We started direct bypass treatment for MMD at the University of Fukui Hospital in September 2009. Of 179 consecutive STA-MCA anastomoses on 95 hemispheres of 71 MMD patients at the University of Fukui Hospital between 2009 and 2023, 49 anastomoses on 26 hemispheres of 21 patients were enrolled in this clinical trial. Three anastomoses on two hemispheres in two patients were excluded because bypass occlusion was confirmed using intraoperative indocyanine green (ICG) video-angiography (VA). Two anastomoses on one hemisphere in one patient were performed using scissor-attached micro-forceps and were occluded probably due to the thrombus formation in the donor artery derived from the injury of STA during the harvest. Although the reconstruction of STA was performed by end-to-end anastomosis, repeated anastomosis under the administration of anti-platelet drugs could not relieve the occlusion. Another anastomosis in another patient conducted by using conventional micro-forceps was occluded probably due to the complexity of anastomosis owing to the discrepancy of diameters between the donor artery (2 mm) and the recipient artery (0.5 mm). Ultimately, 46 anastomoses on 25 hemispheres in 19 patients were included in this study. Anastomosis in the unilateral hemisphere was performed in 14 patients and on both hemispheres in 6 patients. Among 46 anastomoses on 25 hemispheres in this study, single, double, and triple anastomoses underwent 5, 19, and 1 hemispheres, respectively. While both conventional micro-forceps and newly created scissors-attached micro-forceps were used for double or triple bypass in the same operation on 17 hemispheres, only one of them was used for single anastomosis on five hemispheres in five patients and double anastomosis on three hemispheres in three patients. Accordingly, 21 anastomoses on 19 hemispheres of 16 patients were performed using the conventional micro-forceps (conventional group: CG), and 25 anastomoses on 22 hemispheres of 19 patients were performed using the scissors-attached micro-forceps (scissors group: SG; Figure 1). This study was approved by the Ethics Committee of our institute (No. 20150155), and informed consent was obtained from all patients.

The newly developed micro-forceps are 135 mm long and consist of three parts: body, head, and tip (Charmant Co., Ltd., Fukui, Japan, www.charmant.co.jp), which were made of Ti15V3Al3Cr3Sn, stainless steel (SUS304), and aged stainless steel (SUS20J-2), respectively. Each part is fixed and connected using stainless steel screws (SUS304; Figure 2A). The tip is 0.35 mm wide and 0.8 mm long. A small pair of scissors is placed at a distance of 0.8 mm from the tip for cutting ligatures (Figure 2B). This small-shaped tip design makes it possible to deal smoothly with 11-0 ligatures and needles (Figure 2C). The micro-forceps allow the surgeon to suture, tie (Figure 2C), and cut a ligature without exchanging instruments (Figures 2D–F), thus reducing the clamp time.

A large U-shaped scalp incision was made around the ears. After reflection of the scalp and temporal muscles, a frontotemporal craniotomy was performed to expose the frontal lobe, Sylvian fissure, and temporal lobe. In most cases, we performed double anastomosis using the parietal and frontal STAs. The supra-Sylvian M4 portion of the MCA and infra-Sylvian M4 portion immediately beside the Sylvian vein were selected as the recipient arteries. Several lateral branches of the recipient artery were coagulated and sacrificed to obtain a 10 mm long branch-free recipient. After clamping the recipient artery, a circular arteriotomy was performed at its roof.

The STA-MCA anastomosis was performed using 8–10 sutures with 11-0 ligatures. One anastomosis was performed using the conventional micro-forceps, and the other was performed using the scissors-attached micro-forceps. When harvesting two branches of the STA was difficult, a single anastomosis was performed. In this case, the micro-forceps used for the anastomosis were randomly selected by the surgeon.

ICG-VA was performed in all the cases. Occlusion of three anastomoses in two patients was confirmed using ICG-VA at the final stage of the operation. After the STA-MCA anastomosis, the temporal lobe was covered with the temporal muscle using encephalo-myo-synangiosis. During the procedure, the systolic blood pressure was maintained between 100 and 140 mmHg, and the PaCO₂ was strictly maintained between 35 and 45 mmHg. The surgery was performed by KK and four other neurosurgeons.

MR studies were performed before surgery and within 1 week after surgery. Images were obtained using a 3T MR scanner (Signa 3-HD, General Electric, USA) with axial DWI (spin echo EPI, TR/TE = 6,300/64.4 ms, bandwidth = ±250 kHz, FOV = 240, slice thickness/gap = 5/1 mm, matrix = 128 × 256, 2NEX), axial FLAIR sequences (fast spin echo, TR/TE/TI = 10,000/120/2,450 ms, bandwidth = ±35.71 kHz, FOV = 240 mm, slice thickness/gap = 5/1 mm, matrix = 320 × 192, 1NEX), and three-dimensional time-of-flight MR angiography [(3D TOF MRA) (3D SPGR, TR/TE = 22/3.5 ms, flip angle = 18°, band width = ±25 kHz, FOV = 180 mm, matrix = 320 × 192, slice thickness/gap = 1.2/−0.6 mm (ZIP2), number of slabs = 3 (location per slab = 28; overlap, five slices), 1NEX, ASSET Factor = 1.5)].

Anastomotic site infarction (ASI) was defined as an asymptomatic DWI-hyperintense lesion with a diameter <5 mm on the cortex near the Sylvian fissure detected by postoperative 3T MRI. An ASI at the temporal lobe was determined as a small ischemic lesion related to the bypass to the infra-Sylvian MCA (preoperative MRI: Figure 3A, postoperative MRI: Figure 3B). ASIs at the frontal and temporal lobes were determined as ischemic lesions related to the bypass to both the supra- and infra-Sylvian MCAs (preoperative MRI: Figure 3C, postoperative MRI: Figure 3D). Symptomatic large DWI-positive lesions or lesions in the contralateral brain were identified as major infarctions related to bypass surgery for MMD and were distinguished from ASIs (preoperative MRI: Figure 3E, postoperative MRI: Figure 3F).

