Silk Fibroin Closure Eliminates the Incidence of Allergic Contact Dermatitis Compared to Cyanoacrylate Mesh in Total Joint Arthroplasty

Abstract

Background

Allergic contact dermatitis (ACD) from adhesive wound closure systems has garnered particular attention for its potential role in increasing wound complications in total joint arthroplasty (TJA). This study, performed at a high-volume orthopaedic specialty hospital, investigates the incidence of wound complications among 2 adhesive systems: a cyanoacrylate mesh (CM) adhesive and a silk fibroin (SF) adhesive.

Methods

All TJAs from January 2019 to April 2024 with at least 6 wks postoperative follow-up were retrospectively reviewed. Demographics and surgical outcomes were collected and analyzed. Statistical analyses were performed using Fisher’s exact tests and t-tests.

Results

A sample size of 170 CM and 85 SF subjects was calculated to achieve a power of 80%. Of the 257 patients identified (172 CM and 85 SF), 46.7% were females and 53.3% were males, with a mean age of 65.3 ± 9.0 years and a mean body mass index of 28.0 ± 4.6. Bivariate analyses revealed no significant differences in demographics or comorbidities between the CM and SF cohorts, except for frequency of American Society of Anesthesiologists 1 classification (3.9% vs 1.7%; P = .011). The CM cohort exhibited a significantly higher incidence of ACD (6.4% vs 0%; P = .018), while differences in all other clinical outcomes were nonsignificant.

Conclusions

There is a very low incidence of wound complications in TJA when using the SF adhesive for wound closure. There was a statistically significant increase in ACD when using the CM adhesive. SF adhesives appear to be a superior wound closure option to consider in patients undergoing TJA.

Introduction

Adverse events associated with surgical wound closure are of concern and can include, but are not limited to, allergic contact dermatitis (ACD), systemic allergic reactions, drainage, and delayed wound healing that can lead to infection. As such, effective wound closure and protection is critical for optimal healing, as it influences overall recovery and ultimately patient outcomes. The choice of wound closure systems plays a key role in mitigating the aforementioned complications. Traditional wound closure systems, including cyanoacrylate-based adhesives combined with mesh or Steri-Strips (3M, Two Harbors, MN, USA), provide mechanical protection but have also been associated with adverse skin reactions, including ACD, irritation, and wound dehiscence [[1], [2], [3]]. Notably, cyanoacrylate mesh (CM) has been associated with wound dehiscence rates ranging from 1.3% to as high as 8% [[4], [5], [6]]. These complications can lead to increased discomfort, delayed recovery, and a higher likelihood of required secondary intervention. Examples of wound complications, particularly ACD, are shown in Figure 1a and b.

Figure 1.

(a) Photograph of a 59-year-old female patient on day 3 after left anterior total hip arthroplasty. Urticaria, erythema, and inflammation at the operative site dressed with the cyanoacrylate mesh dressing. (b) Photograph of a 62-year-old female patient on day 7 after right total knee arthroplasty. Blistering, erythema, and skin discomfort were noted at the operative site dressed with the cyanoacrylate mesh dressing.

With the recent trend toward outpatient orthopaedic procedures, including joint arthroplasty in the ambulatory settings, as well as a growing emphasis on simplified postoperative wound care, there is increasing interest in alternative wound closure systems that minimize wound complications while maintaining effective barrier protection during the healing process. Silk fibroin (SF), derived from silkworms, possesses unique properties conducive to tissue regeneration and wound healing, making it an attractive candidate for surgical wound closure [[7], [8], [9]]. Among other emerging alternatives, SF-based wound closure systems have shown promise in reducing rates of ACD, and improving wound healing outcomes and postoperative comfort in other surgical contexts [7,10]. However, the application of SF closures in total hip arthroplasty (THA) and total knee arthroplasty (TKA) procedures remains largely unexplored.

