Ju-Young Shin, PhD

 This study was conducted to determine the number of pioglitazone users before and after the issue of the pioglitazone safety warning (intervention) by South Korea’s Ministry of Food and Drug Safety on June 10, 2011.

 A quasi-experimental interrupted time series study was conducted to examine the number of pioglitazone and other antidiabetic drug users between 2009 and 2015.

 We used the National Health Insurance Service—National Sample Cohort database to estimate the number of pioglitazone and other antidiabetic drug users between 2009 and 2015. Relative and absolute changes in the number of drug users were calculated with 95% CIs. Monthly numbers of drug users were presented according to the maximum likelihood estimation method, and a segmented regression analysis was performed to evaluate the effect of the intervention. A Durbin-Watson statistic and Dickey-Fuller test were used to assess autocorrelation and seasonality, respectively.

 A total of 80,724 patients with diabetes, including 12,249 pioglitazone users, were investigated. The relative change in pioglitazone users was —8.13% (95% CI, –8.41% to –7.86%). The intervention was associated with an immediate decrease of 177 pioglitazone users per 1000 patients with diabetes (

 <.05). Without this intervention, the predicted proportion of pioglitazone users would be 90 per 1000 patients with diabetes, which is 1.5-fold higher than the actually observed rate of 60 per 1000 patients with diabetes.

 This intervention led to a moderate decrease in pioglitazone users. Until further evidence is available, caution should be exercised when prescribing pioglitazone to patients with high potential risk of bladder cancer and alternative treatments should be considered.

To our knowledge, no study has been conducted in South Korea to evaluate the effect of the pioglitazone safety warning about its risk of bladder cancer that was released by South Korea’s Ministry of Food and Drug Safety on June 10, 2011.

For pioglitazone, the relative change in the proportion of drug users per 1000 patients with diabetes was —8.13% (95% CI, –8.41% to –7.86%) and the absolute change was –1.04 (95% CI, –1.40 to –0.68) percentage points per 1000 patients with diabetes.

The pioglitazone safety warning was associated with an immediate decrease of 177 pioglitazone users per 1000 patients with diabetes whereas, if the intervention had not been implemented, the proportion of pioglitazone users would have shown an increasing trend.

Pioglitazone is a drug of the thiazolidinedione (TZD) class used to treat type 2 diabetes.

 In June 2011, a French study suggested an increased risk of bladder cancer among patients treated with pioglitazone, resulting in the withdrawal of this drug by the French government. A study conducted in the United States reported similar findings.

 Following these results, the US FDA issued a drug safety communication, followed by warnings from both the European Medicines Agency and the Australian Therapeutic Goods Administration reporting similar concerns. On June 10, 2011, South Korea’s Ministry of Food and Drug Safety (MFDS) also released a safety warning (hereafter referred to as “the intervention”) stating that pioglitazone should be prescribed with caution.

A pioglitazone withdrawal study in France investigating the impact of pioglitazone’s regulatory withdrawal on antidiabetic drug use found no adverse events among patients with diabetes.

 On the other hand, an Australian and British study on TZD utilization trends reported that safety warnings were associated with decreased use of the respective drugs, rosiglitazone and pioglitazone.

 Furthermore, Taiwan’s FDA communicated the possible risks of bladder cancer associated with pioglitazone, and this regulatory action had a positive impact of decreasing the use of pioglitazone among high-risk patients.

The data reported in previous studies had limitations because they were analyzed over a relatively short period of time. Although the withdrawal study in France

was conducted from 2010 to 2014, the drug withdrawal took place in January 2011, thus resulting in a prewithdrawal period of only 12 months.

 This short prewithdrawal period may affect the predicted number of drug users. Another study conducted in Spain examining the effect of rosiglitazone’s safety warning analyzed data collected over only 3 years (2006-2008), an even shorter period (safety warnings for rosiglitazone were released throughout 2007 and 2008).

 With such short observation periods, application of a time lag before or after an intervention may not be feasible. Furthermore, pioglitazone use following the MFDS’ safety warning has yet to be evaluated in Korea.

The current study was therefore conducted to explore the proportion of pioglitazone users before and after the intervention between 2009 and 2015. Furthermore, we evaluated the impact of the intervention on pioglitazone use by comparing the use of other antidiabetic drugs. If numbers of both pioglitazone users and users of other antidiabetic drugs decreased, a cause other than our intervention may be behind it. We hypothesized that with the intervention, the proportion of pioglitazone users would decrease and lead to an increase in the proportion of patients using other antidiabetic drugs, indicating that safety warnings are an effective regulatory measure to prevent use and thus potential adverse outcomes.

This study was conducted using the nationwide population-based National Health Insurance Service—National Sample Cohort (NHIS-NSC) database, which includes approximately 1 million individuals randomly selected from almost the entire Korean population, using national claims data from January 1, 2009, to December 31, 2015 (described in detail elsewhere).

 The NHIS-NSC database contains anonymized patient codes along with sociodemographic characteristics; medical care history; medical care institution types; 

International Classification of Diseases, Tenth Revision 

) diagnosis codes; and drug prescription information (generic name, prescription date, duration, and dosage).

The interrupted time series (ITS) study design, a type of quasi-experimental research, is widely used when evaluating the effectiveness of population-level health interventions implemented at a clearly defined point in time.

 The design generally involves constructing a time series of population-level rates for a particular quality improvement focus and statistically testing for variations in the outcome rate in the time periods before and after implementation of an intervention designed to alter the outcome.

 A segmented regression approach was used with an autoregressive integrated moving average approach to test the effect of the intervention on the outcome of interest by using an appropriately defined impact model, as there was no remaining residual autocorrelation.

Study subjects included all patients found in the NHIS-NSC database 18 years or older with a diagnosis of diabetes (

 codes E10-E14), in both inpatient and outpatient settings, who were prescribed any antidiabetic drug between January 2009 and December 2015. The periods before (January 1, 2009, to June 10, 2011) and after (June 11, 2011, to December 31, 2015) the intervention were defined as shown in 

Exposure was defined as “before” or “after,” relative to the intervention. The proportion of antidiabetic drug users was defined as the number of antidiabetic drug users divided by the total number of patients with diabetes. Use of the study drug, pioglitazone (anatomical therapeutic chemical classification system code, A10BG03), was compared with use of other antidiabetic drugs (comparators), which were classified as (1) rosiglitazone (A10BG02), (2) sulfonylurea derivatives (A10BB) and metformin (A10BA02), (3) dipeptidyl peptidase-4 (DPP-4) inhibitors (A10BH) and glucagon-like peptide-1 (GLP-1) analogues (A10BJ), and (4) insulin analogues (A10A). Metformin, the preferred initial treatment of diabetes, and sulfonylurea, which is used for second-line therapy together with metformin, were grouped because these 2 drug classes are regarded as the most cost-effective and widely prescribed treatment for diabetes.

