兽研简要概述
选择合适尺寸的气管插管（ETT）对犬及其安全至关重要。导管过小可能增加气道阻力并导致气体泄漏或气道保护不充分，而导管过大可能使插管复杂化并损伤气道。临床实践中，导管尺寸常根据犬的体重（BW）选择；然而，体重相近的犬可能因品种不同而具有不同的气道尺寸。首先，作者使用计算机断层扫描比较了体重相近的五个小型犬品种之间的气管直径。接下来，作者回顾了八个品种共512只犬的记录，以了解临床实践中实际使用的导管尺寸与根据体重预测的尺寸之间的差异。第一部分的结果显示，即使在较窄的体重范围内，气管直径也因品种而异。在独立的第二部分队列中，当体重以体重的立方根表示时，体重与临床选择的气管插管尺寸之间的关系因品种而异。在短头品种中，临床选择的气管插管尺寸通常小于基于体重预测的尺寸，且差异程度因品种而异，提示在估计候选气管插管尺寸时考虑品种相关因素可能是有用的。
1. 引言
气管插管是兽医麻醉、急诊医学和重症监护中的一项基本操作[1]。适当选择气管插管（ETT）对于维持气道通畅、提供有效通气以及尽量减少气道管理相关并发症至关重要。一般而言，推荐使用能无阻力插入的最大尺寸导管[2,3]。导管过小可能增加麻醉气体泄漏和气道保护不充分的风险。根据泊肃叶定律，气流阻力随气管插管内径减小而增加；因此，即使导管尺寸小幅减小也可能增加通气阻力。导管过大可能增加插入难度以及黏膜损伤或压力相关气道损伤的风险[1,2,3,4,5,6]。
在小动物临床实践中，气管插管尺寸常通过体重（BW）、外部解剖标志或操作者经验来估计[1,7,8]。既往研究也提出了基于体重预测合适气管插管尺寸的方法[2,4,9]。尽管基于体重的方法为气管插管尺寸估计提供了实用途径，但某些品种可能需要进行调整。特别是，既往研究报道短头犬可能需要比仅根据体重预测更小的气管插管[10]。此外，若干参考文献建议准备预期尺寸的气管插管以及小一号和大一号的导管，以应对插管过程中的尺寸不匹配[1,2,4,5,10]。然而，反复插管尝试可能增加喉部创伤，包括医源性喉麻痹和伴有炎症的气管黏膜损伤[11,12]。因此，在体重之外考虑品种可能为插管前气管插管尺寸选择提供更精细的方法。
为更全面地评估品种相关因素对气道尺寸和临床气管插管选择的影响，本研究分两部分进行。第一部分旨在确定在受限的体重范围内，CT测量的气管直径是否因品种而异。第二部分纳入体重分布更广的犬，以评估品种是否在体重影响之外影响临床选择的气管插管尺寸。
作者假设：气管尺寸因品种而异；体重与临床选择的气管插管尺寸之间的关系因品种而异；临床选择的气管插管尺寸与基于体重预测的尺寸之间的差异程度因品种而异。
2. 材料与方法
2.1. 研究设计与伦理批准
本回顾性研究包括两项独立分析，使用具有不同纳入标准的独立队列。研究方案经首尔国立大学机构动物护理与使用委员会审查并批准（IACUC编号SNU-250421-3）。
2.2. 第一部分：基于CT分析气管直径的品种相关差异
第一部分回顾性审查了2020年1月至2024年12月期间在Ian动物诊断中心接受CT检查的犬的医疗记录。CT检查因与原发性呼吸系统疾病无关的临床指征而进行，最常见的是肿瘤分期/转移评估或腹腔疾病评估，并非专门为本研究而进行。纳入标准为：属于韩国临床实践中常见的五个小型品种之一（马尔济斯犬、贵宾犬、博美犬、西施犬和比熊犬），体重在4至5 kg之间，且医疗记录中采用9分制评分系统记录的体况评分（BCS）为5/9。审查医疗记录以了解呼吸系统疾病病史或临床证据，并审查CT图像以了解结构性气道异常和呼吸系统病变。记录不完整或有任何呼吸系统疾病或结构性气道异常证据的犬被排除。
麻醉管理按照常规临床实践进行，相关数据从病历中获取。根据机构方案，犬通常在麻醉前禁食至少6小时，自由饮水至诱导前约2小时。一般在前臂头静脉放置24G静脉导管（0.7×19 mm；BD Angiocath PlusTM，Becton Dickinson Medical，新加坡，新加坡）。术前用药最常见的为布托啡诺（0.2–0.3 mg kg−1；Butophan Inj.，Myung Moon，首尔，韩国）和咪达唑仑（0.2–0.3 mg kg−1；Midacum Inj.，Myung Moon，首尔，韩国）静脉注射。插管前，通过体外测量从切齿到肩点的距离来估计预期的气管导管插入深度，目标是将气管导管尖端置于胸腔入口附近。随后使用丙泊酚（4–5 mg kg−1；Anepol Inj.，Hana Pharm，首尔，韩国）诱导麻醉。犬取胸卧位，使用带套囊的气管导管（Rüschelit super safety clear，Teleflex Medical，莫里斯维尔，北卡罗来纳州，美国）进行插管。采用最小封闭容积技术进行套囊充气，确认在约2 kPa（约20 cmH2O）压力下密封良好。使用七氟烷（Sojourn Solution；Kyongbo Pharm，牙山，韩国）经氧气通过循环重呼吸系统维持麻醉，氧气流量约为1 L/min。标准术中监测包括心率、血氧饱和度、呼气末二氧化碳和示波法动脉血压。
CT成像作为常规临床实践的一部分进行，采集参数从影像记录中获取。使用CT扫描仪（Somatom go.now；Siemens AG，慕尼黑，德国）获取从鼻尖到膈肌的图像。根据制造商的推荐程序，CT扫描仪在临床成像前于每日开机时进行常规系统校准一次。CT剂量指数测量至少每6个月进行一次。在CT图像采集期间，实施手动通气，并在气道压力维持在2 kPa（约20 cmH2O）的情况下进行短暂屏气。扫描参数在各病例间存在微小差异，但基本一致，通常包括130 kVp、90 mAs、旋转时间0.8秒、1 mm层厚和螺距1。
