Lumbar disk herniation refers to a condition in which the degeneration of intervertebral disks within the lumbar spine ruptures the lumbar spine’s annulus fibrosus and some part or all of nucleus pulposus is herniated, exerting pressure on the nerve root of the dura mater that accompanies a series of symptoms [1]. According to the National Health Insurance Statistical Yearbook 2020, issued by the Health Insurance Review and Assessment Service [2], lumbar disk herniation (M51 Other intervertebral disk disorders) ranked 4th in inpatient frequent disease and 18th in outpatient frequent disease within Korean Medicine. This prominence can be attributed to the fact that Koreans frequently visit Korean Medicine clinics and hospitals for this condition. Diagnostic procedures, including magnetic resonance imaging (MRI) and computed tomography (CT), play a pivotal role in confirming the presence of lumbar disk herniation [1]. The major symptoms of this condition are local pain around the affected area, along with radiating pain in the lower extremities due to pressure exerted by the herniated nucleus pulposus on the nerve roots. In more severe cases, this condition may escalate to cauda equina syndrome [3].

Lumbar disk herniation usually shows improvement through conservative treatment, and many studies underscore the current preference for conservative treatment over surgical therapies [4,5]. Among these conservative treatments, acupuncture treatment has ascended to higher levels of recommendation within international clinical guidelines [6,7]. Acupotomy, a new acupuncture therapy developed by professor Zhu Hanzhang of China in 1976, amalgamates surgical scalpel techniques with acupuncture. This approach has proven effective for old and chronic diseases. In China, the study on acupotomy is very active, with 36 randomized controlled trials related to treatment effects of acupotomy on herniation of intervertebral disk of lumbar spine from 2006 to 2016 [8]. Various studies on acupotomy have been performed in Korea, with a significant portion of these studies reporting clinical results in the context of lumbar disk herniation [9-11]. However, many cases are case reports with limited subjects or comparisons with dry acupuncture. Notably absent are studies delving into the dimensions of treatment effects concerning the frequency or duration of acupotomy sessions. Consequently, a need persists for studies that elucidate the requisite number and duration of acupotomy treatment sessions through systematic exploration.

The authors aim to present the results derived from the statistical analysis of 2 clinical studies on acupotomy for lumbar disk herniation. The first study encompassed 46 patients selected based on predefined inclusion and exclusion criteria [12]. These individuals underwent 4 acupotomy sessions for 2 weeks from April 9, 2018, to December 8, 2018, at Daejeon Korean Medicine Hospital of Daejeon University. The second study comprised 15 patients selected in accordance with inclusion and exclusion criteria, who underwent 6 acupotomy sessions for 3 weeks at Daejeon Korean Medicine Hospital of Daejeon University from May 13, 2021, to November 22, 2021.

1. Participants

The study participants consisted of 46 patients (Group A) who completed the study requirements from the 48 patients selected based on the predefined inclusion and exclusion criteria at Daejeon Korean Medicine Hospital of Daejeon University from April 9, 2018, to December 8, 2018. Additionally, 15 patients (Group B) who met the inclusion and exclusion criteria set forth by Daejeon Korean Medicine Hospital of Daejeon University were included in the study, having undergone selection from May 13, 2021, to November 22, 2021. The specific inclusion and exclusion criteria for each study are detailed below (Tables 1, 2).

Table 1 . Inclusion and exclusion criteria in Group A.

Inclusion criteria

Exclusion criteria

Subjects who meet all of the following criteria can participate in this study: Individuals aged between 19 and 80 years. Individuals who have been diagnosed with herniation of the lumbar intervertebral disk through CT or MRI (within the last 5 years). Individuals experiencing 1 or more of the following symptoms: Lumbar pain, lower extremity pain, lower extremity dysesthesia, lower extremity muscle weakness. Individuals whose pain level at screening is VAS 30 mm or higher. Individuals with no difficulties with language, expression, and concentration. Individuals who can be followed up during the study’s clinical period. Individuals who have voluntarily provided informed consent to participate in clinical research.

Any subject meeting any of the following criteria cannot participate in this study: Those displaying acupuncture hypersensitivity, metal allergies, severe atopic dermatitis, keloid skin, or other forms of skin hypersensitivity. Those requiring surgical treatment due to neurological symptoms such as cauda equine syndrome or sensory/motor paralysis. Those needing hospitalization or experiencing difficulties in reaching the hospital due to severe pain or symptoms hindering mobility. Those with a history of spinal surgery involving metal internal fixation and spinal fusion. Those experiencing musculoskeletal pain in other body parts more severe than the lumbar or lower extremities. Those taking narcotic analgesics, anticonvulsants, corticosteroids, etc. for pain management. Those who have participated in another clinical trial involving interventions within the past 3 months. Pregnant or lactating female or female of childbearing potential who plan to become pregnant during the study period. Those currently receiving or needing active treatment for significant neurological or psychological medical history, alcohol or substance abuse history, or serious underlying diseases. Those with hemorrhagic disease, cardiovascular disease, etc. that can affect hemostasis, including taking anticoagulants or antiplatelet drugs Others who are deemed unsuitable by the person in charge of clinical research.

