Enzymatic chemonucleolysis for lumbar disc herniation—an assessment of historical and contemporary efficacy and safety: a systematic review and meta-analysis

Lumbar disc herniation (LDH) is often managed surgically. Enzymatic chemonucleolysis emerged as a non-surgical alternative. This systematic review and meta-analysis aims to assess the efficacy and safety of chemonucleolytic enzymes for LDH. The primary objective is to evaluate efficacy through “treatment success” (i.e., pain reduction) and severe adverse events (SAEs) rates. Additionally, differences in efficacy and safety trends among chemonucleolytic enzymes are explored. Following our PROSPERO registered protocol (CRD42023451546) and PRISMA guidelines, a systematic search of PubMed and Web of Science databases was conducted up to July 18, 2023. Inclusion criteria involved human LDH treatment with enzymatic chemonucleolysis reagents, assessing pain alleviation, imaging changes, and reporting on SAEs, with focus on allergic reactions. Quality assessment employed the Cochrane Source of Bias and MINORS tools. Meta-analysis utilized odds ratios (OR) with 95% confidence intervals (CI). Among 62 included studies (12,368 patients), chemonucleolysis demonstrated an 79% treatment success rate and significantly outperformed placebo controls (OR 3.35, 95% CI 2.41–4.65) and scored similar to surgical interventions (OR 0.65, 95% CI 0.20–2.10). SAEs occurred in 1.4% of cases, with slightly higher rates in chymopapain cohorts. No significant differences in “proceeding to surgery” rates were observed between chemonucleolysis and control cohorts. Limitations include dated and heterogeneous studies, emphasizing the need for higher-quality trials. Further optimization through careful patient selection and advances in therapy implementation may further enhance outcomes. The observed benefits call for wider clinical exploration and adoption. No funding was received for this review.

www.nature.com/scientificreports/tissue may exert physical compression or inflammation on adjacent nerve roots [2][3][4][5] , thereby inciting symptoms such as debilitating low back pain (LBP) and radiculopathy, with potential repercussions for the patient's overall quality of life (Fig. 1A, B).The conventional management of LDH typically entails surgical interventions, often pursued following unsuccessful conservative treatments 3 .Nonetheless, surgical treatments for LDH may be accompanied by unsatisfactory outcomes in up to 15% of cases while exposing patients to not negligible risks, such as neurovascular injury and chronic post-operative LBP 6 .
In the historical context, chemonucleolysis emerged as a compelling alternative to surgical procedures, aiming to enzymatically degrade herniated disc material and assuage associated symptoms 7 .The inception of enzymatic chemonucleolysis can be attributed to chymopapain (Fig. 1C), which was introduced by Smith and Schwartz in 1964 7 as a less invasive technique for treating LDH.This general proteolytic enzyme demonstrated promising outcomes by mitigating complications in comparison to conventional discectomy [8][9][10] .Despite its initial success, safety concerns surfaced, encompassing adverse reactions such as anaphylaxis and nerve root disturbances 11 .Notwithstanding, the Food and Drug Administration (FDA) retained its approval; however, the primary manufacturer terminated its production for unspecified reasons, rendering it inactive in clinical applications 12 .
Contemporary research has borne witness to a potential renaissance in chemonucleolysis, with condoliase, a chondroitin sulfate ABC endolyase (also known as chondroitinase ABC), emerging as a novel and potentially more targeted alternative.Possessing substrate-specificity for chondroitin sulfate and hyaluronic acid, condoliase selectively degrades proteoglycan-rich tissues while preserving surrounding non-proteoglycan structures 13,14 (Fig. 1C).Approved in Japan since 2018, condoliase has undergone extensive clinical trials affirming its safety and efficacy, positioning it as a prospective advancement in LDH treatment 15,16 .Alternatively, collagenase-based therapies were also explored in the 1990s but have recently garnered some new interest as a re-explored alternative LDH-targeted chemonucleolysis therapies 17 .The evolutionary trajectory of chemonucleolysis underscores the imperative to reassess its efficacy and safety, particularly within the realm of LDH treatment.
In navigating the comparative landscape of chemonucleolysis, this systematic review and meta-analysis endeavor to evaluate the general potential and safety of chemonucleolytic treatments as an alternative, nonsurgical LDH therapy, and specifically, to discern the clinical efficacy of distinct enzymatic products.The