审查CT图像以确认所有纳入犬的气管导管尖端位于胸腔入口附近。使用临床影像数据，在三个解剖水平的横断面CT图像上测量气管直径：颈段气管（C3–4）、胸腔入口（C7–T1）和胸内气管（T3–4），如图1所示。颈部测量在通过C3–C4椎间盘中点的横断面上获取。胸腔入口测量在C7–T1椎间盘中点获取，胸内气管测量在T3–T4椎间盘中点获取。在每个水平，垂直于气管腔长轴测量管腔内垂直径。
2.3. 第二部分：临床选择与基于体重预测的气管导管尺寸的临床分析
对于第二部分，回顾性审查了2023年1月至2025年12月期间在Ian动物诊断中心管理的犬的麻醉记录。扩展的临床队列包括八个品种：马尔济斯犬、贵宾犬、博美犬、西施犬、比熊犬、法国斗牛犬、北京犬和波士顿梗犬。为尽量减少体况的混杂影响，仅纳入在9分制评分系统中记录体况评分为5/9的犬。如果麻醉记录中可获得体重、体况评分和临床麻醉期间使用的气管导管尺寸，则犬符合纳入条件。病历审查确认第二部分的患者使用与第一部分相同的方案进行麻醉。每只犬仅纳入一份麻醉记录进行分析。气管导管选择由三名临床医生使用相同的通用临床方法进行。初始气管导管尺寸基于体重和常规临床评估进行估计。在短头犬中，通常选择比仅根据体重预期的更小的初始候选气管导管，这与先前的证据一致，即基于体重的估计往往高估短头犬临床使用的气管导管尺寸[10]。随后根据通过难易程度和气道密封充分性评估导管适配性。漏气试验由各临床医生使用相同的临床标准进行，气道压力约20 cmH2O时密封良好被视为可接受。如果在插入过程中遇到明显阻力，则选择较小的气管导管；如果漏气试验表明气道密封不充分，则根据临床适当情况向上调整气管导管尺寸。
第二部分队列全程使用与第一部分相同的气管导管产品。临床所选导管尺寸取自麻醉记录。对每只犬，使用 Haider 等[10]报道的公式计算基于体重的预测气管导管内径：导管尺寸（mm）= 2.407 × ln[体重（kg）] + 1.820，其中体重以千克为单位。未取整的预测值用于与临床所选导管尺寸进行主要统计学比较。为便于描述性临床解读，预测值另按最接近的0.5 mm内径增量取整，对应本研究所用导管产品可获得的尺寸增量。为评估临床所选与基于体重预测的导管尺寸之间差异的品种相关性，Haider公式应用于所有品种（包括短头品种），作为共同的基于体重的参考，而非经过验证的临床推荐。差异计算如下：差异 = 临床所选导管尺寸 − 未取整的基于体重预测的导管尺寸。
2.4. 统计分析
所有统计分析均使用SPSS 29.0版（IBM Corp.，Armonk，NY，USA）进行。使用Shapiro–Wilk检验评估数据分布，并根据数据的分布特征选择统计方法。连续变量包括年龄、体重和气管直径，以均值 ± 标准差（SD）表示。导管尺寸以中位数（四分位距）表示。在第一部分中，由两名观察者独立测量颈部、胸廓入口和胸腔内水平CT测量的气管直径。观察者间一致性使用基于双向混合效应模型绝对一致性的组内相关系数（ICC）评估。ICC值报告95%置信区间（CI）。两名观察者测量值的均值用于后续统计分析。对于CT测量气管直径的分析，使用线性混合效应模型以考虑每只犬三个解剖部位的重复测量。犬作为随机截距纳入，品种、解剖部位、性别、年龄和体重作为固定效应纳入。纳入品种×部位交互项以评估气管直径的品种相关差异是否随解剖位置而变化。当发现显著的品种×部位交互作用时，计算估计边际均值，并在每个解剖部位分别进行品种间两两比较，并校正多重比较。模型结果以估计边际均值及95%置信区间报告，p值 < 0.05认为具有统计学显著性。
在第二部分中，构建一般线性模型，以临床所选导管尺寸为因变量，品种为固定因子，体重的立方根为连续协变量。为评估体重的立方根与临床所选导管尺寸之间的关系是否因品种而异，模型中纳入品种×体重立方根交互项。由于交互作用显著，随后检查了各品种特有的体重立方根与临床所选导管尺寸之间的关系，而非假设各品种具有共同斜率。还检查了品种特有的体重分布和范围，以评估品种间的重叠程度。通过目视检查标准化残差对预测值的散点图以评估线性和方差齐性，检查残差直方图和正态Q–Q图以评估残差分布，并使用Cook距离检查强影响观测值，从而评估模型假设。
对每个品种，使用Wilcoxon符号秩检验比较临床所选导管尺寸与未取整的基于体重预测的导管尺寸。报告原始p值。为考虑八个品种特异性比较，应用Bonferroni校正的显著性阈值α = 0.00625（0.05/8）。对每个品种，配对差异另使用Hodges–Lehmann位置偏移估计进行总结，95%置信区间通过20,000次bootstrap重抽样获得。
3. 结果
3.1. 研究人群
3.1.1. 第一部分队列
第一部分共筛选1498份病历。其中，1162只犬因不符合体重标准被排除，256只因不符合体况评分标准被排除，29只因呼吸系统疾病或结构性气道异常被排除，8只因病历或CT图像不完整被排除。因此，最终分析纳入43只犬。所有犬体重均在4至5 kg之间，允许在相对狭窄的体重范围内进行品种特异性CT测量气管直径比较。基线特征总结于表1。
3.1.2. 第二部分队列
第二部分共筛选2830份麻醉记录。其中，2168份记录因不符合体况评分标准被排除，150份记录因麻醉记录不完整被排除。因此，最终分析纳入512只 unique 犬。与第一部分不同，第二部分未限制于狭窄的体重范围，因为临床分析旨在评估常规实践中更广泛体重分布下的导管尺寸选择。所有犬体况评分均为5/9。基线特征总结于表2。
3.2. 第一部分：基于CT的TD品种相关差异分析
