Knee Surg Sports Traumatol Arthrosc DOI 10.1007/s00167-014-2945-1
KNEE
The sizing of hamstring grafts for anterior cruciate reconstruction: intra- and inter-observer reliability Tim Dwyer • Daniel B. Whelan • Amir Khoshbin • David Wasserstein • Andrew Dold • Jaskarndip Chahal • Aaron Nauth • M. Lucas Murnaghan Darrell J. Ogilvie-Harris • John S. Theodoropoulos
•
Received: 21 October 2013 / Accepted: 10 March 2014 Ó Springer-Verlag Berlin Heidelberg 2014
Abstract Purpose The objective of this study was to establish the intra- and inter-observer reliability of hamstring graft measurement using cylindrical sizing tubes. Methods Hamstring tendons (gracilis and semitendinosus) were harvested from ten cadavers by a single surgeon and whip stitched together to create ten 4-strand hamstring grafts. Ten sports medicine surgeons and fellows sized each graft independently using either hollow cylindrical sizers or block sizers in 0.5-mm increments—the sizing technique used was applied consistently to each graft. Surgeons moved sequentially from graft to graft and measured each hamstring graft twice. Surgeons were asked to state the measured proximal (femoral) and distal (tibial) diameter of each graft, as well as the diameter of the tibial
and femoral tunnels that they would drill if performing an anterior cruciate ligament (ACL) reconstruction using that graft. Reliability was established using intra-class correlation coefficients. Results Overall, both the inter-observer and intra-observer agreement were [0.9, demonstrating excellent reliability. The inter-observer reliability for drill sizes was also excellent ([0.9). Excellent correlation was seen between cylindrical sizing, and drill sizes ([0.9). Conclusions Sizing of hamstring grafts by multiple surgeons demonstrated excellent intra-observer and intraobserver reliability, potentially validating clinical studies exploring ACL reconstruction outcomes by hamstring graft diameter when standard techniques are used. Level of evidence III. Keywords
T. Dwyer (&) D. B. Whelan D. Wasserstein J. Chahal A. Nauth M. L. Murnaghan D. J. Ogilvie-Harris J. S. Theodoropoulos University of Toronto Orthopaedic Sports Medicine, Toronto, Canada e-mail: [email protected] T. Dwyer D. B. Whelan J. Chahal M. L. Murnaghan D. J. Ogilvie-Harris J. S. Theodoropoulos Women’s College Hospital, Toronto, Canada T. Dwyer J. S. Theodoropoulos Mt Sinai Hospital, Toronto, Canada D. B. Whelan A. Nauth St Michael’s Hospital, Toronto, Canada A. Khoshbin A. Dold University of Toronto, Toronto, Canada J. Chahal D. J. Ogilvie-Harris Toronto Western Hospital, Toronto, Canada
Hamstring Sizing Reliability
Introduction Recent literature in anterior cruciate ligament reconstruction (ACLR) using hamstring autograft suggests that smaller graft sizes may be correlated with an increased rate of graft failure [8, 10]. A retrospective study by Magnussen et al. [8] of 132 patients undergoing ACLR showed that hamstring autograft size of 7 mm or less was associated with an increased risk of revision surgery. Factors such as height (\147 cm tall), female sex, weight, and thigh circumference, are known to be associated with smaller hamstring graft diameter [3, 12, 13]. Recently, studies have demonstrated the ability of magnetic resonance imaging (MRI) to predict hamstring graft size pre-operatively, potentially allowing a discussion of appropriate graft choice and/or augmentation with patients prior to surgery [1, 2, 15].
123
Knee Surg Sports Traumatol Arthrosc
While a prospective study would be required in order to assess the true failure risk of using smaller hamstring graft diameters, at this time, the intra-observer (same surgeon, different times) and inter-observer (different surgeons) variability of measuring hamstring graft diameters is unknown. Typically the semitendinosus and gracilis are harvested, stripped of any residual muscle belly, doubled over, and whip stitched together creating a four-strand graft. The diameter of this construct is typically sized with the use of sequential cylindrical sizers or testing blocks, in 0.5-mm increments. The objective of this study was to establish the intraand inter-observer reliability of hamstring graft measurement using cylindrical sizing tubes and testing blocks. The hypothesis was that there would be excellent intra-observer and inter-observer reliability with respect to hamstring size measurement.
