Material and methods

Literature searching and selection strategy

Searching strategy included randomized controlled trials in the Pubmed, EMBASE, ELibrary and Cochrane Library databases published for the period from 2008 to August 2018 and devoted to comparison of long-term results of TA and ACSF in surgical treatment of degenerative disease of cervical intervertebral discs. Two researchers analyzed literature data. Disagreements regarding inclusion of data into meta-analysis were resolved via consensus with the entire team of the authors. The trial was carried out in accordance with international recommendations for systematic reviews and meta-analyzes (PRISMA) [15].

At the first stage, literature searching was carried out using the keywords "cervical disk arthroplasty", "cervical total disk replacement", "anterior cervical discectomy and fusion", "cervical spine degeneration", "cervical intervertebral disс degeneration" for English-speaking systems and «тотальная артропластика шейных межпозвонковых дисков», «тотальное эндопротезирование шейных межпозвонковых дисков», «передний шейный спондилодез», «дегенерация шейного отдела позвоночника», «дегенеративное заболевание шейных межпозвонковых дисков» for the e-Library system. Titles of reports were analyzed in manual fashion regarding their accordance to inclusion criteria. At the second stage, we examined abstracts and excluded publications inappropriate to inclusion criteria. At the third stage, we looked at the full text of the articles for their compliance with inclusion criteria and references for relevant studies (Fig. 1).

Inclusion criteria

The following compliance criteria were determined in order to compare the effectiveness of two surgical interventions:

(1) accepted trials: randomized clinical trials devoted to long-term results of TA and ACSF in adults with cervical spine degeneration followed by neurological symptoms (radicular neuralgia, radicular neuritis or radiculopathy);

(2) types of surgical interventions: trials comparing TA of cervical intervertebral discs and ACSF using various implants;

(3) outcomes: analysis of long-term clinical and instrumental results of these interventions; regression of neurological symptoms, quality of life associated with Neck Disability Index (NDI), VAS score of pain in the cervical spine and upper extremities, incidence of adverse events and degeneration of adjacent spinal segments, incidence of redo surgical interventions;

(4) study design: randomized clinical trials with Jadad score 3 [16] and over and postoperative follow-up at least 48 months.

Research bias risk assessment

Each study included in this meta-analysis was evaluated using the “Study Bias Risk Assessment” option of Review Manager 5.3 software (The Nordic Cochrane Center, The Cochrane Collaboration, 2014, Copenhagen, Denmark). The following parameters were applied: (1) data sequence generation, (2) research data concealment (3) blinding procedure, (4) incomplete research data, (5) selective presentation of research results and (6) other bias parameters (Fig. 2).

Statistical analysis

Relative risk (OR) and 95% confidence interval (CI) were calculated for dichotomous variables. Continuous variables were analyzed considering standardized mean difference (SMD) and 95% CI (random-effect model). I² coefficient was applied to assess between-study heterogeneity. Homogeneous trials were determined by I² coefficient<25%, low heterogeneity – 25—50%, moderate heterogeneity – 50—75%, over 75% — high heterogeneity. Study asymmetry was analyzed using a funnel plot and Egger’s regression-based test. Forest plots were constructed using Review Manager 5.3 software (The Nordic Cochrane Center, The Cochrane Collaboration, 2014, Copenhagen, Denmark). Differences were significant at p-value ≤0.05.

Results

Literature searching

According to inclusion criteria, the meta-analysis enrolled 9 randomized controlled trials devoted to assessment of long-term postoperative results in 2439 patients with cervical spine degeneration [17–25]. Overall characteristics of studies are shown in the Table.

Major clinical, instrumental and intraoperative data were extracted from all studies included into meta-analysis. Information on the use of various prostheses and cervical spine implants (ProDisc-C, Mobi-C, Bryan, Prestige, PCM Kineflex/C), as well as cages and autologous bone implants for ACSF surgery was shown.

Clinical parameters

Clinical outcomes of TA and ACSF surgery in patients with cervical spine degeneration in long-term period included: (1) regression of neurological symptoms, (2) NDI-associated quality of life, (3) VAS score of pain in cervical spine and (4) upper limbs. Pooled analysis of showed that TA of cervical intervertebral discs was followed by more common regression of neurological symptoms (RR=1.03, 95% CI 1.00 — 1.06, p<0.00001; I²=9%; Fig. 4).

Pooled analysis also confirmed significantly higher NDI-associated quality of life after TA (SMD=-0.34, 95% CI -0.47 — -0.22, p=0.05; I²=9%; Fig. 5)

and less VAS score of pain in upper limbs (SMD=-0.17, 95% CI -0.31 — -0.03, p=0.02; I²=0%; Fig. 6).

At the same time, there were no significant differences in VAS scores of pain in cervical spine (SMD=-0.13, 95% CI -0.46 — 0.21, p=0.46; I²=78%; Fig. 7).

Redo surgery

Incidence of repeated interventions was significantly lower after TA compared with ACSF (RR=0.52, 95% CI 0.37—0.75, p<0.0005; I²=28%; Fig. 8).

Incidence of redo surgeries on the operated (OR=0.40, 95% CI 0.27—0.58, p<0.00001; I²=3%; Fig. 8) and adjacent spinal segments (OR=0.43, 95% CI 0.26—0.72, p<0.001; I²=0%; Fig. 8) was also significantly lower after TA compared with ACSF.

Degeneration of adjacent spine segments

TA was characterized by significantly lower incidence of degeneration of overlying adjacent spinal segment (RR=0.56, 95% CI 0.42—0.74, p<0.0001; I²=68%; Fig. 9).

