Abstract Background Little is known about the strains at the glenoid near the bone-implant interface in reverse shoulder arthroplasty. The purpose of the current study was to evaluate the strains on the glenoid bone under a compressive load after implantation of three different sizes of metal-backed baseplates in either inferior or superior position in combination with three different sizes of glenospheres and three different glenosphere designs. Methods Three sizes of baseplates (small, medium, large) were implanted in thirty-six paired human cadaveric scapulae either inferior, flush with the glenoid neck, or with a 5 mm superior offset. Glenospheres were available in three sizes (36 mm, 39 mm, 42 mm) and designs (standard, 4 mm lateralized, 2.5 mm inferiorized). Specimens were mounted in a servo-hydraulic testing apparatus at a 60° angle between the glenoid and actuator holding the humeral component. Four strain-gauge rosettes were placed around the glenoid rim to measure strains transferred to the scapular bone under a compressive load (750 N) relative to the various baseplate-glenosphere combinations. Following repeated compression, a load-to-failure test was performed. Results Mean overall registered strains were 161με (range: − 1165 to 2347) at the inferior sensor, −2με (range: − 213 to 90) at the superior sensor, −95με (range: − 381 to 254) at the anterior sensor, and 13με (range: − 298 to 128) at the posterior sensor. Measured bone strains did not show any significant differences across tested baseplate and glenosphere design, size, or positioning combinations (p > 0.05 for all sensors). Furthermore, linear regression analysis did not identify any of the evaluated parameters as an independent influential factor for strains (p > 0.05 for all sensors). Mean load-at-failure was significantly higher in the group of inferior (3347.0 N ± 704.4 N) compared to superior (2763.8 N ± 927.8 N) positioned baseplates (p = 0.046). Conclusion Different baseplate positions, baseplate sizes, glenosphere sizes, and glenosphere design or various combinations of these parameters did not significantly influence the measured bone strains at the glenoid near the bone-implant interface in a contemporary reverse shoulder arthroplasty system. Level of evidence Basic Science Study, Biomechanical Study.

Background: Previous biomechanical studies regarding deltoid function during glenohumeral abduction have primarily used static testing protocols. Hypotheses: (1) Deltoid forces required for scapular plane abduction increase as simulated rotator cuff tears become larger, and (2) maximal abduction decreases despite increased deltoid forces. Study Design: Controlled laboratory study. Methods: Twelve fresh-frozen cadaveric shoulders with a mean age of 67 years (range, 64-74 years) were used. The supraspinatus and anterior, middle, and posterior deltoid tendons were attached to individual shoulder simulator actuators. Deltoid forces and maximum abduction were recorded for the following tear patterns: intact, isolated subscapularis (SSC), isolated supraspinatus (SSP), anterosuperior (SSP + SSC), posterosuperior (infraspinatus [ISP] + SSP), and massive (SSC + SSP + ISP). Optical triads tracked 3-dimensional motion during dynamic testing. Fluoroscopy and computed tomography were used to measure critical shoulder angle, acromial index, and superior humeral head migration with massive tears. Mean values for maximum glenohumeral abduction and deltoid forces were determined. Linear mixed-effects regression examined changes in motion and forces over time. Pearson product-moment correlation coefficients (r) among deltoid forces, critical shoulder angles, and acromial indices were calculated. Results: Shoulders with an intact cuff required 193.8 N (95% CI, 125.5 to 262.1) total deltoid force to achieve 79.8° (95% CI, 66.4° to 93.2°) of maximum glenohumeral abduction. Compared with native shoulders, abduction decreased after simulated SSP (–27.2%; 95% CI, –43.3% to –11.1%, P = .04), anterosuperior (–51.5%; 95% CI, –70.2% to –32.8%, P < .01), and massive (–48.4%; 95% CI, –65.2% to –31.5%, P < .01) cuff tears. Increased total deltoid forces were required for simulated anterosuperior (+108.1%; 95% CI, 68.7% to 147.5%, P < .01) and massive (+57.2%; 95% CI, 19.6% to 94.7%, P = .05) cuff tears. Anterior deltoid forces were significantly greater in anterosuperior (P < .01) and massive (P = .03) tears. Middle deltoid forces were greater with anterosuperior tears (P = .03). Posterior deltoid forces were greater with anterosuperior (P = .02) and posterosuperior (P = .04) tears. Anterior deltoid force was negatively correlated (r = −0.89, P = .01) with critical shoulder angle (34.3°; 95% CI, 32.0° to 36.6°). Deltoid forces had no statistical correlation with acromial index (0.55; 95% CI, 0.48 to 0.61). Superior migration was 8.3 mm (95% CI, 5.5 to 11.1 mm) during testing of massive rotator cuff tears. Conclusion: Shoulders with rotator cuff tears require considerable compensatory deltoid function to prevent abduction motion loss. Anterosuperior tears resulted in the largest motion loss despite the greatest increase in deltoid force. Clinical Relevance: Rotator cuff tears place more strain on the deltoid to prevent abduction motion loss. Fatigue or injury to the deltoid may result in a precipitous decline in abduction, regardless of tear size.

