肠外瘘患者确定性手术前非甲状腺疾病综合征对术后手术部位感染的预测价值

Predictive value of non.thyroidal illness syndrome before definitive operation on postoperative surgical site infection in patients with enterocutaneous fistula

目的探讨肠外瘘患者确定性手术前非甲状腺疾病综合征（NTIS）对术后手术部位感染（SSI）的预测价值。方法采用回顾性病例对照研究方法。收集2014年4月至2016年11月解放军南京总医院收治的264例行确定性手术治疗的肠外瘘患者（181例甲状腺功能正常和83例NTIS患者）的临床资料．行确定性手术治疗后86例患者发生SSI设为SSI组，178例患者未发生SSI设为无SSI组。观察指标：（1）术后发生SSI的相关风险因素分析。（2）术前NTIS对术后SSI的影响。（3）血清游离三碘甲腺原氨酸（FT3）水平对术后SSI的预测能力。正态分布的计量资料采用x±s表示，采用独立样本t检验；计数资料采用绝对数或百分比表示，采用疋。检验。等级资料比较采用非参数检验。多因素分析采用Logistic回归模型。绘制受试者工作特征（ROC）曲线，并计算曲线下面积（AUC）分析血清FT3水平对术后SSI的预测能力。结果（1）术后发生SSI的相关风险因素分析：SSI组患者术前经瘘口肠液丢失量、术前血红蛋白水平、手术切除部位、手术方式分别为〈200mL／24h65例、200—500mL／24h15例、〉500mL／24h6例，（119±36）g/L，胃和十二指肠5例、小肠50例、回结肠31例、结直肠36例（部分患者手术切除多个部位），开腹58例、腹腔镜辅助28例，无SSI组患者分别为135、27、16例，（125±39）g/L，1l、9l、53、71例，127、51例，两组上述指标比较，差异无统计学意义（X2=0．471，t=1．202，X2=0．332，0．422，P〉0．05）。SSI组患者术前瘘口类型、术前血清白蛋白水平、术前NTIS、术中出血量、手术持续时间分别为单发57例、多发29例，（35±8）g/L，36例，〈300mL67例、≥300mL19例，〈3h53例、≥3h33例，无SSI组患者分别为146、32例，（37±9）g／L，47例，161、17例，140、38例，两组上述指标比较，差异有统计学意义（X2=8．089，t=2．422，X2=6．426．7．746，8．547，P〈0．05）。多因素分析结果表明：术前多发肠瘘和术前NTIS是影响肠外瘘患者术后发生SSI的独立影响因素（比值比=1．873．2．464，95％可信区间：1．052～2．671，1．120～4．392）。（2）术前NTIS对术后SSI的影响：术前NTIS患者的术前多发性肠瘘发生率、术前肠内营养时间〉3个月比例、术后SSI发生率分别为31．3％（26／83）、72．3％（60／83）、43．4％（36／83），术后SSI中的浅层切口感染发生率、深层切口感染发生率、器官腔隙感染发生率分别为9．6％（8／83）、21．7％（18／83）、7．2％（6／83），术前甲状腺功能正常患者分别为19．3％（35／181）、57．5％（104／181）、27．6％（50／181）、11．6％（21／181）、3．9％（7／181）、8．8％（16／181）；两者术前多发性肠瘘发生率、术前肠内营养时间〉3个月比例、术后SSI发生率、浅层切口感染发生率、深层切口感染发生率比较，差异均有统计学意义（X2=4．603，5．319，6．426，4．256，4．377，P〈0．05）；而器官腔隙感染发生率比较，差异无统计学意义（X2=0．193，P〉0．05）。（3）血清FT3水平对术后SSI的预测能力：ROC曲线显示血清FT3预测术后SSI的最佳截值点为3．5pmol／L，AUC为0．75，灵敏度为72．6％，特异度为68．7％。结论肠外瘘患者确定性手术前存在的NTIS与术后SSI的发生相关，血清FT3预测术后SSI的最佳截值点为3．5pmol／L。

