Comparative outcomes of robotic versus laparoscopic inguinal hernia repair: a systematic review and meta-analysis

Abstract

To compare clinical, operative, patient-reported, and economic outcomes of robotic versus laparoscopic inguinal hernia repair in adults. We conducted a PRISMA 2020-compliant systematic review and meta-analysis of PubMed, Web of Science, and Google Scholar through June 13, 2025. The protocol was retrospectively registered in PROSPERO (CRD420251075296). Comparative randomized and observational studies of elective robotic versus laparoscopic inguinal hernia repair in adults were eligible. Random-effects meta-analyses used restricted maximum-likelihood estimation with Hartung-Knapp confidence intervals; risk of bias was assessed using RoB 2 and MINORS, and certainty was evaluated using GRADE. Nineteen reports representing 18 unique cohorts and 78,940 participants were included. Nine reports (3,200 participants) contributed recurrence data; robotic repair was associated with a lower observed recurrence risk (RR 0.32, 95% CI 0.17–0.59; I² = 0%), although certainty was low and the randomized evidence was imprecise. Robotic repair required longer operative time (11 studies; MD + 30.04 min, 95% CI 9.87–50.20; I² = 99.6%); after excluding the influential Holleran 2022 cohort, the estimate was + 22.09 min (95% CI 8.87–35.30). Pooled immediate and postoperative-day-1 pain estimates did not show a statistically significant difference. Length of stay was not significantly different in the primary analysis (MD + 0.33 days, 95% CI -0.03 to 0.70), while an influence analysis suggested a small increase after robotic repair (+ 0.18 days). Overall complication estimates were unstable, and no clear differences were found for hematoma, surgical-site infection, or urinary retention. Robotic repair was consistently more expensive across heterogeneous healthcare settings. Robotic inguinal hernia repair was associated with lower observed recurrence in predominantly non-randomized evidence, but this finding should not be interpreted as proof of causal superiority. Robotic repair generally required approximately 22–30 additional operative minutes and incurred higher costs, while pooled pain, length-of-stay, and complication findings were uncertain or clinically small. Technique selection should therefore consider patient and hernia characteristics, surgeon experience, institutional resources, and the low to very low certainty of the available evidence.

Supplementary Information

The online version contains supplementary material available at 10.1007/s11701-026-03728-4.

Introduction

Inguinal hernia repair is among the most frequently performed general surgical procedures worldwide, with millions of operations undertaken each year [1]. Over the past three decades, the field has shifted from traditional open repair to minimally invasive approaches. Laparoscopic techniques—particularly the transabdominal preperitoneal (TAPP) and totally extraperitoneal (TEP) methods—have become well-established due to their association with reduced postoperative pain, lower wound infection rates, faster recovery and improved cosmetic outcomes [2, 3]. Robotic-assisted surgery represents the latest evolution in minimally invasive hernia repair.

Robotic TAPP (R-TAPP) and robotic TEP (R-TEP) offer enhanced visualization, tremor filtration, and superior instrument articulation, improving the surgeon’s ability to perform precise dissection and suturing [4]. Due to these capabilities, there is growing interest in whether robotic platforms offer meaningful clinical benefits over traditional laparoscopy, particularly in complex, bilateral, or recurrent hernia cases [4, 5]. Despite its appeal and technical promise, the adoption of robotic-assisted techniques remains limited due to concerns regarding longer operative times, higher cost, learning curve, and uncertainties surrounding long-term outcomes. Current evidence remains inconsistent: while some studies report improved postoperative outcomes such as reduced pain and faster return to activity, others show minimal clinical advantages when compared to laparoscopy – often at significantly higher costs [6]. Given these inconsistencies, healthcare systems face difficulties justifying the expense of robotic platforms, and surgeons lack clear guidance for selecting the optimal technique in complex cases [7].

We compared robotic-assisted and laparoscopic inguinal hernia repair in adults across operative, clinical, patient-reported, and economic outcomes. We also examined robustness using leave-one-out and influence analyses, study-design and methodological-quality subgroups, and outcome-specific GRADE assessments.

Methods

This review followed PRISMA 2020 guidance [8]. The protocol was submitted to PROSPERO on February 4, 2026 and registered as CRD420251075296 [9]. Because review activities began on June 14, 2025, the registration was retrospective. Ethical approval was not required because only previously published aggregate data were analyzed.

Inclusion criteria

Studies were included if they:

1. Were randomized controlled trials (RCTs) or prospective/retrospective comparative cohort studies;

2. Involved adults aged 18 years or older undergoing elective robotic or laparoscopic inguinal hernia repair (R-TAPP, R-TEP, TAPP, or TEP);

3. Were published in English from 2010 onwards; and

4. Reported at least one primary or secondary outcome.

Exclusion criteria

We excluded studies that:

1. Involved patients younger than 18 years;

2. Focused on open repair or did not provide clear comparison between robotic and laparoscopic groups;

3. Were non-comparative, including case reports or single-arm series;

4. Were reviews, editorials, or conference abstracts without accessible full text; or

5. Represented duplicate publications.

Search strategy

The final search was conducted on June 13, 2025. The PubMed search was: (("Hernia, Inguinal"[Mesh] OR "inguinal hernia"[tiab] OR "inguinal hernias"[tiab]) AND ("Robotic Surgical Procedures"[Mesh] OR robot*[tiab] OR "robot-assisted"[tiab] OR "robotic-assisted"[tiab]) AND ("Laparoscopy"[Mesh] OR laparoscop*[tiab] OR TAPP[tiab] OR TEP[tiab] OR "transabdominal preperitoneal"[tiab] OR "totally extraperitoneal"[tiab])). Filters were English language, adults, and publication from January 1, 2010 through June 13, 2025. No study-design filter was applied.

The Web of Science Core Collection search was: TS = (("inguinal hernia" OR "inguinal hernias") AND (robot* OR "robot-assisted" OR "robotic-assisted") AND (laparoscop* OR TAPP OR TEP OR "transabdominal preperitoneal" OR "totally extraperitoneal")). Results were refined to articles in English published from 2010 through June 13, 2025.

Google Scholar was searched using the reproducible queries "robotic inguinal hernia repair" AND "laparoscopic inguinal hernia repair" and "robot-assisted" inguinal hernia TAPP TEP outcomes. The first 250 results sorted by relevance were screened, consistent with the prespecified Google Scholar yield used in the PRISMA count.

The PICOTS framework was defined as follows: population, adults undergoing elective inguinal hernia repair; intervention, robotic R-TAPP or R-TEP; comparator, laparoscopic TAPP or TEP; primary outcomes, operative time, postoperative pain, and recurrence; secondary outcomes, hospital stay, complications, costs, return to work or activity, and other patient-reported outcomes; timing, perioperative assessment through the longest reported follow-up; and setting, randomized or comparative prospective/retrospective hospital, registry, or administrative-database studies [10]. Only English-language reports were eligible,the potential for language bias was considered in the certainty assessment and limitations. Full database strategies are reproduced in Supplementary Table S1.

