Friction and interfacial mechanics of surgical suturing: needle–tissue penetration, suture drag and friction-control strategies

1 Introduction

Surgical suturing is commonly described as a method of approximating divided tissues, but mechanically it is a sequence of distinct contact events rather than a single action. The surgeon first drives the needle towards the tissue, producing local deformation before the needle tip penetrates and cuts through it. The curved needle is then advanced along its path, the attached thread is pulled through the newly created channel, the wound edges are brought together, and the suture is finally tightened and secured with a knot (Oppici et al., 2023; Ostrander et al., 2024). Studies of suturing performance have therefore treated needle insertion, suture pulling and knot tying as separate subtasks, each requiring controlled motion and force application (Oppici et al., 2023; Ostrander et al., 2024; Dubrowski et al., 2005). Measurements of hand kinematics and tissue reaction forces have further shown that suturing skill is reflected not only in the final appearance of the closure, but also in how movement and load are transmitted to the tissue during needle passage (Dubrowski et al., 2005).

These stages create three principal tribological interfaces: needle–tissue, suture–tissue, and suture–suture contact. At the needle–tissue interface, the applied force must overcome tissue deformation, cutting resistance and friction along the needle surface. Once the needle has passed, the suture slides against the walls of the tissue tract, producing drag that depends on the thread material, diameter, construction and surface characteristics. During knot formation, contact shifts mainly to adjacent suture strands, where friction is required to prevent sliding and preserve the tension used to approximate the wound. This interpretation places surgical suturing within biotribology, which examines friction, wear and lubrication at interfaces involving biological tissues and medical devices (Zhang et al., 2022).

Friction in suturing may therefore be stage-dependent rather than uniformly beneficial or undesirable. Lower resistance during tissue passage may improve handling and reduce local mechanical disturbance. Monofilament sutures generally pass more easily through tissue, whereas braided or multifilament materials tend to provide greater flexibility and favourable handling while generating higher tissue friction and microtrauma (Byrne and Aly, 2019; Tajirian and Goldberg, 2010; Greenberg and Clark, 2009). To manage this trade-off, surface engineering strategies are employed to tune interfacial properties. Clinical handling itself is multidimensional and includes ease of passage, first-throw holding, smooth knot tie-down, knot security, memory and resistance to fraying (Ford et al., 2005). A very smooth suture may reduce drag during passage, but may also require additional throws because its surface provides less resistance to knot slippage (Tajirian and Goldberg, 2010). Additional throws used to improve holding can increase knot bulk and leave more foreign material within the wound. Furthermore, knot security was shown to depend on a combination of suture material, size, surface properties, knot configuration and number of throws rather than friction alone (Greenberg and Clark, 2009; Wong and McGrouther, 2023).

The consequences of these interactions arise from several mechanical, material and procedural factors. Larger sutures introduce more foreign material and may increase tissue reaction. In a canine abdominal-wall model, larger-diameter polyglycolic acid sutures produced greater inflammatory responses, while the smallest tested size still provided sufficient mechanical stability for the investigated closure (Li S. et al., 2023). In 1,000 plastic-surgery outpatients, patient characteristics, wound location and length, and surgeon experience contributed more strongly to local complications than differences among the evaluated suture materials or techniques (Gabrielli et al., 2001). Tribological behaviour is therefore best considered as one mechanical contributor within a broader procedural setting.

Recent developments in medical robotics could provide a useful engineering context for force-aware suturing. Efendi (Efendi, 2026) discussed how vision-guided robotic systems combine object localisation and robotic-arm coordination with force, tactile and position sensing to improve manipulation in medical environments. Similar approaches might also be useful when the object being handled is deformable biological tissue rather than a rigid medical object. Saputra et al. (Saputra et al., 2026) reviewed soft robotic grippers for surgical applications and highlighted features such as compliant materials, variable stiffness and integrated sensing, which could help distribute contact forces and adapt to irregular tissue shapes. These developments could support future robotic suturing systems, where needle motion, tissue manipulation and suture passage all depend on controlled interaction with soft tissue. Combining visual, force and tactile information might therefore help make these interactions more adaptive and better controlled. Robotic suturing is therefore not considered a tribological topic in itself. Studies that directly measure or model needle–tissue or suture–tissue interaction forces are treated as directly relevant mechanical evidence, whereas studies addressing overall suturing performance, trajectory planning, skill assessment, force feedback or general tissue manipulation are included as supporting engineering evidence where they provide context for force regulation and tissue interaction.

1.1 Relationship to previous reviews and unique contribution of the present review

Previous reviews have addressed several components relevant to surgical suturing, but largely from separate mechanical, material or clinical perspectives. Van Gerwen et al. (van Gerwen et al., 2012) reviewed experimental needle–tissue interaction forces and the effects of insertion method, needle characteristics and tissue properties, but the scope extended broadly to needle insertion rather than the complete suturing process. Li et al. (Li Y. et al., 2023) reviewed surgical sutures mainly from the perspective of materials, fibre structure, fabrication, surface modification and functionalisation. Knot mechanics have been considered separately through biomechanical studies of knot security (Wong and McGrouther, 2023), while other reviews have focused on force feedback and autonomous suturing (Oppici et al., 2023; Ostrander et al., 2024) or the broader biotribology of medical devices (Zhang et al., 2022). Thus, although friction and mechanical interaction appear within these reviews, the different interfaces involved in a complete stitch have generally been considered separately.

The present review instead considers suturing as a sequence of connected interfacial events. Its central framework is that the desired frictional behaviour could change as the stitch progresses: relatively low resistance could be beneficial during needle penetration and suture pull-through, controlled sliding might be required while the knot is advanced, and sufficient suture–suture friction could then be necessary to resist slippage after tightening. By linking the needle–tissue, suture–tissue and suture–suture interfaces within this stage-dependent framework, the review brings together penetration mechanics, tissue drag, surface engineering, knot mechanics, experimental and computational assessment, and force-aware robotic suturing. The contribution is therefore not simply a review of suture materials or individual friction measurements, but a mechanical interpretation of how interfacial requirements change across the complete suturing process. Figure 1 summarises the stage-dependent interfacial mechanics considered in the present review, highlighting how the dominant contact mechanisms and desirable frictional behaviour may change from needle penetration to suture passage, knot advancement and final knot locking.

