Introduction

It is estimated that the current prevalence of adult spinal deformity (ASD) is upwards of 60% with expectations for continued increase within an ageing population. In this context, while only a small proportion of these patients ultimately experience symptoms, these symptoms may be severely debilitating [1, 2], significantly affecting patient function and subsequent quality of life [3].

Surgical deformity correction of ASD has been shown to be of significant benefit in the appropriately selected patient [2, 4]. However, potential rates of complications are significant particularly with use of conventional deformity corrective techniques, with all cause complication rates approaching as high as 86% [5,6,7] and an up to 18.5% rate of major complications [5]. Furthermore, the rate of post-operative hardware failure is also considerable and is now considered a leading cause of revision surgery [8]. Minimally invasive deformity corrective techniques have emerged in hopes of mitigating risk, with early data suggesting a more favourable safety profile [9,10,11,12,13,14,15], particularly with use of a circumferential minimally invasive spine (CMIS) technique, even in the setting of severe deformity correction [16].

Despite these advancements, efforts to further decrease complications rates remain of significant importance. In this context, as most patients with ASD, particularly females, are more likely to have metabolic bone disease, optimization of bone health remains a potential area of risk mitigation with deformity correction. Various strategies currently exist that are theorized to improve bone quality in the peri-operative period; however, limited data exist regarding the actual effect of some of these strategies to do so.

Pre-operative or peri-operative use of parathyroid hormone (PTH) analogues including teriparatide and abaloparatide (Forteo and Tymlos) has emerged in recent years as one of the foremost strategies to improve bone quality in the pre/peri-operative periods. Therefore, the purpose of our study was to critically evaluate whether a PTH analogue would improve lumbar bone density in the 1-year peri-operative period. Secondarily, we set out to quantify the prevalence of hardware-related complications in patients who were successfully treated with PTH analogue.

Material and methods

Patient data

A prospectively collected data registry of 254 patients who underwent CMIS correction of ASD from January of 2011 to January of 2020 was retrospectively analysed. Adult spinal deformity was defined by the following criteria: Cobb angle > 20 or SVA > 50 mm or PI-LL > 10. Only patients who were placed on parathyroid hormone analogues for one year in the peri-operative period and with 2 year radiographic and clinical follow-up were included in the study. Informed consent was obtained from all patients. Indications for surgery included progressively worsening axial back pain with or without radicular symptoms or neurogenic claudication. All patients had undergone at least 6 months of conservative therapy prior to consenting to surgery. Institutional Review Board approval from Cedars-Sinai was obtained for this study.

Forty-seven patients were identified who were placed on anabolic PTH therapy in the peri-operative period. Of these, 41 patients had pre-operative and two-year post-operative computed tomography lumbar scans for review. Assessments of bone quality were quantified via a standardized HU measurement protocol (16). Hounsfield units (HU) were measured from L1 to L3 vertebral levels on all patients' pre-operative and post-operative CT scans within the PACS system. Measurements were taken from three different locations within the vertebral body and averaged to obtain a mean HU measurement for each vertebral body. Figures 1, 2 and 3. This methodology has previously been described [17]. Of note, CT scans are routinely obtained in all patients undergoing CMIS deformity correction in order to assess for any pre-existent segmental or facet fusion which would necessitate an osteotomy. Dual-energy X-ray absorptiometric scan (DEXA) was also done to document bone density. Of note, if the DEXA score was less than − 3.5, minimally invasive surgery was contraindicated. Furthermore, if the T-score was between − 2.0 and − 3.5, daily injection of PTH analogue (20 µg of teriparatide or 80 µg of abaloparatide) was immediately started as soon as the patient was scheduled for surgery and could be seen by endocrinology. In patients with a T-score > − 2.0, while PTH therapy was not routinely initiated, if their bone quality was deemed to be poor intra-operatively (surgeon discretion—based on tactile feedback), patients were immediately started on PTH therapy post-operatively. In all patients, PTH therapy was continued for 1-year post surgery.

