Quilting sutures versus conventional closure after modified radical mastectomy with axillary dissection: a GRADE-assessed systematic review and meta-analysis of randomized controlled trials with pre-specified subgroups by quilting extent.

Introduction
Modified radical mastectomy with axillary lymph node dissection remains a definitive surgical intervention for breast cancer management worldwide. However, postoperative seroma formation persists as one of the most frequent complications, with incidence rates ranging from 3% to 85% [1–7]. Seromas constitute a substantial source of patient morbidity through delayed wound healing, increased infection risk, skin flap necrosis, persistent pain, and repeated clinic visits for aspiration [1–3, 7–10]. The clinical ramifications extend beyond immediate postoperative recovery, as seroma management frequently delays the initiation of adjuvant chemotherapy and radiotherapy, potentially compromising oncological outcomes [3, 4, 11, 12]. Furthermore, the economic burden imposed by seroma-related complications underscores the need for evidence-based preventive strategies in contemporary breast surgical oncology [11, 13, 14].
The pathophysiological mechanisms underlying post-mastectomy seroma formation involve surgical trauma to lymphatic structures, inflammatory exudation, and persistence of dead space between elevated skin flaps and the underlying chest wall musculature [1, 2, 6, 7, 15]. Traditional wound closure techniques create a potential cavity for fluid accumulation, driven by ongoing lymphatic leakage and inflammation [3, 6, 7]. The quilting suture technique, characterized by the systematic fixation of mastectomy skin flaps to the underlying pectoralis major muscle using absorbable sutures placed at regular intervals, represents a mechanistic approach to dead space obliteration [2, 4, 5, 7–18]. By mechanically reducing the potential space for fluid accumulation, quilting theoretically addresses the fundamental anatomical factors predisposing seroma development, offering a simple and cost-effective intraoperative intervention [11, 13, 14, 17].
Despite accumulating evidence from individual randomized controlled trials demonstrating the variable efficacy of quilting sutures, substantial heterogeneity persists regarding the optimal technical parameters and comparative effectiveness [1–5, 8–18]. Key technical variables, including the extent of anatomical coverage (pectoral area versus axilla versus both) and the concurrent use or omission of closed-suction drainage, have not been systematically evaluated in comparative studies [11, 13, 17, 18]. Several studies have reported significant reductions in seroma formation rates with quilting compared to conventional techniques [2–5, 7–10, 12, 14–16], while others have demonstrated minimal benefits [9, 17, 18]. Important methodological limitations in the existing literature, including small sample sizes, inconsistent seroma definitions, and insufficient reporting of potential confounders, have precluded definitive conclusions regarding the magnitude of the treatment effect [11, 13, 17].
Prior systematic reviews and meta-analyses have attempted to synthesize the available evidence on quilting sutures versus conventional closure; however, these analyses were limited by the inclusion of heterogeneous mastectomy types, such as breast-conserving surgery and skin-sparing procedures without axillary dissection [11, 13, 18]. Previous meta-analyses have demonstrated significant reductions in seroma rates with quilting but did not specifically isolate modified radical mastectomy with axillary dissection, the surgical context in which seroma risk is highest [11, 13]. Critically, no previous systematic review has performed pre-specified subgroup analyses comparing different quilting techniques or focused exclusively on modified radical mastectomy with complete axillary dissection [11, 13, 17, 18].
This systematic review and meta-analysis is the first comprehensive synthesis of randomized controlled trial evidence specifically evaluating quilting sutures versus conventional closure following modified radical mastectomy with axillary lymph node dissection. Uniquely, this analysis employed pre-specified subgroup stratification by quilting technique, with three studies utilizing pectoral-only quilting and four studies employing comprehensive pectoral-plus-axillary quilting to elucidate whether anatomical coverage influences treatment efficacy [11, 17, 18]. Furthermore, this investigation implemented sensitivity analyses were performed using the leave-one-out approach to examine the impact of individual studies on the observed treatment effects [11, 13, 17, 18]. Most importantly, this systematic review applied the GRADE methodology to assess the certainty of evidence for all critical outcomes, providing a transparent evaluation of confidence in effect estimates to inform evidence-based surgical decision-making [11, 13]. By focusing exclusively on modified radical mastectomy with axillary dissection, incorporating pre-specified quilting technique subgroup analyses, and applying GRADE-assessed certainty ratings, this meta-analysis addresses the critical gap left by previous reviews that included breast-conserving surgery and procedures without axillary dissection.