在线性混合效应模型中，品种与TD显著相关（F(4, 33) = 4.927，p = 0.003），而单独解剖部位不显著（F(2, 76) = 1.007，p = 0.370）。发现显著的品种与部位交互作用（F(8, 76) = 3.705，p = 0.001），表明TD的品种相关差异随解剖位置而变化（图2）。在模型中校正后，年龄（p = 0.105）、体重（p = 0.764）和性别（p = 0.924）与TD无显著关联。观察者间可靠性采用基于双向混合效应模型和绝对一致性定义的平均测量ICC进行评估。ICC值分别为：颈部TD 0.966（95% CI，0.314–0.991），胸廓入口TD 0.831（95% CI，−0.155–0.954），胸腔内TD 0.752（95% CI，−0.186–0.930）。在显著的品种与部位交互作用后进行的部位特异性两两比较显示，在颈部水平，马尔济斯犬、比熊犬和贵宾犬的TD显著大于博美犬（校正p < 0.001、p = 0.007和p = 0.043）。在胸腔内水平，贵宾犬的TD显著大于西施犬（校正p = 0.031）。多重比较校正后，胸廓入口处未发现显著的品种差异。
3.3. 第二部分：体重校正后品种对临床选择ETT尺寸的影响
包含品种、体重立方根及品种与体重立方根交互作用的一般线性模型显著（F(15, 496) = 164.157，p < 0.001），解释了临床选择ETT尺寸变异的83.2%（R2 = 0.832；校正R2 = 0.827）。品种（F(7, 496) = 5.503，p < 0.001）、体重立方根（F(1, 496) = 251.549，p < 0.001）及品种与体重立方根交互作用（F(7, 496) = 6.796，p < 0.001）均显著。显著的交互作用表明，体重立方根与临床选择ETT尺寸之间的关系因品种而异。各品种体重立方根与临床选择ETT尺寸的斜率分别为：马尔济斯犬1.990 mm/kg1/3（95% CI，1.661–2.319），贵宾犬3.049 mm/kg1/3（95% CI，2.713–3.385），博美犬2.316 mm/kg1/3（95% CI，1.988–2.645），西施犬4.216 mm/kg1/3（95% CI，3.415–5.016），比熊犬3.253 mm/kg1/3（95% CI，2.948–3.557），法国斗牛犬2.643 mm/kg1/3（95% CI，1.620–3.665），北京犬2.257 mm/kg1/3（95% CI，0.301–4.212），波士顿梗犬1.233 mm/kg1/3（95% CI，−1.243至3.709）。残差图的目视检查未显示线性或方差齐性的实质性违反。残差分布近似正态，Q–Q图尾部有轻度偏离，未发现高影响力观测值（最大Cook距离 = 0.39）。
3.4. 第二部分：临床选择与基于体重预测的ETT尺寸之间的比较与差异
差异的幅度在各品种间差异很大。在马尔济斯犬、贵宾犬、博美犬和比熊犬中，临床选择的ETT尺寸与基于体重预测值最接近的商业可用0.5 mm ETT尺寸匹配的比例为41.0–49.0%。在西施犬中，完全匹配率较低，为25.5%，中位差异为−0.40 mm。法国斗牛犬中仅2.0%观察到完全匹配，北京犬为0%，波士顿梗犬为12.5%，其中位差异约为−1.0 mm。差异≥1.0 mm在马尔济斯犬、贵宾犬、博美犬和比熊犬中占8.2–14.3%，西施犬中占17.0%，而法国斗牛犬为85.7%，北京犬和波士顿梗犬均为75.0%。效应量在马尔济斯犬、贵宾犬、博美犬和比熊犬中较小，但在西施犬、法国斗牛犬、北京犬和波士顿梗犬中较大（表3）。这些发现表明差异存在显著的品种相关变异，西施犬的负差异不如法国斗牛犬、北京犬和波士顿梗犬明显。
统计分析中，差异计算为临床选择ETT尺寸减去未取整的基于体重预测的ETT尺寸。负值表示临床选择ETT尺寸小于预测值，而正值表示大于预测值。各品种差异值的分布见图3。
4. 讨论
本研究包括两项相关分析，旨在评估品种相关因素对犬ETT尺寸选择的影响。在第一部分中，基于CT的气管测量表明，即使在相似的体重范围内，TD在各品种间也存在显著差异。这一发现提示气道尺寸可能不能仅由体重完全解释，品种相关的解剖变异可能导致气管尺寸的差异。在颈部、胸廓入口和胸腔内水平观察到的差异进一步表明，品种相关的气道变异沿气管走行延伸，而非局限于单一节段。基于CT的评估特别有用，因为它允许在解剖学定义的层面测量TD而不受重叠影响[13,14]，从而为品种间比较提供客观的解剖学信息。
在第二部分中，品种、体重（BW）的立方根以及品种与体重立方根的交互作用与临床选择的ETT尺寸显著相关，表明体重与临床选择ETT尺寸之间的关系因品种而异。由于先前已有报道指出线性气道尺寸随体重立方根缩放[15,16]，且体重立方根也已被评估为犬ETT尺寸的预测因子[17]，本研究采用体重立方根代表体型缩放。品种与体重立方根交互作用的显著性持续存在，提示临床选择ETT尺寸的品种相关差异不能仅由体型缩放来解释。特别是，在短头品种中观察到与基于体重预测的显著偏差，这与先前报道一致[10,18,19]。然而，先前的研究未确定这种偏差的程度在个别短头品种之间是否存在差异。在本研究中，西施犬显示出相对较小的负偏差，而法国斗牛犬、北京犬和波士顿梗犬显示出明显更大的负偏差。这些发现提示，就体重与临床选择ETT尺寸之间的关系而言，短头品种可能并不代表一个同质性群体。这种区别可能具有临床相关性，因为短头犬由于其特征性气道解剖结构而特别易发生困难气管插管。先前报道描述了短头犬插管时阻力更大、插管尝试次数增加以及插管时间延长[20,21]。插管过程中更大的阻力和反复的气道操作可能增加气道损伤的风险，而气道建立延迟可能促进低氧血症的发生或加重。因此，在首次插管尝试时选择合适的ETT尺寸对短头犬可能尤为重要，以尽量减少反复气道操作并促进及时建立气道。本研究发现表明，临床选择与基于体重预测的ETT尺寸之间的偏差在短头品种之间可能存在差异。尽管需要进一步评估临床结局的研究，但这些发现可能支持在完善短头犬ETT尺寸选择标准时进一步考虑品种特异性因素。
