Contemporary outcomes for arterial reconstruction with non-saphenous vein cryo-preserved conduits.

OBJECTIVE: Cryopreserved (CP) products are utilized during challenging cases when autogenous or prosthetic conduit use is not feasible. Despite decades of experience with cadaveric greater saphenous vein (GSV), there is limited available data regarding the outcomes and patency of other CP products, specifically arterial and deep venous grafts. This study was designed to evaluate outcomes of non-GSV CP conduits in patients undergoing urgent, emergent, and elective arterial reconstruction at our institution. We hypothesized that non-GSV CP allografts have adequate patency and outcomes and are therefore a feasible alternative to GSV in settings where autologous graft is unavailable or prosthetic grafts are contraindicated. METHODS: This study was approved by the Institutional Review Board at our institution. We retrospectively reviewed charts of patients undergoing arterial reconstructions using CP conduits from 2010 to 2022. Data collected included demographics, comorbidities, smoking status, indications for surgery, indication for CP conduit use, anatomic reconstruction, urgency of procedure, and blood loss. Time-to-event outcomes included primary and secondary graft patency rates, follow-up amputations, and mortality; other complications included follow-up infection/reinfection and 30-day complications, including return to the operating room and perioperative mortality. Time-to-event analyses were evaluated using product-limit survival estimates. RESULTS: Of 96 identified patients receiving CP conduits, 56 patients received non-GSV conduits for 66 arterial reconstructions. The most common type of non-GSV CP product used was femoral artery (31 patients), followed by aorto-iliac artery (22 patients), and femoral vein (19 patients), with some patients receiving more than one reconstruction or CP product. Patients were mostly male (75%), with a mean age of 63.1 years and a mean body mass index of 26.7 kg/m2. Indications for CP conduit use included infection in 53 patients, hostile environment in 36 patients, contaminated field in 30 patients, tissue coverage concerns in 30 patients, inadequate conduit in nine patients, and patient preference in one patient. Notably, multiple patients had more than one indication. Most surgeries (95%) were performed in urgent or emergent settings. Supra-inguinal reconstructions were most common (53%), followed by extra-anatomic bypasses (47%). Thirty-day mortality occurred in 10 patients (19%). Fifteen patients (27%) required return to the operating room for indications related to the vascular reconstructions, with 10 (18%) cases being unplanned and five (9%) cases planned/staged. Overall survival at 6, 12, and 24 months was 80%, 68%, and 59%, respectively. Primary patency at 6, 12, and 24 months was 86%, 70%, and 62%, respectively. Amputation freedom at 6 months, 12 months, and 24 months was 98%, 95%, and 86%, respectively for non-traumatic indications. CONCLUSIONS: Non-GSV CP products may be used in complex arterial reconstructions when autogenous or prosthetic options are not feasible or available.

Outcomes of upper extremity interventions for chronic critical ischemia.

BACKGROUND: Critical hand ischemia owing to below-the-elbow atherosclerotic occlusive disease is relatively uncommon. The aim of this study was to examine the outcomes in patients presenting with critical ischemia owing to below-the-elbow arterial atherosclerotic disease who underwent nonoperative and operative management. METHODS: A database of patients undergoing operative and nonoperative management for symptomatic below-the-elbow atherosclerotic disease between 2006 and 2016 was retrospectively queried. Patients with critical ischemia (tissue loss and rest pain) were identified. Three management groups were identified: no revascularization (None), endovascular revascularization (Endo), and open revascularization by bypass (Bypass). Patients with acute embolism, active vasculitis, end-stage renal disease, ipsilateral dialysis access complications of steal, and ipsilateral trauma were excluded. RESULTS: One hundred eight patients (56% male; average age, 59 years) presented with symptomatic below-the-elbow disease: 93% presented with digital ulceration and the remainder with rest pain. Eighty-one percent had diabetes and 41% had chronic renal insufficiency (not on dialysis). All underwent catheter-based angiography. Fifty-three patients (49%) had no intervention and subsequently were committed to wound care; 26 of these required no further intervention, 10 had an interval palmar sympathectomy, and 17 underwent either a phalanx or digital amputation. Thirty-four patients (31%) underwent an endovascular intervention with a median of 1.5 vessels (ulnar, radial, or interosseous arteries) intervened on. Technical success was achieved in 29 patients (85%). Of the five technical failures, two went on to bypass, one had a focal endarterectomy and patch angioplasty, and one was treated conservatively. Ten patients in the Endo group required either a phalanx or digital amputation. Twenty-one patients (19%) underwent a saphenous vein bypass (reversed or nonreserved) to the radial in 12 and the ulnar in 11 limbs. In follow-up, 11 patients underwent open or endovascular intervention to maintain patency of the bypass. There were nine phalanx or digital amputations in the Bypass group. No below-the-elbow or above-the-elbow amputations were performed within 30 days. The wound healing rate without amputation was 78% (85 of 108). The predictors of wound healing were technical success of the revascularization, intact palmar arch and presence of digital run-off. The presence of an incomplete arch and poor digital run-off were associated with a phalanx or digital amputation. CONCLUSIONS: Upper extremity interventions for critical ischemia are associated with a high rate of success. Major amputations are rare and the many can be treated nonoperatively. In appropriately selected patients, both endovascular and open interventions have a high rate of success.

