点击显示 收起
加州大学洛杉矶分校(UCLA)退伍军人医院的新研究显示，胰管十二指肠切除术(PD)对胰脏可切除或可手术移除的腺瘤有缓解治疗效果。在美国，每年有37,000例胰脏癌，是癌症死亡的第四名；PD包括切除胰脏和十二指肠，仅适用10%的胰脏癌病患。研究作者UCLA手术结果与品质中心的外科住院医师Irina Yermilov建议，虽然切除是唯一治疗方式，但治愈率相当低— 约80%到90%的病患死亡。这些发现发表于美国外科医学会第93届临床讨论会。
【研究者建立一个新的资料库】
虽然PD显示有一些存活利益且较少再住院，但研究者希望知道，即使治愈率低，手术是否至少可以缓解胰脏癌症状。研究在PD之后长期追踪检视再度住院的人数、时间和地区；再度住院的确定原因；以及辨识可以用来预测再度住院的病患因素。为了完成这些，研究团队整合三个已知的资料库建立一个新的资料库：一个显示在加州癌症注册中心的所有病患，一个显示所有住院病患档案/该州非联邦医院所有可能之再度住院病例，另一个显示该州所有的死亡个案(有可信的死亡率资料)。
在29,523位病患中，1994到2004年间有1,802位病患因为胰脏癌进行PD，被纳入此研究，使用“International Statistical Classification of Diseases and Related Health Problems, Ninth Edition (ICD-9) ”之PD编码，全部都有完整的分期资讯；因为这是长期研究，全部病患中有6% (108)因为在30天内死亡而被排除。研究者辨识所有的再住院资料，直到病患死亡为止，检查所有的 ICD-9 编码，以确定再度住院之原因，使用单一变项和多变项回归，以辨识任何可以用来预测再度住院的因素。
研究对象的平均年龄是66岁，56%是淋巴结阳性，平均存活期间是17个月，存活者追踪资讯期间最长达42.2个月。
这1,802位病患中，19%在30天内，57%在1年内，以及74%在4年内再度住院；46%再度住院是到原先手术医院以外的医院— 资料库中不一定可以追踪到的事情，平均再度住院次数是2次；整体而言，PD之后有77.5%再度住院；22.5%不曾再度住院。
末期胰脏癌病患中，70%到90%有胃肠道阻塞，30%到50%有胆管阻塞，80%到90%感受到腹痛；PD之后，研究中有39%病患因为脱水/营养不良，38%因为贫血，6%到16%因为GI阻塞，7%到12% 因为胆管阻塞，5% 到16%因为腹痛再度住院。
研究并未发现可以预测再度住院的明显病患因素。这77.5%的再度住院者中，56.4%是淋巴结阳性，平均住院天数为 18.2天；其他22.5%未曾再度住院者之中，55.7%是淋巴结阳性，平均住院天数为18.5 (整体P< .05)；两组中各类的肿瘤分期相似，两性或病患种族之间没有明显差异。
【研究的优缺点】
Yermilov医师引述该研究属于强度大、高品质的住院病患和癌症注册资料库，准确捕捉加州的所有再度住院资料，并提供基本资料，可以提供基本讯息，她列出一些研究限制，如ICD-9纪录不全以及缺乏准确的辅助治疗资料。
研究者结论认为，他们的发现提供更准确的PD后住院资讯，所观察的缓解程度和之前发表的文献相当。
约翰霍普金斯外科助理教授，研究谘询者Timothy Pawlik医师表示，这是一个重要的议题；以再度住院来评估PD的缓解程度的敏感性有多少？病患未再度住院的理由可能千奇百怪都有，研究并没有再度住院长度的详细资料，例如住院期间的必要介入方式、这对稍早住院或一年后才住院的病患可能会有所不同。
他进一步表示，我们估计手术期间的失血量？这或许可以作为手术困难度的标记。我认为有些严重缺点缺乏辅助资讯，例如，缓和只有有结果才有意义。什么是我们要尝试缓解的特定目标，缓解这些对改善生活品质的效果有多少？
Yermilov 医师回应表示，因为这是初步研究，我们无法检视所有资料；举例来说，我们没有可以估计资料库中血液损失量的方式，同样的，没有人像我们一样探讨再度住院，之前的研究显示PD之后的再度住院率只有11%到 26%。
Yermilov医师和Pawlik医师宣称没有相关财经关系。
美国外科医学会第93届临床讨论会。手术论坛S11。发表于2007年10月8日。
Resection for Pancreatic Cancer May Provide a Palliative Effect
By Lexa W Lee, ND
Medscape Medical News
Pancreaticoduodenectomy (PD) may have benefits as a palliative therapy in resectable or surgically removable adenocarcinoma of the pancreas, a new study at the University of California, Los Angeles (UCLA), Veterans Administration Hospital, suggests. There are 37,000 cases of pancreatic cancer in the United States per year; it is the fourth leading cause of cancer death. PD, which involves resecting the pancreas and duodenum, is possible in only 10% of pancreatic cancer patients. "While the only cure is resection, the cure rate is low — 80% to 90% of patients die," commented study author Irina Yermilov, MD, surgical resident at the UCLA Center for Surgical Outcomes and Quality. The findings were presented here at the American College of Surgeons 93rd Clinical Congress.
