Back pain occurs in 80% of the worldwide population and can occur at any age [1]. Herniated intervertebral discs (HIVD) are a common cause of back pain and can result in social and economic losses due to difficulties in personal and work life and increased medical expenses [2,3]. According to the National Health Insurance Statistics Yearbook [4], the number of patients treated for other intervertebral disc disorders (M51) gradually increased from 1,847,234 in 2013 to 1,939,137 in 2022. The disease is the 6th and 5th most common cause of hospitalization condition in all medical institutions and Korean medicine medical institutions.

HIVD is a disease in which the annulus fibrosus within the nucleus pulposus ruptures due to external forces or degenerative changes, compressing the spinal dura mater or nerve roots, causing neuropathy such as back pain, radiating pain in the lower extremities, and decreased sensation and muscle weakness in the area controlled by the nerves of the corresponding spinal segment [5]. HIVD is diagnosed by magnetic resonance imaging (MRI) or computed tomography (CT), and the surgery or conservative treatment is selected depending on the degree of herniation of the nucleus pulposus and clinical symptoms. Surgical treatment can reduce pain in a relatively short period of time, although there has been no significant difference in the results of surgery and conservative treatment after 4 years, and back pain can recur even after surgery [6,7]. Various conservative treatments, including acupuncture, moxibustion, herbal medicine, Chuna, pharmacopuncture, physical therapy, medication, and manual therapy, have been reported. Interest in Korean medicine is increasing as published reports have described that HIVD symptoms can be improved by > 80% with Korean medicine treatment [2,8].

Pharmacopuncture, a new type of acupuncture therapy, maximizes the effect through the combination of physical stimulation of acupuncture with the chemical reaction of herbal medicines. It has been used to treat various musculoskeletal disorders by extracting, purifying, diluting, and mixing herbs and injecting them into acupoints [9,10]. According to the clinical practice guidelines of Korean medicine in 2021 [11], bee venom pharmacopuncture is mainly used in HIVD treatment. Additionally, studies have reported the use of Soyeom, Shinbaro, Joongseongouhyul, and Aconiti Ciliare Tuber pharmacopuncture treatments for HIVD [7,8,12-14]. Most studies using pharmacopuncture for HIVD treatment have been retrospective, such as case reports or case–control studies. There is a lack of pragmatic randomized controlled trials (PRCTs) or prospective studies that reflect the actual clinical environment.

Therefore, in this study, we aimed to investigate the effect of pharmacopuncture on HIVD by conducting a prospective PRCT utilizing commonly used bee venom, Soyeom, and Aconitum ciliare Decaisne (ACD) to reflect an actual clinical environment.

1. Participants

The study participants were recruited from June 27, 2023, to August 8, 2023, through a recruitment notice at Daejeon University Cheonan Korean Medicine Hospital. The inclusion and exclusion criteria are presented in Table 1. The participants received a detailed explanation, voluntarily signed a written consent form, and underwent screening. The present study was approved by the Institutional Review Board of the Daejeon University Cheonan Korean Medicine Hospital (approval no. DJUMC-2023-BM-03) and registered with the Clinical Research Information Service (CRIS no. KCT0008542). The present study was conducted in accordance with the principles stipulated in the Declaration of Helsinki.

Table 1 . Study criteria.

Inclusion criteria

1. Age of 19–64 years

2. If lumbar disc herniation has been diagnosed through CT or MRI within the last 6 months

3. VAS score for back or radiating of ≥ 50 mm

4. Agreed to participate in this trial and voluntary signing a written informed consent form

Exclusion criteria

1. Diagnosis of spinal metastases of tumors, acute fractures, and/or spinal dislocations

2. Back pain caused by tumor, fibromyalgia, rheumatoid arthritis, and/or gout

3. History of other chronic diseases, including stroke, myocardial infarction, kidney disease, and/or diabetic neuropathy

4. Currently taking medications, such as steroids, immunosuppressants, and/or psychiatric drugs

5. Pain in other parts worse than the low back pain

6. Pregnancy or breastfeeding

7. Lumbar surgery within the last 3 months

8. Participation in another clinical trial within 4 weeks

9. Difficulty in completing the informed consent form

10. Other reasons, as determined by the researchers

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

2. Interventions

The treatment group received a combination of pharmacopuncture and integrated Korean medicine (IKM), whereas the control group received only the IKM treatment. The IKM treatment included acupuncture, cupping, and infrared therapy.