 Additionally, because both DPP-4 inhibitors and GLP-1 analogues are incretin-based drugs with similar mechanisms of action, they were grouped together as well.

Demographic variables, such as age and gender, were identified from the database. With regard to medical institution type (ie,

institutions that patients visited for diabetes-related healthcare services), the following classification was applied: tertiary hospitals (≥500 beds), general hospitals (30-499 beds), and clinics (<30 beds). With regard to comorbidities, those with a history of ischemic heart disease (

 codes I24 and I25), myocardial infarction (I21), ischemic stroke (I63), hypertension (I10-I15), and cancer (C00-D49) were assessed in the whole study period (ie, the periods before and after the intervention).

Age, gender, medical institution type, and comorbidities were presented as frequencies and proportions. The absolute standardized difference (aSD) was calculated for all categorical variables. The absolute change in drug users was calculated as the difference between the proportions before and after the intervention, whereas the relative change (%) was calculated by dividing the absolute change by the proportion of users before the intervention. 95% CIs were calculated for both absolute and relative changes. To estimate the impact of the intervention, the proportion of monthly antidiabetic drug users was analyzed via ordinary least-squares regression and maximum likelihood estimation.

A segmented regression model was designed using data available for 30 months prior to the intervention. Using preintervention data, the monthly rates over time were projected to predict what would have occurred without the intervention. The dependent variable was the proportion of antidiabetic drug users per 1000 patients with diabetes. Independent variables included time (months), intervention indicator, and time after the intervention (months). The intervention indicator variable was set as a dichotomous variable: 0 (before) or 1 (after). Time after the intervention was a continuous variable representing the number of months after the intervention and was set to 0 for all months prior to the intervention. The equation used for the regression model is as follows

The estimates for “intervention” and “time after intervention” are the main coefficients of interest from the segmented regression analysis, with the former measuring the level of change immediately after the intervention and the latter measuring the trend after the intervention. The assumption of autocorrelation for time-series data was assessed via Durbin-Watson statistics, and seasonality or stationarity was assessed using an augmented Dickey-Fuller test.

 In addition, the numbers of incident and prevalent pioglitazone users per 1000 patients with diabetes were calculated to compare and contrast their trends. Incident pioglitazone users were defined as those having no prescription for pioglitazone within the previous 12 months of the first occurrence of pioglitazone use. Prevalent pioglitazone users were defined as those having a prescription for pioglitazone in each respective month, regardless of their past use. To test the sensitivity and robustness of the study results, a 3-month lag was applied both before and after the intervention.

All statistical analyses were performed using the SAS Enterprise Guide statistical application program provided by the NHIS (release 9.71; SAS Institute Inc; Cary, North Carolina) and accessed through a virtual machine system. A 2-tailed 

 value <.05 or aSD <0.1 was considered statistically significant.

We identified a total of 80,724 patients with diabetes between 2009 and 2015. Among these, 12,249 (15.17%) were pioglitazone users, with men representing a higher proportion of both patients with diabetes and pioglitazone users. There were no statistically significant differences in age before or after the intervention for both groups (aSD >0.1), whereas there were statistically significant differences in medical institutions for only patients with diabetes and in comorbidities for both groups (aSD < 0.1) (

). Proportions of women, patients aged 50 to 59 years, patients aged 70 to 79 years, and visitors to primary care clinics who used pioglitazone decreased after the intervention, with relative changes of —3.74%, –7.13%, –37.85%, and –7.91%, respectively (

In the preintervention period, the proportion of pioglitazone users was 12.77%, and this decreased to 11.73% after the intervention, resulting in relative and absolute changes of —8.13% and –1.04 percentage points, respectively. However, proportions of patients using sulfonylurea and metformin, DPP-4 inhibitors and GLP-1 analogues, and insulin analogues all showed an increase after the intervention, with use of DPP-4 inhibitors and GLP-1 analogues increasing the most (relative change, 209.03%) (

The estimates for “intervention” and “time after intervention” are the main coefficients of interest from the segmented regression analysis. As observed from the “time” estimates, use of all antidiabetic drugs except rosiglitazone showed an increasing trend over the entire study period. The “time after intervention” coefficient showed a decreasing trend for use of pioglitazone, sulfonylurea and metformin, and insulin analogues, although the trend for pioglitazone was not statistically significant. Finally, the coefficient for “intervention” was statistically significant only for pioglitazone, sulfonylurea and metformin, and insulin analogues (

). The intervention was associated with an immediate decrease in pioglitazone users of 177 per 1000 patients with diabetes (

<.05), whereas there was a decrease of 448 for DPP-4 inhibitors and GLP-1 analogues.

From 2009 to 2015, proportions of pioglitazone and insulin analogue users showed a decreasing trend (

), although this trend was not as steep as that of rosiglitazone. Without the intervention, the predicted proportion of pioglitazone users was 90 per 1000 patients with diabetes, whereas the actually observed proportion was 60 per 1000 patients with diabetes (

). Likewise, by applying a 3-month lag both before and after the intervention, similar results were observed, indicating robustness in results (

). The numbers of prevalent and incident pioglitazone users are shown, with the general trend of prevalent users increasing and that of incident users decreasing (

The prevalence of diabetes shown in our study is in agreement with the 8.0% reported by the Diabetes Fact Sheet in Korea in 2016.

 Pioglitazone accounted for 15.2% of all antidiabetic drugs used in Korea in this study. All antidiabetic drugs except rosiglitazone and pioglitazone showed an overall increasing trend of use from 2009 to 2015. After the intervention, the relative change in the proportion of pioglitazone users was —8.13%, with the intervention being associated with an immediate decrease of 177 pioglitazone users per 1000 patients with diabetes. Thus, it can be deduced that safety warnings are effective in decreasing the number of patients using the flagged drug.

