编者按：门静脉高压是肝硬化最常见的并发症，经过国内外研究者多年的艰难探索，其诊疗技术已经达到相当的水平。2017年的AASLD年会上，《国际肝病》特别策划了一场关于如何诊断门脉高压的圆桌会议，南方医科大学南方医院祁小龙教授主持，邀请了来自西班牙巴塞罗那肝脏血流动力学实验室主任、Hepatology副主编Juan Carlos Garcia-Pagan教授和巴塞罗那医院消化内镜中心Andres Cardenas教授，分别从侵入性HVPG检查和非侵入性手段在不同临床需求下的应用、HVPG临床实践的现状、及不同病因对无创技术准确性的影响三个方面进行了探讨。

 

一、侵入性与非侵入性诊断，分别适用于哪些临床需求？

 

祁教授：关于HVPG和无创性检测评估，现在我们已经有了不少可用的工具。对于不同的临床情况，如筛查、诊断、风险分层和预后预测，侵入性和非侵入性手段在其中发挥着什么样的作用？

 

Garcia-Pagán教授：这取决于我们想知道什么。如果我们在随访一位慢性肝病患者，会想知道他是否有门静脉高压的潜在并发症，是否已经发展成临床上显著的门静脉高压症，这时非侵入性评估就可以达到这样的效果，如FibroScan。

 

对于不能明确是否存在临床性门脉高压症的患者，我们可以选择内镜检查。如果内镜检查为阳性，说明患者有明显的门脉高压。如果内镜检查没有定论，可以认为患者没有静脉曲张，但仍不能判断患者的门脉压力梯度是否＞10mmHg，未来发展为静脉曲张的风险是否较高，这时候侵入性评估方法是比较好的，可检测肝静脉压力梯度（HVPG）。

 

HVPG仍然是目前评估患者治疗反应的金标准，尚没有无创手段可替代，FibroScan并不能很好的明确患者的压力梯度数值。我们正在研究可反映患者治疗应答的生物标志物，例如最近发表的数据表明在肝硬化、丙型肝炎以及丙型肝炎治愈患者中，HVPG的降低并不能很好地反映FibroScan发现的病变改善。另外，目前我们无法通过FibroScan评估HVPG对β受体阻滞剂的反应。我们需要继续努力，以期找到能将压力梯度变化联系起来的方法。

 

Cardenas教授：FibroScan可帮助判断慢性肝病患者是否存在肝硬化，如果测量值为10kPa，就意味着有可能是肝硬化；>15kpa说明极有可能存在肝硬化，且合并显著门脉高压；但对于10至15kpa范围的患者，我们不能明确是否合并门脉高压，然而在大多数的情况下，我们又是需要确认的。所以，确诊门静脉高压症最好是通过HVPG，可以联合活检判断患者是否存在肝硬化。

 

在静脉曲张的筛查方面，这些无创性的检查非常好。例如肝硬化患者的FibroScan值＜20，血小板计数＞150000，那么或许能避免内镜检查，因为这些患者更倾向于没有静脉曲张。值得注意的是，FibroScan检查的结果可以提示研究者谁不会有静脉曲张，但不太擅长判断谁有静脉曲张。

 

二、HVPG侵入性检测在临床实践中的现状？

 

祁教授：HVPG仍然是诊断的金标准，从您的角度来看，在目前的国际临床实践中，哪些门静脉高压患者迫切需要有创性HVPG进行危险分层？在这其中，又有多少比例的患者实际接受了有创HVPG检查？以经颈静脉肝内门体静脉分流术（TIPS）早期干预为例，Hepatology杂志最近发表了一项来自法国的真实世界研究表明，肝硬化出血住院患者有33%符合早期TIPS标准，但只有7%的患者实际接受了TIPS，这意味着早期TIPS干预在临床实际中存在很大的差距。

 

Cardenas教授：HVPG是最好的工具，但也有可靠的无创性检查，主要包括Child-Pugh评分等。Child-Pugh评分在10和13之间的患者可能会出现非常高的门静脉压力，这些患者都是最应该接受TIPS治疗的患者，他们不一定需要HVPG检测。另外，对于有急性食管胃底静脉曲张破裂出血且Child-Pugh评分小于14的患者，应先行TIPS治疗而非HVPG检测。

 

Garcia-Pagán教授：根据我自己在转诊中心的经验，有许多患者是表现为诊断困难的患者，这意味着我们对这些病例可能存在偏倚。有时，因为不正确的诊断研究，这些患者将被归类为肝硬化，但是一旦我们做了充分的评估，我们将会得出这样的结论：他们并没有肝硬化，但存在不明原因的门脉高压甚至门静脉血栓。