Methods

Study design
The review protocol was prospectively registered with the International Prospective Register of Systematic Reviews (PROSPERO: CRD420251237379). We conducted a GRADE-assessed systematic review and meta-analysis of randomized controlled trials comparing quilting sutures with conventional skin closure following modified radical mastectomy with axillary lymph node dissection in adult women with breast cancer. The review was designed and reported in accordance with the PRISMA 2020 guidelines. No protocol amendments were made after the PROSPERO registration. Only parallel-group RCTs were eligible for quantitative synthesis; non-randomized or single-arm studies were used, if at all, only to inform the qualitative background and were not pooled.

Literature search strategy
A comprehensive electronic search was performed from database inception to June 1, 2025, without any language restrictions. We systematically searched MEDLINE (via PubMed), Embase, Scopus, Web of Science, Cochrane Central Register of Controlled Trials (CENTRAL), ClinicalTrials.gov, WHO International Clinical Trials Registry Platform (ICTRP), and Google Scholar. The search strategies combined controlled vocabulary (for example, MeSH/Emtree) and free-text terms for “breast neoplasms,” “mastectomy,” “modified radical mastectomy,” “quilting,” “flap fixation,” “skin fixation,” “dead space closure,” “seroma,” and “randomized,” using Boolean operators and filters for human studies and clinical trials. The reference lists of relevant trials and prior reviews were manually searched to identify additional studies. The detailed database-specific search strings and number of records retrieved from each source are presented in the Supplementary Table 1.

Inclusion criteria
We applied strict eligibility criteria to select the studies included in this review. We included parallel-group randomized controlled trials enrolling adult women (≥ 18 years) with histologically confirmed breast cancer undergoing modified radical mastectomy (or equivalent total mastectomy with level I–II axillary lymph node dissection). Trials had to compare quilting (or flap fixation/skin fixation) of mastectomy flaps using sutures to conventional nonquilted closure, with both arms using closed-suction drainages. To be eligible for the meta-analysis, studies were required to report at least one of the predefined outcomes of interest (seroma incidence, need for aspiration, drainage volume or duration, or wound-related complications) with sufficient data to derive the effect estimates.

Exclusion criteria
We excluded studies in which quilting was combined with other anti-seroma interventions that were not used in the control arm (e.g., sclerotherapy and fibrin sealants) unless the incremental effect of quilting could be isolated. Non-randomized designs (cohort, case–control, case series), conference abstracts without full text, animal or cadaveric studies, and trials involving breast-conserving surgery, immediate reconstruction, or sentinel node biopsy alone were excluded from the study. Studies restricted to male breast cancer, mixed surgical populations without separable data for mastectomy with axillary dissection, or lacking usable numerical outcome data after attempts to contact the authors were also excluded from the study.

Data extraction and outcome measures
After deduplication, two reviewers (W.M. and M.E.K.) independently screened the titles/abstracts and full texts against the eligibility criteria. In addition to removing duplicate records, we checked for multiple reports of the same trial and potential overlapping datasets by cross-referencing the author groups, study settings, recruitment periods, sample sizes, and trial registry identifiers (when available). We extracted outcomes only once per unique RCT using the most complete report. Data were extracted in duplicate using a piloted form capturing study characteristics, patient demographics, details of the surgical technique (extent of quilting, suture material, axillary quilting, number and placement of drains), and outcomes. Any disagreements at screening or extraction were resolved through discussion with a third senior author (M. B. M. D.). For dichotomous outcomes, we recorded the number of events and participants randomized to each arm. For continuous outcomes, we extracted means, standard deviations, and sample sizes or approximated means/SDs from medians and ranges/IQRs using standard formulas when necessary.
The primary outcome was the incidence of postoperative seroma, as defined by each trial (clinical or ultrasound-detected collection requiring aspiration or causing symptoms). Secondary outcomes included (1) number of seroma aspirations per patient, (2) total aspirated volume, (3) daily and total drain output, (4) duration to drain removal, (5) operative time, and (6) wound-related complications (hematoma, surgical site infection, flap necrosis, wound dehiscence, and other Clavien–Dindo-graded complications). When multiple time points were reported for drainage, we preferentially used the total cumulative volume and time for the final drain removal. We anticipated that trial-level differences in seroma definitions (clinical vs. sonographic), aspiration thresholds, and drain removal criteria could contribute to heterogeneity; therefore, these were considered when interpreting inconsistency and indirectness within the GRADE framework.

