This systematic review identified and analysed randomised controlled trials of the effectiveness of physiotherapy and rehabilitation after spinal surgery for degenerative spine conditions. The studies consisted of groups that received physiotherapy and rehabilitation after surgery, standard postoperative care, or no postoperative intervention. Thus, it was aimed to show the additional effects of physiotherapy and rehabilitation to surgery more clearly. Answers were also sought to the questions of in which time period after surgery physiotherapy is more effective, for which pathologies, and in which type of surgeries.

Significant improvements in total follow-up time after cervical spinal surgery in both intervention and control groups were recorded in most of the included studies, except for a few parameters in some studies.

In the study by Coronado et al. [16], although there was a significant improvement in mental health scores in the HEP group in the first 6 weeks, there was no improvement over the entire 12-month follow-up period. Wilbaut et al. [8] also reported no improvement in cervical rotation range of motion in any group. In the study by Peolsson et al. [10] the prevalence of patients with arm reflex impairment did not change significantly in the SPT group during the 24-month follow-up period. Hermansen et al. [14] showed that there was no improvement in walking balance in the SPT and SA groups at the end of 6 months. Finally, in the study of Uehara et al. [17] no change was found in the C2–7 sagittal vertical axis distance in both groups.

Although the improvements in most of the outcome measures were similar between the intervention and control groups in the reviewed studies, significant improvements were recorded in favor of the intervention group in certain clinical parameters in some studies. However, there were some results that can be described as surprising.

Wilbault et al. [8] reported that the improvement in the CSQ-CAT score between 3 and 6 months in the SPT group was higher than the improvement in the SA group, but there was no difference between the improvements in the total follow-up period of 6 months. It would not be very realistic to interpret this difference, which occurs only in the three-month period and disappears in the total follow-up period, as a positive effect of structured physiotherapy in terms of catastrophising.

In another study by Wilbaut et al. [9], higher expectation satisfaction was reported in the SPT group than in the SA group at the end of a total follow-up period of 6 months. Expectation fulfillment, in a way, refers to the effectiveness of all surgical and postoperative approaches on the improvements that occur from the patient’s point of view. Patient feedback of all medical interventions is considered as an important success criterion. Therefore, the difference in this parameter can be considered valuable. In addition, the SPT ≥ 50% group showed more improvement in the expectation fulfillment, patient enablement instrument, and neck pain frequency parameters than the SA group in the same follow-up period, suggesting that compliance with the physiotherapy program plays a critical role in treatment effectiveness.

Peolsson et al. [10] found that contrary to expectations, EQ-VAS improvements were higher in the SA group than in the SPT group, both at 3 months and for a total of 24 months. This difference in the improvement in quality-of-life VAS score in favour of the SA group was surprising. At the end of 24 months, the scores of the groups were quite close to each other. The reason for the improvement difference is likely to be that the EQ-VAS baseline scores of the SA group were lower than those of the SPT group.

In the study of Coronado et al. [16], in terms of improvement between groups, the neck pain NRS score was higher in the HEP group than in the UC group in the first 6 weeks. Although the difference in improvement was not maintained during the total follow-up period, considering the high pain intensity at baseline, it can be considered important that the early home exercise program contributed to the reduction of severe pain in the early postoperative period.

In another study, Peolsson et al. [12] reported that the prevalence of participants with impaired arm reflexes decreased more in the SA group than in the SPT group in the 1-year follow-up period. However, this difference disappeared by the end of the total follow-up period of 2 years. When the two groups were analysed, it was observed that the prevalence values of the SPT group were quite close to each other from baseline to 2 years, but in the SA group there was a significant decrease in the first year but an increase between the first and second year. This suggested that the prevalance of patients with impaired arm reflexes in the SA group increased between the first and second year.

When the results were considered overall, it was seen that in some studies, significant improvements in certain clinical parameters were recorded in the intervention group compared to the control group.

When the studies were analysed in this review, there were seen to be variations in many factors such as patient age, initial pathologies, surgeries performed, and the level of surgery.

In the studies of Wilbaut, Peolsson, Svensson, and Hermansen et al. [8,9,10,11,12,13,14], the age range of the patients included in the study was as wide as 18–70 years. McFarland et al. [15] included patients aged between 30 and 75 years, Coronado et al. [16] included those over 21 years of age and did not specify an upper age limit, and Uehera et al. [17] included patients aged 40–89 years. The severity of cervical degeneration and postoperative recovery processes are known to differ in different age groups. Osteophytic structures in the cervical vertebral column that do not cause root symptoms but reduce cervical range of motion are commonly seen in geriatric patients. In addition, the endurance and strength of the neck muscles are likely to be lower than in a young patient. Such factors have the potential to negatively affect adherence to exercise after surgery.In contrast, non-geriatric patients, who are in employment, may not have the time and opportunity to exercise as much as geriatric individuals, who are more likely to be retired. Therefore, such wide age ranges in the studies may be a confounding factor in terms of PT efficacy.