 

三、欧美指南中的非侵入性诊断是否适用于乙肝肝硬化门脉高压？

 

祁教授：肝硬化的不同病因会在门脉高压无创诊断中扮演重要作用。在我们中国门脉高压无创诊断研究组(CHESS)队列数据中，超过80%的患者为乙型肝炎相关肝硬化。然而欧美文献报道中的患者多为丙型肝炎或酒精相关肝硬化，乙型肝炎肝硬化只占较少的比例。这些文献中的无创诊断方法是否仍然适用于乙型肝炎相关肝硬化？

 

Garcia-Pagán教授：FibroScan的阈值可能的确有差别，但我认为没有太多的变化，在全球范围内均适用。我想最主要的问题是测量HVPG的方法，很多中心没有规范操作，不符合最低标准，这意味着存在巨大的可变性，使得数据不准确。

 

对于乙型肝炎患者，我们同时测量患者的HVPG，这种相关性与丙型肝炎相关肝硬化和酒精性肝硬化非常相似，但仍有差别。我认为，如果患者有实质性肝硬化，楔压与门脉压力有很好的相关性，但胆汁淤积性疾病不同。胆汁淤积性疾病患者的楔压不能反映门静脉的压力。

 

Cardenas教授：我同意这一观点。我们需要更多这方面的研究，从目前大多数病毒性肝炎研究看，乙型肝炎与丙型肝炎在肝硬化的评估上没有太大的区别。

 

祁教授：从CHESS研究组的数据来看，基于FibroScan的数据，乙型肝炎与丙型肝炎存在一定差异，其诊断阈值的区分值得研究。

 

Garcia-Pagán教授：可能是阈值的变化，可能需要校正那些阈值。这与NASH的阈值不同于丙型肝炎和酒精性肝炎的阈值一样，因为转氨酶水平影响炎症水平。

 

祁教授：目前，国际上门脉高压研究的样本量都非常有限，平均在100例左右，因此很难对特定病因进行亚组分析，这是我们CHESS研究组正在关注和解决的问题。

 

How to Diagnose Portal Hypertension: InvasiveHVPG vs. Noninvasive measurement?

 

Dr Qi: The topic today is how to diagnose portal hypertension - HVPG versus non-invasive measurement? I will introduce the participants in our discussion. We welcome Professor Juan Carlos Garcia-Pagan from the Hospital Clinic at Barcelona University, Spain. He is Senior Consultant in Hepatology and Associate-Editor at Hepatology. Professor Andres Cardenas is also from the Hospital Clinic at Barcelona University. We look forward to a discussion with one of the world’s top portal hypertension teams.

 

For HVPG and non-invasive assessment, we have a lot of tools available now. For different clinical situations, like screening, diagnosis and predictive prognosis, what are the roles of the invasive and non-invasive techniques?

 

Dr Garcia-Pagan: I think it depends what you want to know. If you are following a patient with a chronic liver disorder, you will want to know if they have the potential complication of portal hypertension and you need to see if they already have developed clinically significant portal hypertension. There are non-invasive methods that allow you to achieve this effectively (FibroScanelastography, for example) and there are times with ultrasound alone measuring portal vein diameter, when that is clear.

 

But there are still a few patients though, despite this, for whom you are unable to definitively say that they have clinical portal hypertension. For these difficult cases, you can decide to do endoscopy and if the endoscopy is positive, you will know that the patient has significant portal hypertension. If the endoscopy in inconclusive, you can say the patient has no varices but you do not know if the pressure gradient is >10mmHg and if the risk for varices to develop in the future is higher or not. So invasive methods are quite good, but there are still more patients for whom you will want to know the gradient and where you need to do HVPG.

 

If you want to look at other issues, the non-invasive methods we have now are not good enough to categorize the clinical value of HVPG. There is the potential to improve on that. We know that FibroScan is not good to say if a patient has a 16, 18 or 19mmHg gradient. Additionally, we are still working with biomarkers to determine what happens in these patients in response to treatment. For example, we have recently published that in patients treated with cirrhosis and hepatitis C and where the hepatitis C is cured, the correlation between the reduction in HVPG is not well mirrored by the reduction in FibroScan findings. In this situation, FibroScan is probably not good enough. Right now, it is also probably not good enough to demonstrate the response of HVPG to beta-blockers. We cannot assess that.

