Caesarean section (CS) is one of the most common obstetric surgeries, with its incidence in India rising from 17.2% in 2015-16 to 21.5% in 2020-211. While generally safe, CS is associated with significant risks of SSIs, occurring in 5–12% of cases globally2, leading to increased morbidity, prolonged hospitalization, and higher healthcare costs. Existing preventive measures focus on preoperative, intraoperative, and postoperative interventions; however, cost-effective solutions remain vital, particularly in resource-limited settings. Human amniotic membrane (HAM), rich in bioactive factors like VEGF, TGF-β, bFGF, PDGF, and SDF-1, has shown promising results in wound healing3. Its properties include promoting angiogenesis, reducing inflammation, minimizing scarring, and enhancing immune tolerance. HAM has been successfully used in diverse surgeries, including plastic4, gastrointestinal5, neurosurgical6, ophthalmic7 and vaginal plastic8 procedures, but its application in CS wound healing remains unexplored. This study investigates the effectiveness of HAM in improving CS wound healing and reducing associated complications.

Study Design Ethics committee approval (IECHR-2023-59-88) and written informed consent from patients was recorded. Clinical trial registry for the study was done at Clinical Trial Registry India (CTRI). Patients who were scheduled to undergo primary CS with gestational age more than 28 weeks were included in the study while those with infections like HIV, HCV, HBV, syphilis, suffering through moderate to severe anemia or thrombocytopenia were excluded from the trial. The patients were randomised into 2 groups by block randomisation, 50 in each group: Group A, Autologous HAM patch dressing group and group B, Control group. Patients were fully counselled in the regional languages about the procedure and pros and cons of the HAM dressing. Following this a written informed consent was recorded. Preparation of HAM patch After extracting placenta by controlled cord traction, it was held with sterile gloved hands and placed over the sterile sheet. The amnion chorion bilayer was separated from the placental tissue manually. The bilayer was placed in a sterile kidney tray where it was washed with 200 ml NS till all the blood clots & debris were removed. The Amnion (which is thin, transparent, shiny, avascular) was separated from the chorion by blunt dissection or simple traction with the help of an assistant. The amnion corresponds to size 10×2 cm approx. was cut by scissors. After skin closure, the HAM patch was placed over the suture line as dressing (Fig 1). Normal saline soaked gauze was placed over the HAM patch to prevent drying of the membrane followed by dry gauze over it. Vaginal toileting was done by povidone-iodine solution. Assessment and Follow up On postoperative Day 3, sterile gauze dressing was removed and condition of AM dressing was analysed and wound healing was assessed using the REEDA score. If the wound was healthy, no further dressing was done and caesarean site was left open as such, and if unhealthy, then appropriate dressing was done. At Day 8 during suture removal, REEDA score was re-assessed. The development of any SSI, defined and classified according to Centres for Disease Control (CDC) criteria 2 was seen. Patient was discharged according to hospital protocol and advised to follow routine wound care at home. Telephonically follow up after discharge was done and the patient was explained to report if she had fever, pain, wound discharge or any other complaints. During a routine PNC visit after 42 days, the patient was re-evaluated for wound healing by REEDA9 score. In presence of signs and symptoms of inflammation/ infection, pus swab/aspirate was collected using aseptic technique in sterile containers and was sent to microbiology lab without delay for further processing. Outcomes Primary outcome measure: REEDA (Redness, Edema, Ecchymosis, Discharge, Approximation) score in both the groups.

Secondary outcome measures 1. Incidence of SSI in both groups. 2. Adverse events like pain, swelling, discharge, dehiscence in both the groups. 3. Organisms associated with caesarean surgical site infection in both the groups. 4. Total length of hospital stay due to wound morbidity in both the groups Statistical Analysis All the data was entered in MS Excel. The scores of wound healing by REEDA score was compared in both groups by unpaired T tests. Incidence rate of surgical site infection and other parameters between the two groups was compared by Chi-squared test/Fisher's exact test. SPSS software was used for analysis version 24.0. P-value <0.05 was taken as significant.