Based on literature evidence [8,12-14], Korean medicine doctors (KMDs) used Soyeom pharmacopuncture for the initial treatment of HIVD, and then mainly performed bee venom pharmacopuncture to maximize pain reduction. Depending on the participant’s condition on the day of the procedure, if there was a concern about hypersensitivity to bee venom pharmacopuncture, the relatively safe ACD pharmacopuncture was selected to avoid hypersensitivity. Bee venom pharmacopuncture (Hanen) at a concentration of 1,000:1 using dried bee venom (Unimed Pharm Inc.), Soyeom pharmacopuncture (Kirin Korean Medicine Industrial Institute), and ACD pharmacopuncture (Kirin Korean Medicine Industrial Institute) were used in the treatment group. A 1-mL disposable syringe (26 G, 1/2 inch syringe; Junglim Medical) and a disposable needle (30 G, 1/2 inch; Junglim Medical) were used.

Considering the participants' pain and area, the depth of needling and type of acupuncture were selected based on the clinical judgment of the KMD, and the needle was injected for 15 minutes (sterilized disposable stainless steel acupuncture needle; Dongbang Medical). The needles used were 0.20 × 30, 0.25 × 30, and 0.40 × 70 mm in size. Infrared light therapy was applied to the lower back for 15 minutes simultaneously with acupuncture treatment. Cupping was performed using the dry cupping technique for 5 minutes in the lumbar region before or after the acupuncture treatment. A licensed KMD with at least 2 years of clinical experience in acupuncture and pharmacopuncture conducted the treatment. The participants received a total of 16 treatments twice a week for 8 weeks and a follow-up evaluation on the 9th week.

As a pragmatic clinical study, the treatment methods, including acupoint, depth of treatment, and type and dose of pharmacopuncture, were not pre-determined and were decided based on the clinical judgment of the KMD, depending on the participant’s symptoms and conditions. The acupoints used during the procedure, the type and dose of pharmacopuncture administered (mL), and the acupoint at which the procedure was performed were retrospectively reviewed and recorded.

3. Study design

The present study was a prospective PRCT. The participants were assigned in a 1:1 ratio to the treatment and control groups using the block randomization method. The randomization codes were distributed by an independent statistician and sealed in an impermeable envelope. The researcher opened the envelope with the same number as the randomization number and assigned the participants to the treatment or control group. Owing to the characteristics of the pragmatic clinical study and intervention used, blinding of the participants and investigators was not applied. Additionally, the number of people conducting the study was limited; hence, blinding of investigators and assessors was not implemented. Given that blinding was not performed, a synergistic effect could have occurred in the research hypothesis [15].

Participants voluntarily signed a written consent form and underwent screening. At screening, the participants’ demographic data, such as sex, age, height, and weight, were recorded. Moreover, vital sign taking, radiating pain assessment, visual analog scale (VAS) measurement, and physical examination were performed; and medical and medication histories were taken. We checked whether they had undergone lumbar MRI or CT within the last 6 months, and if they had no imaging test record, we performed a lumbar CT on the patient at our hospital and confirmed whether he or she has a lumbar disease.

The participants who met the eligibility criteria for this trial were randomly assigned to the treatment or control group and underwent a baseline evaluation at visit 1. In both groups, 16 sessions were conducted twice a week for 8 weeks. After completing the 8-week intervention, a follow-up evaluation was conducted on the 9th week and the trial was terminated (Table 2).

Table 2 . Study schedule and outcome measurements.

Visit	Study period
	Screening	Enrollment	Treatment period																Follow-up
	Week −1 to 0	Week 0 (Baseline)	V1	V2	V3	V4	V5	V6	V7	V8	V9	V10	V11	V12	V13	V14	V15	V16	V17
			Week 1–8																Week 9
Enrollment:
Eligibility screen	×
Informed consent	×
Check MRI or CT of the lumbar spine	×
Allocation		×
Intervention:
Pharmacopuncture with IKM			×	×	×	×	×	×	×	×	×	×	×	×	×	×	×	×
IKM			×	×	×	×	×	×	×	×	×	×	×	×	×	×	×	×
Assessments:
VAS			×	×	×	×	×	×	×	×	×	×	×	×	×	×	×	×	×
PGIC							×				×				×				×
ODI			×				×				×				×				×
No worse than mild pain					×		×		×		×		×		×		×		×
ROM			×	×	×	×	×	×	×	×	×	×	×	×	×	×	×	×	×

The “×” refers to what is done in the given period..