Current techniques for anatomic repair of the dislocated acromioclavicular (AC) joint aim on reconstruction of the AC ligaments and utilize tunnels drilled through the acromion. This improves the stability of the reconstruction but might also increase the risk of fractures at the acromion. The purpose of this study was to evaluate the fracture risk for the acromion after transacromial tunnel placement for anatomic AC joint stabilization procedure. It was hypothesized that the risk of fracture of the acromion is correlated to size and orientation of bone tunnels commonly used for anatomic AC joint reconstruction.

PURPOSE:To evaluate the number of connective tissue progenitor cells (CTPs) and nucleated cells obtained during bone marrow aspiration (BMA) from the proximal humerus using either a fenestrated or a nonfenestrated trocar and determine differences in varying amounts of aspiration volume. The first hypothesis was that the number of CTPs extracted with the fenestrated trocar would be greater due to its potential to extract more cells through its fenestrations. The second hypothesis was that using consecutive aspirations with either trocar would provide a consistent number of CTPs and nucleated cells throughout the aspiration with no significant decrease of cells at the end. METHODS:Patients were eligible for inclusion if they underwent primary or revision arthroscopic rotator cuff surgery, were between 18 and 75 years of age, and signed the informed consent. Between January 2011 and September 2013, 24 patients underwent BMA from the proximal humerus during arthroscopic surgery. They were grouped according to which of 3 different trocars were used for aspiration: (1) nonfenestrated, (2) fenestrated trocar A, and (3) fenestrated trocar H. Four consecutive 12 mL double syringes were used for each aspiration: 1 (0-12 mL), 2 (12-24 mL), 3 (24-36 mL), and 4 (36-48 mL). One milliliter was removed from each syringe (nonconcentrated BMA). The remainder of the BMA was then spun using a centrifuge. BMA and concentrated BMA were brought to the laboratory, counted for nucleated cells (million cells/mL BMA) and cultured for 7 days to obtain colony-forming units (CTPs/million cells). RESULTS:No significant differences were observed in tubes 1 to 4 in the number of nucleated cells in the nonconcentrated and concentrated BMA using the nonfenestrated trocar compared with the fenestrated trocars A and H (all P > .05), except for concentrated BMA tube 3 (P = .014) and tube 4 (P = .003). Nonconcentrated and concentrated BMA from tubes 1 to 4 had a significantly higher CTP prevalence using the nonfenestrated trocar compared with the fenestrated trocars A and H (all P < .05). Most of the times the first tube of each aspiration showed a significantly greater amount of cells and a greater CTP prevalence compared with tubes 2, 3, and 4. CONCLUSIONS:Aspiration from the proximal humerus with the nonfenestrated trocar during BMA was associated with higher prevalence of CTPs, suggesting that more CTPs can be obtained using a nonfenestrated trocar. Furthermore, CTPs can be obtained through all consecutive aspirations with a greater amount in the first tubes. LEVEL OF EVIDENCE:Level II, prospective comparative study.

