A Review of Arterial Grafts Used for Microvascular Arterial Reconstruction

Arterial grafts are sometimes used in microvascular reconstruction and their clinical benefit over standard venous grafts is unknown. To determine arterial graft utilization in clinical microvascular arterial reconstruction, a review of the literature was done. PubMed search resulted with 4,352 finds, and after screening for relevance, 11 articles reporting on 55 arterial grafts were analyzed. All reports were retrospective studies, case reports, and case series, with no randomized controlled trials. Two retrospective series reported better patency of arterial versus venous grafts in upper-limb revascularization for chronic occlusion, but the findings were highly biased. Better patency of arterial grafts did not lead to higher rate of clinical improvement. Antiplatelet and lipid-lowering agents seem to be underused in venous graft recipients and use of no-touch venous grafting has not been reported. Based on the available data, routine use of arterial grafts cannot be recommended. Studies that show better patency of arterial grafts in hand revascularization for chronic vascular insufficiency are retrospective and biased, so a randomized controlled trial is needed.     

Use of small diameter synthetic grafts in carotid bypasses in rats

Vascular prostheses of expanded polytetrafluoroethylene were tested as microvascular synthetic substitutes in a carotid bypass model in rats. Poor patency rates (28%) were obtained with this material when compared with the rates obtained in homologous arterial grafts (100%). It is concluded that expanded polytetrafluoroethylene is unacceptable for use as a 1 mm vascular conduit.     

Carotid bifurcation reconstruction using y-shaped microvascular grafts: an experimental study in rats

Reconstruction of the carotid bifurcation was performed in rats using y-shaped microvascular grafts in an attempt to examine the feasibility of using such grafts in intracranial vascular reconstruction, especially for vessels with bifurcations. Twenty y-shaped microarterial grafts were obtained from the carotid bifurcation of donor rats, while 20 y-shaped microvenous grafts were obtained from the femoral veins of the rats. The patency rate of arterial grafts and venous grafts were 75% and 97%, respectively. Anastomotic aneurysms frequently occurred in both arterial grafts (45%) and venous grafts (60%). Results of histological examinations concluded that these aneurysms were false aneurysms. We discuss the reasons for aneurysm formation and technical issues that might improve the experimental results with regard to patency and aneurysm formation.     

Immunosuppressive treatment of venous allografts

In order to evaluate whether temporary immunosuppressive therapy is able to improve the results obtained with viable venous allografts and achieve better results than with synthetic grafts, 142 arterial reconstructions were performed in mongrel dogs bypassing their ligated femoral arteries. Histological as well as immunological studies were performed and patency determined by weekly palpation and regular angiography. The 6-month cumulative patency rates were: Group I: synthetic grafts (a) Dacron: 48%, (b) plasma-TFE: 53%. Group II: fresh grafts (a) autografts: 100%, (b) allografts: 37%, (c) allografts treated with cyclosporin 4 mg kg-1 daily for 1 month: 74% (100% after 1 month). Group III: grafts preserved in Hanks' solution with 15% DMSO at -160 degrees C for 1 month (a) autografts: 77%, (b) allografts: 35%, (c) allografts treated with methylprednisolone 1 mg kg-1 daily: 38%, (e) allografts treated with cyclosporin and methylprednisolone: 83%. Group IV: human saphenous veins implanted as xenografts and treated with cyclosporin and methylprednisolone: 18%. Immunosuppressive therapy with cyclosporin seems to be able to prevent early thromboses due to rejection seen after implantation of viable fresh or cryopreserved venous allografts, and the results are significantly better than those obtained with synthetic grafts. Tissue matching might further improve these results. This study suggests that cryopreserved venous allografts could be used for the creation of a vein-bank and their use, in combination with tissue typing and temporary immunosuppressive therapy may be warranted for arterial reconstructions when autologous saphenous vein is not available.     