MRI, magnetic resonance imaging; CT, computed tomography; IKM, integrated Korean medicine; VAS, visual analog scale; PGIC, Patient Global Impression of Change; ODI, Oswestry Disability Index; ROM, range of motion..

The primary outcome measure was the mean change in VAS scores from baseline to visit 17. The secondary outcomes were the mean changes in the Oswestry Disability Index (ODI), Patient Global Impression of Change (PGIC), and range of motion (ROM) from baseline to visit 17. For “no worse than mild pain,” a survival analysis was conducted at the point when the VAS decreased by 50% as compared to baseline data. The VAS score for pain and ROM were evaluated at each visit; ODI at visits 1, 5, 9, 13, and 17; report of “no worse than mild pain” at visits 3, 5, 7, 9, 11, 13, 15, and 17; and PGIC at visits 5, 9, 13, and 17. The adverse events (AEs) were monitored at each visit.

4. Sample size

This clinical study investigated the effectiveness of pharmacopuncture as treatment for HIVD. Based on the study by Park et al. [16], the effect size of pharmacopuncture for chronic cervical pain ranges from 0.16 to 0.51. Considering a power of 80% and a conservative effect size of the clinical study of 0.1 to 0.3, we planned to recruit a total of 40 participants, with 20 participants per group, according to Whitehead et al. [17]. In Park et al.’s [16] study, the dropout rate was 3%; hence, in the present study, the dropout rate was set at 5%, and the final sample size was set at 22 people per group, with a total of 44 people.

5. Statistical analyses

Full set and per-protocol analyses were conducted in this study, and the full analysis set was used in the primary analysis. For continuous data, an independent two-sample t-test or Wilcoxon rank sum test was used. For categorical data, the chi-square or Fisher’s exact test was used, and the data were expressed as frequencies and percentages. The significance level for all the statistical tests was set at 5%. Missing data were replaced using the last observation carried forward, which is a type of multiple imputation method. Missing data for the linear mixed model were handled using a mixed-effects model for repeated measures (MMRM) approach.

The marginal mean, standard error, 95% confidence interval (CI), and p-value for the amount of change in each group estimated by using an analysis of the covariance model and of the change between groups are presented. To analyze the trend of changes in the evaluated variables at each time point as compared with the baseline data, the marginal means and their standard errors (or 95% CI) are presented in Figs. 1–3 using the linear mixed effect model by the MMRM method.

Figure 1. Change in the VAS score over time. Differences between the groups were analyzed using a linear mixed-effects model with baseline values as adjustments. All data are expressed as mean difference (95% confidence interval). VAS, visual analog scale.

Figure 2. Changes in the secondary outcomes over time. (A) ODI; (B) PGIC; (C) flexion; (D) extension; (E) rotation, left; (F) rotation, right; (G) lateral bending, left; (H) lateral bending, right. (C–H) refer to the participants’ range of motion. ODI, Oswestry Disability Index; CI, confidence interval; PGIC, Patient Global Impression of Change.

Figure 3. Subgroup analysis. (A) VAS; (B) ODI. VAS, visual analog scale; MD, mean difference; ODI, Oswestry Disability Index.

Additionally, a survival analysis was performed for the participants with at least a 50% reduction in the pain indicators (VAS scores) at each time point as compared with the baseline data. A Kaplan-Meier survival analysis was used to compare the time needed to recover from back pain, which was indicated by at least a 50% reduction in VAS scores after randomization, and the curves were compared using the log-rank test. A Cox proportional hazards model was used to compare the hazard ratios.

Additionally, the data of subgroups within the treatment group were analyzed. The subgroups included a group that mainly used bee venom more than 14 times (hereinafter referred to as the main bee venom group, n = 9) and a group that mainly used Soyeom or ACD more than 11 times (hereinafter referred to as the pharmaco-pack group, n = 6). An analysis of variance (ANOVA) was performed for the three groups along with the control group. If normality was satisfied, one-way ANOVA was performed. Otherwise, the Kruskal-Wallis test was performed. Then, a post hoc test was performed using the Tukey method.