Upon release of the pioglitazone safety warning by the MFDS, a moderate decrease in the proportion of pioglitazone users was observed. The proportion of pioglitazone users actually began to decrease a few months prior to the intervention, possibly due to healthcare provider awareness arising from various study results reporting an increased risk of bladder cancer associated with pioglitazone use and precautionary actions prior to the national regulatory body announcement.

 A study conducted in Korea examining the trends of antidiabetic drug use in adult patients with diabetes from 2002 to 2013 showed a trend similar to that reported in our study, with the use of TZDs decreasing steadily from a peak of 13.0% in 2009 to 6.5% in 2013, which was lower than the use in 2002 (7.3%).

Key events associated with pioglitazone between 2009 and 2015 had either a positive or negative influence on its use. Before the intervention, in July 2010, a study reported an increased risk of cardiovascular (CV) disease with rosiglitazone use, leading to the withdrawal of the drug in September 2010.

 This resulted in a sharp increase in the number of pioglitazone users, as results of previous studies suggested that safety warnings lead to a decrease in the respective drug’s use while concurrently prompting an increase in the use of drugs with similar mechanisms.

 This particular trend was observed worldwide, including in Australia, Denmark, France, Germany, the Netherlands, Taiwan, and the United States.

However, around the time of rosiglitazone withdrawal, in August 2010 and April 2011, respectively, 2 studies reported a risk of fracture and bladder cancer associated with the use of pioglitazone, leading to a decrease in pioglitazone use.

 Following the intervention, the MFDS updated pioglitazone’s label to contain information on its increased risk of bladder cancer in November 2011. Since the intervention, there has been a continuous decrease in the proportions of both prevalent and incident pioglitazone users. However in January 2013, following positive results of the PROactive (Prospective Pioglitazone Clinical Trial in Macrovascular Events) clinical trial, which reported the benefits of pioglitazone for patients with diabetes with a history of CV disease,

 insurance reimbursement was extended to include 30-mg pioglitazone on top of the previously reimbursed 15-mg pioglitazone to lessen the burden of patients with diabetes showing no improvement in blood glucose levels with 15-mg pioglitazone.

 This resulted in an increase in the proportions of both prevalent and incident pioglitazone users. In addition, in November 2013, health insurance benefit coverage criteria to third-line antidiabetic drugs were extended and a pioglitazone complex was released to the market in the following month.

 The aforementioned events resulted in an increase in pioglitazone users.

Despite the intervention, pioglitazone still accounted for 11.73% of all antidiabetic drugs used in the period after the intervention (July 2011 to December 2015), showing only a minor absolute reduction of 1.04 percentage points. A study in France showed trends between 2006 and 2013 of pioglitazone and other antidiabetic drug use similar to those shown in our study, with decreased incidence of first-line noninsulin glucose-lowering drugs (especially TZDs), but DPP-4 inhibitors and metformin showing opposite trends to those found in our study.

 We found no discrepancies regarding the trend of antidiabetic drug use, and the decrease in the proportion of pioglitazone users in Korea was considered a result of the intervention.

Our study showed a significant short-term reduction in the proportion of pioglitazone users after the intervention. Without it, high-risk patients, such as those 65 years or older, men, and those with additional risk factors like having been exposed to aromatic chemicals or smoking, would have been vulnerable to bladder cancer with the use of pioglitazone.

 However, the intervention prevented such vulnerability implicated with the drug from occurring within this high-risk population, thereby minimizing health risks of these patients. Our study showed that the MFDS safety warning for pioglitazone effectively decreased the proportion of pioglitazone users. This finding was also observed in several past studies evaluating the effectiveness of similar interventions.

 However, the observed effect is not universal because others have found contrasting results.

 Nonetheless, in Korea, safety warnings along with the pop-up alert system of drug utilization review have been shown to affect prescribing, thereby reducing the proportion of users of the flagged drugs.

 Thus, further studies are needed to validate the true effectiveness of safety warnings.

The strengths of our study are that, to the best of our knowledge, this is the first population-based study conducted in Korea to examine the temporal trends in the prevalence of pioglitazone users before and after the safety warning issued in June 2011. In addition, we used the nationally representative NHIS-NSC database, which provided a valuable opportunity to investigate and explore the extent of pioglitazone use and its changes over time in Korea. Notably, the NHIS-NSC database underwent strict systematic stratified random sampling with proportional allocation within each stratum by using the individual’s total annual medical expenses as a target variable for sampling, resulting in robust representation of the Korean population.

Despite the strengths of this study, the results should be interpreted with caution considering the following limitations. First, the disease codes listed in the NHIS-NSC database may not represent the participant’s true disease status, as the codes were created for health insurance claims. Moreover, as this was an ITS study, other interventions besides the intervention of interest may have influenced the number of pioglitazone users. It is therefore difficult to ascertain whether the steady decline in the number of pioglitazone users was either accelerated or slowed by factors other than the intervention of interest (eg, by the release of a new and more effective antidiabetic drug).

Regulatory actions, such as the pioglitazone safety warning released by the MFDS, have been shown to reduce the likelihood of prescribing the relevant drug. This population-based study demonstrated decreases in the proportion of pioglitazone users compared with the proportions of those using other antidiabetic drugs over time. However, this decreasing trend appeared to have started before the intervention. Those with high potential risk of bladder cancer should be prescribed pioglitazone with caution or should consider alternative treatments.

The results of our study are relevant to ongoing research investigating and evaluating the effectiveness of regulatory actions taken by national regulatory bodies. Importantly, future studies should assess regulatory actions using various study designs, and comparison of their results would allow for conclusions to be drawn regarding the true effectiveness and impact of a regulatory action.

 School of Pharmacy, Sungkyunkwan University (HEJ, ISO, YHB, JYS), Suwon, Gyeonggi-do, South Korea; Department of Public Health Science, Graduate School of Public Health, and Institute of Health and Environment, Seoul National University (SIC), Seoul, South Korea.

 The authors gratefully acknowledge funding from the Basic Science Research Program through the National Research Foundation, South Korea, by the South Korean government (Ministry of Science and ICT, grant No. 2017K1A3A1A12073341); this study was conducted independently of the funder.