Materials and methods Hamstring tendons (gracilis and semitendinosus) were harvested by a single surgeon from ten (six male and four female) full cadaveric legs up to the top of the thigh, with an age range between 63 and 92). The residual muscle belly was removed in a standard fashion, and the tendons doubled over an Endobutton loop (Smith & Nephew, Andover, MA, USA), creating ten four-strand hamstring grafts. All four strands were whip stitched together over the distal third using a no 1 Ultrabraid suture (Smith & Nephew, Andover, MA, USA); whip stitching was performed in a non-locked fashion in order to minimize bulk. All grafts were pre-tensioned to 15 lb for 5 min on a graft tensioner (Acufex Graftmaster III, Smith & Nephew, Andover, MA, USA)—in order to minimize swelling, grafts were not covered in wet gauze. Two of the grafts were augmented with a tibialis posterior tendon, allowing the creation of two larger grafts in order to increase the spectrum of graft sizes. Ten sports medicine surgeons and fellows sized each graft independently, within the same 1-h time period to minimize potential variation in sizes secondary to drying. Five surgeons regularly performed hamstring ACL reconstruction, two surgeons performed predominantly bone– patella tendon–bone ACL reconstruction, with three orthopaedic sports medicine fellows performing the remainder of sizings. Sizing was performed using either hollow cylindrical sizers in 0.5-mm increments (Acufex, Smith & Nephew, Andover, MA, USA) or block sizers with 0.5-mm increments (Graft Sizing Block, GraFix, ConMed Linvatec, Largo, FL, USA)—the sizing technique was used consistently for each graft. Surgeons moved sequentially from graft to graft and completed the rotation
123
twice, allowing them to measure each hamstring graft twice. Surgeons were asked to state the measured proximal (femoral) and distal (tibial) diameter of each graft, as well as the diameter of the tibial and femoral tunnels they would drill if performing an ACLR using that graft. Independent observers recorded all measurements, blinding the surgeon to the measurements of other surgeons and to their own previous graft measurement. Concurrent validity of graft diameter was sought by measuring graft circumference. Each hamstring graft was placed on a tensioning device (5-lb tension), and the graft circumference measuring using a no 1 suture to encompass the graft in the proximal and distal tendon. Each suture length was then measured with a digital calliper (accurate to 0.1 mm) and converted mathematically to a hamstring diameter. Approval for conducting this study was granted by the Research Ethics Board, Mt Sinai Hospital, University of Toronto. Statistical analysis First and second run tibial measurement, femoral measurement, and mean of the two measurements by surgeon sizing and drill size were examined for intra-observer and inter-observer variation. First and second run measurements were averaged for an ‘overall’ or average measurement. All measurements were summarized using medians and range. To examine intra-observer variation (test–retest by same surgeon), the first and second run measurements of surgeon sizing and drill size, respectively, were compared using intraclass correlation coefficients [ICC (2,1)] for test–retest agreement and their 95 % confidence intervals [4, 7, 11]. The ICC was used rather than Kappa statistics in order to interpret interval data. These comparisons were done for tibial and femoral measurements separately. Inter-observer variation (extent to which different surgeons agreed) was calculated using an ICC (2,k) for variation of the mean of the two measurements of hamstring size and drill size between surgeons [4, 11]. Each of these measurements were also correlated with graft diameter as measured by graft circumference using Pearson’s correlation coefficient (rp) and compared for differences using an adjusted form of the paired t test. p values testing the significance of the Pearson’s correlation coefficient and the paired t test accounted for the clustered nature of the data (i.e. multiple measurements conducted by each surgeon) and were derived from random-effects models. A p value of \0.05 was used to indicate statistical significance. SAS v9.3 (Cary, NC, USA) was used for all statistical analyses.
Knee Surg Sports Traumatol Arthrosc
Post hoc power analysis demonstrated that if it was expected that raters would agree approximately 50 % of the time, then a sample of 25 cadavers would be required (assuming a 40 % relative error rate). As a result, this study may have been potentially underpowered.
As the agreement between observers was not known at the time of conducting this research, a power analysis was conducted in which it was estimated that raters would agree approximately 50 % of the time; assuming a 40 % relative error rate, a sample size of 25 cadavers would be required. Owing to limited number of specimens and cost, 10 cadaveric specimens were used.
