ERCP后感染和抗生素预防：11年间连续的质量改善途径[述评]

xyin77

Lee JG, Lee CE. Infection after ERCP, and antibiotic prophylaxis: a sequential quality-improvement approach over 11 years. Gastrointest Endosc. 2008 Mar;67(3):476-7.
题目：ERCP后感染和抗生素预防：11年间连续的质量改善途径[述评]
作者：John G. Lee, Christopher E. Lee （加利福尼亚大学Irvine医学中心）
来源：胃肠内镜杂志 [Gastrointestinal Endoscopy (GIE)], 2008,67(3):476-478.
大部分患者在ERCP前不需要抗生素预防，但对原发性硬化性胆管炎、假性囊肿和肝门部胆管癌患者给予抗生素预防是合适的，因为对他们而言彻底引流不太可能。
预防性抗生素应用是为了防止ERCP后的感染性心内膜炎和其他感染。尽管ERCP后两类感染（感染性心内膜炎和其他感染）并不常见，该并发症潜在的发病率和死亡率和实施预防性抗生素应用的简便性之间的衡量，促使他们应用（抗生素预防）。我们对内镜检查感染风险的理解在进一步进展中。比如，美国胃肠内镜学会（ASGE）认为对梗阻性胆道系统的ERCP检查是高危操作，对高危心脏条件患者需要预防心内膜炎的发生。然而，美国心脏学会的最新推荐不支持任何对GI内镜检查行抗生素预防，包括梗阻性胆道系统的ERCP。
ERCP后非心源性的感染风险（认为约1%）不常见，使得开展抗生素预防的随机对照研究很困难和不现实。尽管荟萃分析（meta-analysis）证实抗生素预防不减少ERCP后的感染，不推荐（反对）预防性抗生素应用；ASGE则支持对实施ERCP的胆道梗阻或胰腺囊肿患者应用抗生素预防。
本期的胃肠内镜杂志（Gastrointestinal Endoscopy），Cotton等报告了他们长达11年的关于减少ERCP前抗生素预防的质量改善方式的结果。从1994年7月到1996年12月，他们对95%的实施ERCP的患者静脉应用广谱抗生素，回顾发现仅11例（0.48%）发生操作后感染。这促使他们于次年将预防方案改为口服环丙沙星，更在随后4年里，该口服方案仅用于怀疑有胆管或胰管梗阻和免疫抑制的患者。最后，于2002-2005年间，他们只对预计引流不彻底（原发性硬化性胆管炎、假性囊肿、胆囊结石或肝门部胆管癌）或免疫抑制的患者给予口服环丙沙星，对不成功的内镜下引流患者给予静脉环丙沙星预防。作者没有提到接受静脉环丙沙星的患者数，但提示第三阶段期间的操作技术高成功率，达到98%。同样地，我们假设替代的广谱抗生素，除了用于对环丙沙星过敏的患者，也对已用着抗生素的患者继续应用。
该质量改善方案将ERCP患者的抗生素预防应用从95%减至26%，但未显著增加ERCP后的感染。6150例（预防性）治疗患者中25例（0.42%）、5334例未治疗患者中7例（0.13%）发生感染；在最后阶段的2002-2005年间，预防性治疗和未预防性治疗者的感染发生率分别为0.47%（5/1065）和0.13%（4/2974）。
多因素分析中(OR = 5.2; 95% CI, 2.4-11.2)，肝移植术后狭窄是和感染显著相关的唯一变量。整个研究阶段，感染风险显著下降(OR=0.89; 95% CI, 0.80-0.98)，尽管减少了抗生素预防的应用——也许反映了内镜下引流的进步。
我们同意作者的观点，大部分患者在ERCP前不需要抗生素预防。美国心脏学会已经最终澄清心内膜炎预防的问题，但不推荐对GI内镜检查（给予预防性抗生素）。另外，不会有很多人对未成功内镜下引流的患者给予预防性抗生素有争议。尽管对原发性硬化性胆管炎、假性囊肿和肝门部胆管癌患者给予抗生素预防是合理的，对他们而言彻底引流不太可能，本文作者们完成的非常高的引流成功率也反对对这些患者常规抗生素预防，因为大部分将获得成功的引流。我们同意作者的有“可预测无法引流胆道系统”的患者是抗生素预防的适应证，但这具有何种意义还需要进一步阐述。此外，非常高的技术成功率提示，他们单位的大部分患者能获得成功的引流。一个适当的折衷方案可能是延迟抗生素治疗，直到确认内镜下引流失败。
为何先前的肝移植和ERCP后感染相关的原因还不清楚。免疫抑制看起来是个合乎逻辑的原因，但假如这样的话，是否意味着AIDS、肝硬化和其他免疫抑制状态也是抗生素预防的适应证？另一个答案是，与环孢霉素 A应用或血管损伤相关的胆汁成分或胆流的改变，会增强移植后的感染风险，这和免疫抑制无关。作者提出，对先前有胆道操作且免疫抑制的患者给予持续抗生素预防，但他们未清晰地定义这组患者。ASGE没有对免疫受损患者正式推荐或反对抗生素预防方案；我们没有给予常规预防性抗生素，除了引流不佳的患者，无关免疫状态。
报导的低发生率的感染并发症来自受到高度尊敬的、很有专长的、且开展大量ERCPs病例的中心，（这种经验）不能推广到其他医院。Cotton等强调推祝造影剂前抽吸胆汁的重要性，以减少胆管内压力和胆汁??猜龇盗鞯姆缦眨?约巴瓿芍Ъ芤?鳎ǖǖ溃?Ｎ颐且睬苛曳炊远愿蚊挪坎”涞幕颊呤凳┩耆?牡ǖ涝煊埃桓??返氖牵?挥性诘妓糠胖糜谙琳??耍ㄔ诳吹降妓肯喙氐牡ü芏危┖蟛攀凳┑ǖ涝煊埃??廴痉缦战档阶畹汀８?玫姆椒ㄊ遣捎肕RCP和/或EUS，在ERCP前对需要防止支架引流的胆管给予定位，使ERCP操作的造影剂降到最少或不需要。
ERCP后感染发生于0.42%接受单剂量抗生素预防的患者。但不清楚是否更长时间的治疗会降低感染至低危风险患者的0.13%水平。我们的实践（经验）是仅对引流不完全患者给予单剂量静脉广谱抗生素（阿莫西林-舒巴坦或环丙沙星），随后给予5-7天口服抗生素治疗，不管基础免疫状态或胆道污染。
总之，引流不完全是ERCP时预防性抗生素应用的主要原因；是否操作前或后应用（抗生素预防）得看操作单位的成功率。制止抗生素预防、直至明确的引流失败，可以进一步降低高成功率单位的抗生素用量。
Lee JG, Lee CE. Infection after ERCP, and antibiotic prophylaxis: a sequential quality-improvement approach over 11 years. Gastrointest Endosc. 2008 Mar;67(3):476-7.

