This study has several strengths. This study is the first to assess the monitoring patterns of metabolic parameters in adolescents (12–19 years old) using a predominantly primary care large national EMR database. Previous studies have been conducted in the Medicaid population. Another strength of this study is that actual clinical measures such as BMI values, blood pressure readings, total cholesterol values and fasting blood glucose levels in the GE EMR database were used to assess monitoring patterns in children and adolescents on antipsychotics. Previous studies have used current procedural terminology (CPT) codes and not actual clinical measures to assess monitoring patterns in children and adolescents on antipsychotics.11,19
As with any research study there are several limitations. One of the most important limitations of this study is that prescriptions in an EMR database are tracked by prescription orders and medication lists and not by actual prescriptions filled at the pharmacy. We cannot be entirely sure if patients are filling the prescriptions or taking prescriptions prescribed to them, which may have resulted in the misclassification of exposure. Another limitation of the database is that it is predominantly a primary care physician network database and, therefore, healthcare received outside of the primary care setting may not be captured in the database. This dataset likely includes antipsychotic users that are relatively less severe compared with those patients seeking care from psychiatrists. As a result, the results may be biased towards the null since we are automatically excluding sicker or more severe patients.
Another limitation of the study is the missing BMI, lipids, blood glucose and blood pressure values in the GE EMR database regardless of antipsychotic treatment. This limitation may have resulted in a misclassification bias which was assumed to have been equally distributed between the antipsychotic group and the comparison group.
Conclusions
Adolescents on antipsychotics are not being monitored according to guidelines published by the ADA/APA although their metabolic parameters are being monitored moderately more frequently than the age- and gender-matched comparison group. In particular, the majority of adolescents treated with antipsychotics remain under-monitored for BMI, lipids and glucose. Without regular monitoring of metabolic parameters, adolescents on antipsychotics are at higher risk of growing into adulthood with abnormal weight and other metabolic parameters that impact adult obesity and its cardiovascular outcomes.20–23 Clinicians need to be more proactive in monitoring metabolic parameters among all adolescents receiving antipsychotic prescriptions and not just those with pre-existing metabolic conditions. Strategies to improve awareness of ADA/APA guidelines and improve monitoring of metabolic parameters in adolescents on SGAs need to be developed.
ACKNOWLEDGEMENT
The authors would like to thank Xiangyang Ye's assistance in stratified random sampling and matching used in this study.
PRESENTATION
The results of this study have been accepted for podium presentation at the 165th American Psychiatric Association Annual meeting in Philadelphia, PA in May 2012.
CONFLICTS OF INTEREST
This study was funded in part from an unrestricted educational grant from Bristol-Myers Squibb. The authors report no conflict of interest except for authors SG and DB who are employees of the Pharmacotherapy Outcomes Research Center (PORC) which received funding from Bristol-Myers Squibb to conduct this study.
Contributor Information
Sameer R Ghate, Research Assistant Professor.
Christina A Porucznik, Assistant Professor, University of Utah, Salt Lake City, UT, USA.
Qayyim Said, Assistant Professor, University of Arkansas for Medical Sciences, Little Rock, AR, USA.
Mia Hashibe, Assistant Professor, University of Utah, Salt Lake City, UT, USA.
Elizabeth Joy, Medical Director, Clinical Outcomes Research; Adjunct Professor, Family and Preventive Medicine, University of Utah, School of Medicine, Salt Lake City, UT, USA.
Diana I Brixner, Professor and Chair, Dept of Pharmacotherapy, Executive Director, Outcomes Research Centre, College of Pharmacy; Director of Outcomes, Program in Personalized Health Care, University of Utah, Salt Lake City, UT, USA.
REFERENCES
- 1.Lehman AF, Lieberman JA, Dixon LB, et al. Practice guideline for the treatment of patients with schizophrenia, second edition. American Journal of Psychiatry 2004;161(2 Suppl):1–56 [PubMed] [Google Scholar]
- 2.National Institute for Health and Clinical Excellence Bipolar Disorder: the management of bipolar disorder in adults, children and adolescents, in primary and secondary care: quick reference guide. National Institute for Health and Clinical Excellence: London, 2006 [PubMed] [Google Scholar]
- 3.Davidson JR. Major depressive disorder treatment guidelines in America and Europe. Journal of Clinical Psychiatry 2010;71(Suppl E1):e04 DOI: 10.4088/JCP.9058se1c.04gry [DOI] [PubMed] [Google Scholar]
- 4.Olfson M, Blanco C, Liu L, Moreno C, Laje G. National trends in the outpatient treatment of children and adolescents with antipsychotic drugs. Archives of General Psychiatry 2006;63(6):679–85 DOI:63/6/679 [pii] 10.1001/archpsyc.63.6.679 [DOI] [PubMed] [Google Scholar]
- 5.Consensus development conference on antipsychotic drugs and obesity and diabetes. Diabetes Care 2004;27(2):596–601 [DOI] [PubMed] [Google Scholar]
- 6.US Food and Drug Administration Warning about hyperglycemia and atypical antipsychotic drugs. In: FDA Safety News; Show#28, 2004. Available at: www.accessdata.fda.gov/psn/transcript.cfm?show=28 (accessed 08/02/11).
