Een operatie is de standaardbehandeling bij patiënten met CRS wanneer pijn en/of uitval van gevoel en/of kracht in de arm aanhouden met congruente MRI-afwijking. Vaak kent een CRS echter een voorspoedig spontaan herstel (Lyer, 2016). Over het algemeen vindt men dat alleen tot operatieve therapie moet worden overgegaan als conservatief beleid gefaald heeft. Een operatie kan gepaard gaan met complicaties en hoge kosten. Op dit moment is het onduidelijk of CRS beter te behandelen is door middel van chirurgische decompressie van de zenuwwortel of door middel van niet opereren in het algemeen. Wat levert een operatie precies op voor een patiënt en hoe staat dit in verhouding tot complicaties en kosten? Deze module gaat in op de effectiviteit van chirurgische behandeling. Omtrent de timing van chirurgische interventie verwijst de werkroep naar de submodule ‘Timing’.

1.1. Pain (short-, mid- and long term) (critical)

1.2 Quality of life (critical)

1.3 Functioning (short-, mid- and long term) (critical)

1.4 Patient satisfaction (short-, mid- and long term) (important)

1.5 Complications (important)

1.6 Re-operation; 1.7 Return to work, tingling, and adjacent segment disease

2.1 Pain (short-, mid- and long term) (critical)

Quality of life, functioning, patient satisfaction, complications, return to work, tingling, re-operation, and adjacent segment disease

MSQ, Medication Quantitative Scale; NDI, Neck Disability Index; NRS, Numeric Rating Scale; VAS, Visual Analogue Scale

Results

1. Surgery with physiotherapy vs. physiotherapy alone; and surgery vs. physiotherapy

1.1 Arm-pain (critical)

Three studies reported on pain (Enquist, 2013; Enquist, 2017; Persson, 1997). Results are presented in three post-intervention terms: 1.1.1 Short term: until 6 months), 1.1.2. Midterm: >6 months to 12 months, 1.1.3. long term: >12 months to 8 years. In the study by Persson (1997), type of pain (neck- and/or arm-pain) was not otherwise specified. A brief overview of the main characteristics is provided in Table 2.

Table 2. Overview on post-intervention terms

NRS: Numeric rating scale; VAS: Visual Analogue Scale

1.1.1 Short term (post treatment: up to 6 months)

Two studies reported on pain up to six months (Enquist, 2013; Persson, 1997).

• Arm pain (surgery with physiotherapy versus physiotherapy alone)

Enquist (2013) reported results for arm pain reduction using a VAS-scale ranging from 0-100mm. Six months after initiation of the intervention, mean reduction (within group mean change from baseline) was 21.1 (SD 38.4) in the group receiving surgery with physiotherapy and 16.0 (SD 38.6) in the group receiving only physiotherapy. This resulted in a mean difference of 5.1 (95%CI -13.9 to 24.1). Results are depicted in Figure 1.

• Pain (surgery versus physiotherapy)

Persson (1997) reported results for mean current pain intensity using a VAS-scale ranging from 0-100mm. Four months after intervention, mean score was 27.0 (SD 23.0) in the group receiving surgery and 41.0 (SD 28.5) in the group receiving physiotherapy. This resulted in a mean difference of 14.00 (95%CI -0.19 to -27.81). Results are depicted in Figure 2.

Figure 1. Mean reduction for armpain, surgery with physiotherapy versus physiotherapy alone (short term follow-up)

Figure 2. Mean differences for (arm-) pain, surgery versus physiotherapy (short term follow-up)

1.1.2 Midterm (post-treatment >6 months to 12 months)

Two studies reported on pain after 6 months to 12 months (Enquist, 2013; Persson, 1997).

• Arm pain (Surgery with physiotherapy versus physiotherapy alone)

In the study by Enquist (2013), twelve months after initiation of the intervention, mean reduction was 25.1 (SD 40.9) in the group receiving surgery with physiotherapy and 20.3 (SD 41.0) in the group receiving only physiotherapy. This resulted in a mean difference of 4.80 (95%CI -15.43 to 25.03). Results are depicted in Figure 3.

• Pain (surgery versus physiotherapy)

In the study by Persson (1997), sixteen months after initiation of the intervention, mean score was 30 (SD 28.1) in the group receiving surgery and 39 (SD 25.8) in the group receiving physiotherapy. This resulted in a mean difference of 9.00 (95%CI -5.39 to 23.39). Results are depicted in Figure 4.

