Sage Journals: Discover world-class research

Abstract

Study Design

Systematic Review.

Objective

To review the effect of NSAID administration on postoperative outcomes following posterior spinal surgery, with specific attention to pain control, opioid reduction, bleeding risk, and fusion outcomes.

Methods

PubMed, Google Scholar, Cochrane Library, Embase, Scopus, and Web of Science were queried through September 2025. Studies were included if they included adult patients undergoing posterior spinal procedures, including both fusion and non-fusion surgeries. In addition, studies had to report perioperative NSAID use in posterior spinal surgery with relevant clinical outcomes: Postoperative pain, opioid consumption, blood loss, wound complications, nonunion, and hardware failure. Bias was assessed with MINORS and Newcastle–Ottawa Scales.

Results

Eleven studies (396,206 patients) were included. Across posterior fusion and non-fusion procedures, short-term NSAID use consistently reduced postoperative opioid consumption and improved early pain control, with no clinically meaningful increase in bleeding or wound complications. However, prolonged NSAID use was associated with increased risks of nonunion, hardware failure, and wound complications in fusion cohorts. Significant variability in NSAID type, dose, timing, and study design exists.

Conclusions

Short-term NSAID use appears safe and effective as an opioid-sparing strategy in non-fusion posterior spinal surgery. Prolonged or high-dose administration may compromise fusion outcomes. Future high-quality studies are needed to definitively define the optimal postoperative NSAID window, dosing, and administration route.

Keywords

NSAID postsurgical outcomes posterior spinal surgery fusion complications

Introduction

Effective postoperative pain management is critical to optimizing surgical outcomes,¹ as in orthopedic surgery, inadequate analgesia has been linked to increased complications, prolonged hospitalization, and reduced patient satisfaction.^2,3 While opioids remain a cornerstone of postoperative pain control, their extensive side effect profile has driven the shift toward opioid-sparing strategies. Several studies suggest that reduced opioid exposure may improve patient outcomes, as opioids do not always enhance immediate postoperative pain relief and are associated with increased adverse events.^4,5

Nonsteroidal anti-inflammatory drugs (NSAIDs) are commonly used as adjuncts to opioids, providing analgesic and anti-inflammatory benefits without sedative or addictive effects. Their use across orthopedic surgery has been associated with reduced opioid consumption and mitigation of opioid-related complications.^6-9 However, concerns persist that NSAIDs may impair bone healing, based on both preclinical models and clinical observations.¹⁰ NSAIDs inhibit cyclooxygenase enzymes, lowering prostaglandin production — a mechanism that, while beneficial for pain and inflammation, may negatively affect osteogenesis.¹¹ Prior systematic reviews have explored NSAID use in spinal surgery, highlighting potential impacts on bone healing, fusion rates, pain control, and opioid consumption in biochemical, animal, and human models, but emphasize the need for higher-quality, targeted studies.¹²

Posterior spinal fusion remains among the most frequently performed surgeries for degenerative, traumatic, and deformity-related spinal conditions. Existing literature demonstrates that anterior, posterior and transforaminal approaches to surgery may have different risks and outcomes, with anterior approaches being associated with mortality, DVT and GI events, while posterior approaches are associated with an increased need for blood transfusion.^13-15 Given this complication profile, clarifying the risks and benefits of NSAID use, a class of drugs that may worsen bleeding, in this population is critical to improve outcomes. Previous systematic reviews have examined the impact on NSAIDs on spinal surgery broadly, not considering these different complication rates.

This systematic review isolates posterior procedures and provides procedure-specific risk stratification of NSAID use by (1) distinguishing fusion from non-fusion surgeries, (2) separating short term and prolonged dosing-an approach not taken in previous spine NSAID reviews.

Methods

We conducted a systematic review to evaluate the safety and efficacy of perioperative NSAID use in adult patients undergoing posterior spinal surgery. The review was conducted in accordance with the PRISMA 2020 (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) and MINORS (Methodological index for non-randomized studies).

Search Strategy

Literature review utilized the following databases: PubMed, Google Scholar, Cochrane Library, Embase, Scopus, and Web of Science. The search included studies published in English or translated, with no date restrictions, and was last updated on 9/23/2025.

Eligibility Criteria

This study included human studies on patients undergoing posterior spinal surgeries. Surgeries included those conducted in the lumbar, thoracic and cervical regions. Perioperative NSAID use of any type, dose or route were included. Study types included randomized controlled trials, prospective cohort studies, and retrospective cohort studies. Included studies were required to report at least 1 of the following relevant outcome measures: blood loss, pain control, or postoperative complications. Studies that were either explicitly focusing on anterior spinal approaches or lacked clear specification of surgical approach were excluded. In addition, case reports, editorials, narrative reviews and letters were excluded. Pediatric (as defined by age <18 years at the time of surgery), animal, and in-vitro studies were also excluded. Studies not reporting relevant clinical outcome measures were excluded. Lastly, studies without English or translated versions, as well as conference abstracts without accessible data, were excluded.

Study Selection

Two independent reviewers screened all titles and abstracts for relevance. Full-text reviews were conducted for studies that met the inclusion criteria or where eligibility remained unclear. Discrepancies were resolved through discussion.

A total of 1101 studies were identified through database searches. Following title and abstract screening, 61 studies were selected for abstract review, of which 11 met all inclusion criteria and were analyzed to be included in the final analysis. The study selection process is illustrated in a PRISMA flow diagram, with detailed documentation of exclusion criteria at each stage (Figure 1).

Figure 1.

PRISMA 2020 flow diagram depicting study identification, screening, eligibility assessment, and inclusion for systematic review of NSAID use in posterior spinal surgery

Data Extraction

Data extraction was performed independently by 2 reviewers using a standardized collection form. Information collected included.

• Study title, year, authors, design

• Reported clinical outcomes: Bleeding metrics, including estimated blood loss (EBL) and transfusion rates; Pain control, including Visual Analog Scale (VAS) scores

In addition, demographics including cohort age, gender, surgery type, NSAID used and route of administration were recorded. (Table 1).

Table 1.