CT, computed tomography; MRI, magnetic resonance imaging; VAS, visual analog scale..

Table 2 . Inclusion and exclusion criteria in Group B.

Inclusion criteria

Exclusion criteria

Subjects who meet all of the following criteria can participate in this study: Adult males and females aged between 19 and 85 years. Those who have been diagnosed with lumbar intervertebral disk disorder (injury), spinal stenosis, and spondylosis with neuromyopathy through imaging examinations (MRI or CT) within 6 months prior to participating in the clinical trial. (However, if a patient is suspected of having lumbar sacral neuropathy but lacks imaging records during the screening test, an MRI or CT will be conducted to check if the diagnostic criteria are met.) Patients presenting symptoms related to lumbosacral radiculopathy, such as radiculopathy, muscle weakness, paresthesia, etc., or who have been diagnosed with lumbosacral neuropathy through physical examination. Patients with NRS-4 or higher pain/discomfort due to lumbar sacral neuropathy. Patients capable of reading and comprehending the symptom questionnaire and able to provide meaningful answers. Patients who have voluntarily granted written consent, agreeing to clinical plan and follow-up, as approved by the Clinical Research Review Committee.

Any subject meeting any of the following criteria cannot participate in this study: Patients with a history of spinal surgery like lumbar sacral intrametallic fixation or spinal fusion, or who have previously undergone spinal surgery but continue to experience associated pain. Patients exhibiting motor paralysis and neurological symptoms expected to have a challenging recovery, particularly those with cauda equine syndrome or those deemed to necessitate surgical interventionas determined by the researcher. Patients undergoing active drug treatment involving strong opioids for pain management. Patients who received Korean Medicine treatment within 2 weeks before the baseline of the clinical study. (However, if, as determined by the research personnel, a new type of neuropathic symptom differing from existing pain is evident despite recent treatment at another institution, participation in this clinical trial may be permitted.) Those displaying acupuncture hypersensitivity, metal allergies, severe atopic dermatitis, keloid skin, or other forms of skin hypersensitivity. Patients with hemophilia. Patients taking drugs that may cause hemostasis, such as anticoagulants, antiplatelet drugs, and aspirin, and are deemed by the clinical investigator to be unable to discontinue such drugs during the clinical trial period. Those who participated in another clinical study within 30 days prior to screening for this clinical research and received investigational drugs (including placebo). Patients with psychotic disorders, alcoholism, or drug addiction. Pregnant or lactating female or female of childbearing potential who do not intend to use contraception through the clinical trial. Others who are deemed unsuitable for participation in clinical trials by the designated research personnel.

MRI, magnetic resonance imaging; CT, computed tomography; NRS, numeric rating scale..

2. Method

The 2 clinical studies were designed with a parallel design involving 2 groups for comparison. These studies adhered to a control group setup and were carried out with assessors who were blinded to the group assignments. Allocation of participants to groups was achieved using SAS® version 9.4 (SAS Institute) by a statistician unrelated to this study. Furthermore, the produced random allocation table was maintained in confidentiality by a statistician independent of this study. The design of these 2 studies was meticulously crafted to mitigate bias, and this was accomplished by ensuring that those administering acupuncture treatments were not involved in the evaluation of treatment effectiveness.

Both studies were approved by the Institutional Review Board (IRB) of the Clinical Trial Center at Daejeon Korean Medicine Hospital of Daejeon University (DJDSKH-18-BM-02 [12], DJDSKH-21-BM-06). These studies were also duly registered in the clinical research information system (CRIS) overseen by the Korea Disease Control and Prevention Agency (CRIS-KCT0002824 [12], KCT000 6043). Furthermore, the current study comparing the 2 aforementioned studies was similarly granted approval by the IRB of Clinical Trial Center, at Daejeon Korean Medicine Hospital of Daejeon University (DJDSKH-21-E-35).

Patients previously diagnosed with herniated intervertebral disk of lumbar spine through spinal CT or MRI were requested to their medical certificate. Patients without prior diagnosis underwent lumbar spine CT scans for herniation diagnosis. All subjects underwent a comprehensive blood examination including complete blood cell count (CBC), aspartate aminotransferase (AST), alanine aminotransferase (ALT), erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), prothrombin time (PT), and activated partial thromboplastin time (aPTT). Additionally, urine human chorinic gonadotropin (hCG) testing was conducted for females of reproductive age.

Subjects in Group A received 4 acupotomy sessions in 2 weeks and attended 2 additional follow-up appointments. Evaluation was performed on the first treatment day (baseline) and at 2, 4, and 6 weeks post-baseline. Conversely, subjects in Group B underwent 6 acupotomy sessions in 3 weeks, followed by 1 follow-up visit. Evaluation was performed on the first treatment day (baseline) and then at 2, 3, and 7 weeks from the baseline. In cases where treatment and evaluation appointments coincided, the evaluation took precedence and occurred before the treatment session. This was done to mitigate any potential bias resulting from the unique characteristics of acupotomy, which could influence assessments immediately after the treatment session (Figs. 1, 2).

