Low back pain (LBP), which refers to pain from below the second lumbar vertebra where the lumbar nerve ends to the sacroiliac joint and hip joint, is caused by the disruption of the function and balance of the intervertebral disks, joints, ligaments, nerves, and blood vessels [1]. Based on statistics from the Global Burden of Disease, the number of prevalent cases of LBP globally was estimated to be 619 million in 2020, marking a substantial increase of 60.4% from 1990 [2]. Common causes of LBP include lumbar strain, herniated disks, degenerative spondylopathy, spinal stenosis, and fractures [3]. Common medications for LBP management include acetaminophen, nonsteroidal anti-inflammatory drugs (NSAIDs), steroids, antidepressants, and narcotic analgesics (i.e., opioids). Long-term use of these medications may lead to several side effects, such as gastrointestinal disorders, renal problems, and liver damage [4]. Furthermore, complications can arise when conservative treatments fail and surgery is required. In 10%–20% of these cases, postsurgical pain relief and surgical site recovery may not occur [5].

In traditional Korean medicine, various LBP treatments, such as acupuncture (AT), cupping, and herbal medicine, are used. One treatment method, thread embedding acupuncture (TEA), involves implanting polydioxanone (PDO) into the subcutaneous tissue to induce sustained aseptic inflammatory responses, increase the formation and regeneration of collagen and other connective tissues, strengthen tissues, and enhance the body’s self-healing ability. Furthermore, TEA has been widely used [6].

Although most aspects of traditional Korean medicine focus on musculoskeletal diseases, such as LBP, clinical studies on the use of TEA for musculoskeletal diseases are limited compared with research on other conditions and treatments, such as plastic surgery, skin lifting, obesity, insomnia, and urinary dysfunction [2]. Sung et al. [7] performed a meta-analysis of RCTs conducted before January 2018 on the efficacy of TEA for lumbar herniated intervertebral disks (HIVDs) in 2020. On the other hand, Dong et al. [8] reviewed RCTs conducted before 2018 on the efficacy of TEA for chronic lumbar strain in 2020. However, no studies have yet conducted a comprehensive, high-quality trend analysis of the TEA method and its overall effectiveness in the treatment of LBP. Moreover, no studies have analyzed the trends in the use of TEA for LBP over the last 5 years.

To address this gap, we selected RCTs published between February 2018 and July 2023 to investigate the latest trends and efficacy regarding the use of TEA for LBP. This study provided a more detailed analysis of the TEA itself compared with previous studies conducted in 2020. Moreover, we analyzed various diseases that cause LBP, which is not limited to a single disease, and analyzed the efficacy and safety of thread embedding therapy by comparing it with more diverse interventions used in the control groups. We evaluated the studies based on the following categories: (1) year of publication, (2) type of disease, (3) sample size, (4) treatment interventions, (5) treatment frequency and duration, (6) insertion point and depth of thread embedding, (7) types and sizes of threads and needles, (8) evaluation indexes, (9) treatment outcomes, and (10) side effects. Furthermore, we have developed recommendations for future research on the use of TEA for LBP.

1. Database selection and search methods

Studies written in Korean were retrieved from the following domestic academic databases: Oriental Medicine Advanced Searching Integrated System (OASIS), Korean Studies Information Service System (KISS), Research Information Sharing Service (RISS), and Korean Traditional Knowledge Portal (KTKP). International studies were retrieved from PubMed and the China National Knowledge Infrastructure (CNKI).

The search terms varied for each database (Table 1). The search was limited to articles published between February 2018 and July 2023. No limits were placed on language or the underlying causes of LBP.

Table 1 . Search strategies for each database.

Korean database (OASIS, KISS, RISS, and KTKP)	Maesun (in Korean)
CNKI	(“maixianzhi” OR “maixian” OR “maizhen” OR “maicangliaofa”) AND “yaotong” AND “suijiduizhaoshiyan” (all in Chinese)
PubMed	#1. Search: “needle embedding therapy” OR “embedding therapy” OR “catgut embedding” OR “acupoint embedding” OR “thread embedding”
	#2. Search: “RCT” OR “randomized control trial”
	#1 AND #2

OASIS, Oriental Medicine Advanced Searching Integrated System; KISS, Korean Studies Information Service System; RISS, Research Information Sharing Service; KTKP, Korean Traditional Knowledge Portal; CNKI, China National Knowledge Infrastructure..

