影响糖尿病患者听力损害因素的多元Logistic回归分析

Multiple Logistic regression analysis on factors influencing hearing impairment of patients with diabetes

目的多元Logistic回归分析影响糖尿病患者听力损害因素。方法回顾性分析2010年10月至2014年6月入住本院的208例糖尿病患者的临床资料,其中,听力损害44例,听力正常164例。另选择同期于本院体检的健康者122例作为对照组。采用Pearson单因素及多元Logistic回归分析糖尿病组听力正常者与听力损害者体质指数（BMI）、腰围（WC）、糖化血红蛋白（Hb Alc）、总胆固醇（TC）、甘油三酯（TG）、收缩压（SBP）、舒张压（DBP）、高密度脂蛋白胆固醇（HDL-C）、低密度脂蛋白胆固醇（LDL-C）、糖尿病肾病（diabetic nephropathy,DN）及糖尿病视网膜病变（diabetic retinopathy,DR）等指标的差异,比较糖尿病组与对照组研究对象纯音测听、畸变产物耳声发射（distortion product evoked otoacoustic emission,DPOAE）振幅、瞬态诱发性耳声发射（transient evoked otoacoustic emission,TEOAE）振幅及ABR阈上40 d BⅢ潜伏期。结果 1经Pearson单因素分析,糖尿病组听力正常者与听力损害者在WC、TC、TG、SBP、DBP、HDL-C、LDL-C等方面比较差异均无显著性（P〉0.05）,在年龄、病程、BMI、Hb Alc、LDL-C、DN及DR方面比较差异均具有显著性（P〈0.05）;2经多元Logistic回归分析,影响糖尿病患者听力损害的因素包括：年龄、病程、LDL-C、DN及DR;3糖尿组患者双耳低、中、高频纯音测定结果显著高于对照组,且糖尿病组患者左耳测定值均显著低于右耳（P〈0.05）;4糖尿病组患者双耳中、高频DPOAE振幅均显著低于对照组,且糖尿病组患者右耳DPOAE振幅均显著低于左耳（P〈0.05）;5糖尿病组患者双耳中、高频TEOAE振幅均显著低于对照组,且糖尿病组患者左耳中、高频TEOAE振幅均显著高于右耳（P〈0.05）;6糖尿病组患者ABR阈上40 d BⅢ、Ⅴ波潜伏期及Ⅰ~Ⅲ波间期较对照组均显著延长（P〈0.05）。结论影响糖尿病患者听力损害的因素以年龄、病程、LDL-C、DN及DR为主,应加强对上述危险因素的控制与监测;糖尿病患者听力损害以右耳更为常见,且以高频听力损害为主。

Objective To analyze factors influencing the hearing impairment of diabetic patients with multiple Logistic regression. Method Retrospective analysed from October 2010 to June 2014 of 208 cases admitted in our hospital with diabetes, in which hearing impairment in 44 cases, 164 cases of normal hearing. Selected 122 healthy subjects undergoing physical examinations during the same period as control group. Pearson single factor and multiple Logistic were applied to regressively analyze results of body mass index（BMI）, waist circumference（WC）, Hb Alc, total cholesterol（TC）, triglyceride（TG）, systolic blood pressure（SBP）, diastolic blood pressure（DBP）, high-density lipoprotein cholesterol（HDL-C）, low density lipoprotein cholesterol（LDL-C）, diabetic nephropathy（DN） and diabetic retinopathy（DR） and other indexes. Results of pure tone audiometry and amplitude results of distortion product evoked otoacoustic emission（DPOAE） and transient evoked otoacoustic emission（TEOAE）, and threshold 40 db III incubation period of auditory brain-stem response（ABR） of patients in the diabetes group and control group were compared. Result1 Comparative differences in WC, TC, TG, SBP, DBP, HDL-C and LDL-C of patients with normal hearing and hearing impairment in diabetes group via Pearson single factors analysis were not statistically significant（P 0.05）, and the differences in the age, course, BMI, Hb Alc, LDL-C, DN and DR were significant（P 0.05）; 2 the factors influencing diabetic patients with hearing impairment via by multivariate logistic regression analysis include the age, course, LDL-C, DN and DR; 3 determination results of low-frequency, intermediate-frequency and high-frequency pure tone of patients in diabetes group were significantly higher than control group, the measurements of left ear of patients in the diabetes group were significantly lower than those of right ear（P 0.05）; 4 DPOAE amplitude of patients with intermediate frequency and high frequency in diabetes group was significantly less than control group, and DPOAE amplitude of right ears of patients with diabetes group was significantly lower than that of left ears（P 0.05）; 5 TEOAE amplitude of patients in the diabetes group was significantly lower than control group, and intermediate frequency and high frequency TEOAE amplitudes of left ears of patients in the diabetes group were significantly higher than those indexes of right ears of patients（P 0.05）; 6Ⅲ and Ⅴ wave incubation periods and Ⅰ~Ⅲ wave periods of ABR threshold 40 d B of patients in the diabetes group were significantly prolonged when compared to those of patients in control group（P 0.05）. Conclusion Factors influencing the hearing impairment of patients with diabetics mainly involve the age, duration, LDL-C, DN and DR, and the control and monitoring over such factors shall be strengthened. The hearing impairment of right ear over patients with diabetics is further regular, and high frequency hearing impairment is preferred.