Discussion The most important finding of this study is that surgeon measurement of hamstring size using cylindrical sizers is both reproducible and accurate, with intra-observer and inter-observer agreement [0.9. This is important as investigators continue to examine the relationship between hamstring graft diameters and outcomes of ACLR surgery. There is some literature to suggest that smaller diameter hamstring grafts are associated with increased failure rates in ACLR. The results of this study suggest that the possibility of confounding measurement error is low. A retrospective review by Magnussen et al. [8] found that of 132 patients having a hamstring ACLR, 59 had a graft equal to or less than 7 mm in diameter—the failure rate in these patients was significantly higher (13.6 %) than in patients with grafts greater than 8 mm (1 %). In this study, age \20 years was also shown to be significantly correlated with failed surgery, consistent with other studies looking at failed ACLR [5, 6, 14]. Park et al. [10] reported similar findings, following 296 patients after hamstring ACLR and identified improved outcomes when grafts 8 mm or greater were used, compared to those with grafts less than 8 mm. In contrast, a study by Kamien et al. [5] of 98 patients did not identify a correlation between failure rate and hamstring graft diameter. Clearly, this issue needs to be addressed with a large, prospective cohort study—the results of this current study demonstrate that the pooling of
Results The median and ranges for each of the hamstring grafts using cylindrical sizers, drill measurement, and measurement of diameter by way of circumference are seen in Table 1. The inter-observer reliability for the hamstring sizing (95 % confidence interval) was 0.95 (0.9–0.99) for the tibial measurement and 0.92 (0.85–0.98) for the femoral measurement. The intra-observer reliability (test–retest) for the tibial measurement was 0.99 (0.97–1) and 0.99 (0.95–1) for the femoral measurement. The inter-observer correlation for the tibial drill sizes (based upon hamstring sizing) was 0.95 (0.9–0.98) and 0.93 (0.86–0.98) for the femoral drill sizes. The correlation between the hamstring measurement and circumference converted to diameter was 0.84 (p \ 0.001) for the tibial side and 0.91 (p \ 0.001) for the femoral side. The number of times a surgeon measured the graft differently between the first and second measurement varied between 2/10 grafts (4 surgeons), 3/10 grafts (1 surgeon), 4/10 grafts (2 surgeons), 6/10 grafts (1 surgeon), and 9/10 grafts (1 surgeon, exclusively performing BPTB ACLR). The mean difference was 0.57 (range 0.5–1 mm). In 35/39 measurements, the second measurement was smaller, suggesting that grafts may have decreased in size from multiple measurements.
Table 1 Median hamstring graft sizes using cylindrical sizersa and block sizersb, median drill measurement for each graft, and graft circumference converted to diameter
Graft number
1a
Drill measurement median (range)
Circumference measurement converted to diameter
F
F
F
T 7.9
T 8 (7.5–10)
T
8.5 (8–10)
8.5 (7.5–10)
8.5 (8–10)
7.7
2a
6.5 (6.5–7)
6.5 (6.5–7)
6.5 (6.5–7)
6.5 (6.5–7)
6.7
5.8
3a 4a
8.5 (8–9) 10 (9.5–11)
9 (8.5–9.5) 10.5 (10–11)
8.5 (8–9.5) 10.5 (10–11)
9 (8.5–9.5) 10.5 (10.5–11)
8.1 9.4
8.7 9.5
5a
9 (8.5–10)
9.5 (8.5–10)
9.5 (8.5–10)
9.5 (8.5–10)
9.0
9.3
a
F femoral side of graft, T tibial side of graft
Hamstring size median (range)
6
8.5 (8–9)
8.5 (8.5–95)
8.5 (8.5–9)
8.5 (8.5–9)
8.4
8.8
7b
8.5 (7.5–8.5)
8.5 (7.5–8.5)
8.5 (7.5–9)
8.5 (7.5–9)
8.6
8.2
8b
7 (6.5–8)
7 (6.5–7.5)
7 (6.5–8)
7 (6.5–8)
7.3
6.7
9b
8 (7–8.5)
8 (7.5–8.5)
8 (7–8.5)
8 (7.5–8.5)
8.4
6.0
8.6
8.4
10b
8.5 (8–9)
8.5 (8–9)
8.5 (8–9)
8.5 (8–9)
123
Knee Surg Sports Traumatol Arthrosc
results between centres would be feasible, as surgeons are consistent when sizing hamstring grafts. In this study, an excellent correlation was seen between the diameter measured using both cylindrical sizers and open sizing blocks, and recommended drill sizes. Certainly, there is a ‘feel’ to both sizing techniques, with some surgeons tending to upsize drill sizes in order to allow the hamstring graft to pass easily during reconstruction, while others opt for a tight line-to-line drilling in order to allow ‘fit and fill’ of the bony tunnels with the hamstring graft. In the event of differential sizing of the femoral and tibial ends of the graft, some surgeons perform differential drilling and while some surgeons upsize the femoral drill size to match the tibia. However, in this study, the variability between hamstring sizing and drill sizing (femoral side, tibial side, and mean) was minimal, suggesting that surgeons perform line-to-line drilling in the majority of cases. This information will be useful to surgeons conducting large multi-centred studies looking to establish whether there is indeed a link between smaller diameter hamstring diameters and increased revision rates after ACLR. Limitations of this study are related to the viscoelastic behaviour of hamstring tendons—under pressure, these grafts have been shown to decrease in both volume and diameter [9]. As a result, after multiple passages through cylindrical sizers, it is possible that the hamstring diameters would have reduced by the second run of sizing. However, a very good intra-observer and inter-observer reliability was observed, suggesting that this effect was minimal. Power analysis suggested that 25 cadaveric specimens were required; this study may have been underpowered due to the limitations of cost. Finally, in this study, two types of sizing blocks were used—closed cylindrical sizing tubes and open sizing blocks. This difference was mitigated by using the same sizing device for each hamstring—however, by doing so, a direct comparison of the two methods could not be made.