Most patients do not require antibiotic prophylaxis before ERCP, but it seems reasonable to give antibiotic prophylaxis in patients with primary sclerosing cholangitis, pseudocyst, and hilar tumor, for whom complete drainage might not be possible.

Antibiotic prophylaxis is given to prevent infective endocarditis and other infections after ERCP. Although both types of infections are uncommon after ERCP, the potential morbidity and mortality of such complications weighed against the ease of giving prophylactic antibiotics have driven their use. Our understanding of the infectious risk of endoscopy has continued to evolve. For example, the American Society for Gastrointestinal Endoscopy (ASGE) considered ERCP in obstructed biliary systems to be a high-risk procedure requiring endocarditis prophylaxis in high-risk cardiac conditions.1 However, the most recent recommendations from the American Heart Association do not endorse antibiotic prophylaxis for any GI endoscopy, including ERCP in an obstructed biliary system.2

The risk of noncardiac post-ERCP infections (thought to be approximately 1%) is uncommon enough to make randomized controlled studies of antibiotic prophylaxis difficult and impractical. Although a meta-analysis showed that antibiotic prophylaxis did not reduce post ERCP infections, and recommended against it,3 the ASGE endorsed it in patients with biliary obstruction or pancreatic cysts undergoing ERCP.1 The actual clinical use of prophylaxis before ERCP has not been well documented.

In this issue of Gastrointestinal Endoscopy, Cotton et al4 report the results of their quality improvement project that reduced pre-ERCP antibiotic prophylaxis over an 11-year period. They administered broad-spectrum intravenous antibiotics to 95% of patients undergoing ERCP from July 1994 to December 1996 and retrospectively documented only 11 (0.48%) postprocedure infections. This prompted them to change prophylaxis to oral ciprofloxacin for the following year, which was then given only to patients with suspected biliary or pancreatic obstruction and immunosuppression over the next 4 years. Finally, between 2002 and 2005 they gave oral ciprofloxacin only to patients with predicted incomplete drainage (primary sclerosing cholangitis, pseudocyst, gallstones, or hilar tumor) or immunosuppression, and intravenous ciprofloxacin prophylaxis after unsuccessful endoscopic drainage. The authors did not state the number of patients receiving intravenous ciprofloxacin, but a high rate of technical success is implied from the 98% success rate recorded during the third phase. Likewise, we assume that an alternate broad-spectrum antibiotic was given to patients with allergy to ciprofloxacin and that patients already on antibiotics continued them.