Figure 3. Mean difference for arm-pain reduction, surgery with physiotherapy versus physiotherapy alone (mid-term follow-up)

Figure 4. Mean difference for pain, surgery versus physiotherapy (midterm follow-up)

1.1.3 Long term (>12 months to 8 years)

Two studies reported on arm-pain after 6 months to 12 months (Enquist, 2013; Enquist, 2017). Results are depicted in Figure 5.

• Arm pain (surgery with physiotherapy versus physiotherapy alone)At 24 months after intervention (Enquist, 2013), mean reduction of arm-pain was 18.1 (SD 48.0) in the group receiving surgery with physiotherapy and 20.5 (SD 48.3) in the group receiving physiotherapy only. This resulted in a mean difference of -2.40 (95%CI -26.17 to 21.37).
  At 5-8 years after intervention (Enquist, 2017), mean reduction of arm pain was 33.0 (SD 40.9) in the group receiving surgery with physiotherapy and 19.0 (SD 33.3) in the group receiving physiotherapy only. This resulted in a mean difference of 14.00 (95%CI -4.44 to 32.44). Results are depicted in Figure 5.

Figure 5. Mean difference for arm-pain, surgery with physiotherapy versus physiotherapy alone (long term follow-up)

1.2 Quality of life (critical)

One study reported on quality of life (Enquist, 2017). Quality of life was measured using both the EQ-5D (0-1) and the EQ-VAS (0-100), with higher scores indicating better quality of life.

Five to eight years after intervention, mean score increased on the EQ-5D was 0.29 (SD 0.43) in the group receiving surgery with physiotherapy and 0.14 (SD 0.34) in the group receiving physiotherapy alone. This resulted in a mean difference of -0.15 (95%CI -0.05 to 0.35). Results are depicted in Figure 6.

Five to eight years after intervention, mean score on the EQ-VAS was 29 (SD 26.8) in the group receiving surgery with physiotherapy and 30 (SD 28.9) in the group receiving physiotherapy alone. This resulted in a mean difference of 4.00 (95%CI -10.23 to 18.23). Results are depicted in Figure 7.

Figure 6. Mean difference for quality of life (EQ-5D), surgery with physiotherapy versus physiotherapy alone

Figure 7. Mean difference for quality of life (EQ-VAS), surgery with physiotherapy versus physiotherapy alone

1.3 Functioning (critical)

Two studies reported on functioning (Enquist, 2013; Enquist, 2017). Functioning was measured using the Neck Disability Index (0-50), with higher scores indicating worse disability. A percentage of the reduction on the NDI was provided (0-100).

• 1.3.1 Short term (post treatment: up to 6 months)

One study reported on functioning up to six months (Enquist, 2013). Six months after initiation of the intervention, mean reduction was 12.1% (SD 16.9) in the group receiving surgery with physiotherapy and 7.7% (SD 16.9) in the group receiving only physiotherapy. This resulted in a mean difference of 4.40 (95%CI -3.95 to 12.75). Results are depicted in Figure 8.

• 1.3.2 Midterm (post-treatment >6 months up to 12 months)

One study reported on functioning after 6 months to 12 months (Enquist, 2013).

Twelve months after initiation of the intervention, mean reduction was 13.9% (SD 20.2) in the group receiving surgery with physiotherapy and 7.1% (SD 20.2) in the group receiving only physiotherapy. This resulted in a mean difference of 6.80 (95%CI -3.18 to 16.78). Results are depicted in Figure 9.

• 1.3.3 Long term (>12 months to 8 years)

In the study by Enquist (2013), 24 months after intervention, mean score was 14.2 (SD 23.4) in the group receiving surgery with physiotherapy and 11.5 (SD 23.6) in the group with physiotherapy alone. This resulted in a mean difference of 2.7 (95%CI -8.91 to 14.31).

In the study by Enquist (2017), 5-8 years after intervention, mean reduction was 21.0% (SD 19.1) in the group receiving surgery with physiotherapy and 11.0% (SD 19.4) in the group with physiotherapy alone. This resulted in a mean difference of 10.00 (95%CI 0.49 to 19.5). This difference was clinically relevant. Results are depicted in Figure 10.

Figure 8. Mean difference for functioning (NDI), surgery with physiotherapy versus physiotherapy alone (short term follow-up)

Figure 9. Mean difference for functioning (NDI), surgery with physiotherapy versus physiotherapy alone (midterm follow-up)

Figure 10. Mean difference for functioning (NDI), surgery with physiotherapy versus physiotherapy alone (long term follow-up)

1.4 Patient satisfaction (important)

In the study by Enquist (2013) and Enquist (2017), patient satisfaction was measured using the Patient’s Global Assessment. Patients were asked whether after treatment, their neck/arm problems were much better, better, unchanged, worse or much worse. This score was dichotomised into better (defined by “better” or “much better”) and worse. Results are depicted in Table 3 and Figure 11.

