第一部分：中国手术患者围手术期异体红细胞输注情况的纵向观察研究

目的

我国于2014年颁布《围手术期输血指南》，提倡采用限制性输血策略，以减少异体红细胞输注。本研究旨在比较限制性输血策略实施前后，围手术期异体红细胞输注情况的变化。

方法

本研究是一项基于人群横断面数据形成的纵向观察研究，数据来源于医院质量监测系统数据库，研究对象为2013年1月1日至2018年8月31日我国三级医院的手术患者。主要观察指标是围手术期异体红细胞输注。通过构建包含混合效应的Logistic回归模型分析不同年份、地区间的异体红细胞输注情况，并对五类术中失血风险较高的代表性手术（颅内动脉瘤切除术或颅内畸形血管切除术、开胸二尖瓣置换手术、开胸肺叶切除术、开腹胃切除术及髋关节置换术）患者进行分组分析，以控制手术方面的混杂因素。

结果

在我国三级医院的36092159名手术患者中，702067名患者（1.95%, 95% CI：1.94%～1.95%）在围手术期接受了异体红细胞输注。在2013年至2018年间，粗输注率由2.24%（95% CI：2.22%～2.25%）下降至1.84%（95% CI：1.81%～1.87%）；校正人群的性别及年龄差异后，仍可观察到类似的下降趋势。针对五类行代表性手术的单病种患者分析后发现，各手术类型患者的粗输注率也均有下降。颅内动脉瘤切除术或颅内畸形血管切除术患者的粗输注率由9.48%（95% CI：8.18%～10.96%）下降至7.54%（95% CI：5.48%～10.28%）；开胸二尖瓣置换术患者的粗输注率由21.78%（95% CI：21.11%～22.48%）下降至18.93%（95% CI：16.87%～21.19%）；开胸肺叶切除术患者的粗输注率由6.89%（95% CI：6.52%～7.29%）下降至4.37%（95%CI：3.49%～5.46%）；开腹胃切除术患者的粗输注率由7.19%（95% CI：6.90%～7.48%）下降至7.09%（95% CI：6.60%～7.61%）；髋关节置换术手术患者的粗输注率由17.34%（95% CI：16.85%～17.85%）下降至7.03%（95% CI：6.14%～8.02%）。在各类手术中用混合效应Logistic回归模型校正患者性别、年龄及地区因素后，异体红细胞输注率仍呈下降趋势，但有小范围波动。以2016或2017年的输注率最低，2018年的稍有回升。我们还发现不同地区间的粗输注率差异悬殊，以西北地区的相对较高，而华东地区的则相对较低，可相差3~10倍。校正患者年龄、性别及年份因素后，仍存在明显的地区间差异。

结论

2013年至2018年，我国三级医院手术患者围手术期异体红细胞的输注率呈下降趋势，但存在显著的地区间差异。

     第二部分：围手术期输注血浆对非大量输血的手术患者在院期间死亡

     的影响的回顾性队列研究

     目的

血浆的输注指征及具体输注方案尚有争议，却在围手术期经常使用。尚不确定我国目前的血浆输注实践与手术患者不良结局之间的关系。本研究旨在探究我国非大量输血的手术患者输注血浆后的在院期间死亡情况，为围手术期血浆管理提供参考。

方法

本研究是一项回顾性队列研究，数据来源于医院质量监测系统数据库，研究对象为2013年1月1日至2018年8月31日我国三级医院输注红细胞但输注量不超过10单位的手术患者。通过建立多因素Logistic回归模型，分析围手术期输注血浆与在院期间死亡之间的关联。并按照手术类型（心脏外科手术、血管外科手术、胸或腹部手术、骨科手术）及红细胞输注量进行亚组分析，以减少混杂因素的影响。次要结局包括手术切口感染、院内感染、急性呼吸窘迫综合征以及住院时长。

结果

在69319名输注红细胞但不超过10单位的手术患者中，34827名患者（50.24%）在围手术期接受了血浆输注，808名患者（1.17%）在院期间死亡。我们发现围手术期输注血浆与在院期间死亡率增高有关（每输注100 ml血浆Odds ratio [OR]，1.05；95% CI，1.04-1.06)。且在心脏外科手术、血管外科手术、胸或腹部手术患者中，以及围手术期输注>0且≤2单位、>2且≤4单位、>4且≤6单位、>6且≤8单位和>8且≤10单位红细胞的患者中均发现类似的联系。相较于红细胞输注量较高组（>8且≤10单位），输注量较低组（>0且≤2单位）的患者每输注100ml血浆增加的在院期间死亡风险更大。另外，围手术期输注血浆与手术切口感染（每输注100 ml血浆OR，1.03；95% CI，1.02-1.04）、院内感染（每输注100 ml血浆OR，1.03；95% CI，1.02-1.04）和急性呼吸窘迫综合征（每输注100 ml血浆OR，1.03；95% CI，1.00-1.05）的发生危险增高有关，且住院时间更长（每输注100 ml血浆hazard ratio [HR]，1.05；95% CI，1.04-1.07）。

结论

围手术期输注红细胞但输注量不超过10单位的患者中，输注血浆与在院期间死亡等不良结局的发生危险增加有关。

Part I: Temporal and Geographic Variations of Perioperative Red Blood Cell Transfusion: A Longitudinal Observational Study in China from 2013 to 2018

Objective

Restrictive transfusion strategy is recommended in “Perioperative Transfusion Guidelines (2014, China)” to limit the use of allogeneic red blood cell (RBC) transfusion. We aimed to examine whether the transfusion practice had changed.