 

There is work in progress to try and improve on that, but at the moment, for assessing response to treatment, HVPG is still the gold standard and I don’t think there is any good non-invasive method. We need to work on that as long as we need to correlate changes in the pressure gradient.

 

Dr Cardenas: One thing I also want to say is that these tests are also used to determine if a patient has cirrhosis. Sometimes we can have a patient with chronic liver disease where we don’t know if they have cirrhosis or not. FibroScan will help you decide if the patient has cirrhosis. If they have a value of 10kPa, then that tells us there is a possibility of cirrhosis. A value >15kPa says there is a high possibility that there is also significant portal hypertension. The problem is that there is a zone between 10 and 15kPa where we don’t know, and in most cases, we are going to need confirmation. The best confirmation for portal hypertension would be HVPG, with a biopsy to tell you if the patient has cirrhosis or not. In terms of screening for varices, these tests are very good. If you have a low FibroScan reading (<20) and the platelet count is >150000 in somebody with cirrhosis, you can probably avoid endoscopy in three patients because these patients will likely not have varices. It tells you who will not have varices, but not very good at telling you who does have varices.

 

Dr Qi: From your perspective, what percent of portal hypertension patient urgently need invasive HVPG for risk stratification, and of this population, what percent actually receive an invasive procedure. A recent study in Hepatology suggested 30% of patients were indicated to be at high risk for early TIPS, but only 5% receive TIPS, which represents a large gap.

 

Dr Garcia-Pagan: I don’t think there is data on how many patients will improve their management by doing HVPG and where HVPG is not done. I don’t have that data because it is very difficult to know. In my experience at a referral center, there are many patients who present as difficult patients, which means we are biased towards these cases. Sometimes these patients will have been categorized as cirrhotic, and once we do a full evaluation, we conclude that they do not have cirrhosis but have hepato-portal hypertension or even a portal vein thrombosis that went undiagnosed because the incorrect diagnostic study was undertaken. But regarding your question, I don’t think data on who needs it and who gets it do not exist.

 

Dr Cardenas: HVPG is the best tool but there are other non-invasive tests that are reliable mainly in the Child-Pugh class. Someone with a Child-Pugh between 10 and 13 will probably have portal pressures that are very high, and these are the patients we should be most likely to be sending for TIPS (transjugular intrahepatic portosystemic shunt). They don’t necessarily need a HVPG, but if they have acute variceal bleeding and their Child-Pugh score is <14, these patients should undergo a pre-emptive TIPS without the need for a HVPG. It is not something that can be done in every center.

 

Dr Qi: In China, the different etiologies of cirrhosis may play a role in portal hypertension. In China, more than 80% of patients have cirrhosis due to HBV. From the literature, most European patients have HCV or alcohol-related cirrhosis. Are these diagnostic methods still suitable for HBV-related cirrhosis?

 

Dr Garcia-Pagan: The thresholds may be a little different, but globally, it may work. The main problem I think is the way that HVPGs are performed and that they are not conducted properly in a lot of centers. They do not fulfill minimum criteria, which means there is huge variability making the data inaccurate. For the small number of patients where we measure total pressure and wedge pressure at the same time with hepatitis B cirrhosis, the correlation is quite similar to hepatitis C cirrhosis and alcoholic cirrhosis. I think wedged pressure has a good correlation with portal pressure if the patient has parenchymal cirrhosis, which is different from cholestatic disorders, for example. In these patients, where wedge pressure differs, it does not reflect portal pressure. The thresholds for elastography change a little bit, but I don’t think there are that many changes.

 

Dr Cardenas: I agree with that. I think it is a topic we need more information about, but in most of the studies of viral hepatitis, there are not huge differences between hepatitis B and C.

 

Dr Qi: In my experience, based on FibroScan data, there is a difference between HBV and HCV.

 

Dr Garcia-Pagan: They will be changes in thresholds most probably and those thresholds probably need to be corrected. It is the same for NASH where the thresholds are different than hepatitis C and the thresholds for alcohol, because the level of transaminase affects the level of inflammation. But overall, I think the main information can be mirrored from what we currently have for hepatitis C and alcohol.

 

Dr Qi: At this time, most studies have a very limited sample size of around 100 patients, so it is difficult to subgroup analyses for specific etiologies. It is a topic we still need to focus on.

 

Dr Garcia-Pagan: I feel those types of studies will be reproducing existing data and would probably not be accepted in a high impact journal. But I am biased.