Inter-surgeon variability in the identification of clock face landmarks when placing suture anchors in arthroscopic Bankart repair

Abstract

Background

The accuracy of surgeons in utilizing the clock face method for anchor placement has never been investigated. Our hypothesis was that shoulder arthroscopy surgeons would be able to place suture anchors at predetermined positions with accuracy and reliability.

Methods

Ten cadaveric shoulders were used. Five fellowship-trained shoulder arthroscopy surgeons were directed to place a suture anchor at 3:30, 4:30, and 5:30 clock in two shoulders each. The position of the anchors was determined with computed tomography. The accuracy of placement was calculated and data analyzed with one-way analysis of variance. The intraclass correlation coefficients were calculated.

Results

The overall accuracy was 57%. The accuracy of anchor placement at the 3:30 position was 40% (average position 2:24 o’clock), it was 50% at the 4:30 position (average position 3:42 o’clock) and 80% at the 5:30 position (average position 5:03 o’clock). No statistical difference in accuracy between the placement of the superior, middle, and inferior anchors (p = 0.145) was seen. The intraclass correlation coefficient for inter-surgeon reliability was 0.4 (fair) while the intraclass correlation coefficient for intra-surgeon reliability was 0.6 (moderate).

Discussion

The findings of this study suggest a moderate degree of accuracy and fair to moderate inter- and intra-surgeon reliability when using the clock face system to guide anchor placement.

Introduction

Conventionally, shoulder arthroscopy surgeons describe the glenoid fossa as the face of a clock, with the biceps anchor considered the landmark for the 12 o’clock position.^1,2 An imaginary longitudinal axis passing through the biceps anchor and the most inferior margin of the glenoid defines the 12 and the 6 o’clock position, while an antero-posterior line at the widest width of the glenoid defines the 3 and 9 o’clock position.² These landmarks are routinely used in arthroscopic labral repair to describe the extent of labral detachment and to plan the position of anchor insertion on the glenoid face.

The arthroscopic Bankart repair with suture anchors is a reliable technique in selected patients, with failure rates comparable to open repair.^3–5 Some authors advocate that this repair includes the use of a minimum of three suture anchors placed at 3:30, 4:30, and 5:30 o’clock position.^6–9 Such anchor placement on the glenoid rim is thought to enable an anatomic reduction and fixation of the labrum.^9–11 To our knowledge, there are no studies evaluating the accuracy of surgeons in utilizing the clock face method for anchor placement.

This study was conducted with the following aims: (1) to evaluate the accuracy of suture anchor placement at the 3:30, 4:30, and 5:30 o’clock position on the anterior glenoid and (2) to evaluate inter- and intra-surgeon reliability of anchor placement. Our hypothesis was that fellowship trained shoulder arthroscopy surgeons would be able to place suture anchors at predetermined locations on the glenoid face, with accuracy and reliability.

Materials and methods

Shoulder arthroscopy

Ten fresh-frozen cadaveric shoulders (five right and five left) were used, with specimens mounted on custom-made lab trays simulating a beach chair position. The anterior labrum was arthroscopically detached in each shoulder using a shaver and arthroscopic elevator. Five fellowship-trained shoulder arthroscopy surgeons were directed to place three suture anchors (2.3 mm Bioraptor, Smith & Nephew, Andover, Massachusetts) at the 3:30, 4:30, and 5:30 clock position in two shoulders each (one right and one left shoulder), using an anterior-inferior portal established with an outside-in technique (Figure 1). Each surgeon in the study performs a minimum of 50 shoulder arthroscopies per year, with each having performed a minimum of 100 arthroscopic labral repairs.

Figure 1.

Arthroscopic view from the posterior portal showing anchor placement at 3:30, 4:30, 5:30 o’clock position on the glenoid face.

Computed tomography analysis

Following anchor insertion, each shoulder was dissected, and the glenoid removed using an oscillating saw (Figure 2). A computed tomography (CT) scan of each glenoid was performed using 1 mm slices and three-dimensional reconstruction. The center of the glenoid was defined as the intersection of an antero-posterior line drawn at the widest width, and superior-inferior line drawn at the tallest span of the glenoid.¹² The 12 o’clock position was defined as the intersection of the latter line and the superior glenoid (Figure 3). For clock face measurement, the clock face was applied to both left and right glenoids with 3 o’clock on the anterior glenoid and 9 o’clock on the posterior glenoid, as per convention.