Fig. 1
figure 1

Measure of Hounsfield Units off of a Lumbar Computed Tomography Scan at L1

Full size image
Fig. 2
figure 2

Measure of Hounsfield Units off of a Lumbar Computed Tomography Scan at L2

Full size image
Fig. 3
figure 3

Measure of Hounsfield Units off of a Lumbar Computed Tomography Scan at L3

Full size image

All patient demographic information and comorbidities were collected. Radiographic parameters including thoracic kyphosis (TK), lumbar lordosis (LL), pelvic tilt, pelvic incidence (PI), sagittal vertical axis (SVA), and coronal balance were measured pre-operatively, 3 months post-operatively, and at latest follow-up utilizing full length 36″ standing films. PI-LL was additionally calculated off pre-operative and post-operative images. Any peri-operative and post-operative complications were noted.

Patient reported outcome measures including the Visual Analogue Score (VAS), Oswestry Disability Index (ODI), and the Scoliosis Research Society-22 item survey (SRS-22) were collected pre-operatively, at one year, and at latest follow-up.

CMIS protocol

The CMIS protocol has previously been described in detail [18]. In brief, patients with ASD undergo two-staged deformity correction with an intervening 3-day interval: Stage 1) multilevel oblique LLIF and MIS L5-S1 oblique interbody fusion (OLIF) or anterior lumbar interbody fusion (ALIF) Stage 2) MIS pedicle screws with rod contouring and derotation/translation. For the posterior percutaneous rodding, 5.5 mm titanium alloy rods are utilized. Posterior segmental “pars-facet-pars” fusions are then performed at segments without interbody fusion (typically cephalad of L1-2). No patient had any posterior column osteotomy or anterior column release other than at L5-S1 where an ALIF would require release of the ALL. Patients are immediately ambulated after stage 1 and on post-operative day 2, a standing radiograph is obtained. These data are then utilized to help understand how much residual deformity correction is needed during the second stage. Furthermore, any persistence of radicular pain or neurogenic claudication is noted during this time. If symptoms are noted, a lumbar MRI is performed, and a targeted minimally invasive decompression is performed during the 2nd stage.

Imaging

All computed tomography (CT) imaging was obtained with two helical 64-channel scanners (750 HD & Revolution; Generic Electric). CT parameters were as follows: slice thickness 0.625 mm with 0.625 mm intervals, tube voltage (120 kVp), tube current (300 with auto-range of 100–700 mA), and a bone reconstruction algorithm (window width/window level, 2000/500). Of note, these parameters are in alignment with previously established protocol [17].

Statistical analysis

Descriptive statistics were calculated with standard deviations and ranges. Patient characteristics, for both groups were analysed with the use of Chi-square and student T tests. A chi-square test was used for categorical variables, and an independent student T test was used to assess continuous variables. A p < 0.05 was set as our measure of statistical significance.

Results

Patient data

Mean age of patients was 70 years of age (range 52–84; SD = 7). Mean follow-up was 66 months (range 24–132, SD = 33). Twenty-three patients met criteria for severe deformity (Cobb angle > 50 degrees or SVA > 95 mm or PI/LL mismatch > 20 or PT > 30). A relevant list of patient comorbidities potentially impacting bone density is presented in Table 1. The most common patient comorbidities were hypertension, hypercholesterolemia, and hypothyroidism.

Table 1 Patient comorbidities
Full size table

The pre-operative and post-operative spinopelvic parameters are listed within Table 2. Baseline mean pre-operative radiographic parameters were as follows: Coronal cobb of 37.5° (15–70.1, SD 16.1), lumbar lordosis 37° (16.6–68.7, SD 12.8), thoracic kyphosis 40.9° (15.6–61.3, SD 15.8), PI-LL mismatch 20.5° (1.8–49.3, SD 13.1), SVA 67.7 mm (10.5–177.8, SD 47.6), and AVT 42.1 mm (16.5–77.9, SD 18.5). All of these parameters had statistically significant improvement at final post-operative follow-up with a coronal cobb of 12.7° (0–29.4, SD 8.8); p = 0.00000000016), lumbar lordosis 47.8° (26.1–71.6, SD 10.1); p = 0.000001, thoracic kyphosis 48.9° (27.1–67.8, SD 11.7); p = 0.03, PI-LL mismatch 12.6° (0.4–28.5, SD 7.1; p = 0.0006, SVA 39.7 mm (0–163, SD 38.8); p = 0.0005, and AVT 16.3 (0–45.3, SD 12.7); p = 0.000007.

