Benefits of Early Treatment for Patients with Hepatic Myelopathy Secondary to TIPS: A Retrospective Study in Northern China

Transjugular intrahepatic portosystemic shunt (TIPS) is an effective therapy for reducing portal pressure. Hepatic myelopathy (HM), a rare complication of chronic liver diseases, remains obscure in terms of treatment and prognosis. We aimed to determine an optimal treat strategy for patients with HM after TIPS. Twenty-nine patients who developed HM after TIPS were stratified by time-lapse from onset to treatment: group A (n = 16), <6 months; group B (n = 13), ≥6 months. Therapeutic measures included shunt-limiting and medical treatments. Overall survival, lower-limb muscle strength, Fugl-Meyer score, Barthel index, and serum ammonia were recorded. Median survival time in group A or B was 30 months or 16.5 months, respectively (log rank p = 0.0172). All patients in group A obtained improvement in grading of muscle strength (p < 0.0001), Fugl-Meyer score (p = 0.0021), and Barthel index (p = 0.0003), particularly male patients and those subjected to shunt-limiting. Serum ammonia levels were decreased significantly in both group A (p = 0.0007) and group B (p = 0.0007). Collectively, once HM is confirmed after TIPS, active intervention is imperative and urgent, especially within the first 6 months from onset of symptom. TIPS shunt-limiting is particularly beneficial for rehabilitation in patients with early-onset HM.


Result
Totally 36 patients with HM were identified for this retrospective study. Patients with HM accounted for 1.04% (36/3467) of all patients undergoing TIPS during the study period. Any candidates unable to provide informed consent or complete information were excluded from analysis. In total, 29 patients (22 male and 7 female) were proved eligible for final analysis (Fig. 1). All patients displayed similar symptoms and disease evolution, initially presenting as heaviness of the lower limbs. One patient experienced urinary incontinence. Early electromyography examination was normal in 26 patients. All the patients underwent preoperative computed tomography and ultrasound examinations. Symmetrical hyperintensity was identified along lateral pyramidal (corticospinal) tracts bilaterally on regular/FLAIR MR images as a nonspecific finding. Clinical and demographic information of the study population is summarized in Table 1.
Post-treatment cranial MR imaging in 11 patients from group A exhibited slight reversal of previously identified demyelinated areas within basal ganglia and globus pallidus ( Fig. 2A and B).

Complications.
In the course of this study, we focused on changes in perioperative conditions of patients subjected to stent-limiting procedures. The 17 (100%) shunt-limiting surgeries were successfully performed, without serious complications. In group A, one patient experienced severely low platelets (12 × 10 9 /L; normal reference range: 100-300 × 10 9 /L) caused by stress gastrointestinal bleeding on the day following surgery. The bleeding was controlled by platelet transfusions (2 units). The patient's ADL/physical exercise capacities were improved during follow-up. Another patient developed hematemesis 3 months postoperatively (total volume: ~1000 ml). Emergency endoscopy revealed bleeding in gastric ulcer; and the patient recovered well after endoscopic treatment. Three patients developed refractory ascites symptoms (similar to pre-TIPS), which were controlled by diuretics. Two additional patients underwent LT.
Physical strength and function evaluation. Patients in group A had experienced significant improvement than group B in terms of lower-limb muscle strength, physical activity, and Barthel ADL index. All patients in group A routinely exhibited improved lower-limb muscle strength, physical activity, and Barthel scores,