Study selection

Two reviewers (H.N. and A.H.) independently screened titles and abstracts in Rayyan [11], and two reviewers (J.A. and R.W.) independently assessed potentially eligible full texts. Disagreements were resolved by discussion and consensus. Rayyan was used for exact and near-duplicate detection. Google Scholar records were additionally cross-checked against PubMed and Web of Science records using DOI or PMID when available, title, first author, publication year, and cohort characteristics. Multiple publications from the same cohort were linked to one study record and were not counted as independent participants or entered simultaneously in the same meta-analysis. The selection process is shown in Fig. 1.

Fig. 1.

PRISMA 2020 flow diagram. Of 873 records identified, 310 duplicates were removed; 563 records were screened, 51 reports were sought for retrieval, 25 were not retrieved, 26 full-text reports were assessed, and 19 reports representing 18 unique studies were included

Outcome measures

Primary outcomes were operative time in minutes, postoperative pain at prespecified time windows, and hernia recurrence at the longest reported follow-up. Pain was pooled only when instruments and time points were clinically compatible; standardized mean differences were used when VAS and NRS measured the same construct.

Secondary outcomes were length of hospital stay in days, overall postoperative complications, seroma, hematoma, surgical-site infection, urinary retention, return to work or physical activity, quality of life, satisfaction or chronic discomfort, and economic outcomes. Complications were pooled by consistently reported event categories; outcomes that could not be harmonized were synthesized narratively.

Data extraction and quality assessment

Two reviewers independently extracted study and report identifiers, study design, country and setting, sample size, demographic and comorbidity data, hernia characteristics (primary/recurrent and unilateral/bilateral), operative approach, follow-up, operative time, pain instruments and time points, recurrence, individual complications, patient-reported outcomes, return to activity, and cost perspective and components. A third reviewer checked the completed dataset. Overlapping publications were mapped to their parent cohort to prevent double-counting. Disagreements were resolved by consensus.

Risk of bias was assessed independently by two reviewers. RoB 2 was applied at the outcome/report level to the primary and long-term publications of the single RIVAL randomized cohort [12], and MINORS was applied to the 17 observational cohorts [13]. Disagreements were resolved by consensus or consultation with a senior reviewer. The RoB 2 traffic-light plot is presented in Fig. 2 and MINORS scores in Table 3.

Fig. 2.

RoB 2 traffic-light plot for the primary [4] and long-term [24] reports from the single RIVAL randomized cohort. Green indicates low risk and yellow indicates some concerns

Table 3.

MINORS assessment of the 17 observational cohorts

Study ID	Study Design	A clearly stated aim	Inclusion of consecutive patients	Prospective collection of data	Endpoint appropriate to the aim of the study	Unbiased assessment of the study endpoint	Follow-up period appropriate to the aim	Loss to follow-up < 5%	Prospective calculation of study size	An adequate control group (Comparative studies only)	Contemporary groups (Comparative studies only)	Baseline equivalence of groups (Comparative studies only)	Adequate statistical analyses (Comparative studies only)	Total Score
Zayan et al., 2019 [3]	Retrospective cohort study	2	1	0	2	1	2	1	0	2	1	1	2	13
Ayuso et al., 2023 [6]	Case matched compartive study	2	2	1	2	2	2	2	0	2	2	1	2	20
Baginski et al., 2024 [22]	Single-center retrospective review	2	1	0	2	1	1	2	0	1	1	2	2	15
Wadhawan et al., 2025 [23]	Prospective non-randomized comparative study	2	2	2	2	1	1	2	0	2	2	2	1	17
Tatarian et al., 2021 [24]	Retrospective cohort study	2	1	0	2	1	2	2	0	2	1	2	2	17
Kudsi et al., 2023 [25]	Retrospective cohort	2	1	0	2	1	2	1	0	2	2	1	2	16
Okamoto et al., 2023 [26])	Retrospective cohort study	2	1	0	2	1	1	2	0	2	2	2	2	17
Hinojosa-Ramirez et al., 2024 [7]	Retrospective single-center comparative study	2	1	0	2	1	1	2	0	2	2	1	2	16
Muysoms et al., 2021 [27])	Retrospective single-centre comparative study	2	1	0	2	1	1	2	0	2	2	1	2	16
Pai et al., 2025 [28])	Retrospective observational study	2	1	0	2	1	1	2	0	1	2	1	2	15
Choi et al., 2023 [29])	Retrospective comparative cohort	2	2	1	2	1	1	2	0	1	2	1	2	17
Chao et al., 2024 [30]	Multicenter retrospective observational cohort	2	2	0	2	1	1	2	0	2	2	2	2	18
Brosnihan et al., 2025 [31])	Retrospective	2	2	0	2	1	1	2	0	2	2	1	2	17
Peltrini et al., 2023 [32]	Multicenter retrospective observational cohort	2	2	0	2	1	2	2	0	2	2	2	2	19
Khoraki et al., 2020 [5]	retrospective cohort	2	2	0	2	1	1	2	0	2	2	1	2	17
Gerdes et al., 2022 [2])	Prospective comparative cohort	2	2	0	2	1	1	2	0	2	2	1	2	17
Holleran et al., 2022 [1]	Retrospective database cohort	2	2	0	2	1	1	2	0	2	2	1	2	17

MINORS items were scored 0 (not reported/inadequate), 1 (reported but inadequate), or 2 (reported and adequate), for a maximum 2 comparative-study score of 24

Certainty of evidence assessment

Certainty of evidence was evaluated for each pooled outcome using GRADE across risk of bias, inconsistency, indirectness, imprecision, and publication bias [14]. Observational evidence began at low certainty and was downgraded when additional serious limitations were present. Randomized and observational evidence were considered together without double-counting overlapping reports. Outcome-specific judgments and downgrade rationales are summarized in Table 5.

Table 5.

GRADE certainty assessment of outcomes comparing robotic and laparoscopic inguinal hernia repair

Outcome	Studies (participants)	Relative/absolute effect	Risk of bias	Inconsistency	Indirectness	Imprecision	Publication bias	GRADE certainty
Recurrence	9 (3,200)	RR 0.32 (95% CI 0.17–0.59); about 19 fewer recurrences per 1,000 patients (11–23 fewer). I2 = 0%; prediction interval 0.16–0.66	Serious1	Not serious	Not serious	Not serious	Undetermined (< 10 studies)	LOW
Operative time	11 (26,484)	MD + 30.04 min (95% CI 9.87–50.20). After excluding influential Holleran 2022: + 22.09 min (8.87–35.30)	Serious1	Very serious2	Serious2	Not serious	Not detected, but low power (Egger P = 0.506)	VERY LOW
Immediate postoperative pain (operation day/PACU)	2 (1,253)	SMD 0.02 (95% CI − 0.66 to 0.70); I2 = 0%. The CI includes meaningful benefit and harm	Serious1	Not serious	Serious3	Very serious3	Undetermined	VERY LOW
Postoperative pain (POD1)	3 (343)	SMD − 0.47 (95% CI − 1.56 to 0.63); I2 = 81.3%. VAS-only sensitivity: SMD − 0.72 (− 1.81 to 0.36)	Serious1	Serious4	Serious4	Very serious4	Undetermined	VERY LOW
Length of hospital stay	9 (26,111)	MD + 0.33 days (95% CI − 0.03 to 0.70). After excluding Holleran 2022: + 0.18 days (0.06–0.29)	Serious1	Very serious5	Serious5	Serious5	Undetermined	VERY LOW
Any postoperative complication	6 (39,156)	RR 1.68 (95% CI 1.00–2.84); about 10 more events per 1,000 (0–26 more). I2 = 93.4%; excluding Holleran 2022: RR 1.18 (0.90–1.54)	Serious1	Very serious⁶	Serious⁶	Very serious⁶	Undetermined	VERY LOW
Seroma	3 (24,383)	RR 4.41 (95% CI 2.92–6.66); about 7 more seromas per 1,000 (4–11 more). Excluding Holleran 2022: RR 1.85 (0.08–42.23)	Serious1	Serious⁷	Not serious	Serious⁷	Undetermined	VERY LOW
Hematoma	5 (1,618)	RR 0.77 (95% CI 0.43–1.40); about 7 fewer hematomas per 1,000 (18 fewer to 13 more). I2 = 0%	Serious1	Not serious	Not serious	Serious⁸	Undetermined	VERY LOW
Surgical-site infection	5 (1,840)	RR 0.80 (95% CI 0.26–2.42); about 2 fewer infections per 1,000 (6 fewer to 11 more). I2 = 0%	Serious1	Not serious	Not serious	Very serious⁸	Undetermined	VERY LOW
Postoperative urinary retention	4 (667)	RR 0.74 (95% CI 0.39–1.41); about 16 fewer events per 1,000 (38 fewer to 26 more). I2 = 0%	Serious1	Not serious	Not serious	Serious⁸	Undetermined	VERY LOW