Quality assessment
The risk of bias for the included RCTs was independently assessed by W.M. and M.E.K. using the Cochrane RoB 2. The domains evaluated were bias arising from the randomization process, deviations from the intended interventions, missing outcome data, outcome measurement and selective reporting. Each domain and the overall judgment were rated as “low risk,” “some concerns,” or “high risk.” Any discrepancies were adjudicated by M.B.M.D. Risk-of-bias judgments were summarized graphically and used to inform the GRADE certainty ratings for each outcome.

Statistical analysis
All meta-analyses used random-effects models to account for the anticipated clinical and methodological heterogeneity across the trials. For dichotomous outcomes, we calculated risk ratios (RRs) with 95% confidence intervals (CIs) using Mantel–Haenszel random-effects models with restricted maximum likelihood (REML) estimation of τ² and Hartung–Knapp–Sidik–Jonkman (HKSJ) adjustments for small-study uncertainty, where applicable. For continuous outcomes, we used mean differences (MDs) when all studies reported the same measurement scale and standardized mean differences (SMDs) when the scales differed significantly. When only ranges were available, we used the Hozo et al. convention. Studies with zero events in one arm were handled with a treatment-arm continuity correction of 0.5, and the impact of this approach was examined in sensitivity analyses. Between-study heterogeneity was quantified using Cochran’s Q, I², and τ² statistics. Potential small-study effects and publication bias were explored using visual inspection of funnel plots and Egger’s regression test, where ≥ 10 comparisons were available, acknowledging the limited power of these methods when few studies contributed to a given outcome. For the primary outcome of postoperative seroma and the key secondary outcome of the need for aspiration, our primary analyses included all seven randomized trials. Because a single large study (Awad 2019) contributed substantially to between-study heterogeneity without changing the direction of the effect, we also performed a leave-one-out sensitivity analysis excluding this trial. For the GRADE ratings, we used the more conservative pooled estimates from these sensitivity analyses (with lower heterogeneity) to reflect the most stable effect sizes for clinical decision making.

Sensitivity analysis, subgroup analysis, and GRADE certainty of evidence
For the primary outcome (seroma incidence), we performed a leave-one-out sensitivity analysis by sequentially omitting each trial from the random-effects model to evaluate the influence of individual studies on the pooled effect and between-study heterogeneity. Pre-specified technique-based subgroup analyses were undertaken according to the overall quilting strategy: (1) an “aggressive dead-space obliteration” subgroup, comprising trials in which quilting was combined with axillary flap fixation and/or a reduction in the number of drains in the quilting arm (Khater [3], Myint [4], Jayalal [10]; three RCTs); and (2) a “standard pectoral quilting” subgroup, in which quilting was confined to the pectoral fascia and the drain configuration was identical in both arms (Arafa [1], Awad 2019, Özaslan [6], Prakash 2024; four RCTs). Subgroup-specific pooled effects were calculated, and differences between subgroups were explored using a test for interaction, which was interpreted cautiously, given the modest number of studies per stratum. The certainty of evidence for seroma and other key outcomes (need for aspiration, drain duration, and major wound complications) was assessed using the GRADE framework, rated down from high certainty for RCTs when there was an important risk of bias, inconsistency, indirectness, imprecision, or suspected publication bias, and summarized in a GRADE Summary-of-Findings table.

Results
The database search yielded 2237 records, of which 785 duplicates were removed, leaving 1452 unique citations for title/abstract screening (Fig. 1). After excluding 1350 clearly ineligible records, 102 full-text reports were assessed for eligibility. Ninety-five reports were excluded, mainly because they were non-randomized, involved different surgical procedures, or lacked extractable outcome data, leaving seven randomized controlled trials for qualitative and quantitative synthesis (Fig. 1).

