Introduction

Adult scoliosis has been found to occur with increasing frequency as patients age, with rates as high as 68 % in individuals over the age of 60 [1]. Although the majority of these individuals will not seek out medical attention for asymptomatic degenerative scoliosis, the increase in life expectancy of the general population will undoubtedly lead to more elderly patients presenting with spinal deformities seeking reconstructive surgery. Unfortunately, this older patient population also has substantially higher prevalence of chronic disease, which may make any moderate- to-large spinal operation risky [2]. For this reason, it is highly relevant during preoperative counseling to quantify the potential risks and benefits for the patients in this high risk group.

To date, numerous studies have shown that age is an independent risk factor for surgical complications in scoliosis [3–5]. In addition, not only surgical morbidity but also surgical mortality has been associated with older age following spinal surgery as well as differences in alignment [6–9]. However, other studies have shown that older patients may benefit the most from spinal deformity surgery as they often report worse baseline health-related quality of life (HRQOL), greater use of analgesics, and larger functional limitations compared to younger patients [10, 11]. For this reason, surgeons are in need of further data to help risk benefit tradeoffs for the elderly patient with spinal deformity.

In this study, a retrospective review of a multicenter prospective adult spinal deformity database was conducted to compare operative and non-operative management of patients over the age of 75 who presented with symptomatic adult scoliosis. Our hypothesis was that the operative patients would significantly benefit from operative treatment over non-operative treatment.

Methods

Patient population

This study is a retrospective review of a prospectively collected multicenter adult spinal deformity database. Patients were drawn from the International Spine Study Group (ISSG), which is composed of 11 sites across the United States. All patients were enrolled into an Institutional Review Board-approved protocol by each site. Inclusion criteria for the ISSG database were: age >18 years and presence of spinal deformity, as defined by scoliosis Cobb angle ≥20°, sagittal vertical axis (SVA) ≥5 cm, pelvic tilt (PT) ≥25°, and/or thoracic kyphosis (TK) ≥60°. Exclusion criteria included spinal deformity of a neuromuscular etiology and presence of active infection or malignancy. In addition to the database inclusion criteria, patients were included in the present study if (1) ≥4 spinal levels fused and (2) age at time of enrollment was 75 years or greater.

Patients were group based on operative and non-operative management. The decision to pursue operative management was determined between each individual surgeon and patient based on the patient’s goal of care. This decision is very complex and the majority of ASD patients seen in our clinics have received non-operative treatment. They are being seen for possible surgical intervention. The patients presenting with a progressive neurologic compromise, myelopathy, or bowel/bladder incontinence are generally advised to pursue operative treatment. The remaining patients are counseled regarding operative and non-operative management options. In general, patients are encouraged to maximize non-operative treatments, such as physical therapy, steroid injections, and pharmacological therapy.

Data collection, radiographic assessment, and classification

Data collected included demographic data, Charlson comorbidity Index (CCI) surgical data, health-related-quality of life (HRQOL), comorbidities, complications, and reoperations.

Full-length free-standing lateral spine radiographs (36″ cassette) at baseline and 1- and 2-year follow-up were analyzed using validated software [12, 13] (Spineview®, ENSAM, Laboratory of Biomechanics, Paris, France). All radiographic measures were performed at a central location (NYU) based on standard techniques [14] and included: thoracic kyphosis (TK, T4-T12; Cobb angle between superior endplate of T4 and inferior endplate of T12) [15], lumbar lordosis (LL, Cobb angle between superior endplate of L1 and superior endplate of S1) [15], sagittal vertical axis (SVA, C7 plumbline relative to S1), pelvic tilt (PT), and mismatch between pelvic incidence [16, 17] and lumbar lordosis (PI-LL). Cervical radiographic measures included C2–C7 SVA (C2 plumbline relative to C7), C2–C7 lordosis (CL), T1 slope (T1S), and T1S minus CL (T1S-CL).

Based on the above radiographic parameters, patients were additionally stratified by the SRS-Schwab adult spinal deformity classification [18]. The classification is divided into the coronal curve type and sagittal modifiers. The coronal curve type is determined based on the maximal coronal angle measured according to standard Cobb technique. The 4 types include the following: Type T: patients with a thoracic major curve of greater than 30° (apical level of T9 or higher), Type L: patients with a lumbar or thoracolumbar major curve of greater than 30° (apical level of T10 or lower), type D: patients with a double major curve (one T and one L curves), with each curve greater than 30°, and type N: patients with no coronal curve greater than 30° (i.e., no major coronal deformity).

