兽研引言
犬超声引导下胸段竖脊肌平面（ESP）阻滞的研究有限。本研究旨在描述一种在T9横突背外侧缘（DLETP）的横向（TV）超声引导ESP入路，将其扩散范围与纵向（LNG）入路进行比较，并分析胸段横突解剖结构。
方法
本双盲实验研究使用23具犬尸体。对21具尸体的每侧在T9采用TV或LNG入路进行ESP注射，使用0.6 mL/kg染料/造影剂，随后进行计算机断层扫描和解剖。
结果
5具尸体被排除在解剖分析之外。所有TV注射均置于DLETP上方，而18.75%的LNG注射偏向外侧。两种入路均在ESP内产生头尾侧扩散，TV组和LNG组的中位（范围）扩散分别为8（5–14）和8.5（4–13）个椎体。TV组头侧扩散超过尾侧，为4.5（2–6）对3.5（0–7）个椎体（p = 0.027），LNG组为5（3–7）对3（0–6）（p = 0.004）。脊神经背支内侧支（MBDR）染色显示TV和LNG的中位（范围）值分别为5（4–8）和5（0–6）。TV组头侧MBDR染色占优势，为4.5（1–6）对尾侧1（0–3）（p = 0.003），LNG组为4（0–6）对0（0–3）（p < 0.001）。背支外侧支（LBDR）染色显示TV和LNG分别为4（1–9）和4（2–12）。TV组头侧LBDR染色超过尾侧，为3（0–6）对1（0–3）（p = 0.019），LNG组为3（1–8）对1（0–5）（p = 0.034）。腹支染色发生于10/16次TV注射（62.5%）和11/16次（68.75%）LNG注射。两种入路均显示硬膜外和交感干扩散，而LNG发生一次血管内注射。两种入路之间未发现显著差异。DLETP在T10尾侧主要缺失，且小于横突（p < 0.001）。
结论
TV和LNG入路在犬T9产生相当的神经染色和ESP扩散，具有优先头侧分布。TV入路减少了注射物错位。DLETP小于横突，并在T11头侧一致可识别。
[内容请查看原文]
引言
自竖脊肌平面（ESP）阻滞首次在人类中描述以来（1），该技术在兽医和人类医学中引起了越来越多的兴趣，并已在多种动物物种中进行了尸体研究（2–12）。
关于ESP阻滞作用机制的各种理论已在人类中提出（8, 13）。最一致观察到的模式涉及染料扩散至脊神经背支的外侧支（LBDR）和内侧支（MBDR）。这些神经支配脊柱的背外侧，并在犬尸体中一致被染色（2, 4）。进一步扩散至脊神经腹支（VBSN）、椎旁间隙、交感干、硬膜外间隙和淋巴系统已在人类和兽医物种中报道（3, 12, 14–18）。
竖脊肌构成覆盖椎骨和肋骨背外侧的轴上肌群的一部分，包括最长肌、半棘肌和髂肋肌（4）。ESP阻滞最初使用超声（US）引导的纵向旁矢状入路（LNG）进行（1, 2, 4）。最近，横向（TV）入路已在人类和兽医医学中描述（18–21）。迄今为止，仅有一项兽医研究直接比较了同一区域即T12横突的两种入路，发现注射物扩散相当（18）。然而，由于肌肉和椎骨解剖的差异，这些结果不能外推至更头侧的椎骨（22, 23）。
横突背外侧缘（DLETP）是竖脊肌和多裂肌的附着部位，是最初描述用于在犬T9水平进行ESP阻滞的解剖标志（4）。然而，该标志并非在所有胸椎中一致可识别（18），其存在、形态和变异性尚未被全面表征。
本研究的目标是描述一种新的US引导TV入路用于DLETP上方的胸段ESP阻滞，比较LNG和TV US引导ESP阻滞入路在T9的神经染色和注射物扩散，并研究胸椎横突的解剖。我们假设两种入路将产生相似的神经扩散和染料分布。
[内容请查看原文]
材料与方法
本前瞻性、随机、双盲、实验性和尸体解剖研究已在巴伦西亚天主教大学伦理批准下进行（CEEAUCV2210）。
[内容请查看原文]
动物
本研究共纳入23具成年犬尸体，并将其分配至两个独立的实验周（图1）。第一周使用12具尸体，第二周使用11具尸体（图1）。所有尸体均在室温（20–22 °C）下解冻48小时。23具尸体中的2具（每周各1具）被随机分配用于初步解剖学研究。其余21具用于比较传统LNG入路与TV入路在T9水平的注射物扩散情况，并分析胸椎横突的解剖结构（图1）。尸体数量参考了此前发表的类似尸体研究（4, 18, 24–26）。选择比既往研究更多的尸体数量，以提高解剖学观察的精确性和稳健性，并考虑到可能的排除情况。排除标准包括脊柱创伤、任何可能影响ESP通透性的损伤以及因腐败导致ESP和肌肉内气体过度积聚。所有标本均来源于自然死亡或与本研究的无关原因实施安乐死后的尸体。第一实验周的尸体还曾用于一项无关的尸体研究，但未对本次研究评估的解剖结构造成干扰（25）。
[图：图示研究期间尸体分配及竖脊肌平面注射分布的示意图。ESP，]
[内容请查看原文]
初步解剖学评估
每周从23具尸体中分配1具（共2具）用于初步解剖学评估，进行超声检查，随后进行大体解剖剥离，以使研究人员熟悉局部解剖和技术（图1）。将动物置于胸卧姿势，剃除胸背部被毛。一名兽医麻醉师使用连接Samsung HM70EVO便携式超声仪（Samsung®）的16–3 MHz线阵探头进行超声评估，以确定TV入路ESP注射的相关标志。随后按照Herrera-Linares等人（18）描述的方案进行大体解剖剥离。每个实验周由不同的兽医外科医生进行解剖剥离，以识别脊神经的内侧支、外侧支和腹侧支，以及椎旁间隙和交感干（图2a）。
[图：解剖剥离图像。(a) 侧视图显示竖脊肌平面（ESP），并识别]
[内容请查看原文]
超声引导下LNG与TV ESP注射
在23只解冻犬中，第一周和第二周分别有11具和10具尸体接受了超声引导下ESP注射（图1）。每具尸体每侧均接受一次LNG和一次TV ESP注射。所有21只处理犬的注射部位（左侧或右侧）及不同入路（LNG或TV）均使用随机化软件（https://www.randomizer.org）随机分配。第一实验周，所有LNG注射由一名兽医麻醉师完成，所有TV注射由另一名兽医麻醉师完成。第二周，由于第一周实施LNG注射的麻醉师无法到场，所有注射均由第一周负责TV注射的兽医麻醉师完成。所有尸体均置于胸卧姿势，操作前剃除注射区域被毛。使用连接Samsung HM70EVO便携式超声仪（Samsung®）的16–3 MHz线阵探头引导所有注射。
对于LNG入路，将线阵探头平行于脊柱放置于注射同侧、最后一根肋骨紧后方。向头侧移动探头直至识别出最后一根肋骨的骨性声影，并重复此过程向头侧移动直至第9肋可视化。在此水平，将探头向内侧移动，以获得对应于T9 DLETP的不规则高回声形态（图3b, c）。采用平面内技术，将回声绝缘针（22G，80 mm，Ultraplex 360® B. Braun Medical Ltd.）以头尾方向穿过竖脊肌，直至针尖接触DLETP的背侧面（图3a, c）（4, 18, 24）。轻轻将针尖压向骨面，尽可能快速地注射总量为0.6 mL/kg的1:1亚甲蓝（Methylthioninium Chloride Injection 1% w/v；Martindale Pharmaceuticals Ltd.）与碘海醇造影剂（Omnipaque 300；GE Healthcare）混合液。
[图：T9水平超声引导下竖脊肌平面（ESP）注射。纵向（LNG）的技术、图像和标志]