Building upon these findings, and in observation of an increased incidence of ACD using CM for closure at our institution, our study aims to prospectively evaluate the efficacy of an SF dressing compared to a cyanoacrylate adhesive combined with mesh, in reducing superficial wound complications following THA and TKA. Leveraging a case-control retrospective design, we seek to establish the safety and efficacy of SF dressings while investigating its potential to mitigate adverse wound healing events. We hypothesize that, compared to a CM dressing, the SF dressing will decrease the incidence of ACD.

Material and methods

A priori analysis was conducted prior to the start of this study. Based on previously published literature on CM wound dressing outcomes, CM and SF adverse event rates of 5% and 0%, respectively, were chosen [11,12]. For the purposes of statistical modeling, this rate was conservatively adjusted to 0.01% to enable a priori analysis of our study group. An effect size of 0.80 and a CM-to-SF allocation ratio of 2:1 was used in a Fisher’s exact test power analysis. This power analysis test was chosen due to the expected infrequent adverse event rate in this dataset using the SF dressing. A priori analysis demonstrated that a sample size of 170 CM subjects and 85 SF subjects was required to achieve 80% power. Once sufficient sample size was confirmed, institutional review board approval was obtained.

A retrospective review using a total joint arthroplasty (TJA) registry was conducted to identify all patients who received either SF or CM dressings for postoperative wound care after undergoing TJA by a fellowship-trained joint arthroplasty surgeon from June 2023 to March 2024 at a single, high-volume orthopaedic specialty hospital. It is important to note that the SF dressing only became available to our institution in December 2023; thus, patient enrollment was performed in a sequential fashion with complete inclusion of all subjects based on qualification for use. TJA was defined to include any primary or revision THA or TKA, or unicompartmental knee arthroplasty. Exclusion criteria included multitrauma surgical procedures, arthroplasty due to tumor cases, any procedure involving a wound problem at the surgical site prior to TJA, and any procedure with follow-up of less than 6 wks. Data were collected from hospital and clinic records through the institution’s total joint registry.

All primary and revision THAs were performed via a direct anterior approach. All primary and revision TKAs were performed through an anterior midline incision. For postoperative wound care, all patients underwent subcutaneous layer closure with either 2-0 or 3-0 Vicryl (Ethicon, J&J) sutures. All patients underwent skin closure with a running 3-0 monoderm suture and closed suction drainage was not used in any of the cases. The final skin closure application was performed using the SF or CM closure system. The SF closure system, SYLKE Adhesive Wound Closure (SYLKE Inc., La Jolla, CA, USA), is a Food and Drug Administration–registered Class 1, 510(K)-Exempt Medical Device for use on clean surgical wounds and became available to our institution in December 2023. Dermabond Prineo (Ethicon, J&J), the CM closure system, is a cyanoacrylate adhesive combined with a synthetic mesh. On completion of closure in both groups, an occlusive barrier dressing was applied. Removal of the CM wound closure system occurred between 10 and 14 d postoperatively or earlier upon identification of a wound complication. The SF wound closure system remained in place for a minimum of 14 d and was left in place for up to 6 wks or until it was no longer effectively applied to the wound. Hospital records and chart notes were retrospectively reviewed for all patients in which either dressing was used upon closure for the above-stated procedures, specifically for adverse cutaneous events directly related to wound closure systems. ACD was diagnosed as erythema, with or without weeping or pruritus, in a geometric border at the surgical site within 14 d of the index procedure and in the absence of fluctuance or purulence. It is important to note that all patients in this study that were documented with a wound complication, including ACD, were seen either in the clinic setting or emergency department. Treatment of concerns without visual inspection, such as minor concerns that might be handled through phone conversation, were not included in the adverse event study group to ensure accurate and valid data analysis. Recognizing that ACD almost always occurs within 4 wks of application of wound closure technology [[13], [14], [15]], we followed all subjects for a minimum of 6 wks to ensure inclusion of all adverse cutaneous events related to wound closure.