Background: The purpose of this study was to examine biomechanical properties and the degree of radiolucency of two cemented basic glenoid designs for total shoulder arthroplasty. Our hypothesis was that a component with increased micro-motion in the laboratory at time zero would also exhibit a greater amount of radiolucency in patients at a minimum of 2 years post total shoulder arthroplasty.Methods: Thirty cadaveric shoulders were divided into 2 groups (keel vs. peg). The glenoid components were first loaded with a single axial eccentric force of 196 N in all orientations and then with a transversal load of 49 N to simulate in vivo loads with abduction. Displacement of the glenoid component was determined with four different linear variable-differential transducers. In the second phase, 56 antero-posterior x-rays of 52 patients with either the same keeled (n = 24) or pegged (n = 32) glenoid component with a minimum of 24 months follow-up were evaluated for radiolucency.Results: Biomechanically the pegged glenoid showed a significant increase in micro-motion during eccentric axial loading as well as during combined loading in the anterior, posterior, and inferior position as compared to the keeled glenoid (p < 0.05). In contrast all results were significant with greater radiolucency for the keeled glenoid component (p=0.001).Conclusion: While the pegged component exhibited a greater amount of micro-motion during biomechanical testing, radiolucency was greater in patients with a keeled component. These findings provide support for both components from different perspectives and highlight the need for well-constructed studies to determine whether glenoid design has an effect on clinical outcome, because influences are multifactorial and biomechanical forces may not recreate forces seen in vivo.

To evaluate the biomechanical performance of different techniques for CC reconstruction using suture button systems with integrated tendon augmentation. Hypothesis was that (1) reconstructions using a cortical button combined with a biological augmentation (semitendinosus allograft) will demonstrate improved stability than a modified Weaver–Dunn procedure and (2) constructs using two tunnels at the clavicle for fixation will show superior horizontal stability than single-tunnel constructs.

PURPOSE:To evaluate the biomechanical performance of an all-soft suture anchor (JuggerKnot; Biomet, Warsaw, IN) in comparison with a classic solid suture anchor (2.4-mm biocomposite SutureTak; Arthrex, Naples, FL) in an in vitro labral repair model. METHODS:We dissected 12 cadaveric shoulders (mean age, 61 ± 9.4 years), leaving the labrum intact, and bone mineral density was obtained (mean, 0.375 ± 0.06 g/cm(3)). Simulated labral tears were made at the anteroinferior and posteroinferior edges of the labrum. Repairs used 2 all-soft suture anchors (JuggerKnot) or 2 solid anchors with free, high-strength No. 2 suture (FiberWire; Arthrex) spanning the operative construct to load the repair. Differential variable reluctance transducers were used to measure labral displacement for each specimen. The testing protocol consisted of a preconditioning phase at 10 N for 10 cycles (1 Hz) and then a final load-to-failure testing at a rate of 3 mm/min. Labral displacement of 2 mm was determined as the primary outcome. RESULTS:There was no statistical difference (P = .22) in ultimate load to failure and displacement at ultimate failure (anchor pullout) between the all-soft JuggerKnot (146.0 ± 43.0 N and 19.8 ± 5.4 mm, respectively) and the solid SutureTak (171.9 ± 52.6 N and 22.3 ± 6.8 mm, respectively). The solid anchor had a significantly higher ultimate load at 2 mm of labral displacement than the all-soft suture anchor (84.1 ± 19.0 N and 39.2 ± 10.6 N, respectively; P < .001). CONCLUSIONS:Whereas both the solid SutureTak and the all-soft JuggerKnot displayed similar results on ultimate load-to-failure testing, the solid anchor required significantly greater load for 2 mm of labral displacement than the all-soft anchor. CLINICAL RELEVANCE:The all-soft anchor (JuggerKnot) is similar in biomechanical performance to the classic solid anchor (SutureTak) with the exception of load at 2 mm of labral displacement, suggesting micromotion of the device.