Target lesion characteristics in failing vein grafts predict the success of endovascular and open revision.

OBJECTIVES: This study examined the association of anatomic and temporal characteristics of graft-threatening lesions with the efficacy of percutaneous and open graft revision for failing infrainguinal vein grafts. METHODS: Consecutive open and endovascular revisions for graft threatening lesions were reviewed. We evaluated graft durability and individual target lesion response to open and endovascular treatment to determine characteristics that may influence outcomes. Treatment failure was defined as target lesion restenosis or graft occlusion. RESULTS: Eighty-four (58 endovascular, 26 open) infrainguinal vein graft revisions were performed in 67 failing, nonthrombosed infrainguinal vein grafts. Primary assisted graft patency at 5 years was 63% (95% confidence interval [CI], 46% to 77%). Follow-up was 29.5 +/- 19.2 months. Grafts treated for early lesions (<6 months) failed (occlusion or need for additional interventions) more frequently than those with late occurring lesions (P = .03). Overall target lesion revascularization patency was 45% (95% CI, 32% to 58%) at 3 years. Average time to target lesion revascularization failure was 7.5 months, with no significant difference noted between endovascular and open treatment groups. Overall target lesion revascularization patency at 3 years was also not significantly different between open and endovascular groups at 54% (95% CI, 30% to 73%) vs 41% (95% CI, 25% to 56%; P = .15). When divided by early and late-occurring target lesions, endovascular treatment of early lesions was associated with inferior patency compared with open procedures; no difference in patency was seen between treatment groups for late lesions. When divided by target lesion location (anastomotic vs mid-graft), treatment for both proximal and distal anastomotic target lesion was associated with inferior patency compared with mid-graft revision at 32% (95% CI, 17% to 47%) vs 62% (95% CI, 37% to 87%) at 3 years (P = .03). In addition, although results of anastomotic target lesion treatment significantly favored open repair, even open repair of anastomotic target lesions was associated with a <50% patency rate at 3 years. In contrast, mid-graft target lesions treated with open revisions were uniformly successful compared with a 54% patency at 3 years with endovascular treatment (P = .24). Short lesions (<2 cm) fared equally well with either endovascular or open treatment. Univariate analysis noted only anastomotic treatment was associated with significantly increased odds of failure. CONCLUSION: Grafts that develop early lesions fare poorly regardless of treatment modality. Lesions involving anastomoses of failing grafts are better treated with open revision, but patency after treatment of such lesions is still worse than treatment of mid-graft lesions. In contrast, the method of treatment does not influence outcome after treatment of mid-graft target lesions. Thus, endovascular therapy should be reserved for focal, late-appearing lesions involving the mid-graft.

Type 2 endoleaks after abdominal aortic aneurysm stent grafting with systematic mesenteric and lumbar coil embolization.