本研究存在若干局限性。第一，回顾性设计可能引入了与临床医生判断、既往经验、品种相关预期及置入难易度相关的变异性。尽管ETT选择由三位临床医生采用相同的总体临床方法进行，且导管适配性通过漏气试验评估，但临床医生通常对短头犬考虑较小的初始候选ETT。因此，预先存在的基于品种信息的临床选择实践可能导致了第二部分中观察到的品种相关差异。此外，解剖适配性、置入阻力、插管后气道损伤及其他临床结局的客观指标未能系统获取。因此，临床选择的ETT尺寸不一定代表解剖上最优的导管尺寸。第二，第一部分和第二部分基于纳入标准不同的独立队列。因此，第一部分获得的基于CT的气管测量值无法直接匹配或解释为第二部分观察到的临床选择ETT尺寸的直接解剖学解释。第三，由于CT测量是在气管插管后获得的，ETT管体和充气套囊可能影响气管几何形态，尤其是在颈段和胸廓入口水平。此外，仅气管直径并不能决定临床可置入的最大ETT尺寸，因为通过喉部可能代表额外的解剖限制。再者，本研究使用的CT图像以1 mm层厚采集，面内像素大小约为0.5–0.7 mm，扫描仪的几何精度容差约为±0.5至±1.0 mm。因此，用于统计分析的代表CT衍生TD值不应被解释为以与所报告小数位数相对应的精度代表绝对解剖气管直径。第四，基于ICC点估计，观察者间信度总体较高，但宽置信区间表明这些估计值应谨慎解读。最后，尽管对临床选择与基于体重预测的ETT尺寸之间偏差的品种相关差异进行了统计学评估，但这些发现应在前瞻性研究中以更大且品种分布更均衡的群体加以证实。
尽管存在这些局限性，本研究提供了补充性的解剖学和临床证据，表明品种相关变异可能与超越体重 alone 的 ETT 尺寸估算相关。品种相关效应在短头犬中似乎更为明显。然而，在这些品种中观察到的异质性表明，它们可能不构成用于 ETT 尺寸估算的同质群体。需要进一步研究以阐明短头犬临床选择的 ETT 尺寸中品种相关差异与短头构型相关的程度，并确定有助于表征这些差异的客观指标。尽管如此，本研究结果表明，在完善 ETT 尺寸选择方法时，品种相关因素，特别是与短头品种相关的因素，可能值得在 BW 之外加以考虑。
5. 结论
在本研究中，尽管第1部分和第2部分基于不同的队列，但在气道尺寸和临床选择的 ETT 尺寸中均发现了品种相关效应。与既往报告一致，这些效应在短头犬中更为明显，本研究进一步证明品种相关差异的程度在个别短头品种之间存在差异。由于在首次插管尝试中选择合适的 ETT 对短头犬可能尤为重要，这些发现表明在完善 ETT 尺寸选择标准时，品种相关因素可能值得考虑。
查看英文原文▾
Simple Summary
Selecting the correct endotracheal tube (ETT) size is critical for dogs and their safety. A tube that is too small may increase airflow resistance and cause gas leakage or inadequate airway protection, whereas a tube that is too large may complicate intubation and injure the airway. In clinical practice, tube size is often chosen based on the dog’s body weight (BW); however, dogs of similar weight may have different airway dimensions depending on breed. First, the authors used computed tomography scans to compare tracheal diameters among five small dog breeds with similar BW. Next, the authors reviewed records from 512 dogs across eight breeds to see what tube sizes were clinically used in practice compared with the sizes predicted from BW. Results from Part 1 showed that tracheal diameter varied among breeds even within a narrow BW range. In the separate Part 2 cohort, the relationship between body mass and clinically selected ETT size differed among breeds when body mass was expressed using the cube root of BW. In brachycephalic breeds, clinically selected ETT sizes were generally smaller than BW-based predicted sizes, and the magnitude of the discrepancy varied among breeds, suggesting that breed-related considerations may be useful when estimating candidate ETT sizes.