Purpose: To evaluate the effect of entry and exit points of the coracoid tunnel on load to failure and mode of failure, to reduce the incidence of coracoid fractures and acromioclavicular joint repair failures. Methods: This study investigates 5 tunnel placements based on different entry and exit points in the coracoid process: center-center orientation represents perfect placement of the bone tunnel and served as perfect tunnel placement in our study. Four common errors in drilling were then tested and acted as the experimental groups in our study (medial-center, center-medial, lateral-center, and center-lateral). Using 35 cadaveric shoulders (mean age, 68.0 +/- 13.0 years), we tested these 5 tunnel orientations using a single repair technique (cortical button) loaded to failure on an MTS 858 Servohydraulic test system (MTS Systems, Eden Prairie, MN). A control group of 7 cadaveric shoulders without the presence of a coracoid tunnel was also tested to determine the type of fracture pattern that occurred. Results: The coracoids without tunnel drilling fractured in patterns similar to traumatic coracoid injuries. With regard to the 5 tunnel groups, it was found that the loads to failure with center-center and medial-center tunnel placement were significantly higher than those with center-medial, center-lateral, and lateral-center tunnel placement. The failure modes of the former were primarily within the repair constructs, whereas those of the latter were primarily due to bony failure. Conclusions: Our biomechanical results showed a higher peak load to failure with a center-center or medial-center tunnel orientation, which may lessen the risk of coracoid fracture during drilling with a 6-mm cannulated drill bit. Clinical Relevance: Proper trajectory of the drill during formation of a coracoid bone tunnel can help reduce the risk of coracoid process fracture and repair failure.

Background: Arthroscopic procedures for reconstruction of acromioclavicular (AC) joint separations are increasingly used in clinical practice. Multiple surgical techniques exist, but there are still few data on biomechanical performances of commonly used arthroscopic techniques and fixation methods. Hypothesis: Single and double clavicular tunnel reconstructions show comparable primary stability with a modified Weaver-Dunn procedure, and double tunnel constructs show superior horizontal stability. Study Design: Controlled laboratory study. Methods: The AC joints of 40 cadaveric shoulders were tested for anterior, posterior, and superior translation (70-N load) and maximal load to failure (superior) with the MTS 858 Bionix II Servohydraulic testing system. Shoulders were assigned to 4 groups: (1) native (n = 18), (2) coracoclavicular (CC) reconstruction with 1 clavicular and 1 coracoid tunnel (SCT) fixed with a suture pulley and 2 buttons (n = 8), (3) CC reconstruction with 2 clavicular and 1 coracoid tunnel (DCT) fixed with a suture pulley and 3 buttons (n = 8), and (4) modified Weaver-Dunn reconstruction (n = 6). Results: Native specimens showed a mean anterior translation of 7.92 mm (±1.69 mm), a mean posterior translation of 7.84 mm (±2.09 mm), and a superior translation of 4.28 mm (±1.81 mm). Maximal load to failure was 579.44 N (±148.01 N). The SCT technique showed a mean anterior translation of 5.81 mm (±1.16 mm), posterior translation of 8.30 mm (±1.94 mm), and a superior translation of 2.28 mm (±0.52 mm). The maximal load to failure was 591.35 N (±231.17 N). Anterior and superior translations were significantly less compared with the native specimen (P = .005 and P = .003). The DCT technique had an anterior translation of 4.68 mm (±0.6 mm), posterior translation of 6.85 mm (±0.83 mm), and superior translation of 2.09 mm (±0.86 mm). The mean maximal load to failure was 651.16 N (±226.93 N). Anterior and superior translations were significantly less compared with the native specimens (P = .000 and P = .001). No statistically significant differences were shown between SCT and DCT reconstruction for all measurements (P > .05). One reconstruction of the modified Weaver-Dunn procedure failed directly after mounting it into the testing device. The remaining 5 showed a mean anterior translation of 11.36 mm (±3.17 mm), a mean posterior translation of 13.51 mm (±2.21 mm), and a mean superior translation of 3.31 mm (±0.47 mm). Anterior and posterior translations were significantly increased compared with the native specimen (P = .019 and P = .000). The mean maximal load to failure measured 311.13 N (±52.2 N) and was significantly less compared with the native specimen (P = .000). The Weaver-Dunn technique showed significantly less maximal load to failure and more anterior and posterior translation compared with SCT and DCT (P ≤ .05). Conclusion: Isolated reconstruction of the CC ligaments using single and double clavicular tunnel techniques results in a high load to failure for superior translation, which is equal to the native stability, and less translation in all 3 directions as well as higher superior stability when compared with the modified Weaver-Dunn procedure. A potential drawback is the risk of coracoid fracture, as the high load to failure of the device may exceed load to failure of cortical bone prior to device breakage. Clinical Relevance: Single clavicular tunnel arthroscopic reconstructions of the coracoacromial ligaments show good biomechanical results.