2. Inclusion and exclusion process

After the search process was completed, two independent researchers (MYG and LHG) removed the duplicate articles, excluded the studies based on the predetermined criteria, and used RefWorks to perform data selection. Articles related to animal experiments, those in which TEA was not the primary intervention, those that were solely focused on evaluating the cost-effectiveness of TEA, and those whose original full text could not be obtained were excluded from consideration. The first round of article selection was based on the titles and abstracts of the remaining articles. In the second round, the selected documents were reviewed by the aforementioned researchers, who then reconciled their opinions. In cases of disagreement, a third-party was consulted, and a decision was made based on the majority opinion. Ultimately, nine articles were selected for inclusion in the present study (Fig. 1).

Figure 1. Article selection flow chart. CNKI, China National Knowledge Infrastructure; RISS, Research Information Sharing Service; OASIS, Oriental Medicine Advanced Searching Integrated System; KISS, Korean Studies Information Service System; KTKP, Korean Traditional Knowledge Portal; TEA, thread embedding acupuncture.

3. Data extraction

Data on the following topics were extracted from the selected studies: (1) year of publication, (2) type of disease, (3) sample size, (4) treatment interventions, (5) treatment frequency and duration, (6) insertion point and depth of thread embedding, (7) types and sizes of threads and needles, (8) evaluation indexes, (9) treatment outcomes, and (10) side effects. The data were then analyzed to provide a comprehensive overview and evaluation of the research on the use of TEA for LBP.

4. Assessing the quality of the selected articles

An assessment of the risk of bias (RoB) in the relevant literature was conducted using the Cochrane RoB tool presented in the National Evidence-based Healthcare Collaborating Agency (NECA) systematic literature review manual [9].

1. Search results

Since previous studies on the use of TEA for LBP that were published before January 2018 were analyzed, we decided to search for RCTs published between February 2018 and July 2023 [7,8]. Regarding international databases, 2,738 studies were retrieved from PubMed (n = 134) and CNKI (n = 2,604). Regarding domestic databases, 127 studies were retrieved from RISS (n = 80), OASIS (n = 23), KISS (n = 11), and KTKP (n = 13). After removing 84 duplicate studies, the remaining 2,781 studies were evaluated based on the exclusion criteria. We excluded articles that were not RCTs (n = 2,338), focused on diseases other than LBP (n = 346), involved animal subjects (n = 17), or did not include TEA as the main intervention (n = 60). Moreover, the full text versions of 10 articles could not be accessed, and 1 solely focused on the cost-effectiveness of TEA and had missing test results on clinical efficacy; these articles were also excluded. This process led to the final selection of nine RCTs (Fig. 1).

2. Analysis of selected articles

1) Year of publication

Three of the selected studies (44.44%) were published in 2018 [10-12]: one (11.11%) was published in 2019 [13], and three (33.33%) were published in 2020 [14-16]. Moreover, one (11.11%) was published in 2021 [17], whereas another (11.11%) was published in 2022 (Table 2) [18].

Table 2 . Basic characteristics of the articles.