对于TV入路，将探头置于正中线最后一根肋骨尾侧，垂直于脊柱，并略微向注射侧倾斜。然后将探头向头侧移动，直至第13肋的圆形骨性声影出现在图像的外侧。从此处开始，沿背外侧至腹内侧方向追踪该肋骨，直至其与内侧出现的相应横突的高回声表面接触。重复此操作并向头侧移动，直至第9肋显影并与T9横突接触，使第九肋横突关节进入成像窗口（图3e、f）。此定位可清晰显示椎板、横突、DLETP、肋横突关节和肋结节，从而便于精确识别注射部位（图3e、f）。采用平面内技术，将针沿背外侧至腹内侧方向推进，直至针尖到达DLETP的背侧面（图3d、f；补充视频S1）。将针尖轻压于骨面，然后在此位置以先前所述LNG入路的相同速率和相同总容量注射先前使用的相同染料/造影剂混合物。
[内容请查看原文]
计算机断层扫描研究及解剖剥离
注射后立即将每具尸体转移至计算机断层扫描（CT）区域。使用16排多探测器CT扫描仪（Siemens Somatom Scope，慕尼黑，德国）以螺旋模式进行胸腰段CT扫描。CT采集参数包括512×512矩阵、螺距0.9、1.0 mm层厚、180 mAs、130 kV，显示视野根据患者体型调整。使用DICOM查看器（OsiriX MD v14.2.0）以软组织（窗宽[WW]=350 HU，窗位[WL]=40 HU）和骨（WW=1500 HU，WL=450 HU）重建算法审阅图像。扫描由同一名对治疗分组不知情的兽医放射科医师评估。记录横突长度、DLETP的存在及长度、以及过渡椎或肋骨或椎体异常的存在（图4）。测量并比较两种入路在ESP、硬膜外腔、椎旁间隙和淋巴系统内的注射物扩散范围。采用半定量评分系统评估腐败产生的气体程度，以CT图像上竖脊肌平面及周围肌肉内气体的存在为参考。评分定义如下：0=未检测到气体，1=中等但可接受的气体存在，2=过度且不可接受的气体存在（图5）。
[图：从T1到T13的横断面CT扫描，显示横突及横突背外侧缘（w]
[图：代表性横断面CT图像，说明用于评估死后气体产生的半定量评分系统]
所有解剖剥离均在CT后立即由参与初步解剖研究的同一名外科医师按照先前使用的相同技术进行（18）。外科医师对治疗分配不知情。当染料以大于1 cm的长度环形包绕神经时，MBDR、LBDR和VBSN的染色被认为是成功的（27）。还通过计数扩散的椎体数量记录ESP内的纵向扩散范围。如果至少覆盖其长度的50%，则认为椎体扩散完全（28）。
[内容请查看原文]
统计分析
结果以分类变量的n（%）表示，非参数连续数据以中位数（范围）表示，参数数据以均值±标准差（SD）表示。使用Kolmogorov–Smirnov检验评估分析变量的正态性。分类变量使用Fisher精确检验进行比较。对于连续变量，独立样本比较使用Student's t检验（参数数据）或Mann-Whitney U检验（非参数数据），而配对比较使用Wilcoxon符号秩检验（非参数数据）。对于涉及两组以上连续变量的比较，使用单因素方差分析，随后使用Tukey事后检验进行比较。所有检验中p值<0.05被认为具有统计学意义。对于独立样本和配对样本中具有统计学意义的差异，计算Hodges-Lehmann中位数差异95%置信区间（CI）。数据分析和图表生成使用IBM SPSS Statistics for macOS（版本30.0；IBM Corp.）和GraphPad Prism（版本11）。
[内容请查看原文]
结果
[内容请查看原文]
动物
二十一只犬接受了双侧ESP注射。尸体包括4只混种犬（19.05%）、3只灵缇（14.29%）、3只比格犬（14.29%）、2只边境牧羊犬（9.52%），以及比熊犬、布拉科犬、可卡犬、德国牧羊犬、哈士奇、迷你柯利犬、波登科犬、贵宾犬和斯塔福德郡梗各1只（各占4.76%）。平均（±SD）体重为17.69±6.24 kg，体况评分为6（5–8）/9。
经CT评估后，21具尸体中有1具（4.76%）被排除在注射物扩散分析之外，因其肌肉组织内存在多个异物，可能影响ESP通透性。在剩余20具尸体中，因腐败导致气体积聚的半定量评估结果为：20具中9具（45%）评分为0，7具（35%）评分为1，4具（20%）评分为2。4具（20%）评分为2的尸体因ESP及周围肌肉组织内气体积聚过多而被排除在扩散分析之外（图5）。CT评估还显示21具尸体中存在解剖学异常：21具中2具（9.52%）有胸腰过渡椎，1具（4.76%）在T10尾侧腹侧椎板出现骨质新生，5具（23.81%）存在脊椎退行性变，1具（4.76%）患有椎间盘疾病。全部21具尸体均纳入横突的解剖学研究（图4）。
[内容请查看原文]
超声引导下ESP注射与分组
每具尸体每侧均接受TV和LNG两种注射，每组（TV组和LNG组）共21次注射，总计42次注射。按照前述CT评估后的排除标准，42次注射中32次（76.19%）纳入扩散分析，每组剩余16次注射（图1）。TV组所有注射均在预定目标T9处DLETP上进行。相比之下，LNG组16次注射中有3次（18.75%）——1次由一名兽医麻醉师操作、2次由另一名操作者完成——被无意中置于DLETP外侧，导致以腹外侧扩散为主（图6b、d）。
[图：计算机断层扫描（CT）影像学研究显示竖脊肌平面注射后造影剂/染料的分布]
[内容请查看原文]
注射物扩散的解剖分析
解剖记录显示，TV组和LNG组注入ESP的总扩散长度分别为8（5–14）个椎体和8.5（4–13）个椎体。最头侧的扩散在TV组达到T3，在LNG组达到T2；最尾侧的扩散分别达到L3和L2。两组中头侧扩散均见于16/16次注射（100%），而尾侧扩散见于TV组14/16次（87.5%）和LNG组12/16次（75%）。两组的头侧扩散均显著大于尾侧扩散：TV组为4.5（2–6）个椎体 vs. 3.5（0–7）个椎体（p = 0.027，95% CI：−2.5，0），LNG组为5（3–7）个椎体 vs. 3（0–6）个椎体（p = 0.004，95% CI：−3，−1）（图7a）。
[图：每具尸体竖脊肌平面（ESP）横向（TV）和纵向（LNG）入路注射后的扩散]
TV组所有注射（100%）均染色MBDR，LNG组16次注射中有15次（93.75%）染色，染色神经数中位数（范围）分别为5（4–8）和5（0–6）（图2b、7a）。两组中染色最头侧和最尾侧的MBDR分别为T3和T11。TV组中头侧染色见于16/16次注射（100%），染色MBDR中位数为4.5（1–6），而尾侧染色见于13/16次注射（81.25%），中位数为1（0–3）。LNG组中头侧染色见于15/16次注射（93.75%），染色分支中位数为4（0–6），而尾侧染色见于11/16次注射（68.75%），中位数为0（0–3）。两组中头侧染色范围均显著大于尾侧（TV：p = 0.003，95% CI：−4.5，−1.5；LNG：p < 0.001，95% CI：−4，−2）。
两种入路的所有注射（100%）均染色LBDR，TV组和LNG组染色分支数中位数（范围）分别为4（1–9）和4（2–12）（图2b、7a）。在所有尸体中，染色最头侧的LBDR在TV组为T3、在LNG组为T1，而最尾侧在两组均为T13。头侧染色见于TV组14/16次注射（87.5%）和LNG组16/16次注射（100%），染色分支中位数分别为3（0–6）和3（1–8）。尾侧染色见于TV组14/16次注射（87.5%）和LNG组13/16次注射（81.25%），TV组和LNG组染色分支中位数分别为1（0–3）和1（0–5）。LBDR的头侧扩散在TV组（p = 0.019，95% CI：−3.5，−0.5）和LNG组（p = 0.034，95% CI：−3，0）也均显著大于尾侧扩散。