Characteristics of included studies
Across seven randomized controlled trials from Egypt, Turkey, Australia, and India, sample sizes per arm ranged from 21 to 400 adult women undergoing modified radical mastectomy with axillary clearance for invasive breast cancer (Table 1). All trials excluded breast-conserving or reconstructive procedures and commonly excluded patients who had received neoadjuvant chemotherapy or had major uncontrolled comorbidities. Seroma incidence was the primary endpoint in each trial, with drainage-related outcomes (volume and duration) and wound-related complications (aspiration, surgical site infection, flap necrosis, hematoma, wound dehiscence, and length of stay) reported as secondary outcomes (Table 1).
Across all seven RCTs (1,412 patients; 706 in the quilting arm and 706 in the conventional closure arm), baseline characteristics were summarized by trial and arm in Table 2. Age and BMI were reported by intervention arm in most trials and were generally similar between groups, whereas breast cancer–related characteristics (e.g., stage distribution) were variably detailed and not consistently reported by arm; where available, most patients had stage II–III disease, and menopausal status was reported in only one trial. Cardiometabolic comorbidities (e.g., diabetes mellitus and hypertension) were inconsistently reported, but the available arm-level data did not indicate clinically important baseline imbalances (Table 2).
Quilting methods differed across trials with respect to suture material, pattern, and the anatomical extent of flap fixation, including whether axillary dead space was formally addressed; only one trial explicitly described formal axillary quilting. Importantly, operative variables that may influence seroma-related outcomes, particularly mastectomy specimen weight and residual skin-flap thickness, were not reported in a group-comparable manner in the included RCTs and therefore could not be summarized quantitatively. Drain management also varied across trials: several studies used identical drain number/placement in both arms, whereas others reduced drains in the quilting arm (often to a single axillary drain), which is clinically relevant for interpreting drainage-related outcomes and the practical implications of quilting (Table 3).

Methodological quality and study integrity
All seven randomized trials were judged to have at least some risk of bias, with one study having a high overall risk (Özaslan 2010) and six having “some concerns” risk (Fig. 2). Although RoB 2 is outcome-specific, the domain-level concerns identified (particularly lack of blinding and non-standardized seroma assessment) were applicable across the key outcomes in these surgical trials; therefore, we present study-level summaries and applied the same domain concerns when considering outcome-level GRADE downgrading. Domain-level assessment showed that bias arising from the randomization process (D1) was generally low, with adequately described sequence generation and allocation concealment in most trials; however, Jayalal 2025 and Prakash 2024 had insufficient detail (some concerns), whereas Özaslan 2010 used an odd–even allocation method, resulting in a high risk in D1. Deviations from the intended interventions (D2) and missing outcome data (D3) were consistently judged as low risk across all studies, reflecting protocol-adherent perioperative care and minimal loss to follow-up. In contrast, all trials were rated as having some concerns regarding the measurement of outcomes (D4) and selection of the reported result (D5), largely due to unblinded outcome assessment, clinical rather than standardized seroma definitions, and a lack of pre-registered statistical analysis plans. Consequently, the overall risk of bias profile was dominated by “some concerns,” with a single high-risk trial explored in the sensitivity analyses (Fig. 2).

Discussion

Primary outcome

Incidence of postoperative seroma
Quilting sutures significantly reduced postoperative seroma compared with conventional closure in the primary analysis, including all seven trials (RR 0.36, 95% CI 0.23–0.55; I²=65%), with the pooled effect clearly favoring quilting and not crossing the line of no effect (Fig. 3). In a pre-specified leave-one-out sensitivity analysis excluding the large Awad 2019 trial, the effect remained strongly in favor of quilting, with a slightly attenuated but more homogeneous estimate (RR 0.43, 95% CI 0.32–0.57; I²=15%). The GRADE certainty assessment and Summary-of-Findings table were based on the primary pooled estimate, including all eligible trials (RR 0.36, 95% CI 0.23–0.55), whereas the RR 0.43 estimate is presented only as a sensitivity analysis.

Secondary outcomes

Need for seroma aspiration
The need for seroma aspiration was also significantly lower in the quilting group (RR 0.29, 95% CI 0.20–0.42), indicating a relative reduction of approximately 70%(Fig. 4). The confidence interval excludes unity, and heterogeneity is low to moderate (I²=30%), suggesting a robust and consistent benefit of quilting for this outcome.