GRADE categories: high, moderate, low, and very low certainty. Low certainty means the true effect may differ substantially from the estimate; very low certainty means the true effect is likely to differ substantially

¹Most evidence was contributed by observational cohorts susceptible to residual confounding, surgeon-selection effects, learning-curve effects, and outcome ascertainment bias. This limitation established the low starting certainty and was not double-counted

²Operative-time estimates differed markedly across operative definitions, robotic and laparoscopic approaches, laterality, surgeon experience, and database settings (I²=99.6%).

³Immediate pain was informed by only two studies using different instruments/timing, and one study required conversion from median/IQR to mean/SD; the CI spanned important benefit and harm

⁴POD1 pain combined VAS and NRS data, showed substantial heterogeneity (I²=81.3%), and had a very wide CI crossing both bene fit and harm

⁵Hospital stay was strongly influenced by one national database study, showed extreme heterogeneity (I²=96.2%), and was affected by different outpatient and admission pathways

⁶Complication definitions varied; the pooled estimate was unstable, the prediction interval was wide (0.34–8.39), and exclusion of Holleran 2022 removed heterogeneity and attenuated the effect.

⁷The seroma estimate was dominated by Holleran 2022 and became extremely imprecise and non-significant when that study was removed.

⁸Sparse events and wide confidence intervals included clinically important benefit and harm; surgical-site infection involved only 12 total events.

Statistical analysis

Meta-analysis was performed in R using the meta and metafor packages. Dichotomous outcomes were pooled as risk ratios (RRs), and continuous outcomes as mean differences (MDs), with 95% confidence intervals (CIs). Standardized mean differences using Hedges g were used when the same construct was measured on different continuous scales. Random-effects models used restricted maximum-likelihood estimation and Hartung-Knapp confidence intervals [15]. Heterogeneity was quantified using Cochran Q, I², and τ².

Medians, interquartile ranges, ranges, confidence intervals, or standard errors were converted to means and standard deviations only when established methods were applicable [16, 17]. Units were harmonized before pooling,otherwise, findings were synthesized narratively. Continuity corrections and alternative sparse-data models were explored for dichotomous outcomes, and double-zero studies were retained descriptively and examined using risk-difference sensitivity analyses.

Primary analyses included all eligible, non-overlapping reports for each outcome. For outcomes with I² greater than 50%, prespecified leave-one-out and influence diagnostics were used to identify whether a single study materially altered the pooled estimate or heterogeneity; exclusions were based on influence diagnostics rather than statistical significance. Additional sensitivities excluded the randomized cohort, converted summary data, and reports with specific data-quality concerns. Subgroup analyses by methodological quality/design and hernia classification were performed only when at least two informative studies per subgroup were available. Funnel plots and Egger tests were restricted to outcomes with at least 10 independent studies. Two-sided P < 0.05 was considered statistically significant. Costs were synthesized narratively because currencies, price years, perspectives, and included cost components were not comparable.

Results

Study selection

A total of 873 records were identified. After 310 duplicates were removed, 563 records underwent title and abstract screening and 512 were excluded. Fifty-one reports were sought for retrieval; 25 could not be retrieved, leaving 26 full-text reports assessed for eligibility. Seven reports were excluded (3 non-comparative or mismatched designs [17–19], 2 ineligible populations [20, 21], and 2 with insufficient or unreported outcomes [22, 23]). Nineteen reports representing 18 unique studies met the inclusion criteria [1–7, 24–35] (Fig. 1).

Characteristics of included studies

The 19 included reports represented 18 unique cohorts and 78,940 unique adults: 17 observational cohorts and one randomized cohort reported in the primary RIVAL publication and its long-term follow-up. Eleven reports originated in the United States, and one each originated in India, Japan, Spain, Belgium, South Korea, Taiwan, Italy, and Switzerland. Sex was reported for 54,560 unique participants: 50,319 males (92.2%) and 4,241 females (7.8%); 24,380 participants (30.9%) were from reports that did not provide sex distribution. Studies were published from 2019 to 2025. Table 1 summarizes study characteristics and identifies the two RIVAL reports from the same randomized cohort; Table 2 summarizes operative approaches and follow-up.

Table 1.