在背侧椎旁间隙内观察到VBSN完全染色：TV组10/16次注射（62.5%），LNG组11/16次注射（68.75%），两组染色VBSN中位数均为1（0–5）（图2c、7a）。VBSN头侧染色见于TV组5/16次注射（31.25%）和LNG组6/16次注射（37.5%），而尾侧染色见于TV组4/16次注射（25%）和LNG组7/16次注射（43.75%）。解剖中观察到的腹侧椎旁扩散见于TV组7/16次注射（43.75%）和LNG组6/16次注射（37.5%）（图2c）。交感扩散见于TV组6/16次注射（37.5%）和LNG组7/16次注射（43.75%）（图2d、7a）。
编号7、9和13的犬只接受了LNG注射，注射位置误置于DLETP外侧，结果显示要么无MBDR染色，要么仅1或2个MBDR染色，且相对于注射部位，染色位置比预期更靠颅侧（图7a）。在TV组中，编号7的犬只也观察到类似的颅侧偏移，尽管CT分析确认其扩散模式与该入路其他注射相似。编号8的犬只在TV和LNG入路中，首个染色神经根水平也出现颅侧偏移，尽管CT确认其扩散模式与各组其余注射观察到的相当（图7a）。
在尸体解剖的扩散分析中，TV组和LNG组之间未检测到统计学显著差异。
[内容请查看原文]
注射物扩散的CT分析
CT记录的ESP内总扩散长度在TV组和LNG组分别为8（5–15）和8（5–17）个椎体，两组最颅侧扩散均在C7，最尾侧扩散TV组在L3、LNG组在L2（图6a、c，7b）。颅侧扩散显示TV组和LNG组的中位数（范围）分别为4（2–8）和3.5（1–9）个椎体。尾侧扩散TV组为3（1–7）个椎体，LNG组为3（1–6）个椎体。CT分析期间，TV注射中12例（75%）、LNG注射中13例（81.25%）观察到腹侧椎旁扩散，中位扩散分别为1（0–5）和2（0–3）个椎体（图6a、b、d，7b）。TV入路中10例（62.5%）、LNG入路中11例（68.75%）记录到硬膜外扩散，两者中位数均为1（0–15）（图6e，7b）。编号7的犬只与其他动物相比硬膜外扩散增加，两种入路均从T12延伸至C5。LNG组观察到1例（6.25%）部分血管内注射，TV组无（图6b、f）。
CT分析后，TV组和LNG组之间或各组内颅侧与尾侧扩散之间均未检测到统计学显著差异。
[内容请查看原文]
横突的解剖学研究
CT分析横突和DLETP的存在及长度结果见图8表格（图4、8a）。21具尸体中17.5具（83.33%）在T10以尾侧DLETP缺失，21具中3.5具（16.67%）在T11缺失。在每个存在DLETP的椎体中，其长度显著小于横突长度（p < 0.001，95% CI见图8a）（图4、8a、b）。椎体间DLETP分析显示前三个椎体之间以及这些椎体与其余椎体之间存在显著差异（p < 0.001）（图8b）。还观察到尺寸逐渐增大，从T1的2.16 ± 0.58至T11的4.13 ± 1.38（p < 0.001）（图8a、b）。ANOVA检验还显示椎体间横突长度总体存在显著差异（p < 0.001）。然而，组间分析仅显示最后三个椎体与其余椎体之间存在差异（图8a、b）。
[图：横突背外侧缘（DLETP）和横突（TP）测量结果在]
[内容请查看原文]
讨论
正如假设，在T9使用0.6 mL/kg的LNG和TV入路均产生可比的ESP间隙扩散，以及背支外侧支和内侧支的一致染色。然而，由于注射物相对于DLETP的误置，LNG入路导致不良外侧扩散的发生率更高。
CT分析后，少数动物因腐败继发过多软组织气体积聚而被排除，这可能改变了注射溶液的分布（29）。选择解冻期是为了标准化程序并确保与先前犬类研究的可比性（12, 18, 24）。然而，这些研究中未报告可比并发症。这种差异可能由解冻过程中环境温度的差异解释，其他研究未一致报告该温度，可能引入分解速率的变异性（12, 18, 24）。体成分和死亡原因等额外因素可能进一步使标准化复杂化。此外，先前研究缺乏气体积聚的明确排除标准，加上死后分解对注射物扩散的影响目前未知，限制了直接可比性（12, 18, 24）。本研究未评估腐败导致气体水平增加的动物中注射物扩散的改变。因此，需要进一步研究确定尸体分解如何影响此类尸体研究中溶液的分布。研究死后分解对注射物分布的影响可能代表有价值的未来研究方向，改善标准化及与活体受试者的可比性。
LBDR和MBDR是为从正中线到腋线的轴上肌、椎板、韧带、小关节和皮肤提供感觉和运动功能的神经（30, 31）。LBDR和MBDR染色是ESP注射后最常见的发现之一，这可能解释了对涉及这些结构的操作中ESP阻滞所报告的临床镇痛效果（8, 32）。在本研究中，LBDR和MBDR均被一致染色，结果与先前在犬中的研究相似，且两种入路产生相当的染色效果（4, 18）。一个重要发现是，使用两种入路时，MBDR和LBDR染色的颅侧范围以及ESP内染料扩散的颅侧范围均大于其尾侧范围。这与先前的报告一致，并提示注射物在竖脊肌平面内优先向颅侧分布（25）。这些发现可能具有临床意义，因为它们提示将注射位置置于手术部位尾侧可以利用本研究中观察到的优先颅侧扩散。然而，这一假设应在活体动物中加以证实，因为尸体染料分布可能无法完全反映生理条件下的局部麻醉药扩散。
两组均证实了VBSN的染色，该神经为胸壁、肋骨、胸膜和胸骨提供感觉和运动功能（30, 31, 33, 34）。这与在犬、猫和马中的尸体研究以及人类关于VBSN染色和椎旁扩散的报告一致（2, 9, 12, 15, 18, 35, 36）。然而，其他犬和猪的研究发现染色仅限于背支（4, 24）。两组中还观察到硬膜外扩散，其中一只动物的扩散范围比其他动物更广泛。这与一些犬和马尸体研究中的发现相似（9, 18, 25），但与其他犬或猫的研究不一致（4, 12, 24, 37）。本研究中发现的VBSN染色和硬膜外扩散支持ESP阻滞对涉及胸部的手术或疼痛性操作的潜在镇痛效果，并与先前在ESP阻滞后接受胸部操作的犬中的报告一致（10, 38）。此外，两组的注射均导致交感干染色，这与一些人类、犬和猫研究中的发现一致（8, 12, 18, 39），但与其他犬研究不一致（4, 24, 37）。在T9处ESP注射后扩散至交感干可能有助于该技术对胸部内脏痛的镇痛效果（40）。然而，ESP阻滞后单侧交感干染色的临床意义仍有待确定。先前在犬T12处报道的淋巴扩散（18）在此处未出现，可能反映了注射部位的解剖差异（22, 23）。除MBDR和LBDR染色外，这种扩散的变异性与人类研究中报道的情况相似（41–43）。此外，这表明这些扩散模式的潜在机制仍不清楚，差异可能归因于各研究之间针尖位置、注射深度、注射物体积和速度、体况评分和体型、动物年龄以及个体间筋膜解剖的差异。需要进一步的标准化研究以更好地了解这些分布。然而，这些结果连同现有证据支持ESP阻滞对轴上肌、皮肤和椎骨的潜在镇痛效果，而对胸壁、胸腔和内脏、胸骨以及扩展的硬膜外扩散的覆盖仍不一致（10, 11, 32, 38）。