The HRQOL measures collected included pain assessment with an 11 point numerical rating scale (NRS), Oswestry Disability Index (ODI), Short Form-36 (SF36, PCS, and MCS), and Scoliosis Research Society (SRS22) and were obtained preoperatively, 1 year, and 2 years postoperatively. Differences and correlations between age groups for HRQOL at the standard preoperative time and follow-up time points were investigated. Furthermore, the change in HRQOL at the post-operative time points compared with pre-operative values was compared across surgery and age groups. In order to place HRQOL outcomes in a clinically relevant context, minimal clinically import difference (MCID) values have been established for the HRQOL instruments [19, 20]. Analysis for differences in the proportions of patients between operative and non-operative patients reaching MCID for each HRQOL measure was also considered. The MCID values used in the present study included: ODI (−15), PCS (+5.2), SRS Activity (+0.375), SRS Pain (+0.587), SRS Appearance (+0.8), and SRS Mental (+0.42) [19–22]. And lastly, SRS Activity, SRS Pain, SRS Appearance, and SRS Mental scores were compared to age- and sex-matched subjects from a large study of 1346 adult subjects without scoliosis [23].

Complications were recorded and included the number of intra-, peri-, and post-operative complications and associated number of revisions and reoperations. The specific type of complication was recorded and included implant failure, infection, neurologic, cardiopulmonary, vascular, gastrointestinal, operative, renal, wound problems, radiographic, and death. In addition, medical comorbidities were recorded.

Statistical analyses

Continuous variables were described with the mean and standard deviation (SD). Statistical analysis was performed using the Student’s t test and the Wilcoxon rank sum or Kruskal–Wallis tests where appropriate. Patients were also compared to age- and sex-matched subjects without spinal deformity at baseline and 2 years for SRS Activity, SRS Pain, SRS Appearance, and SRS Mental. Frequency analysis used for categorical variables was carried out using the Fisher’s Exact Test. All statistical analyses were conducted using commercially available software (JMP v7.0, SAS Institute, Inc., Cary, NC) and the level of significance was set at p < 0.05 for all.

Results

Patient population

51 patients met inclusion criteria and where eligible for 2-year follow-up, of which 27 (52.9 %) met baseline and 2-year clinical and radiographic follow-up: 12 operative and 15 non-operative (Table 1). Twenty four patients were excluded from the 51 eligible patients due to the following reasons: 10 were lost to follow-up, 7 withdrew from the study, 4 did not have complete baseline data, and 3 non-operative patients crossed over to operative management. 10 out of the 27 (37.0 %) were at least 80 years of age. The non-operative group had a significantly higher percentage of female patients. There were no significant differences between the operative patients and non-operative patients for age, BMI, CCI, and SRS-Schwab coronal curve Type (p > 0.05 for all). The majority of patients (59.3 %) were classified at baseline as SRS-Schwab coronal curve Type N (Table 1).

Table 1 Demographic data and baseline SRS-Schwab coronal curve types for patients ≥75 years of age
Full size table

Comorbidities

Out of the total of 27 patients, 24 (88.9 %) had at least one comorbidity that ranged from 1 to 5. The number of patients with each type of comorbidityis listed in Table 2. The majority of patients had 2 or greater comorbidities (48.1 %). The top 5 most common comorbidities included leg weakness (66.7 %), numbness or tingling in legs (51.9 %), arthritis (48.1 %), hypertension (44.4 %), and osteoporosis (37.0 %).

Table 2 The number of patients with the listed number and types of comorbidities for all patients
Full size table

Surgical data

Surgical characteristics are presented in Table 3. All 12 operative patients underwent a posterior-only approach and posterior instrumentation and fusion. The average length of stay (LOS) in days was 12.9 ± 3.7. Estimated blood loss (EBL) in mL was 2,416.7 ± 1,565.4. The levels fused ranged from 4 to 17 and the UIV and LIV levels ranged from T1 to T11 and L5-ilium, respectively. 11 out of 12 (91.7 %) patients underwent an osteotomy and 11 (91.7 %) had a decompression. Of note, 33.3 % of patients had a pedicle subtraction osteotomy (PSO), while 58.3 % underwent a posterior element osteotomy (ponte or smith-peterson osteotomy, SPO).