1. Participant’s general characteristics

Forty-four people were recruited and registered for this clinical trial; they then underwent screening (treatment group: 22; control group: 22). During the study, one participant in the control group withdrew from the study, resulting in 43 participants who completed the clinical study (Fig. 4). There were no significant differences in the participants’ general characteristics (Table 3).

Figure 4. Study flow chart.

Table 3 . Demographic characteristics.

Characteristic	Control group (n = 22)	Treatment group (n = 22)	p-value*
Sex, female	17 (77)	16 (73)	0.728
Age (y)	42.8 [13.4]	46.4 [10.7]	0.424
Height (cm)	162.9 [9.8]	163.2 [7.9]	0.452
Weight (kg)	63.8 [16.0]	66.7 [17.5]	0.474
Systolic BP (mmHg)	120.6 [14.3]	120.2 [13.0]	0.879
Diastolic BP (mmHg)	74.2 [9.4]	76.5 [10.0]	0.359
Pulse (bpm)	76.0 [11.1]	82.0 [11.6]	0.084
Body temperature (°C)	36.5 [0.2]	36.6 [0.2]	0.057
Radiation pain	10.0 (45)	9.0 (41)	0.761
SLRT, RT (°)	80.5 [11.7]	83.6 [9.0]	0.415
SLRT, LT (°)	82.3 [8.7]	77.7 [16.9]	0.539
Bragard test (+)	2.0 (9)	5.0 (23)	0.412
Patrick test (+)	9.0 (41)	9.0 (41)	> 0.999
ADF (+)	6.0 (27)	8.0 (36)	0.517
APF (+)	3.0 (14)	3.0 (14)	> 0.999

Data are presented as n (%) or mean [standard deviation]..

BP, blood pressure; SLRT, straight leg raise test; RT, right; LT, left; ADF, ankle dorsiflexion; APF, ankle plantar flexion..

*Categorical variable: Fisher’s exact test; continuous variable: Wil­coxon rank sum test..

2. Visual analog scale

In the intra-group comparison, the amounts of VAS score change were −45.5 (95% CI: −55.0, −36.0) and −22.4 (95% CI: −31.9, −12.9) for the treatment and control groups, respectively, which showed significant differences. The adjusted mean difference between the two groups was −23.0, showing a significant difference (p = 0.0013; Table 4, Fig. 1).

Table 4 . Outcome comparison between and within the control and treatment groups.

Variable	Observed value				Change from the baseline
	Control	Treatment	p-value*		Control†	p-value*	Treatment†	p-value*	p-value‡
VAS
Visit 1	70.3 (9.5)	68.9 (7.4)	0.589
Visit 17	46.8 (23.7)	24.5 (20.5)	0.004		−22.4 [−31.9, −12.9]	< 0.0001	−45.5 [−55.0, −36.0]	< 0.0001	0.0013
ODI
Visit 1	32.0 (12.0)	33.6 (14.7)	0.851
Visit 17	21.2 (12.5)	13.4 (7.0)	0.047		−11.5 [−15.8, −7.1]	< 0.0001	−19.5 [−23.8, −15.1]	< 0.0001	0.0121
PGIC
Visit 17	1.7 (0.8)	2.2 (0.8)	0.044		1.7 [1.4, 2.1]	< 0.0001	2.2 [1.9, 2.6]	< 0.0001	0.0419
ROM, flexion
Visit 1	82.3 (14.8)	7.5 (19.9)	0.236
Visit 17	81.6 (15.8)	84.8 (8.5)	0.568		1.3 [−2.9, 5.4]	0.5325	8.3 [4.1, 12.4]	0.0002	0.0229
ROM, extension
Visit 1	24.1 (6.5)	20.9 (8.3)	0.213
Visit 17	26.4 (6.0)	24.8 (6.6)	0.391		3.5 [0.8, 6.2]	0.0121	2.7 [−0.0, 5.3]	0.0519	0.6659
ROM, rotation, Lt
Visit 1	41.4 (5.8)	43.9 (2.1)	0.213
Visit 17	42.3 (7.4)	43.0 (5.5)	0.802		−0.0 [−2.9, 2.8]	0.9877	0.0 [−2.8, 2.9]	0.9877	0.9830
ROM, rotation, Rt
Visit 1	40.7 (7.0)	44.1 (2.0)	0.067
Visit 17	41.8 (8.1)	42.0 (6.1)	0.588		−0.2 [−3.3, 3.0]	0.9201	−0.8 [−3.9, 2.4]	0.6348	0.7962
ROM, lateral bending, Lt
Visit 1	25.9 (6.5)	27.3 (6.1)	0.387
Visit 17	29.1 (2.9)	29.3 (2.3)	0.981		2.6 [1.4, 3.7)	< 0.0001	2.7 [1.5, 3.8)	< 0.0001	0.9174
ROM, lateral bending, Rt
Visit 1	27.5 (4.8)	27.0 (5.9)	0.936
Visit 17	29.1 (2.9)	29.1 (2.9)	> 0.999		1.8 [0.6, 3.0]	0.0046	1.9 [0.7, 3.1]	0.0031	0.9153