 The authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

 Concept and design (HEJ, SIC, JYS); acquisition of data (HEJ, ISO, YHB); analysis and interpretation of data (HEJ, SIC, ISO, JYS); drafting of the manuscript (HEJ, JYS); critical revision of the manuscript for important intellectual content (HEJ, SIC, JYS); statistical analysis (HEJ, ISO, YHB); and administrative, technical, or logistic support (YHB).

Ju-Young Shin, PhD, School of Pharmacy, Sungkyunkwan University, 2066 Seobu-ro, Jangan-gu, Suwon, Gyeonggi-do, South Korea. Email: shin.jy@skku.edu.

1. Waugh J, Keating GM, Plosker GL, Easthope S, Robinson DM. Pioglitazone: a review of its use in type 2 diabetes mellitus [erratum in 

. 2006;66(1):85-109. doi: 10.2165/00003495-200666010-00005.

2. Lewis JD, Ferrara A, Peng T, et al. Risk of bladder cancer among diabetic patients treated with pioglitazone: interim report of a longitudinal cohort study. 

. 2011;34(4):916-922. doi: 10.2337/dc10-1068.

Articles

The pioglitazone safety warning issued in South Korea, which recommended prescribing with careful attention among those with high risk of bladder cancer, led to a moderate decrease in pioglitazone users.

Prescribing Trend of Pioglitazone After Safety Warning Release in Korea

 To examine the extent of codeine prescriptions for children younger than 12 years in Korea and to investigate characteristics associated with pediatric codeine use.

 A retrospective observational study was conducted to examine codeine prescriptions and patients’ characteristics.

 We used the Korea Health Insurance Review and Assessment Service National Patient Sample database. The study participants were patients younger than 12 years who were prescribed codeine as inpatients or outpatients between 2011 and 2016. Pediatric codeine use was defined as codeine prescribed at least once for a child younger than 12 years. The frequency and proportion of pediatric codeine users were analyzed by age group (0-2, 3-6, or 7-11 years), sex, year, region, diagnosis, type of medical institution, and coprescribed medication. Logistic regression analyses were performed to identify characteristics associated with pediatric codeine use.

 Of all patients younger than 12 years, 518,895 (55.8%) and 16,337 (1.7%) were treated with codeine in outpatient and inpatient settings, respectively. Odds of pediatric codeine prescriptions were highest for outpatients at clinics (adjusted odds ratio [OR], 1.19; 95% CI, 1.16-1.21) and public hospitals (adjusted OR, 1.56; 95% CI, 1.28-1.91) and for inpatients at public hospitals (adjusted OR, 8.38; 95% CI, 6.64-10.58).

 Codeine was frequently prescribed for pediatric outpatients in Korea, especially in primary care clinics. Efforts to limit codeine use in children are required to prevent the occurrence of codeine-related adverse events.

Of patients younger than 12 years in Korea, 55.8% and 1.7% were treated with codeine in outpatient and inpatient settings, respectively.

Children were more likely to be treated with codeine in the outpatient setting at clinics and public hospitals.

Efforts to limit codeine use in children are needed to prevent the occurrence of codeine toxicity events.

Codeine, an opioid alkaloid derived from opium, was first identified in 1833.

 Codeine is a prodrug that is converted into morphine by cytochrome P-450 2D6 (CYP2D6) in the liver and binds to the μ-opioid receptor in the central nervous system, thereby producing its analgesic effect.

 Codeine has long been used in children as an analgesic and antitussive agent, and it was the most frequently used opioid medication among European children and adolescents in the early 2000s.

 However, codeine toxicity has been linked to pediatric deaths.

 Life-threatening or fatal toxicity of codeine was attributed to ultrarapid CYP2D6 metabolizers.

Owing to safety concerns, in 2012 the World Health Organization deleted codeine from its list of recommended analgesics based on the risk associated with metabolic rate differences and the insufficient pharmacological effect due to its lower metabolism in children younger than 12 years.

 Subsequently, the European Medicines Agency, the Medicines and Healthcare Products Regulatory Agency of the United Kingdom, and Health Canada also issued safety letters and guidelines recommending against the use of codeine in children younger than 12 years based on its opioid toxicity risk and suggesting a change in the drug labeling.

 In 2013, the FDA added a black box warning to the label of codeine-containing products to restrict the postoperative use of codeine for children, which was elevated to a contraindication in 2017 alerting that codeine should not be used to treat pain or cough in children younger than 12 years.

 Following the initiatives of other countries, the regulatory authority in Korea also issued a safety letter to warn about the risk of codeine use in children and, in November 2014, changed the drug label so that codeine is contraindicated in children younger than 12 years.

Despite repeated warnings regarding weak opioids such as codeine, the off-label use of codeine in young patients appears to persist, emphasizing the urgency of the need to address the issue of its contraindicated use.

 Characteristics of patients (eg, age) and healthcare providers (eg, specialty) were shown to influence opioid analgesic use and off-label prescribing to children in the United States.

 In the United States, codeine prescriptions in the emergency department (ED) were associated with age of the patient and geographic region of the prescriber.

 Odds of pediatric codeine use were higher in children outside of the Northeast United States, with poor physical health status, and with health insurance.

 In Australia, codeine use was higher in areas that were more remote and those with higher percentages of males.

 We reasoned that the characteristics of pediatric codeine use differ by patient setting, because inpatient use of codeine is mostly for management of postoperative pain but it is usually used as an antitussive agent in the outpatient setting.

 Therefore, this study aims to examine the patterns of prescribing codeine for children younger than 12 years and to identify characteristics associated with pediatric codeine prescriptions in outpatient and inpatient settings in Korea using population-based health insurance claims data. In addition, we sought to understand if a change in prescribing practice occurred after the label change for codeine in Korea.

The patterns of prescribing codeine were analyzed using Health Insurance Review and Assessment Service National Patient Sample (HIRA-NPS) data. The primary aim of the HIRA-NPS database is to construct sample data representative of the Korean population for researchers to access by compiling patient visit and prescription data from National Health Insurance claims data. The data are anonymized by replacing personal information with random unique identifiers. The database includes information on general characteristics of the patient (age and sex), medical services (inpatient and outpatient procedures and inpatient prescriptions), diagnoses (specialty codes, diagnosis codes, and names), outpatient prescriptions (generic code, dose, and prescribed days), and medical institution (type, number of beds, and region).