2 Review methodology

2.1 Literature search and study selection

This work was conducted as a structured narrative review of the tribology and biomechanics of surgical suturing. The literature search was carried out in PubMed, Scopus and Web of Science Core Collection. The search covered studies published from January 2000 to July 2026. Reference lists of relevant original articles and review papers were also examined to identify additional studies. During manuscript revision, the literature was checked again for relevant recent publications, including studies recommended during peer review.

The search was organised around the main mechanical stages and interfaces involved in suturing. Search terms included combinations of “surgical suture”, “suturing” and “suture needle” with “tribology”, “friction”, “drag”, “pull through”, “sliding”, “penetration”, “puncture”, “needle tissue interaction”, “suture tissue interaction”, “suture interaction”, “knot security”, “loop security”, “surface coating”, “wear”, “force feedback”, “robotic suturing”, “robotic surgery”, “tissue manipulation”, “finite element” and “computational modelling”. Boolean operators were used to combine the terms and the syntax was adjusted according to the requirements of each database.

Titles and abstracts were screened for relevance to the defined scope, followed by full text assessment where necessary. Studies were included as core evidence when they investigated a mechanical or interfacial aspect directly relevant to surgical suturing. This included needle penetration and cutting forces, tissue deformation, needle tissue friction, suture pull through and tissue drag, suture surface properties, friction controlling coatings, suture-suture friction, knot security, tissue damage during suture passage, and force measurement or modelling during manual or robotic suturing. Experimental studies using biological tissues, ex vivo specimens, animal models, synthetic tissue substitutes and controlled tribological systems were considered together with relevant computational and robotic studies.

The scope was defined by the mechanical relevance of each study to the suturing process rather than by study type alone. Studies dealing primarily with injection, biopsy or other non-suturing needle applications were excluded unless the mechanical approach was directly applicable to a specific component of surgical suturing. Clinical findings were considered where they could be related to handling, tissue loading, tissue damage or knot behaviour, whereas studies concerned only with tensile strength, degradation, antimicrobial activity or clinical outcomes without a relevant mechanical or interfacial component were not treated as primary tribological evidence. Computational studies were included where they provided estimates of tissue deformation, local stress or contact behaviour that may be difficult to measure directly during suturing, or helped examine the effects of material, loading and contact assumptions. Broader studies on surgical robotics, robotic manipulation and soft tissue handling were included as supporting literature when force sensing, force control or tissue manipulation provided relevant engineering context for the mechanical interaction between the instrument, suture and tissue. Staples, tissue adhesives and implant-related tribology were outside the scope of the review. These studies were treated as supporting mechanical evidence rather than as primary tribological studies.

2.2 Synthesis

The included literature covered experimental, animal, ex vivo, computational, clinical and robotic studies with substantially different designs and outcome measures. Therefore, the studies were appraised according to design specific methodological factors rather than using a single numerical quality score. Experimental studies were considered with respect to the tissue or substitute used, needle and suture characteristics, contact geometry, hydration, loading and velocity conditions, force measurement method, repeated measurements and relevant comparison conditions. Computational studies were considered according to material representation, loading and contact assumptions and the reported level of experimental validation. Clinical and robotic studies were considered according to the suturing task, comparison groups or conditions, force measurement approach and reported mechanical outcomes.

Literature was synthesised according to the main tribological interface or mechanical process investigated. Numerical values were compared directly only where the testing configurations and outcome measures were reasonably similar. Where tissue type, hydration, contact geometry, loading, velocity or measurement method differed substantially, the reported values were treated as study-specific and used mainly to identify mechanical trends rather than as directly interchangeable absolute values. A meta-analysis was not performed because of the heterogeneity of the study designs, tissue models, materials and reported outcomes.

3 Needle–tissue penetration mechanics

Needle penetration begins before the tissue surface is broken. Initial contact produces local indentation and viscoelastic deformation, causing the reaction force to rise until a critical puncture point is reached. Bao et al. (Bao et al., 2016) separated this response into a non-rupture phase, dominated by tissue stiffness, and a rupture phase involving relative needle–tissue motion. Frick et al. (Frick et al., 2001) observed a similar force build-up in sheep skin and tendon; skin produced two peaks corresponding to entry through the upper surface and emergence from the underside. Increasing tissue pre-tension raised penetration resistance, whereas varying insertion speed from 1 to 10 mm/s had no significant effect. Bao et al. (Bao et al., 2016) reported that faster insertion reduced the force and time required to puncture the liver, indicating that rate effects depend strongly on tissue type and test conditions.

For clarity, needle insertion force and friction force are treated as different mechanical quantities in this review. Needle insertion force refers to the total force measured during advancement of the needle through tissue. Depending on the stage of insertion and the needle trajectory, this response may include tissue deformation, cutting or fracture at the tip, shaft–tissue friction and additional tissue compression. Friction force refers specifically to the resistance associated with relative sliding at the needle–tissue interface and therefore represents only one component of the total insertion force. After needle passage, suture pull-through force or tissue drag is considered separately and may include interfacial friction, local tissue deformation, ploughing and abrasion around the moving thread.

Following the puncture, the measured load represents several mechanisms. Jackson and Çavuşoğlu (Jackson and Çavuşoǧlu, 2012) modelled the interaction of a curved suture needle as the combined effect of cutting at the tip, friction along the embedded shaft and tissue compression generated when the needle does not follow its natural curvature. The compression term was related to the area swept by the needle during non-ideal motion. For straight-needle insertion, (Okamura et al., 2004) similarly divided resistance into capsule stiffness, shaft friction and cutting, treating the cutting component as approximately constant for a given tissue. Together, these models provide a basis for interpreting total insertion force in terms of tip condition, interfacial friction and driving trajectory. Accordingly, total needle insertion force should not be used interchangeably with friction force unless the frictional component has been separately measured or estimated.