Data are presented as the mean (standard deviation)..

VAS, visual analog scale; ODI, Oswestry Disability Index; PGIC, Patient Global Impression of Change; ROM, range of motion; Lt, left; Rt, right..

*p-values were compared within each group..

^†Estimated marginal mean changes and 95 % confidence intervals controlled for the baseline..

^‡p-values were compared between groups..

3. Oswestry Disability Index

In the intra-group comparison, the amounts of ODI change were −19.5 (95% CI: −23.8, −15.1) and −11.5 (95% CI: −15.8, −7.1) for the treatment and control groups, respectively. The adjusted mean difference between the two groups was −8.0, showing a significant difference (p = 0.0121; Table 4, Fig. 2A).

4. Patient Global Impression of Change

In the intra-group comparison, the amounts of PGIC change were 2.2 (95% CI: 1.9, 2.6) and 1.7 (95% CI: 1.4, 2.1) in the treatment and control groups, respectively, showing significant difference. In the intergroup comparison, the mean difference was 0.5, which was significant (p = 0.0419; Table 4, Fig. 2B).

5. Range of motion

The changes in flexion were significant in the treatment group. Lateral bending (left and right) showed significant results within both each groups, although there were no significant differences in extension or rotation (Table 4, Fig. 2C–H).

6. Survival analysis

At week 9, 16 participants in the treatment group and 7 participants in the control group recovered, and the treatment group took an average of 48.5 days to recover. Recovery in the treatment group was 2.82 times higher than that in the control group (Fig. 5).

Figure 5. Survival analysis.

7. Subgroup analyses

Subgroup analyses were conducted for the VAS and ODI scores. The amounts of VAS score and ODI changes were −47.3 (p < 0.0001) and −20.0 (p < 0.0001) for the main bee venom group, −42.8 (p < 0.0001) and –16.5 (p < 0.001) for the pharmaco-pack group, and −23.6 (p < 0.0001) and −12.2 (p < 0.0001) for the control group. All three groups showed significant differences (Fig. 3).

In the post hoc test between the groups, the difference in the VAS score between the main bee and control groups was −23.8, showing a significant difference (p = 0.029), although there was not significant difference between the pharmaco-pack and control groups and between the main bee venom and pharmaco-pack groups. In the post hoc test between the three groups, the difference in ODI was not significant; however, between the main bee venom and control groups, the ODI score tended to decrease more in the main bee venom group.

8. Adverse events

Altogether, 10 AEs were observed in eight participants during the study period (Table 5). All AEs were mild in severity and included itching, headache, aggravation of the trigger finger, acute enteritis, body aches, and fever. The causality of AEs was classified as follows: “definitely related,” “probably related,” “possible related,” “probably not related,” “definitely not related,” and “unknown.” Itching occurred in seven cases in five participants in the treatment group, and all of them recovered after 2 to 6 days. Five cases were judged to be probably not related to the intervention, either in which the one-off symptoms occurred and were cured or in subjects who occasionally took antihistamines for allergic dermatitis with a medical history, but in two cases in which the same subjects repeatedly complained of itching were judged to be probably related to the intervention. Other AEs were considered to be probably not or definitely not related to the intervention. There were no significant differences in the vital signs or hematological test results between the treatment and control groups. No serious AEs were observed.

Table 5 . AEs.