1. Introduction
Endotracheal intubation is a fundamental procedure in veterinary anesthesia, emergency medicine, and critical care [1]. Appropriate endotracheal tube (ETT) selection is essential for maintaining airway patency, providing effective ventilation, and minimizing complications associated with airway management. In general, the largest tube that can be inserted without resistance is recommended [2,3]. ETTs that are too small may increase anesthetic gas leakage and the risk of inadequate airway protection. According to Poiseuille’s law, airflow resistance increases as ETT internal diameter decreases; therefore, even small reductions in tube size may increase resistance to ventilation. ETTs that are too large may increase insertion difficulty and the risk of mucosal or pressure-related airway injury [1,2,3,4,5,6].
In small animal practice, ETT size is often estimated using body weight (BW), external anatomical landmarks, or operator experience [1,7,8]. Previous studies have also proposed BW-based methods for predicting appropriate ETT size [2,4,9]. Although BW-based methods provide a practical approach to ETT size estimation, adjustment may be needed in some breeds. In particular, previous studies have reported that brachycephalic dogs may require smaller ETTs than those predicted from BW alone [10]. In addition, several references recommend preparing ETTs in the expected size as well as one size smaller and larger in anticipation of mismatch during intubation [1,2,4,5,10]. However, repeated intubation attempts may increase laryngeal trauma, including iatrogenic laryngeal paralysis and tracheal mucosal injury with inflammation [11,12]. Therefore, considering breed in addition to BW may provide a more refined approach to ETT size selection before intubation.
To more comprehensively evaluate breed-related influences on airway size and clinical ETT selection, the present study was conducted in two parts. Part 1 was designed to determine whether CT-measured tracheal diameter differed among breeds even within a restricted body-weight range. Part 2 included dogs across a broader body-weight distribution to evaluate whether breed influenced clinically selected ETT size beyond the effect of body weight.
The authors hypothesized that tracheal size would differ among breeds, that the relationship between body mass and clinically selected ETT size would vary by breed, and that the magnitude of discrepancy between clinically selected and BW-based predicted ETT sizes would differ among breeds.
2. Materials and Methods
2.1. Study Design and Ethical Approval
This retrospective study comprised two independent analyses using separate cohorts with different inclusion criteria. The study protocol was reviewed and approved by the Institutional Animal Care and Use Committee of Seoul National University (IACUC No. SNU-250421-3).
2.2. Part 1: CT-Based Analysis of Breed-Related Differences in Tracheal Diameter
For Part 1, medical records of dogs that underwent CT examination at the Ian Animal Diagnostic Center between January 2020 and December 2024 were retrospectively reviewed. CT examinations were performed for clinical indications unrelated to primary respiratory disease, most commonly for tumor staging/metastatic evaluation or assessment of intra-abdominal disease, and were not performed specifically for this study. Dogs were included if they belonged to one of five small breeds commonly encountered in Korean clinical practice (Maltese, Poodle, Pomeranian, Shih Tzu, and Bichon Frise), weighed between 4 and 5 kg, and had a body condition score (BCS) of 5/9, as recorded in the medical record using the 9-point scoring system. Medical records were reviewed for a history or clinical evidence of respiratory disease, and CT images were reviewed for structural airway abnormalities and respiratory pathology. Dogs with incomplete records or any evidence of respiratory disease or structural airway abnormality were excluded.
Anesthetic management was performed according to routine clinical practice, and relevant data were obtained from the medical records. Dogs were typically fasted for at least 6 h before anesthesia, with free access to water until approximately 2 h before induction, according to institutional protocols. A 24-gauge intravenous catheter (0.7 × 19 mm; BD Angiocath PlusTM, Becton Dickinson Medical, Singapore, Singapore) was generally placed in the cephalic vein. Premedication most commonly consisted of butorphanol (0.2–0.3 mg kg−1; Butophan Inj., Myung Moon, Seoul, Republic of Korea) and midazolam (0.2–0.3 mg kg−1; Midacum Inj., Myung Moon, Seoul, Republic of Korea) administered intravenously. Before intubation, the intended ETT insertion depth was estimated by externally measuring the distance from the incisors to the point of the shoulder, with the aim of positioning the ETT tip near the thoracic inlet. Anesthesia was then induced with propofol (4–5 mg kg−1; Anepol Inj., Hana Pharm, Seoul, Republic of Korea). Dogs were intubated in sternal recumbency using cuffed endotracheal tubes (Rüschelit super safety clear, Teleflex Medical, Morrisville, NC, USA). Cuff inflation was performed using the minimum occlusive volume technique, with confirmation of an adequate seal at approximately 2 kPa (approximately 20 cmH2O). Anesthesia was maintained with sevoflurane (Sojourn Solution; Kyongbo Pharm, Asan, Republic of Korea) delivered in oxygen using a circle rebreathing system, with an oxygen flow rate of approximately 1 L/min. Standard intraoperative monitoring included heart rate, oxygen saturation, end-tidal carbon dioxide, and oscillometric arterial pressure.
CT imaging was performed as part of routine clinical practice, and acquisition parameters were retrieved from imaging records. Images were obtained from the tip of the nose to the diaphragm using a CT scanner (Somatom go.now; Siemens AG, Munich, Germany). The CT scanner underwent routine system calibration once daily at scanner startup before clinical imaging, in accordance with the manufacturer’s recommended procedure. CT dose index measurements were performed at least once every 6 months. During CT image acquisition, manual ventilation was applied, and short breath-hold periods were performed with airway pressure maintained at 2 kPa (approximately 20 cmH2O). Scan parameters showed minor variation among cases but were largely consistent, typically including 130 kVp, 90 mAs, a rotation time of 0.8 s, 1 mm slice thickness, and a helical pitch of 1.