Study	Disease	Sample size	Treatment	Treatment frequency (total sessions)
Goo (2022) [18]	HIVD	E: 35	E: TEA	E: 1/wk for 8 wk (8 sessions)
		C: 35	C: STEA	C: 1/wk for 8 wk (8 sessions)
Fan (2021) [17]	HIVD	E: 47	E: TEA + FN	E: 1/wk for 3 wk (3 sessions)
		C: 48	C: FN	C: 30 min, 1/wk for 3 wk (3 sessions)
Back (2020) [16]	Chronic nonspecific LBP	E: 42	E: TEA + exercise	E: 1/wk for 4 wk (4 sessions)
		C: 42	C: exercise	C: 30 min, 2/wk for 4 wk (8 sessions)
Sung (2020) [15]	CLBP	E: 19	E: TEA + AT	E: 1/wk for 8 wk (8 sessions)
		C: 19	C: AT	C: 20 min, 2/wk or 8 wk (16 sessions)
Yang (2020) [14]	Lumbago	E: 36	E: TEA	E: 1/2 wk for NR (2 sessions)
		C: 36	C: AT	C: 20 min, 7/2 wk for NR (14 sessions)
Lee (2019) [13]	CLMS	E: 35	E: TEA	E: 1/10 d for NR (2 sessions)
		C: 35	C: AT	C: 30 min, 5/wk for NR (15 sessions)
Cheng (2018) [12]	HIVD	E: 165	E: TEA	E: 1/15 d for NR (3 sessions)
		C: 165	C: EA	C: 30 min, 1/d for NR (15 sessions)
Lee (2018) [11]	CLBP	E: 20	E: TEA	E: 1/2 wk for 8 wk (4 sessions)
		C: 20	C: EA	C: 20 ± 5 min, 2/wk for 8 wk (16 sessions)
Mei (2018) [10]	Lumbar muscle strain	E: 30	E: TEA	E: 1/10 d for NR (3 sessions)
		C: 30	C: EA	C: 30 min, 1/3 d for NR (10 sessions)

HIVD, hernia of intervertebral disc; E, evaluation; C, control; TEA, thread embedding acupuncture; STEA, sham thread embedding acupuncture; FN, fire needle; LBP, low back pain; CLBP, chronic low back pain; AT, acupuncture; NR, not reported; CLMS, chronic lumbar muscle strain; EA, electroacupuncture..

2) Type of disease

All studies addressed the types of LBP. Three studies (33.33%) focused on HIVDs [12,17,18], whereas two (22.22%) focused on chronic low back pain (CLBP) [11,15]. Chronic nonspecific LBP [16], lumbago [14], lumbar muscle strain [10], and chronic lumbar muscle strain (CLMS) [13] were each examined in one study.

3) Sample size

Two studies (22.22%) had less than 50 participants [11,15], 6 (66.66%) had 50–100 participants [10,13,14,16-18], and one (11.11%) had more than 100 participants [12]. The total number of participants in the selected studies was 859, with an average of 95.4 (Table 2).

4) Treatment interventions

In every study, TEA was used in the experimental group. In the control groups, three studies (33.33%) used AT [13-15], three (33.33%) used electroacupuncture (EA) [10-12], and one (11.11%) used sham thread embedding acupuncture (STEA) [18], fire needle (FN) [17], and exercise therapy [16]. Furthermore, one study included AT [15], FN [17], and exercise therapy [16] concurrently with TEA in the experimental group (Table 2).

5) Treatment frequency and duration

With regard to the TEA treatment frequency and duration, the participants of four studies (44.44%) were treated once a week; the treatment duration was eight weeks in two studies [15,18], three weeks in one study [17], and four weeks [16] in one study. The participants in two studies (22.22%) were treated every two weeks, and the treatment duration was four [14] or eight [11] weeks. The participants in two other studies (22.22%) were treated every 10 days, and the treatment duration was 20 [13] or 30 [10] days. The participants in one study (11.11%) were treated every 15 days in a 45-day period [12].

6) Insertion point and depth of thread embedding

The TEA insertion points were primarily in the surrounding soft tissues, acupoints, and Ashi points. The most commonly used acupoints, EX-B2 [12,14,15] and BL23 [10,14,16], were used in three studies. In two studies (22.22%), the insertion points were selected based on anatomical locations and illustrations instead of acupoints [11,15]. Moreover, two studies (22.22%) used distal acupoints [11,12].

Sung et al. [15] used bone landmarks as reference points for the insertion points from L4 to the sacrum area and provided anatomical illustrations. Lee et al. [11] used 11 locations in the lumbar area, including intrinsic muscles, multifidus muscles, lumbar erector spinae, and iliolumbar ligament locations, and six locations in the abdominal area, including the rectus abdominis, iliacus, and inguinal ligament locations, and provided counts, directions, and descriptions of the insertions in tables and anatomical illustrations. Cheng [12] used acupoints around the navel, such as Geon-gwae, Gon-gwae, Jin-gwae, and Son-gwae, based on the five elements of tai chi and Bagua acupuncture, but did not specify the exact locations. Goo et al. [18] used 23 acupoints without specifying their exact locations [18]. Three studies (33.33%) also mentioned the insertion depths, which ranged from 2 cm to 4 cm (Table 3) [14-16].