与TV入路相比，LNG技术即使在由不同操作者实施时，也与意外注射至DLETP外侧肋骨的风险更高相关，并改变了注射液的扩散模式。此前犬类研究未在T12报告此现象（18）；然而，尽管T9的横突比其他胸椎更长，DLETP仍相对较小，形成狭窄的注射目标区域。这些发现表明，在具有DLETP的椎体上实施时，TV入路可能优于LNG技术。采用LNG入路时，探头轻微的侧向或内侧移动即可使DLETP移出成像视野，导致肋骨或横突的背侧面被误认为DLETP。相比之下，TV入路提供横断面视图，可同时显示椎板、横突、DLETP、肋横突关节和肋结节。因此，探头向头侧或尾侧移动可将横突间肌和肋间肌纳入视野，降低误置风险。此外，LNG入路中的肌内注射会产生纵向肌纤维的透镜状分离，类似于ESP水分离术，而TV入路中观察到的环形扩散模式更明确地提示误置（19–21）。操作者经验和手工灵巧度也可能导致LNG入路中观察到的误置注射。尽管两位操作者均经验丰富且熟悉LNG入路，但个体技术差异可能影响针具操作和注射物沉积的准确性。LNG注射在预期目标区域之外并非唯一与意外扩散模式相关的因素。两组中少量正确实施的注射显示MBDR染色出现非典型头侧偏移，首个染色神经根相对于注射部位的位置比预期更靠头侧。在一例中，该模式在对侧镜像出现，提示个体解剖变异而非仅技术因素可能独立影响注射液在ESP内的头尾分布。
胸椎横突的形态学评估显示，横突和DLETP长度存在显著变异，且DLETP在不同椎体水平的存在不一致。DLETP在最尾侧胸椎中一致缺失，但缺失的具体椎体水平在尸体间存在差异。这些发现表明，本研究描述的TV入路应仅应用于具有DLETP的椎体，而在该结构缺失时推荐采用Herrera-Linares等人（18）描述的TV入路。
本研究并非没有局限性。死后变化可能影响注射染料溶液的行为和分布，可能与活体动物中观察到的模式不同。活体动物中存在的若干生理过程在尸体中缺失。这些包括血液和淋巴流动、胸腔负压、肌张力、呼吸时胸廓运动以及局部麻醉药扩散，所有这些在体内通常存在，并可能影响注射物质在组织内的整体行为。由于这些过程在死后模型中缺失，本研究观察到的扩散模式应在此局限性下解读，进一步在活体动物中进行研究将有助于确认这些发现。另一局限性是每组实施注射的麻醉师数量。LNG入路由两位麻醉师实施注射，而TV组仅由一位完成所有注射。仅使用一位操作者可能引入与技术水平或对TV技术熟悉程度相关的操作者依赖性偏倚。然而，仅一人具备实施这一新技术所需的培训。因此，为避免第二位操作者以减少针具放置不准确的风险并确保一致性。
本尸体研究结果表明，在具有该解剖结构的椎体上，于第9胸椎DLETP处实施TV和LNG超声引导ESP入路进行ESP注射是可行的。两种技术在ESP内产生相当的多节段扩散，具有优先向头侧神经的染料和扩散分布。然而，与LNG入路相比，TV入路与注射物误置于期望目标区域之外及随后不良侧向扩散的风险较低相关。本研究还显示，DLETP小于横突，且在T11头侧的椎体中一致可识别，尽管也可能存在于T11。因此，本研究描述的TV入路不应在无DLETP的椎体上复制。需要临床研究评估该技术在具有DLETP的胸椎上的镇痛疗效和安全性。
查看英文原文▾
Introduction
Studies of the ultrasound-guided thoracic erector spinae plane (ESP) blocks in dogs are limited. This study aimed to describe a transverse (TV) ultrasound-guided ESP approach at the dorsolateral edge of the transverse process (DLETP) of T9, compare its distribution with a longitudinal (LNG) approach, and analyze the thoracic transverse process anatomy.
Methods
Twenty-three canine cadavers were used in this double-blinded experimental study. An ESP injection using 0.6 mL/kg of dye/contrast was performed on each lateral of 21 cadavers with a TV or LNG approach at T9, followed by computed tomography and dissection.
Results
Five cadavers were excluded from dissection analysis. All TV injections were placed over the DLETP, whereas 18.75% of LNG injections were misplaced laterally. Both approaches produced craniocaudal spread within the ESP, with median (range) spreads of 8 (5–14) and 8.5 (4–13) vertebrae for the TV and LNG groups, respectively. Cranial spread exceeded caudal in the TV group, with 4.5 (2–6) vs. 3.5 (0–7) vertebrae (p = 0.027), and LNG, with 5 (3–7) vs. 3 (0–6) (p = 0.004). Staining of the medial branches of the dorsal rami (MBDR) of the spinal nerves showed median (range) values of 5 (4–8) and 5 (0–6) in TV and LNG. Cranial MBDR staining predominated in TV, with 4.5 (1–6) over caudal with 1 (0–3) (p = 0.003), and LNG, with 4 (0–6) vs. 0 (0–3) (p < 0.001). Lateral branches (LBDR) staining showed 4 (1–9) and 4 (2–12) in TV and LNG. Cranial LBDR staining exceeded caudal in TV, with 3 (0–6) vs. 1 (0–3) (p = 0.019), and LNG, with 3 (1–8) vs. 1 (0–5) (p = 0.034). Ventral branches staining occurred in 10/16 TV injections (62.5%) and 11/16 (68.75%) LNG. Both approaches showed epidural and sympathetic trunk spread while one intravascular injection occurred in LNG. No significant differences were found between approaches. The DLETP was predominantly absent caudal to T10 and smaller than the transverse process (p < 0.001).