Demographics of Individual Studies & Pooled Analysis

Author	Age (pooled mean ± 1SD)	Gender (%)	Surgery type	NSAID type	Total participants (excluding dropouts)
Chaibhuddanugul et al.¹⁶	60.90 ± 6.9	27.7% male, 72.3% female	Multilevel lumbar decompression with posterior lateral fusion	Ketorolac	47
Gaus et al.¹⁷	45.64 ± 11.29	Not recorded	Elective decompression and posterior stabilization	Ibuprofen	39
Gonzalez-suarez et al.¹⁸	52.06 ± 9.28	45.43% male, 54.57% female	Single level and multilevel posterior lumbar fusion	All NSAID types included	72,786
Kraiwattanapong et al.¹⁹	67.43 ± 6.27	51.02% male, 48.98% female	1-2 level spinal process splitting lumbar laminectomy	Ketorolac	49
Li et al.²⁰	59.61 ± 10.18	40% male, 60% female	Single level transforaminal lumbar interbody fusion	Loxoprofen	80
Nakatsuka et al.²¹	63.50 ± 0.918	47.79% male, 52.21% female	1-2 level transforaminal lumbar interbody fusion	All NSAID types included	429
Park et al.²²	52.68 (no SD reported)	32.95% male, 67.05% female	Postereolateral lumbar spinal fusion	Ketorolac	88
Senker et al.²³	Only table given; No exact mean age reported	38.67% male, 61.33% female	1-4 level minimally invasive spinal fusion	All NSAID types included	181
Siribumrungwong et al.²⁴	57.27 ± 10.96	34.38% male, 65.63% female	1-3 level decompressive laminectomy and fusion	Parecoxib and ketorolac	96
Yukawa et al.²⁵	59.98 ± 14.91	51.39% male, 48.61% female	1-2 level lumbar interbody fusion with instrumentation	Diclofenac	72

Risk of Bias Assessment

Two reviewers independently assessed risk of bias for each study using the Methodological Index for Non-Randomized Studies (MINORS) for non-randomized designs and the Newcastle–Ottawa Scale (NOS) for observational cohorts. Disagreements were resolved by consensus with a third reviewer when needed. Study-level scores and domain-level judgments are presented in Table 2).

Table 2.

Risk of Bias Assessment via MINORS and Newcastle-Ottawa Scales

Paper	Zygourakis (2025)	Li (2023)	Chaibhuddanugul (2024)	Nakatsuka (2025)	Senker (2022)	Siribumrungwong (2015)	Park (2005)	Yukawa (2005)	Kraiwattanapong (2020)	Gaus (2023)	Linsday (2022)
Selection
Truly or somewhat representative of average adult in community	1	1	1	1	1	1	1	1	1	1	1
Drawn from same community as exposed cohort	1	1	1	1	1	1	1	1	1	1	1
Secured records for ascertainment of exposure	1	2	2	2	2	2	2	2	2	2	2
Demonstration that outcome of interest was not present at start of study	1	1	1	1	1	1	1	1	1	1	1
Comparability
Study controls for age/sex	1	1	1	1	1	1	0	1	1	1	1
Study controls for at least 3 additional risk factors	0	0	0	0	0	0	0	1	0	0	1
What are they controlling for?	CCI	BMI, spinal level, operative time, drain output, ODI	Operative time, ASA, BMI, ODI, back pain, leg pain, diagnosis, level of fusion	Surgical characteristics Deep vs superficial infection, 30 and 90 day reoperation	Number of levels of levels fused	ASA, surgical duration, EBL, postop bleeding	Smoking, diagnosis, fusion levels, operative time, EBL, walking day, hospital stay	Height/Weight, operative levels, operative time & blood loss	Weight, height, BMI, ASA, spinal level, ODI, RMDQ, ODI, operative time, blood loss, drain output	Surgical duration, EBL, ASA	CCI, tobacco use, BMI
Outcomes
Independent blind assessment	0	2	2	0	0	2	1	1	1	1
Long enough follow up (>3 months)	1	0	0	0	0	0	1	0	1	0
Adequacy follow up, (loss to follow up)	1	2	2	2	2	2	2	2	2	2
Newcastle-ottawa score	7	10	10	8	8	10	9	10	10	9	7
MINORS score	19	23	23	20	19	23	19	20	24	22	17
Clearly stated aim	2	2	2	2	2	2	2	2	2	2	2
Inclusion of consecutive patients	2	2	2	2	2	2	2	2	2	2	2
Prospective collection of data	0	2	2	0	1	2	0	2	2	2	0
Endpoints appropriate to the aim of the study	2	2	2	2	2	2	2	2	2	2	2
Unbiased assessment of study endpoint	1	2	2	1	1	2	1	1	2	1	1
Follow-up period appropriate to the aim of the study	2	1	1	1	1	1	2	1	2	1	2
Loss to follow up less than 5%	2	2	2	2	2	2	2	2	2	2	1
Prospective calculation of the study size	0	2	2	2	0	2	0	0	2	2	0
An adequate control group	2	2	2	2	2	2	2	2	2	2	2
Contemporary groups	2	2	2	2	2	2	2	2	2	2	2
Baseline equivalence of groups	2	2	2	2	2	2	2	2	2	2	1
Adequate statistical analysis	2	2	2	2	2	2	2	2	2	2	2

Data Synthesis

Studies were grouped based on clinical outcome findings. Quantitative pooling was deemed inappropriate due to substantial clinical and statistical heterogeneity among studies. Specifically, included trials varied in NSAID class (ketorolac, ibuprofen, diclofenac, parecoxib), administration route (topical, infiltration, systemic), duration of therapy, and outcome reporting metrics (VAS vs NRS pain scales, inconsistent nonunion definitions). This heterogeneity violated assumptions of homogeneity required for meta-analysis; therefore, results were synthesized qualitatively using a structured narrative framework emphasizing consistency of effect direction.

Results

Eleven studies, including 5 randomized controlled trials and 6 retrospective or database-based cohort studies, were included. Outcomes were categorized into 3 major domains: (1) Opioid-Sparing and Pain Control, (2) Postoperative Complications and Healing (3) Bleeding and Drainage. Across studies, NSAID type, administration route, timing, and duration varied. Several studies reported significant findings across several domains.

Grouped Outcomes

Several studies reported outcomes spanning more than 1 domain. Specifically, Chaibhuddanugul et al, Siribumrungwong et al, Yukawa et al, Li et al, and Nakatsuka et al reported both pain control and drainage-related outcomes. In addition, Gonzalez-Suarez et al reported outcomes related to both postoperative complications and readmissions. See Tables 3-5 for respective outlines.

Table 3.

