对来自犬源型内脏利什曼病流行区甘肃省文县和迭部县的537只家犬血样和来自非流行区天祝县的37只家犬血样,分别采用人用rk39试条、犬用rk39试条和循环抗原（CA）试条进行平行测试,以PCR检测结果为标准比较3种试条的敏感性和特异性。结果显示,574份犬血样中,221份PCR检测结果为阳性,阳性率为38.5%(221/574)。CA试条与犬用、人用rk39试条检测221份PCR阳性血样的敏感度分别为90.0%（199/221）、65.2%（144/221）和64.3%（142/221）,三者与PCR检测结果的差异具有统计学意义（χ² = 141.76、20.31、17.96,P < 0.05）。CA试条与犬用、人用rk39试条检测316份流行区非内脏利什曼病感染犬血样,特异度分别为95.3%（301/316）、99.7%（315/316）和99.4%（314/316）,三者与PCR检测结果的差异具有统计学意义（χ² = 258.85、312.01、308.05,P < 0.05）。CA试条与犬用、人用rk39试条检测流行区有症状犬的结果与PCR检测阳性的符合率分别为88.5%（23/26）、88.5%（23/26）和84.6%（22/26）,三者与PCR检测的差异有统计学意义（χ² = 23.22、23.22、20.50,P < 0.05);检测流行区无症状犬的结果与PCR检测阳性的符合率分别为91.8%（179/195）、62.1%（121/195）和61.5%（120/195）,3种试条与PCR检测结果的差异有统计学意义 （χ² = 136.25、11.33、10.39,P < 0.05）。运用曲线下面积（AUC）方法评估3种试条,其准确性均在0.8以上。kappa一致性检验系数为0.671~0.857,表明3种试条均适用于利什曼感染犬的初筛,CA试条优于犬用、人用rk39试条,是流行地区开展利什曼感染犬筛查较好的检测方法。

Blood samples were collected from 537 domestic dogs in Wenxian County and Diebu County of Gansu Province, where canine visceral leishmaniasis were endemic, and from 37 domestic dogs from Tianzhu County of Gansu Province, the non-endemic area. Blood examination was carried out using three immunochromatographic test strips including human rk39 strip, canine rk39 strip and circulating antigen (CA) strip in parallel to assess the sensitivity and specificity of the methods, and using PCR assay as the standard to compare with. The results showed that among the total 574 dog blood samples, 221 samples were tested Leishmania positive by PCR with the positive rate of 38.5% (221/574). Examining the 221 positive samples with CA strip, the human rk39 strip and canine rk39 strip, the sensitivity was found of 90.0% (199/221), 65.2% (144/221) and 64.3% (142/221), respectively, all of which significantly differed from that of PCR method (χ² = 141.76, 20.31, and 17.96, P < 0.05). The specificity of CA test strip and the two rk39 test strips in testing the blood samples of 316 dogs without Leishmania infection from the endemic areas was 95.3% (301/316), 99.7% (315/316) and 99.4% (314/316), respectively, all of which significantly differed from that of the standard (χ² = 258.85, 312.01, and 308.05, P < 0.05). The positive-result coincidence rate of the CA strip and the canine and human rk39 strips with PCR in dogs with symptoms from the endemic areas was 88.5% (23/26), 88.5% (23/26) and 84.6% (22/26), respectively, all of which significantly differed from that of the PCR method (χ² = 23.22, 23.22, and 20.50, P < 0.05). The positive-result coincidence rate of CA test strip and the canine and human rk39 test strips with PCR in dogs without symptoms in the endemic areas was 91.8% (179/195), 62.1% (121/195) and 61.5% (120/195), respectively, all of which significantly differed from that of the PCR method (χ² = 136.25, 11.33, and 10.39, P < 0.05). Area under the curve (AUC) analysis was used to assess the method accuracy, showing the AUC value of all the 3 methods was above 0.8. The kappa consistency coefficient was 0.671-0.857, indicating that all of the three test strips are applicable for preliminary screening of Leishmania-infected dogs. The CA test strip is superior to the two rk39 test strips and could be a good method for screening Leishmania-infected dogs in endemic areas.