Table 3 Surgical characteristics for patients ≥75 years of age
Full size table

Complications

The percentages listed are out of the total operative patients (n = 12, 100 %). 9 (75.0 %) patients had at least 1 complication with a total of 24 complications (11 major and 13 minor, Table 4). There were 3 (25.0 %) patients that had an intraoperative complication, 6 (50.0 %) with peri-operative, and 8 (66.7 %) with a post-operative complication. 6 patients (50.0 %) had a reoperation during the two-year follow-up period, of which 3 were for a revision. The 3 reoperation indications were coronal malalignment, radiculitis, and deep infection. The 3 revision indications were 2 rod fractures and 1 for PJK. The top 3 major complications included implant (25.0 %), operative (25.0 %), and radiographic (16.7 %, Table 4). The top 4 common minor complications included radiographic (25.0 %) and cardiopulmonary (25.0 %) neurologic (16.7 %), and gastrointestinal (16.7 %). There were neither wound complications nor deaths during the 2 year follow-up period.

Table 4 The number of patients with complications by timing and type
Full size table

Radiographic analysis

Means ± 1 SD are presented in Table 5. All baseline radiographic parameters were statistically similar between operative and non-operative patients (p > 0.05 for all). The operative patients had significantly lower 2-year PI-LL (7.5 ± 13.2 vs. 27 ± 13.9, p = 0.0008). In addition, the operative group had a significant improvement in TK (p = 0.0047), C7SVA (p = 0.0056), and PI-LL (p = 0.0067), at 2 years compared with preoperative values (Table 5). There were no significant differences between baseline and 2-year radiographic parameters for the non-operative patients (p > 0.05 for all).

Table 5 Means ± 1 SD of all radiographic parameters at baseline, 1 and 2 years for patients ≥75 years of age
Full size table

HRQOL analysis

The mean SRS values for age- and sex-matched normal controls [23] used for comparison to the present study population were the following: SRS Activity (4.12 ± 0.02), SRS Pain (4.25 ± 0.05), SRS Appearance (4.17 ± 0.04), and SRS Mental (4.11 ± 0.06). All the patients, as well as within the operative and non-operative groups, had a significantly lower mean baseline and 2 year SRS Activity, SRS Pain, and SRS Appearance compared with the age- and sex-matched normal controls (p < 0.05 for all). In addition, the operative group had significantly worse baseline and 2 year SRS Mental than the age- and sex-matched normal controls (p < 0.05 for both).

The operative patients had worse baseline ODI (p = 0.0096), PCS (p = 0.0034), SRS Activity (p = 0.0114), SRS Appearance (p = 0.0095), and SRS Total (p = 0.0097) than the non-operative patients (Table 6). The operative group had significantly higher 2-year SRS Appearance (p = 0.0306) and 2-year SRS Satisfaction (p = 0.0053) than the non-operative patients. No other 2-year differences were noted (p > 0.05 for all). When compared to baseline values, the operative patients had significant improvement in 2 year ODI (p = 0.0072), PCS (p = 0.0009), SRS Activity (p = 0.0104), SRS Appearance (p = 0.0011), SRS Pain (p = 0.0079), SRS Satisfaction (p = 0.0102), and SRS Total (p = 0.0014). There were no significant differences between baseline and 2-year HRQOL for the non-operative patients (p > 0.05 for all). Furthermore, the operative patients were significantly more likely to reach MCID than the non-operative patients for PCS (66.7 vs. 14.3 %), SRS Activity (81.8 vs. 26.7 %), SRS Pain (81.8 vs. 33.3 %), and SRS Appearance (63.6 vs. 0.0 %, p < 0.05 for all, Table 7).