Summary of Opioid Consumption and Pain Control Studies

Title	Author	Study type	Primary outcome	Secondary outcome	Significant primary outcomes (Treament vs control/Comparator Groups)
Addition of ketorolac to local anesthesia for wound infiltration in multilevel posterior lumbar spinal fusion: A randomized, double-blinded, placebo-controlled trial	Chaibhuddanugul et al.	Randomized controlled trial	Post-op morphine consumption (using PCA), 11-point numeric pain rating (NRS) scale pain rating scale for back/leg pain	Drain volume, postop hemoglobin, creatinine, AST, ALT, complications (nausea/vomiting, gastrointestinal bleeding, delayed wound healing, hematoma), hospital stay	24 hour total morphine consumption: 14.1 ± 8.7 vs 23.1 ± 14.2, P = 0.03
					Mean 96 hour total morphine consumption: 23.5 ± 14.6 vs 37.4 ± 23.8, P = 0.03
					Median NRS pain scale 0-6hrs postop: 2.2 vs 1.4, P < 0.001
Preventive efficacy of ibuprofen and dexamethasone combination for postoperative pain in posterior vertebral stabilization	Gaus et al.	Randomized controlled trial	Pain score using NRS scale at rest and with movement (2,6,12,24hr post op), IL-6 measurement (preop,2hr postop, 24hr postop), total post-op fentanyl use (with PCA)	Surgery duration, bleeding volume	24 hr NRS pain score in motion: 1.77±0.93 vs 2.77 ± 0.44 vs 2.31 ± 0.86, P = 0.005
					24 hr NRS pain score at rest
					1.15 ± 0.38 vs 2.62 ± 0.56 vs 1.46 ± 0.52, P < 0.001
					IL-6 level (pg/ml) 24 hr postop
					0.175 (0.143-0.233) vs 0.286 (0.246-0.318) vs 0.179 (0.007-0.266), P = .044
					Total opioid consumption: 48.46 ± 37.82 vs 226.15 ± 59.52 vs 90.0 ± 24.64, P < 0.001
					*3 comparator groups used; data reported as ibuprofen 800 mg and dexamethasone 10 mg (group B) vs intravenous ibuprofen 400 mg and dexamethasone 10 mg (group A) vs ibuprofen 800 mg alone (group C)
Effect of multimodal drugs infiltration on postoperative pain in split laminectomy of lumbar spine	Kraiwattanapong et al.	Randomized controlled trial	Amount of morphine consumption and visual analogue scale (VAS) for pain	Oswestry disability index (ODI) at 2w and 3 months, roland-morris low back pain and disability questionnaire	48 hr total morphine consumption: 10.0 ± 3.4 vs 24.1 ± 6.3, P < 0.001
				(RMDQ)at 2w and 3m, patient satisfaction, length of hospital stay, and side effects of pain medication (N/V, itching/rash, urinary retention, constipation, overall
					Effect size of VAS at rest 48 hrs postop (95% CI): −2.0 (−2.5, −1.4)
					Effect size of VAS at rest 48hrs postop (95% CI): −2.0 (−2.6−1.4)
					*Values not explicitly stated for VAS scores
Effect of topical nonsteroidal anti-inflammatory drugs around the incision on postoperative pain in transforaminal lumbar interbody fusion surgery: A double-blind randomized controlled trial	Li et al.	Randomized controlled trial	Amount of morphine consumption and visual analogue scale (VAS) for pain	Time of first analgesic demand, operation time, postoperative drain output, side effects of opioids (n/v), postoperative stay, and oswestry disability index (ODI) score at 1m follow up	Time to first PCA: 110.75 ± 51.51 vs 70.75 ± 41.70 (P < 0.001)
	Li et al.	Randomized controlled trial			48 hr total sufentanil citrate usage (μg): 10.20 ± 2.45 vs 13.93 ± 3.51, P < 0.001
A prospective randomized study of preemptive analgesia for postoperative pain in the patients undergoing posterior lumbar interbody fusion: Continuous	Yukawa et al.	Prospective randomized clinical trial	Visual analog pain scale (VAS), supplemental analgesic demands, time to first request	Complications (nausea, vomiting, discontinuation, “side effects”)	Supplemental analgesic use: 3.7 ± 2.0 vs 2.6 ± 2.0 vs 2.1 ± 2.1, P = 0.013
					*3 comparator groups; data reported as diclofenac sodium
					Suppository vs continuous subcutaneous morphine vs continuous epidural analgesia

Data presented as mean ± SD or median (min -max) unless otherwise specified.

Table 4.