Figure 1. Group A study protocol flow. CT, computed tomography.

Figure 2. Group B study protocol flow. CT, computed tomography.

All procedures were administered by the same Korean Medicine Doctor with clinical experience of more than or equal to 5 years. A scalpel (Dongbang Medical Co., Ltd.) of dimensions 0.75 × 80 mm, encompassing a flat blade attached to the tip of the cone-shaped acupuncture needle body, was used. Prior to administration, the Korean Medicine Doctor identified the disease area based on radiographic findings. Treatment points were then selected around the area of discomfort, with the EX-B2 point corresponding to the affected intervertebral disk serving as the main administration point. These procedures took place while the patient was in prone position (Fig. 3). Throughout the duration of each study, all participating subjects were prohibited from undergoing any other treatments apart from those specified within the scope of the study.

Figure 3. Application of acupotomy. (A) Skin disinfection. (B) Needle insertion. (C) Needle manipulation. (D) Disinfection and band-aid application.

3) Assessment

In Group A, pain evaluation involved the examination of changes in patients’ subjective pain degree using changes in visual analog scale (VAS) [13] values at baseline (week 0), week 2, week 4, and week 6. Meanwhile, for Group B, pain evaluation was based on changes in patients’ subjective pain degree measured through changes in numeric rating scale (NRS) [14] values at baseline (week 0), week 2, week 3, and week 7. It is important to note that higher scores on both the VAS and NRS indicate greater severity of pain.

The evaluation of disability in Group A was assessed by changes in scores in the Korean version of the Roland Morris disability questionnaire (RMDQ) at baseline (week 0), week 2, week 4, and week 6. Similarly, the evaluation of disability in Group B was assessed by changes in scores in the Korean version of RMDQ at baseline (week 0), week 2, week 3, and week 7. The RMDQ, designed by Roland M and Morris R in 1983, comprises 24 questions aimed at capturing potential daily disabilities arising from lower back pain. Patients respond with a “yes” or “no” to each question. The advantage of this questionnaire is its simplicity and rapid completion by patients [15]. Higher RMDQ scores signify more severe disability.

The evaluation of quality of life in Group A was assessed by changes in scores in the Korean version of EuroQol 5-dimention (EQ-5D) at baseline (week 0), week 2, week 4, and week 6. Similarly, the evaluation of quality of life in Group B was assessed by changes in scores in the Korean version of EQ-5D at baseline (week 0), week 2, week 3, and week 7. The EQ-5D is a widely used tool for evaluating quality of life, capturing the patient’s current state across 5 domains: mobility, self-care, usual activities, pain/discomfort, and anxiety/depression [16,17]. The lower the EQ-5D score the lower the quality of life.

3. Statistical analysis

The collected data was analyzed using SAS® version 9.4. Prior to conducting the analysis, mean and standard error were calculated for each group’s demographic characteristics using Student’s t-test for continuous data, aiming to discern differences. Similarly, for categorical data, percentile was computed using Fisher’s exact test prior to analysis. To establish the validity of evaluation variables for each group, the average amount of changes in VAS (evaluation NRS per group), RMDQ, and EQ-5D at baseline, week 2, week 3 or 4, and week 6 or 7 were used. The analysis was conducted utilizing analysis of covariance (ANCOVA), where RMDQ and EQ-5D value were integrated as covariates and each group was designated as the fixed factor. Furthermore, the differences in evaluation variables before and after treatment within each group were investigated using paired t-tests. For assessing variations in the trend of changes across visits, a repeated measures analysis of variance (RM ANOVA) was performed. To account for the assumption of sphericity, the Greenhouse–Geisser correction was used.

Furthermore, due to the great difference in the number of subjects between the 2 groups, comprising 46 and 15 participants, respectively, the aforementioned analysis was repeated using propensity score matching (PSM). This process involved matching 10 samples from each group, with baseline VAS (NRS), RMDQ, and EQ-5D values, and sex serving as the correction variables.

All statistical analyses were executed as two-sided test in accordance with the rules, maintaining a significance level of 5 %.

1. Demographic characteristics

During the screening visit (week 0), an assessment of sex, age, height, body weight, and body mass index was examined. Of the 46 subjects in Group A, 21 were male and 25 were female, with an average age of 54.87. Of the 15 subjects in Group B, 3 were male and 12 were female, with an average age of 59.80. Analysis of the differences in demographic characteristics between the 2 groups did not reveal any significant differences (Table 3).

Table 3 . Baseline demographic characteristics.