This quality improvement project reduced the use of prophylaxis for patients undergoing ERCP from 95% to 26% without significantly increasing post-ERCP infections. Infection developed in 25 of 6150 treated patients (0.42%) and in 7 of 5334 untreated patients (0.13%); infections developed in 5 of 1065 patients (0.47%) and in 4 of 2974 patients (0.13%), respectively, during the last phase between 2002 and 2005. The slightly higher rate of infection observed after prophylaxis probably reflects patient selection, because prophylaxis would not have been given in the absence of any known risk factors for infection. Thus, the 0.13% rate could be considered as the true rate of infectious complications resulting from ERCP unrelated to any known underlying risk factors.

Post–liver transplant stricture was the only variable significantly associated with infection on multivariate analysis (OR = 5.2; 95% CI, 2.4-11.2.) Risk of infection fell significantly (OR=0.89; 95% CI, 0.80-0.98) over the study period, despite reducing antibiotic prophylaxis—perhaps reflecting improvements in endoscopic drainage.

We agree with the authors that most patients do not require antibiotic prophylaxis before ERCP. The American Heart Association may have finally settled the issue of endocarditis prophylaxis by not recommending it for any GI endoscopy. Alternatively, not many will argue with giving prophylaxis after unsuccessful endoscopic drainage. Although it seems reasonable to give antibiotic prophylaxis to patients with primary sclerosing cholangitis, pseudocyst, and hilar tumor, for whom complete drainage might not be possible, the very high success rate of drainage achieved by the authors argues against routine prophylaxis in such patients, because almost all will have successful drainage. We agree with the authors that prophylaxis is indicated in patients with a “predictably undrainable biliary system,” but just what this means remains to be clarified. Again, the very high rate of technical success suggests that most patients can be successfully drained in their unit. A reasonable compromise might be to delay antibiotic treatment until endoscopic drainage has clearly failed.

It is unclear why prior liver transplantation was associated with post-ERCP infection. Immunosuppression seems like a logical reason, but if so, does this mean that prophylaxis is also indicated in patients with AIDS, cirrhosis, and other immunosuppressed states? Alternatively, a change in bile composition or flow related to cyclosporine use or vascular injury could have increased the risk of infections after liver transplantation, independent of immunosuppression.5 The authors propose continuing antibiotic prophylaxis in patients with prior biliary manipulation who are also immunosuppressed but did not precisely define this group. The ASGE does not have a formal recommendation for or against prophylaxis in immunocompromised patients; we do not give routine prophylactic antibiotics, except for patients with suboptimal drainage, regardless of the immune status.

The low rate of infectious complications reported is from a highly respected, expert center performing a very large number of ERCPs and may not be generalizable to other practices. Cotton et al4 emphasize the importance of aspirating bile before injecting contrast to decrease the intrabiliary pressure and the risk of cholangiovenous reflux, as well as completing drainage with stenting. We also strongly recommend against performing a full cholangiography in patients with hilar lesions; rather, cholangiography should be performed only after placing a guidewire proximal to the stricture (and only visualizing the segment with the wire) to minimize the risk of contamination. An even better approach is to use MRCP and/or EUS to localize the stentable segments before ERCP and perform the procedure with minimal to no contrast injection.

Post-ERCP infection was observed in 0.42% of patients treated with a single-dose prophylaxis. It is unclear whether longer treatment would have reduced the risk of infection to the 0.13% level observed in the low-risk patients. Our practice is to treat only patients with incomplete drainage with a single dose of intravenous broad-spectrum antibiotics (either ampicillin with sulbactam or ciprofloxacin) followed by 5 to 7 days of oral antibiotic therapy, regardless of underlying immune status or biliary contamination.

In conclusion, incomplete drainage is the main reason for administering antibiotic prophylaxis in ERCP; whether this is done before or after the procedure depends on the success rate of the unit. Withholding antibiotic prophylaxis until drainage clearly fails could reduce antibiotic usage further in units with high success rates.