Conclusion
The TV and LNG approaches produced comparable nerve staining and ESP spread at T9 in dogs, with preferential cranial distribution. The TV approach reduced injectate misplacement. The DLETP was smaller than the transverse process and consistently identified cranial to T11.
[内容请查看原文]
Introduction
Since the erector spinae plane (ESP) block was first described in humans (1), a growing interest in this technique has emerged in both veterinary and human medicine, with cadaveric studies conducted across several animal species (2–12).
Various theories regarding the mechanism of action of the ESP block have been proposed in humans (8, 13). The most consistently observed pattern involves dye spread to the lateral (LBDR) and medial branches (MBDR) of the dorsal rami of the spinal nerves. These nerves innervate the dorsolateral aspect of the spine and have been consistently stained in canine cadavers (2, 4). Further spread to the ventral branches of the spinal nerves (VBSN), paravertebral space, sympathetic trunk, epidural space, and lymphatic system has been reported in both humans and veterinary species (3, 12, 14–18).
The erector spinae muscles form part of the epaxial musculature covering the dorsolateral aspect of the vertebrae and ribs, including the longissimus, semispinalis, and iliocostalis muscles (4). The ESP block was originally performed using an ultrasound (US)-guided longitudinal parasagittal approach (LNG) (1, 2, 4). More recently, a transverse (TV) approach has been described in both human and veterinary medicine (18–21). To date, only one veterinary study has directly compared both approaches over the same area, the transverse process of T12, finding comparable injectate spread (18). However, these results cannot be extrapolated to more cranial vertebrae due to differences in musculature and vertebral anatomy (22, 23).
The dorsolateral edge of the transverse process (DLETP), the insertion site of the erector spinae and multifidus muscles, was the anatomical landmark originally described for performing the ESP block at T9 level in dogs (4). However, this landmark is not consistently identifiable across all thoracic vertebrae (18), and its presence, morphology, and variability have not been comprehensively characterized.
The objectives of this study were to describe a novel US-guided TV approach for the thoracic ESP block over the DLETP, to compare nerve staining and injectate spread between the LNG and the TV US-guided ESP block approaches at T9, and to study the anatomy of the transverse process of the thoracic vertebrae. We hypothesized that both approaches would produce similar nerve spread and dye distribution.
[内容请查看原文]
Materials and methods
This prospective, randomized, double-blinded, experimental and cadaveric anatomical study has been performed under the ethical approval of the Catholic University of Valencia (CEEAUCV2210).
[内容请查看原文]
Animals
A total of 23 adult canine cadavers were included in this study and allocated into two separate weeks (Figure 1). Twelve cadavers were used in the first week and 11 cadavers in the second week (Figure 1). All cadavers were thawed at room temperature for 48 h (20–22 °C). Two cadavers out of the 23, one for each week, were randomly allocated to a preliminary anatomical investigation. The remaining 21 were used to compare the injectate spread between traditional LNG approach with the TV approach at T9 and to analyze the anatomy of the transverse process of the thoracic vertebrae (Figure 1). The number of cadavers was based on similar previously published cadaveric studies (4, 18, 24–26). A larger number of cadavers than in previous studies was selected to improve the precision and robustness of the anatomical observations, and to account for potential exclusions. Exclusion criteria included spinal trauma, any damage that could affect permeability of the ESP and excessive accumulation of gas within the ESP and muscles due to putrefaction. All specimens were obtained after natural death or euthanasia for reasons unrelated to this study. The cadavers from the first experimental week were also used in an unrelated cadaveric study, without interference with the anatomical structures evaluated in the present investigation (25).
[图：Diagram illustrating cadaver allocation and distribution of the erector spinae plane injections during the study. ESP, e]
[内容请查看原文]
Preliminary anatomical assessment
One of the 23 cadavers was allocated each week (two in total) to a preliminary anatomical assessment and underwent ultrasonographic examination followed by gross anatomical dissection to familiarize the investigators with the regional anatomy and technique (Figure 1). With the animal positioned in sternal recumbency, the dorsal thoracic region was clipped. A veterinary anesthetist performed the ultrasonographic evaluation using a 16–3 MHz linear transducer connected to a Samsung HM70EVO portable US machine (Samsung®) to identify the relevant landmarks for the TV approach for the ESP injection. Gross anatomical dissection was then performed following the protocol described by Herrera-Linares et al. (18). A different veterinary surgeon performed the dissections during each experimental week to identify the medial, lateral, and ventral branches of the spinal nerves, as well as the paravertebral space and the sympathetic trunk (Figure 2a).
[图：Images of the anatomical dissection. (a) Lateral views showing the erector spinae plane (ESP) with the identification of]
[内容请查看原文]
Ultrasound-guided LNG & TV ESP injection
Out of the 23 thawed dogs, 11 and 10 cadavers underwent a US-guided ESP injection during the first and second week, respectively (Figure 1). Each cadaver received a LNG and a TV ESP injection in each lateral. The injection sites (left or right lateral) and the different approaches (LNG or TV) were randomly assigned using a randomization software (https://www.randomizer.org) for all the 21 treated dogs. In the first experimental week, all LNG injections were performed by one veterinary anesthetist, and all TV injections were performed by a different veterinary anesthetist. In the second week, because the anesthetist who performed the LNG injections during the first week was unavailable, all injections were performed by the veterinary anesthetist responsible for the TV injections in the first week. All cadavers were positioned in sternal recumbency, and the injection area was clipped prior to the procedure. A 16–3 MHz linear transducer connected to a Samsung HM70EVO portable US machine (Samsung®) was used to guide all injections.
For the LNG approach, the linear transducer was placed parallel to the spine on the ipsilateral side of the injection, immediately caudal to the last rib. The probe was moved cranially until the bony shadow of the last rib was identified, and this process was repeated cranially until the 9th rib was visualized. At this level, the transducer was moved medially to obtain the irregular hyperechoic shape corresponding to the DLETP of T9 (Figures 3b, c). Using an in-plane technique, an echogenic insulated needle (22-gauge, 80 mm, Ultraplex 360® B. Braun Medical Ltd.) was advanced through the erector spinae musculature in a cranial caudal direction until the tip contacted the dorsal aspect of the DLETP (Figures 3a, c) (4, 18, 24). The needle tip was gently pressed against the bone and a total volume of 0.6 mL/kg of a 1:1 mixture of methylene blue (Methylthioninium Chloride Injection 1% w/v; Martindale Pharmaceuticals Ltd.) and iohexol contrast medium (Omnipaque 300; GE Healthcare) was injected as rapidly as possible.