Number of seroma aspirations
Among patients who required aspiration, quilting reduced the mean number of aspirations by approximately one procedure per patient (MD − 1.03, 95% CI − 1.58 to − 0.47), with all individual studies trending in favor of quilting (Fig. 5). However, heterogeneity was very high (I²=95%), indicating substantial variability in the magnitude of this reduction across trials.

Total drainage volume
The total postoperative drainage volume was significantly lower with quilting (MD − 214 mL, 95% CI − 402 to − 27), and the pooled estimate favored quilting, with the CI not crossing zero (Fig. 6). Nevertheless, heterogeneity was extreme (I²=98%), meaning the absolute size of reduction differed widely between studies, and this result should be interpreted with caution. The between-study variance was substantial (τ² = 51963.77; Fig. 6), supporting considerable variability in effect sizes across trials beyond the sampling error.

Duration to drain removal
Quilting showed a nonsignificant trend toward earlier drain removal (MD − 2.01 days, 95% CI − 5.23 to 1.20), with the confidence interval crossing zero and including both potential benefits and no effect (Fig. 7). Heterogeneity was again very high (I²=99%), and no firm conclusion could be drawn for this outcome. The between-study variance was also high (τ² = 13.17; Fig. 7), reinforcing substantial between-trial variability beyond the sampling error.

Surgical-site infection
There was no evidence that quilting altered the risk of surgical-site infection (RR 0.71, 95% CI 0.39–1.32), as the pooled CI crosses 1.0 and the overall effect is statistically nonsignificant (Fig. 8). Heterogeneity was negligible (I²=0%), suggesting that similar SSI rates were observed across all included trials, regardless of the closure technique used.

Flap necrosis
Similarly, quilting did not significantly change the risk of flap necrosis (RR 0.65, 95% CI 0.32–1.32), with the confidence interval spanning a null value (Fig. 9). Heterogeneity was absent (I² = 0%), indicating consistent but nonsignificant differences between the groups in all studies.

Across seven randomized trials, quilting sutures significantly reduced the incidence of postoperative seroma, the need for seroma aspiration, and the number of aspirations, with effect estimates favoring quilting sutures (Figs. 3, 4 and 5). Quilting also resulted in a statistically significant reduction in total drainage volume, although this outcome was accompanied by high heterogeneity (Fig. 6). In contrast, quilting did not significantly shorten the duration of drain removal (Fig. 7). Importantly, quilting did not increase the rates of surgical site infection or flap necrosis, with pooled RRs close to unity and narrow heterogeneity (Figs. 8 and 9). Overall, quilting improved seroma-related outcomes without compromising the safety of the wound.

Sensitivity analysis
As shown in Table 3, removing the large Awad 2019 trial for several outcomes and the outlier Khater 2015 trial for the number of seroma aspirations reduced heterogeneity, particularly for seroma incidence (I² 65%→15%) and need for aspiration (30%→0%), while preserving the clear benefit of the quilting sutures. The magnitude of the effect of the number of aspirations and total drainage volume attenuated slightly but remained statistically significant in favor of quilting. For the duration to drain removal, exclusion of Awad et al.2019 markedly reduced heterogeneity (99%→56%) and converted a non-significant trend into a significant reduction of approximately one day with quilting, suggesting that this endpoint is sensitive to that single influential study. Overall, the leave-one-out analyses supported the robustness of the primary findings, highlighting that a few large or outlying studies contributed disproportionately to the between-study variability.

Subgroup analysis
In the subgroup analysis of postoperative seroma, both quilting strategies significantly reduced seroma compared with conventional closure, with no evidence that the treatment effect differed between the techniques (test for subgroup differences: p = 0.89, I²=0%; Fig. 10). Aggressive dead space obliteration (three trials; n = 104) showed a marked risk reduction (RR 0.31, 95% CI 0.22–0.51; p < 0.00001; I²=9%), whereas standard quilting (four trials; n = 608) also favored quilting but with a slightly smaller effect (RR 0.57, 95% CI 0.36–0.80; p = 0.01; I²=80%). Overall, quilting sutures decreased the seroma risk by approximately two-thirds compared to conventional closure, irrespective of whether an aggressive dead space obliteration technique or standard quilting approach was used (Fig. 10).