Demographic data were collected to assess matching of the 2 cohorts, SF and CM. Variables included gender, age, body mass index (BMI), diagnosis of diabetes, and American Society of Anesthesiologists (ASA) physical status classification. The primary outcome variable collected was the incidence of ACD. Secondary outcome variables included revision surgery due to infection within 90 d and 30-d emergency department visits. The use of postoperative wound treatment modalities, including oral antibiotics, topical and oral steroids, and oral and intravenous antihistamines, was also recorded.

A priori power analysis was run using G∗Power Software (latest ver. 3.1.9.7; Heinrich-Heine-Universität Düsseldorf, Düsseldorf, Germany). Statistical analyses were performed with the statistical programming tool R (R Core Team [2021]; R: A language and environment for statistical computing; R Foundation for Statistical Computing, Vienna, Austria). Fisher’s exact tests were run to compare patient demographics and clinical outcomes, including gender and ASA classification, smoking status, diagnosis of diabetes, peripheral vascular disease, nutritional deficiency, colchicine use, diagnosis of ACD, oral antibiotics use, and 30-day readmission between the 2 wound closure systems. T-tests were run to compare continuous patient demographics, including age and BMI between the 2 wound dressing systems.

Results

A total of 257 patients were identified who underwent TJA during the proposed time frame and had either CM or SF used as part of their wound closure regimen. An example of the SF wound dressing can be seen in Figure 2a and b. There were 172 patients who received CM closure and 85 patients who received SF (Table 1). In total, 158 patients underwent primary TKA (122 CM and 36 SF), 14 underwent unicompartmental knee arthroplasty (11 CM and 3 SF), 2 underwent revision TKA (2 CM and 0 SF), 82 patients underwent primary THA (37 CM and 45 SF), and 1 underwent revision THA (0 CM and 1 SF). There was no significant difference in the number of patients receiving primary or revision TJA between the 2 groups (Table 1).

Figure 2.

(a) Intraoperative photograph of a 59-year-old female patient with application of a silk fibroin wound dressing after right total knee arthroplasty. (b) Photograph of a 61-year old female patient on day 14 after left total knee arthroplasty. The left knee covered by the silk fibroin wound dressing showed no signs of wound healing complications, with appropriate egress of wound exudate at the inferior portion of the incision.

Table 1.

Patient characteristics.

Characteristicsa	Total (n = 257)	Cyanoacrylate mesh (n = 172)	Silk fibroin (n = 85)	P value
Total TKA	174	135	39	.903
Primary TKA	158	122	36
Revision TKA	2	2	0
Unicompartmental TKA	14	11	3
Total THA	83	37	46	.999
Primary THA	2	37	45
Revision THA	1	0	1
Gender				.596
F	120 (46.7)	78 (45.3)	42 (49.4)
M	137 (53.3)	94 (54.7)	43 (50.6)
Age				.342
Mean (SD)	65.3 (9.0)	65.6 (8.4)	64.5 (10.3)
Median [minimum, maximum]	64 [36, 90]	64 [48, 88]	63 [36, 90]
BMI				.139
Mean (SD)	28.0 (4.6)	28.3 (4.69)	27.4 (4.35)
Median [minimum, maximum]	27.8 [17.9, 41.9]	28 [17.9, 41.9]	27.2 [19.3, 38.2]
ASA				.033b
1	10 (3.9)	3 (1.7)	7 (8.2)	.011b
2	200 (77.8)	135 (78.5)	65 (76.5)	.134
3	47 (18.3)	34 (19.8)	13 (15.3)	.383
Diabetes	21 (8.2)	14 (8.1)	7 (8.2)	1
Prediabetes	6 (2.3)	5 (2.9)	1 (1.2)	.667
Smoker, current	8 (3.1)	3 (1.7)	5 (5.9)	.120
Coronary artery disease	16 (6.2)	12 (7.0)	4 (4.7)	.590
Peripheral vascular disease	3 (1.2)	3 (1.7)	0 (0.0)	.553
Lupus/scleroderma	2 (0.7)	2 (1.2)	0 (0.0)	1
Nutritional deficiency	9 (3.5)	7 (4.1)	2 (2.4)	.722
Skin disorders	1 (0.4)	0 (0.0)	1 (1.2)	.331
Colchicine use	2 (0.8)	8 (4.7)	2 (2.4)	.728
Chemotherapy treatment	1 (0.4)	0 (0.0)	1 (1.2)	.331

ASA, American Society of Anesthesiologists; BMI, body mass index; F, female; M, male; SD, standard deviation; THA, total hip arthroplasty; TKA, total knee arthroplasty.

