Pedicle subtraction osteotomy using ultrasonic bone curette yields fewer durotomies, decreased blood loss: a single surgeon’s retrospective experience

Abstract

Background

Pedicle subtraction osteotomy (PSO) is a three column deformity correction technique that can achieve 30 to 40 degrees of correction and is best used for patients with severe rigid sagittal imbalances. Patients report a high degree of satisfaction with quality of life outcomes as a consequence of undergoing PSO. However, complications range between 12% and 40%, including dural tears or neurovascular injuries causing neurological deficit occurring as a consequence of high speed drilling or traction from the spooned retractors used to facilitate exposure. Ultrasonic bone curettes may provide a safer alternative to the drill and retractor approach. The purpose of this retrospective study was to compare outcomes between newer spoonless ultrasonic bone curettage and traditional high speed drilling with paravertebral spoon retractors for pedicle subtraction osteotomies.

Methods

All patients undergoing PSO with a single surgeon across two academic health institutions in the past eight years were reviewed. Preoperative variables obtained were demographics, prior spine surgeries, and etiology, while intraoperative predictor variables were ultrasonic bone curettage use and number of fused levels. The principal outcomes were surgical: duration of surgery, estimated blood loss (EBL), incidence of durotomies, and neurovascular injury. Secondary outcomes were postoperative: length of stay, need for revision surgery, minor complications, and major complications. Multiple logistic regression was used to evaluate associations with binary outcomes, while multiple linear regression evaluated correlations with continuous outcomes.

Results

Overall, 175 patients were reviewed. In this sample, 89 (51%) had prior surgeries, and 45 (26%) were treated with the ultrasonic bone curettage. Intraoperatively, the mean EBL was 1,105 mL (range, 350–3,000 mL), 29 (17%) of participants sustained durotomies, 14 (8%) experienced neurovascular injury, and the mean duration of surgery was 401 mins (range, 135–695 mins). Postoperatively, 35 (20%) experienced minor complications, 25 (14%) experienced major complications, and mean length of stay was 11 days (range, 2–80 days). Ultrasonic bone curettage use correlated with decreased blood loss (β =−270 mL, P=0.004) and was associated with decreased incidence of durotomy [odds ratio (OR) =0.29, P=0.042].

Conclusions

Use of spoonless ultrasonic bone curettage in PSO was associated with significantly less blood loss and fewer durotomies compared to traditional drilling in this single surgeon’s multi-institutional experience.

Highlight box.

Key findings

• Using ultrasonic bone curettage instead of traditional high speed drilling correlated with decreased blood loss (β =−270 mL, P=0.004) and was associated with decreased incidence of durotomy (odds ratio =0.29, P=0.042) for patients undergoing pedicle subtration osteotomy for a single surgeon across two institutions.

What is known and what is new?

• Pedicle subtraction osteotomy (PSO) can achieve 30 to 40 degrees of correction and is best used for patients with severe rigid sagittal imbalances. The complciation rate is high, 12–40%, many of which are dural tears or neurovascular injuries from drilling. Ultrasonic bone curettage (UBC) has been shown to reduce the incidence of durotomy compared to drilling in other spinal pathologies

• This manuscript adds that (I) UBC also decreases incidence of durotomy in PSO, (II) UBC decreases overall blood loss, which has not been demonstrated in previous literature, and (III) operating room time is not significantly different between UBC and traditional drilling.

What is the implication, and what should change now?

• The implication of this study is that ultrasonic bone curettage may make pedicle subtraction osteotomies safer. UBC should be considered as an alternative to high speed drilling and multi center, prospective, multi surgeon studies are needed to further interrogate this finding.

