Surgery versus conservative management for lumbar disc herniation with radiculopathy : A systematic review and meta-analysis

Background: Lumbar disk herniation with radiculopathy (LDHR) appears to be a large and costly problem. The standard procedure regarding the best treatment for LDHR has being between surgery and conservative management. The aim of this study was to compare and summarize evidence regarding the effectiveness of surgery and conservative treatment for individuals with sciatica due to LDH. Methods: This study reviewed all literatures published on individuals with LDHR, who were managed either through surgery or conservative method. Pain and functional disability were the main outcome measures analyzed. A comprehensive search of PubMed, translating research into practice, physiotherapy evidence database (PEDro), and CINAHL was conducted from October 2011 to June 2017. Two independent researchers selected the studies and extracted the data. Methodological quality was assessed using the PEDro scale. Meta-analysis was carried out where suitable. Results: Eight studies involving (n = 1507) participants were included in the review Meta-analysis was conducted for only four studies (n = 784). The meta-analysis showed significant benefit for early surgery than conservative care (−8.01, 95% CI, −9.27–−6.72) in the short-term effect (−0.49, 95% CI, −0.7– −0.28). However, the result for long-term effect did not show any significant difference between surgery and conservative care (1.60, 95% CI, −6.85–10.05). Conclusion: This current evidence suggests that early surgery for individuals with LDH with radiculopathy is better than conservative care in the short-term without any long-term difference. The results of this review should be interpreted with caution as the populations of the included studies were largely heterogeneous.


INTRODUCTION
Low back pain (LBP) appears to be a major problem globally, with the highest prevalence during the middle age life span (1).It leads to physical impairment and poor quality of life for individuals, as well as increased absenteeism and early retirement (2).
Lumbar disc herniation (LDH), defined as localized displacement of disc material beyond the limits of the intervertebral disc, is believed to be a major contributor to the estimated 60-80% of lifetime incidence of LBP in general population (3) and is among the most common causes for sciatica (4).
Sciatica goes together with almost 10% of cases of LBP (5) with a lifetime incidence ranging from 13% to 40% (6).Symptoms of sciatica may be very difficult to deal with because over 50% of people reporting sciatica or radiculopathy indicate a pattern of intermittent presentation, with relapsing being very common (4,6).This pattern has been estimated to increase the prevalence of long-term disability by 10% (7) and to triple the likelihood that people will seek additional medical care (8,9).Thus, the importance of identifying effective treatment strategies for sciatica has been emphasized as it is said to be associated with delayed recovery from LBP, persistent disability, and increased health-care system utilization and costs (4,8,9).
Microdiscectomy and endoscopic surgeries that are minimally invasive are the most common type of surgery used in the management of individuals with LDH with radiculopathy (LDHR).(10).However, an absolute indication for lumbar disc surgery is a progressive neurological deficit commonly associated with the cauda equina syndrome (11).In addition, Cakir et al. (12) stated that the only clear and objective indication for early surgery is the cauda equina syndrome.Furthermore, the same authors also emphasized that there is no any outstanding evidence with regard to the necessity for immediate surgery even in individuals with severe complication.Therefore, the relative indications for discectomy vary between surgeons and patients (13).
According to Ogink et al. (14), it is incumbent on clinicians to discuss the advantages, disadvantages, risks, alternatives, and estimated expected outcomes with patients before any disc surgery.Most often, the primary aim of lumbar disc surgery is to relieve the patient from pain in the leg.Other symptoms, such as back pain and possible muscle weakness in the leg, appear to be more difficult to reduce with surgery.In this regards, the general recommendation, when patients report symptoms from LDH, is to start with non-surgical treatment.A previous research (4) has mentioned that a period of 3 months was enough to show if a conservative management would be successful in the management of LDHR or not.However, they did not mention if it requires any standardization in terms of frequency and expertise as well as specificity of the type of conservative management that is administered.Thus, they mentioned that, if no or little improvement occurred during this period, then the patient would be a good candidate for surgical intervention (4).
The effectiveness of many conservative treatments for LDHR in comparison with surgery is still unclear.This has been in part due to the heterogeneity of the conservative interventions (15,16) and lack of validated outcome measures in early studies (17).A systematic review by Jacobs et al. (18) has collated the published evidence on conservative treatments for LDHR compared with surgery up to October 2009, However, the study was not able to pool results of the findings due to participants' heterogeneity.Shojania et al. (19) recommended that the average survival time of any systematic review is 5.5 years, with 23% of the reviews becoming outdated within 2 years of publication (19).There appears to be increasing and new literatures since 2009 when the last systematic review on LDHR was published.The objective of this present systematic review was to compare and summarize evidence regarding the effectiveness of surgery compared with conservative treatment for patients with LDHR and also identify who benefits more from surgery and who from conservative care.This systematic review was registered with the Prospero database with an ID number (CRD42017071624).