Objective To investigate the predictive value of non-thyroidal illness syndrome （NTIS） before definitive operation on postoperative surgical site infection （SSI） in patients with enteroeutaneous fistula （ECF）. Methods The retrospective case-control study was emrdueted. The clinical data of 264 ECF patients （181 with euthyroidism and 83 with NTIS） who underwent definitive operation in the Nanjing General Hospital of Nanjing Military Command between April 2014 and November 2016 were collected. After definitive operation, 86 with SSI and 178 without SS1 were respectively allocated into the SSI group and non-SSI group. Observation indicators： （1） risk factor analysis of postoperative SSI; （2） effect of preoperative NTIS on postoperative SSI; （3） predictive power of serum fi＇ee triiodothyronine 3 （FT3） level on postoperative SSI. Measurement data with normal distribution were represented as x-＋-s and was analyzed using the t test. Count data were described as absolute number or percentage, and were analyzed using the chi-square test. The comparison of ordinal data was done by the nonparamentrie test. The multivariate analysis was done using the logistic regression model. The receiver operating characteristic （ROC） curve was drawn, and area under the curve （AUC） was calculated for analyzing predictive power of serum FT3 level on postoperative SSI. Results （ 1 ） Risk factor analysis of postoperative SSI ： cases with volume of preoperative intestinal fluid loss through fistula stoma 〈 200 mL/24 hours, from 200 to 500 mL/24 hours and 〉 500 mL/24 hours, preoperative hemoglobin （Hb） level, cases with surgical site located in stomach and duodenum, small intestine, ileocolon and coloreetum, cases with open surgery and laparoseopic surgery were respectively 65, 15, 6, （119＋36）g/L, 5, 50, 31, 36, 58, 28 in the SSI group and 135, 27, 16, （125±39）g/L, 1l, 91, 53, 71, 127, 51 in the non-SSI group, with no statistically significant difference between groups （X2= 0. 471, t = 1. 202, X2 = 0. 332, 0. 422, P〉0. 05）. Cases with preoperative single and multiple fistula stoma, serum albumin （Alb） level, cases with preoperative NTIS, volume of intraoperative blood loss 〈 300 mL and 〉i 300 mL, operation duration 〈 3 hours and ~〉 3 hours were respectively 57, 29, （35_＋8）g/L, 36, 67, 19, 53, 33 in the SS1 group and 146, 32, （37±9）g/L, 47, 161, 17, 140, 38 in the non- SSI group, with statistically significant differences between groups （X2 = 8. 089, t = 2. 422, X2 = 6. 426, 7. 746, 8. 547, P〈0. 05）. Results of multivariate analysis showed that preoperative muhiple intestinal fistula and NTIS were independent factors affecting occurrence of postoperative SSI in ECF patients （odds ratio = 1. 873, 2. 464, 95% confidence interval： 1. 052-2. 671, 1. 120-4. 392）. （2） Effect of preoperative NTIS on postoperative SSI： incidence of preoperative multiple intestinal fistula, proportion of cases with preoperative enteral nutrition time 〉 3 months, incidence of postoperative SS1, postoperative superficial and deep incision infection rates and organ/ space infection rate were respectively 31.3% （ 26/83 ）, 72. 3% （ 60/83 ）, 43.4% （ 36/83 ） , 9.6% （ 8/83 ） , 21.7%（18/83）, 7.2% （6/83） in patients with NTIS and 19. 3% （35/181）, 57.5% （104/181）, 27.6% （50/181）, 11.6%（21/181）, 3.9%（7/181）, 8.8% （16/181） in patients with euthyroidism, with statistically significant differences in incidence of multiple intestinal fistula, proportion of cases with preoperative enteral nutrition time 〉 3 months, incidence of postoperative SS[, superficial and deep incision infection rates （X2 = 4. 603, 5. 319, 6. 426, 4. 256, 4. 377, P〈 0. 05）, and no statistically significant difference in organ/space in{eetion rate （X~＂ =0. 193, P〉0. 05）. （3） Predictive power of serum FT3 level on postoperative SSI： the ROC curve showed that optimal cut-off point of serum FT3 predicting postoperative SSI was 3.5 pmoL/L, AUC, sensibility and specificity were respectively 0. 75, 72. 6% and 68.7%. Conclusion The presence of NTIS is associated with occurrence of postoperative SS1 in patients with ECF before definitive operation, and optimal cut-off point of serum FT3 predicting postoperative SSI is 3.5 pmol/L.