Purpose: The aim of this study was to examine rotator cuff strain and gap size after margin convergence was performed for a large retracted rotator cuff tear. Methods: We tested 20 cadaveric shoulders using a custom shoulder testing system. A large retracted rotator cuff tear was created by removing the supraspinatus muscle-tendon unit to provide a reproducible model. Margin convergence was performed and strain was measured by use of differential variable reluctance transducers in the intact state, after a massive rotator cuff tear was created, and after each of 5 margin convergence sutures were placed. Data were obtained at 0 and 60 of abduction and with internal and external rotational torques applied to the humerus. Gap size was measured before and after margin convergence sutures were placed. Results: Strain was significantly reduced at all degrees of rotation in 0 of abduction after margin convergence sutures were placed (P < .05). There was a significantly significant decrease in gap size with each suture: 50% with the first suture, 60% with the second suture, 67% with the third suture, and 75% with the fourth suture (P < .05). There was only minimal intrinsic rotator cuff tension during knot tying, with each subsequent suture having less of an effect than the previous. Four margin convergence sutures resulted in a mean of 5 mm of anterior humeral head translation. Conclusions: There was a significant decrease in rotator cuff strain and gap size after margin convergence was performed for a large retracted tear. The first margin convergence suture caused the greatest increase in intrinsic rotator cuff tension, with each subsequent suture having a similar but less dramatic effect.

The purpose of our study was to evaluate all contact pressures between the molded ankle-foot orthosis (MAFO) and the subject during activities of daily living. The MAFOs studied are used clinically to reduce plantar contact pressures associated with foot ulcers in adult neuropathic diabetic subjects, alleviating abnormal pressures by redistributing them to low-pressure plantar regions. While effective, MAFOs are often not used by the subject due to weight and comfort issues. An understanding of the contact pressures between the subject and the orthosis is a first step in improving basic MAFO design. Four nonimpaired, young adult males were tested in this study. A right-side MAFO was custom-molded and fitted for each subject by the same orthotist. Real-time pressures were obtained for the entire contact area using the F-Scan pressure measurement system. The data obtained demonstrated high contact pressures along the metatarsals of the foot, around the heel and ankle, and adjacent to the strap attachment sites. No contact pressures were noted along the posterior calf region during any of the activities performed. These data suggest the calf region would be a suitable site for material removal for weight reduction and increased comfort, especially in warm weather. In addition, these data may be useful to orthotists in improving the basic design and to researchers as a starting point for performing complex finite element analysis on the MAFO.

Calcaneocuboid joint pressures were evaluated with eight cadaver specimens. Real-time pressures were recorded using a TekScan 4200 sensor pad at lengths of 0, 5, and 10 mm in both unloaded and 350-newton loaded models. Recorded pressures exceeded 2.3 MPa in the loaded model at 10 mm lateral column lengthening. Although an acceptable procedure in the pediatric population, application of the Evans lateral column lengthening procedure for management of adult acquired flatfoot may generate excessive pressures leading to joint arthrosis. Lengthening by calcaneocuboid distraction arthrodesis may avoid this problem.

Real time contact pressure pads were used to evaluate calcaneal cuboid joint pressures as a result of lateral column lengthening. The Evan’s lateral column lengthening procedure has been used to restore the arch of the human foot after acquired adult flatfoot. There has been some question as to the consequences of this procedure with regards to calcanealcuboid arthrosis. This study evaluated the calcaneal cuboid contact pressure without lengthening, and after 5 and 10 mm lengthening. Each length was also evaluated with a compressive load of 350 N applied to the tibia. Total contact forces and pressures increased statistically as lengthening increased and after load application. Peak pressures also increased. Pressures exceeded 2.3 MPa in the loaded feet at 10 mm lengthening. Patients subjected to lateral column lengthening of 10 mm may be at increased risk for joint arthrosis.