Figure 2.

Glenoid specimen after dissection.

Figure 3.

CT analysis: The center of the glenoid was defined as the intersection of an antero-posterior line drawn at the widest width, and superior-inferior line drawn at the tallest span of the glenoid. The 12 o’clock position was defined as the intersection of this line and the superior glenoid.

The position of the anchors on the glenoid relative to the clock face was determined. The average deviation (in minutes) on the clock face from the target position was calculated for each anchor position. The accuracy was expressed as a percentage value (number of anchors on target with a deviation equal or less than 30 min divided the total number of anchors placed). Lastly, the inter- and intra-surgeon variability was calculated. The inter-surgeon variability was expressed as a range of values including the lowest and highest mean deviation from the target between surgeons. The intra-surgeon variability was expressed as a range of values including the lowest and highest mean deviation from the target within each surgeon group. A radiologist and an orthopedic surgeon reached agreement on all measurements performed using syngo® WebSpace (Siemens Medical Solutions USA, IL).

Descriptive statistics were utilized as appropriate to present the data in this study. The accuracy data obtained were also analyzed with one-way analysis of variance (ANOVA) test in order to determine if there were differences in accuracy at the designated anchor positions. One-way ANOVA is a reliable statistical technique when comparing three or more independent groups when responses are normally distributed. The intraclass correlation coefficients (ICCs) – model 2, individual – were calculated to assess reliability in anchor placement between and among surgeons. ICC is a descriptive statistic used to measure agreement or reliability within a set of data, whereby the measurements used are assumed to be parametric. SPSS version 16.0 software (SPSS, Chicago, IL) was used for statistical analysis.

Results

The overall accuracy of anchor placement was 57% (17/30 anchors placed within 30 min of target position), with the overall mean deviation from the intended target position of 47 min (SD 30 min, range 5–82.5 min). The accuracy of anchor placement at the 3:30 position was 40% (4/10) with mean deviation of 66 min (SD 52 min, range 0–150 min, average position 2:24 o’clock). The accuracy of anchor placement at the 4:30 position was 50% (5/10), with mean deviation from the target position of 48 min (SD 46 min, range 0–150 min, average position 3:42 o’clock). The accuracy of anchor placement at the 5:30 position was 80% (8/10), with mean deviation of 27 min (SD 26 min, range 0–90 min, average position 5:03 o’clock). There was no statistical difference in accuracy between the placement of the superior, middle, and inferior anchors (p = 0.145). Tables 1 to 3 report the deviation from the target analyzed for each surgeon. The ICC for inter-surgeon reliability was 0.4 for the inter-surgeon’s analysis indicating fair correlation, while the ICC for intra-surgeon reliability was 0.6, indicating moderate correlation.

Table 2.

Deviation from the target for each surgeon analyzed by anchor position.

	3:30 anchor deviation (min)	4:30 anchor deviation (min)	5:30 anchor deviation (min)
Surgeon 1	112.5	82.5	52.5
Surgeon 2	90	75	22.5
Surgeon 3	7.5	0	7.5
Surgeon 4	105	60	30
Surgeon 5	15	22.5	22.5

Table 1.

Anchor position: Overall deviation from the target for each surgeon.

	Mean deviation (min)	Standard deviation (min)	Range (min)
Surgeon 1	82.5	60.5	15–150
Surgeon 2	62.5	34.7	0–90
Surgeon 3	5	7.7	0–15
Surgeon 4	65	35	30–120
Surgeon 5	20	12.2	0–30

Table 3.

Deviation from the target with the accepted range of variability (±30 min) for each surgeon analyzed by anchor position.