Discussion
TIPS is a highly efficient means of treating gastrointestinal hemorrhage and refractory ascites due to portal hypertension. Indications for TIPS, including first-line treatment recommendations, have been compiled by de Franchis and Baveno VI 8 . General awareness of post-TIPS complications has grown with worldwide application of TIPS. Complications of shunt stenosis or occlusion have been dramatically reduced by covered stents 9,10 . Symptoms of hepatic encephalopathy may be readily alleviated through ammonia-reducing drugs and restriction on oral protein intake. However, special instances of TIPS-related HM presenting as rare forms of progressive spastic paraparesis have come under clinical scrutiny in recent years [10][11][12] . Although the pathogenesis of HM is not yet understood, extensive portosystemic shunting is clearly implicated. Possibly, circulatory shunting allows byproducts of nitrogenous degradation or a potential neurotoxin to escape hepatic detoxification, causing damage to the spinal cord. Interestingly, Nardone 13 found that motor evoked potentials (MEP) in patients with more advanced disease were not substantially altered by LT, whereas LT conferred notable neurophysiologic and clinical benefit in patients with early stages of HM. Such findings suggest that HM may be reversible if treated early and timely, perhaps avoiding liver transplantation.
In our study, men (22/29, 75.86%) accounted for the preponderance of patients with HM. PSG was rising in both groups (group A: 12/16, 75%; group B: 10/13, 76.92%), especially in male recipients of stent-limiting treatment (group A: 8/9, 88.89%; group B: 6/8, 75%). However, HVPG was comparable in group A and B due to various confounders (e.g., respiratory rate, anesthesia). We presumed that changes in HVPG should parallel to PSG after stent-limiting procedures. Any upper gastrointestinal bleeding and ascites developing after stent-limiting treatment may therefore be attributed to an increase in PSG. All such symptoms experienced by our patients were ultimately remedied without serious or fatal complications. In our department, super-selective embolization is achieved through controllable Coil (Cook Medical) and Interlock (Boston Scientific); according to Wang 14 , Amplatzer vascular plugs (Abbott Laboratories, Chicago, IL, USA) may be used in lieu of coil to occlude large surgically implemented splenorenal shunts.
The muscle strength and activity of the lower limbs of group A were improved compared with those of group B, so the patients of group A accepted the better self-care ability and insisted on outpatient follow-up. However, there were three patients died of lung infection and 1 patient died of heart failure induced by long-term bed in group B. The mortality of group B (4, 30.77%) is greater than group A (2,12.5%). Anymore 2 patients (group B) lost contact in follow up. Therefore, patients in the group A have significantly longer estimated OS than those in the group B. As mentioned above, a majority of patients were male. A significant difference in OS between group A and B was observed, whereas a border-line difference in OS in male patients or male stent-limiting treatment recipients. Because only three female patients (one in group A; two in group B) underwent stent-limiting treatment, statistical analysis was not performed. A larger sample size is guaranteed to explore if stent-limiting treatment can conceivably impact outcomes of female recipients in both groups. The OS of patients receiving medical therapy was comparable by group.
The muscle strength assessment indicated significant improvement in overall patients. There were four male patients in each group (group A: 4/7, 57.14%; group B: 4/5, 80%) who received medical therapy only.  Fundamentally, muscle strength represents function of muscle fibers and cells. Although men generally surpass women in this regard, their recovery from regional impairment can be slowed. Men are typically less compliant than women in the course of medical treatment. Fugl-Meyer and Barthel scores are gauges of lower-limb activity and ADL capacity. Only patients of group A showed obvious improvement in these two closely related indices.
In our study, 17 patients with HM secondary to TIPS underwent shunt-limiting surgery. By reducing internal shunt diameters (and thus portosystemic flow fractions), blood volumes directly entering systemic circulation without hepatic detoxification are lowered, limiting injurious spinal cord exposures. In our previous study 7 ,  patients with HM who undergo TIPS-limiting surgery recovered in lower-limb muscle strength 3-6 months thereafter, with significant improvement in exercise and ADL capacities. At 6-12 months postoperatively, these parameters were partially improved, with no clinical symptom aggravation. However, subjective influences are of relatively minor importance. Lower-limb exercise capacity remains progressive decline as cirrhosis progresses and muscle strength is lost. Abnormal signals in areas of the basal ganglia and globus pallidus were changed as clinical symptoms abated, consistent with descriptions by Chavarria and Wang 15,16 . Conn 17 performed stent blockage in one patient with HM following TIPS, which improved symptoms, emphasizing positive effects of timely shunt closure. Only one of our patients (in group B) suffered from a poor postoperative outcome, despite successful shunt-limiting surgery, succumbing to acute gastrointestinal bleeding and deteriorating liver function. It should be noted that portal vein pressures were rising after shunt-limiting surgery (compared with baseline values), therefore, complete TIPS occlusion was not recommended. Imaging studies of the portal system should also be perfomred after shunt-limiting procedures; and if aberrant varicosities ensue, good care should be taken to prevent new deadly bleeding.
The present study has several limitations. Its retrospective design inherently is prone to selection bias. Furthermore, our data were locally restricted, drawn from patients inhabiting in Northern China. Larger controlled clinical trials are needed to corroborate these results. We believe this pilot study will enhance our knowledge about evaluation and management of patients with HM. To date, this is the largest multicenter study to assess the effects of therapy on limb activity in the context of post-TIPS HM.
In summary, although median survival times in male recipients of HM treatment were independent of symptomology, early stent-limiting surgery significantly improved postoperative outcome. We strongly advocate active therapeutic intervention as soon as feasible once HM is confirmed after TIPS, especially within 6 months of symptoms onset. By adding stent-limiting surgery to medical therapy, lower-limb muscle strength can be improved significantly in these patients; and treatment administered within 6 months rather than later promote better lower-limb activity and ADL capacity by comparison.

Materials and Methods
Clinical data. We  Study design. Patients were diagnosed with myelopathy and corticospinal tract damage after joint consultation (radiologist, physician, and neurologists). Each diagnosis of HM was based on available diagnostic criteria, after excluding other causes of myelopathy 20,21 . we detected symptoms of HM within 1.5-26 months after TIPS; the time from consultation to treatment was ranged from 1.5-38 months. In the previous study 19 , for patients with symptoms of HM for <3 months, clinical efficacy was achieved through conservative medicine; for patients with symptoms of HM for 3-6 months, conservative medical treatment provided clinical efficacy in few of them, whereas others did not respond similarly, requiring combined shunt-limiting surgery for improvement. Patients were thus grouped according to the time of onset of HM symptoms. Because all patients initially experienced symmetric limping gait of both lower limbs, the time-lapse from lower-limb symptom onset to start of clinic treatment was used to stratify patients into group A (<6 months) and group B (≥6 months).
Original TIPS procedures. Shunts between hepatic vein, inferior vena cava, and portal vein were established. After topical disinfection and local anesthesia (1% lidocaine) injection, internal jugular vein was punctured, and an angiographic catheter was inserted to delineate hepatic vein and inferior vena cava. Hepatic vein or hepatic segment of inferior vena cava was then accessed (RUPS-100 apparatus; Cook Medical Inc, Bloomington, IN, USA); subsequently an intrahepatic portal branch was punctured at an appropriate position and angle for setting placement within portal vein. Based on portography and portal venous pressure (measured via catheter),