Analysis Of Influencing Factors Of Aromatase Inhibitor-associated Arthralgia Based On COX Proportional Hazards Model

Objective: To explore the influencing factors of aromatase inhibitor-associated arthralgia during treatment with aromatase inhibitors in postmenopausal breast cancer patients,and to plot the cumulative incidence curve of aromatase inhibitor-associated arthralgia.Methods:The clinical data of 225 postmenopausal breast cancer patients who underwent surgical treatment in the Breast Disease Diagnosis and Treatment Center of the Affiliated Hospital of Qingdao University from July 31,2017 to August 1,2019 were retrospectively collected.Statistical analysis was performed using R 3.6.3,and log-rank test and Breslow test were used to perform univariate analysis on the distribution of aromatase inhibitor-associated arthralgia(AIA);incidence in different subgroups of all possible influencing factors.The proportional hazards(PH)hypothesis test used the graphical method and the Schoenfeld residual method.After the independent variable(history of chemotherapy)that does not satisfy the PH assumption was converted into a time-dependent covariate,it was included in the COX regression model together with the statistically significant variables in the univariate analysis to perform multivariate analysis to find the independent influencing factors of AIA.In statistical analysis,the starting time was the time when the medication was taken;the end event was the occurrence of AIA;the censoring event was that the AIA symptoms did not appear at the end of follow-up;the survival time : For patients with symptoms of AIA,the survival time is from the time of taking the drug to the time of symptoms of AIA.For patients without symptoms of AIA,the survival time is from the time of starting the drug to the end of follow-up.Results: Of the 225 enrolled patients with a median medication duration of 9 months,97(43.1%)developed symptoms of AIA.More than half of the AIA patients developed the disease within 3 months of taking the medication.The cumulative incidences of AIA at 1 month,6 months,and 12 months were 24.4%,36.1%,and 51.2%,respectively.The most common pain areas of AIA are the hands / wrists(75.3%),followed by the knee joint(59.8%),foot / ankle(17.5%),lower back(16.5%),hip joint(4.1%),shoulder joint(4.1 %).In univariate analysis,the distribution of AIA incidence in three subgroups of age,arthralgia history,and chemotherapy history was statistically different(P <0.05),while BMI,education level,menopause time,nature of work,home address,history of radiotherapy,targeted therapy,hypertension,diabetes,aromatase inhibitor type,tumor diameter,lymph node status,ER,PR,HER2,KI67,the 16 variables were not related to the incidence of AIA.The graphical method and Schoenfeld residual method were used to test the PH hypothesis for variables with statistical significance in univariate analysis.The history of chemotherapy does not satisfy PH hypothesis test,so inclusion of interaction between time and variable chemotherapy: chemotherapy * time(time is the number of months of medication)was included in the Cox proportional hazards model.Multi-factor COX analysis including age,history of arthralgia,history of chemotherapy,history of chemotherapy * time showed age(HR = 0.966,95% CI: 0.934～0.999,P = 0.048),history of chemotherapy(HR = 2.303,95% CI: 1.134～4.677,P = 0.021)is an independent influencing factor of AIA,and the history of arthralgia(HR = 1.418,95% CI: 0.899～2.236,P = 0.133)is not an independent influencing factor of AIA.Chemotherapy * time(HR = 0.887,95% CI: 0.788 ～ 0.998,P = 0.045)has statistical significance,indicating that the COX model is established.The risk hazard of independent variable chemotherapy changes with time,and the hazard ratio cannot be simply expressed by HR = Exp(0.834)= 2.303,which should be a function of time HR = Exp(0.834-0.120 * t).It can be seen that the effect of chemotherapy on the incidence of AIA decreases with time.Conclusion: For postmenopausal breast cancer patients taking aromatase inhibitors,younger patients have a higher risk of AIA;patients who have received chemotherapy also have a higher risk of AIA,but the effect of chemotherapy on AIA gradually decreases with time.