Conclusion The sizing of hamstring grafts by multiple surgeons demonstrated excellent intra-observer and extra-observer reliability, suggesting that multi-centre trials examining the relationship between hamstrings graft size and outcomes after anterior cruciate ligament reconstruction are feasible.
123
References 1. Beyzadeoglu T, Akgun U, Tasdelen N, Karahan M (2012) Prediction of semitendinosus and gracilis autograft sizes for ACL reconstruction. Knee Surg Sports Traumatol Arthrosc 20:1293–1297 2. Bickel BA, Fowler TT, Mowbray JG, Adler B, Klingele K, Phillips G (2008) Preoperative magnetic resonance imaging cross-sectional area for the measurement of hamstring autograft diameter for reconstruction of the adolescent anterior cruciate ligament. Arthroscopy 24:1336–1341 3. Boisvert CB, Aubin ME, DeAngelis N (2011) Relationship between anthropometric measurements and hamstring autograft diameter in anterior cruciate ligament reconstruction. Am J Orthop (Belle Mead NJ) 40:293–295 4. Cohen J (1960) A coefficient of agreement for nominal scales. Educ Psychol Meas 20:37–46 5. Kamien PM, Hydrick JM, Replogle WH, Go LT, Barrett GR (2013) Age, graft size, and Tegner activity level as predictors of failure in anterior cruciate ligament reconstruction with hamstring autograft. Am J Sports Med 41:1808–1812 6. Lind M, Menhert F, Pedersen AB (2012) Incidence and outcome after revision anterior cruciate ligament reconstruction: results from the Danish registry for knee ligament reconstructions. Am J Sports Med 40:1551–1557 7. Maclure M, Willett WC (1987) Misinterpretation and misuse of the kappa statistic. Am J Epidemiol 126:161–169 8. Magnussen RA, Lawrence JT, West RL, Toth AP, Taylor DC, Garrett WE (2012) Graft size and patient age are predictors of early revision after anterior cruciate ligament reconstruction with hamstring autograft. Arthroscopy 28:526–531 9. Meyer DC, Snedeker JG, Weinert-Aplin RA, Farshad M (2012) Viscoelastic adaptation of tendon graft material to compression: biomechanical quantification of graft preconditioning. Arch Orthop Trauma Surg 132:1315–1320 10. Park SY, Oh H, Park S, Lee JH, Lee SH, Yoon KH (2013) Factors predicting hamstring tendon autograft diameters and resulting failure rates after anterior cruciate ligament reconstruction. Knee Surg Sports Traumatol Arthrosc 21:1111–1118 11. Shrout PE, Fleiss JL (1979) Intraclass correlations: uses in assessing rater reliability. Psychol Bull 86:420–428 12. Treme G, Diduch DR, Billante MJ, Miller MD, Hart JM (2008) Hamstring graft size prediction: a prospective clinical evaluation. Am J Sports Med 36:2204–2209 13. Tuman JM, Diduch DR, Rubino LJ, Baumfeld JA, Nguyen HS, Hart JM (2007) Predictors for hamstring graft diameter in anterior cruciate ligament reconstruction. Am J Sports Med 35:1945–1949 14. Wasserstein D, Khoshbin A, Dwyer T, Chahal J, Gandhi R, Mahomed N, Ogilvie-Harris D (2013) Risk factors for recurrent anterior cruciate ligament reconstruction: a population study in Ontario, Canada, with 5-year follow-up. Am J Sports Med 41:2099–2107 15. Wernecke G, Harris IA, Houang MT, Seeto BG, Chen DB, MacDessi SJ (2011) Using magnetic resonance imaging to predict adequate graft diameters for autologous hamstring doublebundle anterior cruciate ligament reconstruction. Arthroscopy 27:1055–1059
KNEE
The sizing of hamstring grafts for anterior cruciate reconstruction: intra- and inter-observer reliability Tim Dwyer • Daniel B. Whelan • Amir Khoshbin • David Wasserstein • Andrew Dold • Jaskarndip Chahal • Aaron Nauth • M. Lucas Murnaghan Darrell J. Ogilvie-Harris • John S. Theodoropoulos
•
Received: 21 October 2013 / Accepted: 10 March 2014 Ó Springer-Verlag Berlin Heidelberg 2014