Different types of surgical techniques were applied in the studies included in this review. In seven studies with patients from the same sample, some of the participants underwent anterior discectomy and fusion surgery (ACDF) and some underwent posterior cervical foraminotomy (PCF) with or without laminectomy [8,9,10,11,12,13,14]. The number of fusion levels applied to the patients varied. Of these seven studies, only the study by Peolsson [11] included only patients who underwent ACDF. Similarly, in the studies by Coronado[16] and McFarland et al. [15], only ACDF patients were included. Uehara et al. [17] used a different surgical method (Kurokawa technique) compared to the other studies.

The pathologies of the patients in the studies were quite diverse, both within each study and when all the included studies were considered. Wilbaut, Peolsson, Svensson, and Hermansen et al. [8,9,10,11,12,13,14] defined the inclusion criteria as cervical disc herniation detected on MRI and clinically confirmed radiculopathy symptoms but did not exclude the presence of additional stenosis. In addition, the duration of root symptoms was determined to be at least 2 months. McFarland et al. [15] used a very general statement as an inclusion criterion, such as having no major cervical structural disorder but having reflex, sensory, and motor loss or having major cervical structural disorder regardless of the presence or absence of problems in these three parameters, and did not set any criteria for symptom duration. The widest pathology spectrum was in the study by Coronado et al. [16], which included patients with cervical stenosis, spondylosis, spondylolisthesis, and disc bulging, and no criterion for symptom duration was set. Uehera et al. [17] also included patients with spondylitic myelopathy and posterior longitudinal ligament ossification without using any severity classification and without specifying symptom duration.

One prerequisite for drawing a common conclusion from a group of publications is that the studies should have a certain level of homogeneity. Therefore, the heterogeneous structure of the literature does not allow an inference to be made about whether PT is effective after cervical spinal surgeries.

When PT should be started after cervical spine surgery, in which type of patients it is more useful, and in which type of surgeries the results are better are important questions to which the scientific world is seeking answers.

Two studies are worthy of particular note with regard to the initiation of PT. It was observed that the physiotherapy program was started immediately after surgery in Coronado and Uehera’s publications and 6 weeks after surgery in other publications. In the study by Coronado, it is noteworthy that an improvement was observed in the important parameter of neck/arm pain NRS in a short period of 6 weeks. However, the authors mentioned the limitation of insufficient sample size. With the exception of the McFarland study, it can be seen that the sample size of those two studies was smaller than that of the other studies. This may have contributed to the fact that the effectiveness of the physiotherapy program was not statistically revealed. Moreover, unlike the other studies, Uehara used a method that did not incise a long and wide muscle such as the semispinalis capitis muscle. Eliminating a potential source of pain and dysfunction may have masked the beneficial effects of physiotherapy. Although it is not realistic to say that there is a positive effect of starting a physiotherapy program earlier based on only two studies, this issue could be answered with future appropriately designed studies.

One of the important factors in the characteristics of the patients included in the studies was age. When the mean age of the patients included in all the studies was considered, the patients in the study by Uehra et al. were considerably older than in the other nine studies, where the patients were of a more similar age. Therefore, the Uehera et al. study can only be compared with the other included studies to analyze the effect of the age factor on treatment efficacy. The common outcomes with some of the other studies were neck pain severity, neck flexor/extensor muscle strength, and cervical range of motion. Significant improvements were noted in all other studies with these outcomes, including the study by Uehera et al. In conclusion, based on a single study, it would not be an appropriate conclusion to say that the age factor has no effect on treatment efficacy.

Another important parameter of patient characteristics is the initial pathology and its severity. To be able to answer the question of which patients with what type and severity of pathology will gain optimum benefit from treatment, the pathology should be clearly categorized in the studies to be compared. For example, if all the patients in the included studies had radiculopathy symptoms due to cervical disc herniation, but the number of levels differed, it would be possible to comment on the efficacy of treatment based on the level or severity of the pathologies. Or, conversely, there should be a situation where the same number of levels were involved but the pathologies were different. However, no such pure classification of pathology was made in the included studies. This heterogeneity in the studies makes it impossible for definitive conclusions to be drawn on this issue. For similar reasons, no conclusions could be made about treatment efficacy in terms of symptom duration.

Another question to be answered is which surgical technique will provide better results. The surgery performed in Uehera’s study differs from the other studies. In all the other studies, the ACDF technique was applied to the majority of the patients and PCF technique to a minority, whereas in the study by Uehara et al. laminectomy surgery (Kurokawa method) was performed by preserving the semispinalis capitis muscle. One of the most important differences of the Kurokawa method from the ACDF and PCF techniques is the absence of vertebral fusion. This important difference is remarkable, and prevents the comparison of common results of the Uehera et al. study with other studies. The most important obstacle to this is that the pathologies in the Uehera study are different from the degenerative pathologies in other studies, such as spondylotic myelopathy and posterior longitudinal ligament ossification. These two pathologies, unlike those in the other studies, usually cause medulla spinalis compression, not root compression. Therefore, it would be unrealistic to compare the results of the Kurokawa method with ACDF and PCF surgical results.

