Fumiko Ohtsu

Development of a Facts Database for Evidence- Based Adverse Reaction Diagnosis

Purpose: We have been building the CARPIS (Case Reports of Adverse Drug Reactions and Poisoning Information System) database since 1987 and have been using CARPIS primarily as a literature database. The purposes of this study were to revamp CARPIS as a fact database for adverse drug reactions (ADRs) and to develop methodology for identifying symptoms as part of evidence-based adverse reaction diagnosis. Revamp: Fact data in case reports, such as causative drugs, ADRs, and patient backgrounds, were analyzed to develop functions for the fact database and to establish different data fields. We developed our own dictionaries and data-tabulating function. Outcome: It was possible to calculate the incidence of subjective symptoms for ADRs, list causative drugs, and identify risk factors using the revamped CARPIS. The incidence of each subjective ADR symptom was compared with the incidence of objective symptoms associated with particular ADRs to identify subjective symptoms that could estimate ADRs. We succeeded in creating new information regarding the symptoms for evidence-based adverse reaction diagnosis. Conclusion: We were successful in revamping CARPIS as a novel fact database for ADRs and in developing methodology for identifying symptoms for evidence-based adverse reaction diagnosis.

[Methodology for Estimating the Risk of Adverse Drug Reactions in Pregnant Women: Analysis of the Japanese Adverse Drug Event Report Database].

Safety information regarding drug use during pregnancy is insufficient. The present study aimed to establish an optimal signal detection method to identify adverse drug reactions in pregnant women and to evaluate information in the Japanese Adverse Drug Event Report (JADER) database between April 2004 and November 2014. We identified reports on pregnant women using the Standardised MedDRA Queries. We calculated the proportional reporting ratio (PRR) and reporting odds ratio (ROR) of the risk factors for the two known risks of antithyroid drugs and methimazole (MMI) embryopathy, and ritodrine and fetal/infant cardiovascular events. The PRR and ROR values differed between all reports in the JADER database and those on pregnant women, affecting whether signal detection criteria were met. Therefore we considered that reports on pregnant women should be used when risks associated with pregnancy were determined using signal detection. Analyses of MMI embryopathy revealed MMI signals [PRR, 159.7; ROR, 669.9; 95% confidence interval (CI), 282.4-1588.7] but no propylthiouracil signals (PRR, 1.98; ROR, 2.0; 95%CI, 0.3-15.4). These findings were consistent with those of reported risks. Analyses of fetal/infant cardiovascular events revealed ritodrine signals (PRR, 2.1; ROR, 2.1; 95%CI, 1.4-3.3). These findings were also consistent with reported risks. Mining the JADER database was helpful for analyzing adverse drug reactions in pregnant women.

Physicochemical Properties of Causative Drugs Associated with RenalNephrotoxicity

Background: In this study, we examined the association between drug-induced renal nephrotoxicity and the physiochemical properties of the causative drugs we extracted from CARPIS in order to provide drug information about potential severe risks at the time of and after marketing approval. Methods: We designated the nephrotoxicity-associated drugs as the case drug group (126 drugs), and all other drugs were set as the control drug group (915 drugs). We compared the physicochemical properties of the group. We performed univariate logistic regression analysis on each investigation item, and then performed multivariate stepwise logistic regression analysis on the items which were p<0.2 using univariate logistic regression analysis. Results: Model 1, which uses a logP value, showed that the odds ratio of pKa at less than 7 was 2.46, and that at less than 7-8 was 2.01. The odds ratio of logP value at less than 0 was 1.67. MW at less than 300-400 was 0.57. Model 2, which uses a logD value, showed that logD at less than 0 was 2.23. The odds ratio of pKa at less than 7 was 2.34, and at less than 7-8 was 2.04. The odds ratio of MW at less than 300-400 was 0.56. Conclusion: The results clearly showed the risk of renal problems associated with water-soluble drugs. The information could be useful to consider potential risks of renal problems associated with water-soluble drugs at the time of drug approval and after marketing and to compensate for the lack of information.

Relationship between Physicochemical Properties of Medical Supplies and Serious Adverse Drug Reactions Listed in the Package Inserts.