GRADE certainty of evidence
Using the primary pooled estimates across all included RCTs, quilting sutures reduced postoperative seroma (RR 0.36, 95% CI 0.23–0.55) and the need for seroma aspiration (RR 0.29, 95% CI 0.20–0.42), corresponding to approximately 200 and 192 fewer events per 1,000 patients, respectively (Table 4). The evidence for reductions in total drainage volume, number of aspirations, and duration to drain removal was directionally consistent but downgraded to low certainty due to substantial heterogeneity and imprecision. For SSI and flap necrosis, event rates were low, and confidence intervals crossed the null, yielding low-certainty evidence without a definitive effect.
Indirectness was not downgraded because all trials evaluated the target population (women undergoing mastectomy with axillary dissection), intervention (quilting/flap fixation), and clinically relevant seroma outcomes. Variability in seroma definitions and drain protocols was considered primarily under inconsistency rather than indirectness. Risk-of-bias downgrading reflected a pervasive lack of blinding/non-standardized outcome assessment, and one trial was at high risk in the randomization process (Özaslan 2010). Publication bias was rated “undetected”; however, because < 10 studies contributed to each outcome, formal statistical assessment (e.g., Egger’s test) was not reliable and was not emphasized.

Downgrading rules used:

Risk of bias – Serious (− 1 level): Most RCTs had “some concerns” due to lack of blinding and outcome measurement, and one trial (Özaslan 2010) had a high risk in the randomization process; outcomes were downgraded once for risk of bias.

Inconsistency – Serious (− 1 level): Marked heterogeneity (I² ≥ ~50% with clearly different point estimates) for the number of aspirations, total drainage volume, and duration until drain removal → downgraded for inconsistency.

Indirectness: Trials directly evaluated the target population, intervention, comparator, and clinically relevant outcomes; variability in seroma/drain protocols was considered under inconsistency.

Imprecision – Serious (− 1 level) when applicable: Wide CIs and/or few events (particularly for SSI and flap necrosis) and outcomes with CIs, including null or broad ranges → downgraded for imprecision.

Publication bias: Rated “undetected”; †formal tests not emphasized because < 10 studies/outcomes.