Preserved arterial grafts in microvascular surgery

The patency rates of arterial grafts preserved by immersion in 70% alcohol and arterial grafts preserved by ficin digestion and dialdehyde tanning were compared with the patency rate of fresh autogenous vein grafts in 5-mm defects in the femoral arteries of 50 rats. The overall patency rate for the fresh vein grafts was 90%. The patency rate at 2 weeks for the alcohol preserved arterial grafts was 40%; the patency rate for the dialdehydetanned arterial grafts at 2 weeks was 30%. The difference between the rates for the fresh vein grafts and the alcohol-preserved and dialdehyde-tanned grafts was statistically significant (P < 0.01 and P < 0.001, respectively). At 2 months there was no statistically significant difference in the rates between the autogenous vein grafts and the preserved arterial grafts, probably because of recanalization of grafts that had occluded primarily. From the data, it is concluded that fresh autogenous vein grafts are still superior to preserved arterial grafts in microvascular surgery.     

Modified anastomotic technique for 1 millimeter internal diameter polytetrafluoroethylene arterial grafts in the rat.

Reported patency rates after standard end-to-end anastomoses for microvascular prosthetic grafts have been inconsistent and usually disappointing. A modified anastomotic technique is described in which the prosthetic graft is invaginated inside the arterial lumen. In this study of 6 cm lengths of 1 mm internal diameter polytetrafluoroethylene femoro-femoral bypass grafts in the rat, 6 (40%) of 15 grafts with standard anastomoses were patent at 6 months compared to 28 (90%) of 31 grafts using the modified anastomotic technique (P < 0.001). With invagination of the prosthetic graft inside the arterial lumen, reliable high patency rates can be achieved with microvascular prostheses long enough for potential clinical applications.     

An experimental study of microarterial grafts to arteries--compared with microvenous grafts

In spite of all the successful advances in microvascular surgery, the use of arterial graft is still a subject of controversy. To elucidate the advantages and indications of arterial grafts in microsurgery, venous and arterial grafts were transplanted to the femoral artery of the rat, and the former was investigated histopathologically one, two, three, four, six and eight weeks, and the latter one, two, three and four weeks after transplantation. As the results, intimal thickening was noted in neither of the two grafts. In the venous grafts, endotherial cells were almost separated two weeks after transplantation, and the media showed fibrous replacement. In contrast, the intima remained unchanged and the media maintained its layers of smooth muscle cells in the arterial grafts. The arterial grafts showed little post-transplantation change and maintained their pre-transplantation structure, retaining contracting and dilating function. Based on these results, arterial grafts might be applied clinically in microsurgery.     

Saphenous vein versus PTFE for above-knee femoropopliteal bypass. A review of the literature.

The autogenous saphenous vein is considered the best bypass graft material for arterial bypasses below the inguinal ligament. However, a synthetic graft or prosthesis is considered an acceptable alternative, especially when the distal anastomosis is situated above the knee. Some studies even suggest that patency rates for vein and synthetic grafts are comparable, whereas others indicate that a vein graft is superior to a prosthetic graft, even above the knee. To test the hypothesis that both vein grafts and synthetic prostheses are equally beneficial in the above-knee position, we performed a systematic review of available studies comparing the patency of saphenous vein and polytetrafluoroethylene (PTFE) as bypass material. English and German medical literature from 1966 to 2002 was searched using Medline, and 25 articles meeting our inclusion and exclusion criteria were selected. The patency of venous bypasses was superior to that of PTFE bypasses at all time intervals studied. After 2 years, the primary patency rate of venous bypasses was 81% as compared to 67% for PTFE bypasses, and after 5 years it was 69 and 49%, respectively. After 5 years, the secondary patency of PTFE bypasses reached 60%. When only randomized trials were considered, venous bypasses were again superior to PTFE bypasses at all intervals studied. After 2 years, the primary patency rate of venous and PTFE bypasses was 80 and 69%, respectively, and after 5 years it was 74 and 39%, respectively. Since both randomized and retrospective studies comparing venous with PTFE bypasses showed that vein grafts were 'better' than PTFE prostheses, the null hypothesis that there is no difference between the two types of graft material was rejected (p=0.008). We conclude from this systematic review that if a saphenous vein is available, a venous bypass should be chosen at all times, even if patients have an anticipated short life expectancy (<2 years). If the saphenous vein is absent or not suitable for bypass grafting, PTFE is a good alternative as bypass material.     