Introduction

Spinal osteotomies are of a group of techniques used by spine surgeons to correct varying degrees of deformities (1). Pedicle subtraction osteotomy (PSO) is a three column intervention that can achieve 30 to 40 degrees of correction and is best used for patients with severe rigid sagittal imbalances (2,3). Initially solely used in the treatment of ankylosing spondylitis (4), PSO has seen significant growth in cases and applications over the past decade. The increased utilization of PSO as a technique can be partially attributed to improved safety (5) and effectiveness (6): patients report a high degree of satisfaction with quality of life outcomes as a consequence of undergoing PSO (7). However, given the complexity and invasiveness of this technique, perioperative and postoperative complications remain frequent. Most recently, the overall complication rate has been estimated to be between 12% and 40% (8,9). The majority of these are intraoperative complications such as dural tears or neurovascular injuries, often occurring as a consequence of high speed drilling or traction from the spooned retractors used to facilitate exposure (10). The most common postoperative complication is neurologic deficit (3,8).

One innovation in surgical equipment that could improve PSO surgical outcomes is the ultrasonic bone curette. This device is designed to reduce damage to surrounding tissue from the mechanical and thermal energy generated by traditional high speed drill when cutting through bone while also reducing the risk of damage to adjacent neurovascular structures through dislodgement of the cutting surface, such as when traditional drills slide or chatter (11). A large review of safety outcomes across a heterogenous cohort of spinal decompressions established no significant difference in complications compared to conventional drilling (12). Furthermore, it has been associated with shorter operative times and reduced blood loss when compared to conventional high speed drilling in a number of applications, including endoscopic assisted craniosynostosis (13) and primary lumbar decompression (14). In some instances, its unique design with respect to precision and mechanism has allowed it to safely enable previously underutilized surgical techniques, such as total uncinectomy for complete anterior decompression of cervical nerve roots (15).

Our team has developed a technique for PSO by which the cancellous vertebral wedge and lateral cortical walls of the target vertebra are removed via ultrasonic curette, obviating the need for high speed drilling during this step as well as removing the need for spoon retraction and potentially reducing the risk of durotomy and neurovascular injury. The purpose of this study is thus to retrospectively compare outcomes of ultrasonic bone curettage-facilitated spoonless PSO to conventional drilling technique in a single surgeon’s experience at two academic medical centers. Positive findings in surgical outcomes would justify investigating the spoonless ultrasonic bone curettage impact on PSO outcomes prospectively. We present this article in accordance with the STROBE reporting checklist (available at https://jss.amegroups.com/article/view/10.21037/jss-2026-1-0020/rc).

Methods

Participants, inclusion criteria

This IRB-approved study (IRB 19-0726, Biomedical Institutional Research Board, Office of Human Research Ethics, UNC Chapel Hill; IRB Pro00090408 Biomedical Institutional Research Board, Office of Human Research Ethics, Duke University) reviewed patients of a single surgeon who underwent PSO at two academic medical centers in the past 8 years. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. Informed consent was not necessary given that all data was deidentified and retrospective. All surgeries in which a PSO was performed were included. PSO patients were identified initially by CPT code, after which the operative note was reviewed for confirmation.

Variables & outcomes

Chart review identified several preoperative, intraoperative, and postoperative areas of comparison. Preoperative characteristics were number of fused segments, etiology (infectious, degenerative, iatrogenic, traumatic), spinal region of involvement, and first vs. repeat spine surgery. Intraoperative variables were operative time in minutes, blood loss in mL, incidence of durotomy, and incidence of neurovascular injury. Postoperative outcomes included major complications (neurologic deficit, deep wound infection, myocardial infarction, pneumonia, respiratory distress, deep vein thrombosis, pulmonary embolus, death), minor complications (ileus, superficial infection, urinary tract infection, arrhythmias, epidural hematoma, and superficial deep vein thrombosis), and need for revision surgery. Complications were evaluated from postoperative day (POD) 0 to POD 90.

