Since the first use of arthroscopy for lumbar spinal surgery in 1990 [1], endoscopic lumbar spinal surgery has evolved both technically and instrumentally [2,3]. The first and most common technique used today was the transforaminal approach, which uses the “Kambin’s triangle” formed by the transversing root, exiting root, and caudal pedicle [1]. In this technique, an endoscope is inserted into the lumbar intervertebral disc using a transforaminal approach and the ruptured disc is removed using straight, upbiting, and deflectable forceps [1]. The first of several randomized controlled trials (RCTs) was published in 1993 [4]. Although the results were promising and favored transforaminal endoscopic discectomy, it was not a general surgical technique and required specialized training [5]. Dr. Yeung systematized and standardized the surgical procedure, which aided in its increased use by spine surgeons [3,6]. As the ruptured disc was removed after resection of the disc material inside the disc, this technique is termed the inside-out technique. Ruetten et al. [2] introduced the outside-in technique in which a herniated disc is resected from the epidural space with minimal injury made to the healthy disc. Transforaminal endoscopic discectomy had initial limitations in cases of large-sized or highly migrated discs. However, these hurdles were overcome with the development of an endoscopic surgery system and evolution of surgical techniques [5,7-13]. Although the inside-out and outside-in techniques have technical differences, they do not significantly influence surgical outcomes. In this regard, all approaches made through Kambin’s triangle are considered transforaminal endoscopic lumbar discectomy (TELD) [14]. This study aimed to review evidence of TELD in systematic reviews and RCTs.

This narrative review included systematic reviews and RCTs that compared TELD with open discectomy (OD), microdiscectomy (MD), or tubular retractor–assisted microendoscopic discectomy (MED). The PubMed database was searched using the following keywords: for RCTs, ((((lumbar) AND (spine)) AND (endoscope)) AND (randomized[Title])) AND (trial[Title])); and for systematic reviews, ((((lumbar) AND (spine)) AND (endoscope)) AND (systematic[Title/Abstract]))). After the initial search, 24 RCTs and 88 systematic reviews were included. Relevant papers selected by two spine surgeons (C.H.K. and W.T.Y.) included 6 RCTs [4,5,15-18] and 4 systematic reviews [19-22].

The following issues were addressed: 1) clinical short-and long-term outcomes; 2) recurrence rate; 3) complication rate; 4) length of hospital stay; 5) anesthesia method; and 6) cost-effectiveness.

1. Short-and long-term clinical outcomes

The first RCT comparing TELD and MD was published in 1993 by Dr. Mayer et al. [4] As a preliminary study, the authors allocated 20 patients to each group [4]. Although there may be a chance of selection bias, the disappearance of sciatica (80% after TELD versus 65% after microdiscectomy) and return to the preoperative occupation (95% after TELD versus 72% after microdiscectomy) were advantages of TELD [4]. Another RCT comparing TELD (n = 30) and OD (n = 30) was published by Hermantin et al. in 1999 [5]. The clinical outcomes were satisfactory in both groups without a significant intergroup difference (97% vs. 93%, respectively) [5]. Gibson et al. [15] compared TELD (n = 70) and MD (n = 70) and showed that the Oswestry disability index (ODI) score, back/leg pain, and Short Form-36 scores were significantly improved without intergroup differences except in leg pain. Reported leg pain was lower in the TELD group than in the MD group at 2 years postoperatively (1.9 ± 2.6 vs. 3.5 ± 3.1, P = 0.002) [15]. Chen et al. published two papers under the same protocol with 1-and 2-year follow-up periods [16,17]. They randomly assigned 250 patients to TELD and MED and assessed ODI, body pain, functional status, European Quality of Life-5 dimensions, and back/leg pain. The outcomes did not differ during the 2-year follow-up period [16,17]. Tacconi et al. [18] compared TELD and MD and showed significantly lower postoperative back pain and lower median interval from surgery to autonomous mobilization after TELD versus MD. The authors analyzed postoperative magnetic resonance imaging and concluded that less damage to the posterior muscular structure enabled patients to mobilize more quickly with less back pain after TELD than after MD [18].