Summary of Postoperative Complications & Healing Papers

Title	Author	Study type	Primary outcome	Secondary outcome	Significant primary outcomes (OR; 95% CI)	Significant secondary outcomes (OR; 95% CI)	Grouping parameters
Association between nonsteroidal anti-inflammatory drugs use and surgical outcomes following posterior lumbar fusion: A medical claims database analysis	Gonzalez-suarez et al.	Retrospective cohort study	30-day readmissions, LOS, pseudoarthrosis/hardware failure, and wound complications (hemorrhage, hematoma, infection, and dehiscence) based on time starting NSAIDs postop	30-day readmissions, pseudoarthrosis/hardware failure, and wound complications (hemorrhage, hematoma, infection, and dehiscence), evaluating dispersing intervals (1-30d, 31-90d, 91-180d, 181-365d, 366-720d) *LOS not evaluated	NSAIDs started <72hrs postop	<72hr Postop Dispersed 1-30d	Treatment group received NSAIDs postop; control group no use of NSAIDs postop
					Wound complications	Wound complications: OR = 0.571; (0.391 - 0.832)
					OR = 0.692, (0.538 - 0.889)	72hr - 90d postop
					Length of stay: OR = 0.736, (0.706 - 0.768)	Dispersed 1-30d
					NSAIDs started 72hrs - 90d postop	Readmission: OR = 1.313 (1.125 - 1.532)
					Pseudoarthrosis and hardware failure: OR = 1.042, (1.017 - 1.068)	Wound Complications:OR = 1.427 (1.286 - 1.583)
					Wound complications	Dispensed 181-365d
					OR = 1.069 (1.035-1.105)	Pseudoarthrosis and hardware failure: OR = 1.212 (1.114 - 1.317)
					NSAIDs started 90d - 1y postop	90d -1y postop
					Pseudoarthrosis and hardware failure: OR = 1.024 (1.009 - 1.039)	Dispensed 91-180d
						Pseudoarthrodesis and hardware failure: OR = 1.154 (1.035 - 1.287)
The effects of ketorolac injected via PCA on spinal fusion	Park et al.	Retrospective cohort study	Radiographic evidence of nonunion (1-4 scale with grades of “probably not solid” and “definitely not solid” classified as a nonunion) measured 1y postop	1-4 scale of clinical outcomes (based on pain, physical activity ability, analgesia required) scores of 1 and 2 are considered unsatisfactory	>1y nonunion: OR = 5.64, P < 0.05 (CI not reported)	Treatment group had a significantly less patients with satisfactory clinical outcome at 1 year postop satifactory clincal outcomes: 76.7% vs 93.1%, P < 0.05 (OR/CI not reported)	Treatment group received ketorolac + fentanyl in PCA; control received only fentanyl
Nonsteroidal anti-inflammatory drugs in the acute post-operative period are associated with an increased incidence of pseudarthrosis, hardware failure, and revision surgery following single-level spinal fusion	Lindsay et al	Retrospective cohort study	Incidence of pseudarthrosis after single-level posterior lumbar spine instrumentation and fusion	Incidence of hardware failure & incidence of revision surgery	NSAID	Hardware failure	NSAID usage
					Pseudoarthrosis: OR = 1.49, (1.41-1.57)	NSAID: 2.47% vs 1.45%; OR = 1.73,	COX-2 inhibitor usage
					COX-2	COX-2 inhibitor: 2.16% vs 1.56%; OR = 1.39	Ketorolac use (inpatient)
					Pseudoarthrosis: OR = 1.45, (1.31-1.61)	Revision surgery	Control groups: No NSAID/COX-2 inhibitor use)
					NSAID use associated with increased pseudarthrosis (7.34% vs 5.06%; OR = 1.49	NSAID: 9.89% vs 5.24%; OR = 1.98,
					COX-2 inhibitor use associated with increased pseudarthrosis (7.50% vs 5.30%; OR = 1.45	COX-2 inhibitor: 10.3% vs 5.72%; OR = 1.89
						NSAID
						Hardware failure: OR = 1.73 (1.5-1.90)
						Revision surgery: OR = 1.98 (1.89-2.08)
						COX-2
						Hardware failure: OR = 1.39 (1.15-1.68)
						Revision surgery: OR = 1.89 (1.72-2.07)

Data reported as Treatment vs Control.

Table 5.

Summary of Bleeding and Drainage Papers

Title	Author	Study type	Primary outcome	Secondary outcome	Significant primary outcomes	Significant secondary outcomes	Grouping parameters
Preoperative NSAID use is associated with a small but statistically significant increase in blood drainage in TLIF procedures	Nakatsuka et al.	Retrospective cohort study	Drain output	LOS, postop complications, return to OR, DVT, UTI, infection	There was a significantly greater drain output in treatment group (not clinically significant)	No significant differences in secondary outcomes	Treatment group received NSAIDs with 7 days of procedure; control group did not
	Nakatsuka et al.	Retrospective cohort study	Drain output		Linear regression NSAID on drain output (mL/d): 21.44580 ± 9.66705, P = 0.027	No significant differences in secondary outcomes
Nonsteroidal anti-inflammatory drugs (NSAID) do not increase blood loss or the incidence of postoperative epidural hematomas when using minimally invasive fusion	Senker et al.	Retrospective cohort study	Blood loss	Hematoma formation	No significant intraoperative blood loss or postop drainage difference; remained insignificant when accounting for gender and number of fused spinal levels	3 patient formed postop hematoma, only 1 was from NSAID group; population too small for statistical analysis; assumption made there was no difference	Treatment group received NSAIDs within 24 hrs of surgery, control group stopped at least 3 days prior
Comparing parecoxib and ketorolac as preemptive analgesia in patients undergoing posterior lumbar spinal fusion	Siribumrungwong et al.	Randomized controlled trial	Verbal numerical pain score (VNRS) at hours 0,1,2,3,4,6,12,18; amount of postop morphine required (using PCA); drain output	Side effects of analgesia (dyspepsia, nausea/vomiting, constipation, dizziness, respiration depression [respiratory rate less than 8 breaths per minute], and pruritus)	Ketorolac: Significantly less pain reported at 0 and 1 hr post op paracoxib: Significantly less pain at 0 hours postop	No difference between NSAID interventions in pain score. No difference in any group in total morphine use. No difference in drain output between 3 groups
					Ketorlac vs Control
					Pain at 0 hr postop: 6.1 vs 8.5, P < 0.05
					Pain at 1 hr postop: 5.3 vs 6.9, P < 0.05
					Parecoxib vs Control
					Pain at 0 hr postop: 6.3 vs 8.5, P < 0.05

Opioid Consumption and Pain Control

Six studies examined the relationship between perioperative NSAID use and opioid consumption, as well as pain control, and all demonstrated statistically significant reductions in opioid use.

Chaibhuddanugul et al reported that patients receiving ketorolac with bupivacaine infiltration had a 44% reduction in morphine consumption during the first 24 h (P < 0.03) and a 35% reduction over 96 hours compared to controls (P < 0.03), with no significant differences in NRS pain scores.

Gaus et al found that patients receiving ibuprofen combined with dexamethasone had significantly lower NRS pain scores at all postoperative time points, both at rest and with movement, compared to those receiving either agent alone. This group also showed reduced total opioid consumption (48.46 ± 37.82 vs 226.15 ± 59.52 vs 90.0 ± 24.64, P < 0.001).

Kraiwattanapong et al observed that NSAID infiltration significantly reduced total morphine use and VAS pain scores at all postoperative intervals, with no significant differences in ODI scores at 2 weeks or 3 months.

Li et al demonstrated that topical NSAID administration led to significantly lower opioid consumption and VAS pain scores at all time points, a significantly longer time to first PCA demand (110.75 ± 51.51 vs 70.75 ± 41.70, P < 0.001), and fewer opioid-related side effects, though no other secondary outcomes reached significance.

Lastly, Yukawa et al found that patients receiving subcutaneous morphine with diclofenac sodium had significantly lower VAS pain scores at time zero postoperatively, though they required more supplemental analgesia within the first 72 hours. There were no significant differences in time to first analgesic request or complication rates (Table 3).