Intraoperative blood loss quantification

Intraoperative blood loss in this study was not derived from visual surgical estimation alone; rather, it was meticulously quantified using the Neptune Waste Management System (Stryker, Kalamazoo, MI, USA), a closed-circuit, digital suction-and-fluid management device that continuously aggregates suctioned intraoperative fluid into a sealed canister and displays cumulative volume in 20 mL increments, with a reported measurement accuracy of approximately ±50 mL—substantially superior than the ±10% accuracy of conventional graduated canisters and the 30–50% error inherent to purely visual estimation (16,17). Estimated blood loss (EBL) was then calculated by subtracting irrigation volume from the Neptune-measured total and adding gravimetrically weighed sponge/gauze blood content, an approach shown in prior comparative studies to approximate true blood loss more reliably than visual estimation (18,19).

Ultrasonic curettage-facilitated spoonless PSO technique

The ultrasonic curettage-facilitated spoonless technique is largely analogous to traditional technique, differing at the removal of the pedicle, cancellous vertebral wedge, and lateral cortical walls of the vertebra (Figures 1,2A-2C) (2). With conventional technique, a high power drill with paravertebral spoon retractors is used to ensure safe visualization of adjacent neurovascular structures. With the ultrasonic curette, which has no rotational force that could dislodge it from the target, spoonless retraction is not required for safe target removal, which is the primary difference from traditional PSO (Figures 1,2D). The remainder of the surgical steps are completed in the fashion of traditional PSO (2).

Figure 1.

Comparing PSO workflows. ALL, anterolateral ligament; PSO, pedicle subtraction osteotomy.

Figure 2.

PSO ultrasonic curette. PSO, pedicle subtraction osteotomy.

Statistical analysis

Statistical analyses were conducted in R and Graph Prism 10. Summary statistics were generated for the imaging outcomes and perioperative chart variables based on their frequency. Binary variables (such as occurrence of a complication) were processed by assigning values of 0 for “No” and 1 for “Yes” prior to multivariate analysis. Multiple logistic regression was used to evaluate associations between predictor variables and each of the specified nonparametric outcomes (durotomies, neurovascular injury, postoperative complications, need for repeat surgery), while multiple linear regression was used to evaluate correlations between predictors and each of the specified parametric outcomes (EBL, duration of surgery, length of stay). Interactions between ultrasonic bone curettage use and number of levels fused was evaluated for blood loss and operative time. No random effects were considered and complete case analysis applied for missing data. No sensitivity analyses were conducted.

Results

Comparing preoperative outcomes (Table 1)

Table 1. Comparing preoperative characteristics.

Variable	Traditional approach	Ultrasonic curette	T score	P value
Sample size, n	130	45	n/a	n/a
Number of fused segments (mean)	7.07	6.94	0.36	0.72
Infectious (%)	36	28	0.96	0.34
Degenerative (%)	40	43	−0.36	0.72
Iatrogenic (%)	4	13	−2.01	0.05
Traumatic (%)	20	16	0.56	0.58
Cervical involvement (%)	4	7	−0.65	0.52
Thoracic involvement (%)	69	65	0.52	0.60
Lumbar involvement (%)	91	83	1.5	0.14
Prior operations (%)	62	47	1.79	0.08

n/a, not applicable.

One hundred and thirty patients underwent traditional drilling while 45 were treated with the ultrasonic bone curette. Baseline demographic and structural characteristics were largely similar between groups. The mean number of fused segments was comparable (7.07 vs. 6.94, P=0.72). Distribution of etiologies, including infectious (36% vs. 28%, P=0.34), degenerative (40% vs. 43%, P=0.72), and traumatic (20% vs. 16%, P=0.58), did not significantly differ. The only variable approaching statistical significance was the incidence of iatrogenic deformity, which trended lower in the ultrasonic group (4% vs. 13%, P=0.05). Cervical, thoracic, and lumbar involvement rates were not significantly different, though lumbar pathology was numerically more frequent in the ultrasonic group (91% vs. 83%, P=0.14). Prior operations were more common in the ultrasonic group (62% vs. 47%), but this difference did not reach statistical significance (P=0.08).

Principal (intraoperative) outcomes (Tables 2,3)

Table 2. Outcome summary data.