Discussion
This systematic review and meta-analysis of randomized controlled trials exclusively evaluating quilting sutures versus conventional closure following modified radical mastectomy with axillary lymph node dissection demonstrated a statistically significant and clinically meaningful reduction in postoperative seroma formation [1–7]. A pooled analysis of seven randomized controlled trials encompassing 1,412 patients revealed that quilting reduced seroma incidence by approximately 64% (RR 0.36), with moderate-certainty GRADE evidence translating to 250 fewer seromas per 1,000 patients than conventional closure [8–12]. This benefit exceeds that reported in previous meta-analyses that included heterogeneous surgical populations, such as breast-conserving surgery and procedures without axillary dissection, thereby establishing quilting as a highly effective intervention specifically in the highest-risk surgical context [11, 13, 19–21]. The consistency of the effect across geographically diverse trials from Egypt, Turkey, Australia, and India underscores the generalizability and robustness of these findings across different healthcare settings and surgical practice patterns [1–4, 6–12]. In interpreting these findings, it is important to distinguish the RCT-based results synthesized here from additional observational data and broader quilting literature, which we used to contextualize but not to drive pooled effect estimates.
The mechanistic rationale underlying the efficacy of quilting lies in the systematic obliteration of the dead space between the mastectomy skin flaps and the pectoralis major muscle, thereby eliminating the anatomical reservoir for lymphatic fluid accumulation [1–3, 6, 7, 15, 17, 18]. By mechanically approximating tissue planes with multiple rows of absorbable sutures placed at regular 2–3 cm intervals, quilting addresses the fundamental pathophysiological driver of seroma formation: persistent dead space combined with ongoing lymphatic leakage from disrupted axillary lymphatic channels [3, 6, 7, 11, 14, 16, 17]. This direct anatomical intervention targets the underlying mechanism of seroma formation and complements passive measures such as closed-suction drainage, which evacuates fluid but does not prevent ongoing lymphatic leakage or eliminate dead space [11, 17, 22–24]. While our analysis showed clear reductions in clinically meaningful seroma outcomes (including a lower need for aspiration), drainage-related continuous outcomes, particularly total drainage volume, showed substantial between-trial variability and imprecision. Therefore, the magnitude of drainage reduction should be interpreted cautiously despite the overall direction favoring quilting [2–10, 12, 15, 16].
Interpretation of drainage-related endpoints warrants additional caution because quilting was evaluated largely as an adjunct to closed-suction drainage rather than a drain-free strategy. Across trials, drains were typically used in both arms, and drain number/placement and removal thresholds were not uniform between studies and, in some trials, differed by arm. Consequently, reductions in total drainage volume or time to drain removal may reflect both dead-space obliteration and differences in drain strategy, and should not be interpreted as evidence that quilting alone eliminates the need for drainage. In contrast, seroma incidence and aspiration outcomes more directly reflect the clinical effectiveness of dead-space closure and are less dependent on drain configuration.
Our prespecified subgroup analysis comparing aggressive dead space obliteration (combining pectoral and axillary quilting with drain reduction) versus standard pectoral quilting revealed that both strategies significantly reduced seroma formation, with no statistically significant difference between the techniques (test for interaction p = 0.89) [3, 4, 10, 12]. This finding suggests that the primary benefit derives from pectoral area quilting, which addresses the largest potential space after mastectomy, whereas the additional complexity of comprehensive axillary quilting may not confer a proportional incremental benefit in all clinical contexts [11, 14, 17, 18, 25]. However, the aggressive dead space obliteration subgroup demonstrated lower heterogeneity (I²=9%) than the standard quilting subgroup (I²=80%), suggesting that more comprehensive anatomical coverage combined with intentional drain reduction may yield more consistent outcomes across different surgical practices [3, 4, 10, 12, 14]. These observations support a tailored approach, wherein surgeons may confidently adopt pectoral quilting as a minimum standard while considering extended axillary quilting for patients at particularly high risk of seroma formation, such as those with an elevated body mass index, extensive axillary dissection, or significant comorbidities [8, 11, 14, 16, 17, 26].
The clinical implications of seroma reduction extend far beyond immediate postoperative recovery, directly impacting oncological care pathways and patient quality of life throughout the breast cancer treatment continuum [3, 4, 11–14]. Seroma-related complications can delay the initiation of adjuvant chemotherapy and radiotherapy [3, 4, 11, 12, 14, 24]. By reducing seroma formation and the need for repeated aspirations, quilting may support timely progression through postoperative care pathways and may reduce follow-up visits and associated patient burden; however, these downstream implications were not directly measured as trial endpoints in the included RCTs [4, 11–13–14, 24, 27, 28–32]. Furthermore, the significant reduction in the number of aspirations per patient (MD − 1.03) translates to fewer needle punctures through healing surgical sites, consequently lowering the risk of secondary infection, chronic pain, and psychological distress associated with prolonged wound management [1, 2–3, 7–9–10, 12, 14, 15–16].
Our analysis demonstrates that quilting reduces seroma without compromising wound healing safety, as evidenced by the nonsignificant differences in surgical site infection (RR 0.71) and flap necrosis rates (RR 0.65) between the quilting and conventional closure groups [1–4, 6–10, 12, 13, 24]. This safety profile directly addresses longstanding surgical concerns that increased suture burden and tissue manipulation during quilting may impair skin flap perfusion or introduce an additional infection risk [11, 13, 17, 25, 27]. The absence of increased flap necrosis despite mechanical fixation to the underlying fascia provides reassurance that quilting, when performed with an appropriate surgical technique and avoiding excessive tension on the suture lines, preserves an adequate blood supply to elevated skin flaps [6, 15, 18, 24, 25]. Moreover, the modest increase in operative time associated with quilting (approximately 10–15 min based on the included trials) represents a clinically acceptable trade-off given the substantial downstream benefits of reduced postoperative complications and healthcare utilization [4, 11–14, 16, 24].
Our sensitivity analyses employing the leave-one-out approach revealed that the primary findings remained robust even when individual studies were sequentially excluded, although the large Awad et al. trial exerted considerable influence on the between-study heterogeneity for several outcomes [2, 11, 13, 17, 18]. Removal of this 800-patient trial reduced heterogeneity for seroma incidence from 65% to 15% while preserving the direction and statistical significance of the treatment effect, suggesting that the observed variability primarily reflects differences in outcome measurement methodology rather than true effect modification [2, 11, 13]. Similarly, exclusion of the Khater trial, which employed a unique running suture technique with formal axillary quilting, substantially reduced heterogeneity in the number of aspirations while maintaining a significant benefit [3, 11, 13, 14, 17]. These sensitivity observations underscore the importance of standardizing seroma definitions and measurement protocols in future trials, as inconsistent diagnostic criteria (clinical palpation versus routine ultrasonography) contribute to the apparent effect heterogeneity across studies [11, 13, 17, 19–21].
The GRADE certainty of evidence assessment provides a transparent evaluation of confidence in effect estimates, with moderate-certainty evidence supporting the efficacy of quilting for seroma reduction and the need for aspiration. In contrast, continuous outcomes such as drainage volume and duration were downgraded to low certainty due to substantial heterogeneity and imprecision [11, 13]. The downgrading of the risk of bias primarily reflects unavoidable limitations in the blinding of surgical interventions and outcome assessment, as well as one high-risk trial employing non-random odd-even allocation [6, 11, 13, 17]. Importantly, no evidence of publication bias was detected through visual inspection of funnel plots and Egger’s regression testing, although the limited number of included trials (n = 7) constrained the statistical power of these assessments [11, 13, 19–21]. The moderate-certainty GRADE ratings for primary outcomes indicate that future research is unlikely to substantially change our confidence in the estimated treatment effect, thereby supporting the integration of quilting into standard surgical practice for modified radical mastectomy with axillary dissection [11, 13, 17, 20, 21].