The HIRA-NPS database contains claims data on a stratified random sample of 3% of all patients covered by the National Health Insurance system; the mandatory health insurance system covers approximately 98% of all inhabitants of Korea.

 Stratified by sex and age, the patient sample was divided into a total of 32 strata (2 for sex and 16 for age) before random extraction. Different patients are selected for the sample data each year, and the patient samples are updated annually because they are developed through the accumulation of claims data over a yearlong cycle. HIRA, a government agency, has constructed and maintained the database annually since 2009. The database is publicly available to qualified researchers through HIRA’s website.

The study participants were patients younger than 12 years of age who received care as inpatients or outpatients in Korea between January 1, 2011, and December 31, 2016. The study drugs were drugs that contain codeine phosphate as the only analgesic and antitussive agent or codeine combination products that contain acetaminophen or ibuprofen as an active ingredient and are approved and listed as reimbursable drugs. Unlike in the United States and some European countries where low-strength codeine products are sold as over-the-counter medications, all codeine-containing products are sold as prescription-only medications in Korea. Inpatient prescriptions of codeine were analyzed separately from outpatient prescriptions owing to differences in dosage and the use of codeine for the management of postoperative pain.

Pediatric patients were defined as patients younger than 12 years. Pediatric codeine use was defined as the prescribing of codeine for children younger than 12 years. The proportion of pediatric patients who received a prescription for codeine was calculated by dividing the number of codeine users younger than 12 years by the number of total pediatric patients. The proportion of pediatric codeine users was calculated by patient setting (outpatient vs inpatient). The frequency and proportion of pediatric codeine use were analyzed by age group (0-2, 3-6, and 7-11 years), year, region, diagnosis, type of medical institution, and coprescribed medication.

The quarterly proportion of patients treated with codeine was plotted for inpatients and outpatients younger than 12 years. In addition, the quarterly proportion of patients treated with antibiotics was plotted to examine if it coincided with the proportion of codeine prescriptions according to seasonality. The numbers of pediatric patients and of pediatric codeine users were calculated by age group, sex, region, and clinical specialty. Pediatric patients were divided into 3 age groups (0-2, 3-6, and 7-11 years). Region was classified as metropolitan, city, or rural. Based on the classification of administrative regions provided by the government, we classified 5 megacities with surrounding satellite towns as metropolitan areas. Clinical specialties were grouped into pediatrics and nonpediatrics. The type of medical institution was divided into tertiary hospitals, general hospitals, clinics, and public hospitals.

 test was performed for categorical data. Multivariate logistic regression was performed to identify characteristics associated with pediatric codeine prescriptions in each patient setting (outpatient and inpatient). Independent variables were sex, region, diagnosis, type of medical institution, and coprescribed medication. An adjusted odds ratio (OR) with 95% CI was calculated for each variable. All statistical analyses were performed using SAS version 9.3 (SAS Institute Inc; Cary, North Carolina). The study protocol was approved by the Sungkyunkwan University Institutional Review Board (IRB No. 2016-10-015).

A total of 929,358 and 957,746 patients younger than 12 years were treated in outpatient and inpatient settings, respectively, between 2011 and 2016 (

). Of these patients, 518,895 (55.8%) and 16,337 (1.7%) patients were codeine users in those settings. The mean ages of these users were 5.5 and 4.2 years, respectively. Infants 2 years and younger comprised 22.7% and 40.3% of all pediatric codeine users in the outpatient and inpatient settings, respectively. In the outpatient setting, 51.1% of pediatric codeine users were treated in metropolitan areas. The most frequent diagnoses for pediatric codeine users were acute bronchitis (83.4% of outpatients and 91.0% of inpatients) and acute upper respiratory infection (81.0% and 88.6%, respectively). Antibiotics were coprescribed in 76.8% and 73.7% of outpatients and inpatients, respectively. In the outpatient and inpatient settings, 82.9% and 48.1% of pediatric codeine users were treated at clinics.

In both outpatient and inpatient settings, boys were as likely to be prescribed codeine as girls (

). The odds of pediatric codeine prescribing in the outpatient setting were higher in nonmetropolitan areas than in metropolitan regions. Children diagnosed with acute upper respiratory infection were 1.62 to 2.22 times more likely to be prescribed codeine than those without the diagnosis. Pediatric outpatients were more likely to receive codeine at clinics (adjusted OR, 1.19; 95% CI, 1.16-1.21) and public hospitals (adjusted OR, 1.56; 95% CI, 1.28-1.91) than at tertiary hospitals; inpatients were more likely to receive codeine at public hospitals (adjusted OR, 8.38; 95% CI, 6.64-10.58) than at tertiary hospitals. The odds of pediatric codeine prescriptions were increased with coprescribing of antibiotics but were decreased with coprescribing of opioid analgesics in both settings.

The proportion of pediatric patients treated with codeine showed little change in the temporal trend (

). The proportion of patients prescribed antibiotics tended to fluctuate together with the proportion of pediatric outpatients prescribed codeine. There appears to be seasonality for the proportion of outpatient codeine users.

By analyzing population-based National Health Insurance claims data, we showed that 55.8% and 1.7% of patients younger than 12 years were treated with codeine in the outpatient and inpatient settings, respectively. The extent of codeine prescriptions for children appears to be remarkably higher in Korea than in other countries. Between 2000 and 2005, the prescription rate of codeine for children younger than 12 years by primary care providers in the United Kingdom was 0.11 patients per 1000 person-years, whereas the rate was 1.32 in adolescent patients aged 12 to 17 years.

 The prescription rate of codeine for children in the ED in the United States was 3.3% in patients aged 3 to 7 years and 4.5% in those aged 8 to 12 years.

 Another US study showed that although an acetaminophen—codeine combination was among the most frequently used drugs among all hospitalized patients, it was not frequently used for patients younger than 12 years.

We showed that codeine was frequently prescribed in primary care clinics in the Korean outpatient setting. Codeine was frequently prescribed for treatment of upper respiratory infections, acute bronchitis, and asthma. Moreover, children who were prescribed antibiotics were more likely to be prescribed codeine than those who were not. Simultaneous fluctuations in the proportion of codeine users in the outpatient setting and the proportion of antibiotic users revealed that codeine was used to treat upper respiratory infections. These findings suggest that codeine was used as an antitussive agent. The prevalence of bronchitis and asthma is high among children with the common symptom of cough and, therefore, codeine is commonly used in children as an antitussive agent.