Group	Characteristic	Overall (n = 44)	Control group (n = 22)	Treatment group (n = 22)	p-value
AE summary	Patients with any AE	8 (18.0)	1 (4.5)	7 (32.0)	0.046
	Patients with severe AEs	0 (0)	0 (0)	0 (0)	> 0.999
	Patients with AEs leading to discontinuation	0 (0)	0 (0)	0 (0)	> 0.999
	Patients with investigator-defined AEs	2 (4.5)	0 (0)	2 (9.1)	0.488
	Patients with serious AEs	0 (0)	0 (0)	0 (0)	> 0.999
AE symptoms	Itching	5 (11.0)	0 (0)	5 (23.0)	0.048
	Headache	1 (2.3)	0 (0)	1 (4.5)	> 0.999
	Trigger finger	1 (2.3)	1 (4.5)	0 (0)	> 0.999
	Acute enteritis	1 (2.3)	0 (0)	1 (4.5)	> 0.999
	Body ache	1 (2.3)	0 (0)	1 (4.5)	> 0.999
	Hot flush	1 (2.3)	0 (0)	1 (4.5)	> 0.999

Data are presented as n (%)..

AE, adverse event..

9. Interventions

Among the 352 pharmacopuncture sessions, bee venom, Soyeom, and ACD were administered to the patients in 222 (63.0%), 122 (34.7%), and 8 (2.3%) sessions, respectively. The types and doses of pharmacopuncture used in the treatment groups are presented in Fig. 6. Both Soyeom and ACD were administered at a dose of 1 mL. Hypersensitivity reactions were confirmed by a skin test prior to the first bee venom pharmacopuncture. The initial dose was mainly 0.1 mL; however, as per the participant responses, it was diluted with distilled water or started at a smaller amount of 0.005-0.200 mL in some cases. A 0.001–1.000 mL (average of 0.22 mL) of bee venom (1,000:1) was administered depending on the participant’s skin reaction. The bee venom doses of < 0.05, 0.05–0.49, 0.50–0.99, and 1 mL were administered in 42 (18.9%), 151 (68.0%), 28 (12.6%), and 1 (0.5%) session, respectively. Pharmacopuncture was performed at the following acupoints: GV3, EX-B5, EX-B8, and EX-B2. In the early stages of the trial, Soyeom was mainly used, and the number of bee venom doses gradually increased toward the end. Duplicate acupuncture needle types were allowed; the type of acupuncture needles used in the participants are listed in Table 6. The acupoints utilized were as follows: bilateral BL23, BL24, BL25, BL26, BL27, BL53, BL54, EX-B2, and GB30, and unilateral EX-B5, GV3, and EX-B8.

Figure 6. Types and does of pharmacopuncture in the treatment group.

Table 6 . Type of acupuncture needles used.

Type (thickness × length; mm)	Control group (n = 436)	Treatment group (n = 422)
0.20 × 30	133 (31)	160 (38)
0.25 × 30	210 (48)	192 (45)
0.40 × 75	93 (21)	70 (17)

Data are presented as n (%)..

10. Ethical statement

The present study was performed according to the guidelines stipulated in the Helsinki Declaration and the Korean Clinical Practice Guidelines. This clinical trial protocol was reviewed and approved by the Institutional Review Board of the Daejeon University Cheonan Korean Medicine Hospital (approval no.: DJUMC-2023-BM-03). All participants voluntary provided written informed consent after receiving a full explanation of the potential benefits and risks of the intervention prior to study participation.

In the present study, the data of 44 participants diagnosed with lumbar HIVD within the previous 6 months were examined to determine the effectiveness of pharmacopuncture administered during a 9-week period.

Pharmacopuncture is widely used in clinical settings to treat various diseases, particularly musculoskeletal disorders. Many studies have reported on the use of pharmacopuncture as a treatment for HIVD. HIVD patients receiving Soyeom pharmacopuncture as an initial treatment have shown significant improvements [12]. Previous studies have reported significant improvements in the VAS score, ODI, and ROM for HIVD patients receiving bee venom pharmacopuncture than for those receiving acupuncture [18,19]. ACD pharmacopuncture as a treatment for HIVD has been reported to be reduce pain without any side effects [14]. Most previous studies have evaluated the effectiveness of a single pharmacopuncture treatment for HIVD; however, no pragmatic study using various types of pharmacopuncture treatments as in actual clinical practice has been reported. As a PRCT, the present study evaluated the therapeutic effect of pharmacopuncture for HIVD in an environment where the KMD can change the type of pharmacopuncture at each visit, reflecting the real scenario in clinical practice.