Outcome	Ultrasonic curette (n=45)			Traditional drilling (n=130)
	Mean	SD		Mean	SD
Operative time (min)	384.87	111.79		406.12	133.70
Surgical bleeding (mL)	914.44	420.58		1,183.46	628.69
Durotomy (%)	7	n/a		20	n/a
Neurovascular injury (%)	2	n/a		4	n/a
Major complications (%)	9	n/a		10	n/a
Minor complications (%)	9	n/a		18	n/a
Length of stay (days)	10.07	9.48		11.39	10.39
Repeat surgery (%)	22	n/a		24	n/a

n/a, not applicable; SD, standard deviation.

Table 3. Relationships of variables to primary (intraoperative) outcomes.

Variable	Primary outcomes, statistic (95% CI), P value
	Estimated blood loss†	Duration of surgery†	Durotomy‡
Use of ultrasonic bone curettage	β =−270 (−453, −87), P=0.004**	β =−20 (−63, 23), P=0.35	OR =0.29 (0.07, 0.90), P=0.042*
Number of fused levels	β =116 (77, 154), P<0.001***	β =17 (8, 26), P<0.001***	OR =1.04 (0.85, 1.23), P=0.51
Thoracic involvement	β =−190 (−376, −4), P=0.049*	β =−40 (−85, 3), P=0.07	OR =0.71 (0.26, 1.92), P=0.36
Lumbosacral involvement	β =105 (−128, 347), P=0.39	β =−31 (−25, 87), P=0.28	OR =0.74 (0.22, 2.71), P=0.41
Prior operations	β =−101 (−263, 60), P=0.21	β =24 (−61, 14), P=0.21	OR =−1.44 (0.62, 3.4), P=0.48

^†, assessed via multiple regression; ^‡, assessed via multiple logistic regression. *, P<0.05; **, P<0.01; ***, P<0.001. CI, confidence interval; OR, odds ratio.

All principal outcome summary data are summarized in Table 2. For the traditional drilling group, the mean operating room time was 406 min, the mean blood loss was 1,183 mL, the incidence of durotomy was 20%, and the incidence of neurovascular injury was 4%. For the ultrasonic curette group, the mean operating room time was 385 minutes, the mean blood loss was 914 mL, the incidence of durotomy was 7%, and the incidence of neurovascular injury was 2%. Table 3 presents the statistics and relevant P values. For intraoperative EBL, multiple regression identified significant positive with number of fused levels (β =116, P=0.004). Significant negative correlations were identified with ultrasonic bone curettage use (β =−270, P=0.004). Variables that did not significantly correlate with EBL included thoracic involvement, lumbosacral involvement, and prior surgeries.

Multiple regression analysis of duration of surgery identified significant positive correlations with number of fused levels (β =17, P<0.001). There were no significant negative correlations. Variables that did not significantly correlate with duration of surgery included use of spoonless ultrasonic bone curettage, thoracic involvement, lumbosacral involvement, and prior operations.

With respect to durotomies, multiple logistic regression identified no positive associations. Spoonless ultrasonic bone curettage was negatively associated with durotomy [odds ratio (OR) =0.29, P=0.042]. Variables that did not significantly associate with durotomies included number of fused levels, thoracic involvement, lumbosacral involvement, and prior operations.

Lastly, for neurovascular injury, there was insufficient sample size to analyze relationships using multiple logistic regressions. Attempting to incorporate etiologies (infectious, degenerative, traumatic, iatrogenic, tumor) into the models raised issues with multicollinearity, and so these variables were excluded. There was insufficient sample size to evaluate the relationship of cervical involvement.

Secondary (postoperative) outcomes (Tables 2,4)

Table 4. Relationships of variables to secondary (postoperative) outcomes.