Researchers Created a New Database
Although PD has been shown to have some survival benefit with fewer readmissions, the researchers wanted to know whether the procedure could at least palliate symptoms of pancreatic cancer in view of the low cure rate. The study examined long-term follow-up after PD by examining the number, timing, and location of readmissions; determining reasons for readmissions; and identifying patient factors that may predict for readmission. To accomplish this, the research team created a new database by cross-linking 3 known databases: one that showed all patients with pancreatic cancer in the California Cancer Registry, another that showed all inpatient files/all possible readmissions to nonfederal hospitals in the state, and a third that showed all deaths in the state (for reliable mortality data).
Out of a total 29,523 patients, 1802 patients who had PD for pancreatic cancer from 1994 to 2004 were selected for the study, using International Statistical Classification of Diseases and Related Health Problems, Ninth Edition (ICD-9), codes for PD. All had complete staging information; as this was a long-term study, 6% of the patient total (108) with a mortality period of 30 days was excluded. The researchers identified all readmissions until death, examined all ICD-9 codes to determine reasons for readmission, and used univariate and multivariate regression to identify any factors that might predict for readmission.
The median age of the study group was 66 years, 56% were node-positive, median survival period was 17 months, and follow-up information for the survivors was available for 42.2 months.
Of the 1802 patients, 19% were readmitted within 30 days, 57% within a year, and 74% within 4 years. Forty-six percent were readmitted at a hospital other than where they were admitted for surgery — something that is not usually tracked in databases. The median number of readmissions was 2. Overall, 77.5% of the group was readmitted after PD; 22.5% were never readmitted.
Of patients suffering advanced pancreatic cancer, 70% to 90% have gastrointestinal obstruction, 30% to 50% have biliary obstruction, and 80% to 90% suffer abdominal pain. After PD, 39% patients in the study were readmitted for dehydration/malnutrition, 38% for anemia, 6% to 16% for GI obstruction, 7% to 12% for biliary obstruction, and 5% to 16% for abdominal pain.
No significant patient-level factors for predicting readmissions were found. Of the 77.5% readmitted, 56.4% were node-positive, with an 18.2-day mean length of stay; of the 22.5% never-readmitted group, 55.7% were node-positive, with an 18.5-day mean length of stay (P< .05, for all). Staging of tumors was similar at all levels between the 2 groups. There were no significant differences between genders or among races or patient ages.
Strengths and Shortcomings of Study
Dr. Yermilov cited the strengths of the study as the large, high-quality inpatient and cancer registry database, which accurately captures all hospital readmissions within California and provides baseline data, which can provide benchmark information. She listed some of the limitations of the study as ICD-9 undercoding and a lack of accurate adjuvant treatment data.
The researchers concluded that their findings provided a more accurate look at hospitalizations after PD and that palliation at the observed population levels agrees with previously published literature.
Timothy Pawlik, MD, assistant professor of surgery at Johns Hopkins and study discussant, said, "This is an important topic. How sensitive actually is readmission as a tool to evaluate how palliative PD is? There could be all sorts of reasons why a patient is not readmitted. The study didn't have details about length of stay for readmissions, such as interventions necessary during stay, may be very different for patients admitted earlier on than maybe a year later."
He continued, "Was estimated blood loss at time of surgery looked at? It could be a surrogate marker for difficulty of case. I think there are some serious shortcomings with lack of adjuvant data [for example,] kind of treatment. Palliation only makes sense relative to an end. What specific end are we trying to palliate, and how effective are we palliating those things that affect quality of life?"
Dr. Yermilov responded, "Since this was a preliminary study, we couldn't look at all the data. We had no access to estimated blood loss in the database, for example. Still, no one has looked at readmissions in the way we have. Previous studies have only shown an 11% to 26% readmission rate after PD."
Dr. Yermilov and Dr. Pawlik have disclosed no relevant financial relationships.
American College of Surgeons 93rd Clinical Congress: Surgical Forum S11. Presented October 8, 2007.