Postoperative Complications & Healing

Three studies assessed postoperative complications and healing in relation to perioperative NSAID use. Gonzalez-Suarez et al evaluated 30-day readmissions, length of stay (LOS), wound complications, nonunion, and hardware failure in posterior fusion surgery, stratifying results by the timing and duration of NSAID administration. They found that NSAIDs initiated within 72 hours postoperatively were associated with decreased wound complications (OR = 0.692, 95% CI 0.538 - 0.889) and shorter LOS (OR = 0.736, 95% CI 0.706 - 0.768). However, NSAIDs started between 72 hours and 90 days postoperatively were linked to increased wound complications (OR = 1.069, 95% CI 1.035-1.105) and nonunion (OR = 1.042, 95% CI 1.017 - 1.068), while NSAIDs initiated between 91 days and 1 year postoperatively were associated with higher rates of nonunion and hardware failure (OR = 1.024, 95%CI 1.009 - 1.039).

Park et al investigated radiographic nonunion as the primary outcome, graded by CT, with secondary outcomes including pain and function. Patients receiving ketorolac and fentanyl via PCA exhibited significantly higher nonunion rates (OR = 5.64, P < 0.05) and fewer satisfactory clinical outcomes (76.7% vs 93.1%, P < 0.05) at 1 year compared to those receiving fentanyl alone.

Lastly, Lindsay et al²⁶ examined nonunion rates following posterior lumbar spinal fusion, with secondary outcomes of hardware failure and revision surgery. Patients who used NSAIDs or COX-2 inhibitors within 6 weeks of surgery had significantly higher rates of hardware failure (OR = 1.73, 95% CI 1.5-1.90) and revision procedures (OR = 1.98, 95% CI 1.89-2.08). However, a separate inpatient ketorolac subgroup within the same study did not show an increased risk of postoperative complications (Table 4).

Bleeding and Drainage (Table 3)

Three studies evaluated perioperative NSAID use in relation to blood loss, drain output, and early postoperative recovery metrics. Nakatsuka et al identified drain output as the primary outcome and found that a 7-day postoperative NSAID taper was associated with significantly increased drain output. However, no significant differences were observed in length of stay, postoperative complications, reoperation rates, or DVT formation.

Senker et al focused on perioperative blood loss and postoperative drainage as primary outcomes, with hematoma formation as a secondary endpoint. They found no significant differences in blood loss or drainage, even when adjusting for gender and number of fused levels. Three cases of postoperative hematoma were reported, 1 of which occurred in the NSAID group, but statistical analysis of hematoma formation was not possible due to the small sample size.

Siribumrungwong et al assessed VNRS pain scores at multiple early postoperative intervals, with secondary outcomes including side effects, total morphine use, and drain output. Patients receiving ketorolac had significantly lower pain scores at 0 and 1 hour postoperatively (6.1 vs 8.5, P < 0.05 and 5.3 vs 6.9, P < 0.05, respectively), and those receiving parecoxib showed significantly lower pain scores at 0 hours (6.3 vs 8.5, P < 0.05). No significant differences were seen between NSAID groups at later time points, and there were no significant differences in total morphine use, drain output, or side effect profiles (Table 5).

Discussion

Summary

This review summarizes the current evidence around NSAID use in posterior spinal surgery, providing some evidence that short-term use reduces opioid requirements and improves pain without increasing bleeding or wound complications. In addition, in fusion cohorts, current evidence suggests that prolonged or high-dose exposure is associated with higher risk of nonunion, hardware failure, and wound complications. It is essential to note that current studies vary widely in NSAID type, route of administration, and duration, limiting the ability of researchers to make definitive claims regarding dosage or timing without further study. This posterior specific, timing aware framing is distinct from prior reviews that pooled anterior and posterior approaches.

NSAID effects on fusion are plausibly time-dependent because the biologic tasks of the hours after surgery (hemostasis, nociception, and acute inflammation) differ from those governing osteogenesis in the following weeks^27,28 In the early phase, COX inhibition reduces peripheral and central sensitization and limits opioid requirements without materially impacting osteoprogenitor activity. Beyond this window, persistent suppression of PGE₂ and downstream RUNX2/osterix signaling may impair endochondral ossification, angiogenesis (VEGF), and RANKL/OPG balance—mechanisms tightly coupled to graft incorporation. This phase-specific model reconciles our finding of short-term analgesic benefit with longer-term associations between extended NSAID exposure and nonunion in fusion cohorts.

Our synthesis supports an exposure–response hypothesis across 3 axes: duration (hours vs weeks), dose (standard vs high), and route (systemic vs local/topical). Trials using topical or limited inpatient dosing produce consistent analgesia with no signal for bleeding or nonunion, whereas outpatient continuation beyond discharge—particularly oral, systemic regimens—tracked with increased odds of pseudarthrosis and hardware failure in fusion. This suggests that cumulative systemic exposure in cohorts undergoing more invasive procedures, rather than a single perioperative dose in less invasive surgery, is the primary biologic driver of risk.

Fusion vs Non-fusion Procedure Distinction

Patients undergoing decompression-only procedures (eg, discectomy, limited laminectomy) typically require only brief analgesic courses, rarely have drains, and have no risk of nonunion or instrumentation failure.^27,28 Decompression procedures lack a biologic fusion mandate; analgesic goals dominate and brief NSAID courses align with those goals. In this context, short NSAID courses appear low risk and beneficial. In contrast, posterior fusion procedures introduce biologic and mechanical healing demands that render the construct vulnerable to dose- and duration-dependent NSAID effects. These procedures require angiogenic ingrowth and osteoblast differentiation at the graft–host interface. Any therapy that repeatedly suppresses COX-2/PGE₂ may attenuate these processes.¹² Thus, pooling decompression and fusion obscures procedure-specific biology and explains why “harm” signals appear only when fusion is required and exposure is prolonged. Stratifying by procedure type prevents dilution of fusion-specific risk signals and avoids the contradictory conclusions that arise when all posterior procedures are pooled, allowing providers to make clear clinical decisions based on a patient’s surgery specific needs.

Pain and Opioid Consumption

Six studies reported on postoperative pain and opioid consumption outcomes. Five of 6 studies found significant improvement in pain outcomes when NSAIDs were used in the perioperative period. Chaibhuddanugul et al, the only study that observed no change in postoperative pain score, still observed a decrease in total opioid usage, underscoring consistent evidence across both spine and non-spine literature that brief NSAID courses are effective opioid-sparing agents.