Circumferential choledochoplasties with autologous venous and arterial grafts

Circumferential choledochoplasties with vascular grafts have rarely been attempted either experimentally or in clinical practice. In this study, choledochoplasties using autologous venous and arterial grafts were performed in rats. Sixty-four rats were randomly selected into five treatment groups: A) venous interpositional graft replacement of a choledochus gap without a stent; B) venous graft with prolene stent; C) venous graft with polyethylene stent; D) arterial graft; E) a control group with simple resection between ligatures in the choledochus. The operative mortality in treatment groups B, C, D, and E, was 0, and 13% in group A. At 12 weeks follow-up, all the rats in group E had died, whereas, 52.2% (P <.05) of the rats in group A, 30% of the rats in group B, 57% of the rats in group C, and 92.8% of the rats in group D survived treatment. Surviving animals were sacrificed at 3 months for further examination. The morphology and caliber of the common bile duct of these rats were normal in 25% of the rats in group A, 33% of the rats in group B, 25% of the rats in group C, and 84.6% of the rats in group D. Proximal dilations were found in the rats presenting with abnormal morphology. The dilations were less marked in the group treated by arterial choledochoplasties. Laboratory and clinical cholestatic parameters were within normal ranges in the presence of common bile duct dilations less than four times the normal duct caliber. Electron microscopic examination of the venous and arterial graft at 3 months follow-up revealed a fibrous ring composed of collagen fibers, fibroblasts, and remnants of elastic fibers. Regenerated ductal epithelium encompassed both types of grafts. Epithelialization was more pronounced in venous grafts as compared to arterial grafts. Biliary epithelium was able to colonize the venous grafts and resume cell specialization and function as in normal biliary epithelium. The most satisfactory results were obtained using venous grafts with stents or by using arterial grafts. © 1993 Wiley-Liss Inc.     

Vascular graft selection

Twenty to thirty per cent of patients with arterial injuries and some patients with venous injuries require interpositional grafting. The first choice of grafting material for both arterial and venous injuries is autogenous vein. Injuries to large vessels such as the aorta and superior vena cava may necessitate synthetic prostheses. Synthetic aortic prostheses have excellent long-term patency rates, but the same materials are much less likely to remain patent in the vena cava. Panel or spiral grafts constructed from saphenous vein appear to be the best replacement for this vessel. Autogenous veins are present in different diameters ranging from a mean of 6.4 mm in the saphenous vein to a mean of 1.8 cm in the internal jugular vein. The thickest autogenous vein is the saphenous vein, and thus it is preferred for medium-sized and small arteries. The authors prefer the larger 7.5-mm cephalic vein for replacement of medium-sized veins. In the absence of suitable saphenous vein, the cephalic vein is also the choice for arterial interposition grafts. Although there are few reports of the use of arterial autografts in vascular trauma, the surgeon should be aware that autografts may be ideal for vascular injuries in children and for isolated injuries with severe contamination. Finally, the use of synthetic grafts in injuries where adequate tissue coverage is not possible may result in immediate limb salvage, but the incidence of limb loss in this situation will be extremely high.     

Autogenous artery and vein grafts in common carotid-supraclinoid carotid anastomoses

Microsurgical techniques employing either autogenous saphenous arterial grafts or venous grafts were used for common carotid-supraclinoid carotid anastomoses (CC-SCAs) in dogs. The CC-SCA provides a large volume of blood to a major vessel supplying the brain. Within 10 to 14 days postoperatively, angiography and postmortem examination showed arterial patency in 9 (69%) of 13 dogs and venous patency in 2 (25%) of 8 dogs. The discrepancy in patency between the arterial and venous grafts may be the result of the differences in construction of arteries and veins, technical failure, shear stress, turbulence, and/or boundary layer separation. The application of autogenous arterial grafts (i.e., radial, hypogastric) in humans seems justified.     

The effect of twisting on microanastomotic patency of arteries and veins in a rat model.

The authors examined the effect of twisting on the patency of microvascular anastomoses 3 days after surgery. A total of 69 male Wistar rats were divided randomly into four groups. The femoral arteries and veins were dissected for a standard distance. A total of 69 microarteriorrhaphies and 68 microvenorrhaphies were performed at 0 deg and with twist of the vessel ends of 90, 180, and 270 deg. Three-day patency rates for arterial microanastomoses were 100% with a 0-deg twist, 80.9% with a 90-deg twist, 68.4% with a 180-deg twist, and 64.2% with a 270-deg twist. Three-day patency rates for venous microanastomoses were 100% with a 0-deg twist, 85% with a 90-deg twist, 28.5% with a 180-deg twist, and 25% with a 270-deg twist (p = 0.047 for arteries, p = 0.001 for veins). These data are statistically significant. Moreover, assuming the risk of thrombosis to be 1 for microanastomosis without twisting, the odds ratio for the risk of vessel thrombosis for 270-deg twisting (the maximal examined degree of arterial and venous twist in the current study) is 10.08 for arterial anastomosis and 226.85 for venous anastomosis.     