Table 6 Means ± 1 SD of all health-related quality of life measures at baseline, 1 and 2 years for patients ≥75 years of age
Full size table
Table 7 The percentage of patients that reached minimum clinically important difference (MCID) from baseline to 2 years post-operative for Oswestry disability index (ODI), physical component score (PCS) from the SF36, and the Scoliosis Research Society (SRS) questionnaire
Full size table

Discussion

We conducted a retrospective study comparing groups of elderly ASD patients who were treated non-operatively and operatively. Of the surgical cohort, 88.9 % had at least one comorbidity and 75 % had at least one complication. However, these patients had a greater improvement in outcomes than their non-operative counterpart, and this improvement was significant over 2 years. Therefore, this study suggests that the operative management of patients with spinal deformity over the age of 75 may be justified in a select group of patients despite a high comorbidity and complication rate. Our results are in agreement with a previous study in which Bridwell et al. [10] prospectively compared changes in HRQOL for non-operative and operative treatment of ASD over a 2-year follow-up period; the older surgical population (60–80 years) experienced significant HRQOL improvements following treatment, while its non-operative counterpart did not. Our data are valuable in the current environment, in which health care systems are trying to quantify the risks and benefits of various medical treatments or surgical interventions based on value obtained per cost [24].

With the global expansion of the elderly population (≥65 years), the demand for surgical treatment of adult spinal deformity will significantly increase, especially by 2030 [25]. Unfortunately, particularly in the context of ASD surgery, numerous studies correlate older age to a higher rate of morbidity and mortality [3, 4, 6, 8]. Elderly adults typically exhibit inferior bone quality and a greater degree of degeneration of various spinal structures, lengthening the operative time and increasing the risk of postoperative complications, especially pseudarthrosis, instrument-related failure, and proximal junctional kyphosis [3, 5, 26]. Moreover, patients over 50 are usually burdened with additional systemic diseases such as cardiovascular disease and diabetes, resulting in an increased risk for peri-operative medical complications [2].

Despite increased comorbidities and surgical risks, elderly patients often have worse baseline HRQOL measures and thus may benefit more significantly from surgery than younger patients. For example, in a retrospective cohort study assessing improvements in HRQOL for operative treatment of ASD, Smith et al. [11] found that, in comparison to the younger population (25–44 years), the older population (65–85 years) started with lower baseline QOL scores and experienced significant improvements with respect to ODI and leg pain numerical rating scale (NRS), as well as a trend toward greater improvements with respect to SF-12 PCS, SRS-22, and back pain NRS.

The present study suggests operative treatment may benefit select elderly patients with ASD by decreasing disability and increasing health status, however, it is important to note that some elderly patients may be poor surgical candidates due to being too unhealthy, having very poor bone quality, or having a lack of a sufficient social support. Thus, the elderly patients that do undergo surgical intervention are a select group despite the high comorbidity rates. They present with worse baseline HRQOL than the non-operative patients and therefore were likely operated on as a result. Although the populations may be different, it is clear throughout the literature that operative patients have worse baseline HRQOL and disability than non-operative patients and thus this does not invalidate the comparison [10, 27–31]. Furthermore, we do not necessarily advocate operative management for all elderly patients with symptomatic ASD. Non-operative methods should be attempted first and all means exhausted before the discussion of surgery intervention. During this discussion, several key factors should be considered, including the severity of the patient’s symptoms, the impact of these symptoms on function and quality of life, the overall health of the patient, and the willingness of the patient to accept the risks of surgery. The complication rate is much higher in the elderly following ASD surgery [32–35] and patients should understand the associated complications and their respective treatments before proceeding.

There are limitations in this study worth mentioning such as patients lost to follow-up. The retrospective nature of this study inherently diminishes the quality of the data; the lack of prospective randomization with regard to treatment selection (operative vs. non-operative) may have created biased cohorts. Specifically, surgeons may have selected patients that they deemed suitable for surgery or patients may have self-selected themselves for or against surgery. Moreover, the relatively small patient population restricts the statistical power of the results obtained, especially with respect to the magnitude of the risk–benefit ratio for surgery (i.e., number of complications and complication rate versus improvements in HRQOL measures). It also restricts the ability to identify certain patients that are candidates for operative treatment beyond the radiographic indication. Future work with larger studies may provide the means to assess this. And lastly, this study did not account for compensatory mechanisms in either cohort, which may also provide valuable differences in these populations [36, 37].

Conclusion

In the largest series to date comparing operative and non-operative management of adult spinal deformity in elderly patients greater than 75 years of age, reconstructive surgery can provide significant improvements in pain and disability over a two-year period. When compared to a similar cohort of elderly patients managed non-operatively over a similar time period, patients undergoing surgery had significantly improved HRQOL measures over baseline values and compared to the non-operative cohort, despite substantial operative morbidity. Furthermore, operative patients were more likely to reach MCID than non-operative patients. When counseling elderly patients with ASD, such data may be used to prognosticate for this complicated age group.