Patient characteristics Mean age of cases and controls was comparable (26.12 ± 3.31 years vs 26.7 ± 3.74 years, p = 0.413). There was no significant difference in the frequency in respective age groups (p=0.944) and BMI of both the groups (4.5 ± 2.02 kg/m² vs. 24.8 ± 2.52 kg/m², p=0.522). Mean family monthly income in cases was comparable Rs 19420 ± 6797.63 vs. Rs 20720 ± 6963.58 in controls, with no significant difference between them (p = 0.347). Compared to controls, cases had comparable distribution of socio-economic status: upper middle (18% in cases vs. 22% in controls), lower middle (28% vs. 26%), upper lower (22% vs. 32%) and lower class (32% vs. 20%) (p = 0.467). Compared to controls, cases had a comparable distribution of booking status: booked (70% in cases vs. 74% in controls) and unbooked (30% vs. 26%) (p = 0.656). Compared to controls, cases had comparable distribution of gestational age majority being at term (54% in cases, 66% in controls)( p=0.787) and parity: primigravida (50% in cases vs. 54% in controls), Gravida 2 (34% vs. 32%) and Gravida ≥3 (16% vs. 14%) (p=0.917). Management characteristics Both the groups had a comparable distribution of elective/emergency caesarean sections: elective (20% in cases vs. 14% in controls) and emergency (80% vs. 86%) (p = 0.424). In both the groups most common indications for caesarean were fetal indications (76% in cases vs. 70% in controls) (p = 0.499). The mean duration of surgery in cases was 43.8 ± 3.72 minutes vs. 42.4 ± 5.91 minutes in controls, with no significant difference observed (p = 0.16). The distribution of maternal indications for caesarean was comparable between cases and controls. Antepartum hemorrhage was the most common indication, observed in 33.33% vs 40% (p = 0.945). Severe preeclampsia with uncontrolled BP followed (41.67% vs. 26.67%), antepartum eclampsia (16.67% vs. 20%), and placenta previa (8.33% vs. 13.33%). Fetal indications for caesarean were also comparable between cases and controls. Fetal distress was the leading cause, observed in 23.68% vs 28.57% (p = 0.945). This was followed by malpresentation (18.42% vs. 20%), cephalopelvic disproportion (13.16% vs. 14.29%), and meconium-stained liquor with poor Bishop’s score (10.53% vs. 14.29%). Other indications included second-stage arrest (13.16% vs. 8.57%), failed induction of labor (10.53% vs. 5.71%), cord prolapse (7.89% vs. 2.86%), and fetal growth restriction with absent end-diastolic flow (2.63% vs. 5.71%). Outcome Group A (cases) had a significantly higher proportion of patients without SSI (94% vs. 76%) and significantly lower proportion with inpatient infection (2% vs. 8%) and readmission (4% vs. 16%) (p = 0.042). Classification of SSI was comparable between cases and controls, with 100% superficial in cases vs. 75% in controls and deep infections 0% vs. 25% (p = 1) (Table 1). Distribution of organisms isolated was also comparable (p=0.71). Significant differences in REEDA scores were noted on days 3, 8, and 42. On day 3, the median score for cases was 2 (1–2), which was significantly lower than the control score of 2 (2–3) (p = 0.02). Similarly, on day 8, the median score for cases was 3 (3–4), significantly lower than the control score of 4 (3–5) (p = 0.042). By day 42, the median score for cases had decreased to 1 (0–1), remaining significantly lower than the control score of 1 (1–1) (p = 0.021). (Table 2, Fig 2). Mean duration of hospital stay was significantly lower in cases at 4.02 ± 0.96 days compared to 5.26 ± 3.22 days in controls (p = 0.011) Discussion The use of autologous amniotic membrane as a dressing has been extensively studied in various types of wounds like vascular foot ulcers, burn wounds, diabetic foot ulcers, corneal ulcer. Mohseni et al.10 demonstrated amniotic membrane dressings reduce pain and analgesic need in cesarean wounds. This first randomized trial evaluates autologous HAM's safety and efficacy in cesarean wound healing and SSI prevention. Randomization ensured comparable patient characteristics, attributing outcome differences solely to the intervention, independent of demographics, socioeconomic factors, clinical variables, or comorbidities. Pertinently for efficacy in wound healing, we assessed REEDA score. Therefore, this study also brought forth the utilization of this easy to use scoring system. Compared to controls, cases had statistically lower REEDA scores at the follow-up time periods up to day 42, starting from day 3 (Fig 3). This showed that inflammation was significantly decreased with HAM dressing as compared to normal wound dressing not only post operatively but also at follow up, highlighting the superior nature of HAM to the normal dressing. Amniotic membrane dressings significantly reduce pain and analgesic needs across wound types. Mohseni et al.10 reported lower VAS scores in cesarean wounds: 24 hours (4.6 vs. 5.6, p=0.026), 36 hours (3.5 vs. 4.4, p=0.026), and 48 hours (2.8 vs. 3.8, p=0.004). Mermet et al.11 observed less pain in 15 chronic ulcer patients treated with HAM for 90 days. Mohammadi et al.12 found reduced pain and analgesic use in 124 burn patients. Adly et al.13 showed HAM outperformed polyurethane dressings in 46 burn patients, improving recovery and reducing pain. These studies affirm HAM's therapeutic and cost-effective benefits. In this study when compared to controls, cases had significantly lower surgical site infections (6% vs. 24%, p=0.023) leading to lesser readmissions (4% vs. 16%) (p=0.042) and significantly lower duration of hospital stay (4.02 ± 0.96 vs. 5.26 ± 3.22 days, p=0.011). This can be attributed to better wound healing and allowing for discharge of the patient and saves a lot of cost for the patient workup. Likewise in the study by Mohseni F et al,10 no case of SSI was reported after HAM application in caesarean wounds. SSI was superficial in 12 patients, with 3 in cases and 9 in controls. Similar findings have been reported in previous studies, for instance, Stanirowski et al14 reported that SSI was superficial in total 19 cases, out of which 5 were in cases and 14 were in controls. In both groups in our study, the two most frequently isolated organisms were MRSA and E. coli. Other commonly isolated organisms in other studies include Klebsiella pneumoniae, Proteus mirabilis, Enterobacter cloacae, Escherichia coli, Staphylococci MSSA. Overall, no adverse events were seen on HAM application. Fetal outcomes were similar, with comparable NICU admissions (18% vs. 20%), with pneumonia being most common cause (55.5% vs. 60%), and mean birth weights (3.1 vs. 3.03 kg). This single-center study had limitations, including a small sample size, lack of double-blinding, and potential biases from institutional protocols and microbial environments, possibly limiting generalizability to broader healthcare settings. This study suggests autologous HAM dressings reduce SSI, advocating further research comparing HAM with advanced dressings in larger trials.