Effects of a venous cuff at the venous anastomosis of polytetrafluoroethylene grafts for hemodialysis vascular access

INTRODUCTION AND METHODS: The most frequent complication of polytetrafluoroethylene (PTFE) arteriovenous grafts for hemodialysis is thrombotic occlusion due to stenosis caused by intimal hyperplasia. This complication is also known for peripheral bypass grafts. Because the use of a venous cuff at the distal anastomosis improves the patency of peripheral bypass grafts, we considered that it might also improve the patency of PTFE arteriovenous grafts. Therefore, a randomized multicenter trial was carried out to study the effect of a venous cuff at the venous anastomosis of PTFE arteriovenous grafts on the development of stenoses and the patency rates. RESULTS: Of the 120 included patients, 59 were randomized for a venous cuff. The incidence of thrombotic occlusion was lower in the cuff group (0.68 per patient-year) than in the no-cuff group (0. 88 per patient-year; P =.0007). However, the primary and secondary patency rates were comparable. The cuff group tended to have fewer stenoses at the venous and arterial anastomoses when examined with duplex scan. Graft failure was higher in patients with an initial anastomosing vein diameter smaller than 4 mm (7 of 18 [39%]) than in those with a vein diameter of 4 mm or larger (16 of 88 [18%]; P =. 052). Local edema, skin atrophy, and obesity yielded a higher risk on graft failure (23% vs 11%). CONCLUSION: A venous cuff at the venous anastomosis of PTFE arteriovenous grafts for hemodialysis reduced the incidence of thrombotic occlusions; stenosis at the venous anastomosis was reduced. However, this did not result in a better patency rate. Therefore, the venous cuff should not be used routinely. Initial vein diameter and local problems (edema, obesity, or skin atrophy) appear to be the most important risk factors for graft failure.     

The utility of the microvascular anastomotic coupler in free tissue transfer

BACKGROUND:

The microvascular anastomosis remains a technically sensitive and critical determinant of success in free tissue transfer. The microvascular anastomotic coupling device is an elegant, friction-fit ring pin device that is becoming more widely used.

OBJECTIVE:

To systematically review the literature to examine the utility of the microvascular coupler in free tissue transfer.

METHODS:

A comprehensive database search was performed to identify eligible publications. Inclusion criteria were anastomotic coupler utilization and free-tissue transfer. Recorded information from eligible studies included patient age, follow-up, radiation history, number of free-flaps and failure rates, reconstruction subsites, number of coupled venous and arterial anastomoses, coupling time, conversion to sutured anastomosis, coupler size and thrombosis rates.

RESULTS:

Twenty-five studies reporting on 3207 patients were included in the analysis. A total of 3576 free-flaps were performed within the following subsites: 1103 head and neck, 2094 breast, 300 limb or body, and 79 nonspecified. There were only 26 reported flap failures (0.7%). A total of 3497 venous and 342 arterial coupled anastomoses were performed. The primary outcome measure was thrombosis rates, and there were 61 venous (1.7%) and 12 arterial (3.6%) thromboses reported. Mean coupling time was 5 min, and 30 anastomoses (0.8%) were converted to suture.

CONCLUSION:

Flap survival and revision-free application of the microvascular coupler occurred in more than 99% of cases. There is a substantial time savings with coupler use. Venous and arterial thrombosis rates are comparable with the best results achieved by sutured anastomosis and, when used by experienced surgeons, the coupler achieves superior results.     