Nellensteijn et al. [19] first systematically reviewed previous studies that compared TELD and open MD, including 1 RCT, 7 non-RCTs, and 31 observational studies. TELD and MD showed similar clinical outcomes, and leg pain decreased by 89% and 87% after TELD and MD, respectively. The overall improvement rate did not differ significantly between TELD and MD (84% vs. 78%) [19]. Ding et al. [21] reviewed 17 retrospective studies and showed no difference in leg pain between TELD and MD (mean difference, 0.13; 95% confidence interval [CI], 0.03–0.22). However, the mean postoperative ODI score was significantly lower after TELD than after MD (mean difference, -0.59; 95% CI, -1.11 to -0.08, P = 0.02) [21]. In addition, the length of incision (mean difference, –3.74; 95% CI, –4.28 to –3.19; P < 0.00001) and amount of bleeding (mean difference, –63.66; 95% CI, –77.65 to –49.67; P < 0.00001) were significantly lower after TELD than after MD [21]. Gadjradj et al. [20] recently compared TELD and open MD in a systematic review. They included 9 (quasi) RCTs and 5 observational studies [20]. The quality of evidence was moderate, suggesting no difference in leg pain at intermediate (standard mean difference [SMD], 0.05; 95% CI, –0.10 to 0.21) or long-term (SMD, 0.11; 95% CI, 0.53–0.30) follow-up [20]. There was also moderate quality evidence suggesting no difference in functional status at intermediate (SMD, 0.09; 95% CI, 0.07–0.24) or long-term (SMD, 0.11; 95% CI, 0.24–0.45) follow-up. There was an expectation of less back pain after TELD than MD due to less trauma incurred by the back muscles [18]. Indeed, patients treated with TELD reported lower levels of back pain due to lesser damage to the muscular structure [18]. However, in intermediate-and long-term follow-up periods, Gibson et al. [15] showed no difference in back pain.

2. Complication and recurrence rates

Nellensteijn et al. [19] reported overall complication rates of 1.5% after TELD and 1% after MD. Ding et al. [21] also showed that overall complication rates that did not differ between TELD and MD (mean difference, 0.52%; 95% CI, 0.26–1.04). Zahang et al. [20] reported no difference in the incidence of complications [22]. Blood loss was significantly lower after TELD than after MD. Ding et al. [21] showed significantly lower blood loss after TELD than MD (mean difference, –63.66 cc; 95% CI, –77.65 to –49.67).

One study reported recurrence rates of 2-10% without an intergroup difference between TELD and open surgery [20]. Nellensteijn et al. [19] showed that the recurrence rate of TELD was 6.8%, while that of MD was 4.7%. Chen et al. [16] showed reported reoperation rates of 3.36% after TELD and 4.1% after MED within a 2-year follow-up. however, Gibson et al. [15] showed that the relative risk of revision after TELD was 2.62 (95% CI, 0.49–14.0) compared to MD, although the difference was not significant (P = 0.44, Fisher’s exact test).

3. Lengths of hospital stay

The mean hospital stay was significantly shorter after TELD than after MD [20]. Hermantin et al. [5] showed that the mean postoperative disability duration before the return to work was significantly longer after OD than after TELD (49 vs. 27 days). In addition, the duration of narcotics use was longer after OD than after TELD [5]. Gibson et al. [15] showed a shorter hospital stay after TELD than after MD (0.7 ± 0.7 vs. 1.4 ± 1.3 days, P < 0.001). Ding et al. [21] showed that, postoperatively, patients stayed in bed longer after MD than after TELD (mean difference, 90.19 hours; 95% CI, 73.56–106.82); thus, the mean hospital stay was longer after MD than after TELD (mean difference, 5.9 days; 95% CI, 4.59–7.21). Zhang et al. [22] reported a longer mean hospital stay after MD than after TELD (mean difference, 8.41 days; 95% CI, 6.56–10.26; P < 0.01).

4. Anesthesia methods and cost-effectiveness

No study has specifically analyzed the method of anesthesia or the cost-effectiveness of TELD.

Gadjradi et al. [20] published an updated systematic review in 2020 that showed concordant results. Leg pain, back pain, functional status, and rate of recurrence were similar between TELD and MD, but TELD featured a shorter mean hospitalization period and minimal blood loss. The included studies are summarized in Tables 1 and 2.

Table 1 . Summary of randomized controlled trials (RCTs).