In the present study, the following three types of pharmacopuncture were used: bee venom, Soyeom, and ACD pharmacopuncture. Bee venom pharmacopuncture uses venom within the venom sac. The main components of bee venom (melittin, apamin, and phospholipase A2) regulate the body’s immune function and increase blood flow, providing anti-inflammatory, analgesic, antiviral, and antibacterial effects as well as resulting in immune system regulation and blood circulation promotion. It is used to treat a variety of musculoskeletal pain and neuropathic diseases, such as HIVD, rheumatoid arthritis, stenosis, and sequelae of spinal surgery [20-22]. Soyeom pharmacopuncture alleviates inflammation and treats whiplash injuries, supraspinatus tendinitis, subdeltoid bursitis, complex regional pain syndrome, ankle tendinitis, and HIVD [12,23,24]. ACD pharmacopuncture contains 1 μg/mL of ACD and aconitine, the main ingredient of ACD, and has excellent analgesic effects; however, it is toxic and caution is required for its use, as there is a risk of death due to poisoning when taken orally. ACD pharmacopuncture, prepared through distillation, extraction, filtration, drying, and dilution, has few side effects, and it has been used to treat HIVD, full-thickness rupture of the supraspinatus tendon, and complex regional pain syndrome [14,25,26].

Many studies have reported the effectiveness of bee venom pharmacopuncture for the treatment of HIVD [3,6,8,13,18,21,22], and a previous report has shown that Soyeom pharmacopuncture improved early acute pain [12]. Moreover, ACD pharmacopuncture reduces pain and has no side effects, making it safer to use than bee venom pharmacopuncture, which can cause hypersensitivity [14]. Therefore, the present study mainly used Soyeom pharmacopuncture in the early stage of treatment and bee venom pharmacopuncture thereafter. ACD pharmacopuncture was applied in cases suspected of having hypersensitivity to bee venom pharmacopuncture based on the participant’s condition on the day of treatment or his body’s reaction after pharmacopuncture. Bee venom pharmacopuncture was the most frequently used intervention. Given that the sensitivity to bee venom pharmacopuncture varies from person to person, a skin test was performed, and the dose was diluted with distilled water or increased depending on the skin' reaction. Redness and itching may occur; therefore, the dosage cannot be specified; although it, was confirmed that bee venom pharmacopuncture was effective even in small amounts as compared with other pharmacopuncture methods.

The primary outcome of this study was the change in the VAS scores for low back pain. VAS is mainly used to quantify subjective pain [27]. Given that the subjective evaluation standards for pain did not match among the participants, the treatment effects were evaluated based on the amount of change at visit 17 compared with the baseline values. The VAS score change was significant in both groups; especially at visit 10, a significant difference in the VAS score change was observed between the treatment and control groups. To evaluate the effect of each type of pharmacopuncture, the treatment group was further divided into the main bee venom and pharmaco-pack subgroups, and their outcomes were analyzed. The main bee venom group included patients who underwent bee venom pharmacopuncture more than 14 times, which was approximately 90% of the number of trials. If the pharmaco-pack group underwent Soyeom or ACD pharmacopuncture more than 12 times, the number of participants would reduce to 2. Therefore, the number of participants was set to ≥ 11 to ensure a minimum number of participants for whom statistics could be used. When classifying the treatment group into subgroups, because the number of participants was small, it was not possible to clearly divide each pharmacopuncture treatment by a certain number of sessions, and since ACD pharmacopuncture was administered a very small number of times overall, it could not be analyzed individually. As a result, the main bee venom and control groups showed significant differences in the outcomes. The difference in the outcomes between the main bee venom and pharmaco-pack groups was not significant because of the small sample size; however, the VAS score of the main bee venom group was more likely to improve than the pharmaco-pack group, and the pharmaco-pack group also included a small number of participants who underwent bee venom pharmacopuncture. Considering that the average VAS score reduction tended to increase as the number of bee venom pharmacopuncture treatments increased, rather than with the increase in the number of Soyeom or ACD pharmacopuncture treatments during the treatment period, the significant difference in the outcomes observed between the treatment and control groups was due to treatment with the bee venom pharmacopuncture. Thus, it is highly likely that the bee venom pharmacopuncture had the greatest effect.