Table 3 . Analysis of thread embedding acupuncture.

Study	Disease	Treatment site	Insertion depth (cm)	Thread material, length	Needle size
Goo (2022) [18]	HIVD	23 acupoints	NR	PDO, 40 or 60 mm	29 gauge
Fan (2021) [17]	HIVD	BL40, BL60, GB30, and GB34	NR	NR	NR
Back (2020) [16]	Chronic nonspecific LBP	(Bilateral) BL23 BL28, EX-B6, and GV14	2.5–3.5	PDO, 2/10 mm	0.9 diameter
Sung (2020) [15]	CLBP	(Bilateral) EX-B2 of the L4–5 and L5–S1	4	PDO, 40 mm	29 gauge
		3 to 4 cm of the L3 and S1 spinous processes toward L1
		3 to 4 cm of the L4 transverse process toward the iliac crest
Yang (2020) [14]	Lumbago	Bilateral BL23, L4-5 EX-B2, EX-B8	2	PDO, 1 cm	8 ho
Lee (2019) [13]	CLMS	(Direction where pain is observed) BL23, BL25, EX-B7	NR	PDO or cagut,1 cm	9 ho
Cheng (2018) [12]	HIVD	Geon-gwae, Gon-gwae, Jin-gwae, Son-gwae, EX-B2, GB30, GB31, GB34, BL37, and BL57	NR	PDO 2 cm	NR
Lee (2018) [11]	CLBP	17 acupoints (11 on the lumbar region, 6 on the abdomen)	NR	PDO, 50 or 90 mm	29 gauge
Mei (2018) [10]	Lumbar muscle strain	BL23 BL25, and Ashi points	NR	PDO, 15–20 mm	7 or 8 ho

HIVD, hernia of intervertebral discs; NR, not reported; PDO, polydioxanone; LBP, low back pain; CLBP, chronic low back pain; CLMS, chronic lumbar muscle strain..

7) Types and sizes of threads and needles

In seven studies (77.77%), the type of thread used was PDO [10-13,15,16,18], whereas one of these studies (11.11%) also used catgut [13]. Yang [14] described a thread used as an absorbable suture without specifying its name, which was classified as PDO. One study (11.11%) did not specify the TEA thread material used [17].

The thread length ranged from 10 to 90 mm. Five studies (55.55%) [10,12-14,16] used threads with a length of 1–2 cm, and one (11.11%) [15] used a thread with a length of 4 cm. Two studies (22.22%) used two different thread lengths; Lee et al.’s [11] threads had a length of 50–90 mm, while Goo et al.’s [18] threads had a length of 40–60 mm. One study (11.11%) did not specify the thread length used [17].

Three studies published in South Korea used 29-gauge needles for TEA [10,15,18]. Two studies (22.22%) [12,17] did not provide needle specification details. In China, “ho” is commonly used as the standard unit for needles. Four studies (44.44%) used needles ranging from 7 ho to 9 ho [11,13,14,16]. On the other hand, two studies (22.22%) used eight ho needles (Table 3) [10,14].

8) Evaluation indexes

A total of 14 evaluation indexes were used in the selected studies, and the number of indexes used in each study ranged from 2 to 5. The visual analogue scale (VAS) was used in all nine studies, whereas three studies used the VAS as the primary outcome [11,15,18]. The Oswestry Disability Index (ODI) was used in five studies (55.55%). Effective rates were used in three studies. The Roland and Morris Disability Questionnaire (RMDQ), European quality of life 5-dimensions (EQ-5D), EuroQol visual analogue scale (EQ-VAS), Japanese Orthopaedic Association (JOA) score, and present pain intensity scale (PPI) were each used twice. The finger-to-floor distance (FFD), minimal clinically important difference (MCID), Chinese Oswestry Disability Index (CODI), lumbar flexor and extensor muscle endurance test, spinal activity score, and lumbar mobility score were each used once (Table 4).

Table 4 . Results and adverse events.