We evaluated the results of our policy of systematic coil embolization of the inferior mesenteric artery (IMA) and/or lumbar arteries (LAs) prior to endovascular abdominal aortic aneurysm (AAA) repair (EVAR). We retrospectively reviewed all patients undergoing EVAR over a 4-year period at one hospital. Results were analyzed using uni- and multivariate analyses. Fifty-five male patients with an average age of 71 years were evaluated. Follow-up averaged 15 +/- 13 months. The IMA was either coiled or occluded in 30 cases. One or more LAs were coiled in 29 patients. An average of 1.3 LAs per patients were coiled (range 0-6). There were no immediate or late complications from coiling. At last follow-up, 14 AAAs showed no change in diameter, one increased by 2 mm, and the remainder (n = 40) decreased by 7.5 +/- 6 mm in maximal diameter. Only five (9%) type 2 endoleaks were detected during follow-up. Three were associated with AAA size increase. Four of the five were treated with additional coiling, with good results. By logistic regression, neither endoleak occurrence nor AAA shrinkage correlated with LA or IMA coiling. However, by multivariate analysis, completeness of lumbar coiling correlated negatively with aneurysm shrinkage (p = 0.04) and IMA coiling correlated positively with aneurysm shrinkage (p = 0.04). Coil embolization of the IMA and/or LAs prior to EVAR can be safely accomplished in a large number of cases and is associated with a low incidence of type 2 endoleaks. We cannot at present demonstrate a benefit to LA embolization in terms of endoleak prevention or AAA shrinkage. However, IMA embolization may be of benefit in terms of AAA shrinkage.

Are type II endoleaks after endovascular aneurysm repair endograft dependent?

PURPOSE: Most clinical end points after endovascular aneurysm repair (EVAR) are endograft-specific, but type II endoleaks have been assumed to be an unavoidable consequence of the repair method and independent of the type of endograft used. Some recent data have suggested that the rate of type II endoleaks may also be graft-dependent. We reviewed a large clinical experience with six endografts to determine the behavior of type II endoleaks and whether they are graft-specific. METHODS: All elective EVAR cases from five university institutions from 1996 to 2003 were retrospectively analyzed. Endografts used in <50 patients were excluded. Endoleaks were diagnosed and classified from contrast-enhanced computed tomography (CT) scans by the treating surgeons. Results of angiography and interventions for endoleaks were tracked. The rate of type II endoleaks was compared among endografts at 1, 6, and 12 months, and yearly thereafter. Statistical significance was defined as P < .05. RESULTS: During the study period, 1909 patients underwent elective EVAR and had an adequate imaging follow-up at one of the specified time points. At 1 month, the overall rate of type II endoleak was 14.0% (range, 9.8% to 25.2%.) The Excluder had a significantly higher incidence of type II endoleaks at 1 month but was similar to most other grafts during longer follow-up. At 6 months, the overall rate of type II endoleak was 16.3% (range, 8.3% to 16.8%). The Talent and Lifepath had an apparent lower initial rate of type II leaks, but this was only significant for the Talent at 6 months compared with Excluder, Zenith, and Ancure, and at 1 year compared with Excluder and Zenith. No graft had a long-term statistically significant difference in the rate of type II endoleak formation. Intervention rates varied by institution and graft type but in general were quite low. Of 25 successful interventions (Ancure, 12; AneuRx, 8; Excluder, 2; Lifepath, 2; Zenith, 1; Talent, 0), 21 were performed during the first year. Interim spontaneous resolution, defined as a negative CT scan after a CT positive for endoleak, was high, especially in the first year. Resolution of type II endoleaks occurred in 54 (33%) of 164 between 1 and 6 months, in 37 (33%) of 112 between 6 and 12 months, in 20 (35%) of 57 from 12 to 24 months, and in 5 (20%) of 25 between 24 and 36 months. The various grafts had a nearly identical pattern, but the rates were highest for the Talent. Late appearing endoleaks, defined as a positive CT after a negative CT, were frequent. At 6 months, 44 (30%) of 147 type II endoleaks were newly diagnosed. The rates were 37 (35%) of 107 at 12 months, 15 (27%) of 56 at 2 years and 5 (25%) of 20 at 3 years. No conversions to open repair for type II endoleaks were noted in the first 4 years. The thrombus burden could not be determined in this analysis. CONCLUSIONS: Type II endoleaks occur in nearly 15% of patients treated by EVAR. The early incidence varies only slightly with graft type. The long-term prevalence and clinical significance are masked by different treatment patterns, spontaneous resolution, newly evident endoleaks, and aneurysm size at initial treatment.

Can computed tomography scan findings predict "impending'' aneurysm rupture?