Figure 11. Risk ratios for a better score on the Patient’s Global Assessment, surgery with physiotherapy versus physiotherapy alone

1.5 Complications (important)

One study reported on complications (Engquist, 2013). No surgery related complications (e.g. onset of neurological deficit, thromboembolism, unexpected bleeding, infection) were reported. These results could not be evaluated using the GRADE-methodology.

1.6 Re-operations (important)

Enquist 2013 reported no re-operations. After 5-8 years (Enquist, 2017), no participants from the surgery group needed another operation. In the non-surgery group, 8 participants underwent surgery.

1.7 Return to work, tingling, re-operation and adjacent segment level disease (important)

None of the RCTs assessed the effect of surgery on these outcomes in patients with cervical radiculopathy.

1. Level of evidence of the literature

1.1 Arm-pain (critical)

• Short term (post treatment: up to 6 months)

The level of evidence regarding the outcome measure pain (short term) started as high because it was based on RCTs and was downgraded by three levels to very low because lack of blinding in all studies (-1, risk of bias), and crossing of both thresholds of clinical decision-making (Enquist, 2013) (-2, imprecision).

• Midterm (post-treatment >6 months to 12 months)

The level of evidence regarding the outcome measure pain (midterm) started as high because it was based on RCTs and was downgraded by three levels to very low because lack of blinding in all studies (-1, risk of bias), and crossing of both thresholds of clinical decision-making (Enquist, 2013) (-2, imprecision).

• Long term (>12 months to 8 years)

The level of evidence regarding the outcome measure pain (long term) started as high because it was based on RCTs and was downgraded by three levels to very low because lack of blinding in all studies (-1, risk of bias), and crossing of both thresholds of clinical decision-making (Enquist, 2013) (-2, imprecision).

1.2 Quality of life (critical)

The level of evidence regarding the outcome measure quality of life started as high because it was based on RCTs and was downgraded by three levels to very low because lack of blinding in all studies (-1, risk of bias), and crossing of both thresholds of clinical decision-making (-2, imprecision).

1.3 Functioning (critical)

•  1.3.1 Short term (post treatment: up to 6 months)

The level of evidence regarding the outcome measure functioning (short term) started as high because it was based on RCTs and was downgraded by two levels to low because lack of blinding (Enquist, 2013) (-1, risk of bias), and crossing of one threshold of clinical decision-making (-1, imprecision).

• 1.3.2 Midterm (post-treatment >6 months to 12 months)

The level of evidence regarding the outcome measure functioning (midterm) started as high because it was based on RCTs and was downgraded by two levels to low because lack of blinding (Enquist, 2013) (-1, risk of bias), and crossing of one threshold of clinical decision-making (-1, imprecision).

• 1.3.3 Long term (>12 months to 8 years)

The level of evidence regarding the outcome measure functioning (long term) started as high because it was based on RCTs and was downgraded by two levels to low because lack of blinding in all studies (-1, risk of bias), and crossing of one threshold of clinical decision-making (Enquist, 2013) (-1, imprecision).

1.4 Patient satisfaction (important)

• 1.4.1 Short term (post treatment: up to 6 months)

The level of evidence regarding the outcome measure patient satisfaction (short term) started as high because it was based on RCTs and was downgraded by two levels to low because lack of blinding in all studies (-1, risk of bias), and crossing of one threshold of clinical decision-making (Enquist, 2013) (-1, imprecision).

• 1.4.2 Midterm (post-treatment >6 months to 12 months)

The level of evidence regarding the outcome measure patient satisfaction (midterm) started as high because it was based on RCTs and was downgraded by two levels to low because lack of blinding in all studies (-1, risk of bias), and crossing of one threshold of clinical decision-making (Enquist, 2013) (-1, imprecision).

• 1.4.3 Long term (>12 months to 8 years)

The level of evidence regarding the outcome measure patient satisfaction (long term) started as high because it was based on RCTs and was downgraded by three levels to very low because lack of blinding in all studies (-1, risk of bias), and crossing of both thresholds of clinical decision-making (Enquist, 2013) (-2, imprecision).

1.5 Complications (important)

The level of evidence of the outcome measure complications could not be GRADED due to a lack of data.

1.6 Re-operations (important)

The level of evidence regarding the outcome measure complications started as high because it was based on RCTs and was downgraded by three levels to very low because lack of blinding in all studies (-1, risk of bias), and very few events (Enquist, 2013) (-2, imprecision).