Study	Disease	Evaluation index	Results	Adverse events
Goo (2022) [18]	HIVD	1. VAS (LBP)	1. S-E, S-C (p < 0.05), S-D after 16 wk (p < 0.05)	NR
		2. VAS (RP)	2. NS-D, S-E, S-C (p < 0.05)
		3. ODI	3. NS-D, S-E, S-C (p < 0.05)
		4. RMDQ	4. NS-D, S-E, S-C (p < 0.05)
		5. EQ-5D	5. NS-D, S-E, S-C (p < 0.05)
		6. EQ-VAS
Fan (2021) [17]	HIVD	1. JOA	1. S-D (p < 0.05), S-E (p < 0.05), S-C (p < 0.05)	NR
		2. VAS	2. S-D (p < 0.05), S-E (p < 0.05), S-C (p < 0.05)
		3. Effective rate	3. S-D (p < 0.005) (E: 95.83/C: 76.6)
Back (2020) [16]	Chronic nonspecific LBP	1. VAS	1. S-D (p < 0.05), S-E (p < 0.01), S-C (p < 0.01)	NR
		2. ODI	2. S-D (p < 0.05), S-E (p < 0.01), S-C (p < 0.01)
		3. FFD	3. S-D (p < 0.05), S-E (p < 0.01), S-C (p < 0.01)
		4. Lumbar flexor and extensor muscle endurance tests	4. NS-D, S-E (p < 0.05), S-C (p < 0.05)
		5. Effective rate	5. S-D (p < 0.05) (E: 97.6/C: 85.7)
Sung (2020) [15]	CLBP	1. VAS	1. S-E, S-C (p < 0.001), S-D (p = 0.013)	NR
		2. MCID	2. S-D (p < 0.05)
		3. RMDQ	3. NS-D, S-E (p < 0.001), S-C (p < 0.05)
		4. EQ-5D	4. NS-D, S-E (p < 0.01), S-C (p < 0.002)
		5. EQ-VAS	5. NS-D, S-E (p < 0.001), S-C (p < 0.05)
Yang (2020) [14]	Lumbago	1. NRS	1. S-D (p = 0.001), S-E (p < 0.001), S-C (p < 0.001)	NR
		2. ODI	2. S-D (p = 0.002), S-E (p < 0.001), S-C (p < 0.001)
Lee (2019) [13]	CLMS	1. VAS	1. NS-D, S-E (p < 0.05), S-C (p < 0.05), S-D after 1 mo (p < 0.05)	NR
		2. ODI	2. NS-D, S-E (p < 0.05), S-C (p < 0.05), S-D after 1 mo (p < 0.05)
		3. Spinal activity score	3. NS-D, S-E (p < 0.05), S-C (p < 0.05), S-D after 1 mo (p < 0.05)
Cheng (2018) [12]	HIVD	1. VAS	1. S-E, S-C (p < 0.001), S-D (p < 0.05)	NR
		2. JOA	2. S-E, S-C (p < 0.001), S-D (p < 0.05)
		3. Effective rate	3. S-D (p < 0.05) (E: 93.33/C: 83.33)
Lee (2018) [11]	CLBP	1. VAS	1. S-D after 8 wk (p = 0.01), S-E, S-C (p < 0.001)	E: none
		2. PPI	2. NS-D, S-E, S-C (p < 0.001)	C: itching (n = 2)
		3. ODI	3. NS-D, S-E, S-C (p < 0.001)
		4. SF-MPQ	4. NS-D, S-E, S-C (p < 0.001)
Mei (2018) [10]	Lumbar muscle strain	1. VAS	1. NS-D, S-E (p < 0.01), S-C (p < 0.01), S-D after 1 mo (p < 0.01)	NR
		2. PPI	2. NS-D, S-E (p < 0.01), S-C (p < 0.01), S-D after 1 mo (p < 0.01)
		3. CODI	3. NS-D, S-E (p < 0.01), S-C (p < 0.01), S-D after 1 mo (p < 0.01)
		4. Lumbar mobility score	4. NS-D, S-E (p < 0.01), S-C (p < 0.01), S-D after 1 mo (p < 0.01)

HIVD, hernia of intervertebral discs; VAS, visual analogue scale; LBP, low back pain; RP, radiating pain; ODI, Oswestry Disability Index; RMDQ, Roland and Morris Disability Questionnaire; EQ-5D, European quality of life 5-dimensions; S-E, significant difference in the experiment group compared with that before treatment; S-C, significant difference in the control group compared with that before treatment; S-D, significant difference between the two groups after treatment; NS-D, not statistically different between the two groups after treatment; JOA, Japanese Orthopaedic Association; FFD, finger-to-floor distance; CLBP, chronic low back pain; MCID, minimal clinical important difference; NRS, numeric rating scale; CLMS, chronic lumbar muscle strain; PPI, present pain intensity scale; SF-MPQ, Short-Form McGill Pain Questionnaire; E, evaluation; C, control; NR, not reported..