Several findings on computed tomography (CT) scans of intact aneurysms have been taken to suggest "imminent'' or "impending'' aneurysm rupture. Often these are identified incidentally in asymptomatic patients when an urgent operation was not planned and may even be ill advised. The authors evaluated whether these signs can truly predict short-term aneurysm rupture. A computerized medical archival system was reviewed from August 1994 to August 2004. Patients with aortic aneurysms and official CT scan reports of "impending rupture'' were reviewed. CT films and reports were reviewed for aneurysm characteristics, while computerized medical records were reviewed for patient demographics, comorbidities, symptoms, documented subsequent rupture, and operative findings. Signs of "impending rupture'' included the crescent sign, discontinuous circumaortic calcification, aortic bulges or blebs, aortic draping, and aortic wall irregularity. Rupture occurring within 2 weeks of the index CT was defined as supporting the "imminent'' label. Forty-five patients with aortic aneurysms and CT stigmata of "impending rupture'' were identified. Five patients with additional signs of suspicious leak and 1 with an infected previously repaired aneurysm were excluded. Of 39 intact aneurysms, 26 (67%) were infrarenal, 2 (5%) were suprarenal, and the remaining 11 (28%) were thoracoabdominal. The patient group had more women than expected (19/39, 49%) and larger aneurysms (mean diameter, 6.8 +/- 1.4 cm). Mean age was 74 years. Ten patients underwent elective repair within the first 2 weeks after the index CT scan (mean, 4 days), precluding adequate observation for early rupture. None had intraoperative signs of rupture. Early rupture: 2 of the 29 remaining patients ruptured within 72 hours of the CT scan, for a positive predictive value of 6.9%. One additional patient ruptured 7 months later after declining an early intervention. No Rupture: 26 patients were observed an average of 246 days (range, 14 days to 3 years) without evidence of rupture. Fourteen were repaired electively 2 weeks to 3 years after the index CT scan, and 12 never underwent repair, mostly because of severe associated comorbidities, and were observed a mean of 394 days without rupture. Although they should be taken seriously, CT signs of "impending rupture'' alone are poor predictors of short-term aortic aneurysm rupture, and alternative terminology is needed until better predictors can be identified.

Is abdominal aortic aneurysm repair appropriate in oxygen-dependent chronic obstructive pulmonary disease patients?

BACKGROUND: The life expectancy of patients with oxygen-dependent chronic obstructive pulmonary disease (COPD) is significantly reduced, but the risk of any intervention is considered prohibitive. However, severe COPD may increase the risk of abdominal aortic aneurysm (AAA) rupture. We reviewed our experience with AAA repair in oxygen-dependent patients to determine whether operative risk and expected long-term survival justify surgical intervention. METHODS: A retrospective review of 44 consecutive patients with oxygen-dependent COPD undergoing AAA repair over an 8-year period was performed. Information was recorded for survival, length of follow-up, patient age, medical comorbidities, pulmonary function tests, and operative approach. Survival data were analyzed by Kaplan-Meier curves and compared with published cohorts of oxygen-dependent patients and the natural history of untreated aneurysms. RESULTS: Twenty-four patients underwent endovascular aneurysm repair (EVAR), and 20 underwent open procedures (14 retroperitoneal and 6 transabdominal). The mean AAA diameter was 6.1 cm (range, 5-9.5 cm). The mean age was 71.4 years, and 82% of patients were male. Operative mortality was 0%. The mean length of stay was 11.2 days for open procedures and 4.3 days for EVAR (significantly longer than that for standard-risk patients). The mean survival time was 37.9 months (range, 2-91 months). Preoperative medical comorbidities, type of repair, and pulmonary function tests were not predictive of survival. Postoperative morbidity was significantly higher with open repair. Long term survival was comparable to historical series of the natural history of O2 dependent patients without AAA but better than untreated 6 cm AAA cohorts. At 42 months, almost 50% of patients in our study group were still alive, compared to 20% survival at 34 months for those with untreated 6 cm AAAs. CONCLUSIONS: It is reasonable to continue to offer AAA repair to home oxygen-dependent COPD patients who are ambulatory and medically optimized and who are without untreated coronary artery disease. Although EVAR may be the most suitable treatment for oxygen-dependent COPD patients, our results show that even open repair may be safely performed in this population, with acceptable results.

Use of endoluminal aortic stent-grafts for the repair of abdominal aortic aneurysms.