Variable	Secondary outcomes, statistic (95% CI), P value
	Length of stay†	Minor complications‡	Major complications‡	Revision surgery‡
Use of ultrasonic bone curettage	β =−1.5 (−5, 2.0), P=0.40	OR =0.48 (0.13, 1.37), P=0.51	OR =0.96 (0.25, 3.00), P=0.11	OR =0.88 (0.37, 1.99), P=0.16
Number of fused levels	β =−0.3 (−1.0, 0.5), P=0.47	OR =1.10 (0.90, 1.34), P=0.68	OR =1.18 (0.93, 1.51), P=0.56	OR =1.2 (0.97, 1.38), P=0.51
Thoracic involvement	β =3.4 (−0.2, 7.0), P=0.06	OR =0.64 (0.24, 1.72), P=0.13	OR =0.64 (0.18, 2.26), P=0.10	OR =0.85 (0.36, 2.04), P=0.48
Lumbosacral involvement	β =−0.5 (−5.2, 4.2), P=0.82	OR =1.17 (0.31, 5.72), P=0.09	OR =0.54 (0.13, 2.83), P=0.08	OR =0.53 (0.19, 1.56), P=0.46
Prior operations	β =0.78 (−2.3, 3.9), P=0.62	OR =0.61 (0.25, 1.44), P=0.15	OR =0.75 (0.26, 2.11), P=0.13	OR =1.57 (0.75, 3.31), P=0.53

^†, assessed via multiple regression; ^‡, assessed via multiple logistic regression. CI, confidence interval; OR, odds ratio.

All principal outcome summary data are summarized in Table 2. For the traditional drilling group, the incidence of major complications was 10%, the incidence of minor complications was 18%, the mean length of stay was 11 days, and the incidence of repeat surgery was 24%. For the ultrasonic curette group, the incidence of major complications was 9%, the incidence of minor complications was 9%, the mean length of stay was 10 days, and the incidence of repeat surgery was 22%. All secondary outcome statistics and relevant p values are presented in Table 4. For length of stay, multiple regression identified no significant positive or negative correlations. Variables that did not significantly correlate were use of ultrasonic bone curettage, number of fused levels, thoracic involvement, and lumbosacral involvement.

Multiple logistic regression analysis of minor complications, major complications, and revision surgery as outcomes identified no significant positive or negative associations. Variables analyzed that did not significantly associate with these outcomes were use of spoonless ultrasonic bone curettage, number of fused levels, thoracic involvement, and lumbosacral involvement. Once again, attempting to incorporate etiologies (infectious, degenerative, traumatic, iatrogenic, tumor) into the models was compromised by multicollinearity, and so these variables were excluded. Additionally, there was once again insufficient sample size to evaluate the relationship of cervical involvement.

Discussion

This retrospective study compared outcomes between spoonless ultrasonic bone curettage and traditional drilling and spoon retraction for PSO in a single surgeon’s experience at two academic medical centers over the past 8 years. While a number of studies have evaluated perioperative and postoperative complications for PSO, and yet others have established the safety and efficacy of ultrasonic bone curettage in various neurosurgical procedures, to date, none have evaluated application of the device to PSOs. Furthermore, none have published extensively on modifying the PSO technique to remove spoon retraction, a potential source of neurovascular injury.

Outcomes

Our perioperative data has both distinct similarities and interesting differences to other reports. The mean operative time was 6.6 hours in this series, which in other reports have ranged from 7.3 to 12.2 hours. This finding can partially be attributed lower number of levels fused on average in this study, 7, compared to other studies, which have ranged from 8 to 10 (8,9,20). A positive finding of this study was the association of ultrasonic curettage use with decreased blood loss. Indeed, the average blood loss in this study, 1,105 mL, is lower than that of other reports, which have ranged from 1,515 to 3,500 mL (8,9,20). The association between use of ultrasonic bone curettage and decreased blood loss has been established in a number of different investigations (11,13,14), and the protective effect is likely magnified in longer procedure like PSO. The reason why ultrasonic bone curettage can significantly reduce blood loss is derived from its intrinsic mechanism. This device works by linear oscillation of a thin blade at 22.5 kilohertz ultrasonic frequency, which generates heat that cauterizes vessels in the bone as it cuts, minimizing the contribution of osteotomy to surgical blood loss (21). Number of fused levels, being associated with lower surgical duration, could have additionally contributed to the lower average EBL.