Strengths and limitations
This review has important strengths: it is an RCT-only synthesis restricted to modified radical mastectomy with complete axillary lymph node dissection, reducing clinical heterogeneity seen in prior reviews; it includes a prespecified quilting technique subgroup analysis and leave-one-out sensitivity analyses; and it applies RoB2 and GRADE to support transparent certainty judgments.
Limitations include technical and perioperative variability across trials (extent of quilting, suture approach, and drain protocols), unavoidable blinding constraints inherent to surgical RCTs, including one trial with a high risk of randomization, and substantial heterogeneity and imprecision for drainage-related continuous outcomes. Importantly, the included RCTs predominantly evaluated quilting as an adjunct to closed-suction drainage (and in some trials with fewer drains in the quilting arm); therefore, the current evidence supports the benefit in a “drained” setting but does not determine whether drains can be safely omitted when quilting is performed. The limited number of trials constrained formal publication bias assessment, and long-term patient-centered outcomes were inconsistently reported. Future multicenter RCTs with standardized seroma definitions and drain removal criteria are needed.

Conclusions
This GRADE-assessed systematic review and meta-analysis provides moderate-certainty evidence that quilting sutures significantly reduce postoperative seroma formation (RR 0.36, 95% CI 0.23–0.55; ≈200 fewer seromas per 1,000 patients) and the need for aspiration (RR 0.29, 95% CI 0.20–0.42; ≈192 fewer patients requiring aspiration per 1,000) following modified radical mastectomy with axillary lymph node dissection. Quilting also reduced the number of aspirations and total drainage volume, without increasing the rates of surgical site infection or flap necrosis compared with conventional closure. Prespecified subgroup analysis demonstrated that both aggressive dead space obliteration (combining pectoral and axillary quilting with drain reduction) and standard pectoral quilting significantly reduced seroma formation, with no statistically significant difference between the techniques, suggesting that pectoral area quilting alone provides substantial benefit, whereas extended axillary quilting may be reserved for high-risk patients. Sensitivity analyses confirmed the robustness of the findings across different study populations and quilting techniques, while the GRADE assessment provided a transparent evaluation indicating that future research is unlikely to substantially change confidence in the estimated treatment effect for primary outcomes. Therefore, quilting sutures should be considered for routine use in this setting as a simple, low-cost, and safe intraoperative technique that directly targets the pathophysiological mechanism of seroma formation through mechanical dead space obliteration. Future multicenter RCTs using standardized seroma definitions, aspiration thresholds, and drain removal criteria would meaningfully strengthen the evidence base and refine the optimal quilting technique parameters.

Supplementary Information