Use of superficial femoral vein in the treatment of infected aortoiliofemoral prosthetic grafts

PURPOSE: To assess the use of the superficial femoral vein for the reconstruction of the aortoiliofemoral sector in the treatment of prosthetic infections. METHODS: From December 1995 to November 1999, 12 patients with infection involving a synthetic vascular prosthesis were submitted to thirteen surgical procedures involving partial or total resection of a synthetic vascular prosthesis and restoration of arterial flow with an in situ femoral vein. RESULTS: The overall operative mortality rate was 15.3% (2 out of 3 patients who underwent total graft replacement and 0 out of 9 patients who underwent unilateral graft replacement). Major amputations related to the surgical procedures were performed in two cases (an 87.5% rate of limb salvage). The series was followed up on average for 22 months (range: 6-65 months). No patient presented clinically significant edema or signs of chronic venous insufficiency in the lower limbs used as donors of autogenous venous grafts. All surviving patients presented complete resolution of the infectious signs and symptoms and none of them presented late thrombosis of the venous graft. CONCLUSION: The use of the superficial femoral vein is a good surgical alternative for the treatment of prosthetic infections with minimal venous morbidity of the lower limbs used as venous graft donors. The autogenous venous grafts present good long-term patency and excellent adaptation to the aortoiliofemoral position.     

Rat femoral vein-to-vein grafts as a microvascular practice model: factors that influence patency

The rat femoral vein has become a standard model for microvascular anastomosis practice as well as for research investigations into various manipulative influences upon venous patency. Although vein grafts to the rat artery are the focus of many experimental studies, few reports have investigated vein-to-vein grafts. This study examines the short-term patency (1-3 days) of vein grafts interpositioned to the rat femoral vein. Several factors are studied for their ability to alter the patency rate. The results indicate that patency is difficult to achieve following the creation of a second serial anastomosis (construction of a vein-to-vein graft). Systemic heparin consistently improves patency rates, and trends toward increased patency are seen for reduction in longitudinal tension of the repaired vessel. It is also implied that a deeper understanding of venous hemodynamics and thrombotic events may lead to improved methodologies in the clinical applications of microvenous repairs. These results indicate that rat vein-to-vein grafts may provide a challenging microvascular training model, while introducing the novice microsurgeon to some of the complicating factors encountered with microvascular grafting.     

Microvenous graft length determination for arterial repair

Varying lengths of femoral vein grafts from 36 New Zealand rabbits were used to bridge defects in the ipsilateral femoral artery in an attempt to estimate the correct length of vein graft needed to replace an arterial defect. Harvested venous segments contracted by 35% to 55% of their in vivo length. Short grafts subjected to linear tension had poor patency rates. Increasing the length improved patency. Long redundant grafts were initially tortuous, but lost their redundancy when studied long term. Silicone perfusion studies, however, revealed a persistence of luminal irregularity and tortuosity. Histologically, arterialized vein grafts thickened by accumulation of fibrous tissue in the medial layer, but otherwise maintained identifiable venous morphological characteristics. Scanning electron microscopic studies identified endothelial changes in vein grafts after arterialization. Based on these results, the ideal microvenous graft should be approximately 35% longer than the arterial defect to be bridged when measured in vivo to permit optimal balance between linear and radial deforming forces.     

Maintaining long-term vessel patency in microvascular surgery using tissue-type plasminogen activator

The primary cause of free flap failure remains vascular thrombosis at the microanastomosis site. Four-hour local infusion of tissue plasminogen activator (t-PA) has been proved to effectively lyse thromboses in microvascular studies of animals; however, rethrombosis occurs once the infusion of t-PA has been terminated. The present study was designed to examine the efficacy of 48 hours of a continuous local t-PA infusion in maintaining long-term venous patency. Our previously described modified arterial inversion graft reanastomosed into the venous system was used in rabbits to form venous thrombi. Three mg of t-PA was infused over 48 hours in eleven rabbits. Seven of eleven grafts were patent at 48 hours and four remained patent at 1 week. In comparing the patency rates in this study with the overall patency rate using the modified arterial inversion graft model (1/22), there are statistically significant differences at both 48 hours (p less than 0.001) and 7 days (p less than 0.05). We conclude that lengthening the infusion time of t-PA may increase the long-term patency rate in this animal model.     