Characteristic	Group A (n = 46)	Group B (n = 15)	p-value
Sex (M/F)*	21 (45.65)/25 (54.35)	3 (20.00)/12 (80.00)	0.1272
Age (y)†	54.87 (1.63)	59.80 (2.77)	0.1296
Interval estimate	51.63, 58.11	54.06, 65.54
Height (cm)†	162.28 (1.20)	158.02 (2.12)	0.0785
Interval estimate	159.90, 164.66	153.63, 162.41
Weight (kg)†	65.45 (1.61)	61.29 (2.02)	0.1716
Interval estimate	62.25, 68.65	57.11, 65.48
BMI (kg/m2)†	24.75 (0.40)	24.47 (0.42)	0.6260
Interval estimate	23.95, 25.55	23.61, 25.33

Data expressed as number (%) or mean (standard error)..

M, male; F, female..

*Fisher’s exact test, †Student’s t-test..

As for demographic characteristics after PSM, the revised composition included 1 male and 9 females in Group A, with an average age of 58.00. For Group B, this adjusted distribution encompassed 2 males and 8 females, with an average age of 61.00. There was no significant difference in the analysis of the differences in demographic characteristics between the 2 groups (Table 4).

Table 4 . Baseline demographic characteristics after propensity score matching.

Characteristic	Group A (n = 10)	Group B (n = 10)	p-value
Sex (M/F)*	1 (10.0)/9 (90.0)	2 (20.0)/8 (80.0)	0.9999
Age (y)†	58.00 (2.52)	61.00 (4.13)	0.5221
Interval estimate	52.58, 63.42	52.13, 69.87
Height (cm)†	158.65 (2.34)	158.66 (2.90)	0.9977
Interval estimate	153.96, 163.34	152.33, 164.99
Weight (kg)†	59.88 (2.19)	60.82 (2.95)	0.7931
Interval estimate	54.87, 64.89	54.60, 67.04
BMI (kg/m2)†	23.81 (0.85)	24.07 (0.57)	0.7913
Interval estimate	21.99, 25.63	22.85, 25.29

Data expressed as number (%) or mean (standard error)..

M, male; F, female..

*Fisher’s exact test, †Student’s t-test..

2. Efficacy assessment

Upon assessing the average amount of change in the primary variable’s baseline value within Groups A and B prior to any interventions, the analysis revealed a significant difference in the baseline VAS (NRS) values between the 2 groups. Conversely, the baseline values of RMDQ and EQ-5D demonstrated no significant difference (p < 0.05) (Table 5).

Table 5 . Analysis of homogeneity.

Variable	Group A			Group B		Mean difference	p-value
	Mean	SE		Mean	SE
VAS (NRS)	5.64	0.23		6.93	0.46	−1.29	0.0071*
RMDQ	7.00	0.59		7.60	1.85	−0.60	0.7541
EQ-5D	0.74	0.02		0.67	0.05	0.07	0.1218

SE, standard error; VAS, visual analog scale; NRS, numeric rating scale; RMDQ, Roland Morris disability questionnaire; EQ-5D, EuroQol 5-dimention..

*p < 0.05..

Subsequent analysis of the average amount of change in the baseline values after PSM indicated that there was no significant difference in the baseline VAS (NRS), RMDQ, and EQ-5D values between the 2 groups (Table 6).

Table 6 . Analysis of homogeneity after propensity score matching.

Variable	Group A			Group B		Mean difference	p-value
	Mean	SE		Mean	SE
VAS (NRS)	6.23	0.56		6.20	0.49	0.03	0.9665
RMDQ	6.30	1.28		6.30	1.78	0.00	0.9999
EQ-5D	0.710	0.044		0.710	0.040	0.00	0.9207

SE, standard error; VAS, visual analog scale; NRS, numeric rating scale; RMDQ, Roland Morris disability questionnaire; EQ-5D, EuroQol 5-dimention..

In Group A, the VAS scores showed a significant decrease continuously from baseline to weeks 2, 4, and 6. Similarly, in Group B, the NRS scores showed significant decrease continuously from baseline to weeks 2 and 3, but did not show significant difference in comparison between the 2 groups (p < 0.05) (Table 7). Furthermore, while the VAS (NRS) scores showed significant changes within each respective group, there was no significant interaction effect between the 2 groups (p < 0.05) (Table 8).

Table 7 . Analysis of pain assessments.

VAS (NRS)	Group A	Group B	Mean difference†	p-value†
Baseline	5.64 (5.19, 6.09)	6.93 (5.99, 7.88)
Week 2	4.49 (3.88, 5.11)	5.67 (4.61, 6.73)	−0.30 (−1.41, 0.82)	0.5936
Difference‡	−1.14 (−1.74, −0.55)	−1.27 (−1.84, −0.69)
p-value‡	0.0003*	0.0003*
Week 3 (B) or 4 (A)	4.23 (3.56, 4.91)	4.73 (3.86, 5.61)	0.20 (−1.05, 1.44)	0.7499
Difference‡	−1.40 (−2.05, −0.75)	−2.20 (−3.12, −1.28)
p-value‡	< 0.0001*	0.0001*
Week 6 (A) or 7 (B)	3.93 (3.23, 4.63)	5.47 (3.93, 7.00)	−0.86 (−2.34, 0.62)	0.2517
Difference‡	−1.71 (−2.40, −1.01)	−1.47 (−2.97, 0.04)
p-value‡	< 0.0001*	0.0557

VAS, visual analog scale; NRS, numeric rating scale..