Table 2 Radiographic outcomes
Full size table

The mean pre-operative L1 bone density measurement improved significantly when compared to final 2-year follow-up values (p = 0.041), from 96 to 185HU. Mean pre-operative L2 vertebral body HU also improved from 138 to 169HU (p = 0.049). The L3 vertebral measurement improved from 151 to 183HU (p = 0.024). Five patients (12%) experienced no bone density increases despite completion of peri-operative PTH analogue therapy. See Table 3 for a comparison between pre-operative HU measurements to post-operative HU measurements.

Table 3 Hounsfield units per level
Full size table

Mean T-Score was − 2.2 (− 0.4 to − 3.1, SD 0.7). This includes nine patients with T-scores > − 2.0 who were determined to have poor bone quality intra-operatively by the senior author. These patients received PTH analogue therapy immediately following surgery.

All patients reported improvements in patient reported outcomes Table 4.

Table 4 Clinical outcomes
Full size table

There were 2 patients who developed PJF (4.8%). Both of these patients had incomplete PTH therapy. The first patient received PTH for only 1 month and stopped due to patient misunderstanding regarding the need for continued therapy. She ended up needing a revision fusion for symptomatic PJK 1 year post-operatively. The second patient received PTH for 3 months and stopped as she was unable to afford any more treatments. She ultimately developed a compression fracture at T11 with concomitant loosening of her hardware at the same level 2 years following surgery. These latter two patients’ pre-operative and post-operative radiographic parameters are presented in Table 5. In patients who continued PTH therapy for the full course, there were 6 cases of radiographic PJK by final follow-up. There were no incidences of screw loosening or screw pull out in this patient group (Table 6). Only one patient experienced nausea from PTH therapy. There were no other PTH related adverse events.

Table 5 Radiographic parameters of the two patients who did not complete PTH therapy
Full size table
Table 6 Mechanical complications
Full size table

Discussion

Common complications following ASD correction are the failure of implanted hardware to provide adequate stabilization and pseudarthrosis, which in turn are potential sources of significant post-operative morbidity and common reasons for revision surgery [19]. Even while comparatively lower with use of CMIS when compared to open deformity correction, the rate of hardware failure is still 10.1% in patients undergoing CMIS for ASD correction. Additionally, proximal junctional issues remain an issue following CMIS, with a cited prevalence of 2.7%. While these latter complications and procedural sequelae are often multifactorial in aetiology, poor bone quality has been frequently cited as a major contributing factor.

In this context, as patients with ASD, particularly females, are more likely to have metabolic bone disease, optimization of bone health remains a potential area of risk mitigation in patients undergoing deformity correction. Furthermore, as these patients tend to be less mobile and oftentimes require use of assistive aids immediately following surgery, they are intrinsically at higher risk of sustaining falls. Consequently, they are at high risk of sustaining fragility fractures which are also associated with significant morbidity and risk for mortality [20]. Therefore, bone health optimization in this patient population potentially yields an important secondary benefit in this regard. While various strategies currently exist that are theorized to improve bone quality in the peri-operative period, limited data exist that quantifies the ability of these aforementioned strategies to do so.

Pre-operative or peri-operative teriparatide and abaloparatide (Forteo and Tymlos) use has emerged in recent years as a novel strategy to improve bone quality in the pre- and peri-operative period in patients undergoing spinal fusion. Both are parathyroid hormone analogues that are administered in an intermittent fashion that results in an anabolic effect on bone formation by preferentially stimulating osteoblast activity. Data have shown that use of these anabolic agents decreases the risk of vertebral compression and hip fractures [21]. Furthermore, they have been shown to increase spinal bone mineral density in patients with osteoporosis [21, 22]. However, these latter findings were the result of approximately 2 years of continuous treatment with PTH analogue. In practice, however, this may be impractical, as patients who are in severe debilitating pain are oftentimes unable to delay surgery for 2 years for anabolic therapy. Therefore, in our practice, in patients with poor bone quality base off of T-scores (T-score − 2.0 to − 3.5), we have shortened the lag time to surgery following initiation of PTH with favourable results. In this regard, our study is one of the first, to study the effects of shorter term PTH therapy on bone quality, particularly in the setting of CMIS for ASD correction.