[图：Ultrasound-guided erector spinae plane (ESP) injection at T9. Technique, images, and landmarks of the longitudinal (LNG)]
For the TV approach, the transducer was placed caudal to the last rib at the midline, perpendicular to the spine, and slightly tilted toward the side of the injection. The probe was then moved cranially until the rounded bony shadow of the 13th rib appeared on the lateral aspect of the image. From this point, the rib was traced in a dorsolateral-to-ventromedial direction until it contacted the hyperechoic surface of the corresponding transverse process that appeared medially. This procedure was repeated cranially until the 9th rib was visualized contacting the transverse process of T9, placing the ninth costotransverse joint within the imaging window (Figures 3e, f). This positioning allowed clear visualization of the lamina, transverse process, DLETP, costotransverse joint, and rib tubercle, thereby facilitating precise identification of the injection site (Figures 3e, f). Using an in-plane technique, the needle was advanced in a dorsolateral-to-ventromedial direction until the tip reached the dorsal aspect of the DLETP (Figures 3d, f; Supplementary video S1). The needle tip was gently pressed against the bone and the same dye/contrast mixture previously used was then injected at this location at the same rate and with the same total volume previously described for the LNG approach.
[内容请查看原文]
Computed tomography study & anatomical dissection
Immediately after the injections, each cadaver was transferred to the computed tomography (CT) area. A thoracolumbar CT scan was performed using a 16–slice multidetector CT scanner (Siemens Somatom Scope, Munich, Germany) in helical mode. CT acquisition parameters included a 512 × 512 matrix, a pitch of 0.9, 1.0–mm slice thickness, 180 mAs, 130 kV, and a display field of view adjusted to patient size. Images were reviewed using a DICOM viewer (OsiriX MD v14.2.0) with soft tissue (window width [WW] = 350 HU, window level [WL] = 40 HU) and bone (WW = 1500 HU, WL = 450 HU) reconstruction algorithms. The scans were evaluated by the same veterinary radiologist blinded to the treatment group. The length of the transverse process, presence and length of DLETP, and presence of transitional vertebrae or ribs or vertebrae abnormalities were recorded (Figure 4). The extent of injectate spread within the ESP, epidural space, paravertebral space, and lymphatic system was measured and compared between the two approaches. A semiquantitative scoring system was used to assess the degree of gas produced by putrefaction, using the presence of gas within the erector spinae plane and surrounding musculature as a reference on the CT images. Scores were defined as follows: 0 = no gas detected, 1 = moderate but acceptable presence of gas, and 2 = excessive and unacceptable presence of gas (Figure 5).
[图：Transverse CT scans from T1 to T13 showing the transverse process and the dorsolateral edge of the transverse process (w]
[图：Representative transverse CT images illustrating the semiquantitative scoring system used to assess postmortem gas produ]
All anatomical dissections were performed immediately after CT by the same surgeon involved in the preliminary anatomical study, following the same technique as previously used (18). The surgeons were blinded to treatment allocation. Staining of the MBDR, LBDR, and VBSN was considered successful when dye circumferentially surrounded the nerve over a length greater than 1 cm (27). The longitudinal extent of spread within the ESP was also recorded by counting the number of vertebrae of the spread. Vertebral spread was considered complete if at least 50% of its length was covered (28).
[内容请查看原文]
Statistical analysis
Results are presented as n (%) for categorical variables, median (range) for non-parametric continuous data, and mean ± standard deviation (SD) for parametric data. Normality of the analyzed variables was assessed using the Kolmogorov–Smirnov test. Categorical variables were compared using Fisher's exact test. For continuous variables, independent-samples comparisons were made using Student's t-test (parametric data) or the Mann-Whitney U test (non-parametric data), while paired comparisons were made using Wilcoxon signed-rank test (non-parametric data). For comparisons of continuous variables involving more than two groups, a one-way ANOVA was used, followed by Tukey's post hoc test for comparisons. A p-value < 0.05 was considered statistically significant for all tests. Hodges-Lehmann median difference 95% confidence interval (CI) was calculated for statistically significant differences in both independent and paired samples. For data analysis and figure generation, IBM SPSS Statistics for macOS (Version 30.0; IBM Corp.) and GraphPad Prism (Version 11) were used.
[内容请查看原文]
Results
[内容请查看原文]
Animals
Twenty-one dogs received a bilateral ESP injection. Cadavers included 4 mixed-breed dogs (19.05%), 3 Greyhounds (14.29%), 3 Beagles (14.29%), 2 Border Collies (9.52%), and 1 each of Bichon Frise, Braco, Cocker Spaniel, German Shepherd, Husky, Miniature Collie, Podenco, Poodle, and Staffordshire Terrier (4.76% each). The mean (± SD) body weight was 17.69 ± 6.24 kg with a body condition score of 6 (5–8) out of 9.
After CT evaluation, 1 of the 21 cadavers (4.76%) was excluded for the analysis of the injectate spread due to multiple foreign bodies within the musculature that could potentially affect ESP permeability. In the remaining 20 cadavers, the semiquantitative assessment of gas accumulation due to putrefaction yielded a score of 0 in 9 cadavers (45%), 1 in 7 cadavers (35%), and 2 in 4 cadavers (20%) out of 20. The 4 cadavers (20%) with a score of 2 were excluded for analysis of the spread due to excessive gas accumulation within the ESP and surrounding musculature (Figure 5). CT evaluation also revealed anatomical abnormalities among the 21 cadavers: 2 out of 21 (9.52%) had a transitional thoracolumbar vertebra, 1 (4.76%) exhibited bone neoformation of the ventral lamina caudal to T10, 5 (23.81%) presented spondylosis, and 1 (4.76%) had intervertebral disc disease. All the 21 cadavers were included for the anatomical study of the transverse process (Figure 4).
[内容请查看原文]
Ultrasound-guided ESP injections & groups
Each cadaver received both, a TV and an LNG injection on each side, resulting in a total of 21 injections per group (TV and LNG) and 42 injections overall. After applying the previously described exclusion criteria following CT evaluation, 32 of the total 42 injections (76.19%) were included in the analysis of spread, leaving 16 injections per group (Figure 1). All injections in the TV group were performed over the intended target, the DLETP at T9. In contrast, 3 injections out of the 16 (18.75%) in the LNG group, 1 by a veterinary anesthetist and 2 by the other, were inadvertently placed lateral to the DLETP resulting in predominantly ventrolateral spread (Figures 6b, d).
[图：Computed tomography (CT) imaging study showing distribution of contrast/dye solution following the erector spinae plane ]
[内容请查看原文]
Dissection analysis of injectate spread
Dissection-recorded total spread length into the ESP was 8 (5–14) and 8.5 (4–13) vertebrae for the TV and LNG groups, respectively. The most cranial spread reached T3 in the TV group and T2 in the LNG group, whereas the most caudal spread reached L3 and L2, respectively. Cranial spread was found in 16/16 (100%) in both groups while caudal was present in 14/16 (87.5%) in the TV group and 12/16 (75%) in LNG. Cranial spread was significantly greater than caudal spread in both the TV group with 4.5 (2–6) vs. 3.5 (0–7) vertebrae (p = 0.027, 95% CI: −2.5, 0), and the LNG group with 5 (3–7) vs. 3 (0–6) vertebrae (p = 0.004, 95% CI: −3, −1) (Figure 7a).