†Least squares mean difference and p-value were analyzed using the ANCOVA, with the baseline scores as covariates and the groups as the fixed factor. ‡Mean difference and p-value were analyzed using the paired t-test for the baseline value and each point value (week 2, week 3 or 4, week 6 or 7)..

*p < 0.05..

Table 8 . Pain assessments between the groups over time.

Source	F-value	p-value	Contrast	p-value
Group	5.00	0.0292*	Week 2	0.8204
Week†	15.46	< 0.0001*	Week 3 or 4	0.2018
Group × week†	1.18	0.3153	Week 6 or 7	0.7416

†Using the Greenhouse–Geisser epsilon correction..

*p < 0.05..

Post-PSM analysis yielded the following outcomes: VAS scores of Group A showed significant decrease continuously from baseline to weeks 4 and 6. In parallel, NRS scores of Group B showed significant decrease continuously from baseline to weeks 2 and 3. Upon comparing the 2 groups, there was significant difference in week 2 (p < 0.05) (Table 9). Furthermore, the VAS (NRS) scores showed significant changes over time (week), yet no significant interaction effect between Group A and Group B (p < 0.05) (Table 10).

Table 9 . Analysis of pain assessment after propensity score matching.

VAS (NRS)	Group A	Group B	Mean difference†	p-value†
Baseline	6.23 (5.03, 7.43)	6.20 (5.14, 7.26)
Week 2	5.77 (4.26, 7.28)	4.70 (3.69, 5.71)	1.04 (0.05, 2.04)	0.0412*
Difference‡	−0.46 (−1.07, 0.15)	−1.50 (−2.34, −0.66)
p-value‡	0.1232	0.0030*
Week 3 (B) or 4 (A)	5.38 (3.62, 7.14)	4.30 (3.18, 5.42)	1.05 (−0.24, 2.24)	0.1039
Difference‡	−0.85 (−1.47, −0.23)	−1.90 (−3.09, −0.71)
p-value‡	0.0131*	0.0056*
Week 6 (A) or 7 (B)	5.26 (3.37, 7.15)	5.10 (2.75, 7.45)	0.13 (−2.15, 2.40)	0.9091
Difference‡	−0.97 (−1.84, −0.10)	−1.10 (−3.32, 1.12)
p-value‡	0.0325*	0.1341

VAS, visual analog scale; NRS, numeric rating scale..

†Least squares mean difference and p-value were analyzed using the ANCOVA, with the baseline scores as covariates and the groups as the fixed factor. ‡Mean difference and p-value were analyzed using the paired t-test for the baseline value and each point value (week 2, week 3 or 4, week 6 or 7)..

*p < 0.05..

Table 10 . Pain assessments between the groups over time after propensity score matching.

Source	F-value	p-value	Contrast	p-value
Group	0.46	0.5060	Week 2	0.0366
Week†	5.27	0.0152*	Week 3 or 4	0.0940
Group × week†	1.22	0.3033	Week 6 or 7	0.9033

†Using the Greenhouse–Geisser epsilon correction..

*p < 0.05..

In Group A, RMDQ scores showed significant decrease continuously from baseline to weeks 2, 4, and 6. Conversely, in Group B, the RMDQ scores did not show significant decrease from baseline. However, there was a significant difference in week 2 and week 6 or 7 when comparing the 2 groups (p < 0.05) (Table 11). Moreover, RMDQ scores showed significant changes across time periods (weeks), though no significant interaction effect was detected between Group A and Group B (p < 0.05) (Table 12).

Table 11 . Analysis of disability assessments.

RMDQ	Group A	Group B	Mean difference†	p-value†
Baseline	7.00 (5.82, 8.18)	7.60 (3.76, 11.44)
Week 2	5.35 (4.34, 6.36)	7.73 (4.08, 11.39)	−2.00 (−3.83, −0.16)	0.0334*
Difference‡	−1.65 (−2.77, −0.53)	0.13 (−1.20, 1.47)
p-value‡	0.0048*	0.8338
Week 3 (B) or 4 (A)	4.72 (3.66, 5.77)	7.07 (3.43, 10.71)	−1.98 (−4.00, 0.03)	0.0532
Difference‡	−2.28 (−3.52, −1.04)	−0.53 (−1.90, 0.84)
p-value‡	0.0006*	0.4179
Week 6 (A) or 7 (B)	4.07 (3.10, 5.03)	7.27 (3.01, 11.52)	−2.82 (−4.93, −0.70)	0.0100*
Difference‡	−2.93 (−4.19, −1.68)	−0.33 (−1.82, 1.16)
p-value‡	< 0.0001*	0.6387

RMDQ, Roland Morris disability questionnaire..