Methods

This study was a longitudinal population-based study. Using the Hospital Quality Monitoring System database, patients undergoing surgeries between January 1st, 2013 and August 31st, 2018 in tertiary hospitals of mainland China were included. The main outcome was allogeneic RBC transfusion. Year- and area-specific crude prevalence and age/sex -adjusted prevalence with 95% confidence intervals (CI) of allogeneic RBC transfusion were estimated using mixed-effects logistic regression models. Subgroup analysis was performed for patients undergoing surgeries with higher risk of bleeding (intracranial aneurysmectomy/arteriovenous malformation resection, open surgery for mitral valve replacement, open surgery for lobectomy, open surgery for gastrectomy and hip replacement) to control surgical confounding.

Results

Of 36092159 patients, a total of 702067 (1.95%, 95% CI 1.94%-1.95%) patients received allogeneic RBC transfusion. From 2013 to 2018, the crude prevalence of RBC transfusion decreased from 2.24% (95%CI: 2.22%~2.25%) to 1.84% (95%CI: 1.81%~1.87%). The decrease remained robust after adjusting for age and sex. Similar trends also observed in patients undergoing intracranial aneurysmectomy/arteriovenous malformation resection [9.48% (95%CI: 8.18%~10.96%) to 7.54% (95%CI: 5.48%~10.28%)], open surgery for mitral valve replacement [21.78% (95%CI: 21.11%~22.48%) to 18.93% (95%CI: 16.87%~21.19%)], open surgery for lobectomy [6.89% (95%CI: 6.52%~7.29%) to 4.37% (95%CI: 3.49%~5.46%)], open surgery for gastrectomy [7.19% (95%CI: 6.90%~7.48%) to 7.09% (95%CI: 6.60%~7.61%)] and hip replacement [17.34% (95%CI: 16.85%~17.85%) to 7.03% (95%CI: 6.14%~8.02%)] respectively. After adjusting for age, sex and geographic differences, the decline in RBC transfusion prevalence still existed over time. The prevalence was lowest and below the average level in 2016 or 2017 while rose a bit in 2018. Geographically, the RBC transfusion prevalence showed wide variations. At the regional level, the crude prevalence was highest in Northwest China, while lowest in Eastern China. The highest prevalence was 3~10 times of the lowest. Similar variations occurred in each surgery subgroups after controlling for age, gender and year.

Conclusion

From 2013 to 2018, the prevalence of allogeneic RBC transfusion in surgical patients from tertiary hospitals in mainland China decreased, possibly due to the implementation of restrictive transfusion strategy. There still exists great variations of perioperative allogeneic RBC transfusion among different geographic regions.

Part II: Associations between Perioperative Plasma Transfusion and In-hospital Mortality in Patients without Massive Transfusion

Objectives

Plasma transfusions are frequently prescribed for surgical patients, although their indications lack a strong evidence base. The impact of perioperative plasma transfusion on outcomes of surgical patients, especially in those without massive transfusion, remains unclear. The objective of this study was to examine the association between perioperative plasma transfusion and in-hospital mortality and morbidity in patients with who are exposed to red blood cells (RBC) but less than 10 units.

Methods

This was a retrospective cohort study. Using the Hospital Quality Monitoring System database, surgical patients transfused with RBC but less than 10 units between January 1st, 2013 and August 31st, 2018 in tertiary hospitals of mainland China were included. We built multivariable logistic regression models for perioperative plasma transfusion and in-hospital mortality. Subgroup analysis was performed for patients undergoing surgery with higher risk of plasma transfusion (cardiac, vascular, thoracic or abdominal and orthopedic surgery) and for patients with various units of RBC transfusion to control confounding. The secondary outcomes included surgical site infection (SSI), nosocomial infection, acute respiratory distress syndrome (ARDS) and length of hospital stay (LOS).

Results

Of 69319 patients, 34827 (50.24%) received plasma transfusion and 808 patients died during hospitalization (1.17%). Perioperative plasma transfusion was associated with an increased risk of in-hospital death (Odds ratio [OR] per 100ml plasma transfusion, 1.05; 95%CI, 1.04-1.06). Similar association was also observed in patients undergoing cardiac, vascular, thoracic or abdominal and orthopedic surgery. Dividing patients by the units of RBC transfusion, association between plasma transfusion and mortality still existed. And compared with patients who were transfused with RBC no more than 2 units, patients transfused with RBC more than 8 units suffered more from per 100ml plasma transfusion. Moreover, plasma transfusion was also associated with an increased risk of SSI (OR per 100ml plasma transfusion, 1.03; 95% CI, 1.02-1.04), nosocomial infection (OR per 100ml plasma transfusion, 1.03; 95% CI, 1.02-1.04), ARDS (OR per 100ml plasma transfusion, 1.03; 95% CI, 1.00-1.05) and longer LOS (hazard ratio [HR] per 100ml plasma transfusion, 1.05; 95% CI, 1.04-1.07).

Conclusions

In patients transfused with red blood cells but less than 10 units, perioperative plasma transfusion is associated with a higher risk of in-hospital mortality and morbidity.