The CT images were reviewed to confirm that the ETT tip was positioned near the thoracic inlet in all included dogs. Using clinical imaging data, TD was measured on transverse CT images at three anatomical levels: the cervical trachea (C3–4), thoracic inlet (C7–T1), and intrathoracic trachea (T3–4), as shown in Figure 1. The cervical measurement was obtained at the transverse plane through the midpoint of the C3–C4 intervertebral disc. The thoracic inlet measurement was obtained at the midpoint of the C7–T1 intervertebral disc, and the intrathoracic tracheal measurement was obtained at the midpoint of the T3–T4 intervertebral disc. At each level, the internal vertical luminal diameter was measured perpendicular to the long axis of the tracheal lumen.
2.3. Part 2: Clinical Analysis of Clinically Selected and BW-Based Predicted ETT Size
For Part 2, anesthetic records of dogs managed at the Ian Animal Diagnostic Center between January 2023 and December 2025 were retrospectively reviewed. The expanded clinical cohort comprised eight breeds: Maltese, Poodle, Pomeranian, Shih Tzu, Bichon Frise, French Bulldog, Pekingese, and Boston Terrier. To minimize the confounding effects of body condition, only dogs with a recorded BCS of 5/9 on the 9-point scoring system were included. Dogs were eligible if BW, BCS, and the ETT size used during clinical anesthesia were available in the anesthetic records. A review of the medical records confirmed that the patients in Part 2 underwent anesthesia using the same protocol as those in Part 1. Only one anesthetic record per dog was included in the analysis. ETT selection was performed by three clinicians using the same general clinical approach. The initial ETT size was estimated based on BW and routine clinical assessment. In brachycephalic dogs, a smaller initial candidate ETT was typically selected than would be expected from BW alone, consistent with previous evidence that BW-based estimates tend to overestimate clinically used ETT size in brachycephalic dogs [10]. Tube fit was subsequently evaluated according to ease of passage and adequacy of airway sealing. The leak test was performed using the same clinical criterion across clinicians, and an adequate airway seal at an airway pressure of approximately 20 cmH2O was considered acceptable. If marked resistance was encountered during insertion, a smaller ETT was selected, whereas if the leak test indicated an inadequate airway seal, the ETT size was adjusted upward as clinically appropriate.
The same ETT product as that used in Part 1 was used throughout Part 2 cohort. Clinically selected tube size was obtained from the anesthetic records. For each dog, the BW-based predicted ETT internal diameter was calculated using the equation reported by Haider et al. [10]: ETT size (mm) = 2.407 × ln[BW (kg)] + 1.820, where BW is body weight in kilograms. The unrounded predicted value was used for the primary statistical comparison with clinically selected ETT size. For descriptive clinical interpretation, the predicted value was additionally rounded to the nearest 0.5 mm internal-diameter increment, corresponding to the available size increments of the ETT product used in this study. To evaluate breed-related differences in the discrepancy between clinically selected and BW-based predicted ETT sizes, the Haider equation was applied to all breeds, including brachycephalic breeds, as a common BW-based reference rather than as a validated clinical recommendation. The discrepancy was calculated as follows:Discrepancy = clinically selected ETT size − unrounded BW-based predicted ETT size.
2.4. Statistical Analyses
All statistical analyses were performed using SPSS version 29.0 (IBM Corp., Armonk, NY, USA). Data distribution was assessed using the Shapiro–Wilk test, and statistical methods were selected according to distributional characteristics of the data. Continuous variables including age, body weight, and TD are presented as mean ± standard deviation (SD). ETT size is presented as median (interquartile range). In Part 1, CT-measured TDs at the cervical, thoracic inlet, and intrathoracic levels were measured independently by two observers. Interobserver agreement was assessed using the intraclass correlation coefficient (ICC) based on a two-way mixed-effects model for absolute agreement. ICC values were reported with 95% confidence intervals (CIs). The mean value of the two observers’ measurements was used for subsequent statistical analyses. For the analysis of CT-measured TD, a linear mixed-effects model was used to account for repeated measurements obtained from the three anatomical sites within each dog. Dog was included as a random intercept, while breed, anatomical site, sex, age, and BW were included as fixed effects. A breed-by-site interaction term was included to evaluate whether breed-related differences in TD varied according to anatomical location. When a significant breed-by-site interaction was identified, estimated marginal means were calculated and pairwise breed comparisons were performed separately at each anatomical site with adjustment for multiple comparisons. Model results are reported as estimated marginal means with 95% confidence intervals, and a p-value < 0.05 was considered statistically significant.
In Part 2, a general linear model was constructed with clinically selected ETT size as the dependent variable, breed as a fixed factor, and the cube root of BW as a continuous covariate. To evaluate whether the relationship between the cube root of BW and clinically selected ETT size differed among breeds, a breed-by-cube root of BW interaction term was included in the model. Because the interaction was significant, breed-specific relationships between the cube root of BW and clinically selected ETT size were subsequently examined rather than assuming a common slope across breeds. Breed-specific BW distributions and ranges were also examined to assess the degree of overlap among breeds. Model assumptions were evaluated by visual inspection of standardized residuals plotted against predicted values for linearity and homoscedasticity, inspection of residual histograms and normal Q–Q plots for residual distribution, and examination of influential observations using Cook’s distance.
For each breed, clinically selected ETT size was compared with the unrounded BW-based predicted ETT size using the Wilcoxon signed-rank test. The raw p-values are reported. To account for the eight breed-specific comparisons, a Bonferroni-adjusted significance threshold of α = 0.00625 (0.05/8) was applied. For each breed, the paired discrepancy was additionally summarized using the Hodges–Lehmann estimate of location shift, with 95% confidence intervals obtained from 20,000 bootstrap resamples.
