Does quality of novice hand-tied square knots improve with repetition during a single training session?

Participants

This prospective study was approved by the University of Tennessee Institutional Review Board (UTK IRB-20-05723-XP). First year veterinary students enrolled in an elective, pass-fail surgery course were recruited as novice trainees for the study. Two surgeons certified by the American College of Veterinary Surgeons also participated as expert control subjects; both had more than 15 years of surgical experience.

Suture materials

A 500-yard (457.2 meter) length of 30-pound (13.6 kg) test, 0.022 inch (0.55 mm) diameter nylon monofilament fishing line (Suffix Superior, Rapala VMC Corp., 17200 Vaaksy, Finland) was cut into fifty 30-foot-long pieces. Each piece was cut into 20 sutures 45.7 cm (18 inches) in length. One random suture from each piece was collected and placed into a bag to produce 50 bags that each contained 20 sutures representative of the entire 500 yards of monofilament line.

Tying apparatus

Tying stands were fashioned from 1.5 cm (5/8 inch) diameter, polyethylene pipe (SharkBite PEX pipe, Cullman, Alabama, USA) and four cm (1 5/8 inch) diameter pipe insulation (Everbilt pipe insulation, Armacell, Chapel Hill, NC, USA). The pipe was cut into 30.5 cm (12 inch) long pieces, each of which was labelled with a number. The pipe insulation was cut into three cm (1.2 inch) long segments, and a segment was positioned near each end of the pipe to elevate the pipe off of the work surface. Each pipe was positioned so that the numbered end was toward the participant’s left. Pipes were secured to work surfaces with a rope that ran through the pipe and was tied under the surface. Additionally, the ends of the pipes were taped to the work surface to prevent rolling or shifting.

Trainee instruction

Each trainee was seated at a work station, and the trainee’s name and pipe number were recorded. Each trainee was given a bag of 20 sutures, a pair of suture scissors, and a strand of string as a practice suture. Before the experiment began, a board-certified surgeon (KMT) instructed the trainees on hand ties as part of the elective course. All trainees were taught to hold the excess suture in their left hand and use their right hand to pass and release the free end of the suture. Verbal instructions were illustrated with a PowerPoint presentation that included an overview of hand ties and knots; a video of a two-handed, hand-tied loop starting with a surgeon’s throw and followed by three square throws (a 2=1=1=1 configuration (Muffly et al., 2010)); and then step-by-step instruction with word descriptions, photos, and videos of each step. During the step-by-step instruction, trainees performed each specific step as the video segment played, then paused until the next step was introduced. At the end of the presentation, the full video was played as a repeated loop, and the trainees practiced hand ties for 10 min using the provided string. The instructor was available for input during this time, but proficiency was not judged at this point. The trainees then cut and removed the string knots from their pipes.

Knot tying

Trainees used the prepared 18-inch-long sutures to hand tie twenty 2=1=1=1 knotted loops (one knotted loop per suture), starting on the left side of the pipe and proceeding sequentially to the right. Trainees were asked to cut the suture ends, leaving them at least 5 mm in length. Trainees were given up to 50 min to complete the task. Once the trainee had completed the knotted loops, the rope ends were untied from the work surface and tied together over the top of the pipe and pipe insulation to prevent loss of any knotted suture loops. Two surgeons (KMT and PYM) also hand-tied twenty 2=1=1=1 knotted loops using the same materials as the trainees, except that the pipes were secured to a work bench that was at standing height, since both surgeons routinely performed surgery while standing.

Informed consent

Once tying was complete, trainees were asked if they would be willing to submit their materials for a study testing the strength and security of hand-tied, knotted loops. Trainees were informed participation was voluntary and would have no effect on grade and that trainee names would not be associated with pipe numbers during testing. Trainees who were interested in participating placed their secured pipes in a plastic container that was sealed and stored in a temperature-controlled room until testing. Subsequent to course grade submission, an e-mail survey detailing the project and requesting electronic consent was sent to each trainee who had submitted materials for testing so that trainees could privately confirm or deny participation in the study.

Testing

Testing was performed with a mechanical testing machine (Instron 5965 dual column table top AVE Extensometer model 2663-901, Instron, Norwood, Massachusetts) consistent with ASTM standard D2256 (Yalcin, 2021). To limit environmental and operator effects, all mechanical testing was performed on the same day by the same investigator (AGD). Before it was tested, each knotted loop was carefully slid off the pipe without handling the knots or pulling on suture ends. For each specimen, the knotted loop was positioned around two mechanical arms with the knot situated to the right, halfway between the two arms. The loop was gently stretched at five mm/ minute to a preload of 5 N of tension to prevent it from slipping or shifting. The change in loop length during knot consolidation, known as extension under preload, was recorded. The machine was zeroed at the new length and tension, and a single test to failure was then performed: the loop was stretched at a rate of 100 mm/minute until it opened. The maximum load and the maximum change in length (extension) before loop failure were recorded.