Needle geometry and operating conditions further alter the mechanical response. Larger needles and triangular cutting tips produced greater puncture forces than smaller, round-bodied needles, while perpendicular entry gave the lowest resistance among the angles (Bao et al., 2016). (Jiang et al., 2014) also found that diameter, tip shape, insertion velocity and tissue structure influenced the force profile; skin and muscle resisted insertion more strongly than fat, and vascular structures introduced marked fluctuations. Needle size and shape should therefore be considered together with bite depth and tissue architecture. An optimisation framework incorporating needle diameter, arc shape, entry and exit points, and path constraints generated feasible trajectories for robotic suturing (Pedram et al., 2017). Supporting experiments with passive curved needles showed that curvature-following insertion, compared with linear advancement, reduced average axial force by 45.87%, transverse force by 55.76% and bending moment by 57.36% in tissue-mimicking gel (Kim and Hutapea, 2025). Robotic execution also depends on grasp position, which influences dexterity, torque demand and the need for regrasping (Li et al., 2015).

Force-aware planning provides a means of controlling these interactions. Jackson and Çavuşoğlu (Jackson and Cavusoglu, 2013) compared robotic trajectories intended to follow the needle arc and minimise unnecessary tissue displacement. Their later work used an unscented Kalman filter to estimate deformation- and friction-related force states during needle passage, creating a basis for trajectory adjustment during execution (Jackson et al., 2016). (Zhong et al., 2019) addressed the needle-induced movement of the intended exit point by coordinating one robotic arm for insertion and another for active tissue manipulation, while online estimation of deformation improved target reachability.

Sensing is also relevant to manual and surgeon-controlled robotic suturing. Kil et al. (Kil et al., 2022) combined force, torque, motion and video data, finding that tangential and orthogonal force measures helped distinguish experienced surgeons from residents during open suturing. In robotic model tasks, contact-force feedback reduced median maximum force from 3.4 to 2.8 N without reducing the suturing score (Yamasaki et al., 2024). Abdominal-wall experiments further showed that needle size and device design affected forceps pressure, traction and pull-through time (Börner et al., 2025). In ex vivo bladder and aortic suturing, novice surgeons using force-feedback instruments applied lower mean forces, made fewer errors and completed the tasks faster (Servais et al., 2025).

4 Suture–tissue friction and tissue drag

Once the needle has created a passage, the attached thread slides against a newly damaged and deformable tissue surface. The resulting resistance includes adhesion at the interface, viscoelastic deformation around the suture and, particularly for rough multifilament threads, local ploughing or abrasion. Capstan experiments provide controlled information on sliding against artificial skin and indicated a dominant adhesive contribution (Zhang et al., 2017a). Suture pull-through measurements using the penetration friction apparatus complement this configuration by incorporating circumferential tissue compression, interfacial shear, sub-surface deformation and the radial normal force exerted around the moving thread (Zhang et al., 2018a; Zhang et al., 2018b).

Suture construction influences both the magnitude and stability of friction. In capstan tests, the surface roughness was 4.25 ± 1.01 µm for nylon monofilament, 35.13 ± 5.31 µm for braided silk and 14.98 ± 2.84 µm for coated Vicryl (Zhang et al., 2017a). Despite its smoother surface, nylon produced a higher apparent coefficient of friction than coated Vicryl in that contact arrangement, showing that roughness alone does not determine sliding behaviour. In soft-material tribology, smoother monofilaments increase the real contact area with flat substrates, thereby increasing adhesive friction. In contrast, PFA pull-through testing produced a mean drag force of 0.25 N for nylon and 0.42 N for PGA multifilament of comparable size, a reduction of approximately 40% for the monofilament. The PGA surface was also considerably rougher than nylon, at 20.0 ± 0.5 and 6.8 ± 0.3 µm, respectively (Zhang et al., 2018a). These findings indicate that multifilament architecture becomes particularly important when individual strands can engage, compress or scrape the walls of the tissue channel.

Experiments using porcine muscle, liver and artificial skin further showed that uncoated braided silk consistently generated more friction than monofilament polypropylene or coated braided Vicryl. Relative to silk, polypropylene reduced the measured coefficient by 5%–59%, while the Vicryl coating reduced it by 1.5%–34%, depending on the tissue and operating condition (Zhang et al., 2017b). A clinical comparison involving ten oral-surgery patients also rated polypropylene more favourably than silk for intraoperative handling, with mean visual analogue scores of 96.4 ± 4.0 and 60.0 ± 17.2 mm, respectively (Dragović et al., 2018). This clinical handling assessment complements the quantitative friction measurements and supports the practical relevance of construction-dependent drag. Conjugate spinning provides another approach, retaining the smooth exterior of a monofilament while reducing stiffness and improving handling and knot performance (Jung et al., 2007).

Diameter increases drag by producing greater tissue displacement and normal force. When the PGA multifilament diameter increased from 0.19 to 0.55 mm, the measured suture pull-through force increased from approximately 0.35–1.15 N (Zhang et al., 2018b). However, friction did not increase proportionally with the estimated normal force: using a uniaxial deformation model, the friction-to-normal-force ratio declined from 2.29 to 0.42 as the normal force increased from 0.31 to 2.56 N. Suture architecture is equally important. For braided silk, suture–suture friction ranged from approximately 11.5–18.1 cN across different twist levels, braid angles and patterns. The highest value occurred with six twists per inch, a 34.8° braid angle and a 1/1 pattern, whereas a 3.7-twist, 28.8°, 1/2 construction produced the targeted 15 cN handling response (Sivanesan et al., 2022).

Operational conditions further modify tissue drag. Increasing normal load from 0.1 to 0.5 N reduced the coefficient of friction by 36% in one suture–artificial-skin pair, while increasing velocity from 5 to 25 mm/s raised it by 26% (Zhang et al., 2017b). Hydration was contact-dependent. Blood and interstitial fluid in porcine tissues reduced adhesion relative to dry artificial skin (Zhang et al., 2017b), whereas fetal bovine serum increased the apparent coefficient for Vicryl from 0.40 to 0.77 in capstan testing. Serum increased friction by 17.8% for nylon, 42.7% for silk and 88.9% for Vicryl, but simultaneously reduced the amplitude of stick–slip fluctuations (Zhang et al., 2017a).