	3:30 w/in 30 min	4:30 w/in 30 min	5:30 w/in 30 min
Surgeon 1	0/2	1/2	1/2
Surgeon 2	0/2	0/2	1/2
Surgeon 3	2/2	2/2	2/2
Surgeon 4	0/2	0/2	2/2
Surgeon 5	2/2	2/2	2/2

Discussion

Our hypothesis was that an experienced arthroscopic surgeon would be able to insert suture anchors at predetermined locations on the glenoid according to the clock face system with a high degree of accuracy and reliability was rejected. The mean deviation from the intended target position was higher than expected in 43.4% of cases (13 out of 30 anchors). The greatest deviation from the target position was observed for the superior/3:30 anchor (66 min), while the least deviation was seen with the placement of the inferior/5:30 anchor (mean deviation 27 min). The accuracy of placing the 5:30 anchor was also highest at 80%, falling to 40% for the 3:30 anchor.

The reasons for this limited accuracy could be searched in surgeon training, could be secondary to poor reliability of the CT measurement technique, or due to inherent unreliability of the clock face system. With regard to surgeon training, each surgeon was fellowship trained in arthroscopic shoulder surgery and had considerable experience in labral repair using suture anchors. While variations in portal placement might be responsible for the variation in anchor positioning,^13–15 the highest variability with anchor placement in this study was consistently seen with the most superior anchor, suggesting that portal placement was not a factor in this variability. At this time, to our knowledge, the intra- or inter-observer reliability of clock face measurement using CT is unknown.

The results of this study generate a degree of skepticism regarding the use of the clock face method for anchor insertion. Anatomically, shoulder arthroscopy surgeons use the biceps anchor, originating from the supraglenoid tubercle, as landmark for the 12 o’clock position. Several anatomic variants have been described including origins both anterior and posterior to the supraglenoid tubercle;^16–19 such variation from the normal anatomy could certainly lead to surgeon variation in intra-operative clock face estimation. The described CT-based method for assigning the 12 o’clock position on the glenoid is independent of the supraglenoid tubercle.¹² However, variations between arthroscopic determination of the 12 o’clock position on the glenoid and CT designation would not explain the high degree of both inter- and intra-observer variability.

A previous study that aimed to determine the reliability of orthopedic shoulder surgeons in identifying intra-articular structures in shoulder arthroscopy demonstrated that while the inter-rater reliability of identifying the anterior labrum was very good, it was poor for the glenoid and anterior-inferior glenohumeral ligament, and intermediate for the remainder of structures.²⁰ The results of this study certainly support our own findings that surgeons are not able to identify clock face landmarks with the high degree of accuracy needed. To be fair, it is not certain that positioning the anchors at 3:30, 4:30, and 5:30 is critical for the outcome of labral repair – the results of this study may suggest that good clinical outcomes may be relatively independent of anchor placement in view of the actual variability.

This study has some limitations. First, while there was no statistical difference in accuracy in anchor placement at the different target positions, it may be that our study was underpowered to detect such a difference. Second, the placement of anchors in a right versus a left shoulder might affect the inter-surgeon variability. In order to minimize this bias each surgeon inserted the anchors at pre-determined glenoid positions in one right and one left shoulder. Third, the axis chosen anatomically to define clock face position on the CT might not be the same axis used by surgeons at the time of arthroscopy. While the identification of the biceps as 12 o’clock position can be consistent, there are no objective anatomic landmarks for the 3 and 6 o’clock positions. Each surgeon tried to identify the 3 o’clock position at widest margin and the 6 o’clock position at the lowest margin of the glenoid. This factor might also explain the differences found between surgeons; however, it certainly reinforces the concept that using the clock face method to guide anchor placement is associated with high inter-surgeon variability. Finally, the use of cadaveric limbs may have limited correlation to the clinical setting – a study involving the analysis of anchor position in glenoids and correlation to clinical outcome might be more valid.

In conclusion, the findings of this study suggest a moderate degree of accuracy and fair to moderate inter- and intra-surgeon reliability when using the clock face system during shoulder arthroscopy to guide suture anchor placement. The clock face method does not provide a reliable and accurate method for placement of suture anchors. Care should be taken in interpreting surgical reports that use this measure pedantically.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Funding for this study was provided from the University of Toronto Orthopaedic Sports Medicine Program research budget.