†Least squares mean difference and p-value were analyzed using the ANCOVA, with the baseline scores as covariates and the groups as the fixed factor. ‡Mean difference and p-value were analyzed using the paired t-test for the baseline value and each point value (week 2, week 3 or 4, week 6 or 7)..

*p < 0.05..

Table 12 . Disability assessments between the groups over time.

Source	F-value	p-value	Contrast	p-value
Group	2.97	0.0900	Week 2	0.0913
Week†	4.82	0.0081*	Week 3 or 4	0.1315
Group × week†	2.68	0.0679	Week 6 or 7	0.0297

†Using the Greenhouse–Geisser epsilon correction..

*p < 0.05..

Analysis after PSM showed that RMDQ scores for both groups demonstrated no significant changes when comparing the 2 groups (Table 13). Notably, there were no significant changes in RMDQ, and no significant interaction effect between the 2 groups (Table 14).

Table 13 . Analysis of disability assessment after propensity score matching.

RMDQ	Group A	Group B	Mean difference†	p-value†
Baseline	6.30 (3.56, 9.04)	6.30 (2.48, 10.12)
Week 2	6.20 (3.02, 9.38)	5.90 (2.82, 8.98)	0.30 (−1.61, 2.21)	0.7442
Difference‡	−0.10 (−1.51, 1.31)	−0.40 (−1.99, 1.19)
p-value‡	0.8760	0.5830
Week 3 (B) or 4 (A)	6.70 (3.43, 9.97)	5.70 (1.62, 9.79)	1.00 (−1.18, 3.18)	0.3468
Difference‡	0.40 (−0.78, 1.58)	−0.60 (−2.54, 1.34)
p-value‡	0.4620	0.5025
Week 6 (A) or 7 (B)	5.60 (2.16, 9.04)	6.00 (1.16, 10.84)	−0.40 (−2.47, 1.67)	0.6888
Difference‡	−0.70 (−2.17, 0.77)	−0.30 (−2.09, 1.49)
p-value‡	0.3100	0.7128

RMDQ, Roland Morris disability questionnaire..

†Least squares mean difference and p-value were analyzed using the ANCOVA, with the baseline scores as covariates and the groups as the fixed factor. ‡Mean difference and p-value were analyzed using the paired t-test for the baseline value and each point value (week 2, week 3 or 4, week 6 or 7)..

Table 14 . Disability assessments between the groups over time after propensity score matching.

Source	F-value	p-value	Contrast	p-value
Group	0.01	0.9180	Week 2	0.7530
Week†	0.31	0.7702	Week 3 or 4	0.3326
Group × week†	0.58	0.5943	Week 6 or 7	0.7004

†Using the Greenhouse–Geisser epsilon correction..

Within Group A, EQ-5D showed significant increase continuously from baseline to weeks 2, 4, and 6. Meanwhile, in Group B, the EQ-5D scores showed significant increase in week 3 compared to baseline. However, no significant change was found when comparing the 2 groups (p < 0.05) (Table 15). Furthermore, EQ-5D scores showed significant change across time periods (weeks), but no significant interaction effect was identified between the 2 groups (p < 0.05) (Table 16).

Table 15 . Analysis of quality of life assessments.

EQ-5D	Group A	Group B	Mean difference†	p-value†
Baseline	0.736 (0.699, 0.772)	0.672 (0.575, 0.769)
Week 2	0.775 (0.744, 0.806)	0.700 (0.623, 0.778)	0.042 (−0.013, 0.097)	0.1302
Difference‡	0.034 (0.006, 0.073)	0.029 (−0.030, 0.088)
p-value‡	0.0221*	0.3165
Week 3 (B) or 4 (A)	0.793 (0.763, 0.823)	0.769 (0.734, 0.803)	0.002 (−0.047, 0.050)	0.0947
Difference‡	0.058 (0.025, 0.090)	0.097 (0.018, 0.176)
p-value‡	0.0008*	0.0197*
Week 6 (A) or 7 (B)	0.794 (0.765, 0.823)	0.736 (0.631, 0.840)	0.032 (−0.037, 0.101)	0.3586
Difference‡	0.058 (0.020, 0.097)	0.064 (−0.026, 0.154)
p-value‡	0.0037*	0.1511

EQ-5D, EuroQol 5-dimention..

†Least squares mean difference and p-value were analyzed using the ANCOVA, with the baseline scores as covariates and the groups as the fixed factor. ‡Mean difference and p-value were analyzed using the paired t-test for the baseline value and each point value (week 2, week 3 or 4, week 6 or 7)..

*p < 0.05..

Table 16 . Quality of life assessments between the groups over time.

Source	F-value	p-value	Contrast	p-value
Group	3.67	0.0604	Week 2	0.7451
Week†	8.48	0.0001*	Week 3 or 4	0.2689
Group × week†	0.86	0.4458	Week 6 or 7	0.8973

†Using the Greenhouse–Geisser epsilon correction..

*p < 0.05..