3. Results
3.1. Study Population
3.1.1. Part 1 Cohort
A total of 1498 medical records were screened for Part 1. Of these, 1162 dogs were excluded because they did not meet the BW criterion, 256 because they did not meet the BCS criterion, 29 because of respiratory disease or structural airway abnormalities, and 8 because of incomplete medical records or CT images. Consequently, 43 dogs were included in the final analysis. All dogs weighed between 4 and 5 kg, allowing breed-specific comparison of CT-measured tracheal diameters within a relatively narrow BW range. Baseline characteristics are summarized in Table 1.
3.1.2. Part 2 Cohort
A total of 2830 anesthetic records were screened for Part 2. Of these, 2168 records were excluded because they did not meet the BCS criterion, and 150 records were excluded because of incomplete anesthetic records. The final analysis therefore included 512 unique dogs. Unlike Part 1, Part 2 was not restricted to a narrow BW range, because the clinical analysis was intended to evaluate ETT size selection across the broader BW distribution encountered in routine practice. All dogs had a BCS of 5/9. Baseline characteristics are summarized in Table 2.
3.2. Part 1: CT-Based Analysis of Breed-Related Differences in TD
In the linear mixed-effects model, breed was significantly associated with TD (F(4, 33) = 4.927, p = 0.003), whereas anatomical site alone was not significant (F(2, 76) = 1.007, p = 0.370). A significant breed-by-site interaction was identified (F(8, 76) = 3.705, p = 0.001), indicating that breed-related differences in TD varied according to anatomical location (Figure 2). Age (p = 0.105), BW (p = 0.764), and sex (p = 0.924) were not significantly associated with TD after adjustment in the model. Interobserver reliability was assessed using average-measures ICCs based on a two-way mixed-effects model with an absolute-agreement definition. The ICC values were 0.966 (95% CI, 0.314–0.991) for cervical TD, 0.831 (95% CI, −0.155–0.954) for thoracic inlet TD, and 0.752 (95% CI, −0.186–0.930) for intrathoracic TD. Site-specific pairwise comparisons following the significant breed-by-site interaction showed that, at the cervical level, TD was significantly greater in Maltese, Bichon Frise, and Poodle dogs than in Pomeranians (adjusted p < 0.001, p = 0.007, and p = 0.043, respectively). At the intrathoracic level, Poodles had a significantly greater TD than Shih Tzus (adjusted p = 0.031). No significant breed differences were identified at the thoracic inlet after adjustment for multiple comparisons.
3.3. Part 2: Effect of Breed on Clinically Selected ETT Size After Adjustment for BW
The general linear model including breed, the cube root of BW, and the breed-by-cube root of BW interaction was significant (F(15, 496) = 164.157, p < 0.001) and explained 83.2% of the variance in clinically selected ETT size (R2 = 0.832; adjusted R2 = 0.827). Breed (F(7, 496) = 5.503, p < 0.001), the cube root of BW (F(1, 496) = 251.549, p < 0.001), and the breed-by-cube root of BW interaction (F(7, 496) = 6.796, p < 0.001) were significant. The significant interaction indicated that the relationship between the cube root of BW and clinically selected ETT size differed among breeds. Breed-specific slopes relating the cube root of BW to clinically selected ETT size were 1.990 mm/kg1/3 (95% CI, 1.661–2.319) in Maltese, 3.049 mm/kg1/3 (95% CI, 2.713–3.385) in Poodles, 2.316 mm/kg1/3 (95% CI, 1.988–2.645) in Pomeranians, 4.216 mm/kg1/3 (95% CI, 3.415–5.016) in Shih Tzus, 3.253 mm/kg1/3 (95% CI, 2.948–3.557) in Bichon Frises, 2.643 mm/kg1/3 (95% CI, 1.620–3.665) in French Bulldogs, 2.257 mm/kg1/3 (95% CI, 0.301–4.212) in Pekingese, and 1.233 mm/kg1/3 (95% CI, −1.243 to 3.709) in Boston Terriers.Visual inspection of residual plots did not indicate substantial violations of linearity or homoscedasticity. The residual distribution was approximately normal, with mild deviations at the tails of the Q–Q plot, and no highly influential observations were identified (maximum Cook’s distance = 0.39).
3.4. Part 2: Comparison and Discrepancy Between Clinically Selected and BW-Based Predicted ETT Size
The magnitude of discrepancy varied substantially among breeds. The clinically selected ETT size matched the nearest commercially available 0.5 mm ETT size to the BW-based predicted value in 41.0–49.0% of Maltese, Poodle, Pomeranian, and Bichon Frise dogs. In Shih Tzus, exact match was lower at 25.5%, with a median discrepancy of −0.40 mm. Exact match was observed in only 2.0% of French Bulldogs, 0% of Pekingese, and 12.5% of Boston Terriers, in which the median discrepancies were approximately −1.0 mm. Differences of ≥1.0 mm occurred in 8.2–14.3% of Maltese, Poodle, Pomeranian, and Bichon Frise dogs and in 17.0% of Shih Tzus, compared with 85.7% of French Bulldogs and 75.0% of both Pekingese and Boston Terriers. Effect sizes were small in Maltese, Poodle, Pomeranian, and Bichon Frise dogs, but large in Shih Tzus, French Bulldogs, Pekingese, and Boston Terriers (Table 3). These findings indicate marked breed-related variation in the discrepancy, with Shih Tzus showing a less pronounced negative discrepancy than French Bulldogs, Pekingese, and Boston Terriers.
For statistical analyses, discrepancy was calculated as clinically selected ETT size minus the unrounded BW-based predicted ETT size. Negative values indicate that the clinically selected ETT size was smaller than the predicted value, whereas positive values indicate that it was larger. The distribution of these discrepancy values according to breed is shown in Figure 3.