1.7 Return to work, tingling, re-operation and adjacent segment level disease (important)

The level of evidence regarding these outcomes was not graded because of lack of data.

2. Surgery vs. cervical collar

2.1 Arm-pain (critical)

One study reported on current pain (Persson, 1997) using a VAS-scale ranging from 0-100mm. Type of pain (neck- and/or arm-pain) was not otherwise specified by the authors.

• 2.1.1 Short term (post treatment: up to 6 months)

Four months after intervention, mean score was 27.0 (SD 23.0) in the group receiving surgery and 48.0 (SD 23.2) in the group wearing a cervical collar. This resulted in a mean difference of -21.00 (95%CI -33.34 to -8.68). This difference was clinically relevant. Results are depicted in Figure 12.

• 2.1.2. Long term (>12 months to 8 years)

Sixteen months after intervention, mean score was 30.0 (SD 28.1) in the group receiving surgery and 35.0 (SD 23.6) in the group wearing a cervical collar. This resulted in a mean difference of -5.00 (95%CI -18.8 to 8.84). This difference was not clinically relevant. Results are depicted in Figure 13.

Figure 12. Mean difference for pain, surgery versus cervical color (short term follow-up)

Figure 13. Mean difference for pain, surgery versus cervical collar (midterm follow-up)

Quality of life, functioning, patient satisfaction, complications, return to work, tingling, re-operation, and adjacent segment disease

The level of evidence regarding these outcomes was not graded because of lack of data.

2. Level of evidence of the literature

2.1 Pain (critical)

• Short term (post treatment: up to 6 months)

The level of evidence regarding the outcome measure pain (short term) started as high because it was based on RCTs and was downgraded by two levels to low because lack of blinding in all studies (-1, risk of bias), and crossing of both thresholds of clinical decision-making (Persson, 1997) (-1, imprecision).

• Midterm (post-treatment >6 months to 12 months)

The level of evidence regarding the outcome measure pain (midterm) started as high because it was based on RCTs and was downgraded by two levels to low because lack of blinding in all studies (-1, risk of bias), and crossing of both thresholds of clinical decision-making (Persson, 1997) (-1, imprecision).

A systematic review of the literature was performed to answer the following question: What is the efficacy of surgical anterior decompression compared to conservative management in patients with cervical radiculopathy?

P: Patients with cervical radiculopathy

I: Surgical decompression of the nerve root (anterior microforaminotomy, ACD, ACDF, ACDP)

C: Conservative treatment (e.g. physiotherapy, cervical collar, PRF, corticosteroids);

O: Patient satisfaction, arm-pain, quality of life, return to work, tingling, functioning, complications, re-operation, adjacent segment level disease

Relevant outcome measures

The guideline development group considered arm-pain, quality of life and functioning as critical outcome measures for decision making; and patient satisfaction, return to work, tingling, complications, re-operation, and adjacent segment level disease as important outcome measures for decision making.

The working group defined the outcome measures as follows:

• Pain: VAS
• Functioning: Neck Disability Index (NDI)
• Quality of life: SF-36 or 1-10 scale

A priori, the working group did not define other outcome measures but used the definitions used in the studies.

The working group defined a 10% difference for both continuous outcome measures and dichotomous outcome measures informing on relative risk (RR ≤ 0.91 and ≥ 1.10) as minimal clinically (patient) important differences. This decision was based on the minimal important change scores described in the article by Ostelo (2008), in accordance with the Dutch Lumbosacral Radicular syndrome guideline (NVN, 2020).

Search and select (Methods)

The databases Medline (via OVID) and Embase (via Embase.com) were searched with relevant search terms from inception until 25 April 2022. The detailed search strategy is depicted under the tab Methods. The systematic literature search resulted in 582 hits. Studies were selected based on the following criteria:

• Systematic review (searched in at least two databases, and detailed search strategy, risk of bias assessment and results of individual studies available), randomized controlled trial comparing surgical decompression of the nerve root with conservative management;
• Patients aged ≥ 18 years;
• Full-text English language publication;
• Studies including ≥ 20 patients (ten in each study arm); and
• Studies according to PICO.

Initially, 29 studies were selected based on title and abstract screening. After reading the full text, 26 studies were excluded (see the table with reasons for exclusion under the tab Methods), and three studies were included.

Results

Three studies were included in the analysis of the literature. A comprehensive overview of study characteristics is depicted in Table 1. Important study characteristics and results are summarized in the evidence table. The assessment of the risk of bias is summarized in the risk of bias table.