One of the most interesting findings of this investigation was the association of the spoonless ultrasonic bone curettage technique with fewer intraoperative durotomies. Although hypothesized to reduce the risk of this complication, existing studies in other spinal procedures have not shown a significant difference in dural tears between ultrasonic bone curettage and traditional drilling (11,14). In our dataset, a total of 17% of patients experienced a durotomy which is similar to other PSO studies which have shown occurrence of dural tears to range between 18% and 25% (8,10), though the incidence of durotomy in the cohort that underwent ultrasonic bone curettage was a mere 7%. Again, the mechanism of the ultrasonic bone curettage through linear oscillation offers an explanation. By eliminating the rotary macromotion of a drill’s burr, the ultrasonic bone curettage reduces the risks of slipping off the cutting surface, entrapping important soft tissues, and penetrating too deeply (22).

Our multiple regression did not identify any associations of the selected variables major postoperative complications, minor postoperative complications, or revision surgeries. Previous investigations of these complications in other surgeries showed no difference in these outcomes comparing ultrasonic bone curettage to conventional high speed drilling (11,12,14). The incidence of major and minor complications, 14% and 20%, respectively, for a combined incidence of 34% is comparable to existing PSO literature, in which estimations have ranged from 12% to 40% in recent years (8,9).

Limitations

The principal limitation of this study was the lack of radiographic data in comparing the degree of correction between the two techniques. Another limitation is its retrospective design, which is inherently subject to selection bias, missing data, and unmeasured confounders. Multivariate regression is prone to confounding, particularly in such a heterogenous dataset, and given thus associations should be interpreted with caution. Given the heterogeneity of the cohort, more complex and synergistic relationships between variables may be obfuscated, and other important variables contributing to these outcomes may have been omitted. Furthermore, the use of multiple statistical comparisons increases the risk of type I error, and no formal correction was applied. As this study reflects the experience of a single surgeon, generalizability is limited, and surgical technique or experience may confound the effect of ultrasonic bone curettage. The study may also have been underpowered for major complications, minor complications, and need for revision surgery. Future prospective, multi-surgeon studies with radiographic outcome comparison, power calculations, and appropriate adjustment for multiple testing will help validate and extend these findings.

Conclusions

Use of a spoonless technique facilitated by ultrasonic bone curettage of the target vertebral body was associated with fewer durotomies and decreased EBL compared to traditional high speed drilling in this single surgeon’s multi-institutional experience with PSO. While decreased blood loss with ultrasonic curette has been established in the literature, this is one of the first studies to suggest a relationship with improved outcomes with respect to durotomies as well as the first to demonstrate safety and efficacy without the use of spoon retraction. While no associations were found with minor or major postoperative complications, this study may not have been sufficiently powered to uncover those relationships given their relative rarity. Additionally, this study did not analyze radiographic correction, a significant limitation for comparison of the two techniques, or rate of fusion. Further study shall explore the effect of ultrasonic bone curettage on radiographic and clinical outcomes prospectively across multiple surgeons to increase external validity and strengthen causality of these results.

Supplementary

The article’s supplementary files as

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study has been approved by the Ethics Committees (IRB 19-0726, Biomedical Institutional Research Board, Office of Human Research Ethics, UNC Chapel Hill; IRB Pro00090408 Biomedical Institutional Research Board, Office of Human Research Ethics, Duke University). The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. Informed consent was not necessary given that all data was deidentified and retrospective.

Data Sharing Statement

Available at https://jss.amegroups.com/article/view/10.21037/jss-2026-1-0020/dss

jss-12-08-134-dss.pdf ^{(71.7KB, pdf)}

DOI: 10.21037/jss-2026-1-0020