Overall, these findings indicate that suture–tissue friction could arise from the combined effects of surface topography, real contact area, material chemistry, tissue deformation and the interfacial fluid environment rather than from roughness alone. A rough multifilament surface could increase local engagement, ploughing and abrasion during penetration, whereas a smoother monofilament might still generate appreciable adhesive friction if a relatively large real contact area develops (Zhang et al., 2017a; Zhang et al., 2018a). Tissue compliance could further influence this contact by changing the deformation and normal force around the moving suture, while differences in tissue composition and hydration might alter adhesion and interfacial shear (Zhang et al., 2017a; Zhang et al., 2017b). Surface chemistry and coatings could modify wettability, fluid interaction and adhesion, although changes in chemistry might have only a limited effect on friction when the underlying roughness and contact area remain similar (Zhang et al., 2020a; Zhang et al., 2020b). Similarly, biological fluids might reduce direct adhesion in some tissue contacts but could increase resistance in others through suture swelling, surface adsorption or formation of a viscous interfacial film (Zhang et al., 2017a; Zhang et al., 2017b). The relative contribution of these mechanisms might therefore depend on the particular combination of suture structure, surface condition, tissue properties and lubrication state.

Irregular sticking and release may transfer fluctuating forces to tissue. After PGA passage, the damaged area at the entry boundary was 0.72 mm² compared with 0.29 mm² at the exit, and multifilament sutures retained more tissue debris than nylon (Zhang et al., 2018a). Newly developed nanoyarns illustrate how construction can reduce this interaction: porcine-skin drag averaged 13 ± 15 mN for a single PCL nanoyarn and 28 ± 26 mN after braiding, compared with 63 ± 10 mN for commercial PDO, with the reported values considered alongside the different sample diameters (Hassinger et al., 2026). Damage may also arise through cutting under sustained tension. In a rabbit model, the greatest loss of loop tension occurred within the first minute and correlated inversely with tissue collagen content (Klink et al., 2011). Elastic TPU sutures reduced relative tension loss to 0.24–0.38 cN, compared with 6.78–8.76 cN for PVDF threads (Helmedag et al., 2021). Barbed sutures distributed traction differently and produced less visible myocardial damage than conventional closures in SEM observations (Shiga et al., 2024). In the tendon, anchorage increased from 5.4 N at 1 cm to 7.45 N at 2 cm and then reached a plateau at longer purchase lengths (Gemalmaz, 2019).

5 Friction-control and surface-engineering strategies

Surface treatment is used to reduce resistance during needle penetration or suture passage without sacrificing knot stability. A simple approach is reported to apply a temporary lubricant immediately before use. (Walling et al., 2005) passed braided polyglactin through antibiotic ointment and reported smoother tissue passage with less drag. This practical technique highlights the potential value of intraoperative lubrication, with application consistency and lubricant retention representing useful parameters for future quantitative assessment.

Commercial braided sutures commonly use polymeric lubricants to fill spaces between filaments and smooth the outer surface. Storch et al. (Storch et al., 2002) compared coated polyglactin 910 with and without incorporated triclosan and found no significant differences in ease of passage, tie-down smoothness, first-throw holding or knot security. Thus, antimicrobial functionality was added without measurably disrupting the existing tribological balance of the polyglactin/calcium-stearate coating.

Needle coatings have produced clearer reductions in interfacial resistance. A plasma-polymerised aminosilane coating applied to curved suture needles for 5 minutes lowered the measured needle penetration force by up to 35% compared with conventional silicone-coated controls (Muro-Fraguas et al., 2026). Longer plasma exposure improved coating continuity and adhesion, and the layer remained stable during repeated insertions. In a related but non-suturing-specific study, a polydopamine–PTFE–activated-carbon coating reduced maximum insertion and extraction forces by 62% and 64% in a PVC phantom and by 49% and 30% in bovine kidney. Atomic-force microscopy indicated a 65.7% reduction in surface friction (Patel et al., 2021). These insertion and extraction forces represent the overall resistance measured during needle motion, whereas the AFM measurement in (Patel et al., 2021) separately characterised surface friction.

For threads, chitosan coating reduced the dynamic coefficient of friction of braided silk across normal loads of 50–150 g and sliding speeds of 50–150 mm/min. Increasing chitosan concentration also produced modest improvements in tenacity and knot strength (Viju and Thilagavathi, 2013). For PGA multifilament sutures, dopamine and carboxymethyl-chitosan treatments produced coefficients of friction of 1.37 and 1.47, respectively, compared with 1.38 before treatment. The corresponding friction forces were approximately 0.75, 0.80 and 0.75 N (Zhang et al., 2020a). The study therefore showed that the coating primarily modified surface coverage and filament interstices while maintaining friction within a similar range.

A dopamine–graphene oxide coating produced a comparable frictional response. The measured coefficients were 1.38 for untreated, 1.37 for dopamine-coated and 1.42 for graphene-oxide-coated sutures, while the respective friction forces remained close to 0.72–0.78 N (Zhang et al., 2020b). Coating uniformity, wettability and mechanical strength improved, with the penetration results indicating that these multifunctional benefits were achieved while retaining a similar level of tissue-drag resistance.

Antibacterial modification also influences handling-related properties. In coated braided silk, antibacterial processing changed static and dynamic suture–suture friction by less than 16.1% and 32.8%, respectively, and increased tissue-drag resistance and bending stiffness by approximately 50% (Chen et al., 2015a). A PCL–levofloxacin coating similarly showed that the processing sequence determined whether filament interstices remained exposed or were filled. Coating after braiding produced a smoother, more monofilament-like surface and lower sliding resistance, while the accompanying reduction in thread–thread friction emphasised the importance of considering first-throw holding (Chen et al., 2015b). Antimicrobial monofilament sutures coated with PCL, calcium stearate and levofloxacin retained a relatively smooth surface and adequate knot strength, providing a basis for future tissue-drag characterisation (Sriyai et al., 2021).