 Codeine has long been recommended as an antitussive agent that acts in the central nervous system against acute and chronic cough among children.

 However, evidence exists that codeine should not be used for cough due to pneumonia in infants owing to the risk of toxicity.

 Our findings and those of other studies suggest that physicians should be guided against prescribing codeine for children in the outpatient setting based on recent evidence and international guidelines.

 Our findings showed that in both outpatient and inpatient settings, the proportion of pediatric patients who were prescribed codeine was lower in 2013 than in 2011 and 2012, which may be attributable to a safety letter issued by the Ministry of Food and Drug Safety of Korea in 2012.

 After a slight uptick in 2014, the proportion fell back to even lower levels in 2015 and 2016, suggesting the possible impact of the label change in November 2014. Nonetheless, the extent of codeine prescriptions for children younger than 12 years remained substantial; approximately half of pediatric outpatients were treated with codeine in 2015 and 2016 after the label change. These findings highlight the need for more efforts to reduce pediatric codeine use in Korea. Similarly, since the release of the 2006 guidelines on codeine use among children by the American Academy of Pediatrics in the United States, prescribing of codeine for children and adolescents 17 years and younger did not decrease much.

 In contrast, the prescription rate of codeine decreased by approximately 60% in Italy after the issuance of a safety letter against use of codeine in infants younger than 2 years in 2007.

The results of the logistic regression analysis showed that pediatric codeine use tends to be greater in public hospitals than in tertiary hospitals in both outpatient and inpatient settings and that outpatient codeine use is higher in clinics and in tertiary hospitals. Our findings also showed that pediatric outpatients are more likely to be treated with codeine in metropolitan areas than in nonmetropolitan areas; the reverse is true for pediatric inpatients. These findings highlight the areas of intervention for clinical practice change. Efforts to improve prescribing behaviors should be directed toward clinics where a majority of pediatric patients were treated with codeine in the outpatient setting. Furthermore, efforts should be focused on outpatient use of codeine in the metropolitan areas.

By using population-based insurance claims data, this study analyzed the patterns of codeine prescriptions in children younger than 12 years in Korea. All drugs containing codeine in Korea are prescription-only medications and use of these drugs is recorded in the HIRA database used for this study. This makes it possible to accurately measure children’s exposure to these drugs. Nonetheless, this study also has limitations. First, although we showed that the proportions of pediatric patients treated with codeine tended to decrease in 2015 and 2016, these findings are only descriptive. Second, it is possible that codeine may have been prescribed as a nonreimbursable drug in the ED. Furthermore, the use of codeine for which claims have not been filed to the payer would not have been captured in the HIRA database. If children younger than 12 years visit the ED more frequently than older patients, this study may have underestimated the use of codeine prescriptions for children. Third, our finding of the high proportion of pediatric patients who were prescribed codeine in the outpatient setting should be interpreted with caution due to the actual practice of patient care in Korea. All patients who receive injectable drugs are classified as inpatient, and injectable use is high in Korea.

Despite these limitations, this study generated valuable evidence that codeine has been widely prescribed for children younger than 12 years in Korea, especially in the outpatient setting, even after the 2014 label change. Our findings suggest that various efforts to further limit pediatric use of codeine should be developed and implemented. One such measure shown to be effective in the United States was a clinical intervention program including prescriber education.

 The effective use of nonopioid analgesics may significantly reduce or even eliminate the need for opioids such as codeine.

 Healthcare providers should be guided on alternative pain management in children based on recent evidence on safety and efficacy.

 Based on our findings, such efforts to restrict prescriptions of codeine among young patients should not only reach major sources of prescriptions, such as metropolitan areas, but also include general hospitals and clinics in nonmetropolitan areas in Korea. Meanwhile, informing parents of pediatric patients about the risks of pediatric codeine use may also be helpful to influence clinical practice. In addition, it is important on the nationwide level to systematically monitor codeine use and harms in the pediatric population and to develop policy measures to encourage healthcare providers to comply with contraindication of codeine in pediatric patients for analgesic and antitussive effects; an example would be limiting reimbursement for off-label use of codeine.

The guidelines published by many international and national organizations based on risk-benefit assessment of codeine recommend that codeine should not be prescribed for children younger than 12 years. This study showed that codeine was frequently prescribed for children younger than 12 years despite the actions taken to restrict the use of codeine in that age group in Korea as well as in many other countries. Children were more likely to be treated with codeine at clinics and public hospitals in the outpatient setting. Our findings suggest that codeine is used for its antitussive effect among children in the outpatient setting. Further efforts to limit codeine use, especially as an antitussive agent in primary care clinics in the outpatient setting, are needed to prevent the occurrence of codeine toxicity events among children younger than 12 years.

No specific funding was granted. Ms Kim and Dr Song wrote the first draft of the manuscript and are listed as co—first authors. Each author listed in the byline has seen and approved the submission of this version of the manuscript and takes full responsibility for the manuscript.

 School of Pharmacy, Sungkyunkwan University (DK, DY, JYS), Suwon, Gyeonggi-do, Korea; Kongju National University, College of Nursing and Health (IS), Gongju, Korea.

 Concept and design (DK, JYS); acquisition of data (DK, IS); analysis and interpretation of data (DK, IS, DY); drafting of the manuscript (DK, IS); critical revision of the manuscript for important intellectual content (DY); statistical analysis (DK, DY); administrative, technical, or logistic support (JYS); and supervision (JYS).

 Ju-Young Shin, PhD, School of Pharmacy, Sungkyunkwan University, 2066 Seobu-ro, Jangan-gu, Suwon, Gyeonggi-do, South Korea. Email: shin.jy@skku.edu.

1. Madadi P, Koren G. Pharmacogenetic insights into codeine analgesia: implications to pediatric codeine use. 

. 2008;9(9):1267-1284. doi: 10.2217/14622416.9.9.1267.

2. Williams DG, Hatch DJ, Howard RF. Codeine phosphate in paediatric medicine. 

. 2001;86(3):413-421. doi: 10.1093/bja/86.3.413.