Abstract Purpose The objective of this study was to establish the intra- and inter-observer reliability of hamstring graft measurement using cylindrical sizing tubes. Methods Hamstring tendons (gracilis and semitendinosus) were harvested from ten cadavers by a single surgeon and whip stitched together to create ten 4-strand hamstring grafts. Ten sports medicine surgeons and fellows sized each graft independently using either hollow cylindrical sizers or block sizers in 0.5-mm increments—the sizing technique used was applied consistently to each graft. Surgeons moved sequentially from graft to graft and measured each hamstring graft twice. Surgeons were asked to state the measured proximal (femoral) and distal (tibial) diameter of each graft, as well as the diameter of the tibial
and femoral tunnels that they would drill if performing an anterior cruciate ligament (ACL) reconstruction using that graft. Reliability was established using intra-class correlation coefficients. Results Overall, both the inter-observer and intra-observer agreement were [0.9, demonstrating excellent reliability. The inter-observer reliability for drill sizes was also excellent ([0.9). Excellent correlation was seen between cylindrical sizing, and drill sizes ([0.9). Conclusions Sizing of hamstring grafts by multiple surgeons demonstrated excellent intra-observer and intraobserver reliability, potentially validating clinical studies exploring ACL reconstruction outcomes by hamstring graft diameter when standard techniques are used. Level of evidence III. Keywords
T. Dwyer (&) D. B. Whelan D. Wasserstein J. Chahal A. Nauth M. L. Murnaghan D. J. Ogilvie-Harris J. S. Theodoropoulos University of Toronto Orthopaedic Sports Medicine, Toronto, Canada e-mail: [email protected] T. Dwyer D. B. Whelan J. Chahal M. L. Murnaghan D. J. Ogilvie-Harris J. S. Theodoropoulos Women’s College Hospital, Toronto, Canada T. Dwyer J. S. Theodoropoulos Mt Sinai Hospital, Toronto, Canada D. B. Whelan A. Nauth St Michael’s Hospital, Toronto, Canada A. Khoshbin A. Dold University of Toronto, Toronto, Canada J. Chahal D. J. Ogilvie-Harris Toronto Western Hospital, Toronto, Canada
Hamstring Sizing Reliability
Introduction Recent literature in anterior cruciate ligament reconstruction (ACLR) using hamstring autograft suggests that smaller graft sizes may be correlated with an increased rate of graft failure [8, 10]. A retrospective study by Magnussen et al. [8] of 132 patients undergoing ACLR showed that hamstring autograft size of 7 mm or less was associated with an increased risk of revision surgery. Factors such as height (\147 cm tall), female sex, weight, and thigh circumference, are known to be associated with smaller hamstring graft diameter [3, 12, 13]. Recently, studies have demonstrated the ability of magnetic resonance imaging (MRI) to predict hamstring graft size pre-operatively, potentially allowing a discussion of appropriate graft choice and/or augmentation with patients prior to surgery [1, 2, 15].
123
Knee Surg Sports Traumatol Arthrosc
While a prospective study would be required in order to assess the true failure risk of using smaller hamstring graft diameters, at this time, the intra-observer (same surgeon, different times) and inter-observer (different surgeons) variability of measuring hamstring graft diameters is unknown. Typically the semitendinosus and gracilis are harvested, stripped of any residual muscle belly, doubled over, and whip stitched together creating a four-strand graft. The diameter of this construct is typically sized with the use of sequential cylindrical sizers or testing blocks, in 0.5-mm increments. The objective of this study was to establish the intraand inter-observer reliability of hamstring graft measurement using cylindrical sizing tubes and testing blocks. The hypothesis was that there would be excellent intra-observer and inter-observer reliability with respect to hamstring size measurement.
