Objectives: The response to coumarin derivatives warfarin and acenocoumarol is marked by high inter-individual and inter-temporal variability, which can lead to serious adverse events especially in elderly patients. This variability is due to both genetic and clinical factors. Polymorphisms of genes CYP2C9 and VKORC1, encoding for the metabolizing enzyme cytochrome P450 2C9 and the target enzyme vitamin K epoxide reductase, respectively, account for about 40% of the inter-individual variability. Clinical factors such as age, sex, comorbidity, concomitant therapy, weight and smoking status also influence coumarin derivatives dose requirements. Actually, anticoagulants are prescribed with manual dosage or with the help of computer-assisted programs. These programs usually suggest the dose on the only basis of patients INR measurements, without taking into consideration clinical and genetic features, this explains why a considerable number of patients, especially the elderly, fail to reach a stable and optimal anticoagulation. In our study we show that combining genetic and clinical information can result in segmenting patients in order to obtain an optimal therapy. Objectives of our study are: (1) introducing an index called drug sensitivity (Dsens) to capture the dose-INR relationship which better characterizes the patient behavior; (2) profiling patients on the basis of their genetic, clinical and therapeutic features. Materials and methods: We conducted an observational study on patients who are prescribed oral anticoagulant therapy in two specialized centers. We have collected in a structured database data of 3,949 patients; in this study we focus on the subset of 1,013 elderly patients (from 65 to 99 years). Each patient is characterized by the following features: age, sex, smoke status, weight, height, creatinine clearance, anticoagulant drug used, medical evidence to anticoagulation therapy, concomitant therapy and a time series of INR measurements and the corresponding anticoagulant doses. So far we have genotyped 561 elderly patients according to polymorphism of CYP2C9 and VKORC1 genes. The sensitivity of each patient is computed by the ratio between INR and dose variations as follows: Dsens = (1 + PDINRi%/Ndays)/(1 + PDdosei%/7) where DINRi% = (INRi+1 – INRi)/INRi and Ddosei% = (dosei+1–dosei)/dosei. Using this index we have classified patients in three Dsens classes (Wrong Responders, Correct Responders and Excess Responders). Note that a lower value of Dsens means that patient is responding in a wrong way to therapy; on the other side higher values of Dsens indicate that patient is responding to therapy in an excessive way. Whereas, patients falling in correct responders’ class have a more predictable drug response behavior. We also evaluated CYP2C9 and VKORC1 allelic variant frequencies distribution in the population of the three Dsens classes. Finally we have applied two machine learning classification algorithms [Support Vector Machines (SVM) and Bayesian Networks (BN)] and we compared the results obtained with a classical INR based classification with those obtained with the proposed Dsens based classification. Results: Characterization of patients in Fig. 1 with the distribution of Dsens and in Table 1 with the CYP2C9 and VKORC1 allelic variant frequencies distribution in the three Dsens classes. Classification results are depicted in Table 2. Conclusion: Profiling patients with classification methods based on Dsens index can identify groups of patients homogeneous with respect to the dynamics of INR. This is the basis for a new dosing algorithm.

Evaluation of the response to oral anticoagulant drugs : an observational study of 1,013 elderly patients / G. Ogliari, I. Giordani, A. Mihalich, C. Lodigiani, D. Castaldi, A. Dubini, P. Ferrazzi, A. Di Blasio, E. Messina, F. Archetti, L. Rota, D. Mari. - In: INTERNAL AND EMERGENCY MEDICINE. - ISSN 1828-0447. - 5:suppl. 2(2010 Dec), pp. S95-S96. ((Intervento presentato al 111. convegno Congresso Nazionale della Società Italiana di Medicina Interna tenutosi a Roma nel 2010 [10.1007/s11739-010-0505-3].

Evaluation of the response to oral anticoagulant drugs : an observational study of 1,013 elderly patients

G. Ogliari
Primo
;
D. Mari
Ultimo
2010

Abstract

Objectives: The response to coumarin derivatives warfarin and acenocoumarol is marked by high inter-individual and inter-temporal variability, which can lead to serious adverse events especially in elderly patients. This variability is due to both genetic and clinical factors. Polymorphisms of genes CYP2C9 and VKORC1, encoding for the metabolizing enzyme cytochrome P450 2C9 and the target enzyme vitamin K epoxide reductase, respectively, account for about 40% of the inter-individual variability. Clinical factors such as age, sex, comorbidity, concomitant therapy, weight and smoking status also influence coumarin derivatives dose requirements. Actually, anticoagulants are prescribed with manual dosage or with the help of computer-assisted programs. These programs usually suggest the dose on the only basis of patients INR measurements, without taking into consideration clinical and genetic features, this explains why a considerable number of patients, especially the elderly, fail to reach a stable and optimal anticoagulation. In our study we show that combining genetic and clinical information can result in segmenting patients in order to obtain an optimal therapy. Objectives of our study are: (1) introducing an index called drug sensitivity (Dsens) to capture the dose-INR relationship which better characterizes the patient behavior; (2) profiling patients on the basis of their genetic, clinical and therapeutic features. Materials and methods: We conducted an observational study on patients who are prescribed oral anticoagulant therapy in two specialized centers. We have collected in a structured database data of 3,949 patients; in this study we focus on the subset of 1,013 elderly patients (from 65 to 99 years). Each patient is characterized by the following features: age, sex, smoke status, weight, height, creatinine clearance, anticoagulant drug used, medical evidence to anticoagulation therapy, concomitant therapy and a time series of INR measurements and the corresponding anticoagulant doses. So far we have genotyped 561 elderly patients according to polymorphism of CYP2C9 and VKORC1 genes. The sensitivity of each patient is computed by the ratio between INR and dose variations as follows: Dsens = (1 + PDINRi%/Ndays)/(1 + PDdosei%/7) where DINRi% = (INRi+1 – INRi)/INRi and Ddosei% = (dosei+1–dosei)/dosei. Using this index we have classified patients in three Dsens classes (Wrong Responders, Correct Responders and Excess Responders). Note that a lower value of Dsens means that patient is responding in a wrong way to therapy; on the other side higher values of Dsens indicate that patient is responding to therapy in an excessive way. Whereas, patients falling in correct responders’ class have a more predictable drug response behavior. We also evaluated CYP2C9 and VKORC1 allelic variant frequencies distribution in the population of the three Dsens classes. Finally we have applied two machine learning classification algorithms [Support Vector Machines (SVM) and Bayesian Networks (BN)] and we compared the results obtained with a classical INR based classification with those obtained with the proposed Dsens based classification. Results: Characterization of patients in Fig. 1 with the distribution of Dsens and in Table 1 with the CYP2C9 and VKORC1 allelic variant frequencies distribution in the three Dsens classes. Classification results are depicted in Table 2. Conclusion: Profiling patients with classification methods based on Dsens index can identify groups of patients homogeneous with respect to the dynamics of INR. This is the basis for a new dosing algorithm.
warfarin ; acenocoumarol ; elderly ; anticoagulant drugs ; observational
Settore MED/09 - Medicina Interna
dic-2010
Società Italiana di Medicina Interna
Article (author)
File in questo prodotto:
Non ci sono file associati a questo prodotto.
Pubblicazioni consigliate

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/2434/153634
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact