Splanchnic denervation for refractory upper abdominal cancer pain: a systematic review and meta-analysis of randomized controlled trials

Abstract

Background

Splanchnic denervation (SND) is an alternative interventional strategy for refractory upper abdominal cancer pain. This systematic review and meta-analysis evaluated the longitudinal analgesic efficacy, opioid-sparing effects, and safety of SND compared with those of conventional control modalities.

Methods

Randomized controlled trials (RCTs) comparing SND with controls were identified by searching PubMed, Embase, CENTRAL, Scopus, and Web of Science through December 2025. The primary outcome was change from baseline (CfB) in pain intensity at 12 weeks. Secondary outcomes included CfB in morphine equivalent daily dose, quality of life (QoL), and procedure-related adverse events.

Results

Six RCTs (n = 337) were included. At 12 weeks, SND demonstrated analgesic efficacy (standardized mean difference [SMD] = −0.36; 95% confidence interval [CI], −1.20 to 0.48; P = 0.323) and opioid-sparing effects (SMD = −1.01; 95% CI, −2.16 to 0.14; P = 0.071) comparable to control groups. At 4 weeks, QoL outcomes were comparable between groups (SMD = 0.70; 95% CI, −0.28 to 1.67; P = 0.127). SND was associated with a significantly lower risk of hypotension (risk ratio [RR] = 0.45; 95% CI, 0.25 to 0.80; P = 0.007) and diarrhea (RR = 0.24; 95% CI, 0.10 to 0.54; P < 0.001) than celiac plexus neurolysis.

Conclusions

SND appears to be a clinically effective, technically feasible, and safe interventional strategy for refractory upper abdominal cancer pain. SND achieved efficacy comparable to conventional controls across 24 weeks. SND may serve as a clinically meaningful option in multimodal pain management for patients with advanced abdominal malignancies.

INTRODUCTION

Cancer pain remains a critical clinical challenge, significantly impairing the quality of life (QoL) [1]. While pharmacological interventions, particularly opioids, form the cornerstone of the World Health Organization analgesic ladder, many patients face suboptimal pain relief or debilitating side effects, necessitating dose escalation and complex polypharmacy [2]. In cases of refractory cancer pain, interventional pain procedures (IPPs) have progressed from last-resort measures to integral components of comprehensive, multimodal pain management [3,4].

Celiac plexus neurolysis (CPN) has long been an established IPP for visceral pain originating from abdominal malignancies, demonstrating efficacy in reducing both pain intensity and opioid use [5–7]. However, the anatomical proximity of the celiac plexus to abdominal tumors means that the procedure is frequently performed in regions compromised by tumor infiltration and distorted anatomy [8,9], which can limit effectiveness of CPN or pose technical challenges [10,11].

Splanchnic denervation (SND), a therapeutic approach targeting the thoracic splanchnic nerves, offers an alternative strategy. These nerves are the principal neural conduits for transmitting nociceptive signals from the upper abdominal viscera [12]. Preganglionic sympathetic fibers are positioned superior and posterior to the celiac plexus, with a high incidence of morphological convergence at a consistent anatomical site on the anterior T11–L1 vertebral body [13]. This focal consolidation of the nerve bundle facilitates the application of SND, which can be performed using traditional chemical neurolysis (SNN) or, notably, radiofrequency ablation (SNRF). Both modalities aim to denervate the visceral nociceptive pathways—SNN via the injection of neurolytic agents and SNRF via thermal radiofrequency (RF) ablation—by targeting this concentrated neural cluster.

Clinical studies increasingly support the efficacy of both SNN and SNRF for managing refractory upper abdominal malignant pain [10,11,14–18]. However, most studies have evaluated SNN and SNRF separately, and no comprehensive systematic analysis or meta-analysis has compared the unified SND approach against other established pain control modalities, such as CPN or conventional medical management. Therefore, this systematic review and meta-analysis was conducted to rigorously compare the analgesic efficacy and opioid-sparing effects of SND against other pain control strategies for refractory upper abdominal malignant pain. The authors hypothesized that SND would be as effective as conventional control modalities in reducing pain intensity and opioid requirements, while offering a superior safety profile with fewer procedure-related complications.

MATERIALS AND METHODS

A systematic review and meta-analysis was performed according to the Preferred Reporting Items for Systematic Reviews and Meta Analyses (PRISMA) statement [19] for conducting and reporting systematic reviews of randomized controlled trials (RCTs). The pre-defined protocol of this study was registered in the International Prospective Register of Systematic Reviews (PROSPERO: CRD420251248584). This study is a systematic review and meta-analysis of published data; therefore, informed consent was not required.

1. Search strategy

A comprehensive literature search was performed across electronic databases, including PubMed, Embase, CENTRAL, Scopus, and Web of Science databases, from their respective dates of inception to December 1, 2025. The detailed search strategies for each database are presented in Supplementary Table 1. To ensure a comprehensive search, the authors manually screened the bibliographies of relevant studies, including systematic reviews, and identified clinical trials. Additionally, a targeted web-based search was conducted to identify RCTs that were not indexed in the primary electronic databases.

2. Study selection

Two independent reviewers screened titles and abstracts of the retrieved studies, followed by full-text review of potentially eligible articles. Any discrepancies during the selection process were resolved through discussion or, if necessary, through consultation with a third reviewer. The inclusion criteria were as follows: (1) RCTs; (2) studies involving patients with chronic cancer pain from upper abdominal malignancies; and (3) studies evaluating percutaneous SND (SNN and SNRF). The exclusion criteria were as follows: (1) non-RCTs, such as observational studies, retrospective studies, case reports, and review articles; (2) studies with incomplete data or those available only as abstracts; and (3) studies involving non-percutaneous methods, such as surgical splanchnicectomy.

3. Data extraction

Two authors independently extracted data from the included RCTs using a standardized form. The following information was collected from each study: (1) study characteristics, including first author, publication year, sample size, and demographics; (2) baseline clinical status, such as initial pain intensity measured by visual analog scale (VAS) or numerical rating scale (NRS) scores and morphine equivalent daily dose (MEDD); (3) intervention protocols for SND, specifying the modality (SNN or SNRF) and detailed procedures (e.g., anatomical target level and volume/concentration of agents or ablation parameters such as temperature and duration); (4) control group protocols, specifying the control modality (e.g., active comparators, placebo or sham interventions, or conventional medical management alone) and detailed procedures; and (5) clinical outcomes, including longitudinal data for pain intensity, MEDD, QoL scores at pre-specified time points, and incidence of acute procedure-related adverse events.

4. Outcome measures

The primary outcome was analgesic efficacy at 12 weeks, assessed by change from baseline (CfB) in pain intensity.

To comprehensively evaluate the multidimensional clinical impact, the following secondary outcomes were analyzed: (1) analgesic efficacy at 4 weeks, 8 weeks, and 24 weeks; (2) opioid sparing effects at 4 weeks, 8 weeks, 12 weeks, and 24 weeks, measured as the CfB in MEDD; (3) absolute QoL scores at 4 weeks, regardless of the specific assessment instrument used; and (4) incidence of procedure-related adverse events, such as transient hypotension, diarrhea, and procedural backache. For studies reporting multiple QoL domains, a composite mean and standard deviation (SD) were calculated to represent global QoL. A correlation coefficient of 0.5 was assumed to pool the variances of these disparate domains [20]. Furthermore, score directions were synchronized such that higher values consistently indicated improved QoL (e.g., multiplying values by −1 where necessary).

5. Quality assessment

Two independent authors assessed the risk of bias using the Revised Cochrane Risk of Bias tool for randomized trials (RoB-2) [21]. This tool evaluates five domains: (1) bias from the randomization process; (2) bias due to deviations from intended interventions; (3) bias due to missing outcome data; (4) bias in the measurement of the outcome; and (5) bias in the selection of the reported result. The overall risk of bias was categorized as “low risk,” “some concerns,” or “high risk.” Any disagreements were resolved through consensus or through consultation with a third senior author. Furthermore, the certainty of evidence for each pooled outcome was determined using the Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) system, rating evidence as “high,” “moderate,” “low,” or “very low” [22].

6. Statistical analysis

Statistical analyses were performed using R software (version 2025.05.1; Posit Software, PBC) with the “meta” and “metafor” packages. For analgesic efficacy, opioid-sparing effects, and QoL scores, the standardized mean differences (SMDs) with 95% confidence intervals (CIs) were calculated using Hedges’ g to account for small sample sizes and potential differences in measurement scales. For procedure-related adverse events, risk ratios (RRs) with 95% CIs were calculated to evaluate relative risks between groups.

For data presented only in graphical format, all numerical parameters, including time-dependent sample sizes (N), mean values, and measures of dispersion, were extracted using WebPlotDigitizer (version 5). To handle missing or non-reported values, the following estimation methods were employed: (1) medians and interquartile ranges were converted to means and SDs using Wan’s formula [23], or the median was utilized directly as a proxy for the mean in cases where additional dispersion metrics were unavailable [20]; (2) missing SDs for absolute mean values at specific time points were imputed using the mean of SDs from other included trials within the same intervention group [24]. Specifically, for QoL outcomes, SDs were derived from studies employing identical measurement instruments, sourced either from other included trials or from clinically comparable published studies; (3) when SDs for CfB values were not reported, they were calculated using a correlation coefficient (r) of 0.5 [25,26].

Inter-study heterogeneity was assessed using Cochran’s Q test and the I² statistic. A fixed-effect model was employed when heterogeneity was considered low (I² < 50%); otherwise, a random-effects model was used (I² ≥ 50%). Given the anticipated limited number of eligible RCTs, the Hartung–Knapp–Sidik–Jonkman adjustment was applied to all random-effects meta-analyses, as this method provides more conservative and accurate CIs when the number of studies is small, compared to the conventional DerSimonian–Laird approach [27]. To enhance the transparency of the authors’ data estimation process and ensure the clinical reliability of the pooled results, sensitivity analyses were integrated into the overall assessment. First, the correlation coefficient (r) was varied to 0.25 and 0.75 to evaluate the impact of data uncertainty on the pooled findings. Second, a sensitivity analysis via outlier exclusion was performed to evaluate the potential disproportionate impact of influential trials on inter-study heterogeneity and to verify the overall stability of the synthesized outcomes.

Furthermore, a systematic, multi-tiered analytical approach was utilized specifically for the 12-week primary outcomes—analgesic efficacy and opioid-sparing effects. First, meta-regression analyses were performed to identify potential sources of inter-study heterogeneity, with clinical moderators including baseline mean values, intervention modality, control type, age, and sex. Second, sensitivity analyses were conducted by specifically selecting trials that utilized CPN as the control group. Concurrently, a subgroup analysis was performed within this refined subset to compare potential outcome disparities between SNN and SNRF modalities. Finally, potential publication bias was assessed using Egger’s regression test. Statistical significance was set at P < 0.05.

RESULTS

1. Study selection

A total of 893 potentially relevant records were identified through initial electronic database searches. Additionally, five records were identified through other methods, including websites (n = 2) and manual citation searching (n = 3). After removing 320 duplicate records, 573 reports from databases were screened by title and abstract, leading to the exclusion of 569. While all four reports from the databases were successfully retrieved, two reports identified through other methods could not be retrieved. Consequently, seven articles (four from databases and three from other methods) underwent a rigorous full-text assessment for eligibility. Of these, one article was excluded because it compared two different SND techniques (SNN vs. SNRF) [16]. Finally, six RCTs involving 337 patients were included (Fig. 1) [28–33]. The detailed characteristics of the included studies are summarized in Table 1. All six studies were two-arm RCTs: one study compared SND with a sham block [30], whereas the remaining five directly compared SND with CPN [28,29,31–33]. Four RCTs utilized SNN [28–31], and the remaining two employed SNRF [32,33].

Fig. 1.

PRISMA flow diagram of the study selection process. SNN: splanchnic neurolysis, SNRF: splanchnic nerve radiofrequency ablation, PRISMA: Preferred Reporting Items for Systematic Reviews and Meta Analyses.

Table 1.

Characteristics of included studies

Author (year)	Group	N	Age (yr)	Sex (M/F)	VAS/NRS	MEDD	Intervention arm			Control arm
							Modality	Details of procedure		Modality	Details of procedure
Süleyman Ozyalçin et al. [28] (2004)	2	39	59.1 ± 7.7	26/13	6.1 ± 0.7	13.4 (N/A)a	SNN	T11, bilateral; 4.5 mL ethanol 96% + 1.5 mL lidocaine 1%, each		CPN	L1, left; 30 mL of ethanol 96% + 10 mL of lidocaine 1%
Shwita et al. [29] (2015)	2	60	47.5 ± 19.9	34/26	8.3 ± 2.5	195.0 ± 52.4	SNN	T11, bilateral; 10 mL of 70% alcohol, each		CPN	L1 bilateral; 20 mL of 70% alcohol, each
Dong et al. [30] (2021)	2	96	60.7 ± 11.1	54/42	6.7 ± 1.2	126.9 ± 46.9	SNN	T12–L1, intervertebral disc; 12–18 mL of absolute alcohol		Sham	T12-L1, intervertebral disc; 12–18 mL of normal saline
Subramaniam et al. [31] (2021)	2	30	54.8 (N/A)	22/8	7.4 (N/A)a,b	72.1 (N/A)a,b	SNN	T12, bilateral; 10 mL of 99.9% alcohol, each		CPN	L1, bilateral; 10 mL of 99.9% alcohol, each
Amr Erfan et al. [32] (2024)	2	52	41.2 ± 11.9	23/29	5 [5–6]c	N/Ad	SNRF	T11, bilateral; 80°C, 2 minutes, each		CPN	L1, bilateral; 5 mL of 100% alcohol, each
Agarwal et al. [33] (2025)	2	60	N/Ae	N/Ae	7.7 ± 1.2	55.1 ± 8.9b	SNRF	T11 and T12, bilateral; 80°C, 90 seconds, each		CPN	L1, left; 5 mL of 2% lignocaine + 15 mL of 100% alcohol

Age, VAS/NRS, and MEDD are presented as baseline values (pre-intervention measurements). Values are presented as means ± SDs, medians [IQRs], or numbers.

N: number of participants, M: male, F: female, VAS: visual analog scale, NRS: numerical rating scale, MEDD: morphine equivalent daily dose, SNN: splanchnic neurolysis, CPN: celiac plexus neurolysis, N/A: not available, SNRF: splanchnic nerve radiofrequency ablation, SD: standard deviation, IQR: interquartile range.

^aThe SD for baseline mean values was not available in the original report and was imputed using the mean of SDs from the other included trials within the same intervention group. ^bNumerical data were extracted from figures using WebPlotDigitizer. ^cReported as median [IQR]; converted to mean and SD using Wan’s formula for meta-analysis. ^dOnly tramadol consumption was reported; total MEDD was not assessed in the study. ^eDemographic data for this study were not reported in the original text; for the purpose of meta-regression, these were imputed using the weighted mean values from the other included studies.

2. Analgesic efficacy

At 4 weeks post-intervention, pooled data from six studies [28–33], involving 311 patients, showed that analgesic efficacy did not significantly differ between the SND and control groups (SMD = −0.61; 95% CI, −1.75 to 0.52; P = 0.223; I² = 87.6%; random-effects model; Fig. 2A).

Fig. 2.

Forest plots comparing analgesic efficacy between the SND and control groups. Analgesic efficacy, measured as the CfB in pain intensity (VAS/NRS scores), was compared at (A) 4 weeks, (B) 8 weeks, (C) 12 weeks (primary endpoint), and (D) 24 weeks. SMDs with 95% CIs are presented using a random-effects model. SD: standard deviation, SMD: standardized mean difference, CI: confidence interval, CPN: celiac plexus neurolysis, SND: splanchnic denervation, CfB: change from baseline, VAS: visual analog scale, NRS: numerical rating scale.

At 8 weeks, pooled data from four studies [28–31], involving 169 patients, showed no significant differences between the groups (SMD = −0.48; 95% CI, −2.34 to 1.37; P = 0.467, I² = 82.8%; random-effects model; Fig. 2B).

At 12 weeks, pooled data from six studies [28–33], involving 244 patients, revealed comparable analgesic efficacy between the SND and control groups (SMD = −0.36; 95% CI, −1.20 to 0.48; P = 0.323; I² = 74.7%; random-effects model; Fig. 2C).

At 24 weeks, pooled data from four studies [29,30,32,33], involving 154 patients, demonstrated no significant differences in analgesic efficacy between the two groups (SMD = −0.52; 95% CI, −1.37 to 0.33; P = 0.147; I² = 62.5%; random-effects model; Fig. 2D).

Sensitivity analyses using varied r values (0.25 and 0.75) and the exclusion of Süleyman Ozyalçin et al. [28] as a potential outlier, consistently showed no significant differences between groups across all time points (Supplementary Figs. 1–3).

3. Opioid-sparing effects

At 4 weeks post-intervention, pooled data from five studies [28–31,33], involving 259 patients, indicated that opioid-sparing effects did not significantly differ between the SND and control groups (SMD = −0.58; 95% CI, −1.60 to 0.43; P = 0.185; I² = 87.6%; random-effects model; Fig. 3A).

Fig. 3.

Forest plots comparing opioid-sparing effects between the SND and control groups. Opioid-sparing effects, measured as the CfB in opioid consumption (MEDD), were compared at (A) 4 weeks, (B) 8 weeks, (C) 12 weeks (primary endpoint), and (D) 24 weeks. SMDs with 95% CIs are presented using a random-effects model. SD: standard deviation, SMD: standardized mean difference, CI: confidence interval, CPN: celiac plexus neurolysis, SND: splanchnic denervation, CfB: change from baseline, MEDD: morphine equivalent daily dose.

At 8 weeks, pooled data from four studies [28–31], involving 169 patients, revealed no significant difference in opioid-sparing effects between the groups (SMD = −0.68; 95% CI, −2.20 to 0.84; P = 0.250, I² = 88.0%; random-effects model; Fig. 3B).

At 12 weeks, pooled data from five studies [28–31,33], involving 192 patients, demonstrated comparable opioid-sparing effects between the SND and control groups (SMD = −1.01; 95% CI, −2.16 to 0.14; P = 0.071; I² = 82.6%; random-effects model; Fig. 3C).

At 24 weeks, pooled data from three studies [29,30,33], involving 129 patients, revealed no significant difference in opioid-sparing effects between the two groups (SMD = −0.40; 95% CI, −2.28 to 1.48; P = 0.456; I² = 76.9%; random-effects model; Fig. 3D).

Sensitivity analyses using r values of 0.25 and 0.75 yielded results that were consistent with the primary findings at all time points (Supplementary Figs. 4, 5).

4. Meta-regression, sensitivity, subgroup, and publication bias analyses at the 12-week primary endpoint

At 12 weeks, meta-regression analysis revealed that baseline pain intensity was the sole significant moderator for analgesic efficacy, accounting for 100% of the observed heterogeneity (P = 0.010). No significant moderators were identified for opioid-sparing effects.

Sensitivity analyses restricted to the CPN-controlled trials confirmed the primary findings. In these refined analyses, analgesic efficacy (SMD = −0.40; 95% CI, −1.60 to 0.80; P = 0.409; I² = 79.7%; random-effects model; Fig. 4A) [28,29,31–33] and opioid-sparing effects (SMD = −0.83; 95% CI, −2.42 to 0.76; P = 0.195; I² = 84.7%; random-effects model; Fig. 4B) [28,29,31,33] remained comparable between the SND and CPN groups.

Fig. 4.

Sensitivity and subgroup analyses at the 12 week primary endpoint. A sensitivity analysis was performed by selecting trials that utilized CPN as the control group. Subgroup analysis was concurrently conducted to compare interventional modalities (SNN vs. SNRF) within this subset. (A) Sensitivity and subgroup analyses comparing analgesic efficacy: in the CPN-controlled sensitivity analysis, the magnitude of pain reduction was comparable between the SND and CPN groups. In the subsequent analysis, no significant difference was identified between the SNN and SNRF modalities. (B) Sensitivity and subgroup analyses comparing opioid-sparing effects: in the CPN-controlled sensitivity analysis, opioid reduction was comparable between the SND and CPN groups. However, the subgroup analysis revealed a significant difference between the modalities, favoring SNRF over SNN. SD: standard deviation, SMD: standardized mean difference, CI: confidence interval, SNN: splanchnic neurolysis, CPN: celiac plexus neurolysis, SNRF: splanchnic nerve radiofrequency ablation, SND: splanchnic denervation.

Subgroup analysis within this subset revealed contrasting findings at 12 weeks. For analgesic efficacy, no significant difference was observed between the SNN and SNRF modalities (P = 0.836; Fig. 4A). However, for opioid-sparing effects, a significant subgroup difference was identified (P < 0.001; Fig. 4B), suggesting a potential clinical advantage of SNRF over SNN in reducing daily opioid requirements at 12 weeks post-intervention.

Egger’s regression test indicated no significant evidence of publication bias for 12-week outcomes (analgesic efficacy: t = −0.90, df = 4, P = 0.421; opioid-sparing effects: t = 0.57, df = 3, P = 0.610).

5. Quality of life

Six studies, involving 311 patients, reported QoL scores at 4 weeks post-intervention [28–33]. Global QoL was assessed using a variety of validated instruments across the included trials as follows: a composite mean of the Patient Satisfaction Scale and Performance Status [28], the Global Health Status/QoL scale of the EORTC QLQ-C30 [29,31], the total score from the FACT-G [32], and composite means calculated from the physical and mental component summaries for the SF-36 [30] and from the four primary domains for the WHOQOL-BREF [33]. The meta-analysis showed no significant difference between the SND and control groups (SMD = 0.70; 95% CI, −0.28 to 1.67; P = 0.127; I2 = 85.3%; random-effects model; Fig. 5).

Fig. 5.

Forest plot comparing QoL between the SND and control groups. Analytic comparison of absolute QoL scores was performed at 4 weeks post-intervention, measured by validated scales. SMDs with 95% CIs are presented using a random-effects model. SD: standard deviation, SMD: standardized mean difference, CI: confidence interval, CPN: celiac plexus neurolysis, QoL: quality of life, SND: splanchnic denervation.

Notably, a sensitivity analysis performed by excluding the study by Amr Erfan et al. [32] resulted in a marked reduction in I² to 6.0%. Under the fixed-effects model, this analysis demonstrated that the SND group had significantly higher QoL scores compared to the control group (SMD = 0.37; 95% CI, 0.12 to 0.62; P = 0.004; Supplementary Fig. 6).

6. Adverse events

Safety profiles regarding procedure-related complications were compared between the SND and control groups. The SND group had a significantly lower risk of hypotension than the control group (RR = 0.45; 95% CI, 0.25 to 0.80; P = 0.007; fixed-effects model; Fig. 6A) [28,29,31–33]. Similarly, the risk of diarrhea was significantly reduced in the SND group (RR = 0.24; 95% CI, 0.10 to 0.54; P < 0.001; fixed-effects model; Fig. 6B) [28,31–33]. For these outcomes, fixed-effects models were used due to low inter-study heterogeneity. No significant difference was identified between the two groups for backache (RR = 1.45; 95% CI, 0.43 to 4.88; P = 0.399; random-effects model; Fig. 6C) [29,31–33]. No other major or life-threatening complications were reported across the included studies.

Fig. 6.

Forest plots comparing procedure-related adverse events between the SND and control groups. The relative risks of (A) hypotension, (B) diarrhea, and (C) backache were compared using RRs with 95% CIs. Fixed-effects or random-effects models were applied based on the observed inter-study heterogeneity for each outcome. RR: risk ratio, CI: confidence interval, CPN: celiac plexus neurolysis, SND: splanchnic denervation.

7. Risk of bias

As illustrated in Fig. 7, the overall risk of bias was classified as “low” in one study [30] and as having “some concerns” in the remaining five studies [28,29,31–33]. Two trials provided only general statements about randomization without detailing specific methods for sequence generation or allocation concealment, leading to “some concerns” for the randomization process [31,32]. Except for the study by Dong et al. [30], which utilized a sham procedure, all other studies were classified as having “some concerns” regarding deviations from intended interventions. This rating primarily reflects the inherent structural difficulty of blinding interventionists within these clinical trial settings.

Fig. 7.

The risk of bias was assessed via the Cochrane Risk of Bias tool.

8. Quality of evidence

As summarized in Supplementary Table 2, the certainty of evidence for analgesic efficacy at 12 weeks, opioid-sparing effects at 12 weeks, and QoL at 4 weeks was rated “very low” due to significant inconsistency and imprecision. In contrast, procedure-related adverse events demonstrated relatively higher certainty: the evidence for hypotension was rated “high,” for diarrhea “moderate,” and for backache “very low.”

DISCUSSION

This meta-analysis was conducted to provide a comprehensive evaluation of SND compared to conventional control modalities for refractory upper abdominal malignant pain. The results demonstrated that longitudinal changes in pain intensity and opioid consumption from baseline were comparable between the SND and control groups across all evaluated time points. Similarly, QoL scores at 4 weeks post-intervention showed no significant differences between the groups. Notably, however, the risks of hypotension and diarrhea were significantly lower in the SND group than in the control group. These findings suggest that although SND provides analgesic efficacy and opioid-sparing effects comparable to conventional methods such as CPN, it may provide a superior safety profile by reducing procedure-related adverse events.

Patients with inoperable upper abdominal malignancies frequently experience severe pain throughout their disease trajectory, often requiring IPPs when conservative management provides limited pain relief [7]. The celiac plexus has long been the primary interventional target because it serves as a critical relay station for nociceptive signals from various upper abdominal organs, including the pancreas, liver, gallbladder, stomach, small bowel, and proximal transverse colons [34]. Consequently, CPN has been the most extensively utilized technique, with decades of evidence supporting its analgesic and opioid-sparing efficacies [5–7].

However, CPN presents significant technical challenges and anatomical limitations. The target is located deep within the retroperitoneum, often necessitating complex needle manipulation and advanced technical proficiency [35,36]. Consequently, serious adverse events such as retroperitoneal bleeding, abscess formation, and ischemic neurological injury can occur [36]. Beyond these technical risks, the clinical efficacy of CPN is constrained by anatomical distortions in the celiac region. Sequential studies by De Cicco et al. [8,37] demonstrated that anatomical distortions around the celiac plexus significantly hinder the optimal spread of the injectates compared to those with normal anatomy. Furthermore, Akhan et al. [38] established that the severity of tumoral invasion in the celiac region is a critical predictor of CPN success, with extensive invasion leading to shorter and less effective pain relief.

Considering these findings, SND offers significant anatomical and technical advantages over CPN. Unlike the celiac plexus, which forms a dense, irregular network around the celiac artery within the retroperitoneal space, the long-path thoracic splanchnic nerves are situated more superiorly and posteriorly in the supradiaphragmatic area [39]. Previous cadaveric studies have shown that although these nerves exhibit morphological diversity, they consistently descend along the lateral aspect of the vertebral bodies and predominantly converge into a focal intersection zone at a fixed landmark near the T11–L1 level before penetrating the diaphragm [13,39,40].

Given these frameworks, the present findings have substantial clinical implications. The meta-analysis demonstrated that SND achieved efficacy comparable to the control group, sustaining a consistent magnitude of reduction in pain intensity and opioid consumption across all evaluated time points. Furthermore, the sensitivity analyses specifically comparing SND with CPN at 12 weeks confirmed no significant differences in analgesic efficacy or opioid-sparing effects.