Materials and methods Hamstring tendons (gracilis and semitendinosus) were harvested by a single surgeon from ten (six male and four female) full cadaveric legs up to the top of the thigh, with an age range between 63 and 92). The residual muscle belly was removed in a standard fashion, and the tendons doubled over an Endobutton loop (Smith & Nephew, Andover, MA, USA), creating ten four-strand hamstring grafts. All four strands were whip stitched together over the distal third using a no 1 Ultrabraid suture (Smith & Nephew, Andover, MA, USA); whip stitching was performed in a non-locked fashion in order to minimize bulk. All grafts were pre-tensioned to 15 lb for 5 min on a graft tensioner (Acufex Graftmaster III, Smith & Nephew, Andover, MA, USA)—in order to minimize swelling, grafts were not covered in wet gauze. Two of the grafts were augmented with a tibialis posterior tendon, allowing the creation of two larger grafts in order to increase the spectrum of graft sizes. Ten sports medicine surgeons and fellows sized each graft independently, within the same 1-h time period to minimize potential variation in sizes secondary to drying. Five surgeons regularly performed hamstring ACL reconstruction, two surgeons performed predominantly bone– patella tendon–bone ACL reconstruction, with three orthopaedic sports medicine fellows performing the remainder of sizings. Sizing was performed using either hollow cylindrical sizers in 0.5-mm increments (Acufex, Smith & Nephew, Andover, MA, USA) or block sizers with 0.5-mm increments (Graft Sizing Block, GraFix, ConMed Linvatec, Largo, FL, USA)—the sizing technique was used consistently for each graft. Surgeons moved sequentially from graft to graft and completed the rotation
123
twice, allowing them to measure each hamstring graft twice. Surgeons were asked to state the measured proximal (femoral) and distal (tibial) diameter of each graft, as well as the diameter of the tibial and femoral tunnels they would drill if performing an ACLR using that graft. Independent observers recorded all measurements, blinding the surgeon to the measurements of other surgeons and to their own previous graft measurement. Concurrent validity of graft diameter was sought by measuring graft circumference. Each hamstring graft was placed on a tensioning device (5-lb tension), and the graft circumference measuring using a no 1 suture to encompass the graft in the proximal and distal tendon. Each suture length was then measured with a digital calliper (accurate to 0.1 mm) and converted mathematically to a hamstring diameter. Approval for conducting this study was granted by the Research Ethics Board, Mt Sinai Hospital, University of Toronto. Statistical analysis First and second run tibial measurement, femoral measurement, and mean of the two measurements by surgeon sizing and drill size were examined for intra-observer and inter-observer variation. First and second run measurements were averaged for an ‘overall’ or average measurement. All measurements were summarized using medians and range. To examine intra-observer variation (test–retest by same surgeon), the first and second run measurements of surgeon sizing and drill size, respectively, were compared using intraclass correlation coefficients [ICC (2,1)] for test–retest agreement and their 95 % confidence intervals [4, 7, 11]. The ICC was used rather than Kappa statistics in order to interpret interval data. These comparisons were done for tibial and femoral measurements separately. Inter-observer variation (extent to which different surgeons agreed) was calculated using an ICC (2,k) for variation of the mean of the two measurements of hamstring size and drill size between surgeons [4, 11]. Each of these measurements were also correlated with graft diameter as measured by graft circumference using Pearson’s correlation coefficient (rp) and compared for differences using an adjusted form of the paired t test. p values testing the significance of the Pearson’s correlation coefficient and the paired t test accounted for the clustered nature of the data (i.e. multiple measurements conducted by each surgeon) and were derived from random-effects models. A p value of \0.05 was used to indicate statistical significance. SAS v9.3 (Cary, NC, USA) was used for all statistical analyses.
Knee Surg Sports Traumatol Arthrosc
Post hoc power analysis demonstrated that if it was expected that raters would agree approximately 50 % of the time, then a sample of 25 cadavers would be required (assuming a 40 % relative error rate). As a result, this study may have been potentially underpowered.
As the agreement between observers was not known at the time of conducting this research, a power analysis was conducted in which it was estimated that raters would agree approximately 50 % of the time; assuming a 40 % relative error rate, a sample size of 25 cadavers would be required. Owing to limited number of specimens and cost, 10 cadaveric specimens were used.