[图：Spread following transverse (TV) and longitudinal (LNG) approaches of the erector spinae plane (ESP) injections in each ]
MBDR were stained in all injections (100%) in the TV group and in 15 of 16 injections (93.75%) in the LNG group, with a median (range) of 5 (4–8) and 5 (0–6) nerves stained, respectively (Figures 2b, 7a). The most cranial and caudal MBDR stained were T3 and T11 in both groups. In the TV group, cranial staining was present in 16/16 (100%) injections with a median of 4.5 (1–6) MBDR stained, compared with 13/16 (81.25%) injections and 1 (0–3) caudally. In the LNG group, cranial staining was present in 15/16 (93.75%) injections with a median of 4 (0–6) branches stained, compared with 11/16 (68.75%) injections and a median of 0 (0–3) branches caudally. Cranial staining was significantly more extensive than caudal in both groups (TV: p = 0.003, 95% CI: −4.5, −1.5); LNG: p < 0.001, 95% CI: −4, −2).
LBDR were stained in all injections (100%) for both approaches, with a median (range) of 4 (1–9) and 4 (2–12) branches stained for the TV and LNG groups, respectively (Figures 2b, 7a). Across all cadavers, the most cranial LBDR stained was T3 in the TV group and T1 in the LNG group, while the most caudal was T13 in both groups. Cranial staining was present in 14/16 (87.5%) injections in the TV group and 16/16 (100%) in the LNG group, with a median of 3 (0–6) and 3 (1–8) branches stained, respectively. Caudal staining occurred in 14/16 (87.5%) injections in the TV group and 13/16 (81.25%) in the LNG group, with a median of 1 (0–3) and 1 (0–5) branches stained in the TV and LNG group, respectively. Cranial spread of LBDR also significantly exceeded caudal spread in the TV (p = 0.019, 95% CI: −3.5, −0.5) and the LNG group (p = 0.034, 95% CI: −3, 0).
Complete staining of the VBSN within the dorsal paravertebral compartment was observed in 10/16 injections (62.5%) in the TV group and 11/16 injections (68.75%) in the LNG group, with a median of 1 (0–5) VBSN stained in both groups (Figures 2c, 7a). Cranial staining of the VBSN was present in 5/16 (31.25%) injections in the TV group and 6/16 (37.5%) in the LNG group while caudal staining was observed in 4/16 (25%) injections in the TV group and 7/16 (43.75%) in the LNG group. Ventral paravertebral spread observed during dissection occurred in 7/16 (43.75%) injections in the TV group and 6/16 (37.5%) in the LNG group (Figure 2c). Sympathetic spread was observed in 6/16 (37.5%) and 7/16 (43.75%) injections in the TV and LNG groups, respectively (Figures 2d, 7a).
Dogs 7, 9, and 13 received LNG injections misplaced lateral to the DLETP and showed either no MBDR staining or staining of only 1 or 2 MBDR, located more cranially than expected relative to the injection site (Figure 7a). A similar cranial shift was observed in dog 7 within the TV group, although CT analysis confirmed a spread pattern similar to the other injections using this approach. Dog 8 also showed a cranial shift in the level of the first stained nerve root with the TV and LNG approaches, despite CT confirming a spread pattern comparable to that observed in the remaining injections for each group (Figure 7a).
No statistically significant differences were detected between the TV and LNG groups in the spread analysis on cadaveric dissection.
[内容请查看原文]
CT analysis of injectate spread
CT-recorded total spread length within the ESP was 8 (5–15) and 8 (5–17) vertebrae for the TV and LNG, respectively, with the most cranial spread at C7 in both groups and the most caudal at L3 and L2 for TV and LNG, respectively (Figures 6a, c, 7b). Cranial spread showed a median (range) spread of 4 (2–8) and 3.5 (1–9) vertebrae for the TV and LNG group, respectively. Caudal spread presented 3 (1–7) vertebrae in the TV and 3 (1–6) in the LNG group. Ventral paravertebral spread was observed in 12 TV injections (75%) and 13 (81.25%) LNG during CT analysis, with a median spread of 1 (0–5) and 2 (0–3) vertebrae, respectively (Figures 6a, b, d, 7b). Epidural spread was recorded in 10 (62.5%) using the TV approach and 11 (68.75%) using LNG, with a median of 1 (0–15) for both (Figures 6e, 7b). Dog number 7 showed an increased epidural spread compared with the other animals, extending from T12 to C5 for both approaches. One (6.25%) partial intravascular injection was visualized in the LNG group with none in the TV (Figures 6b, f).
No statistically significant differences were detected between the TV and LNG groups or between cranial and caudal spread within each group after CT analysis.
[内容请查看原文]
Anatomical study of the transverse process
CT analysis of the presence and length of the transvers process and the DLETP are shown in the table in Figure 8 (Figures 4, 8a). The DLETP was absent caudal to T10 in 17.5 out of 21 cadavers (83.33%) and at T11 in 3.5 out of 21 cadavers (16.67%). In each vertebra where the DLETP was present, it was significantly smaller than the length of the transverse process (p < 0.001, 95% CIs are shown in Figure 8a) (Figures 4, 8a, b). Analysis of the DLETP between vertebrae showed significant differences among the first three vertebrae, as well as between these and the remaining vertebrae (p < 0.001) (Figure 8b). A progressive increase in size was also observed, ranging from 2.16 ± 0.58 at T1 to 4.13 ± 1.38 at T11 (p < 0.001) (Figures 8a, b). The ANOVA test also demonstrated overall significant differences in the transverse processes length between vertebrae (p < 0.001). However, the analysis between groups showed only differences between the last three vertebrae and the rest (Figures 8a, b).
[图：Results of the measurements of the dorsolateral edge of the transverse process (DLETP) and transverse process (TP) in th]
[内容请查看原文]
Discussion
As hypothesized, both the LNG and TV approaches using 0.6 mL/kg at T9 produced comparable spread through the ESP compartment and consistent staining of the lateral and medial branches of the dorsal rami. However, the LNG approach resulted in a greater incidence of undesired lateral spread due to misplacement of the injectate relative to the DLETP.
Following CT analysis, few animals were excluded due to excessive soft tissue gas accumulation secondary to putrefaction, which could have altered the distribution of the injected solution (29). The thawing period was selected to standardize the procedure and ensure comparability with previous canine studies (12, 18, 24). However, no comparable complications have been reported in those studies. This discrepancy may be explained by differences in ambient temperature during the thawing process, which were not consistently reported in other studies and could have introduced variability in decomposition rate (12, 18, 24). Additional factors such as body composition and cause of death may further complicate standardization. Furthermore, the absence of explicit exclusion criteria for gas accumulation in previous studies, combined with the currently unknown effect of post-mortem decomposition on injectate spread, limits direct comparability (12, 18, 24). The alteration in injectate spread in animals with an increased level of gas due to putrefaction was not assessed in this study. Therefore, further studies are needed to determine how cadaveric decomposition influences the distribution of solutions in this type of cadaveric research. Investigating the effect of postmortem decomposition on injectate distribution could represent a valuable future line of research, improving standardization and comparability with live subjects.