4. Discussion
The present study comprised two related analyses designed to evaluate breed-related influences on ETT size selection in dogs. In Part 1, CT-based tracheal measurements demonstrated that TD differed significantly among breeds even within a similar BW range. This finding suggests that airway size may not be fully explained by BW alone and that breed-related anatomical variation may contribute to differences in tracheal dimensions. Differences observed at the cervical, thoracic inlet, and intrathoracic levels further suggest that breed-related airway variation extends along the tracheal course rather than being confined to a single segment. CT-based assessment was particularly useful because it allowed TD to be measured at anatomically defined levels without superimposition [13,14], thereby providing objective anatomical information for comparison across breeds.
In Part 2, breed, the cube root of BW, and the breed-by-cube root of BW interaction were significantly associated with clinically selected ETT size, indicating that the relationship between body mass and clinically selected ETT size varied among breeds. Because linear airway dimensions have previously been reported to scale with the cube root of BW [15,16], and the cube root of BW has also been evaluated as a predictor of ETT size in dogs [17], the cube root of BW was used to represent body-size scaling in the present study. The persistence of a significant breed-by-cube root of BW interaction suggests that breed-related differences in clinically selected ETT size were not explained solely by body-size scaling. In particular, significant discrepancies from BW-based predictions were observed in brachycephalic breeds, consistent with previous reports [10,18,19]. However, previous studies did not determine whether the magnitude of this discrepancy differed among individual brachycephalic breeds. In the present study, Shih Tzus showed a relatively small negative discrepancy, whereas French Bulldogs, Pekingese, and Boston Terriers showed substantially larger negative discrepancies. These findings suggest that brachycephalic breeds may not represent a homogeneous group with respect to the relationship between body weight and clinically selected ETT size. This distinction may be clinically relevant because brachycephalic dogs are particularly susceptible to difficult tracheal intubation owing to their characteristic airway anatomy. Previous reports have described greater resistance during tracheal intubation, as well as increased numbers of intubation attempts and longer intubation times, in brachycephalic dogs [20,21]. Greater resistance and repeated airway manipulation during intubation may increase the risk of airway injury, whereas delayed airway establishment may contribute to the development or worsening of hypoxemia. Therefore, selecting an appropriately sized ETT on the first intubation attempt may be particularly important in brachycephalic dogs to minimize repeated airway manipulation and facilitate timely airway establishment. The present findings indicate that the discrepancy between clinically selected and BW-based predicted ETT sizes may differ among brachycephalic breeds. Although further studies evaluating clinical outcomes are needed, these findings may support further consideration of breed-specific factors when refining ETT size-selection criteria for brachycephalic dogs.
This study has several limitations. First, the retrospective design may have introduced variability related to clinician judgment, prior experience, breed-related expectations, and ease of insertion. Although ETT selection was performed by three clinicians using the same general clinical approach and tube fit was assessed using a leak test, clinicians typically considered a smaller initial candidate ETT in brachycephalic dogs. Therefore, pre-existing breed-informed clinical selection practices may have contributed to the observed breed-related differences in Part 2. In addition, objective measures of anatomical fit, insertion resistance, post-intubation airway injury, and other clinical outcomes were not systematically available. Therefore, clinically selected ETT sizes may not necessarily represent anatomically optimal tube sizes. Second, Parts 1 and 2 were based on independent cohorts with different inclusion criteria. Therefore, CT-based tracheal measurements obtained in Part 1 cannot be directly matched to, or interpreted as a direct anatomical explanation for, the clinically selected ETT sizes observed in Part 2. Third, because CT measurements were obtained after endotracheal intubation, the ETT shaft and inflated cuff could have influenced tracheal geometry, particularly at the cervical and thoracic inlet levels. In addition, tracheal diameter alone does not determine the maximum ETT size that can be inserted clinically, because passage through the larynx may represent an additional anatomical limitation. Furthermore, the CT images used in this study were acquired at a 1 mm slice thickness with an in-plane pixel size of approximately 0.5–0.7 mm, and the geometric accuracy tolerance of the scanner was approximately ±0.5 to ±1.0 mm. Hence, the CT-derived TD values used for statistical analysis should not be interpreted as representing the absolute anatomical tracheal diameter with a level of precision corresponding to the reported decimal places. Fourth, interobserver reliability was generally high based on the ICC point estimates, although the wide confidence intervals indicate that these estimates should be interpreted cautiously. Finally, although breed-related differences in the discrepancy between clinically selected and BW-based predicted ETT size were statistically evaluated, these findings should be confirmed in prospective studies with larger and more balanced breed groups.
Despite these limitations, the present study provides complementary anatomical and clinical evidence that breed-related variation may be relevant to ETT size estimation beyond body weight alone. Breed-related effects appeared to be more evident among brachycephalic dogs. However, the heterogeneity observed among these breeds suggests that they may not constitute a homogeneous group for ETT size estimation. Further studies are needed to clarify the extent to which breed-related differences in clinically selected ETT size among brachycephalic dogs are associated with brachycephalic conformation and to identify objective measures that may help characterize these differences. Nevertheless, the present findings suggest that breed-related factors, particularly those associated with brachycephalic breeds, may warrant consideration in addition to BW when refining ETT size-selection approaches.
5. Conclusions
In the present study, although Parts 1 and 2 were based on separate cohorts, breed-related effects were identified in both airway dimensions and clinically selected ETT size. Consistent with previous reports, these effects were more evident in brachycephalic dogs, with the present study further demonstrating that the magnitude of breed-related differences varied among individual brachycephalic breeds. Because selecting an appropriate ETT on the first intubation attempt may be particularly important in brachycephalic dogs, these findings suggest that breed-related factors may warrant consideration when refining ETT size-selection criteria.