Background 

Fractures of the proximal fifth metatarsal are common, and often they are classified using a three-part scale first proposed by Lawrence and Botte. A clear consensus on prognosis and treatment for these fractures is lacking, particularly for fractures in the middle classification, Zone 2; the reliability of the classification scheme itself may be partly at fault for this. The intra- and interrater reliability of the classification itself has never been established, and it remains unclear whether the three-part classification of fifth metatarsal fractures can be applied consistently enough to guide treatment.

Questions/purposes 

When used by experienced orthopaedic surgeons, (1) What is the overall interrater reliability of the three-part Lawrence and Botte classification of fifth metatarsal base fractures? (2) What is the overall intrarater reliability of the three-part Lawrence and Botte classification of fifth metatarsal base fractures? (3) What are these same metrics for the individual transitions within the classification (Zone 1-2 and Zone 2-3)?

Methods 

Thirty sets of initial presentation radiographs representing an equal number of fractures originally diagnosed by treating clinicians as Zone 1, Zone 2, and Zone 3 were evaluated and classified by three orthopaedic surgeons specializing in foot and ankle surgery and eight foot and ankle fellows to determine interrater reliability. Two weeks later, the same set of radiographs were reevaluated in random order to determine intrarater reliability. Kappa values for interrater and intrarater reliability were calculated. Additionally, the individual transitions between zones were separately analyzed by calculating kappa values for a hypothetical two-part classification based on each transition.

Results 

The three-part Lawrence and Botte classification of fifth metatarsal fractures demonstrated an overall interrater agreement of κ = 0.66 (observed agreement 77% versus chance agreement 33%). Intrarater reliability for the 11 surgeons ranged from κ = 0.60 to κ = 0.90. A two-part scheme divided by the transition between Zone 1 and Zone 2 demonstrated an interrater agreement of κ = 0.83, and a two-part scheme divided by the transition between Zone 2 and Zone 3 demonstrated a much lower interrater reliability of κ = 0.66.

Conclusion 

The three-part Lawrence and Botte classification system demonstrated a concerningly low level of interrater reliability with an observed agreement of 77% compared with a chance agreement of 33%. The primary source of concern is the assessment of the interface between Zone 2 and Zone 3, which proved much less reliable than that between Zone 1 and Zone 2. This suggests that previous studies of isolated Zone 1 fractures likely contain a homogeneous fracture cohort, whereas studies of Zone 2 or Zone 3 fractures are likely to include a mixture of fracture types. In practice, the consensus treatment of fifth metatarsal fractures differs based on whether they represent a more proximal, avulsive injury or a more distal injury from indirect trauma. Our data suggest that the Lawrence and Botte classification should be abandoned. Further work should focus on developing a new classification scheme that demonstrates improved interobserver reliability and more directly corresponds to this treatment paradigm.

Level of Evidence 

Level III, diagnostic study.

中文翻译：

---

第五跖骨基部骨折 Lawrence 和 Botte 分类系统的评估者间和评估者内可靠性是多少？

背景 

第五跖骨近端骨折很常见，通常使用劳伦斯和博特首先提出的三部分评分法对其进行分类。对于这些骨折的预后和治疗缺乏明确的共识，特别是对于中间分类（2 区）的骨折；分类方案本身的可靠性可能是造成这种情况的部分原因。分类本身的内部和组间可靠性尚未确定，并且尚不清楚第五跖骨骨折的三部分分类是否可以一致地应用以指导治疗。

问题/目的 

当经验丰富的骨科医生使用时，(1) 第五跖骨基底部骨折的三部分 Lawrence 和 Botte 分类的整体评分者间可靠性是多少？(2) 第五跖骨基底部骨折三部分 Lawrence 和 Botte 分类的总体评估者内可靠性是多少？(3) 对于分类内的各个转换（区域 1-2 和区域 2-3），这些相同的指标是什么？

方法 

由三名专门从事足部和踝关节手术的骨科医生和八名足部和踝关节研究员对 30 组初始表现 X 光片进行评估和分类，这些 X 线照片代表了临床医生最初诊断为 1 区、2 区和 3 区的相同数量的骨折，以确定可靠性。两周后，以随机顺序重新评估同一组射线照片，以确定评估者内部的可靠性。计算了评估者间和评估者内可靠性的 Kappa 值。此外，通过计算基于每个转变的假设两部分分类的 kappa 值，分别分析区域之间的各个转变。

结果 

第五跖骨骨折的三部分 Lawrence 和 Botte 分类显示，总体间一致性为 κ = 0.66（观察一致性为 77%，偶然一致性为 33%）。11 名外科医生的评分者内可靠性范围为 κ = 0.60 至 κ = 0.90。由区域 1 和区域 2 之间的过渡划分的两部分方案显示出 κ = 0.83 的人间一致性，而由区域 2 和区域 3 之间的过渡划分的两部分方案显示出低得多的人间可靠性 κ = 0.66 。

结论 

由三部分组成的 Lawrence 和 Botte 分类系统显示，观察者间的可靠性水平相当低，观察到的一致性为 77%，而机会一致性为 33%。主要关注点是对 2 区和 3 区之间界面的评估，事实证明该评估的可靠性远低于 1 区和 2 区之间的界面。这表明之前对 1 区孤立骨折的研究可能包含同质骨折群，而对 2 区或 3 区裂缝的研究可能包括多种裂缝类型的混合。在实践中，第五跖骨骨折的共识治疗方法有所不同，具体取决于第五跖骨骨折是代表更近端的撕脱性损伤还是来自间接创伤的更远端损伤。我们的数据表明应放弃劳伦斯和博特分类。进一步的工作应集中于开发一种新的分类方案，该方案证明观察者间的可靠性得到提高，并且更直接地对应于这种治疗范式。

证据水平 

III级，诊断研究。