Abdominal aortic aneurysms affect approximately 1.5% of the United States population. Randomized trials recommend repair when the maximal aneurysm diameter is 5.5 cm or greater. Since the first report of this technique in 1991, endovascular repair has become the preferred method for elective therapy of AAA disease. This has been a direct result of reported decreased hospital length of stay, reduced patient recovery time and improved survival. The application of endovascular aneurysm repair (EVAR) requires a thorough understanding of aneurysm anatomy, which is critical for appropriate patient selection. In particular the surgeon must be familiar with device-related criteria for proximal fixation and distal fixation as well as access vessels size, tortuosity, and calcification which can often be the limiting factor in the application of EVAR. Although the reported results of EVAR indicate significant advantages when compared with conventional open repair, it is critical to have an understanding of the particular complications associated with EVAR. The development of endoleaks, reports of stent migration and stent fracture as well as the development of limb stenosis and/or occlusion have been reported in up to 20% of patients treated with EVAR and thus necessitate appropriate long-term surveillance protocols.

Effectiveness of coiling in the treatment of endoleaks after endovascular repair.

BACKGROUND: Persistent endoleaks are a common problem following endovascular repair of abdominal aortic aneurysms, and the best method of treatment has been an issue of debate. Some experimental evidence has suggested that coiling may not be an effective method because it allows transmission of pressure across the coils with continued expansion of the sac. We reviewed our experience with endoleak coiling to assess the degree of clinical success of this treatment. METHODS: A retrospective review of patients with type I or type II endoleaks treated solely by coiling over a 7-year period (1997-2003) was performed. All endoleaks had been observed for at least 6 months prior to intervention to detect spontaneous resolution. All coils were delivered by selective catheterization of the endoleaks. For Type II endoleaks, the branches were all coiled at their junction with the sac when feasible and the endoleak cavity was packed. Clinical success was defined as cessation of endoleak on follow-up computed tomography as well as no further aneurysmal growth (> or =5mm minor axis). RESULTS: Twenty-eight patients had their endoleaks treated only with coils. There were 22 Ancure, 2 Excluder, 2 AneuRx, and 2 Lifepath endografts in this patient cohort. Procedural morbidity was 0%. Mean follow-up after coiling for all patients was 18 months (range, 1-60 months) while mean follow-up for patients with a type I endoleak was 24 months. Clinical success was achieved in 15 of 19 (79%) patients with type II endoleaks and 8 of 9 (89%) patients with type I. Three patients, all with type I endoleak, required more than 1 episode of coiling, while 2 others, both with type II lumbar endoleaks, required repeat angiography due to inability to access the leak during the first attempt. There were 2 proximal and 6 distal type I endoleaks (2 aortic, 6 iliac) successfully treated while the type II successes included 8 inferior mesenteric artery and 7 sole lumbar endoleaks. Five patients continued to show evidence of endoleak over time: 2 endoleaks were associated with aneurysm growth leading to conversion in 1 patient, 2 patients with type II endoleaks are stable, and the sole type I endoleak with continued perigraft flow has shown significant shrinkage of the sac and continues to be observed 18 months later. No ruptures were noted during follow-up. CONCLUSION: Coiling as the sole method of endoleak management may be a suitable treatment option in selected patients. Clinical success can be expected in over 80% of patients with type II and select type I endoleaks, with minimal morbidity.

Same-evening discharge after carotid endarterectomy: our initial experience.

PURPOSE: The purpose of this study was to review the initial implementation of a same-evening discharge algorithm for patients undergoing carotid endarterectomy (CEA). METHOD: We conducted a retrospective review of a prospective database of patients undergoing CEA over 3 years. RESULTS: From January 2000 to December 2002, 207 patients underwent CEA, of which 186 qualified for same-evening discharge. Fifty-nine patients (32%) who qualified were discharged to home the same evening; none had an adverse event after discharge. The most common reason for patients not to be discharged the same evening was exiting the operating room too late (n = 63, 34%). Thirteen patients chose to stay overnight, and 11 patients did not go home secondary to physician choice. None of these patients experienced any adverse sequelae during the overnight stay. CONCLUSION: Same-evening discharge after CEA is safe and feasible in selected patients. Currently, nearly one third of our patients are discharged within 8 hours of CEA. With appropriate scheduling, patient education, and increasing physician awareness, most patients can be discharged to home the same evening after CEA.