Significant improvements in the mean L1 Hounsfield units were noted when comparing pre-operative values (96; SD 55) to post-op values (185 SD 102). According to previously published data, this indicates a categorical change from osteoporosis (HU < 100) to normal mineral density (HU > 160) from pre-op to post-op in our patient population [17, 23]. Only 5 patients (12%) demonstrated no change in Hounsfield units after completing their PTH treatment. While two patients (4.8%) in this series ultimately developed PJF, both patients did not complete their full PTH analogue treatment. One only took PTH for 3 months (PJK at 2-year post-op) and stopped as she was unable to afford any more treatments. The other one took it only for 1 month (PJK at 1-year post-op) and stopped due to patient misunderstanding regarding the need for continued therapy. Of patients who continued with the full course of therapy, 6 developed radiographic PJK by final follow-up but there were no incidences of hardware failure including screw loosening, or screw pull out.

While the DEXA scan has become the standard for assessing bone density, there are some notable limitations. Specifically, as it an areal based measurement, it may not give accurate depictions of bone density in patients with smaller spines. Furthermore, accuracy may be limited in patients who are obese, with significant spondylosis and bony sclerosis, or those with concomitant aortic calcifications, as these latter characteristics falsely elevate DEXA scores [24, 25] In this context, the measurement of Hounsfield units on a lumbar CT has been proposed as a valid measure of bone quality. Significant correlations between the use of Hounsfield units and a traditional DEXA scan have been demonstrated. The former, however, offers the added benefit of being able to filter out sclerotic bone that may confound DEXA scores [24, 26]. Furthermore, as obtaining CT scans of the lumbar spine are oft a routine part of the pre-surgical planning, this offers a convenient and cheap modality to assess lumbar spine bone density in real-time.

Study limitations

There are some notable limitations with our study. Firstly, although data were pulled from a prospectively collected database, this was nonetheless, a retrospective study and is subject to all the respective inherent limitations. Surgeon selection bias in this particular patient population could have contributed to more optimal results. Of note, there is currently no defined protocol regarding the optimal timing of PTH analogue administration prior to surgery. In general, we chose to start patients on anabolic PTH therapy immediately after they were indicated for surgery. Consequently, the lead time for pre-operative PTH administration for patients ranged anywhere from approximately 2 to 4 weeks. The exception to this was, of course, in the 9 patients who were indicated for anabolic therapy based on the senior author's intra-operative assessment of bone quality. The selection of which PTH analogue the patient received was at the sole discretion of the treating endocrinologist. All patients received one year of continuous administration of PTH analogue post-operatively with the two notable exceptions previously discussed. Despite some inconsistency in the timing of administration of PTH in the peri-operative period, our study is still one of the first to study the effects of shorter term PTH therapy on bone quality, particularly in the setting of CMIS for ASD correction. Ultimately, future studies are warranted to protocolize the ideal peri-operative bone health optimization strategy. Further, with a low incidence of hardware failure and with inherent categorical heterogeneity further limiting the numbers of each mode of failure, more sophisticated multivariable risk-analyses were not feasible. Finally, clinicians should bear in mind that the calculation of Hounsfield Units (HU) relies heavily on various CT parameters including tube voltage, current, slice thickness and intervals, reconstruction algorithms, amongst other factors [17, 27, 28]. Importantly, we recommend that surgeons and researchers familiarize themselves with their own institutions’ scanner parameters and scanning protocols utilized. Nonetheless, in spite of these potential limitations, our study is one of the first of its size to study the effect of PTH analogues on bone quality in patients undergoing CMIS deformity correction using lumbar CT scans.

Conclusion

The incidence of PTH analogues failing to increase bone density in our series was low at 12%. This study shows that PTH analogues may be a powerful adjunct for increasing bone density and may help to mitigate the risk of mechanical complications in patients undergoing deformity correction with minimally invasive techniques. Future comparative studies are warranted to confirm these latter findings and to potentially protocolize the ideal peri-operative bone health optimization strategy.