Study selection
Article hits underwent automatic duplicate screening using the CADIMA software 21 .Subsequently, three researchers (JS, ST, and LA) independently screened the titles and abstracts of the identified articles against predetermined inclusion criteria.Articles suggested for inclusion by at least one of the three reviewers proceeded to the second round of full-text screening.In this stage, identical researchers conducted a comprehensive assessment to confirm adherence to the previously stated inclusion and exclusion criteria.Additionally, papers meeting the criteria were considered for inclusion in the meta-analysis if they incorporated a control group treated with (micro)discectomy, placebo or sham injection, or general conservative interventions.The article screening workflow has been reported in a PRISMA flow diagram (Fig. 2).

Risk of bias
The Cochrane Source of Bias tool 23 was employed to assess the quality of randomized controlled trials, while the Methodological Index for Non-Randomized Studies (MINORS) scheme 24 was utilized to assess the risk of bias in non-randomized clinical trials.To avoid imprecision, the included papers were rated independently by one reviewer, confirmed by a second reviewer, and eventually validated by a third reviewer.

Statistical analysis
Meta-analysis was performed using odds ratios (OR) with 95% confidence intervals (Cl) to describe dichotomous variables.The level of significance (p) was set at 0.05.Heterogeneity among comparisons was calculated according to the I 2 test and was rated as "low" (I 2 ≤ 25%), "moderate" (I 2 = 26-74%), or "high" (I 2 ≥ 75%).When I 2 was ≤ 50%, a fixed-effect model was employed for analysis, whereas a random effect model was used in case of statistically significant values ≥ 50%.Pooled estimates were calculated with the Mantel-Haenszel model for treatment success rate (vs.discectomy) and rates of proceeding to surgery, while inverse variance was used for treatment success rate (vs.sham), SAE rates, and all the outcomes investigated in single-arm studies included for meta-analysis.Due to the presence of < 10 articles per each investigated outcome in comparative studies, publication bias was not evaluated.Meta-analysis was performed using the Review Manager software (v. 5.

Literature screening
The initial literature search yielded a total of 2092 articles, resulting in 1816 articles for screening following duplicate removal (Fig. 2).Then, 1714 studies were excluded through title and abstract screening, and 6 reports could not be found, with 96 studies eventually considered for full text screening.Out of these studies, 44 were excluded (reviews, n = 4; case reports and case series including < 10 patients, n = 6; inappropriate outcomes, n = 14; inappropriate interventions, n = 7; inappropriate comparator, n = 1; inappropriate study population, n = 12).Furthermore, 10 additional papers were identified through hand citation searching and screened.Finally, 62 papers met the inclusion criteria (Table 1).

Risk of bias
The MINORS tool was employed to assess the quality of evidence of included nonrandomized clinical trials, with an average score of 8/16 for noncomparative studies and 16/24 for comparative studies, indicating a serious risk of bias.Likewise, according to the Cochrane Source of Bias tool, the risk of bias in included randomized controlled trials was also significant, with an average score of 18/26 (Supplemental item 5).
Interestingly, the prevalence of allergic reactions varied among different chemonucleolytic agents.Condoliasetreated cohorts exhibited a slightly higher incidence of allergic reactions, such as rashes, affecting 4.1% of patients, in contrast to 2.4% in chymopapain-treated cohorts and 0.0% in collagenase-treated cohorts (Table 3, supplementary item 11).Notably, it is important to highlight the absence of detailed specifications regarding preventative allergy suppressant medication in the condoliase group, while most studies in the chymopapain cohorts clearly implemented such preventative measures.