A univariate analysis was performed with Pearson's chi-squared test, Fisher's exact test for categorical variables, or using the Mann–Whitney U-test for numeric variables. Forward and backward stepwise logistic regression analyses with the Akaike information criterion (AIC) were carried out to determine the associations of potential factors and the occurrence of ASI. The cutoff values for the receiver operating characteristic (ROC) analysis using the area under the curve (AUC) were calculated using Benis' method. All statistical analyses were performed using the JMP software (Version 10, SAS Institute Inc., Cary, NC, USA) and R (R Foundation for Statistical Computing, Vienna, Austria), with an error probability of <0.05.

There were 16 patients [(mean age, 35.4 ± 18.4 years; male-to-female ratio (M:F) = 3:13)] in the CG (21 anastomoses cases on 19 hemispheres) and 19 (mean age 33.0 ± 20.1 years, M:F = 2:17) in the SG (25 anastomoses cases on 22 hemispheres). There were 10 patients with preprocedural in the CG group and 11 in the SG group, respectively. There was one patient recent ischemic stroke in each group. There were 12, 3, and 1 patients with a preoperative modified Rankin score (preop mRS) of 0, 1, and 2, respectively in CG (mean 0.31±0.60). There were 13, 3, 2, and 1 patients with preop mRS of 0, 1, 2, and 3, respectively, in SG (mean 0.52 ± 0.90). There were 2, 13, and 1 patients with the preoperative Suzuki stage of 2, 3, and 4, respectively, in CG (mean 2.94 ± 0.44). There were 6, 11, 1, and 1 patients with the preoperative Suzuki stage of 2, 3, 4, and 5, respectively, in SG (mean 2.84 ± 0.76) (8). Regarding the outcome of bypass function, there were 7, 7, and 2 patients with Matsushima grades A, B, and C, respectively, determined by postoperative angiography at 3 months after the last surgery in CG. There were 9, 8, and 2 patients with Matsushima grades A, B, and C, respectively, in SG. The mean follow-up period of CG and SG were 67.0 ± 18.4 and 69.4 ± 15.9 months, respectively (9). There were 12 and 4 patients with mRS at the final follow-up of 0 and 1, respectively, in CG (mean 0.25 ± 0.45). There were 13, 4, and 1 patients with preop mRS of 0, 1, 2, and 3, respectively, in SG (mean 0.47 ± 0.84). There was no significant difference in the number of patients, number of anastomoses, number of males, mean age, the presence of preprocedural infarction, recent ischemic stroke, preop mRS, Suzuki stage, Matsushima grade of postoperative angiography, mean follow-up period, and mRS at the final-follow up between the two groups. In a patient-oriented comparison, the mean number of sacrificed lateral branches of the recipient artery in CG (21 anastomose) and SG (25 anastomose) were 2.4 ± 1.1 and 2.9 ± 2.2, respectively. The mean number of sacrificed lateral branches with a diameter larger than 200 mm in CG and SG were 0.14 ± 0.36 and 0.2 ± 0.58, respectively. The mean number of sutures was 10.0 ± 1.2 and 10.0 ± 1.8, respectively. There was also no significant difference in the number of sacrificed lateral branches, the number of sacrificed lateral branches with a diameter larger than 200 μm, and the number of sutures between the two groups (Table 1).

The clamp time of the recipient artery regarding the CG and SG was 41.0 ± 11.6 min and 24.2 ± 6.9 min, respectively. The clamp time in the SG was significantly shorter than that in the CG (p < 0.001; Figure 4A). Furthermore, the clamp time per suture in the CG and SG was 4.2 ± 1.4 and 2.4 ± 0.5 min, respectively. The clamp time per suture was significantly shorter in the SG than in the CG (p < 0.001; Figure 4B). The occurrence rates of ASI in the CG and SG were 67 and 16%, respectively. The occurrence of ASIs in CG and SG was 52.3 and 20.0%, respectively. The occurrence of ASIs was more significant in the CG than in the SG (p = 0.0312; Figure 4C).

Forward and backward stepwise logistic regression analyses with AIC revealed the occurrence of ASI were predicted by the combination of the number of sacrificed lateral branches of the recipient artery, the number of sacrificed lateral branches with a diameter larger than 200 μm, and the clamp time with the minimal AIC value (AIC = 54.17). Among them, the clamp time was the only significant predictive factor [odds ratio, 1.087; 95% confidence interval (CI): 1.03–1.16, p = 0.0080; Table 2].

The ROC analysis revealed that the clamp time was a significant predictive factor for the occurrence of ASI (p = 0.0041; AUC = 0.7346; cutoff value, 33.2 min; sensitivity, 68.8%; specificity, 54.6%; Figure 5A). Moreover, the clamp time per suture was a significant predictive factor for the occurrence of ASI (p = 0.0047; AUC = 0.7271 cutoff value, 3.6 min; sensitivity, 58.3%; specificity, 57.1%; Figure 5B).

This study had some limitations. We conducted double anastomosis by using conventional micro-forceps for one anastomosis and scissors-attached micro-forceps for another anastomosis in the same operation on 17 hemispheres of 25 hemispheres in this study. Therefore, 33 anastomose of 46 anastomose in this study were performed using both micro-forceps in the same hemisphere in the same person. Approximately 75% of the patient clinical parameters in CG and SG were the same. That is because we avoided the logistic regression analysis regarding the prediction of ASI occurrence including patient clinical parameters in this study. In addition, it was a retrospective study with a relatively small sample size. Further prospective studies with more patients are required to confirm our results.