^aReported as n (%) unless otherwise stated.

^bP < .05.

Demographics

The 2 cohorts showed no significant differences in baseline characteristics, including mean age, gender distribution, mean BMI, diabetes status, nicotine use, cardiovascular disease, lupus/scleroderma diagnosis, nutritional deficiency, skin disorders, colchicine use, or chemotherapy treatment (all P > .05). There was a statistically significant difference in ASA score distribution between cohorts (P = .033), with a significantly higher number of ASA 1 patients in the SF cohort compared to the CM cohort (P = .011). The distributions of all other ASA categories did not differ significantly (P > .05). A full breakdown of patient characteristics can be found in Table 1.

Clinical outcomes

There was a significantly greater incidence of ACD in the CM cohort compared to the SF cohort (P = .018). Of the 172 patients receiving CM wound dressings, 11 (6.4%) developed ACD within 6 wks of the procedure. In the SF cohort, 0 (0%) of the 85 patients developed ACD. There were no significant differences in the rate of secondary complications between the CM and SF cohorts, and no patients in either group experienced wound dehiscence, persistent would discharge, or superficial infection. Seven patients in the CM cohort (4.1%) were prescribed oral antibiotics as a precautionary measure following TKA due to blistering (n = 1), allergic reaction (n = 4), papular rash (n = 1), and prior wound incision and drainage of superficial hematoma of the operated knee (n = 1). No patients were prescribed oral antibiotics in the SF cohort. In the CM cohort, 2 patients (1.2%) required readmission within 30 d. One was admitted for nonspecific acute pain in the right knee and the other for a rash accompanied by a nonspecific skin eruption. In the SF cohort, 1 patient (1.2%) was readmitted within 30 d for lower back pain. A comprehensive summary of clinical outcomes can be found in Table 2.

Table 2.

Outcome variables.

Outcomes, n (%)	Overall (N = 257)	Cyanoacrylate mesh (N = 172)	Silk fibroin (N = 85)	P value
Allergic contact dermatitis	12	11 (6.4)	0 (0.0)	.018a
Oral antibiotics used	12	7 (4.1)	0 (0.0)	.099
Topical steroid used	8	8 (4.7)	0 (0.0)	.056
Oral steroid use	3 (1.2)	3 (1.7)	0 (0.0)	.553
Oral antihistamine started	3 (1.2)	3 (1.7)	0 (0.0)	.553
Wound dehiscence	0 (0.0)	0 (0.0)	0 (0.0)	1
Persistent wound drainage	0 (0.0)	0 (0.0)	0 (0.0)	1
Superficial infection	0 (0.0)	0 (0.0)	0 (0.0)	1
30-day readmission	3 (1.2)	2 (1.2)b	1 (1.2)c	1

^aP < .05.

^b(1) nonspecific acute pain of right knee, (1) rash and other nonspecific skin eruption.

^c(1) low back pain.

Discussion

ACD is a known complication associated with surgical wound closure systems, among various surgical specialties, and can be associated with patient morbidity and dissatisfaction. While studies have introduced SF dressings as an option for wound dressing in plastic surgery or general surgery, there is a paucity of literature comparing the outcomes of SF vs other medical adhesives in orthopaedic surgery, specifically TJA. In this series of 257 TJA cases, the CM cohort revealed a significantly higher incidence of ACD (P = .018). Only those patients who required an in-person medical visit were considered as qualifying for an adverse event in this study group. The treatment of wound concerns without visual inspection, such as minor issues that might be handled through documented or undocumented phone conversation, were not included as an adverse event to ensure accuracy of the clinical dataset. For this reason, the authors suspect that the true incidence of all ACD events is most likely greater than the values reported in this manuscript.