Analysis after PSM showed that in Group A, EQ-5D scores showed a significant increase in week 6 compared to baseline. However, in Group B, the EQ-5D scores did not show significant increase when compared to baseline. Notably, there was no significant difference in the comparison between the 2 groups (p < 0.05) (Table 17). EQ-5D scores showed significant changes over the time periods (weeks), yet no significant interaction effect was observed between the 2 groups (p < 0.05) (Table 18).

Table 17 . Analysis of quality of life assessment after propensity score matching.

EQ-5D	Group A	Group B	Mean difference†	p-value†
Baseline	0.718 (0.624, 0.812)	0.712 (0.626, 0.799)
Week 2	0.725 (0.615, 0.835)	0.734 (0.662, 0.805)	−0.014 (−0.087, 0.059)	0.6961
Difference‡	0.007 (−0.049, 0.063)	0.022 (−0.034, 0.077)
p-value‡	0.7900	0.4046
Week 3(B) or 4(A)	0.754 (0.669, 0.839)	0.778 (0.741, 0.815)	−0.027 (−0.092, 0.038)	0.3917
Difference‡	0.036 (−0.022, 0.094)	0.066 (−0.007, 0.139)
p-value‡	0.1914	0.0725
Week 6(A) or 7(B)	0.778 (0.714, 0.841)	0.784 (0.665, 0.904)	−0.010 (−0.116, 0.096)	0.8430
Difference‡	0.060 (0.001, 0.118)	0.072 (−0.034, 0.178)
p-value‡	0.0476*	0.1579

EQ-5D, EuroQol 5-dimention..

†Least squares mean difference and p-value were analyzed using the ANCOVA, with the baseline scores as covariates and the groups as the fixed factor. ‡Mean difference and p-value were analyzed using the paired t-test for the baseline value and each point value (week 2, week 3 or 4, week 6 or 7)..

*p < 0.05..

Table 18 . Quality of life assessments between the groups over time after propensity score matching.

Source	F-value	p-value	Contrast	p-value
Group	0.03	0.8573	Week 2	0.6787
Week†	3.67	0.0265*	Week 3 or 4	0.4786
Group × week†	0.15	0.8994	Week 6 or 7	0.8205

†Using the Greenhouse–Geisser epsilon correction..

*p < 0.05..

Acupotomy entails the utilization of a needle thicker than the usual acupuncture needle, equipped with a blade at its tip. This technique effectively alleviates localized pain by delivering strong stimulation. It achieves this by addressing chronic adhesions or disturbances in blood circulation around nerves [18].

This study’s objective involved the evaluation of pain reduction, alleviation of disability linked to herniated intervertebral disk of lumbar spine, and the enhancement of quality of life. For this purpose, evaluation metrics such as VAS, NRS, RMDQ, and EQ-5D were used and statistically analyzed. In particular, VAS and NRS, which serve as indices for pain evaluation, were used based on the study by Song et al. [19]. Despite the different units used in pain evaluation by VAS and NRS, the severity of pain is expressed as an absolute value, disregarding the unit. As both values constitute ordinal variables, it was possible to statistically compare them. Additionally, these 2 values were similar across various pain types, severity levels, evaluation time, and repeated evaluation. Their correlation was also notably strong. Therefore, based on the insights from the aforementioned study, a pain evaluation comparison was conducted between Group A and Group B.

In Group A, VAS scores showed significant decrease from the start of the study through week 6. In Group B, NRS scores showed significant decrease from the start of the study until week 3, but not in week 7. In comparison between groups, the differences in pain reduction were not statistically significant. Group A showed a significant decrease in RMDQ values from the start of the study to week 6. Conversely, Group B did not show any significant decrease in RMDQ values. Furthermore, in comparison between groups, the difference in the improvement of disability between the 2 groups showed a significant difference in week 2 and week 6 or 7. Group A showed significant increase in EQ-5D values from the start of the study to week 6, whereas Group B showed significant increase in EQ-5D values from the start of the study to week 3, but not in weeks 2 and 7. However, upon comparing the 2 groups, the difference in the improvement of quality of life was not significant. Analysis on the trend changes over the study period for each group revealed that the 3 indices—VAS (NRS), RMDQ, and EQ-5D—did not show any significant changes.

Interpreting the outcomes of the aforementioned analysis in a straightforward manner proved challenging due to the potential bias from the difference in the sample sizes between the 2 groups. Notably, a significant discrepancy was observed in the baseline VAS (NRS) values. Consequently, to establish the closest semblance between the 2 groups and facilitate a more accurate comparative analysis, PSM was adopted. However, given the limited sample size within Group B, it was determined that correcting the discrepancy via the adjustment of select physical characteristics and baseline values of the evaluation indices would prove more effective than incorporating many correction variables. Therefore, to create more aligned groups, sex and baseline VAS (NRS), RMDQ, and EQ-5D values were used as correction variables.