These findings are in line with literature outside of spinal surgery: Doleman et al reviewed 71 randomized control trials involving NSAID use in primarily abdominal, orthopaedic, and dental surgery. They found evidence that NSAIDs used in the perioperative period decreased postoperative pain up to 48 hours and decreased total morphine requirements.²⁹ In addition, Kein et al conducted a randomized control study in patients undergoing lumbar surgery with a treatment group receiving pregabalin and celecoxib postoperatively, leading to a significant decrease in VAS score and morphine consumption.³⁰

Postoperative Complications

Three fusion-focused studies (Gonzalez-Suarez, Park, Lindsay) consistently demonstrated a timing- and dose-dependent risk of nonunion and hardware complications. While early NSAID exposure may be protective, prolonged use clearly increases these risks. Notably, short inpatient ketorolac courses did not appear harmful, underscoring the importance of both timing and duration of administration in guiding NSAID safety.

Previous literature supports the temporal nature of NSAID’s impact on fusion in spinal surgery. A meta-analysis published by Dodwell et al isolated 7 studies and found no increased risk of nonunion after exposure to NSAID therapy.³¹ Importantly, in their combined analysis, when moderate-quality long-bone studies were considered separately, NSAID exposure was associated with a markedly increased risk of nonunion (OR 4.4, 95% CI 2.5-7.8). In contrast, spine studies alone did not show a significant effect (OR 2.2, 95% CI 0.8-6.3). Further subanalyses from Dodwell et al demonstrated that high- or low-dose i.v./i.m. Ketorolac and standard oral NSAIDs were not associated with increased nonunion risk, provided exposure was short and well documented (OR 0.8, 95% CI 0.4-1.4). However, studies with prolonged or poorly defined NSAID use revealed significantly increased nonunion rates (P = 0.001). These findings reinforce that both dose and exposure duration are critical determinants of NSAID safety in the perioperative spine setting. In addition, it points to ambiguity in the current guidelines regarding NSAID administration. This supports our finding that prospective research must be conducted to clearly define the route of administration, duration, and type of NSAID and its impact on fusion rates.

Sivaganesan et al, evaluating a variety of spinal surgery fusion modalities, also found little evidence that NSAIDs increased the risk of nonunion in spinal surgery in the short term, while chronic NSAIDs resulted in a higher rate of nonunion. The authors utilized 7 publications with levels of evidence of II and III to conclude that low-dose NSAIDs at a dosing period of 48 hours postoperatively are beneficial to postoperative pain without negative effects on bone fusion.¹²

This variance in findings between posterior spinal surgery and broader literature demonstrate a need for prospective studies that evaluate the risk of specific NSAIDs at specific intervals and their impact on postoperative outcomes in posterior spinal surgery.

Blood Loss & Drainage

Of the 4 studies examining these blood loss and drainage, Gaus et al and Senker et al found no increased postoperative bleeding volumes, while Li et al and Nakatsuka et al found no increased drain output.

This aligns with research outside of posterior spinal surgery. Weng et al found that dosing with parecoxib before and up to 2 days after total hip arthroplasty did not increase blood loss in the postoperative period.³² Laoruengthana et al compared perioperative blood loss in patients receiving either ketorolac or parecoxib. They found a 117.32 mL greater amount of blood loss in the ketorolac group without a difference in blood transfusion rate, suggesting a difference that was not clinically significant.³³ A meta-analysis by Zhang et al analyzed 4 randomized control studies on spinal fusion and found that continued aspirin use in the perioperative period did not increase risk of blood loss or postoperative blood transfusion.³⁴ These studies in conjunction provide evidence that perioperative NSAID use carries little risk of exacerbating postoperative bleeding.

Heterotopic Ossification

While the included studies did not report heterotopic ossification (HO) as a postoperative outcome, NSAIDs are well-established in orthopedic surgery as prophylaxis against HO, particularly following arthroplasty and disc replacement.^32,35-40 However, the applicability of these findings to poster, v6bior spinal fusion remains uncertain. Previous literature has found that while the incidence of HO is high in posterior fusions, (up to 81% at 2-5 years postoperatively), the clinical significance is minimal and may not warrant intervention.⁴¹ Considering that NSAIDs negatively impact fusion rates in posterior spinal surgery, the risk of nonunion and hardware complications outweighs any potential benefit of preventing radiographically evident but clinically insignificant HO.

Strengths and Limitations

NSAID type, dose, timing, and administration route varied substantially (ketorolac, ibuprofen, diclofenac, parecoxib via systemic, topical, or wound infiltration). Timing ranged from intraoperative through 1 year. This heterogeneity, coupled with outcome reporting variability, limited comparability and precluded meta-analysis. The evidence base comprises both randomized trials and retrospective database analyses, each with distinct limitations. The randomized trials were small and underpowered for rare complications, while the retrospective and claims-based studies, though larger, relied on administrative coding that may incompletely capture NSAID exposure or nonunion events. Despite generally high MINORS and NOS scores, unmeasured confounding cannot be excluded. These factors limit causal inference and underscore the need for prospective, exposure-standardized trials. While only 11 studies met inclusion, this reflects the paucity of posterior-specific literature and underscores the need for targeted prospective studies to identify dosing and duration thresholds associated with harm. Although prior systematic reviews have faced potential duplication among national registries, the present analysis included discrete institutional and database cohorts (PearlDiver, MarketScan, and individual hospital studies) from separate countries and years. Therefore, patient-level overlap is unlikely but should be considered.

Future Directions

Future research should prioritize well-designed prospective or randomized trials to strengthen the evidence base. Studies should clearly define NSAID exposure by type, dose, route, and duration, with standardized timing windows for preoperative, intraoperative, and postoperative administration. Establishing these parameters is essential to enable valid comparisons and to identify dose- and duration-specific thresholds that may influence fusion integrity and postoperative outcomes.

Disclosures

One author has multiple financial relationships with companies involved in surgical device development and perioperative therapeutics. Please see disclosure forms for specific dollar amounts. This study was conducted independently of any of the aforementioned companies and did not receive external funding. Reporting these affiliations serves to remain transparent. ICMJE disclosure forms have been submitted by all authors. No external funding was received for this work.

Conclusion

This systematic review provides some evidence that short courses of NSAIDs may be effective for pain control and opioid use reduction in non-fusion posterior spinal surgery. Retrospective studies examining the impact of NSAIDs on fusion rates provide some evidence that prolonged use of NSAIDs may be associated with higher nonunion rates and hardware related complication risks.