Effect of adenoviral titer and instillation pressure on gene transfer efficiency to arterial and venous grafts ex-vivo

OBJECTIVE: Adenoviral-mediated gene transfer to arterial and venous grafts has potential in the treatment of a number of vascular diseases. Despite widespread use of these vectors to mediate gene transfer to blood vessel walls, the optimal transduction conditions for each type of vessel has yet to be determined. Our objective was to study the effect of adenoviral titer and instillation pressure on efficiency of gene transfer to arterial and venous grafts ex-vivo. METHODS: Jugular vein and carotid artery segments of 8 cm were harvested from Yorkshire Cross pigs. Tissue culture media or different titers of an adenoviral vector encoding human placental alkaline phosphatase (hpAP) were instilled into venous and arterial grafts at 0 mm Hg or 80 to 100 mm Hg of pressure and bathed externally in the same solution at 37 degrees C for 30 minutes. The grafts were rinsed, opened longitudinally, and incubated in culture media at 37 degrees C for 48 hours. Grafts were fixed and stained for hpAP transgene expression to quantitate percent luminal transduction or homogenized for alkaline phosphatase (AP) activity to determine total transmural transduction. RESULTS: For venous grafts, the percent luminal area stained for hpAP was greatest with 10(8) plaque-forming units/mL at 0 mm Hg (81% +/- 7%) and decreased with increasing titers (53% +/- 9% at 10(9) pfu/mL and 44% +/- 11% at 5 x 10(9) pfu/mL; n = 7; P <.05). No increase in percent luminal area stain was achieved with an instillation pressure of 80 to 100 mm Hg at any viral titer. The inverse finding was observed in arterial grafts. For arterial grafts, the greatest percent luminal area stained was achieved with 5 x 10(9) pfu/mL at 80 to 100 mm Hg (76% +/- 7%). An instillation pressure of 80 to 100 mm Hg increased the percent luminal area stained at 10(8) pfu/mL from 31% +/- 9% to 66% +/- 8% (n = 8; P =.01). For venous grafts, total AP activity peaked with 10(9) pfu/mL at 0 mm Hg and decreased with an instillation pressure of 80 to 100 mm Hg (30.6 +/- 9.7 U/mg versus 10.9 +/- 2.5 U/mg; n = 7; P <.01). However, for arterial grafts, total AP activity peaked with 5 x 10(9) pfu/mL (0 mm Hg) and increased with an instillation pressure of 80 to 100 mm Hg (32.8 +/- 9.9 U/mg versus 63.4 +/- 20.5 U/mg; n = 8; P <.05). CONCLUSION: High transduction efficiency can be achieved with adenoviral-mediated gene transfer of arterial and venous grafts. Gene transfer with the vascular graft's physiologic pressure conditions improved transduction efficiency for the artery (80 to 100 mm Hg) and vein (0 mm Hg). Comprehensive analysis of adenoviral transduction conditions is important to realize the full promise of adenoviral-mediated gene transfer.     

Clinical experience with non-penetrating vascular clips in free-flap reconstructions.

To date, the gold standard for performing a microvascular anastomosis has been the penetrating suture with attached needle. During the last two decades, non-penetrating techniques have been introduced, including the Unilink system for end-to-end anastomoses, and the VCS clip-applier system for both end-to-end and end-to-side anastomoses. The aim of this study was to compare the results of different techniques used to create microvascular anastomoses in free-flap reconstructions. Between January 1995 and October 1999, we performed 474 microvascular anastomoses in 216 consecutive free-tissue transfers. The anastomosis techniques included manual sutures (42%), Unilink rings (34%) and VCS clips (24%). Seven combined sutured-clipped anastomoses were excluded from further analysis. The mean anastomotic time when rings were applied was significantly shorter than when using clips (P 0.0001) or sutures (P 0.0001). Venous anastomoses using clips took less time than those using sutures (P 0.05). There were 19 anastomotic failures, all of which lead to early flap failure. Ten flaps were salvaged by early reoperation; nine flaps were lost. Three more flaps were lost as a result of other causes, bringing the flap survival rate down to 94.4%. Early flap failure was caused by failure of the arterial anastomosis in eight cases; all of them were sutured (these represented 5% of all arterial anastomoses with sutures). None of the clipped arterial anastomoses failed. Early flap failure was caused by failure of the venous anastomosis in 11 patients. Three of these anastomoses were sutured (representing 6% of all venous anastomoses with sutures), seven were anastomosed with rings (representing 5% of all venous anastomoses with rings) and one was clipped (representing 2% of all venous anastomoses with clips). Both the VCS clip-applier system and the Unilink system are easy to handle and allow fast microvascular anastomoses without intraluminal penetration. The patency rate of clipped vessels is at least as good as the patency rates of vessels anastomosed using sutures or rings.