3. Takahama K, Shirasaki T. Central and peripheral mechanisms of narcotic antitussives: codeine-sensitive and -resistant coughs. 

4. Neubert A, Verhamme K, Murray ML, et al; TEDDY Network of Excellence. The prescribing of analgesics and non-steroidal anti-inflammatory drugs in paediatric primary care in the UK, Italy and the Netherlands. 

. 2010;62(3):243-248. doi: 10.1016/j.phrs.2010.04.006.

5. Friedrichsdorf SJ, Nugent AP, Strobl AQ. Codeine-associated pediatric deaths despite using recommended dosing guidelines: three case reports. 

. 2013;9(2):151-155. doi: 10.5055/jom.2013.0156.

6. Gasche Y, Daali Y, Fathi M, et al. Codeine intoxication associated with ultrarapid CYP2D6 metabolism [erratum in 

. 2004;351(27):2827-2831. doi: 10.1056/NEJMoa041888.

7. Kelly LE, Rieder M, van den Anker J, et al. More codeine fatalities after tonsillectomy in North American children. 

. 2012;129(5):e1343-e1347. doi: 10.1542/peds.2011-2538.

WHO Guidelines on the Pharmacological Treatment of Persisting Pain in Children with Medical Illnesses

. Geneva, Switzerland: World Health Organization; 2012.

9. Medicines and Healthcare Products Regulatory Agency. Codeine for analgesia: restricted use in children because of reports of morphine toxicity. United Kingdom government website. gov.uk/drug-safety-update/codeine-for-analgesia-restricted-use-in-children-because-of-reports-of-morphine-toxicity. Published July 2013. Accessed July 7, 2017.

Codeine was frequently prescribed for children in Korea despite the actions taken to restrict its use in that age group in Korea and other countries.

Pediatric Codeine Prescriptions in Outpatient and Inpatient Settings in Korea

 This study aimed to identify the prevalence and predictors of hypoglycemia in patients with type 2 diabetes (T2D) using South Korea’s nationwide healthcare database.

 Retrospective cohort and nested case-control analyses were conducted to estimate the prevalence and predictors of hypoglycemia, respectively.

 A cohort of 2,273,481 patients with T2D was followed to estimate the 1-year prevalence of hypoglycemia. Total hypoglycemia was identified using outpatient, inpatient, or emergency department visit data containing a diagnosis code for hypoglycemia. Severe hypoglycemia was defined as an event with inpatient admission or emergency care. Within the T2D cohort, cases with hypoglycemia were identified, and up to 4 controls were randomly selected after matching by sex, age, and cohort entry date. Possible predictive factors included insurance type, medical institution type, hypoglycemic history, antidiabetic drugs, Charlson Comorbidity Index score, and diabetic complications. We conducted conditional logistic regression analyses to estimate adjusted odds ratios (aORs) and 95% CIs to identify predictors of hypoglycemia.

 The prevalences of total and severe hypoglycemia were 1.38% and 0.96%, respectively. Those with a history of hypoglycemia had the highest risk for a further hypoglycemic event (aOR, 16.71; 95% CI, 15.62-17.88). Use of combination therapy with insulin and sulfonylurea was highly associated with severe hypoglycemia (aOR, 15.09; 95% CI, 13.60-16.74). Among diabetic complications, the presence of nephropathy was the greatest predictive factor (aOR, 1.79; 95% CI, 1.73-1.85).

 Patients with a history of hypoglycemia or receiving combined antidiabetic therapy must be appropriately managed to achieve optimal glycemic control without significant risk of hypoglycemia.

Our study evaluated the prevalence and predictors of hypoglycemia among patients with type 2 diabetes (T2D) using the nationwide South Korean healthcare database.

The prevalence of total and severe hypoglycemia in patients with T2D was 1.38% and 0.96%, respectively.

A history of hypoglycemia was a strong predictive factor, approximately 16.71 times higher, for a further hypoglycemic event, and combination therapy of insulin and sulfonylurea was highly associated with severe hypoglycemia.

Patients with hypoglycemic history or having combined antidiabetic therapy need to be appropriately managed to achieve optimal glycemic control without significant risk of hypoglycemia.

Maintaining adequate glycemic control in patients with diabetes is essential for preventing micro- and macrovascular complications and premature death. However, intensive glycemic control can increase the risk of hypoglycemia.

 Aggressively achieving an optimal glycated hemoglobin (A1C) level is associated with a 3-fold increased risk of hypoglycemia.

 Hypoglycemia is a potentially severe adverse effect of antidiabetic treatment and can negatively impact physical, mental, and social aspects of a patient’s life. Short-term effects of hypoglycemia include an increased risk of accidents or falls and decreased cognitive function. Long-term consequences include decreased productivity, increased risk for cardiovascular diseases, and fear of future episodes.

 Furthermore, hypoglycemia can lead to a significant cost burden on healthcare systems and society.

 Patients who experience hypoglycemia are less satisfied with, and less likely to adhere to, antidiabetic treatment, which threatens the efficacy of treatment.

 Therefore, preventing hypoglycemia is important in diabetes management.

Previous studies’ results have documented possible predictors of hypoglycemia, including a past history of hypoglycemic events, antidiabetic drugs, and diabetic complications. A history of hypoglycemia is known as the most important risk factor for subsequent hypoglycemic events.

 Hypoglycemia can occur as a side effect of some antidiabetic drugs that increase insulin production.

 Associated comorbidities (eg, liver or kidney disorders and micro- or macrovascular complications), behavioral risk factors (eg, alcohol consumption, exercise, and missed meals), and diabetes duration are also common risk factors for hypoglycemia.

Hypoglycemia is a well-recognized complication in patients with type 1 diabetes (T1D), but it is often underestimated in patients with type 2 diabetes (T2D).

 Previous studies on T2D have focused on patients using insulin or specific oral medications or who visited specific medical institutions, rather than on the population as a whole.

 Furthermore, heterogeneity of study design, data collection methods, and target population make comparison of findings difficult across studies. Therefore, it is necessary to identify the risk of hypoglycemia in the entire population with T2D to provide an efficient management plan for optimal glycemic control with the lowest possible risk of hypoglycemia. This study aimed to evaluate the prevalence and predictive factors of hypoglycemia among patients with T2D using nationwide claims data.