Discussion The most important finding of this study is that surgeon measurement of hamstring size using cylindrical sizers is both reproducible and accurate, with intra-observer and inter-observer agreement [0.9. This is important as investigators continue to examine the relationship between hamstring graft diameters and outcomes of ACLR surgery. There is some literature to suggest that smaller diameter hamstring grafts are associated with increased failure rates in ACLR. The results of this study suggest that the possibility of confounding measurement error is low. A retrospective review by Magnussen et al. [8] found that of 132 patients having a hamstring ACLR, 59 had a graft equal to or less than 7 mm in diameter—the failure rate in these patients was significantly higher (13.6 %) than in patients with grafts greater than 8 mm (1 %). In this study, age \20 years was also shown to be significantly correlated with failed surgery, consistent with other studies looking at failed ACLR [5, 6, 14]. Park et al. [10] reported similar findings, following 296 patients after hamstring ACLR and identified improved outcomes when grafts 8 mm or greater were used, compared to those with grafts less than 8 mm. In contrast, a study by Kamien et al. [5] of 98 patients did not identify a correlation between failure rate and hamstring graft diameter. Clearly, this issue needs to be addressed with a large, prospective cohort study—the results of this current study demonstrate that the pooling of
Results The median and ranges for each of the hamstring grafts using cylindrical sizers, drill measurement, and measurement of diameter by way of circumference are seen in Table 1. The inter-observer reliability for the hamstring sizing (95 % confidence interval) was 0.95 (0.9–0.99) for the tibial measurement and 0.92 (0.85–0.98) for the femoral measurement. The intra-observer reliability (test–retest) for the tibial measurement was 0.99 (0.97–1) and 0.99 (0.95–1) for the femoral measurement. The inter-observer correlation for the tibial drill sizes (based upon hamstring sizing) was 0.95 (0.9–0.98) and 0.93 (0.86–0.98) for the femoral drill sizes. The correlation between the hamstring measurement and circumference converted to diameter was 0.84 (p \ 0.001) for the tibial side and 0.91 (p \ 0.001) for the femoral side. The number of times a surgeon measured the graft differently between the first and second measurement varied between 2/10 grafts (4 surgeons), 3/10 grafts (1 surgeon), 4/10 grafts (2 surgeons), 6/10 grafts (1 surgeon), and 9/10 grafts (1 surgeon, exclusively performing BPTB ACLR). The mean difference was 0.57 (range 0.5–1 mm). In 35/39 measurements, the second measurement was smaller, suggesting that grafts may have decreased in size from multiple measurements.
Table 1 Median hamstring graft sizes using cylindrical sizersa and block sizersb, median drill measurement for each graft, and graft circumference converted to diameter
Graft number
1a
Drill measurement median (range)
Circumference measurement converted to diameter
F
F
F
T 7.9
T 8 (7.5–10)
T
8.5 (8–10)
8.5 (7.5–10)
8.5 (8–10)
7.7
2a
6.5 (6.5–7)
6.5 (6.5–7)
6.5 (6.5–7)
6.5 (6.5–7)
6.7
5.8
3a 4a
8.5 (8–9) 10 (9.5–11)
9 (8.5–9.5) 10.5 (10–11)
8.5 (8–9.5) 10.5 (10–11)
9 (8.5–9.5) 10.5 (10.5–11)
8.1 9.4
8.7 9.5
5a
9 (8.5–10)
9.5 (8.5–10)
9.5 (8.5–10)
9.5 (8.5–10)
9.0
9.3
a
F femoral side of graft, T tibial side of graft
Hamstring size median (range)
6
8.5 (8–9)
8.5 (8.5–95)
8.5 (8.5–9)
8.5 (8.5–9)
8.4
8.8
7b
8.5 (7.5–8.5)
8.5 (7.5–8.5)
8.5 (7.5–9)
8.5 (7.5–9)
8.6
8.2
8b
7 (6.5–8)
7 (6.5–7.5)
7 (6.5–8)
7 (6.5–8)
7.3
6.7
9b
8 (7–8.5)
8 (7.5–8.5)
8 (7–8.5)
8 (7.5–8.5)
8.4
6.0
8.6
8.4
10b
8.5 (8–9)
8.5 (8–9)
8.5 (8–9)
8.5 (8–9)
123
Knee Surg Sports Traumatol Arthrosc
results between centres would be feasible, as surgeons are consistent when sizing hamstring grafts. In this study, an excellent correlation was seen between the diameter measured using both cylindrical sizers and open sizing blocks, and recommended drill sizes. Certainly, there is a ‘feel’ to both sizing techniques, with some surgeons tending to upsize drill sizes in order to allow the hamstring graft to pass easily during reconstruction, while others opt for a tight line-to-line drilling in order to allow ‘fit and fill’ of the bony tunnels with the hamstring graft. In the event of differential sizing of the femoral and tibial ends of the graft, some surgeons perform differential drilling and while some surgeons upsize the femoral drill size to match the tibia. However, in this study, the variability between hamstring sizing and drill sizing (femoral side, tibial side, and mean) was minimal, suggesting that surgeons perform line-to-line drilling in the majority of cases. This information will be useful to surgeons conducting large multi-centred studies looking to establish whether there is indeed a link between smaller diameter hamstring diameters and increased revision rates after ACLR. Limitations of this study are related to the viscoelastic behaviour of hamstring tendons—under pressure, these grafts have been shown to decrease in both volume and diameter [9]. As a result, after multiple passages through cylindrical sizers, it is possible that the hamstring diameters would have reduced by the second run of sizing. However, a very good intra-observer and inter-observer reliability was observed, suggesting that this effect was minimal. Power analysis suggested that 25 cadaveric specimens were required; this study may have been underpowered due to the limitations of cost. Finally, in this study, two types of sizing blocks were used—closed cylindrical sizing tubes and open sizing blocks. This difference was mitigated by using the same sizing device for each hamstring—however, by doing so, a direct comparison of the two methods could not be made.