9) Treatment outcomes

There were three studies on HIVD. Goo et al. [18] used VAS to evaluate HIVD by dividing the condition into two categories: LBP and radiating pain. They also used the ODI, RMDQ, EQ-5D, and EQ-VAS. The results for all five evaluation indices revealed significant effects in the experimental and control groups after 8 weeks of treatment (p < 0.05); however, no significant difference was observed between the groups. Moreover, Fan and Zhou’s [17] VAS and JOA results showed a significant effect in each group after 3 weeks of treatment (p < 0.05). Furthermore, the effective rates were 95.83% in the experimental group and 76.6% in the control group, indicating a significant difference between the groups (p < 0.05). Meanwhile, Cheng’s [12] VAS, JOA, and effective rate results revealed significant effects in the experimental and control groups after 30 days of treatment (p < 0.001) as well as significant differences between the groups (p < 0.05).

Two studies have focused on CLBP. Sung et al. [15] applied five evaluation indices. The VAS results revealed significant effects in the experimental and control groups after 8 weeks of treatment (p < 0.001), and a significant difference was observed between the groups (p = 0.013). The MCID results also showed significant differences between the groups (p < 0.05). The RMDQ, EQ-5D, and EQ-VAS results showed significant effects in the experimental and control groups after 8 weeks, but did not reveal any significant differences between the groups. Lee et al. [11] used VAS, SF-MPQ, PPI, and ODI and found significant effects in the experimental and control groups after 8 weeks of treatment (p < 0.001). However, only the VAS results indicated a significant difference between the groups (p = 0.01).

In a study of lumbar muscle strain, Mei [10] used VAS, PPI, CODI, and the lumbar mobility score and found significant effects in the experimental and control groups after 8 weeks of treatment (p < 0.01). No significant difference was observed between the groups.

Back et al.’s [16] research on chronic nonspecific LBP revealed significant effects in the experimental and control groups after 4 weeks of treatment (p < 0.01) in all five indices, and a significant difference was observed between the groups (p < 0.05). Moreover, the lumbar flexor and extensor muscle endurance tests showed significant effects in the experimental and control groups (p < 0.05) after 4 weeks, but no significant difference was observed between the groups. Furthermore, the effective rate was 97.6% in the experimental group and 85.7% in the control group (p < 0.05).

Yang [14] used numeric rating scale (NRS) and ODI to study lumbago and found significant effects in the experimental and control groups after four weeks of treatment (p < 0.001). The NRS and ODI results revealed significant differences between the groups (p = 0.001 and p = 0.002, respectively). Lee [13] used VAS, ODI, and the spinal activity score and found significant effects in the experimental and control groups after four weeks of treatment (p < 0.05). Furthermore, each of these indexes revealed a significant difference between the groups (p = 0.001, p = 0.002, and p < 0.05, respectively) (Table 4).

10) Side effects

Only one (11.11%) of the selected studies reported adverse events [11]. However, it was not about TEA but about the incidence of pruritus after EA treatment in the control group (Table 4).

11) Assessment of the RoB in RCTs

We used the revised Cochrane RoB2 tool to assess the RoB in the selected studies (Figs. 2, 3). Although all nine studies were randomized, four did not mention the specific method of randomization [10,12,16,17]. Three studies have mentioned concealment of the order of assignment and have rated it as low risk [11,15,18]. In relation to the intended intervention, six studies showed that the intervention was a mediated therapy because the experimental group was not blinded [11,12-14,16,17]. Nonetheless, all nine studies determined that the evaluation tool for the effectiveness of the intervention was appropriate and rated it as low risk. All studies rated missing outcome data as low risk because no missing outcomes or participants were recorded. With regard to the measurement of the outcome, four studies did not report whether the outcome assessor knew about the intervention and therefore rated it as high risk [10,12-14]. Regarding the selection of the reported results, eight studies did not mention whether they were analyzed before blinding was removed, so they were rated as having some concern [10-14,16-18].