Discussion
Chemonucleolysis has a historical context rooted in the pursuit of alternatives to conventional surgical interventions for LDH, with chymopapain being an early but subsequently discontinued option due to safety concerns 11 .Our findings shed light on the evolving landscape of chemonucleolysis, encompassing newer agents like condoliase and re-explored interest in collagenase-based therapies.Here our meta-analysis showed that chemonucleolysis significantly outperforms placebo in terms of treatment success, demonstrating its efficacy in alleviating LDH-associated radiculopathy and disability.The comparable rates of treatment success between chemonucleolysis and surgical intervention suggests that chemonucleolysis may be an effective standalone intervention that can be considered as an alternative to surgical interventions.
Concerns about SAEs (particularly anaphylactic shock) associated with chemonucleolysis were addressed in our analysis.While chymopapain exhibited a higher rate of severe anaphylactic reactions and infectious SAEs, condoliase and collagenase treatments demonstrated safer profiles, with no reported anaphylactic shocks and lower overall SAE rates.No prophylactic allergy treatments are provided with condoliase administration in Japan and none of the included studies provided detailed specifications regarding allergy suppressant medication.The potential benefit of such regimen is a possible avenue of research to potentially further enhance its safety profile.Condoliase, with its substrate-specificity for chondroitin sulfate and hyaluronic acid, presents a more targeted Vol:.( 1234567890 enzymatic alternative, potentially minimizing off-target effects observed with earlier chemonucleolytic agents.The exceptional specificity of condoliase for chondroitin sulfate, renders it an exceptionally selective agent within the intervertebral disc space 13,14 .This specificity stands in contrast to chymopapain or collagenase, as illustrated in Fig. 1C, emphasizing the expected superior safety profile associated with condoliase.Furthermore, condoliase may have additional implications, as it may simultaneously aid in nerve and spinal cord repair.Chondroitin sulfate is produced in response to damage in nerve tissue and hinders nerve growth and axon proliferation.Notably, chondroitinase ABC, a product of condoliase, has demonstrated efficacy in promoting functional recovery following spinal cord injury 86 .This dual capability of selectively targeting LDH-related pathology while potentially supporting neural repair underlines its potential therapeutic scope.Despite promising observations, it is imperative to acknowledge the limitations of the current body of evidence, warranting a call for further comprehensive studies.The controlled studies included in this review, while providing valuable insights, exhibit certain drawbacks.The majority, especially those pertaining to chymopapain, are dated, with a considerable proportion conducted before the 2000s.Furthermore, a prevalent retrospective nature, relatively small cohort sizes, and high-risk of bias scores contribute to the overall limitations.Additionally, the heterogeneity in outcome measures and patient indications across studies, particularly in comparisons involving various chemonucleolytic agents, further underscores the need for caution in drawing definitive conclusions.To advance our understanding of the potential of chemonucleolytic enzymes, future Figure 3. Meta-analysis forest plots depicting the assessment of (in order) the rate of treatment success of chemonucleolysis treated patients compared to surgical discectomy and placebo-treated patients, followed by plots demonstrating the rate of proceeding to surgery compared to surgical discectomy and placebo-treated patients.
Vol.:(0123456789)  22 , and type of LDH have been suggested to enhance treatment outcomes 13,15,16,69 .Other aspects such as the type of contrast agent used, have also been shown to potentially influence enzymatic activity 87 .In addition, there is very little literature reporting long-term imaging evaluations of discs treated with chemonucleolysis (Supplemental item 4), and long-term clinical outcomes and more objective assessments remain largely undetermined 88 .Therefore, future work may consider revising the enzymatic destruction of herniated disc tissue through supplementation with regenerative treatments 16,89,90 e.g., co-injection of cells 91,92 , biomaterials 93,94 , extracellular vesicles 95,96 , or plateletrich plasma (PRP) 97 to support the restoration of the disc over time, particularly regarding younger patients 16,79 .For example, previous animal studies have shown the capacity of PRP products to reverse condoliase-mediated disc deterioration 98 .Here, longer follow-ups are critical to fully grasp the impact of the enzymatic digestion of disc tissue on long-term spinal health 16 .Alternative chemonucleolysis methods, not involving enzymatic Table 3. Tabular overview of reported adverse events and the rate of patients proceeding to surgery.Values are given as the number of events (n) within the patient population (N) and the corresponding percentage (%).Cohort summaries (cumulative totals) are in bold.*Study involving a group of collagenase and a group of chymopapain injections.Abbreviations: ns-not specified, SAE-Severe adverse events.www.nature.com/scientificreports/digestion, such as ethanol 99 or ozone 100 are also being explored and future reviews should seek to compare efficacy and safety outcomes of these approaches to enzymatic chemonucleolysis-treated LDH.