Beyond analgesia, SND demonstrated a superior safety profile. The significantly lower risk of procedure-related adverse events, such as hypotension and diarrhea, establishes SND as a more predictable and stable intervention, especially relevant for older patients or those with poor performance status. For these vulnerable populations, minimizing systemic hemodynamic instability and gastrointestinal distress is as crucial as pain control, positioning SND as a compelling and safer alternative to conventional celiac plexus interventions.

A key advantage of SND highlighted by the author’s analysis is its technical versatility, especially its compatibility with RF ablation. Unlike the celiac plexus, where the dense neural network is inextricably linked with major vascular structures [9], the focal convergence of splanchnic nerves allows precise and safe application of thermal energy [16]. The subgroup analysis suggested a potential advantage of SNRF over SNN in reducing opioid requirements, consistent with previous findings that RF ablation may produce a more pronounced opioid-sparing effect compared to chemical neurolysis [16]. While this finding is currently based on a limited number of trials, it highlights the clinical possibility that the localized energy delivery via RF ablation may offer distinct therapeutic benefits compared to chemical neurolytic agents.

This study builds upon the comparative framework established by Lu et al. [35]. Unlike previous single-arm meta-analyses [41,42], Lu et al. [35] provided significant insights directly comparing SNN and CPN in abdominal cancer pain. While their work established a valuable foundation, the present research advances the evidence by employing a more rigorous and homogenized comparative framework, thereby strengthening the clinical understanding of SND.

First, only RCTs were included, to ensure a higher level of evidence by eliminating potential confounding factors inherent in observational data. Furthermore, the authors’ analysis focused solely on percutaneous procedures, excluding surgical splanchnicectomy to provide findings directly applicable to interventional pain management. Second, a distinctive strength of this study is the structured longitudinal assessment of both pain intensity and opioid consumption at synchronized time points (4, 8, 12, and 24 weeks). By analyzing CfB rather than absolute post-procedural values, a more reliable assessment of the clinical trajectory and temporal stability of SND was provided. Third, beyond analgesic efficacy, a comprehensive meta-analysis of QoL and procedure-related adverse events was performed supported by GRADE-based certainty assessments. This established a multidimensional safety and efficacy profile, offering a clearer clinical guidance for managing refractory cancer pain. Fourth, the present study employed a statistically rigorous approach, incorporating meta-regression and sensitivity analyses to explore and account for heterogeneity, thereby strengthening the robustness of its conclusions.

Despite these advancements, some limitations warrant consideration. First, the total number of included RCTs and overall sample size were relatively small, which may have limited the statistical power to detect subtle differences between groups. Second, the certainty of evidence for primary efficacy outcomes was rated “very low” according to the GRADE system. This rating reflects the inherent structural limitation of achieving perfect double-blinding in interventional procedural trials, which primarily contributed to the risk of bias. Nevertheless, the longitudinal consistency of the authors’ findings across 24 weeks suggests that SND remains a clinically reliable intervention for managing refractory cancer pain. Third, inclusion of both sham-controlled and CPN-controlled RCTs introduced heterogeneity in the control arm. However, the meta-regression analysis at the 12-week primary endpoint revealed that control modality explained none (0%) of the observed heterogeneity. Fourth, although the subgroup analysis suggested that SNRF may be a promising alternative to chemical neurolysis, the limited number of included trials precludes a definitive conclusion. Further large-scale RCTs are necessary to firmly establish the comparative efficacy of SNRF in clinical practice. Finally, the evaluation of QoL was limited to the short-term period to ensure the inclusion of the largest available patient cohort, which may not fully represent the long-term impact of the intervention.

In conclusion, SND appears to be a clinically promising and safe interventional strategy for managing refractory upper abdominal cancer pain. The authors’ findings specifically demonstrate that SND maintained a reliable therapeutic trajectory, achieving analgesic efficacy comparable to the conventional control modalities throughout a 24-week longitudinal period. Furthermore, its technical feasibility in the context of distorted anatomy, combined with a superior safety profile, highlights SND as a viable alternative within the multimodal management of advanced abdominal malignancies. Further large-scale, high-quality randomized trials are warranted to confirm these findings.

SUPPLEMENTARY MATERIALS

Supplementary materials can be found via https://doi.org/10.3344/kjp.26106.