Study	Study period	Treatment methods	Number of patients (TELD/other)	Mean age	Clinical outcomes	Complication rate	Recurrence rate	Hospital stay
Mayer 1993 [4]		TELD vs. MD	20/20	40	Disappearance of sciatica (80% after TELD and 65% after MD), Return to the preoperative occupation (95% after TELD, and 72% after MD)
Hermantin 1999 [5]	-	TELD vs. OD	60 (30/30)	40	Satisfactory in both groups without a significant difference 97% (TELD) vs. 93% (OD)			Significantly longer after OD (49 days) vs. after TELD (27 days)
Gibson 2017 [15]	2006-2015	TELD vs. MD	140 (70/70)	41	ODI, back pain, SF-36: significantly improved in both groups without differences Leg pain: lower in TELD than in MD at postoperative 2 years 1.9 ± 2.6 (TELD) vs. 3.5 ± 3.1 (MD) (P = 0.002)	No significant difference	RR of revision after TELD = 2.62 (95% CI, 0.49–14.0, P = 0.44)	Shorter hospital stay after TELD than after MD (0.7 ± 0.7 vs. 1.4 ± 1.3 days, P < 0.001)
Chen 2018 [17]	2013-2016	TELD vs. MED	153 (80/73)	41	ODI score, bodily pain, functional status, EQ-5D, back/leg pain: no significant difference during the 1-year follow-up period	13.75% (TELD) vs. 16.44% (MD) (P = 0.642)	Reoperation rate: 6.25% (TELD) vs. 4.11 (MED) (P = 0.818)
Chen 2020 [16]	2013-2016	TELD vs. MED	241 (119/122)	41	ODI score, bodily pain, functional status, EQ-5D, back/leg pain: no significant difference during the 2-year follow-up period.	13.44% (TELD) vs. 15.57% (MD) (P = 0.639)	Reoperation rate: 3.36% (TELD) vs. 4.1% (MED) (P = 0.167)
Tacconi 2020 [18]	2017-2019	TELD vs. MD	50 (25/25)	44	Significantly lower postoperative back after TELD than MD Less damage to the posterior muscular structure enabled patients to mobilize more quickly with less back pain after TELD than MD

MD: microdiscectomy; MED: microendoscopic discectomy; OD: open discectomy; ODI: Oswestry disability index; SF-36: short-form healthy survey indices; TELD: transforaminal endoscopic lumbar discectomy..

Table 2 . Summary of systematic reviews.

Study	Surgical methods	Number of studies	Clinical outcomes	Complication rate	Recurrence rate	Hospital stay
Nellensteijn 2010 [19]	TELD vs. MD	39 (1 RCT, 7 non-RCTs, 31 observational studies)	Similar clinical outcomes Leg pain decrease: 89% (TELD) vs. 87% (MD) The overall improvement: 84% (TELD) vs. 78% (MD)	No significant difference Overall 1.5% after TELD and 1% after MD	6.8% (TELD) vs. 4.7% (MD)
Ding 2018 [21]	TELD vs. MD	17 (all retrospective studies)	Leg pain: No significant difference between TELD and MD (MD, 0.13; 95% CI, 0.03 – 0.22). Postoperative ODI score: significantly lower after TELD than MD (MD, –0.59; 95% CI, –1.11 to –0.08, P = 0.02) Length of incision: significantly lower after TELD (MD –3.74; 95% CI –4.28 to –3.19; P < 0.00001) Intraoperative bleeding: significantly lower after TELD (MD –63.66, 95% CI –77.65 to –49.67; P < 0.00001)	No significant difference between TELD and MD (MD, 0.52%; 95% CI, 0.26 – 1.04)		Longer after MD than after TELD (MD, 5.9 days; 95% CI, 4.59 – 7.21)
Zhang 2018 [22]	TELD vs. MD	9 (5 RCTs, 4 retrospective studies)	Significantly lower intraoperative bleeding after TELD than MD	No significant difference in complication rate	No significant difference between TELD and MD (RR = 1.77, 68% CI 0.66, 4.8, P = 0.26)	Longer after MD than after TELD (MD, 8.41 days; 95% CI, 6.56 – 10.26, P < 0.01)
Gadjradj 2020 [20]	TELD vs. MD	14 (9 RCTs, 5 observational studies)	Leg pain: no significant difference at intermediate follow-up (SMD, 0.05; 95% CI, –0.10 –0.21) and at long-term follow-up (SMD, 0.11; 95% CI, 0.53 – 0.30) Functional status: no significant difference at intermediate follow-up (SMD, 0.09; 95% CI, 0.07 – 0.24) and at long-term follow-up (SMD, 0.11; 95% CI, 0.24 – 0.45)

MD: microdiscectomy; MED: microendoscopic discectomy; OD: open discectomy; ODI: Oswestry disability index; RCT: randomized controlled trials; SF-36: short-form healthy survey indices; TELD: transforaminal endoscopic lumbar discectomy..

The current study reviewed previous systematic reviews and RCTs to summarize clinical outcomes, complications, and length of hospital stay. The clinical outcomes and rates of complications did not differ between TELD and OD, MD, or MED; however, the hospital stay was unanimously shorter after TELD than after MD. Various factors, such as less postoperative pain, less injury to the posterior spinal muscle, and less dependency on narcotics, may have shortened the length of hospital stay.