At the same time, we did not obtain significant differences in the incidence of degeneration of underlying segment in both groups (RR=0.56, 95% CI 0.28—1.09, p=0.09; I²=93%; Fig. 9).

Adverse events

Pooled analysis of the incidence of adverse events demonstrated their slight predominance after TA (RR=0.64, 95% CI 0.42—0.99, p=0.04; I²=0%; Fig. 10).

Discussion

ACSF surgery is still the "gold" standard of surgical treatment of patients with cervical spine degeneration although TA of intervertebral discs has been actively used in spinal surgery over the past ten years [26, 27]. The secondary role of TA in surgical treatment of patients with cervical spine degeneration is largely determined by the absence of clear data of high clinical efficacy of this procedure in long-term postoperative period [28]. There are a large number of meta-analyses comparing clinical efficacy and safety of TA and ACSF surgery in patients with cervical spine degeneration. However, most of these meta-analyses are based on the study of randomized clinical trials with a short-term period of postoperative follow-up. Considering the peculiarities of the methodology of previous studies, we have analyzed controlled randomized trials comparing the long-term results of TA and ACSF surgery in patients with cervical spine degeneration.

We found that long-term outcomes after TA are superior to those after ACSF surgery in patients with cervical spine degeneration. Redo surgery is the most important indicator of clinical and economic effectiveness of any surgical intervention due to its direct impact on the quality of life, additional financial costs for patients and/or state health care system. TA was followed by significantly less incidence of repeated interventions compared with ACSF surgery. Our data are largely consistent with the results of meta-analyses of A. Hu et al. [9] and Y. Wu et al. [11]. However, these researches enrolled randomized clinical trials with small sample size.

Degeneration of adjacent intervertebral discs is one of the common causes of repeated hospitalizations and surgical interventions after ACSF surgery [29]. In contrast to ACSF surgery, TA of the intervertebral discs is associated with normal biomechanics of cervical spine and even distribution of the load between spinal segments [30]. However, there are no clear evidence in favor of reducing the incidence of degeneration of adjacent segments after TA. Pooled analysis clearly demonstrated that TA results significantly less incidence of degeneration of the overlying adjacent spinal segment. On the other hand, incidence of degeneration of the underlying spinal segment was similar after TA and ACSF surgery. It is worth noting that studies comparing incidence of degeneration of adjacent segments were characterized by high heterogeneity. In our opinion, high heterogeneity of these studies was due to various diagnostic criteria for degeneration of adjacent intervertebral discs and the number of operated spinal segments. It is known that degeneration of adjacent segments confirmed by instrumental survey does not significantly correlate with clinical and neurological symptoms [31, 32]. In this regard, prospective randomized trials with large sample size and long-term postoperative follow-up are necessary for more objective assessment and comparison of the incidence of degeneration of adjacent segments after TA and ACSF surgery.

Limitations

There are certain limitations of this meta-analysis. Firstly, meta-analysis enrolled only 9 randomized clinical trials with small sample size that obviously influenced the results of statistical analysis. Secondly, the trials were characterized by certain methodological drawbacks, in particular, the absence and/or violation of the blinding procedure that affected the reliability of the results too. Thirdly, postoperative follow-up period was significantly less after ACSF surgery (p=0.038). This feature directly influenced the results of pooled analysis.

Conclusion

TA is characterized by significantly better clinical efficacy in patients with cervical spine degeneration in long-term postoperative period compared with ACSF surgery. Undoubtedly, further meta-analysis enrolling high-quality randomized clinical trials with long-term postoperative follow-up of patients after TA and ACSF surgery for cervical spine degeneration are required.

Financing

The study did not have financial support.

No conflict of interests to declare.

Authors’ participation:

Byvaltsev V. A. — design and methodology of research, final editing of the manuscript;

Stepanov I. A. — literature review, statistical analysis;

Aliev M. A. — literature review, statistical analysis;

Aglakov B. M. — literature review, statistical analysis;

Yusupov B. R. — literature review, statistical analysis;

Konovalov N. A. — final editing of the manuscript.

Commentary

The article is devoted to analysis of long-term outcomes of surgical treatment of cervical spine degeneration. Considering increased life expectancy of people in developed countries, greater number of patients with degenerative spine diseases is expected. New surgical technologies appear in response to this tendency. First of all, effectiveness of certain technology is determined by postoperative outcomes. The authors performed a meta-analysis of long-term outcomes of surgical treatment of patients with cervical spine degeneration and compared two technologies — total arthroplasty and anterior cervical spinal fusion surgery. A meta-analysis enrolled 9 randomized controlled clinical trials comprising 2439 patients with cervical spine degeneration. Long-term postoperative results were analyzed. The study was carried out in accordance with international recommendations for systematic reviews and meta-analyses (PRISMA). It should be noted that the authors use the Review Manager 5.3 software “Research Bias Risk Assessment” to evaluate each study included in this meta-analysis. The authors showed that total arthroplasty is characterized by significantly better clinical efficacy in patients with cervical spine degeneration in long-term postoperative period. I would like to note that consideration of arthroplasty as a secondary approach, in my opinion, is not entirely correct. The same is true for the term "gold standard". The last decade has clearly shown that medicine is becoming more personalized (“4P medicine”). Today we have the opportunity to choose a treatment strategy considering the goals and individual characteristics of a particular patient. In this context, similar safety of both methods is more important than data on superiority of one method over another. In general, the article is devoted to a relevant issue due to thorough methodology and evidence-based analysis. The report may be recommended for the use in educational programs and lectures as an example of correct approach.

A.G. Nazarenko (Moscow, Russia)