Characteristics of 19 included reports representing 18 unique study cohorts

Study	Publication year	Study design	Country	Sample size	Robotic group	Laparoscopic group	Age range	Baseline comorbidities	Selected outcomes and reported P values
Zayan et al. [3]	2019	Retrospective cohort study	USA	105	37 (Male; 37) (Female; 0)	68 (Male; 59) (Female; 9)	18 to 89	HTN: LIHR: 16, RIHR: 16, CAD: LIHR: 6, RIHR: 4, DM: LIHR: 5, RIHR: 2	Operative time: RIHR vs. LIHR ( p < 0.001), Recurrence: (p = 0.295), 1-year CCS: (p = 0.296), Follow-up duration: (p = 0.019)
Ayuso et al. [6]	2023	Case-matched comparative study	USA	282	141 (Male; N/A) (Female; N/A)	141 (Male; N/A) (Female; N/A)	NR	LIHR group: 2.9 ± 2.5, RIHR group: 2.6 ± 2.2 ( DM: RIHR: 5%, LIHR: 7.8%) (COPD: RIHR: 0%, LIHR: 2.1%) (CAD: RIHR: 5.7%, LIHR: 5%)	Operative Time (LIHR vs RIHR): p = 0.36 Total Charges (LIHR vs RIHR): p < 0.01 Recurrence (LIHR vs RIHR): P = 0.38
Miller et al. [21]	2023	Randomized controlled trial follow-up report (RIVAL)	USA	102	48 (Male; N/A) (Female; N/A)	54 (Male; N/A) (Female; N/A)	NR	NR	VAS 1-year: p = 0.30 VAS 2-year: p = 0.17 Recurrence: p = 1.0 Physical activity: p > 0.05 at all time points
Baginski et al. [22]	2024	Single-center retrospective review	USA	298	245 (Male; 223) (Female; 22)	53 (Male: 49; Female: 4)	NR	Diabetes: L-TEP: 6, R-TAPP: 27 COPD: L-TEP: 0, R-TAPP: 2 Hypertension: L-TEP: 23, R-TAPP: 104 ESRD: L-TEP 0, R-TAPP 3 (1.2%) Immunosuppressed: L-TEP: 1, R-TAPP: 16 Smoking (n): L-TEP:3, R-TAPP: 22	Hernia recurrence: R-TAPP group (1.6%) vs. L-TEP (9.3%), P = 0.01 , Postoperative pain: R-TAPP (5.3%) vs. L-TEP (13.2%), P = 0.01 , complications: R-TAPP (4.5%) vs. L-TEP (14.8%), P = 0.01
Wadhawan et al. [23]	2025	Prospective non-randomized comparative study	India	83	24 (Male; 24) (Female; 0)	59 (Male; 55) (Female; 4)	NR	N/A	Operative time: P < 0.001; VAS pain: P < 0.05 for reported early comparisons; 1-month CCS: P = 0.0118
Tatarian et al. [24]	2021	Retrospective database cohort with propensity-adjusted analyses	USA	36,124	559 (Male; 504) (Female; 55)	35,565 (Male; 33,032) (Female; 2533)	23–90 years	comorbidity: RIHR 65.3% LIHR: 38%, HTN: RIHR: 43.7% LIHR: 23.8%, COPD: RIHR: 10.9% LIHR: 5.9%, DM: RIHR: 9.1% LIHR: 4.9%, Obesity: RIHR: 7.5% LIHR: 2.4%	p-values: Complications: R-IHR 9.3% vs L-IHR 1.1% ( highest in RIHR): p < 0.0001 LOS: R-IHR 1.74 days vs L-IHR 0.19 days (longest in RIHR): p < 0.0001 Recurrence R-IHR 6.5% vs L-IHR 3.9% ( higher in RIHR) p = 0.0455. No significant difference in complications, readmission, or length of stay (p > 0.1) Recurrence was higher in robotic (p = 0.0455) Unadjusted results showed longer hospital stay and higher complication, readmission, and ED visit rates in robotic (all p < 0.0001)
Kudsi et al. [25]	2023	Retrospective cohort	USA	1,153	547 (Male; 500) (Female; 47)	606 (Male; 568) (Female; 38)	Both groups; (20–90)	LIHR 250; RIHR 224 CAD: LIHR 47; RIHR 56 DM: LIHR 56; RIHR 58	Operative time (p < 0.001) Peritoneal breach (p < 0.001) Recurrence (p = 0.013)
Okamoto et al. [26]	2023	Retrospective cohort study with propensity-score matching	Japan	160	80 (Male; 76) (Female; 4)	80 (Male; 75) (Female; 5)	61–75 years	Matched group: DM: RIHR:3, LIHR: 4, unmatched: DM: RIHR: 3, LIHR: 59	Operative time: p = 0.087 Pain: p = 0.11 Recurrence: p = 1.00 Learning curve (experience vs operative time): p < 0.001
Hinojosa-Ramirez et al. [7]	2024	Retrospective single-center comparative study	Spain	98	20 (Male; 19) (Female; 1)	78 (Male; 61) (Female; 17)	NR	ASA Classification) - Robotic (n = 20): ASA I: 1 (5%) ,ASA II: 14 (70%) ,ASA III: 5 (25%) -Laparoscopic (n = 78): ASA I: 6 (7.8%) ,ASA II: 56 (72.7%) ,ASA III: 15 (19.5%)	Longer surgical time and higher cost in robotic group (p < 0.001) , Longer hospital stay in robotic group (p = 0.007)
Muysoms et al. [27]	2021	Retrospective single-centre comparative study	Belgium	676	404 (Male; 377) (Female; 27)	272 (Male; 237) (Female; 35)	18–90 years	COPD: 21 (lap), 28 (robotic) DM: 11 (lap), 25 (robotic) Renal disease: 5 (lap), 6 (robotic) AAA: 2 (lap), 2 (robotic) Liver disease: 1 (lap), 2 (robotic)	Cost difference: P < 0.001, Wound infections (lower in robotic): P = 0.001 ,Hospital stay (shorter in robotic): P = 0.014
Pai and Hsieh [28]	2025	Retrospective observational study	USA	647	215 (Male; 193) (Female; 22)	432 (Male; 311) (Female; 121)	NR	(HTN RIHR: 149, LIHR:290) (DM RIHR:60, LIHR: 115) (Chronic pulmonary disease RIHR:43, LIHR: 106) (Chronic kidney disease RIHR: 22, LIHR: 33) (Rheumatic disease RIHR: 3, LIHR: 6)	-Complications: p < 0.019 - Length of stay:p < 0.001 - Hospital costs: p < 0.001
Choi et al. [29]	2023	Retrospective comparative cohort	South Korea	100	50 (Male; 50) (Female; 0)	50 (Male; 49) (Female; 1)	~ 20–85 years	HTN: 44, DM: 14, Cardiovascular: 15, Pulmonary: 10, Liver: 6, Renal:0, Cerebral:4, Other cancer history: 10, BPH: 20	- Postoperative pain score (POD1): RIHR lower than LIHR (p = 0.015) - Surgical cost: RIHR significantly higher than LIHR (p < 0.001)—ASA score: p = 0.002—HTN prevalence: p = 0.001
Chao et al. [30]	2024	Multicenter retrospective observational cohort	Taiwan	1,042	279 (Male; 255) (Female; 24)	763 (Male; 696) (Female; 67)	> 18	25 patients (7 RIHR + 18 LIHR) had CCI 3–7, 1017 ( 272 RIHR + 745 LIHR) had CCI 0–2	- Operative time: p < 0.001 - Opioid use (pain proxy): p = 0.001 - Reoperation rate: p = 0.016
Brosnihan et al. [31]	2025	Retrospective study	USA	13,611	7306 (Male; 6698) (Female; 608)	6305 (Male; 5747) (Female; 558)	27- 89 years	(DM RIHR:664, LIHR:535) (COPD RIHR:170, LIHR: 137) (CHF RIHR: 157, LIHR: 111) (HTN RIHR: 2894, LIHR: 2178) (ESRD on HD RIHR: 26, LIHR: 16) (Metastatic cancer RIHR:15, LIHR: 14) (bleeding disorder RIHR: 95, LIHR: 112)	-Operative time (RIHR vs LIHR/OIHR): RIHR longer p- value < 0.001 -30 day morbidity (RIHR vs OIHR):RIHR and LIHR lower (1.0% vs 1.6%) p-value 0.001 -30 day mortality (RIHR vs OIHR): Similar (very low overall), p-value 0.004 -Readmission rate: RIHR lowest (1.2%),p-value < 0.001 -Superficial SSI: RIHR and LIHR lower than OIHR, p-value < 0.011 -Reoperation rate (30-day): p-value < 0 .556
Peltrini et al. [32]	2023	Multicenter retrospective observational cohort	Italy	120	40 (Male; 35) (Female; 5)	80 (Male; 71) (Female; 9)	N/A	(Cardiovascular RTAPP: 16, LTAPP: 32) (Pulmonary RTAPP: 9, LTAPP: 15) (Urologic RTAPP: 8, LTAPP: 16) (Gastrointestinal RTAPP: 9, LTAPP: 12)	- Operative time: p < 0.001—Length of stay: p < 0.001—Pain score (VAS): p = 0.13—Recurrence: None in either group
Khoraki et al. [5]	2020	Retrospective cohort	USA	183	45 (Male; 42) (Female; 3)	138 (Male; 133) (Female; 5)	18–85 years	(DM RIHR: 2, LIHR: 14) (HTN RIHR: 15, LIHR: 43) (Hyperlipidemia RIHR: 8, LIHR: 29) (COPD/Asthma RIHR: 4, LIHR:8) (Obstructive sleep apnea RIHR:3, LIHR:1) (Chronic kidney disease, RIHR:4, LIHR: 2) (Atrial fibrillation, RIHR: 1, LIHR: 4) (Gastroesophageal reflux disease RIHR:5, LIHR: 15) (Congestive heart failure RIHR: 0, LIHR: 2) (Cerebrovascular accident RIHR: 1, LIHR:1)	- Operative time: longer for robotic p < 0.01 - Severe complications (Clavien-Dindo grade III–IV): higher for robotic p = 0.01 - Reoperation rate:higher for robotic, p = 0.01 - Total hospital cost: higher for robotic, p < 0.01
Gerdes et al. [2]	2022	Prospective comparative cohort	Switzerland	58	29 (Male; 27) (Female; 2)	29 (Male; 24) (Female; 5)	21–82 years	N/A	- Postoperative pain (VRS): 100% of r-TAPP vs 89.6% of l-TAPP reported low or no pain, p = 0.11
Prabhu et al. [4]	2020	Randomized controlled trial primary report (RIVAL)	USA	102	48 (Male; 44) (Female; 4)	54 (Male; 48) (Female; 6)	NR	Comorbidities were matched between groups (age, sex, BMI, comorbidities were balanced), but no specific number of comorbidities reported	Operative Time: p < 0.001, Hospital Cost: p < 0.001, Surgeon Frustration (NASA TLX): p = 0.004, total cost p < 0.001, operating room cost p < 0.001, BMI diffriences between groups p = 0.01
Holleran et al. [1]	2022	Retrospective database cohort	USA	24,098	6063 (Male; N/A) (Female; N/A)	18,035 (Male; N/A) (Female; N/A)	35–89 years	(CHF RIHR: 592, LIHR: 883) (History of angina RIHR: 28 LIHR: 103) ( prior MI RIHR: 1178, LIHR: 8855) ( HTN RIHR: 3628, LIHR: 8929) (CVA RIHR: 324, LIHR: 745) (PAD RIHR: 141, LIHR: 263) ( COPD RIHR: 589, LIHR: 1616) ( sleep apnea RIHR: 1683, LIHR: 2253) (DM RIHR: 1137, LIHR: 1727) ( dialysis within 2 weeks pre-op RIHR: 19, LIHR: 41) ( acute renal failure RIHR: 4, LIHR: 6) (bleeding disorder RIHR: 231, LIHR: 398) ( Disseminated cancer RIHR: 48, LIHR: 31)	Operative time: p < 0.001, Length of stay: p < 0.001, Complication rates: p < 0.001, Unplanned return to OR: p < 0.001, Pulmonary complications: p < 0.001, Renal complications: p < 0.001, Infectious complications: p < 0.001, Systemic complications (VTE, sepsis): p < 0.001