Other polymeric coatings have primarily targeted degradation and biological performance. PLLA- and PCL-coated PDO sutures developed smooth, uniform surfaces and retained strength for longer in vivo (Wu et al., 2022). Bioactive-glass/graphene-oxide coatings were similarly evaluated for coating stability, knot performance and regenerative potential (Öksüz et al., 2023). These studies broaden the role of surface engineering beyond drag reduction and provide suitable platforms for future pull-through and suture–tissue friction measurements.

An alternative strategy is to alter the interface after insertion. Light-activated tissue-integrating sutures produced a baseline drag of approximately 38 kPa, comparable to commercial PGA. After laser activation, substantially greater force was required to remove them from the skin than PGA or silk sutures, indicating increased post-placement tissue anchorage rather than permanently high insertion friction (Ghosh et al., 2019). This approach illustrates the desired tribological sequence: low resistance during passage, followed by controlled interfacial fixation after positioning.

Overall, no single coating strategy could be considered uniformly advantageous because the desired effect depends on whether the target is reduced penetration resistance, lower tissue drag, improved biological function or post-placement stability. Low-friction needle coatings could provide relatively clear reductions in insertion resistance, with plasma-based coatings also showing encouraging stability during repeated penetration (Muro-Fraguas et al., 2026; Patel et al., 2021). For suture threads, chitosan could reduce sliding resistance while maintaining or modestly improving mechanical properties (Viju and Thilagavathi, 2013), whereas dopamine-, carboxymethyl-chitosan- and graphene-based coatings produced relatively small changes in friction despite modifying surface coverage, wettability or other functional properties (Zhang et al., 2020a; Zhang et al., 2020b). Antimicrobial and polymeric coatings could add biological or degradation-related benefits, but might also alter bending stiffness, tissue drag or suture–suture friction, which could influence handling and knot behaviour (Chen et al., 2015a; Chen et al., 2015b; Sriyai et al., 2021; Wu et al., 2022; Öksüz et al., 2023). Durability is also unevenly characterised across these strategies, as some coatings have been examined after repeated penetration or in vivo exposure, whereas others have mainly been assessed under short-term laboratory conditions (Muro-Fraguas et al., 2026; Wu et al., 2022; Öksüz et al., 2023). Repeated passage could progressively affect coating integrity and its friction-reducing function, and possible coating degradation or transfer to the tissue interface should therefore be evaluated alongside mechanical performance. The most suitable coating should not only reduce penetration or tissue resistance, but also preserve coating integrity, flexibility, mechanical strength and knot security without introducing undesirable biological effects.

6 Friction trade-off with knot security

A surgical knot does not become secure merely because adjacent strands are brought into contact. Tightening generates normal pressure where the filaments cross, thereby allowing suture–suture friction to resist subsequent sliding. In polypropylene monofilaments, Johanns et al. measured a dynamic friction coefficient of 0.20 ± 0.02 and showed that knot strength depended on the combined effects of friction, tightening pre-tension and permanent filament deformation (Johanns et al., 2023). Plastic bending retained the constricted knot geometry after the free ends were released, whereas insufficient friction allowed elastic recovery and knot dilation. In this system, permanent deformation therefore contributed to knot stabilisation, but should not be interpreted as a universal requirement for frictional locking. For tight knots, strength increased approximately in proportion to the number of throws and friction coefficient and followed a power-law dependence on pre-tension with an exponent of 1.56 ± 0.23. The analysis was conducted under dry conditions, and lower resistance would be expected in the presence of biological fluids.

Knot security is distinct from loop security. The former describes resistance to slippage after the knot is completed, whereas the latter concerns maintenance of tissue approximation while successive throws are placed. Sliding arthroscopic knots showed lower security before reinforcement, and adding three reversing half-hitches on alternating posts markedly reduced slippage. With Ethibond, this shifted the main failure process from knot movement towards filament elongation (Lo et al., 2004). Material strength did not directly predict knot behaviour: unknotted FiberWire failed at 250.0 ± 17.8 N compared with 128 ± 3.8 N for Ethibond, yet FiberWire knots continued to slip at considerably higher loads (Lo et al., 2004).

Tightening force interacts with the knot configuration to determine strand contact and locking. Surgeons applied tying forces from 3 to 63 N in one study (Neuhofer et al., 2014). When square knots were tightened at 50 N, mean slippage was 6 ± 2.9 mm for FiberWire, 3 ± 1.5 mm for PDS and 4 ± 0.9 mm for Vicryl. Five-throw FiberWire knots resolved completely at a mean load of 222 ± 24.9 N, indicating that six or seven throws provided greater security for this low-sliding-resistance material (Neuhofer et al., 2014). Additional initial twists can similarly increase the contact length and friction of a surgeon’s knot. Modiri et al. (Modiri et al., 2025) observed improving holding resistance up to approximately five twists, followed by a plateau associated with increasing knot bulk and geometric variability.

The required number of throws is material- and configuration-dependent. Three-throw nonidentical sliding knots were untied substantially more frequently than six-throw knots in both braided polyglactin and monofilament polydioxanone (Ivy et al., 2004). In a separate series of 225 square knots, four to six throws produced greater failure loads than three, while five throws provided the best tested compromise between security and knot volume (Muffly et al., 2011). Similarly, it was found that square and surgeon’s knots generally required four throws, whereas sliding knots required five; security values were comparable between five and six throws (Silver et al., 2016).

Suture composition changes the balance between frictional locking and mechanical strength. Across seven 4–0 materials, measured knot security ranged from 8.7 to 11.9 N, and synthetic monofilaments showed a greater tendency to untie than braided materials (Kim et al., 2007). (Türker et al., 2011) reported that No. 5 FiberWire had the highest failure load, 625 ± 30 N, but also the greatest slippage, 14 ± 1.9 mm. No. 2 Ethibond slipped only 0.1 ± 0.3 mm despite its lower tensile capacity. More recent testing confirmed that four-throw knots were generally dependable with polyfilament sutures, whereas monofilaments required at least five throws with multiple blocking half-hitches; 51% of the tested monofilament knots showed some opening (Romeo et al., 2023). Figure 2 shows the knot configurations investigated in this work.