To the authors’ knowledge, previous reports on SF dressings have been restricted to animal models and other surgical specialties [7,12,[16], [17], [18], [19]]. A full literature search of previous clinical studies that assessed outcomes of silk-derived wound dressings can be found in Table 3. Zhang et al. evaluated SF biomaterials in animal models, demonstrating reduced average wound healing times and better skin regeneration compared to standard commercial dressings [10]. Rouhani et al. reported a significantly lower rate of discomfort, skin irritation, and itching in patients who received the SF dressing compared to those who received the CM dressing at all time points up to 4 wks postoperatively. Consequently, none of their patients required pharmaceutical intervention in the SF group, while 45.8% of patients in the CM group required either an antibiotic or steroid prescription [7]. Similarly, across 71 patients who underwent 2-stage, implant-based breast reconstruction and received the silk-derived SERI Surgical Scaffold (Allergan Inc., Medford, MA), Fine et al also found low rates of infection (1.4%) [19]. A second prospective trial of 48 patients conducted by Rouhani et al corroborates these findings, reporting a significantly lower incidence of skin erythema (P = .002) in patients using a silk dressing compared to those using Steri-Strips following abdominoplasty, reduction mammaplasty, or brachioplasty procedures. Notably, 75% of patients had a partial or total detachment of Steri-Strips, while 18.8% of patients experienced a partial detachment of the silk dressing and 0% experienced a total detachment (P < .001) [12].

Table 3.

Literature review of silk fibroin outcomes.

Author (year)	Surgical procedure(s)	Comparator	Total N (female N)	Study design	Mean age, years (range)	Clinical follow-up	Silk fibroin-associated complications
Rouhani et al. (2023)	Abdominoplasty, belt lipectomy, mastopexy, reduction mammoplasty	CM (Dermabond Prineo [Ethicon, Inc., Somerville, NJ])	25 (25)	Prospective, randomized, single-blinded trial	40.7 (19-66)	<1 wk to 8 wks	Skin discomfort (4%), itching (8%), rash (0%), topical/oral steroids or oral/intravenous antibiotics (0%)
Rouhani et al. (2024)	Abdominoplasty, belt lipectomy, mastopexy, reduction mammoplasty	Steri-Strips (3M, Two Harbors, MN, USA)	48 (48)	Prospective, randomized, single-blinded trial	43.6 (19-76)	<1 wk to 8 wks	Skin discomfort (4.2%), irritation or itching (4.2%), rash (2.1%), topical/oral steroids or oral/intravenous antibiotics (0%)
Fine et al. (2015)	Two-stage, implant-based breast reconstruction	None	71 (71)	Prospective, multicenter, single-arm trial	Median 48 (26-74)	<1 wk to 12 mo	Tissue necrosis (8.5%), hematoma (7%), seroma (7%), cellulitis (4.2%), breast infection (1.4%), implant loss (4.2%), capsular contracture (1.4%)
De Vita et al. (2014)	Direct-to-implant after skin-sparing mastectomy	None	15 (15)	Retrospective case series	Median 21 (41-62)	Mean 9.5 mo (range: 4-13 mo)	Loss of implant due to partial flap necrosis (4.7%) Seroma (4.7%) Hematoma self-limiting (4.7%)

CM, cyanoacrylate mesh.

One potential reason for the reduced attachment time in the Steri-Strip group may be due to the higher likelihood of developing ACD; patients who experience wound complications may be more likely to remove the wound dressings, thus negating their purpose as a mechanical barrier. Furthermore, variability in wound closure strip application techniques and intended function may contribute to inconsistent outcomes, as highlighted by a survey of the Canadian Society of Plastic Surgery, which found a lack of consensus among surgeons on best practices for their use [20]. Within the same study, 70% of respondents experienced complications with wound closure strips, most commonly blistering (82.7%). Finally, a prospective comparison of CM wound dressing and interrupted polypropylene sutures for wound closure in patients undergoing total ankle arthroplasty found CM to be associated with a higher rate of wound complications, including ACD (2.5% vs 0%) and wound dehiscence (8.3%-2.8%), and a significantly higher rate of surgical site infections (P = .035) [21].