METHODOLOGY Evidence acquisition
The databases of PubMed, translating research into practice database, physiotherapy evidence database (PEDro), and the cochrane library were searched from June 2011 to June 2017.The MeSH criteria for PubMed search strategy was used (Table 1).In PEDro, simple search was conducted, combining search terms separately.Manual searches of the reference list was also conducted.

Study selection
Covidence trial version was used by the two independent reviewers (MSD and BB) to carry out the electronic database searches and screened the title and abstracts.Full copies of potential eligible papers were also retrieved and screened by the two independent reviewers (BB and MSD).

Exclusion criteria
This review excluded any study which participants had LDHR with known cause of the problem.These include the following: Individuals with systemic inflammatory diseases, spinal stenosis, spondylolisthesis, spine fractures, tumors, infections, or osteoporosis.

Data extraction
Data extracted from the included studies were study design, sample size, sex, age, participants, interventions, outcomes, and follow-up.Information was also retrieved directly from the study of Jacobs et al. (18).

Quality assessment
The methodological quality score of the reviewed studies is reported in Table 2. Rating of trials and risk of bias was carried out using the PEDro Methodological Quality Scale due to its high validity and reliability (20) (Appendix 1).Previous authors have shown that studies scoring ≥6 of 10 were often considered to be of high quality (20,21) (Table 2).

Data analysis
The following headings were used to extract data for the table of evidence: Author, year of publication, study population, type of interventions, design, outcome measures, results, and conclusion.Comparison was done on the same reported outcomes and all the data were pooled using RevMan 5 software.
I 2 statistic was used to assess for any statistical difference between-study heterogeneity, and any value ≥75% was considered high while ≤25% are said to be low while 50% was considered moderate heterogeneity.Funnel plots were assessed to identify the publication bias (Figures 1-3).

RESULTS
The overall search resulted in eight studies that met the inclusion criteria.Initial database search produced 257 citations, of which 10 were appropriate for full-text review.Figure 4 shows the complete study selection process.Four studies met the inclusion criteria of the present study which were not part of the 2011 review, while another four studies were drawn from the 2011 study, making a total of eight studies in the present review.

Characteristics of included studies
Table 3 a shows summary of the characteristics of the included studies with their findings.Two

Study quality and bias
The PEDro scores of the included studies ranged from 5 to 7, with a mean score of 6.8 (Table 2).All participants were randomly allocated, and all studies provided adequate results and analysis.All studies studies compared early surgery with prolonged conservative care for 6 months followed by surgery if needed (22,23).Five studies contrasted surgery with usual conservative care (15,17,18,24,25) and one study contrasted surgery with manipulation (26).concealed allocation and seven studies assessed baseline comparability.No study blinded participants, therapists, and outcome assessors.With all studies, the greatest possible source of bias was related to blinding.Four publications scored >6 (16,22,25,26) along with three (15,23,24) from the 2011 review totaling 7 studies of high quality.

Data synthesis
Due to inherent heterogeneity among the included studies, only four studies were pooled for inclusion into meta-analysis.Two studies (15,23) from the 2011 review and the other two studies (18,27) from the remaining included studies.However, meta-analysis for this review was conducted in two phases.The first phase involved the pooling of two studies (22,27) that compared early surgery with prolonged conservative care followed by surgery if needed.These studies are homogenous in participants' characteristics, interventions, and outcomes.Data for these studies were pooled for short-term (8 weeks) and long-term (52 weeks) effects on disability (Roland-Morris Disability Questionnaire), pain (visual analog scale), and global perceived recovery (7-point Likert scale).
Similarly, the second phase involved pooling the remaining two studies (15,16) that contrasted surgery with usual conservative care.These studies, however, like those in the first phase were homogenous in participants' characteristics, interventions, as well as outcomes.Data for these studies were equally pooled into meta-analysis for only long-term (2 years) effects on BP (SF-36 BP), PF (SF-36 PF), and functional disability (Oswestry Disability Index).