Values are report-level. Prabhu 2020 and Miller 2023 are the primary and long-term reports from the same 102-participant RIVAL randomized cohort; these participants were counted once in the unique total of 78,940. NR, not reported. Group counts were checked against report totals; no unreported P values were imputed

Table 2.

Operative approach and follow-up details

Study	Robotic intervention	Laparoscopic comparator	Operative time and/or follow-up
Zayan et al. [3]	Robotic inguinal hernia repair (approach not further specified)	Laparoscopic inguinal hernia repair	RIHR: 2 h (1:45–2:15), LIHR: 58 min ( 0:54–1:03)
Ayuso et al. [6] (2023)	LIHR and RIHR using TAPP approach with synthetic mesh (mainly 3D Max); peritoneal closure with tacks (LIHR) vs suture (RIHR)	Laparoscopic transabdominal preperitoneal repair (L-TAPP)	LIHR: 100.6 ± 38.3, RIHR: 109.1 ± 50 min
Miller et al. [21] (2023)	Robotic TAPP vs Laparoscopic TAPP inguinal hernia repair	Laparoscopic inguinal hernia repair	1-year and 2-year follow-up
Baginski et al. [22]	L-TEP, R-TAPP	Laparoscopic inguinal hernia repair	N/A
Wadhawan et al. [23]	RAS: TAPP (100%) LS: eTEP (54.24%), TEP (25.42%), TAPP (20.34%)	laparoscopic inguinal hernia repair	Mean operative time RAS: 89.83 min; LS: 67.71 min
Tatarian et al. [24] (2021)	R-IHR, L-IHR, O-IHR	laparoscopic inguinal hernia repair	N/A
Kudsi et al. [25] (2023)	LIHR: L-TEP; RIHR: R-TAPP	laparoscopic inguinal hernia repair	RIHR: 78.8 ± 29.1 min LIHR: 72.7 ± 28.5 min
Okamoto et al. [26]	R-TAPP (robotic transabdominal preperitoneal repair using da Vinci system) L-TAPP (laparoscopic transabdominal preperitoneal repair)	Laparoscopic (L-TAPP)	RIHR: 99.5 min, LIHR: 89 min
Hinojosa-Ramirez et al. [7] (2024)	Robotic TAPP vs Laparoscopic TAPP	Laparoscopic (L-TAPP)	RIHR: 86 ± 33.1 min LIHR: 40 ± 14.5 min (p < 0.001)
Muysoms et al. [27]	Conventional laparoscopic inguinal hernia repair, Robot-assisted laparoscopic inguinal hernia repair	Conventional laparoscopic inguinal hernia repair	RIHR: 56.2 min, LIHR: 57.6 min
Pai et al. [28]	RIHR, LIHR	laparoscopic inguinal hernia repair	N/A
Choi et al. [29]	Robotic TAPP, Laparoscopic TAPP	laparoscopic inguinal hernia repair	RIHR: 30.2 ± 11.9 min LIHR: 31.5 ± 10.3 min
Chao et al. [30] (2024)	RIHR (TAPP,TEP), LIHR (TAPP,TEP)	laparoscopic inguinal hernia repair	N/A
Brosnihan et al. [31]	RIHR, LIHR	laparoscopic inguinal hernia repair	N/A
Peltrini et al. [32] (2023)	Robotic transabdominal preperitoneal (RTAPP) inguinal hernia repair, Laparoscopic transabdominal preperitoneal (LTAPP) inguinal hernia repair	Laparoscopic transabdominal preperitoneal (LTAPP) inguinal hernia repair	N/A
Khoraki et al. [5] (2020)	RIHR, LIHR	laparoscopic inguinal hernia repair	N/A
Gerdes et al. [2]	Robotic TAPP (r-TAPP), Laparoscopic TAPP (l-TAPP) inguinal hernia repair	Laparoscopic TAPP (l-TAPP) inguinal hernia repair	N/A
Prabhu et al. [4]	Robotic TAPP, Laparoscopic TAPP inguinal hernia repair	Laparoscopic TAPP inguinal hernia repair	N/A
Holleran et al. [1] (2022)	Robotic, laparoscopic, and open inguinal hernia repair	laparoscopic inguinal hernia repair	N/A