Figure 2. Risk of randomized controlled trials bias summary.

Figure 3. Risk of randomized controlled trials bias.

This study searched domestic and international databases for RCTs that were published between February 2018 and July 2023. To determine the latest trends in the use of TEA for LBP, the selected studies were examined in 12 categories with RoB.

All selected studies focused on LBP, including three on HIVD [12,17,18]. Two studies (22.22%) have examined CLBP. Sung et al. [15] studied patients experiencing CLBP over 3 months, whereas Lee et al. [11] investigated the presence of lumbar disk herniation and spinal stenosis in patients with CLBP over 6 months. The other four studies addressed lumbar muscle strain, chronic nonspecific LBP, lumbago, and CLMS; only the included cases of functional LBP; and excluded specific causes, such as infection, inflammation, tumors, and fracture [10,13,14,16]. According to imaging tests, such as computed tomography and magnetic resonance imaging, the participants in these four studies had no neurological symptoms or abnormalities.

According to the statistics from 2016, HIVD is the most common cause of LBP in people under the age of 59 due to modern society’s increasing number of office workers who sit for long hours and use computers and smartphones; this number is increasing every year [19]. Functional LBP, which occurs easily, is another common disease of the lumbar spine. Given these trends, the proportion of patients with LBP who visit traditional Korean treatment has increased, and research on this condition is actively ongoing.

Regarding sample size, one of the selected studies had 330 participants; this was the only study with more than 100 participants [13]. In RCTs with large sample sizes, the difference in the number of participants in each group is usually negligible compared with the total number of participants, which is generally acceptable. In RCTs with small sample sizes, the risk and bias could increase [20]. Considering this phenomenon, future research should include larger sample sizes and examine various cases of TEA therapy to increase the validity and minimize the bias of such studies.

In the control groups, various treatments, such as AT, STEA, EA, FN, and exercise therapy, were administered. For example, Goo et al. [18] used STEA in the control group; the thread was removed from the needle and temporarily inserted into the same insertion point as the one used in the experimental group. Moreover, Back et al. [16] used various exercise therapy activities, such as cobra pose, plank position, and limb crossing, to improve the core strength of the control group through the stimulation of the muscles around the spine and abdomen. In three studies, AT, exercise therapy, and FN were also used as concurrent treatments in the experimental group [15-17].

Although Korean traditional therapy and exercise therapy were performed in the control groups in all selected studies, it was challenging to compare these approaches with drug therapy interventions, such as oral drugs and injections, which are frequently used as conservative treatments for LBP in clinical practice, or physical therapy methods, such as interferential current therapy [21]. Thus, future research should investigate the effectiveness of TEA in clinical practice by employing interventions that are widely used in clinical practice in control groups.

Regarding treatment frequency, TEA was administered at intervals of 7–15 days with a mean interval of 10.1 days, and most of the TEA insertion points used were soft tissues, acupoints, and Ashi points around the affected area. Only three studies (33.33%) have investigated TEA depth [14-16]. To ensure the safety and reproducibility of TEA approaches, future research should focus on the insertion points, depths, and directions used.

PDO was the commonly used type of thread in seven (77.7%) of the nine selected studies [10-13,15,16,18]. All Korean studies used 29-gauge needles with a diameter of 0.33 mm and Chinese needles with diameters of 7, 8, 9, or 0.9 mm. Different standard units for the size of the needles, such as “No.” and “ho,” were used in China, and only Yang [14] described the exact size of the No. 8 needle as 0.8 × 120 mm. In China, the number of a needle is determined by their diameter. For example, Back et al. [16] used the No. 9 needle with a diameter of 0.9 mm [22]. When TEA was introduced in Korea in the 1960s, it was performed in China by performing an incision in the affected area, burying the thread, and then suturing it. In Korea, it was noted that this method makes it difficult to maintain hygiene and increases vulnerability to infection. Therefore, in the early 2000s, a self-insertion needle with an attached thread was developed; this method is currently used [23]. As a result, it is possible to use a thin needle, such as 29-gauge needles in Korea, whereas, in China, there are still some cases in which the aforementioned injection method is used. Given the differences in the thickness of the needles used in these countries, it is challenging to compare treatment effects between Korea- and China-based studies.