Surgical repair of recurrent anterior shoulder instability requires secure fixation of the separated inferior glenohumeral complex to bone. Many techniques of fixation are in use for both arthroscopic and open repair. The specific aim of this study was to compare the initial failure strength of eight repair techniques using a previously described canine model of Bankart repair. Intact capsule-to-bone complexes failed at the bony interface at 236 N. Traditional Bankart repair failed at 122.1 N (2 sutures) and 74.7 N (1 suture), Acufex TAG rod (Acufex Microsurgical, Mansfield, MA) at 143.5 N (2 sutures) and 79.8 N (1 suture), transglenoid suture technique (2 sutures) at 166.6 N, Mitek GII (Mitek, Norwood, MA) (1 suture) at 96.4 N, Zimmer Statak (Zimmer Inc, Warsaw, IN) (1 suture) at 95.2 N, and Acufex bioabsorpable Suretac at 82.2 N. The two-suture repairs were statistically equivalent in strength to each other, as were the one-suture repairs and the Suretac device. Two-suture repairs were significantly stronger than the one-suture repairs (P < .01) failure. In the single-suture specimens, failure occurred by suture breakage in 46% (18 of 39) of specimens and soft-tissue failure around the suture in 54% (21 of 39), Failure in the two-suture techniques primarily occurred by soft-tissue failure (23 of 25) and this proved a statistically significant difference (P < .003). No device broke or pulled out of bone. Our results indicate that in a soft-tissue-to-bone repair model (1) pullout of suture anchors is a rare event and suggests that pullout strength of suture anchors should not be the sole basis of comparison of one device to another in Bankart repairs; (2) suture techniques and anchor devices that allow for two sutures exhibit a stronger initial pullout strength than one-suture techniques and devices; (3) suture anchor techniques are equivalent in strength to suture-alone techniques; and (4) all repair techniques are significantly weaker than undisturbed (control) specimens indicating that the ultimate outcome of a Bankart repair, open or arthroscopic, is dependent on physiological repair of soft-tissue to bone. Further study is required to document the strength of soft-tissue-to-bone repair with time. These results are repair specific. Soft-tissue-tobone repair in other areas, i.e., rotator cuff, may show different failure mechanics.

We have used a dye-flow visualization technique to analyze the substructure of flow separation in a plastic model of the human carotid bifurcation under steady and pulsatile flow. Under steady conditions at a physiologic flow split (Q external carotid/Q common carotid = 0.30) and Reynolds number (500), a large region of separated fluid developed along the outside wall of the sinus, opposite the flow divider. Yellow dye injected into the boundary layer upstream from the bifurcation traveled slowly along the wall of the common carotid and entered directly into the separation. Blue dye injected into the central, high-velocity streamlines in the common carotid impacted on the flow divider, then traveled circumferentially and entered the separation. Mixing of these two sources was documented by the appearance of green fluid, which lingered in the region of separation. Pulsatile flow resulted in a smaller region of separation; mixing still occurred. Flow separation at the carotid bifurcation is a site of mixing of fluids previously subjected to prolonged low-shear wall contact and brief high-shear wall contact. Separation is itself a site of low shear, but this study reveals a mechanism whereby low and high shear may act independently or synergistically to explain the link between flow separation and atherogenesis.