The LBDR and MBDR are the nerves that provide sensation and motor function to the epaxial musculature, vertebral lamina, ligaments, facet joints and skin from the midline to the axillary line (30, 31). Staining of the LBDR and MBDR is one of the most commonly observed findings after an ESP injection and may explain the clinical analgesic effect of the ESP block reported for procedures involving these structures (8, 32). In the present study, both the LBDR and MBDR were consistently stained with similar results to previous studies in dogs, and both approaches producing comparable staining (4, 18). An important finding was the greater cranial extent of MBDR and LBDR staining, and of dye spread within the ESP, relative to their caudal extent using both approaches. This is consistent with previous reports and suggests a preferential cranial distribution of the injectate within the erector spinae plane (25). These findings may have clinical relevance, as they suggest that positioning the injection caudal to the surgical site could take advantage of the preferential cranial spread observed in the present study. However, this hypothesis should be confirmed in live animals, as cadaveric dye distribution may not fully reflect local anesthetic spread under physiological conditions.
Staining of the VBSN, the nerves that provide sensation and motor function to the thoracic wall, ribs, pleura and sternum, was also demonstrated in both groups (30, 31, 33, 34). This is consistent with cadaveric studies in dogs, cats, and horses and with human reports of VBSN staining and paravertebral spread (2, 9, 12, 15, 18, 35, 36). Nevertheless, other canine and porcine studies found staining confined to the dorsal rami (4, 24). Epidural spread was also observed in both groups, with one animal presenting a more extensive spread compared to the rest of the animals. This is similar to findings in some cadaveric studies in dogs and horses (9, 18, 25), but not with others in dogs or cats (4, 12, 24, 37). Staining of the VBSN and the epidural spread identified in this study supports the potential analgesic effect of the ESP block for surgical or painful procedures involving the thorax, and is consistent with previous reports in dogs undergoing thoracic procedures following ESP block (10, 38). In addition, injections in both groups resulted in staining of the sympathetic trunk, consistent with findings from some studies in humans, dogs, and cats (8, 12, 18, 39) but not with others in dogs (4, 24, 37). Spread to the sympathetic trunk following ESP injection at T9 may contribute to the analgesic effects of the technique for thoracic visceral pain (40). However, the clinical relevance of unilateral sympathetic trunk staining following ESP block remains to be determined. Lymphatic spread, previously reported at T12 in dogs (18), was absent here, possibly reflecting anatomical differences at the injection site (22, 23). This variability of the spread, aside from the staining of the MBDR and LBDR, mirrors that reported in human studies (41–43). Moreover, it indicates that the mechanisms underlying these spread patterns remain unclear, with discrepancies likely attributable to differences in needle tip placement, injection depth, injectate volume and speed, body condition score and size, age of the animal, and inter-individual fascial anatomy across studies. Further standardized research is needed to better understand these distributions. Nevertheless, these results, together with current evidence, support the potential analgesic effect of the ESP block for the epaxial musculature, skin, and vertebrae, while coverage of the thoracic wall, thoracic cavity and viscera, sternum, and extended epidural spread remains inconsistent (10, 11, 32, 38).
In contrast to the TV approach, the LNG technique was associated with a greater risk of unintentional injection over the rib lateral to the DLETP, even when performed by different operators, altering the pattern of injectate spread. This was not reported at T12 in previous canine studies (18); however, even though the transverse process at T9 is longer than at other thoracic vertebrae, the DLETP remains comparatively small, creating a narrow target area of injection. These findings suggest that the TV approach may offer advantages over the LNG technique when performed at vertebrae with a DLETP. With the LNG approach, minor lateral or medial probe movements can displace the DLETP from the imaging field, allowing the dorsal surface of the rib or transverse process to be mistaken for it. In contrast, the TV approach provides a cross-sectional view that simultaneously displays the vertebral lamina, transverse process, DLETP, costotransverse joint, and rib tubercle. Consequently, cranial or caudal probe displacement brings the intertransverse and intercostal muscles into view, reducing the risk of misplacement. Additionally, intramuscular injection in the LNG approach produces a lenticular separation of longitudinal muscle fibers that resembles ESP hydrodissection, whereas the circumferential spread pattern seen in the TV approach more clearly indicates misplacement (19–21). Operator experience and manual dexterity may also have contributed to the misplaced injections observed with the LNG approach. Although both operators were experienced and familiar with the LNG approach, differences in individual technique could have affected needle manipulation and the accuracy of injectate deposition. LNG injections performed outside the intended target area were not the only ones associated with unexpected spread patterns. A small number of correctly performed injections in both groups showed an atypical cranial shift in MBDR staining, with the first stained nerve root located more cranially than expected relative to the injection site. In one case, this pattern was mirrored contralaterally, suggesting that individual anatomical variation, rather than technique alone, may independently influence the craniocaudal distribution of the injectate within the ESP.
Morphological evaluation of the transverse process of the thoracic vertebrae revealed marked variation in transverse process and DLETP length, as well as inconsistency in DLETP presence across vertebral levels. The DLETP was consistently absent in the most caudal thoracic vertebrae, though the specific vertebral level of absence varied between cadavers. These findings suggest that the TV approach described in this study should be applied only at vertebrae with DLETP, while the TV approach described by Herrera-Linares et al. (18) is recommended where this structure is absent.
This study is not exempt from limitations. Post-mortem changes may influence the behavior and distribution of injected dye solutions, potentially differing from patterns observed in live animals. Several physiological processes present in live animals are absent in cadavers. These include blood and lymphatic flow, negative thoracic pressure, muscle tone, thoracic movement during breathing, and local anesthetic diffusion, all of which are normally present in vivo and may contribute to the overall behavior of injected substances within tissue. As these processes are absent in a post-mortem model, the spread patterns observed in this study should be interpreted with this limitation in mind, and further studies in live animals would be valuable to confirm these findings. Another limitation was the number of anesthetists performing the injections in each group. While two anesthetists performed the injections for the LNG approach, only one carried out all injections in the TV group. Using only one operator may introduce operator-dependent bias related to skill level or familiarity with the TV technique. However, only one individual possessed the necessary training to perform this novel technique. Therefore, a second operator was avoided to reduce the risk of inaccurate needle placement and ensure consistency.
The results of this cadaveric study demonstrated that the TV and LNG US-guided ESP approaches performed over the DLETP of the 9th thoracic vertebra were feasible for administering an ESP injection at vertebrae where this anatomical structure is present. Both techniques produced comparable multisegmental spread within the ESP, with a preferential cranial nerve dye and spread distribution. However, the TV approach was associated with a lower risk of injectate misplacement outside the desired target area and subsequent undesired lateral spread, compared with the LNG approach. This study also showed that the DLETP was smaller than the transverse process and was consistently identified in vertebrae cranial to T11, although it may also be present at T11. Therefore, the TV approach described in this study should not be replicated in vertebrae without a DLETP. Clinical studies are warranted to evaluate the analgesic efficacy and safety of this technique at thoracic vertebrae with a DLETP.