LIHR, laparoscopic inguinal hernia repair; RIHR, robotic inguinal hernia repair; TAPP, transabdominal preperitoneal; TEP, totally extraperitoneal; L-TEP, laparoscopic TEP; R-TAPP, robotic TAPP; NR, not reported

Risk of bias assessment

The randomized evidence comprised one RIVAL cohort with a primary report (Prabhu 2020) and a 1- and 2-year follow-up report (Miller 2023). RoB 2 was therefore interpreted at the report/outcome level rather than as two independent trials. The primary report was judged at low risk of bias overall; the follow-up report had some concerns because handling of incomplete long-term outcome data was insufficiently described (Fig. 2).

MINORS was applied to the 17 observational cohorts (Table 3). Scores ranged from 13 to 20 of 24, indicating moderate to relatively high methodological quality. Recurrent weaknesses were retrospective data collection, lack of blinded outcome assessment, and absence of prospective sample-size calculation. Because MINORS scores do not eliminate residual confounding, pooled observational findings were interpreted cautiously.

Primary outcomes

Hernia recurrence

Nine reports comprising 3,200 participants contributed recurrence data (13 events among 1,381 robotic repairs and 51 among 1,819 laparoscopic repairs). Robotic repair was associated with a lower observed recurrence risk (RR 0.32, 95% CI 0.17–0.59; P = 0.0003), corresponding to a 68% relative reduction. Heterogeneity was absent (I² = 0%, τ² = 0), and the prediction interval remained below the null (0.16–0.66) (Fig. 3).

Fig. 3.

Random-effects forest plot of hernia recurrence after robotic versus laparoscopic inguinal hernia repair. Risk ratios below 1 favor robotic repair; the diamond shows the pooled estimate, and I2 and τ2 quantify heterogeneity

Leave-one-out estimates ranged from RR 0.28 to 0.38 and remained below 1.0 (Figure S1). Excluding the RIVAL follow-up report, which was the only randomized recurrence contribution, produced a similar estimate (RR 0.30, 95% CI 0.14–0.64). In subgroup analysis, moderate-quality observational studies favored robotic repair (RR 0.26, 95% CI 0.13–0.50; I² = 0%), whereas the high-quality observational subgroup (Ayuso 2023 only: RR 2.00, 95% CI 0.18–21.81) and randomized subgroup (Miller 2023 only: RR 1.26, 95% CI 0.08–19.49) each contained one imprecise report. The subgroup-difference test was not significant (P = 0.16) (Figure S2).

A clinically interpretable laterality or primary-versus-recurrent recurrence subgroup analysis was not possible. Only two entirely unilateral cohorts supplied risk-ratio-informative recurrence data; one bilateral study had no events in either arm and the other had internally inconsistent visit-specific recurrence reporting. The planned interaction tests were therefore not performed, and this evidence gap is reported rather than replaced by unstable subgroup estimates.

Operative time

Eleven studies including 26,484 participants (7,095 robotic and 19,389 laparoscopic repairs) contributed to the primary operative-time analysis. Robotic repair required longer operating time (MD + 30.04 min, 95% CI 9.87–50.20; P = 0.0078). Heterogeneity was extreme (I² = 99.6%, τ² = 888.27), reflecting differences in time definitions, operative technique, laterality, learning-curve phase, surgeon experience, and database setting (Fig. 4).

Fig. 4.

Random-effects forest plot of operative time after robotic versus laparoscopic inguinal hernia repair. Positive mean differences indicate longer robotic operating time; I2 and τ2 quantify heterogeneity

Leave-one-out estimates remained statistically significant and ranged from + 22.09 to + 33.26 min. Influence diagnostics identified Holleran 2022 as influential (Figure S3); after its exclusion, the pooled difference was + 22.09 min (95% CI 8.87–35.30; P = 0.0043), although heterogeneity remained substantial (I² = 92.3%, τ² = 300.63) (Figure S4). Excluding the single randomized cohort yielded MD + 30.11 min (95% CI 7.49–52.72), indicating that the direction was not dependent on randomized evidence. Primary and sensitivity results are shown side by side in Supplementary Table S2.

After exclusion of the influential Holleran 2022 cohort, subgroup estimates were + 13.83 min for high-MINORS observational studies (95% CI -52.87 to 80.53; I² = 50.4%), + 23.77 min for moderate-MINORS observational studies (95% CI 3.64–43.89; I² = 94.2%), and + 29.39 min for the randomized report (95% CI 19.10–39.69). There was no evidence of effect modification by quality/design (P for subgroup difference = 0.108) (Figure S5). Hernia-classification subgroup analyses were not performed because study-level reports did not provide compatible stratified means and variances.

The operative-time funnel plot showed no visually apparent asymmetry, and Egger’s test was not significant (P = 0.506) (Fig. 5); however, the small number of studies limited power to detect publication bias.

Fig. 5.

Funnel plot for the operative-time meta-analysis. Egger’s regression test did not detect asymmetry (P = 0.506), although power was limited by the small number of studies

Postoperative pain

Twelve reports assessed postoperative pain, with substantial variation in instruments, timing, analgesic protocols, and outcome definitions. Continuous VAS and NRS scores were pooled only within clinically aligned time windows; CCS, verbal categories, neuropathic pain, analgesic consumption, and composite pain/medication outcomes were synthesized separately. Negative SMDs favored robotic repair.

Two studies assessed immediate postoperative or PACU pain (1,253 participants) and showed no difference (SMD 0.02, 95% CI -0.66 to 0.70; P = 0.818; I² = 0%) (Fig. 6A). Three studies (343 participants) contributed postoperative-day-1 data; the pooled estimate favored robotic repair but was imprecise and not statistically significant (SMD -0.47, 95% CI -1.56 to 0.63; P = 0.208; I² = 81.3%) (Fig. 6B). A VAS-only sensitivity analysis yielded SMD -0.72 (95% CI -1.81 to 0.36; P = 0.075; I² = 0%) (Fig. 6C). Okamoto 2023 was the only NRS-only postoperative-day-1 study and reported a median score of 2 (IQR 1–3) in both groups.

Fig. 6.

Random-effects analyses of postoperative pain after robotic versus laparoscopic inguinal hernia repair: (A) immediate postoperative/PACU pain; (B) postoperative-day-1 pain measured using VAS or NRS; and (C) VAS-only postoperative-day-1 sensitivity analysis. Negative standardized mean differences favor robotic repair

Beyond postoperative day 1, pooling was not appropriate because repeated measurements would have double-counted participants and several studies measured non-equivalent constructs. Wadhawan 2025 reported lower VAS scores and analgesic use through postoperative day 7 and a lower 1-month CCS score, whereas the RIVAL reports found no significant differences at 1 week, 30 days, or 1–2 years in pain, neuropathic pain, or SF-36 bodily pain. Other studies generally reported no sustained differences.