For statistical purposes, each loop was categorized as one of four types based on its mode of failure: 1. Broke at knot—the suture broke at or near the knot without any slippage of the knot or loosening of the loop; 2. Slipped then broke—the knot slipped during testing as the throws tightened down to each other, but tension/load remained at ≥ 60% of maximum and the loop broke under tension; 3. Slipped—a portion of the knot remained intact, but the loop loosened enough to produce a decrease in tension/load of ≥40%, ending the test; and 4. Untied—the knot completely untied and the loop opened, ending the test (Maturello et al., 2014; Avoine et al., 2016; Drabble et al., 2021). Mode of failure was based on the shape of the load–displacement curves (Figs. 1, 2, 3 and 4) and visual inspection of knotted loops (Avoine et al., 2016). Data from any sutures that failed during stretching under preload would be excluded from initial analysis. Any preloaded, knotted loops that failed when ≤5N of load was applied during single test to failure were categorized as “dangerous” (at risk for failure in a clinical situation) based on previous studies (Ching et al., 2013; Drabble et al., 2021; Ind, Shelton & Shepherd, 2001). For statistical purposes, knots also were categorized as secure (broke at knot or slipped then broke) or insecure (slipped or untied) based on the type of failure (Babetty, Sumer & Altintas, 1998; Wasik, Cross & Voss, 2013). A previously published power analysis indicated 10 knotted loops were required in each arm to detect a 5 N difference and a minimum of 17 knotted loops were needed to detect a temporal difference while maintaining a power of 80% with a type I error rate of 5% (Muffly et al., 2012; Ind, Shelton & Shepherd, 2001).

The difference between surgeons and trainees as well as the differences among modes of failure on extension at preload, maximum extension at failure, and maximum load to failure of knotted loops were evaluated using mixed model analysis for randomized block design. Diagnostic analysis on residuals was conducted to verify the normality and equal variance assumptions using Shapiro–Wilk test and Levene’s test. The difference between surgeons and trainees in the number of each type of knot tied, as well as the differences among mode of failure depending on the temporal group (e.g., first and last ten loops; or first, second, third, and fourth set of five loops), were analyzed using Poisson regression. Post hoc multiple comparisons were performed with Tukey’s adjustment. Statistical significance was identified at the level of 0.05. All analyses were conducted in SAS 9.4 TS1M7 (SAS institute Inc., Cary, NC).

Modes of failure

All 640 knotted loops tolerated a preload of 5 N; thus, none were excluded from data analysis. Based on the modes of failure (Tables 1 and 2), trainees tied 140 secure knots (23.3%) and 460 insecure knots (76.6%), and surgeons tied 11 secure knots (27.5%) and 29 insecure knots (72.5%).

When all knotted loops were compared as two temporal groups (first and last ten knotted loops) or four temporal groups (first, second, third, and fourth set of five knotted loops), there were no significant differences in the numbers of secure and insecure knots or in the modes of failure of temporal groups. Within temporal groups there were no significant differences in modes of failure or numbers of secure and insecure knots for surgeons and trainees (p > 0.44 for all values).

Results of mechanical testing

With single test to failure, knots of 54 loops were classified as dangerous, including 53/600 (8.8%) of trainee knotted loops and 1/40 (2.5%) surgeon knotted loops (p = 0.16). All dangerous knots failed by untying. For trainees, 29 dangerous knots occurred in the first ten knotted loops, and 24 occurred in the second ten (p = 0.47).

Overall, knotted loops tied by surgeons had less extension under preload than those tied by trainees (p = 0.0269). Among trainee and surgeon groups, the amount by which the loop extended under preload was not associated with the mode of failure of the loop (e.g., breaking, slipping, or untying; p > 0.89), its knot category (secure, insecure, or dangerous; p ≥ 0.88) or whether the specimen was one of the first ten or last ten loops tied (p = 0.2738).

Overall, there was no difference between surgeons and trainees in terms of the maximum amount that loops extended before failure (p = 0.3379) or the maximum load measured before the loops failed (p = 0.8827). Mode of failure did have an effect on maximum extension and maximum load, however (Tables 1 and 2). Among both groups of participants, loops tied with secure knots tolerated greater loads (p < 0.0001) and extended more (p ≤ 0.0331) before failure than those tied with insecure knots. For both groups, loops that broke at the knot tolerated the greatest load before failing (p ≤ 0.0104), and those that untied had the greatest amount of extension before failing (p ≤ 0.0264).