To evaluate the improvement in participants’ daily life function, pain score, and satisfaction at visit 17, as compared with the baseline data, the changes in ODI, PGIC, report of “no worse than mild pain,” and ROM were selected as secondary outcomes. The ODI was used to evaluate the degree of disability in daily life caused by pain [28]. The changes in ODI were significant in both the treatment and control groups, and from visit 13 onward, the treatment group showed significant improvements as compared the control group, confirming that the daily life functions improved to a greater extent in a shorter period of time in the treatment group. The PGIC is a method for assessing the degree of improvement in pain intensity as compared with that before study participation [29]. There was a significant change in PGIC in both the treatment and control groups, and a significant difference was observed between the two groups at visit 13. Additionally, reports of “no worse than mild pain” reflected the degree of improvement and satisfaction with weekly treatment, as compared with that before starting the intervention [30]. Although the index was considered to correlate with the VAS score, the results of this study did not indicate consistency with the VAS score, indicating a limitation of the subjective evaluation. Therefore, we performed survival analysis by redifining recovery as a 50% decrease in VAS compared to the baseline VAS. The average number of days needed to recover in the treatment group was 48.5 days, and accordingly, pharmacopuncture treatment for HIVD required at least 7 weeks. There was a significant improvement in ROM during flexion in the treatment group, but no significant difference was observed in the other movements. HIVD is clinically characterized by limited ROM and increased pain during flexion, which can be ameliorated with pharmacopuncture treatment.

In the safety assessment, AEs were reported significantly more frequently in the treatment group (n = 7) than in the control group (n = 1), with a higher incidence of itching observed in the treatment group (n = 5) than in the control group. Two of the participants with itching were thought to be probably related to the intervention because of the possibility of hypersensitivity to the bee venom. Mild itching can occur normally due to a hypersensitivity to bee venom and is considered to be a positive response to treatment rather than a side effect. Therefore, when using bee venom pharmacopuncture, it is necessary to carefully adjust the drug’s dosage to ensure safe administration of the intervention. However, since itching occurred in seven out of the 352 total pharmacopuncture sessions, pharmacopuncture could be considered a relatively safe treatment. The other symptoms had no causal relationship with the intervention. There were no dropouts due to adverse reactions or serious AEs.

Our study has several limitations. First, a bias may potentially exist due to the lack of blinding of participants, researchers, and evaluators. We hope that this bias can be reduced in follow-up studies that would apply blinding. Second, the generalizability of the study results is limited, as the study only analyzed a small sample size of 44 and had a relatively short follow-up period of 9 weeks. Third, because the present study involved outpatients, there were variables other than the intervention treatment that could not be controlled. Fourth, given that the study analyzed cases receiving a combination of treatments rather than a single intervention and that there were cases receiving a mixed of type of pharmacopuncture, it is difficult to reveal the effects of a specific pharmacopuncture. However, the present investigation is significant because it confirmed the effectiveness and safety of combined pharmacopuncture treatment through a pragmatic clinical study that reflected a clinical environment for HIVD. In the future, further studies with a large sample size and relatively long follow-up period are needed to determine the effectiveness of each pharmacopuncture treatment for HIVD.

The analyses using a linear mixed model by the MMRM approach revealed that pharmacopuncture showed significantly superior efficacy at 9 weeks of treatment based on the obtained VAS score, ODI, PGIC, survival analysis data, and ROM (flexion) at this time point. For the treatment of HIVD, we confirmed that pharmacopuncture, especially bee venom, was more effective for pain reduction, daily life function improvement, shortening the treatment period, increasing treatment satisfaction, and improving flexion, as compared with IKM treatment consisting of acupuncture, cupping, and infrared therapy alone.

Conceptualization: JHK, HL. Investigation: YYC. Methodology: YYC, HWR. Project administration: JHK, HL. Supervision: JHK, HWR. Writing - original draft: YYC. Writing - review & editing: HWR, JHK, HL.

The authors have no conflicts of interest to declare.

None.

This study was conducted according to the guidelines stipulated in the Helsinki Declaration and the Korean Clinical Practice Guidelines. This clinical trial protocol was reviewed and approved by the Institutional Review Board of the Daejeon University Cheonan Korean Medicine Hospital (IRB no.: DJUMC-2023-BM-03). All participants voluntarily provided written informed consent after receiving a fully explanation of the potential benefits and risks of the interventions prior to study participation.