We utilized the nationwide healthcare database of the Health Insurance Review and Assessment Service (HIRA; Seoul, South Korea) between January 1, 2011, and December 31, 2013. HIRA is a governmental agency established to evaluate the accuracy of claims for National Health Insurance (covers approximately 96.7% of the overall population in South Korea) and National Medical Aid (covers approximately 3.3% of the population).

 The HIRA database is generated during claim reimbursement for healthcare services. This database is representative of the total population in South Korea and has advantages for generalization to the population. It includes information regarding demographic variables; all medical services provided, along with diagnostic codes (Korean Classification of Diseases version 7 [KCD-7]); and all prescription medications dispensed.

Patient records and information were anonymized and deidentified before the analysis. This study was approved by the institutional review board of Sungkyunkwan University. Informed consent was waived by the board.

Study Design and Patient Selection Criteria

Two methods were used to evaluate the prevalence and predictors of hypoglycemia among patients with T2D. First, a retrospective cohort study was used to estimate the 1-year prevalence of hypoglycemia in these patients. The study population included patients 20 years or older and diagnosed with T2D, defined as having a T2D diagnosis code (KCD-7 code E11) and receiving treatment with antidiabetic drugs. The index period for selecting patients was January 1, 2011, to December 31, 2012, to ensure a 1-year follow-up period for estimating the prevalence. We excluded patients diagnosed with T1D (KCD-7 code E10) or gestational diabetes (KCD-7 code O24).

Second, a nested case-control study was conducted to identify the predictors of hypoglycemia within the cohort with T2D. The cohort entry date was defined as the first date of receiving antidiabetic drugs, with a diagnosis of T2D between July 1, 2011, and December 31, 2013. A 6-month period was applied to assess possible predictive factors. Within the eligible cohort, we identified cases with hypoglycemia that occurred after the cohort entry date. For each case, we defined the index date as the date of the first eligible claim for a hypoglycemia-related visit. Controls were selected using individual matching. For each case, up to 4 controls matched for sex, age (±5 years), and cohort entry date were randomly selected without replication from patients with T2D without a diagnosis of hypoglycemia after the cohort entry date. The index date for each control was considered equal to the index date of their matched case.

The 1-year prevalence of hypoglycemia was defined as the percentage of patients who experienced 1 or more hypoglycemic episode during the 1-year follow-up period after being screened for T2D. Total hypoglycemia was identified on the basis of the first medical encounter for hypoglycemia (outpatient, inpatient, or emergency department [ED] visits) containing any of the following KCD-7 diagnosis codes: E11.63 (T2D, with hypoglycemia), E12.63 (malnutrition-related diabetes mellitus, with hypoglycemia), E13.63 (other specified diabetes mellitus, with hypoglycemia), E14.63 (unspecified diabetes mellitus, with hypoglycemia), E16.0 (drug-induced hypoglycemia without coma), E16.1 (other hypoglycemia), or E16.2 (hypoglycemia, unspecified). Also, the setting of hypoglycemia diagnosis was used to capture information on the severity of hypoglycemia. Severe hypoglycemia was defined as an event with inpatient admission or emergency care. Annual prevalence was also presented from 2011 to 2013.

We considered the following factors as potential predictors of hypoglycemia: insurance type, medical institution type, hypoglycemia history, antidiabetic drugs, Charlson Comorbidity Index (CCI) score, and diabetic complications.

Insurance type was classified as insured or medical aid beneficiaries. It was used as a proxy indicator of income levels because the medical aid program is intended for low-income households.

 The type of medical institution was categorized as clinic, hospital, general hospital, or tertiary hospital based on the most-visited medical institution during the treatment period. A history of hypoglycemia was defined as having a claim for an outpatient, inpatient, or ED visit for hypoglycemia occurring during the 6-month period before cohort entry.

Antidiabetic drugs were evaluated using prescription data from the 30 days preceding the index date. Medications included insulin, sulfonylureas, nonsulfonylurea secretagogues, biguanides, α-glucosidase inhibitors, dipeptidyl peptidase-4 (DPP-4) inhibitors, thiazolidinediones, glucagon-like peptide-1 (GLP-1) analogues, and sodium glucose cotransporter-2 (SGLT-2) inhibitors. Thiazolidinediones, GLP-1 analogues, and SGLT-2 inhibitors were combined into 1 category (other monotherapy) due to limited sample size. When using 2 or more multiple drug combinations, we categorized with and without insulin and/or sulfonylureas.

The CCI score, a weighted index that accounts for the number and severity of comorbidities, was used as a general marker of comorbidity. The score was assessed using algorithms by Quan et al during the 6-month period before the index date.

 Diabetic complications were assessed using methods published by Park et al during the 6-month period prior to the index date.

 We defined diabetic complications as having at least 2 outpatient visits or an inpatient visit with a hospital stay of 2 or more days, including diagnosis for the following complications: retinopathy, nephropathy, neuropathy, peripheral vascular disease, cerebrovascular disease, or cardiovascular disease.

In the analysis of prevalence, baseline characteristics were expressed as means and SDs or frequencies and percentages. We measured 1-year prevalence of hypoglycemia over 3 years, and each year for 2011, 2012, and 2013. We also conducted subgroup analysis to identify differences in prevalence according to age group, sex, insurance type, medical institution type, antidiabetic drugs, CCI score, and diabetic complications.

In the analysis of predictors, descriptive statistics were used to compare the baseline characteristics between selected cases and matched controls. For categorical variables, frequencies and percentages were reported. Differences between cases and controls were analyzed using Pearson’s χ

 test (or Fisher’s exact test). For continuous variables, means and SDs were presented. Mean differences between groups were analyzed using Student’s 

 test. We conducted a conditional logistic regression to identify independent predictors of hypoglycemia. Conditional logistic regression analysis is a method used to minimize bias arising from unconditional logistic analysis of matched data.

The prevalence and predictors of hypoglycemia in South Korean patients with type 2 diabetes were evaluated using a nationwide healthcare database.

Author | Ju-Young Shin, PhD

Articles

Prescribing Trend of Pioglitazone After Safety Warning Release in Korea

Pediatric Codeine Prescriptions in Outpatient and Inpatient Settings in Korea

Prevalence and Predictors of Hypoglycemia in South Korea