Akin to prior studies in other surgical specialties, this study demonstrates a statistically significant difference in ACD and topical steroid use between the SF cohort and the CM cohort during the postoperative period, with zero instances of both ACD and topical steroid use in the SF subgroup. It should also be mentioned that we did not observe wound healing concerns in SF subgroup, including the risk of wound dehiscence in high-tension anterior knee incisions used in TKA. Our study is the first to expand the application of SF dressings to the field of orthopaedic surgery.

Similar to previous manuscripts, findings from our study support no incidence of ACD using an SF dressing and may be attributed to the exceptional biocompatibility, low immunogenicity, and hemostatic properties of SF [9,12,22,23]. Studies on SF wound dressings highlight that its structure promotes oxygen exchange at the surgical site and allows for the escape of wound exudate, thus enhancing wound healing outcomes [12,24]. Moreover, SF films, which can elongate approximately 15% when dry and 27% when wet, help distribute shear forces evenly across the application site, thereby reducing the risk of blister formation [12,25].

Limitations of this manuscript include the retrospective design. Also, this study is limited to a single total joint registry from 1 institution and reflects the practice of a single surgeon, which may impact the generalizability of the findings. In addition, the lack of statistical significance of secondary outcome variables between the 2 cohorts is likely explained by the extremely low incidence of secondary outcome events in the SF cohort. Also, although the CM cohort had higher values in almost all secondary outcome variable categories compared to the SF cohort, they did not approach statistical significance most likely due to sample size. Despite these suggested limitations, the SF cohort of patients included in this study has all successfully achieved satisfactory wound healing outcomes.

The observations of our study provide enthusiasm to support the safe and effective use of SF dressings in orthopaedic surgery, particularly in TJA. The notable reduction in the incidence of ACD with SF dressings, as compared to a commonly used alternative, not only enhances patient safety and satisfaction but also has the potential to substantially reduce hospital and patient costs by reducing wound-related complications [26,27]. Future studies should aim to further investigate the application of SF dressings across various orthopaedic procedures and other surgical specialties.

Conclusions

There is a very low incidence of wound complications in TJA when using the SF dressing wound closure system. There was a statistically significant decrease in ACD when using the SF dressing compared to the cyanoacrylate adhesive mesh dressing. SF may be a superior wound closure option to consider in the care of patients undergoing TJA.

Conflicts of interest

Robert Gorab received royalties from DePuy Synthes Products, Inc.; received speakers bureau/paid presentations for Synthes GMBH; and is a paid consultant for Medical Device Business Services, Inc. and DePuy Synthes Products, Inc. Steven Barnett received royalty fees from DePuy Synthes Products, Inc. and OMNIlife science, Inc. and is a paid consultant for DePuy Synthes Products, Inc., Medical Device Business Services, Inc., and OMNIlife science, Inc. Susanne Porter is a paid consultant for Medical Device Business Services, Inc. Travis Scudday is a paid consultant for OMNIlife science, Inc. and Medical Device Business Services, Inc. All other authors declare no potential conflicts of interest.

For full disclosure statements refer to https://doi.org/10.1016/j.artd.2025.101668.

CRediT authorship contribution statement

Justin P. Moo Young: Writing – review & editing, Writing – original draft, Investigation, Conceptualization. Viraj A. Deshpande: Writing – review & editing, Writing – original draft, Formal analysis, Data curation. Susanne Porter: Writing – review & editing, Writing – original draft, Conceptualization. Edward J. Quilligan: Writing – review & editing, Writing – original draft, Resources, Project administration, Conceptualization. Travis Scudday: Writing – review & editing, Writing – original draft, Supervision, Conceptualization. Steven Barnett: Writing – review & editing, Writing – original draft, Supervision, Conceptualization. Robert Gorab: Writing – review & editing, Writing – original draft, Supervision, Methodology, Investigation, Conceptualization.