One of the most important handicaps in the studies included in the review is the uncertainty of the compliance rates of the participants with the physiotherapy and rehabilitation program. In all the included studies by Wilbaut, Peolsson, Svenson, and Hermansen, patients in the structured physiotherapy group visited their physiotherapist once a week during the first week of the treatment program and twice a week in the following weeks. In fact, in the structured physiotherapy program, they performed the exercises as part of a home program [8,9,10,11,12,13,14]. Wilbault et al. [8] reported inadequate physiotherapist supervision in rehabilitation programmes, and Svensson et al. [13] and Hermansen et al. [14] reported inadequate participation in structured physiotherapy programmes as a handicap. In all of these studies, the rate of participants with ≥ 50% attendance in the structured physiotherapy program was 67%. It was also found that approximately 10% of the participants in the SPT group never started the treatment program. In the study by McFarland et al. [15] it was reported that the patients were discharged 1–2 days after surgery, received 1–2 sessions of physiotherapy during hospitalization, and then performed the exercises on their own. Two follow-up interviews were conducted during the follow-up period, and the questions of the patients about the physiotherapy program were answered by the physiotherapists responsible for them. However, it was not reported in the article to what extent the patients complied with the physiotherapy programs. In that publication, it was stated as a limitation that 1–2 sessions of face-to-face training and telephone follow-up were insufficient to provide deep cervical flexor muscle activation training to the ECS group. In the study by Coronado et al. [16], it was reported that the exercises were guided by a physiotherapist in weekly telephone interviews, and the participants completed an exercise diary. It was mentioned however that this patient follow-up system was inadequate for the detection of exercise compliance. Similarly, in the study by Uehara et al. [17], it was emphasised that patients were asked to record their daily exercises in a diary, but no information was given about checking the patients’ exercise diaries and their level of compliance with exercise. However, the article reported concerns about the reliability of using exercise diaries to measure patients’ level of compliance with exercise. There were two studies in which an SPT ≥ 50% group was identified and analysed. In the 2018 publication by Wilbaut et al. more positive results were reported in the SPT ≥ 50% group than in the SA group in notable parameters such as neck pain frequency, high expectation fulfillment, and the PEI score measuring the ability to cope with the current health problem and life [8, 9]. Although it is difficult to conclude from a single study that increased participation in exercise leads to more favourable outcomes, when all the included studies are considered, it can be said that the uncertainty of the level of attendance in exercise is one of the biggest barriers to drawing conclusions about the effectiveness of physiotherapy.

In the studies included in this review, specific exercise types, repetitions, number of sets, and intensities in the physiotherapy and rehabilitation programs generally applied were not described in detail. The same structured physiotherapy program was used in all the studies by Wilbaut, Peolsson, Svenson and Hermansen [8,9,10,11,12,13,14]. Although the structured physiotherapy program (SPT) specifies weekly progression, the exercises to be performed are described in general terms in the SPT program diagram and in the text, such as “non-resistance exercises to activate the deep cervical neck muscles or progressive low-load exercises to increase neuromuscular control of the trunk and scapular muscles”. For example, many exercises can be included under the definition of progressive low-load exercises to increase neuromuscular control of the scapular muscles. Different exercises for the same purpose may have advantages and disadvantages compared to each other and may differ in terms of suitability for patients. For other parameters of the exercises, inadequate descriptions, such as only “2–3 times a day” were used. This kind of lack of information can be considered an important shortcoming for trials in which the main aim is to assess the effectiveness of physiotherapy and rehabilitation after surgery. A detailed explanation of this is given in the study published by Peolsson in 2019, which can only be accessed with a DOI number [10]. However, when the source of the relevant exercise visuals is examined, these exercises can be seen to have been designed for a study published in 2013 that examined the effectiveness of neck-specific exercises after whiplash injuries [19]. In addition, there is a discrepancy between the exercise program in the visuals and the exercise program in the publications. The other six trials conducted in the same group of patients did not use this source. In the studies by McFarland [15], Coronado [16] and Uehara et al. [17], the exercises were explained in more detail and it was possible to make clearer comments about the physiotherapy program applied. However, Coronado et al. reported that the participants in the usual care group, which was the control group, were referred to physiotherapy after the 6th week. Therefore, this can be said to be a factor that raises doubt about the objective comparison of the outcomes of the two groups.

With the exception of the study by Coronado et al. [16] within-group changes in follow-up times were analyzed with the two groups combined in all the other studies. Wilbaut et al. [8, 9] reported within-group change over time in SPT-SA groups combined and SPT ≥ 50%-SA groups combined, and Peolsson [10,11,12], Svenson [