1. Previous RCTs and systematic reviews

Many systematic reviews and RCTs compared lumbar endoscopic with conventional surgery, but only a few studies have specifically focused on TELD versus conventional surgical techniques such as OD, MD, or MED [4,5,15-21]. The first prospective study comparing TELD and MD was published in 1993 by Mayer et al. [4] This study showed a successful result of TELD and might have triggered further studies; in fact, several RCTs and systematic reviews were serially published. Hermantin et al. [5] performed an RCT (30 patients in each group) to compare TELD and OD. The clinical success rate exceeded 95% in both groups, but the mean hospital stay was much shorter in the TELD group [5]. Gibson et al. [15] showed similar clinical outcomes for TELD and MD, but in contrast to previous RCTs, lower leg pain was observed in the TELD group at 2 years postoperative. However, a high reoperation rate was a concern, although the intergroup difference was insignificant [15]. This study may be the only one to emphasize the possibility of different recurrence rates between TELD and MD (RR of revision after TELD = 2.62; 95% CI, 0.49 after TELD. Kim et al. [23] analyzed national data and commented on the risk of early reoperation after TELD versus open discectomy. This issue requires further analysis in a large number of patients to overcome the type II error. Nonetheless, the rapid recovery and short hospital stay may offset the potentially high reoperation rate [15].

TELD was not strongly recommended in the systematic review of Nellensteijn et al. [19]. However, many subsequent systematic reviews did not specifically include TELD, and further evidence to strongly support TELD was not published until 2018. Ding et al. [21] performed a meta-analysis of retrospective studies, but the inherent selection bias of retrospective studies and heterogeneity among them did not improve the quality of evidence. Zhang et al. [22] performed a systematic analysis that compared TELD with MD, but their study included both TELD and interlaminar endoscopic lumbar discectomy (IELD) despite the title of the paper.

Gadjradj et al. [20] recently updated a systematic review and included only TELD. The results showed no difference in clinical outcomes [20]. The authors also commented that differences in lifestyles in various cultures, health insurance status, measuring tools for patient-reported outcomes, and time differences made the studies heterogeneous [20]. In addition, the expectation of better outcomes amid the higher cost of TELD than MD may have influenced the postoperative outcomes [20]. Therefore, high-quality evidence for TELD could not be retrieved from previous RCTs and systematic reviews.

There may be a question of whether lumbar endoscopic discectomy (ED), including TELD and IELD, could replace MD. Mathu et al. [24] answered the question in a systematic review. Superiority was established at the 95% CI for ED versus MD in terms of functional outcomes, such as ODI score (P = 0.008), duration of surgery (P = 0.023), and length of hospital stay (P < 0.001), but there was significant heterogeneity [24]. Other outcomes, such as the visual analog scale score for back pain (P = 0.860) and leg pain (P = 0.495), MacNab classification (P = 0.097), recurrence rates (P = 0.993), reoperation rates (P = 0.740), and return-to-work period (P = 0.748), did not differ significantly between ED and MD [24].

Although we tried to find evidence of the method of anesthesia and its cost-effectiveness, these issues were not addressed in the RCTs or systematic reviews. Regarding the method of anesthesia, Perez-Roman et al. [25] compared spinal and general anesthesia for lumbar discectomy. They showed that spinal anesthesia was superior in operation/anesthesia time, overall complication rate, postoperative pain score, need for postoperative analgesia, and length of hospital stay [25]. In this regard, the advantage of local or spinal anesthesia may be applicable to TELD, but this was not supported by evidence [26].

Cost is an important issue in accepting TELD as a standard surgical procedure [27-30]. However, direct and indirect medical expenses vary among countries, and health insurance affects cost-effectiveness. Nevertheless, this issue requires specific addressing before TELD can be accepted as the standard surgical technique.

2. Limitations

The present study reviewed and summarized previous studies, but it has several limitations. First, although there have been several RCTs and systematic reviews, the strength of the recommendation was moderate due to a lack of robust studies [20]. Therefore, instead of repeating the same systematic review, we serially presented previous studies to facilitate access to the evidence created by the previous RCTs and systematic reviews. Second, TELD and other open surgical techniques differed in surgical trajectory and blinding the surgical techniques was impossible [15]. Therefore, the expectation of better outcomes and the high cost of TELD may have influenced the postoperative outcomes, but these factors were not controlled in previous RCTs [20]. Finally, although many RCTs and systematic reviews recommended further studies to improve the level of evidence, a strong recommendation could not be retrieved, probably due to similar outcomes between TELD and MD with the current instruments measuring patient-reported outcomes. There might have been a floor effect of the instruments, and different clinical outcomes might not have been detected [21,31]. We suggest the development of a new instrument to overcome the possible floor effect of the current instruments [31]. Despite these limitations, we hope that this review will provide readers with a better perspective of TELD.

The clinical outcomes of TELD and OD, MD, and MED seemed similar with a moderate quality of evidence. However, the evidence was not robust, and further RCTs are required to improve its quality.

This work was supported by the New Faculty Startup Fund from Seoul National University; and the SNUH Research Fund [grant number 04-2021-0540]. The authors appreciate the statistical advice provided by the Medical Research Collaborating Center at Seoul National University Hospital.

No potential conflict of interest relevant to this article was reported.

None.