Analgesic outcomes were also inconsistent. Chao 2024 reported lower rescue opioid use after robotic repair, while Kudsi 2023 reported similar PACU pain and opioid requirements. Baginski 2024 reported less postoperative pain using a composite definition that included patient report or a new/continued analgesic prescription rather than a validated scale. Overall, pooled evidence did not establish a significant immediate or postoperative-day-1 pain advantage, although selected individual studies suggested a possible short-term benefit that requires confirmation using standardized instruments and time points.

Secondary outcomes

Length of hospital stay

Nine studies including 26,111 participants reported length of stay. In the primary random-effects analysis, robotic repair was associated with a nonsignificant numerical increase of 0.33 days (approximately 7.9 h; 95% CI -0.03 to 0.70; P = 0.0697; I² = 96.2%, τ² = 0.213) (Fig. 7). Holleran 2022 was influential; after its exclusion, the pooled estimate was + 0.18 days (approximately 4.3 h; 95% CI 0.06–0.29; P = 0.0096; I² = 43.2%) (Figure S6). Although statistically significant in the influence analysis, a difference of about 4 h is likely to have limited patient-level clinical importance and may primarily reflect discharge pathways and institutional workflow.

Fig. 7.

Random-effects forest plot of length of hospital stay after robotic versus laparoscopic inguinal hernia repair. Positive mean differences indicate longer stay after robotic repair

Overall complications

Six studies included 39,156 participants and 795 total complications. The primary estimate suggested a possible increase after robotic repair but was borderline and highly heterogeneous (RR 1.68, 95% CI 1.00–2.84; P = 0.052; I² = 93.4%; prediction interval 0.34–8.39) (Fig. 8A). After exclusion of influential Holleran 2022, the estimate attenuated to RR 1.18 (95% CI 0.90–1.54; I² = 0%) (Figure S7). The available evidence therefore did not establish a clear difference in overall complications.

Fig. 8.

Random-effects forest plots of complications after robotic versus laparoscopic inguinal hernia repair: (A) any postoperative complication; (B) seroma; (C) hematoma; (D) surgical-site infection; and (E) postoperative urinary retention. Risk ratios below 1 favor robotic repair

Seroma

Three studies with 24,383 participants reported seroma. The primary analysis favored laparoscopy (RR 4.41, 95% CI 2.92–6.66; P < 0.0001; I² = 21.1%) (Fig. 8B), but the finding was dominated by Holleran 2022. After that study was omitted, the estimate was extremely imprecise and nonsignificant (RR 1.85, 95% CI 0.08–42.23) (Figure S8). The apparent seroma signal was therefore study-dependent and should be interpreted cautiously.

Hematoma

Five studies with 1618 patients reported 22 hematomas following robotic repair and 24 hematomas following laparoscopic repair. There was no significant difference between approaches (RR 0.77, 95% CI 0.43 to 1.40; P = 0.392; I² = 0%; prediction interval 0.33 to 1.79) (Fig. 8c). Leave-one-out estimates were also nonsignificant, suggesting that the null result was not due to a single study (Figure S9 in the supplementary materials).

Surgical site infection

Five studies with 1,840 patients reported five surgical-site infections following robotic repair and seven following laparoscopic repair. The pooled estimate showed no significant difference (RR 0.80, 95% CI 0.26 to 2.42; P = 0.691; I² = 0%; prediction interval 0.17 to 3.85) (Fig. 8d). All leave-one-out analyses were still nonsignificant (Figure S10 in the supplementary materials).

Postoperative urinary retention

Four studies with 667 patients reported 14 cases of urinary retention following robotic repair and 24 cases after laparoscopic repair. There was no statistically significant difference (RR 0.74, 95% CI 0.39 to 1.41; P = 0.359; I² = 0%; prediction interval 0.26 to 2.12) (Fig. 8e), and the results did not change with leave-one-out analyses (Figure S11 in the supplementary materials).

Patient-reported outcomes and functional recovery

Patient-reported outcomes were too heterogeneous to pool. In the RIVAL randomized cohort, SF-36 domains, physical and mental component scores, activity, and cosmesis were similar at early and 1–2-year follow-up. Wadhawan 2025 reported a lower 1-month CCS score after robotic repair, whereas Zayan 2019 reported similar CCS scores postoperatively and at 1 year. Hinojosa-Ramirez 2024 found no statistically significant difference in return to work or return to physical activity. Overall, patient-reported evidence did not demonstrate a consistent long-term advantage for either approach.

Economic outcomes

Eight studies reported economic outcomes. Meta-analysis was not appropriate because currencies, price years, healthcare systems, cost perspectives, reimbursement arrangements, and included components differed substantially. Findings were therefore grouped by country/setting and by cost component (Table 4).

Table 4.

Structured narrative synthesis of economic outcomes by country, healthcare setting, and cost component

Study	Country/setting	Economic metric/perspective	Robotic	Laparoscopic	Absolute/relative difference	Statistical result	Main driver and interpretive caveat
[4]	USA; 6 academic/affiliated centers	Provider costs: total, OR, and disposable/reusable supplies	Total median $3,258 (IQR 2,568–4,118); OR $1,401 (1,234–1,933); supplies $1,784 ± 1,149	Total median $1,421 (1,196–1,930); OR $879 (714–1,052); supplies $623 ± 434	+ $1,837 total (2.29 ×); + $522 OR (1.59 ×); + $1,161 supplies (2.86 ×)	All P < 0.001	Higher OR resource use and instruments/supplies; randomized evidence, but short 30-day horizon
[5]	USA; tertiary academic MIS unit	2017 hospital costs plus allocated robotic capital/maintenance	Total $9,994 ± 2,498; supplies $1,588 ± 1,298; device allocation $3,106/case	Total $5,995 ± 3,049; supplies $1,380 ± 565	+ $3,999 total (1.67 ×); supplies + $208 (1.15 ×)	Total P < 0.01; supplies P = 0.31	Increment attributed to operative time and robotic capital/maintenance ($2,130 + $976); readmission/reoperation costs excluded
[6]	USA; urban teaching hospital	Institutional charges, not actual costs	$52,068 ± 13,774	$41,549 ± 8,848	+ $10,519 (1.25 ×)	P < 0.01	Billing charges cannot be interpreted as resource cost or reimbursement; no component breakdown
[31]	USA; National Inpatient Sample, obese inpatients	Survey-weighted total hospital cost; adjusted regression	Weighted mean $78,000 ± 2,300 SE	Weighted mean $58,100 ± 2,600 SE	Crude + $19,900; adjusted + $5,620 (95% CI $4,800–6,440)	P < 0.001 adjusted	Administrative inpatient population; survey-weighted estimates and residual confounding limit comparison with ambulatory repair
[3]	USA; single-center mixed inguinal/ventral cohort	Charges, direct, indirect, and total cost	Charges $45,459; direct $5,017; indirect $3,373; total $7,832	Charges $30,247; direct $4,638; indirect $3,622; total $8,605	Charges + $15,212; direct + $379; indirect -$249; total -$773	P < 0.001; 0.017; 0.058; 0.050, respectively	Mixed hernia population; higher charges did not translate into higher total cost, illustrating non-equivalence of cost metrics
[30]	Belgium; single-payer hospital	Mean index-operation cost and index + readmission cost; capital excluded	Index €2,612; with readmission €2,653; materials €1,039; consumables €1,317	Index €1,963; with readmission €2,078; materials €408; consumables €648	Index + €649 (33.1%); with readmission + €575; materials + €631; consumables + €669	Index P < 0.001	Materials/consumables drove the increment; stay and physician fees were similar/slightly lower; capital exclusion likely understates robotic cost
[7]	Spain; public National Health System	Direct procedure, stay, and total hospitalization costs	Procedure €2,810.15 ± 218.10; total €3,792.49 ± 326.06	Procedure €725.81 ± 158.91; total €1,497.20 ± 375.68	Procedure + €2,084.34 (3.87 ×); total + €2,295.29 (2.53 ×)	Both P < 0.001	Robotic procedural inputs and OR occupancy were dominant; public-system unit costing
[32]	South Korea; single-center	Surgery fee only; hospitalization and ancillary expenses excluded	$3,814.80 ± 172.90	$209.60 ± 27.50	+ $3,605.20 (18.2 ×)	P < 0.001	Highly setting-specific fee differential; not a comprehensive hospital-cost estimate