Autologous HAM is a novel technique for CS wound healing and allows for early patient discharge with lesser hospital stay and less REEDA score. The autologous HAM patch dressing significantly improves primary Caesarean wound healing compared to standard dressing methods. Such dressings can be especially useful in lower resource settings where SSI rates are very high and advanced dressing materials are unaffordable. Acknowledgement None

1. Neethi Mohan V, Shirisha P, Vaidyanathan G, Muraleedharan VR. Variations in the prevalence of caesarean section deliveries in India between 2016 and 2021 - an analysis of Tamil Nadu and Chhattisgarh. BMC Pregnancy Childbirth. 2023;23(1):622. 2. Shea SK, Soper DE. Prevention of cesarean delivery surgical site infections. Obstet Gynecol Surg. 2019;74(2):99–110. 3. Duan-Arnold Y, Uveges TE, Gyurdieva A, Johnson A, Danilkovitch A. Angiogenic potential of cryopreserved amniotic membrane is enhanced through retention of all tissue components in their native state. Adv Wound Care. 2015;4(9):513-22. 4. Lipovy B., Hladik M., Stourac P., Forostyak S. Case report: Wound closure acceleration in a patient with toxic epidermal necrolysis using a lyophilised amniotic membrane. Front Bioeng Biotechnol. 2021;9:649317. 5. Mamede KM, Sant'anna LB. Antifibrotic effects of total or partial application of amniotic membrane in hepatic fibrosis. Acad Bras Cienc. 2019;91:e20190220. 6. De Weerd L, Sjavik K, Pedersen LK, Weum S, Hennig RO. Triple use of autologous amnion graft in the treatment of meningomyelocele and split cord malformation. Plast Reconstr Surg Glob Open. 2020;8:e2539. 7. Caporossi T, Molle A, Carlà MM, Picardi SM, Gambini G, Scampoli A, et al. Applications of human amniotic membrane patching assisted vitrectomy in the management of postoperative PVR in complex retinal detachments. J Clin Med. 2023;12(3):1137. 8. Vatsa R, Bharti J, Roy KK, Kumar S, Sharma JB, Singh N, et al. Evaluation of amnion in creation of neovagina in women with Mayer-Rokitansky-Kuster-Hauser syndrome. Fertil Steril. 2017;108:341–5. 9. Alvarenga MB, Francisco AA, de Oliveira SM, da Silva FM, Shimoda GT, Damiani LP. Episiotomy healing assessment: Redness, Oedema, Ecchymosis, Discharge, Approximation (REEDA) scale reliability. Rev Lat Am Enfermagem. 2015;23(1):162-8. 10. Mohseni F, Saem J, Sekhavati E, Molazem Z, Tabrizi R. Amniotic membrane for pain control after cesarean section. Crescent J Med Biol Sci. 2018;5(3):198-202. 11. Mermet I, Pottier N, Sainthillier JM, Malugani C, Cairey-Remonnay S, Maddens S, et al. Use of amniotic membrane transplantation in the treatment of venous leg ulcers. Wound Repair Regen. 2007;15(4):459-64. 12. Mohammadi AA, Riazi H, Hasheminasab M, et al. Amniotic membrane dressing vs conventional topical antibiotic dressing in hospitalized burn patients. Iran Red Crescent Med J. 2009;11(1):66-70. 13. Adly OA, Moghazy AM, Abbas AH, Ellabban AM, Ali OS, Mohamed BA. Assessment of amniotic and polyurethane membrane dressings in the treatment of burns. Burns. 2010;36(5):703-10. 14. Stanirowski PJ, Bizoń M, Cendrowski K, Sawicki W. Randomized controlled trial evaluating dialkylcarbamoyl chloride impregnated dressings for the prevention of surgical site infections in adult women undergoing cesarean section. Surg Infect (Larchmt). 2016;17(4):427-35.