Short-term effect
Meta-analysis showed significant benefit for early surgery versus conservative care followed by surgery if needed for short-term disability (−3.−1.52-−1.02)(Figure 5).There was no any significant difference between groups for disability, leg pain, back pain, and global perception of recovery with a between-study heterogeneity ranging from high to negligible (I 2 = 0%, 0%, 0%, and 100%), respectively.However, the overall short-term effect favored early surgery (−8.01, 95% CI, −9.27-−6.72),but the result has no clinical significance (I 2 = 100%) due to the high rate of heterogeneity of participants.

Long-term effect
The meta-analysis result for early surgery versus conservative care followed by surgery if needed for long-term effect showed significant improvement for disability (−0.40, 95% CI, −0.50-−0.30)and back pain (−2.30, 95% CI, −2.72-−1.88)with no significant benefit for either surgery or prolonged conservative care for leg pain (−0.00, 95% CI, −0.37-0.37)and global perception of recovery (−0.05, 95% CI, −0.34-0.24)(Figure 6).There was no significant difference between groups for disability, leg pain, back pain, and global perception of recovery with between-study heterogeneity ranging from high to negligible (I 2 = 0%, 0%, 0%, and 100%), respectively.The study did not favor or preferred any intervention in terms of clinical benefit on a long-term basis.

DISCUSSION
This current study identified and reviewed eight studies that compared surgery with conservative care in the management of individuals with LDHR.Due to high heterogeneity of the included studies, only four studies were pooled into meta-analysis.Two studies (23,25) contrasted early surgery with prolonged conservative care followed by surgery if needed.The outcome of this review revealed that early surgery is better than prolonged conservative care for short-term but not different in the longterm effects.This outcome may be possible as some of the patients (39% for Peul et al. (23) and 44% for Peul et al. (25)) in the prolonged conservative care group had to cross-over to surgery due to persistent sciatica or increasing leg pain.Moreover, another reason for the above result could have been that the patients in the early surgery group had more severe symptoms that they could not cop up with the prolonged hospital visits of the prolonged conservative management.
The meta-analysis result of the other two pooled studies (15,16) that contrasted surgery with conservative management did not favor either surgery or non-operative management.However, in addition to more cross-over from conservative treatment to surgery than cross-over from surgery to conservative treatment, patients in the surgical group had more severe symptoms than patients in the conservative treatment group.Furthermore, the conservative treatment protocol was not standardized in all the studies which are in contrast to surgery in which standard open discectomy with examination of the involved nerve root was used.This lack of conservative treatment standardization coupled with heterogeneous patient populations may be responsible for the non-beneficial effect of conservative treatment.

Comparison with other reviews
This current review differed from the previous review (18) that compared surgery with conservative management for LDHR.However, differences in inclusion criteria and search strategies between our review and 2011 review seemed to result in a considerably different collection of trials.For example, only four of the eight trials in our review were included in the 2011 review.This seemed to be attributable to the different databases used as well as increase in new trials published.In addition, the different collection trials in the previous review led to some differences in evidence summaries.Although our review was able to do meta-analysis for four studies, 2011 review could not do it due to heterogeneity of the included trials.
The most common methodological flaws of the trials included in this review were failure to blind participants, therapists, and outcome assessors.Future trials should aim at having a single or double RCT.Another limitation common in the included trials is failure to standardize the conservative treatment  is pertinent, therefore, that subsequent trials should focus on the standardization of the non-operative management.
protocols.Only one study (26) compared microdiscectomy with manipulative therapy, and all other studies did not standardize the conservative care.It

FIGURE 1 .
FIGURE 1. Flow chart of study selection process.

FIGURE 2 .
FIGURE 2. Surgery versus prolonged conservative care for short-term effect.

FIGURE 3 .
FIGURE 3. Funnel plot for surgery versus prolonged conservative care for short-term effect.

FIGURE 5 .
FIGURE 5. Funnel plot for surgery versus prolonged conservative care for Long-term effect.

FIGURE 4 .
FIGURE 4. Surgery versus prolonged conservative care for long-term effect.

FIGURE 6 .
FIGURE 6. Surgery versus usual conservative care for long-term effect.

FIGURE 7 .
FIGURE 7. Funnel Plot for surgery versus usual conservative care for long-term effect.

TABLE 1 .
PubMed search strategy

TABLE 2 .
Rating of trials on the PEDro methodological quality scale

TABLE 3 .
Characteristics of included study