All studies revealed significant differences between the experimental and control groups for one or more indexes. Lee et al. [11] found that only the VAS showed significant differences between the groups after 8 weeks of treatment (p = 0.01) and after 16 weeks of follow-up (p = 0.013). None of the indexes used by Lee [13] and Mei [10] revealed a significant difference between groups immediately after the experiment; however, all index results in these studies revealed significant differences between groups 1 month after the end of the experiment (p < 0.01). Goo et al. [18] used five indices, but only the VAS of LBP showed a significant difference between the groups after 16 weeks of observation (p < 0.005), and the effect in the experimental group after 16 weeks of follow-up indicated that TEA was effective in the long-term.

Similar to the studies conducted in 2020, TEA showed a significant efficacy in treating lower back pain; however, the long-term effects of TEA were only mentioned in this study. Four studies (44.4%) showed that TEA was effective even after follow-up [10,11,13,18]. Since the long-term effects could be a strength of TEA compared with other interventions, follow-up is necessary in future studies.

In three studies, pain was subjectively evaluated by respondents in the FFD [16], lumbar flexor and extensor muscle endurance tests, spinal activity score [13], and lumbar mobility score [10] and used as an objective evaluation by comparing the results before and after treatment. Future research should increase the use of objective evaluation indexes to develop an accurate comparison of the effectiveness of TEA before and after treatment.

Only one (11.11%) of the studies mentioned adverse events and did not focus on the side effects of TEA, but on pruritus caused by EA in the control group [11]. Since the majority of studies did not report any side effects at all, future studies should include detailed descriptions of the side effects of TEA so that countermeasures can be developed and the safety of this procedure can be improved.

We used the Cochrane RoB2 tool to assess the RoB in the nine included studies, with uncertain results for studies with insufficient information. The intended intervention section was rated as low risk because the intervention could not be blinded by the practitioner, and the assessment tool for the intervention in each study was considered appropriate. The measurement of the outcome section was rated as high risk even though the evaluators in the three studies were unaware of the intervention due to insufficient information.

One limitation of the present study was that we included various types of diseases causing LBP compared with previous studies; however, the number of selected studies on each cause of LBP was insufficient. Moreover, we attempted a more detailed analysis of TEA itself than in the studies published in 2020 by including categories, such as the length of the threads, needles, insertion point, depth, and direction, but the selected studies did not fully describe the exact TEA treatment methods used. Furthermore, these studies included only one report about side effects, which should be addressed in future research to ensure the safety and reproducibility of TEA. Most of the selected studies were also conducted in a single country (i.e., China). Moreover, the inclusion and exclusion criteria limited the focus of the present study such that the case reports on the use of TEA for LBP and studies on TEA for the treatment of other diseases were not analyzed. Therefore, some of the latest trends in TEA may have been omitted from this analysis.

Overall, the present study aimed to provide the latest analysis and limitations of TEA on LBP and encourage researchers to publish studies that improve the quality of the literature on this topic.

Based on 9 studies published between February 2018 and July 2023, TEA appeared to be a safe treatment and may effectively relieve pain in LBP. The research trends of TEA on LBP were analyzed and summarized based on several categories. This study encourages future high-quality research by highlighting limitations such as an insufficient number of studies, limited publishing countries, and unreported or omitted data on TEA of the included studies.

Conceptualization: YGM, HJL, JHL, YKL, SCL, JSK. Data curation: YGM, HGL. Formal analysis: All authors. Funding acquisition: JSK. Investigation: All authors. Project administration: All authors. Supervision: All authors. Visualization: YGM, HGL. Writing – original draft: All authors. Writing – review & editing: All authors.

The authors have no conflicts of interest to declare.

This achievement was carried out with the support of the research fund of Jeahan Oriental Medical Academy and department of Korean Medicine at Daegu Haany University in 2023.

This research did not involve any human or animal experiment.