Analyzing the 2 groups, formed through the utilization of PSM, unveiled that Group A showed a significant decrease in VAS values from the start of the study to weeks 4 and 6. Conversely, in Group B, there was a significant decrease in NRS values from the start of the study to week 3, but not in week 7. It was identified that pain increased during the follow-up on week 4 in Group B, with 6 sessions of acupotomy compared to Group A with 4 sessions of acupotomy. Comparing the 2 groups, there was a significant difference in pain reduction between the 2 groups in week 2, but not in week 3 or 4 and week 6 or 7. This means that, regardless of the number of acupotomy sessions, the severity of pain during the final follow-up was similar in the 2 groups. As the total number of acupotomy sessions might not consistently correlate directly with treatment effectiveness, adequate number of acupotomy sessions should be determined based on the severity of symptoms and patient characteristics. Notably, there was no significant decrease in RMDQ values in either Group A or Group B. Moreover, when comparing the 2 groups, the observed difference in disability improvement was not significant. Notably, the RMDQ values exhibited a fluctuating pattern of increase and decrease across the 2 groups, making it challenging to attribute the effect of acupotomy to disability improvement. The reason behind the absence of discernible significant effects could be attributed to the small sample size within Group B, even after the application of PSM for correction. It’s noteworthy that Group A, prior to PSM correction, exhibited a significant decrease in RMDQ values. Group A demonstrated significant increase in EQ-5D values from the start of the study to week 6, while Group B did not show any significant increase. Comparing the 2 groups, the difference in the improvement in quality of life was not significant. This means that during the final follow-up, the efficacy 4 acupotomy sessions in Group A in terms of improving quality of life surpassed that of 6 acupotomy sessions in Group B. This underscores the point that a higher number of sessions does not always guarantee treatment satisfaction. When analyzing the trend changes over the study period between the groups, 3 indices—VAS (NRS), RMDQ, and EQ-5D—did not show any significant changes.

Acupotomy is an invasive treatment characterized by substantial stimulation, showcasing remarkable efficacy for chronic diseases even with a limited number of sessions conducted over a short span. Therefore, upon acupotomy, adequate treatment period and number of sessions should be planned for each patient. Lengthy treatment periods and many treatment sessions could potentially impose a burden on pain relief and improvement of quality of life. Consequently, it is important to establish the number of acupuncture sessions, considering the patient’s threshold. However, it is difficult to conclude that the treatment outcomes of acupotomy are not directly proportional to the number of acupotomy sessions since the sample sizes are small. Moreover, given that the variance between 4 and 6 sessions is not substantial, further studies to delve into the relationship between treatment sessions and the efficacy of acupotomy.

Certain limitations are associated with this study. The follow-up period post-treatment completion was relatively short in both groups, making it challenging to determine the sustained effects of acupotomy treatment. Additionally, the sample sizes for each group—46 and 15—were relatively small. To address this, a reduced number of 10 samples were extracted from each group for analysis. However, the significance of this study lies in its capacity to derive conclusive outcomes through a comparative and statistical analysis of pain alleviation, reduction in disability, and enhancement of quality of life concerning the frequency and duration of acupotomy sessions for herniated intervertebral disk of lumbar spine patients. The hope is that the results from this study will serve as valuable clinical evidence and data, propelling further investigations into acupotomy for lumbar spine intervertebral disk herniation. Anticipating a range of studies exploring Korean Medicine treatment technologies, the author eagerly awaits further progress in this domain.

In this study, the acupotomy treatment has a statistically significant effect on pain reduction in patients with lumbar disc herniation. However, there was no statistically significant difference in pain reduction, functional impairment, and quality of life improvement between 'acupotomy 4 times in 2 weeks' and 'acupotomy 6 times in 3 weeks'.

It is hoped that further studies on the number of acupotomy and treatment effects will be conducted in the future.

Conceptualization: BSK, YIK, TYP. Data curation: BSK, YIK, TYP. Formal analysis: BSK, YIK, JHJ. Funding acquisition: BSK, YIK, TYP. Investigation: BSK, YIK, HKK. Methodology: BSK, YIK, JHJ. Project administration: BSK, YIK, JHJ. Resources: BSK, YIK, JHJ. Software: BSK, YIK, HKK. Supervision: BSK, YIK, JHJ. Validation: BSK, YIK, TYP. Visualization: BSK, YIK, TYP. Writing – original draft: BSK, YIK. Writing – review & editing: All authors.

The authors have no conflicts of interest to declare.

This work was supported by clinical research grant from Daejeon Korean Medicine Hospital of Daejeon University in 2021.

Both studies were approved by the Institutional Review Board (IRB) of the Clinical Trial Center at Daejeon Korean Medicine Hospital of Daejeon University (DJDSKH-18-BM-02, DJDSKH-21-BM-06). These studies were also duly registered in the clinical research information system (CRIS) overseen by the Korea Disease Control and Prevention Agency (CRIS-KCT0002824, KCT0006043). Furthermore, the current study comparing the 2 aforementioned studies was similarly granted approval by the IRB of Clinical Trial Center, at Daejeon Korean Medicine Hospital of Daejeon University (DJDSKH-21-E-35).