Conclusion The sizing of hamstring grafts by multiple surgeons demonstrated excellent intra-observer and extra-observer reliability, suggesting that multi-centre trials examining the relationship between hamstrings graft size and outcomes after anterior cruciate ligament reconstruction are feasible.
123
References 1. Beyzadeoglu T, Akgun U, Tasdelen N, Karahan M (2012) Prediction of semitendinosus and gracilis autograft sizes for ACL reconstruction. Knee Surg Sports Traumatol Arthrosc 20:1293–1297 2. Bickel BA, Fowler TT, Mowbray JG, Adler B, Klingele K, Phillips G (2008) Preoperative magnetic resonance imaging cross-sectional area for the measurement of hamstring autograft diameter for reconstruction of the adolescent anterior cruciate ligament. Arthroscopy 24:1336–1341 3. Boisvert CB, Aubin ME, DeAngelis N (2011) Relationship between anthropometric measurements and hamstring autograft diameter in anterior cruciate ligament reconstruction. Am J Orthop (Belle Mead NJ) 40:293–295 4. Cohen J (1960) A coefficient of agreement for nominal scales. Educ Psychol Meas 20:37–46 5. Kamien PM, Hydrick JM, Replogle WH, Go LT, Barrett GR (2013) Age, graft size, and Tegner activity level as predictors of failure in anterior cruciate ligament reconstruction with hamstring autograft. Am J Sports Med 41:1808–1812 6. Lind M, Menhert F, Pedersen AB (2012) Incidence and outcome after revision anterior cruciate ligament reconstruction: results from the Danish registry for knee ligament reconstructions. Am J Sports Med 40:1551–1557 7. Maclure M, Willett WC (1987) Misinterpretation and misuse of the kappa statistic. Am J Epidemiol 126:161–169 8. Magnussen RA, Lawrence JT, West RL, Toth AP, Taylor DC, Garrett WE (2012) Graft size and patient age are predictors of early revision after anterior cruciate ligament reconstruction with hamstring autograft. Arthroscopy 28:526–531 9. Meyer DC, Snedeker JG, Weinert-Aplin RA, Farshad M (2012) Viscoelastic adaptation of tendon graft material to compression: biomechanical quantification of graft preconditioning. Arch Orthop Trauma Surg 132:1315–1320 10. Park SY, Oh H, Park S, Lee JH, Lee SH, Yoon KH (2013) Factors predicting hamstring tendon autograft diameters and resulting failure rates after anterior cruciate ligament reconstruction. Knee Surg Sports Traumatol Arthrosc 21:1111–1118 11. Shrout PE, Fleiss JL (1979) Intraclass correlations: uses in assessing rater reliability. Psychol Bull 86:420–428 12. Treme G, Diduch DR, Billante MJ, Miller MD, Hart JM (2008) Hamstring graft size prediction: a prospective clinical evaluation. Am J Sports Med 36:2204–2209 13. Tuman JM, Diduch DR, Rubino LJ, Baumfeld JA, Nguyen HS, Hart JM (2007) Predictors for hamstring graft diameter in anterior cruciate ligament reconstruction. Am J Sports Med 35:1945–1949 14. Wasserstein D, Khoshbin A, Dwyer T, Chahal J, Gandhi R, Mahomed N, Ogilvie-Harris D (2013) Risk factors for recurrent anterior cruciate ligament reconstruction: a population study in Ontario, Canada, with 5-year follow-up. Am J Sports Med 41:2099–2107 15. Wernecke G, Harris IA, Houang MT, Seeto BG, Chen DB, MacDessi SJ (2011) Using magnetic resonance imaging to predict adequate graft diameters for autologous hamstring doublebundle anterior cruciate ligament reconstruction. Arthroscopy 27:1055–1059