Knot failure may also occur through irreversible filament extension rather than interfacial sliding. Prolene and PDS began plastically deforming at lower loads than braided Ethibond, Vicryl and FiberWire, and Prolene provided the least resistance to tendon-repair gapping (Jordan et al., 2016). In repaired cadaveric flexor tendons, knot unravelling accounted for 50%–88% of failures and occurred after initial repair-site gapping and failure of the peripheral stitch (Mistry et al., 2016). Frictional locking can prevent such progression. In a vascular model, adding one locking throw increased leakage pressure from 148.7 ± 109.4 to 332.7 ± 25.2 mmHg for a square throw and from 62.5 ± 46.2 to 323.0 ± 21.0 mmHg for a surgeon’s throw (Wood and Monnet, 2024).

The following studies provide supporting mechanical evidence rather than direct surgical-suturing evidence. Although they examine elastic rods, capstan configurations or general physical knots rather than surgical sutures in biological tissues, they provide useful mechanistic insight into contact pressure, tension transfer, frictional locking, filament deformation and knot geometry. Their findings are therefore used here to support interpretation of surgical knot mechanics rather than as direct evidence of clinical knot performance.

Recent studies from general filament and knot mechanics provide additional mechanistic insight into how contact develops as a knot is tightened. (Grandgeorge et al., 2021) showed that contact between two tightly crossing filaments is not associated with a uniform pressure distribution. Instead, the contact pressure can be strongly localised along curved contact regions, while increasing load can deform the filament cross-sections and redistribute the pressure. Similar behaviour was observed for an elastic rod wrapped around a rigid cylinder, where increasing load changed the contact from a localised point to an extended contact region. Bending stiffness, filament thickness and the evolving contact geometry therefore influence how force is transmitted through a wrapped filament (Grandgeorge et al., 2022). Recently, (Sabater et al., 2026) showed that the previously reported increase in sliding strength with tying tension does not require plastic deformation. Their experiments, finite-element simulations and elastica model indicated that this behaviour mainly arises from the simultaneous increase in normal contact force and contact length during tightening. Once the available contact length became fully developed, the response approached a more linear relation with tying tension. These findings distinguish the general frictional-contact mechanism from material-specific effects of permanent deformation. Increasing normal contact force and contact length during tightening can increase sliding resistance even in an elastic system, whereas plastic deformation, when present, can provide an additional mechanism for retaining the tightened geometry.

Sano et al. (2022) showed that tension was not uniform through a knot. In the clove hitch, they identified alternating contact regions with different mechanical roles. Tension decreased progressively in capstan-like filament–cylinder contacts, whereas much sharper tension drops occurred at self-contact or “nip” regions where one filament segment pinched another. A more general analytical treatment of knot friction showed that self-locking depends not only on the friction coefficient but also on the additional normal pressure created when one segment is compressed between neighbouring segments, together with the total contact curvature and knot topology (Leuthäusser, 2024). (Aymon et al., 2025) recently identified a similar localised locking unit in the bowline knot. Most of the tension loss occurred within this region, where one filament segment wrapped around two central segments, and the predicted stability was highly sensitive to friction. These findings suggest that knot locking is produced by specific high-contact regions that progressively reduce the transmitted tension, rather than by uniform friction acting throughout the whole knot.

Filament deformation can further determine whether a tightened knot remains stable or changes configuration under load. Johanns et al. (Johanns and Reis, 2024) showed that stopper-knot failure can occur through friction-induced torsional deformation rather than simple strand sliding. Increasing load generated local twisting at self-contact regions and eventually caused the knot to capsize, while tighter knots resisted this transition to higher loads because of the greater torsional stiffness of the shorter loop segments and higher contact forces. A recent slipknot study also showed that pre-tightening force and the number of loops alter the contact area and therefore the peak force required to release the knot. The release process occurred in two mechanical stages involving contact, friction and elastic deformation before a topological rearrangement caused rapid opening (Xue et al., 2025). Together, these observations show that knot security cannot be described by friction alone; bending, twisting, contact evolution and configuration changes can either strengthen the locking state or provide a pathway for knot release.

Overall, the combined surgical-suture evidence and supporting general knot-mechanics literature indicate that knot security develops through a coupled sequence of contact, tension transfer and filament deformation. Pre-tension increases contact length and normal pressure, while friction redistributes tension through wrapped and self-contacting regions. Local pinching and locking regions can produce large tension drops, whereas bending, plastic deformation or torsional deformation can either preserve the tightened geometry or promote geometric instability and capsizing. Knot security should therefore not be interpreted from the friction coefficient alone, but from its interaction with contact geometry, pre-tension, filament stiffness and deformation, and the particular locking topology of the knot.

At present, a universal or clinically validated optimal friction range could not be identified from the available literature. Reported friction values depend on suture material, surface structure, hydration, tightening force and knot configuration, and are not always expressed using the same mechanical quantity. For example, a dynamic friction coefficient of 0.20 ± 0.02 was reported for polypropylene under a specific dry knot-testing condition (Johanns et al., 2023), whereas braided silk showed suture–suture friction values of approximately 11.5–18.1 cN depending on its structural parameters (Sivanesan et al., 2022). These values could therefore be considered configuration-specific rather than general targets. A clinically meaningful friction range might instead need to be defined for a particular material–knot–tissue combination and linked to outcomes such as acceptable tissue drag, knot slippage and loop-tension retention.

The central tribological requirement is therefore selective friction: low resistance is desirable while the thread passes through tissue and the knot is advanced, but sufficient strand-to-strand friction must develop after tightening. Material selection, pre-tension and knot architecture must be considered together, because improving tissue glide without modifying the locking sequence can increase the risk of knot relaxation or complete unravelling.