Spectrum of cystic neoplasms of the pancreas and their surgical management.

BACKGROUND: Owing to increased awareness and the widespread use of abdominal ultrasonography and computed tomography, an increasing number of cystic neoplasms are being identified. Cystic neoplasms of the pancreas are divided into the following 4 main groups: serous cystic neoplasms, mucinous cystic neoplasms, solid pseudopapillary neoplasms, and intraductal pancreatic mucinous neoplasms. OBJECTIVE: To review our experience with cystic neoplasms of the pancreas at our institution from January 1992 through September 2002. METHODS: Medical records were reviewed for age, sex, clinical signs and symptoms, diagnosis, surgical treatment, morbidity, mortality, and histologic features. RESULTS: Seventy-three patients (49 women and 24 men) underwent surgical resection of a cystic neoplasm of the pancreas from January 1992 through September 2002. The most common presenting symptom was abdominal pain. Other symptoms included nausea, emesis, weight loss, jaundice, and pancreatitis. Most patients (73%) had no complications. The most common complication (10%) was pancreatic fistula. There were 3 deaths. CONCLUSIONS: Cystic neoplasms of the pancreas are an increasing entity. Long-term survival of patients with these tumors is generally better than that of patients with adenocarcinoma of the pancreas and mandates aggressive resectional therapy in most patients. Resection of these tumors can be done with resultant low morbidity and mortality rates.

The effect of intraoperative duplex on the management of postoperative stroke.

BACKGROUND: Stroke after carotid endarterectomy (CEA) may be a result of intraoperative ischemia, embolism, or thrombosis at the operative site. Intraoperative duplex should eliminate the occurrence of a severe internal carotid artery (ICA) thrombosis and, thus, negate the benefit of reoperation. This article will detail the results of our evolving treatment algorithm for immediate versus delayed post-CEA neurologic deficit (ND). METHODS: We studied patients who had an ND after CEA from 1988 to 2000. Results. Thirty-two patients (3.2%) had a post-CEA ND (26 related stroke or transient ischemic attack, 6 other); 31 had a satisfactory intraoperative duplex post-CEA, 1 was not tested. Fifteen patients awoke from operation with a related deficit, 5 of whom were re-explored and all had a patent ICA. One patient without lateralizing signs who was not re-explored had extensive thrombosis at postmortem. The remaining 9 all had a duplex-proven patent ICA. Ten patients had a lucid interval before their related ND. Six patients were re-explored and all had thrombosed ICAs; 5 of the 6 improved postthrombectomy. Four patients were not re-explored for various reasons; a carotid thrombosis was not later diagnosed in any of these patients. CONCLUSIONS: Intraoperative and postoperative duplex has modified our treatment of post-CEA stroke. No longer are all patients re-explored. Patients with a normal intraoperative duplex who awaken with an immediate stroke do not usually have occlusive thrombus and routine re-exploration does not benefit these patients. Patients who have an ND develop after a lucid period may have a thrombosed ICA despite a normal intraoperative duplex, and unless there is a timely normal duplex, re-exploration is recommended and appears to benefit these patients.

Distal pancreatectomy: does the method of closure influence fistula formation?

The appropriate closure of the pancreatic remnant after distal pancreatectomy is still debated. Suture techniques, stapled closure, and pancreaticoenteric anastomosis all have their supporters. In this study we have reviewed our data from distal pancreatectomy to determine whether the type of remnant closure or underlying pathologic process had any relation to postoperative fistula formation. We performed a retrospective chart review of patients undergoing distal pancreatectomy at our institution between 1993 and 2001. The charts were reviewed for morbidity and mortality. These were then related to the type of closure of the pancreatic stump. From 1993 to 2001 a total of 86 patients underwent distal pancreatectomy. Data were available on 85 patients. Indications for surgery were pancreatic tumor (69%), pancreatitis (14%), trauma (7%), and extra pancreatic disease (9%). Pancreatic fistula occurred in 14 per cent (N = 12), intra-abdominal abscess in 8 per cent (N = 7), and wound infection in 2 per cent (N = 2). There was no mortality in the series. The incidence of pancreatic fistula formation was not related to method of closure of the pancreatic remnant nor to the underlying pathologic process. Postoperative pancreatic fistulas will close spontaneously even without total parenteral nutrition.