Conclusion
Our comprehensive analysis underscores the evolving landscape of chemonucleolysis as a viable non-surgical intervention for LDH.With newer agents like condoliase exhibiting enhanced safety profiles and promising efficacy, the reevaluation of chemonucleolytic enzymes offers a valuable therapeutic avenue.Further research, particularly large randomized controlled trials, is imperative to solidify the evidence base and refine the clinical application of these enzymatic interventions, fostering their broader adoption in LDH management.We advocate for wider adoption and thorough evaluation of these techniques in clinical settings, particularly in Europe and the USA, to validate and refine their potential and determine the place of chemonucleolysis in the clinical toolbox to mend LDH.

Identification of studies via databases and registers Identification of studies via other methods Identification Screening Included
General study characteristics extracted included: first author name, year of publication, country, funding source (if any), study design, sample size, age, sex, follow-up range (including individual timepoints, minimum, maximum, and mean, when reported), indication for chemonucleolysis, previous and concurrent treatments, technical description of the intervention (i.e., experimental product, enzyme source, injected volume and concentration, injection site, needle gauge, and number of discs injected).In comparative studies, characteristics of the control groups and respective treatments (i.e., sham injections or surgical discectomy) were recorded.Complications, success/failure rates, likelihood of undergoing further surgery, and SAEs were assessed in all included patients.Treatment success was defined as a good to excellent outcome as subjectively reported by included patients, whereas worsening or no improvement results were considered a treatment failure.An SAE was defined as a post-injection complication leading to or risking death, severely prolonging hospitalization, or causing lasting disability.Additionally, specific complications of anaphylactic shock, allergic reactions and infectious complications were separately recorded.Imaging changes (e.g., Pfirrmann classification 22 , T2-intensities at magnetic resonance imaging, disc height index, or LDH-volume/size) were also acquired.Data presented graphically were extracted using WebPlotDigitizer version 4.7 (https:// autom eris.io/ WebPl otDig itizer, by A. Rohatgi) to approximate reported scores.Imaging outcomes were assessed at baseline and 3-, 6-, 12-, and 24-months post-transplantation, with average values, variability, and population size documented for each time point.

Table 2 .
Tabular overview of treatment success as indicated by improvement in pain outcomes.Values are given as the number of events (n) within the patient population (N) and the corresponding percentage (%).Cohort summaries (cumulative totals) are in bold.*Study involving a group of collagenase and a group of chymopapain-treated patients , #self-declared improvement, †as defined by Modified MacNab criteria, ‡Based on > 50% pain score (leg or back) improvement compared to baseline, ˆBased on > 20 mm VAS change, ‹Surgery-cohort applied in both collagenase and chymopapain from Brown et al. (1989).Abbreviations: (L)-Low dose, (M)-Medium dose, (H)-High dose, ns-not specified.
research endeavors should prioritize the design and implementation of higher-quality trials.Large randomized controlled trials, characterized by rigorous methodologies and standardized reporting, hold the potential to offer more robust and generalizable insights into the efficacy and safety of these enzymatic chemonucleolytic interventions for LDH.Moreover, further optimizations of chemonucleolysis techniques and study designs are warranted.For example, optimal patient selection criteria, considering factors like age, Pfirrmann grade ContinuedVol:.(1234567890) Scientific Reports | (2024) 14:12846 | https://doi.org/10.1038/s41598-024-62792-8www.nature.com/scientificreports/