7 Experimental and computational assessment of suturing tribology and its tissue implications

Experimental assessment of suturing encompasses two mechanically different events: penetration of tissue by the needle and subsequent sliding of the thread through the puncture tract. Conventional capstan and linear-sliding tests provide controlled comparison of suture surfaces under prescribed normal loads. The penetration friction apparatus complements these methods by measuring needle puncture and suture pull-through sequentially while allowing the skin substitute to deform around the moving suture (Zhang et al., 2017c). Friction values obtained from different test configurations should therefore be interpreted within their experimental context. The purpose of comparing these studies was not to establish a common absolute friction scale, but to examine how different methods characterise suturing-related friction and whether consistent mechanical trends emerge across them.

With braided PGA, the measured suture pull-through force remained near 0.4 N when the needle-to-suture diameter ratio exceeded 2.5, but increased towards 1.2 N at smaller ratios. Surface imaging also revealed adhered debris and a distinct worn track after passage, indicating that the measured pull-through resistance included contributions from interfacial sliding as well as deformation and progressive damage to the counterface (Zhang et al., 2017c). Table 1 compares representative experimental studies of suture–tissue friction, highlighting differences in test configuration, tissue model, loading conditions and reported friction metrics.

TABLE 1. Representative experimental studies characterising suture–tissue friction and related testing conditions.

Force platforms provide a complementary assessment of the loads transferred during complete suturing tasks. In a laparoscopic box trainer, novices generated maximum and mean forces of 4.7 ± 1.3 and 2.1 ± 0.6 N, respectively, compared with 2.6 ± 0.4 and 0.9 ± 0.3 N among experienced surgeons (Horeman et al., 2010). The SurgForce system extended this approach by calculating mean and peak force, force instability, momentum, threshold exceedances and force energy from the recorded signal. Nine of eleven parameters differed across novice, intermediate and expert groups (Montoya-Alvarez et al., 2020). Within this system, force energy was expressed in N²s as an accumulated force-signal metric, while displacement-based integration provides the corresponding mechanical work-to-failure measure. Technique also affects force transfer: when a standardised 5-N wound was closed, the mean maximum pulling force ranged from 3.46 ± 0.61 N for a pulley suture to 8.11 ± 1.00 N for a horizontal mattress configuration (Marsidi et al., 2020).

The selected tissue model strongly influences the measured response. Synthetic membranes and training pads offer repeatable conditions, while ex vivo tissues add anisotropy, hydration, heterogeneity and time-dependent behaviour. Experiments on eight porcine tissues found that the first visible damage occurred at approximately 11.43 N in fascia and 1.25 N in the fallopian tube; small bowel tolerated about 1.5 times the load of large bowel (Rodrigues et al., 2012). These macroscopic thresholds provide tissue-specific reference values for interpreting applied force. Species also influences pull-through behaviour. Pig and sheep flexor tendons resisted continued suture cheese-wiring more strongly than human tendons, with the differences becoming greater after 5 and 10 mm of splitting (Peltz et al., 2017). These results support model-specific interpretation when translating animal-tissue measurements to human repair. Accordingly, values obtained from silicone, artificial-skin or other synthetic models should primarily be interpreted as configuration-specific laboratory measurements for comparative and mechanistic assessment. They should not be considered clinically transferable friction, force or damage thresholds without validation in the relevant biological tissue under physiologically representative hydration, tension, loading and suturing conditions. Similarly, findings from animal or ex vivo tissues provide greater biological relevance but still require appropriate validation before being generalised to human clinical performance.

The functional value of interfacial resistance depends on the stage of suturing. A mesh suture designed to distribute load and encourage tissue ingrowth required 4.27 ± 0.42 N to pull through ballistic gel, compared with 2.23 ± 0.48 N for conventional polypropylene. In porcine abdominal walls, it also increased eight-day repair strength from 160 ± 56 to 320 ± 57 N and work to failure from 7.3 ± 3.7 to 24.6 ± 14.2 J (Dumanian et al., 2015). Greater post-placement engagement therefore supported repair strength, whereas low resistance remained desirable during placement. The relevant tribological objective is stage-dependent: smooth passage followed by adequate anchorage after tightening.

Within the robotic literature, direct interaction studies and broader performance studies provide different levels of mechanical evidence. Measurements or models of forces during needle–tissue interaction inform the interface mechanics directly, whereas overall task-force and performance measures are used here primarily to provide engineering context for force-aware suturing. Force sensing is particularly important in minimally invasive and robotic suturing, where direct tactile information is reduced. In robotic interrupted suturing, integrated feedback lowered mean force from 1.53 ± 0.55 N without feedback to 1.04 ± 0.27 N at the highest setting. Maximum force decreased by 55%, and 82% of surgeons achieved a reduction greater than 25% (Awad et al., 2024). These results demonstrate improved mechanical force regulation during the investigated suturing task; however, whether such reductions translate into less tissue damage or improved patient-level healing remains to be established.

Computational models provide access to contact stresses that are difficult to measure experimentally. An early three-dimensional skin model predicted that an S-shaped excision reduced postoperative extrusion by approximately 40% compared with a spindle-shaped closure (Yoshida et al., 2001). A multiscale Kessler tendon model incorporated orthotropic tissue behaviour and a suture–tendon friction coefficient of 0.005, predicted high-stress regions that corresponded to experimentally observed acellular tissue zones (Rawson et al., 2015). For facial suspension sutures, increasing the number of anchoring cones reduced local equivalent strain by about 20% and stress by approximately 60%, with cadaveric validation differences of up to 18% (Mousavi et al., 2023). More recent fixed-grid modelling represents individual sutures as discrete springs, allowing their number and position to be changed without remeshing and preserving local stress concentrations near insertion sites (Holm et al., 2024). Figure 3 shows the FE predictions of strain and stress during wound closure using different numbers of sutures. This spring-based representation is suited to evaluating wound-closure mechanics and suture placement. A robotic-suturing model explored needle forces up to 8 N and compared with experimental values of 0–1.4 N, illustrating how simulated force levels vary with material parameters, force location and contact assumptions (Vargas et al., 2025).