- 7.FDA USFaDA Zyprexa (olanzapine): use in adolescents. 2010. Available at: www.fda.gov/Safety/MedWatch/SafetyInformation/SafetyAlertsForHumanMedicalProducts/ucm198402.htm (accessed 03/11/2010).
- 8.Haupt DW, Rosenblatt LC, Kim E, Baker RA, Whitehead R, Newcomer JW. Prevalence and predictors of lipid and glucose monitoring in commercially insured patients treated with second-generation antipsychotic agents. American Journal of Psychiatry 2009;166(3):345–53 DOI:appi.ajp.2008.08030383 [pii] 10.1176/appi.ajp.2008.08030383 [DOI] [PubMed] [Google Scholar]
- 9.Morrato EH, Cuffel B, Newcomer JW, Lombardo I, Kamat S, Barron J. Metabolic risk status and second-generation antipsychotic drug selection: a retrospective study of commercially insured patients. Journal of Clinical Psychopharmacology 2009;29(1):26–32 DOI: [pii] [DOI] [PubMed] [Google Scholar]
- 10.Hsu C, Ried LD, Bengtson MA, Garman PM, McConkey JR, Rahnavard F. Metabolic monitoring in veterans with schizophrenia-related disorders and treated with second-generation antipsychotics: findings from a Veterans Affairs-based population. Journal of the American Pharmaceutical Association 2003;48(3):393–400 DOI:N01234155308VG48 [pii] 10.1331/JAPhA.2008.07007 [DOI] [PubMed] [Google Scholar]
- 11.Morrato EH, Nicol GE, Maahs D, et al. Metabolic screening in children receiving antipsychotic drug treatment. Archives of Pediatric and Adolescent Medicine 2009;164(4):344–51 DOI:164/4/344 [pii] 10.1001/archpediatrics.2010.48 [DOI] [PubMed] [Google Scholar]
- 12.Brixner DI, Said Q, Corey-Lisle PK, et al. Naturalistic impact of second-generation antipsychotics on weight gain. Annals of Pharmacotherapy 2006;40(4):626–32 DOI:aph.1G564 [pii] 10.1345/aph.1G564 [DOI] [PubMed] [Google Scholar]
- 13.McAdam-Marx C, Bouchard J, Aagren M, Conner C, Brixner DI. Concurrent control of blood glucose, body mass, and blood pressure in patients with type 2 diabetes: an analysis of data from electronic medical records. Clinical Therapy 2011;33(1):110–20 DOI:S0149–2918(11)00025–7 [pii] 10.1016/j.clinthera.2011.01.018 [DOI] [PubMed] [Google Scholar]
- 14.LaFleur J, McAdam-Marx C, Alder SS, et al. Clinical risk factors for fracture among postmenopausal patients at risk for fracture: a historical cohort study using electronic medical record data. Journal of Bone and Mineral Metabolism 2011;29(2):193–200 DOI: [DOI] [PubMed] [Google Scholar]
- 15.King G, Langche Z. Logistic regression in rare events. Political Analysis 2001;9:135–63 [Google Scholar]
- 16.RELOGIT: Rare Events Logistic Regression [computer program]. Version 1.1 Harvard University: Cambridge, MA, 1999 [Google Scholar]
- 17.Correll CU, Manu P, Olshanskiy V, Napolitano B, Kane JM, Malhotra AK. Cardiometabolic risk of second-generation antipsychotic medications during first-time use in children and adolescents. JAMA 2009;302(16):1765–73 DOI:302/16/1765 [pii] 10.1001/jama.2009.1549 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Panagiotopoulos C, Ronsley R, Davidson J. Increased prevalence of obesity and glucose intolerance in youth treated with second-generation antipsychotic medications. Canadian Journal of Psychiatry 2009;54(11):743–9 [DOI] [PubMed] [Google Scholar]
- 19.Morrato EH, Druss BG, Hartung DM, et al. Small area variation and geographic and patient-specific determinants of metabolic testing in antipsychotic users. Pharmacoepidemiology and Drug Safety 2011;20(1):66–75 DOI: 10.1002/pds.2062 [DOI] [PubMed] [Google Scholar]
- 20.Baker JL, Olsen LW, Sorensen TI. Childhood body-mass index and the risk of coronary heart disease in adulthood. New England Journal of Medicine 2007;357(23):2329–37 DOI:357/23/2329 [pii] 10.1056/NEJMoa072515 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Bhargava SK, Sachdev HS, Fall CH, et al. Relation of serial changes in childhood body-mass index to impaired glucose tolerance in young adulthood. New England Journal of Medicine 2004;350(9):865–75 DOI: 10.1056/NEJMoa035698350/9/865 [pii] [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Sinaiko AR, Donahue RP, Jacobs DR, Jr, Prineas RJ. Relation of weight and rate of increase in weight during childhood and adolescence to body size, blood pressure, fasting insulin, and lipids in young adults. The Minneapolis Children's Blood Pressure Study. Circulation 1999;99(11):1471–6 [DOI] [PubMed] [Google Scholar]
- 23.Srinivasan SR, Myers L, Berenson GS. Predictability of childhood adiposity and insulin for developing insulin resistance syndrome (syndrome X) in young adulthood: the Bogalusa Heart Study. Diabetes 2002;51(1):204–9 [DOI] [PubMed] [Google Scholar]