The precise location and progression of anastomotic hyperplasia and its possible relationship to flow disturbances was investigated in femoro-femoral Dacron grafts in 28 dogs. In 13 grafts, the outflow from the end-to-side downstream anastomosis was bidirectional (BDO), and in 15 it was unidirectional (UDO) (distally). Grafts were electively removed at intervals of two to 196 days or at the time of thrombosis. Each anastomosis and adjacent artery was perfusion-fixed and sectioned sagittally. The mean sagittal section was projected onto a digitized pad, and the total area of hyperplasia internal to the arterial internal elastic lamina and within the adjacent graft was integrated by computer. The location of the hyperplasia was compared with previously established sites of flow separation and stagnation. The observation was made that hyperplasia is significantly greater at the downstream, as compared with the upstream, anastomosis in both groups (BDO = p less than 0.001 and UDO = p less than 0.001) (analysis of variance for independent groups). Furthermore, this downstream hyperplasia was progressive with time (BDO p less than 0.01) (UDO p less than 0.01); Spearman Rank Correlation. There was no significant increase in the extent of downstream hyperplasia where flow separation was known to be greater (BDO). Five grafts failed (three BDO, two UDO), as a result of complete occlusion of the downstream anastomosis by fibrous hyperplasia. Transmission electron microscopy showed the hyperplasia to consist of collagen-producing smooth muscle cells. Anastomotic hyperplasia is significantly greater at the downstream anastomosis, is progressive with time, and is the primary cause of failure of Dacron arterial grafts in this model. Quantitative analysis of downstream anastomotic hyperplasia may be a valuable measure of the biocompatibility of Dacron grafts.

The precise location and progression of anastomotic hyperplasia and its possible relationship to flow disturbances was investigated in femoro-femoral Dacron grafts in 28 dogs. In 13 grafts, the outflow from the end-to-side downstream anastomosis was bidirectional (BDO), and in 15 it was unidirectional (UDO) (distally). Grafts were electively removed at intervals of two to 196 days or at the time of thrombosis. Each anastomosis and adjacent artery was perfusion-fixed and sectioned sagittally. The mean sagittal section was projected onto a digitized pad, and the total area of hyperplasia internal to the arterial internal elastic lamina and within the adjacent graft was integrated by computer. The location of the hyperplasia was compared with previously established sites of flow separation and stagnation. The observation was made that hyperplasia is significantly greater at the downstream, as compared with the upstream, anastomosis in both groups (BDO = p less than 0.001 and UDO = p less than 0.001) (analysis of variance for independent groups). Furthermore, this downstream hyperplasia was progressive with time (BDO p less than 0.01) (UDO p less than 0.01); Spearman Rank Correlation. There was no significant increase in the extent of downstream hyperplasia where flow separation was known to be greater (BDO). Five grafts failed (three BDO, two UDO), as a result of complete occlusion of the downstream anastomosis by fibrous hyperplasia. Transmission electron microscopy showed the hyperplasia to consist of collagen-producing smooth muscle cells. Anastomotic hyperplasia is significantly greater at the downstream anastomosis, is progressive with time, and is the primary cause of failure of Dacron arterial grafts in this model. Quantitative analysis of downstream anastomotic hyperplasia may be a valuable measure of the biocompatibility of Dacron grafts.

The flow fields of model side-to-end vascular anastomoses, such as those used at the proximal end of axillofemoral and femoropopliteal grafts were analyzed. Clear plastic models of anastomoses of various angles (15°, 20°, 30°, 45°, 90°, 135°, 150°, 160°, 165°) were subjected to flow splits (flow exiting from the side branch/inflow) of 0.00 to 1.00, using water under steady-flow conditions. Nearwall flow was visualized by selective dye injection upstream. Regions of reversed, slow-moving nearwall fluid (separated flow) opposite the side-branch origin were evident at all angles, occurring at a flow split of 0.28 at 15° and progressively lower values at higher angles. A narrow ring of separated fluid formed about the anastomosis, entering the side branch. At flow splits of over 0.45, disturbed flow in the anastomotic region occurred in models with angles over 45°, due to the deflection of the mainstream toward the side branch, obliterating the near-wall flow evident at lower flow splits. The lack of large regions of near stagnation evident in our models at flow splits normally seen in clinical grafts (0.50 and above), especially at sharply angled anastomoses, may in part explain the clinical observation that hyperplasia occurs primarily at distal, as compared with proximal, anastomoses. Furthermore, the added platelet-wall contact caused by the disturbed flow at the proximal anastomosis may contribute to downstream hyperplasia.