Repeated contact also changes the interacting surfaces. SEM analysis found deformation in 53.1% of 192 clinically used dental needles. No deformation was detected after one or two stitches, but the prevalence rose from 25% after three stitches to 95.8% after eight; triangular tips deformed more than tapered tips (Cumbo et al., 2025). After implantation, absorbable sutures likewise developed material-specific surface changes, including filament cleavage in polyglactin, crater formation in poliglecaprone and delayed cracking in polydioxanone (Andrade et al., 2006). The associated tissue response reflects the combined contributions of mechanical trauma, surface degradation and material chemistry. An independent cell study found early pro-inflammatory upregulation with Vicryl, Ti-Cron, FiberWire and silk-based Perma-Hand, whereas Monocryl, nylon and Ethibond were comparatively inert (Lock et al., 2017).

8 Research gaps and future directions

Future research in suturing tribology could be prioritised according to both scientific importance and practical feasibility. In the short term, standardised friction testing, controlled wet-tissue experiments and repeated-passage evaluation of coatings could be among the most immediately addressable areas because they can be investigated using existing experimental methods. Separation of cutting, deformation and friction components could also be advanced through combined experimental and computational studies, although reliable validation would remain important. Linking mechanical measurements with tissue damage and healing might represent a medium-term priority because it requires more complex biological models and longitudinal assessment. Adaptive robotic control could be a longer-term objective, as its development might depend on reliable force sensing, validated interaction models and tissue-specific mechanical limits established by the preceding studies. A priority for the field is possibly the development of standardised methods for measuring needle and suture friction. Figure 4 shows the principal research gaps and priority directions for improving the experimental reliability, biological relevance and clinical translation of tribology- informed surgical suturing.

Capstan, linear-sliding and penetration-friction tests use different contact geometries, loading conditions and output variables; harmonised reporting would improve direct comparison. Future protocols could specify tissue hydration, pre-tension, insertion angle, pulling velocity, needle-to-suture diameter ratio and the number of repeated passages. Reporting both peak puncture force and steady pull-through force would also help distinguish penetration from sliding behaviour (Zhang et al., 2017c).

Greater physiological realism is equally important. Straight insertion into flat silicone sheets provides a controlled experimental baseline, while surgical needles typically follow a curved path through hydrated and tensioned tissue. Experiments using fresh porcine cornea showed that circular-trajectory robotic suturing produced more stable force profiles and lower out-of-plane distortion torque than manual insertion (Yang et al., 2012). Similar models can be developed for skin, fascia, bowel and tendon, with controlled tissue tension and clinically representative entry–exit geometry.

Measured needle insertion force combines tissue deformation, cutting and shaft friction. Analytical and state-estimation models can separate these components conceptually (Jackson and Çavuşoǧlu, 2012; Jackson et al., 2016), and further direct experimental validation would strengthen component-level interpretation. Plasma and composite coatings reduce insertion resistance (Muro-Fraguas et al., 2026; Patel and Hutapea, 2024); repeated-passage studies can determine how long this benefit is retained and characterise any transfer of coating material to the tissue interface. Future studies should therefore report total insertion force and any experimentally isolated or model-estimated friction component separately.

Mechanical measurements could also be linked more directly with surgical outcomes. A PDA–PTFE–activated-carbon coating reduced maximum insertion force from 3.54 ± 0.052 to 1.82 ± 0.023 N and reduced histological tissue-damage area from 0.201 ± 0.063 to 0.121 ± 0.067 mm² (Patel and Hutapea, 2024). This straight-needle bovine-kidney model provides a useful basis for analogous studies using curved suturing needles and additional tissue types. Future work could correlate friction, surface damage, inflammation and healing in tissue-specific models. Until such relationships are established, laboratory friction and force values should be regarded as model-specific mechanical metrics rather than clinically validated target ranges.

Finally, needle–suture selection could integrate tissue strength with material properties, because visible damage thresholds vary markedly among tissues (Rodrigues et al., 2012). Robotic systems currently use force feedback to reduce overall loading (Awad et al., 2024). Future closed-loop control could estimate friction, cutting and deformation separately and adapt the insertion speed, trajectory or gripping force during each stitch.

9 Conclusion

Surgical suturing involves a sequence of tribological interactions rather than a single penetration event. Needle insertion combines tissue deformation, cutting and shaft friction, while subsequent thread passage generates drag through adhesion, viscoelastic deformation and, particularly for rough or multifilament sutures, local abrasion (Jackson and Çavuşoǧlu, 2012; Zhang et al., 2017a; Zhang et al., 2017c). During knot formation, friction becomes beneficial: tightening increases strand-to-strand contact pressure and contact length, allowing friction to resist slippage, while permanent filament deformation may provide additional stabilisation in materials where it occurs (Johanns et al., 2023; Sabater et al., 2026). Thus, friction should be controlled according to the stage of suturing rather than simply minimised.

For suture and needle design, the main requirement is low resistance during penetration and pull-through without compromising handling, coating integrity or knot security. Needle geometry, suture diameter, surface construction and tissue properties should be considered together, as the same combination may behave differently in skin, fascia, bowel or tendon (Bao et al., 2016; Rodrigues et al., 2012). Surface treatments can reduce insertion resistance, and their durability and safety should be evaluated after repeated passage. Where possible, mechanical improvements should also be linked to histological outcomes. Coated needles have shown that lower insertion force can coincide with reduced tissue damage (Patel and Hutapea, 2024). Future work should prioritise standardised testing in hydrated, tensioned and anatomically realistic tissues, experimental separation of cutting and friction forces, and the direct correlation of tribological metrics with healing. Force-measuring robotic systems should ultimately adapt needle trajectory, speed and gripping force in response to the changing tissue–instrument interface (Jackson et al., 2016; Awad et al., 2024).

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Hamed Rajabi, London South Bank University, United Kingdom

Reviewed by: Isak Karabegović, University of Bihać, Bosnia and Herzegovina

Chao Li, Peking University Third Hospital, China