CI, confidence interval; IQR, interquartile range; MIS, minimally invasive surgery; OR, operating room; RCT, randomized controlled trial; SE, standard error. Values are means ± SD unless otherwise stated. Currency values were not converted across price years or jurisdictions

In the United States, the RIVAL trial reported higher median total, operating-room, and supply costs for robotic repair. Khoraki 2020 reported a $3,999 higher total hospital cost, attributed mainly to operating time and allocated robotic capital/maintenance. Ayuso 2023 reported higher institutional charges, which are not equivalent to actual resource costs. In the National Inpatient Sample, Pai 2025 reported higher survey-weighted inpatient hospital costs for robotic repair among patients with obesity, but this administrative inpatient cohort is not directly representative of routine ambulatory repair. In Europe, Muysoms 2021 and Hinojosa-Ramirez 2024 similarly found higher robotic procedure or hospitalization costs, driven primarily by consumables, materials, and operating-room resources.

Choi 2023 reported a large difference in surgery fees in South Korea, but hospitalization and ancillary costs were excluded. Zayan 2019 illustrated the importance of distinguishing charges from costs: robotic repair generated higher charges and direct cost, but total cost was numerically lower in a mixed inguinal/ventral cohort. Across settings, robotic platforms, instruments/consumables, and operating-room use were the most consistent incremental cost drivers. No study used a common cost-effectiveness framework linking incremental cost to quality-adjusted life-years, recurrence prevented, or another shared effectiveness measure.

Certainty of evidence

Certainty was low for recurrence because the pooled estimate was dominated by non-randomized studies susceptible to residual confounding despite negligible statistical heterogeneity. Certainty was very low for operative time, postoperative pain, length of stay, overall complications, seroma, hematoma, surgical-site infection, and urinary retention because of retrospective designs, inconsistency, outcome-measurement variation, sparse events, and/or imprecision (Table 5).

Discussion

This review synthesized 19 reports from 18 unique cohorts comprising 78,940 adults. Robotic repair was associated with a lower observed recurrence risk in the pooled analysis, but the certainty of this finding was low, most evidence was observational, and the single randomized recurrence estimate was imprecise and did not show a clear difference. The result should therefore be interpreted as an association rather than evidence that the robotic technique itself prevents recurrence.

Robotic repair required approximately 30 additional operative minutes in the primary analysis and about 22 min after exclusion of an influential national database study. This difference is statistically robust in direction but clinically context-dependent: it may have limited direct importance to an individual patient, yet it can affect operating-room throughput, staffing, anesthesia exposure, and resource allocation at scale. Similarly, the primary length-of-stay estimate was not significant; the influence analysis suggested about 4.3 additional hours after robotic repair, a modest difference that may reflect discharge policy more than recovery.

Pooled immediate and postoperative-day-1 pain estimates did not establish a significant benefit, and longer-term pain and quality-of-life outcomes were generally similar. Overall complications were also uncertain, while the apparent seroma increase was driven by one large cohort. Economic findings were more consistent: robotic repair incurred higher procedural or hospital expenditure in most settings, mainly through platform, instrument, consumable, and operating-room costs. These findings broadly agree with earlier reviews that reported longer operating time and higher cost without consistent perioperative superiority [36, 37].

Several limitations restrict interpretation. Seventeen of 18 unique cohorts were observational, creating confounding by indication and selection bias related to patient complexity, hernia laterality or recurrence, and surgeon choice. Surgeon learning curve, annual robotic volume, institutional pathways, and operative-time definitions varied substantially. Follow-up duration and recurrence ascertainment were inconsistent, aggregate data prevented patient-level adjustment, and subgroup analyses by primary/recurrent and unilateral/bilateral hernia were usually infeasible. Twenty-five reports could not be retrieved and therefore could not be assessed; because they supplied no outcome data, a quantitative exclusion sensitivity analysis was not possible. Restriction to English-language publications and screening only the first 250 Google Scholar results may have introduced language or retrieval bias. Multiple reports from the RIVAL cohort required careful linkage to avoid double-counting. Finally, economic studies used heterogeneous price years, currencies, perspectives, and definitions and did not permit formal cost-effectiveness analysis.

Future multicenter randomized trials should use a prespecified core outcome set, standardized definitions of operative time and complications, blinded or registry-based recurrence ascertainment, and long-term patient-reported follow-up. Trials should stratify or report results by primary versus recurrent and unilateral versus bilateral hernia, surgeon experience, and institutional volume. Multi-country economic evaluations should use uniform costing perspectives, price-year adjustment, and incremental cost-effectiveness outcomes.

Taken together, the current evidence does not support a universal claim of robotic superiority. The observed recurrence association is hypothesis-generating, while the longer operating time and higher cost are more consistent findings. Selection of approach should be individualized to patient and hernia characteristics, surgeon expertise, local capacity, and the value placed on uncertain potential benefits relative to resource use.

Conclusion

Robotic inguinal hernia repair was associated with lower observed recurrence in pooled, predominantly observational evidence, but certainty was low and causal superiority was not established. Robotic repair generally required 22–30 additional operative minutes and cost more, while pooled pain, length-of-stay, and complication outcomes were uncertain or clinically small. Decisions should be individualized, and adequately powered randomized studies with standardized long-term and economic outcomes are needed.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contribution

All authors contributed substantially to the conception, design, literature review, data collection, analysis, interpretation of results, and manuscript preparation. All authors reviewed, revised, and approved the final manuscript and agreed to its submission for publication.

Data availability

The data supporting the findings of this study were extracted from previously published studies included in this systematic review and meta-analysis. All data analyzed during this study are included in the published article and its supplementary materials. Additional information is available from the corresponding author upon reasonable request.

Declarations

Conflicts of interest

The authors declare no competing interests.