Footnotes

Acknowledgements

Mitchell Ng serves as a paid consultant for Styker, Johnson & Johnson, Curafix, Pacira, Sage, Alafair, Next Science, Bonitti, Ferghana & HOF.

ORCID iDs

Rajendra Singh

Mitchell Ng

Ethical Considerations

This article does not contain any studies with human or animal participants.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

Articles included in this review are all taken from peer reviewed, published journals.*

Appendix

References

van Boekel

RLM

Warlé

Nielen

RGC

, et al. Relationship between postoperative pain and overall 30-day complications in a broad surgical population: an observational study. Ann Surg. 2019;269(5):856-865. doi:10.1097/SLA.0000000000002583

Yang

MMH

Far

Riva-Cambrin

Sajobi

Casha

. Poor postoperative pain control is associated with poor long-term patient-reported outcomes after elective spine surgery: an observational cohort study. Spine J. 2024;24(9):1615-1624. doi:10.1016/j.spinee.2024.04.019

LWT

Suh

Chen

, et al. Early postoperative pain after total knee arthroplasty is associated with subsequent poorer functional outcomes and lower satisfaction. J Arthroplast. 2021;36(7):2466-2472. doi:10.1016/j.arth.2021.02.044

King

Geiger

, et al. American society for enhanced recovery and perioperative quality initiative joint consensus statement on perioperative opioid minimization in opioid-naïve patients. Anesth Analg. 2019;129(2):567-577. doi:10.1213/ANE.0000000000004194

Fiore

El-Kefraoui

Chay

, et al. Opioid versus opioid-free analgesia after surgical discharge: a systematic review and meta-analysis of randomised trials. Lancet. 2022;399(10343):2280-2293. doi:10.1016/S0140-6736(22)00582-7

Chunduri

Aggarwal

. Multimodal pain management in orthopedic surgery. J Clin Med. 2022;11(21):6386. doi:10.3390/jcm11216386

Duong

Ponniah

VanDeCapelle

, et al. Effect of a postoperative multimodal opioid-sparing protocol vs standard opioid prescribing on postoperative opioid consumption after knee or shoulder arthroscopy. JAMA. 2022;328(13):1326. doi:10.1001/jama.2022.16844

Hansen

Carlino

Saunee

Dasa

Bhandutia

. Modern perioperative pain management strategies to minimize opioids after total knee arthroplasty. Orthop Clin N Am. 2023;54(4):359-368. doi:10.1016/j.ocl.2023.05.002

Day

Boryan

. Current evidence-based approaches to multimodal pain control and opioid minimization after arthroscopic and knee preservation surgery. J Am Acad Orthop Surg. 2024;32(1):e24-e32. doi:10.5435/JAAOS-D-23-00342

10.

Zhang

Cai

. High-dose ketorolac affects adult spinal fusion. Spine. 2011;36(7):E461-E468. doi:10.1097/BRS.0b013e3181dfd163

11.

Lambrechts

D’Antonio

Heard

, et al. The inhibitory effect of NSAIDs and opioids on spinal fusion. Spine. 2024;49(12):821-828. doi:10.1097/BRS.0000000000004959

12.

Sivaganesan

Chotai

White-Dzuro

McGirt

Devin

. The effect of NSAIDs on spinal fusion: a cross-disciplinary review of biochemical, animal, and human studies. Eur Spine J. 2017;26(11):2719-2728. doi:10.1007/s00586-017-5021-y

13.

Makanji

Schoenfeld

Bhalla

Bono

. Critical analysis of trends in lumbar fusion for degenerative disorders revisited: influence of technique on fusion rate and clinical outcomes. Eur Spine J. 2018;27(8):1868-1876. doi:10.1007/s00586-018-5544-x

14.

Memtsoudis

Vougioukas

Gaber-Baylis

Girardi

. Perioperative morbidity and mortality after anterior, posterior, and anterior/posterior spine fusion surgery. Spine. 2011;36(22):1867-1877. doi:10.1097/BRS.0b013e3181c7decc

15.

Polly

Santos

ERG

Mehbod

. Surgical treatment for the painful motion segment. Spine. 2005;30(16 Suppl):S44-S51. doi:10.1097/01.brs.0000174529.07959.c0

16.

Chaibhuddanugul

Weerakul

Laoruengthana

Varakornpipat

Sudbanthad

Mahatthanatrakul

. Addition of ketorolac to local anesthesia for wound infiltration in multilevel posterior lumbar spinal fusion: a randomized, double-blinded, placebo-controlled trial. Spine. 2024;49(24):1716-1721. doi:10.1097/BRS.0000000000004998

17.

Gaus

Darise

Ala

Ahmad

Musba

Tan

. Preventive efficacy of ibuprofen and dexamethasone combination for postoperative pain in posterior vertebral stabilization: a randomized controlled study. Anaesth Pain Intensive Care. 2023;27(5):548-554. doi:10.35975/apic.v27i5.1851

18.

Gonzalez-Suarez

Green

Cavagnaro

Moya

Zygourakis

Desai

. Association between nonsteroidal anti-inflammatory drugs use and surgical outcomes following posterior lumbar fusion: a medical claims database analysis. Int J Spine Surg. 2025;19(2):224-236. doi:10.14444/8732

19.

Kraiwattanapong

Arnuntasupakul

Kantawan

Woratanarat

Keorochana

Langsanam

. Effect of multimodal drugs infiltration on postoperative pain in split laminectomy of lumbar spine: a randomized controlled trial. Spine. 2020;45(24):1687-1695. doi:10.1097/BRS.0000000000003679

20.

Wang

Yuan

Tian

Wang

Liu

. Effect of topical nonsteroidal anti-inflammatory drugs around the incision on postoperative pain in transforaminal lumbar interbody fusion surgery: a double-blind randomized controlled trial. Pain Physician. 2023;26(5):467-473.

21.

Nakatsuka

Kim

Protopsaltis

Fischer

. Preoperative NSAID use is associated with a small but statistically significant increase in blood drainage in TLIF procedures. Clin Spine Surg. 2025;38:E521-E527. doi:10.1097/BSD.0000000000001795

22.