We sought to clarify the relationship between the physicochemical properties of each medical supply and serious adverse drug reactions listed in the package inserts, by reviewing new information. We investigated 1) 1078 medicines currently available on the domestic Japanese market by using physicochemical data, such as cLogD, molecular weight (MW), and pKa and 2) the serious adverse drug reactions stated in the package inserts and the presence or absence of serious renal and liver disorders, as well as mental, extrapyramidal, and skin disorders. The renal disorders data showed: cLogD<0, adjusted odds ratio (aOR)=2.00; MW values ≥500, aOR=2.28; and pKa<7.4, aOR=1.95-2.06. The liver disorders data showed: pKa<8.4, aOR=1.83-1.95, and MW values ≥300, aOR=1.47-1.87. The mental disorders data showed: cLogD≥0, aOR=2.12, and MW values<400, aOR=2.46-2.85. The extrapyramidal disorders data showed: pKa≥6.4, aOR=4.50-11.32; cLogD≥0, aOR=4.71; and MW values<500, aOR=7.95-15.08. The skin disorders data showed: cLogD<0, aOR=1.46; MW values ≥500, aOR=1.69; and pKa<6.4, aOR=1.65 or<7.4-8.4, aOR=1.59. This information will be useful for investigating the relationships between new drugs entering the market and their potential future adverse drug reactions, and for establishing both precautionary and medical observational standards.

Exploring Risk Factors that Contribute to the Onset of Ritodrine-Associated Serious Adverse Drug Reactions

Background: Ritodrine is a drug used for threatened premature labor. The severe adverse drug reactions associated with ritodrine are known to be pneumonedema, leukopenia, and rhabdomyolysis, but there have been few investigations on the risk factors. We performed a case-control study and selected case reports as a case group and healthy pregnant women in clinical practice as a control group. Methods: We extracted the onsets of pneumonedema, leukopenia, and rhabdomyolysis associated with ritodrine from case reports in Japan as a case group. We selected healthy pregnant women with ritodrine administration in clinical practice as a control group. We investigated their age, medical history; Pregnancy Induced Hypertension (PIH), multiple pregnancies, concomitant drugs administered, and maximum rate of ritodrine infusion, and examined the association with those factors with the onset of adverse drug reactions by logistic regression analysis. Results: The results of the case group showed: pneumonedema (28 cases); leukopenia (25 cases); rhabdomyolysis (21 cases). The risk factors significantly associated with pneumonedema are a medical history of the cardiovascular system, PIH, multiple pregnancy, and concomitant treatment with steroids, which all match with the precautions in ritodrine’s package insert. The factors associated with leukopenia are its administration longer than 7 days and the concomitant treatment with Mg. The factors associated with rhabdomyolysis are multiple pregnancies and a concomitant treatment with Mg. Conclusion: Risk factors for the onset of pneumonedema match the descriptions in the ritodrine package insert, and can be explained by pharmacological actions. Thus, this study could elucidate the risk factors for rare adverse drug reactions limited to pregnant women. The onsets of leukopenia and rhabdomyolysis were caused by physiological changes by pregnancy and its progression of disease state and ritodrine’s pharmacological action, and were suggested the possibility of risk factors.

Development and Validation of Estimate Equations for Adverse Drug Reactions Using Risk Factors and Subjective Symptoms

The purpose of this study was to develop and validate estimate equations for preventing adverse drug reactions (ADRs). We conducted five case-control studies to identify individual risk factors and subjective symptoms associated with the following five ADRs: drug-induced ischemic heart disease; renal damage; muscle disorder; interstitial pneumonia; and leucopenia. We performed logistic regression analysis and obtained eight regression equations for each ADR. We converted these to ADR estimate equations for predicting the likelihood of ADRs. We randomly selected 50 cases with non-individual ADRs from the Case Reports of Adverse Drug Reactions and Poisoning Information System (CARPIS) database of over 65000 case reports of ADRs, and assigned these cases to a validation case group. We then calculated the predictive probability for 50 cases using the eight estimate equations for each ADR. The highest probability for each ADR was set as the probability of each ADR. If the probability was over 50%, the case was interpreted as ADR-positive. We calculated and evaluated the sensitivity, specificity, and positive likelihood ratio of this system. Sensitivity of the estimate equations for muscle disorder and interstitial pneumonia were ≥90%. Specificity and positive likelihood ratios of estimate equations for renal damage, interstitial pneumonia and leucopenia were ≥80% and ≥5, respectively. Our estimate equations thus showed high validity, and are therefore helpful for the prevention or early detection of ADRs.