Park

Moon

Park

Lee

. The effects of ketorolac injected via patient controlled analgesia postoperatively on spinal fusion. Yonsei Med J. 2005;46(2):245-251. doi:10.3349/ymj.2005.46.2.245

23.

Senker

Aspalter

Trutschnig

Franke

Gruber

Stefanits

. Nonsteroidal anti-inflammatory drugs (NSAID) do not increase blood loss or the incidence of postoperative epidural hematomas when using minimally invasive fusion techniques in the degenerative lumbar spine. Front Surg. 2022;9:1000238. doi:10.3389/fsurg.2022.1000238

24.

Siribumrungwong

Cheewakidakarn

Tangtrakulwanich

Nimmaanrat

. Comparing parecoxib and ketorolac as preemptive analgesia in patients undergoing posterior lumbar spinal fusion: a prospective randomized double-blinded placebo-controlled trial. BMC Muscoskelet Disord. 2015;16(Issue 1):59. doi:10.1186/s12891-015-0522-5

25.

Yukawa

Kato

Ito

Terashima

Horie

. A prospective randomized Study of preemptive analgesia for postoperative pain in the patients undergoing posterior lumbar interbody fusion continuous subcutaneous morphine, continuous epidural morphine, and diclofenac sodium. Spine. 2005;30(21):2357-2361.

26.

Lindsay

Travis

Ryu

Wright

Jung

. Nonsteroidal anti-inflammatory drugs in the acute post-operative period are associated with an increased incidence of pseudarthrosis, hardware failure, and revision surgery following single-level spinal fusion. Spine. 2023;48(15):1057-1063. doi:10.1097/BRS.0000000000004695

27.

Smorgick

Baker

Bachison

Herkowitz

Montgomery

Fischgrund

. Hidden blood loss during posterior spine fusion surgery. Spine J. 2013;13(8):877-881. doi:10.1016/j.spinee.2013.02.008

28.

Ogura

Dimar

JR Gum

, et al. Hidden blood loss following 2- to 3-level posterior lumbar fusion. Spine J. 2019;19(12):2003-2006. doi:10.1016/j.spinee.2019.07.010

29.

Doleman

Leonardi-Bee

Heinink

, et al. Pre-emptive and preventive NSAIDs for postoperative pain in adults undergoing all types of surgery. Cochrane Database Syst Rev. 2021;6(6):CD012978. doi:10.1002/14651858.CD012978.pub2

30.

Kien

Geiger

van Chuong

, et al. Preemptive analgesia after lumbar spine surgery by pregabalin and celecoxib: a prospective study. Drug Des Dev Ther. 2019;13:2145-2152. doi:10.2147/DDDT.S202410

31.

Dodwell

Latorre

Parisini

, et al. NSAID exposure and risk of nonunion: a meta-analysis of case-control and cohort studies. Calcif Tissue Int. 2010;87(3):193-202. doi:10.1007/s00223-010-9379-7

32.

Weng

Xiao

, et al. Pain management using perioperative administration of parecoxib for total hip arthroplasty: a randomized, double-blind, placebo-controlled trial. [Epub ahead of print]. https://www.painphysicianjournal.com/

33.

Laoruengthana

Rattanaprichavej

Reosanguanwong

Chinwatanawongwan

Chompoonutprapa

Pongpirul

. A randomized controlled trial comparing the efficacies of ketorolac and parecoxib for early pain management after total knee arthroplasty. Knee. 2020;27(6):1708-1714. doi:10.1016/j.knee.2020.10.005

34.

Zhang

Wang

Liu

Sun

. Safety of continuing aspirin therapy during spinal surgery: a systematic review and meta-analysis. Medicine (Baltim). 2017;96(46):e8603. doi:10.1097/MD.0000000000008603

35.

Park

Kang

Shin

Chung

Lee

. Heterotopic ossification following lumbar total disc replacement. Int Orthop. 2011;35(8):1197-1201. doi:10.1007/s00264-010-1095-4

36.

Sheng

Liu

Meng

. Incidence of heterotopic ossification at 10 years after cervical disk replacement: a systematic review and meta-analysis. Spine. 2023;48(13):E203-E215. doi:10.1097/BRS.0000000000004674

37.

Wagner

Kang

Helgeson

. Heterotopic ossification after transforaminal lumbar interbody fusion without bone morphogenetic protein use. Spine J. 2014;14(11):2783-2784. doi:10.1016/j.spinee.2014.06.032

38.

Wen

Chen

Deng

Chen

. Effectiveness of indomethacin in preventing heterotopic ossification: a systematic review and meta-analysis of randomized controlled trials. J Orthop Surg Res. 2024;19(1):589. doi:10.1186/s13018-024-05086-z

39.

Yang

Wen

Chen

, et al. Effectiveness and priority of irradiation and six NSAIDs in prevention heterotopic ossification after total hip arthroplasty: a network meta-analysis of randomized controlled studies. Front Pharmacol. 2025;16:1601349. doi:10.3389/fphar.2025.1601349

40.

Joice

Vasileiadis

Amanatullah

. Non-steroidal anti-inflammatory drugs for heterotopic ossification prophylaxis after total hip arthroplasty: a systematic review and meta-analysis. Bone Joint Lett J. 2018;100-B(7):915-922. doi:10.1302/0301-620X.100B7.BJJ-2017-1467.R1. PMID: 2995419.

41.

Tian

, et al. Incidence of heterotopic ossification after implantation of interspinous process devices. Neurosurg Focus. 2013;35(2):E3. doi:10.3171/2013.3.FOCUS12406

The Impact of Non-Steroidal Anti-inflammatory Drugs on Postoperative Outcomes in Posterior Spinal Surgery: A Systematic Review

Abstract

Study Design

Objective

Methods

Results

Conclusions

Keywords

Introduction

Methods

Search Strategy

Eligibility Criteria

Study Selection

Data Extraction

Risk of Bias Assessment

Data Synthesis

Results

Grouped Outcomes

Opioid Consumption and Pain Control

Postoperative Complications & Healing

Bleeding and Drainage (Table 3)

Discussion

Summary

Fusion vs Non-fusion Procedure Distinction

Pain and Opioid Consumption

Postoperative Complications

Blood Loss & Drainage

Heterotopic Ossification

Strengths and Limitations

Future Directions

Disclosures

Conclusion

Footnotes

Acknowledgements

ORCID iDs

Ethical Considerations

Funding

Declaration of Conflicting Interests

Data Availability Statement

Appendix

References