Defining a Surgical Invasiveness Threshold for Increased Risk of a Major Complication Following Adult Spinal Deformity Surgery

Brian J. Neuman, MD; Andrew B. Harris, BS; Eric O. Klineberg, MD; Richard A. Hostin, MD; Themistocles S. Protopsaltis, MD; Peter G. Passias, MD; Jeffrey L. Gum, MD; Robert A. Hart, MD; Michael P. Kelly, MD; Alan H. Daniels, MD; Christopher P. Ames, MD; Christopher I. Shaffrey, MD; Khaled M. Kebaish, MD

Disclosures

Spine. 2021;46(14):931-938. 

In This Article

Results

Patient and Treatment Characteristics

Mean (± standard deviation) patient age was 60 ± 15 years, and 78% of patients were women. Half of patients (n = 287) had undergone previous spine surgery. Patients had the following mean preoperative radiographic parameters: SVA, 6.4 ± 7.3 cm; PI-LL, 16° ± 21°; PT, 24° ± 11°; SS, 31° ± 12°; TK, 15° ± 0.88°; and LL, 23° ± 0.61° (Table 2).

Mean preoperative ODI value was 44 ± 18, and SRS-22r score was 2.8 ± 0.7. By the 2-year postoperative timepoint, the mean ODI value improved to 28 ± 20 (P < 0.001), and the mean SRS-22r score improved to 3.7 ± 0.8 (P = 0.034).

The mean number of spinal levels fused was 11 ± 4.3, and 22% of patients (n = 127) underwent three-column osteotomy (Table 2). Mean estimated blood loss (EBL) was 1651 ± 1492 mL, and mean ASD-SR score was 91 ± 36 points. A total of 422 patients (74%) were admitted to the intensive care unit postoperatively.

Nonresponder Analysis

We performed a nonresponder analysis, comparing the final analytic cohort with patients who were eligible for 2-year follow-up but did not have complete follow-up data. Compared with included patients, those lost to follow-up were slightly older (61 vs 59 years, P < 0.010), had greater SVA (7.5 ± 2.9 cm vs. 6.4 ± 7.3 cm, P = 0.006), and were less likely to be female (69% vs. 79%, P < 0.001). We found no significant differences between patients who were included versus those lost to follow-up in terms of preoperative ODI value (P = 0.230) or SRS-22r Total score (P = 0.104).

ASD-SR Cutoff Score

An overall ASD-SR cutoff score of 90 was determined (Figure 1), with 278 patients (48%) above or equal to the cutoff, and 296 patients (52%) below the cutoff. Example radiographs are shown for patients above and below the ASD-SR cutoff score (Figure 2A-H).

Figure 1.

Receiver operating characteristic (ROC) curve showing the cutoff point that optimizes sensitivity and specificity for predicting major complications after adult spinal deformity surgery.

Figure 2.

Radiographs of two patients who underwent adult spinal deformity procedures that were below and above the Adult Spinal Deformity Surgical and Radiographic (ASD-SR) cutoff score of 90 points, respectively. Preoperative anteroposterior (A) and lateral (B) and postoperative anteroposterior (C) and lateral (D) radiographs of a patient whose procedure had an ASD-SR score of 63. This patient underwent primary arthrodesis from T9 to the pelvis (30 points) with 6 Smith-Petersen osteotomies (6 points) and two levels of decompression (2 points), iliac fixation (2 points), and anterior lumbar interbody fusion (ALIF) (8 points). Treatment resulted in a 1.1-cm increase in sagittal vertical axis (0.2 points), a 9° decrease in thoracic kyphosis (4.5 points), a 2° decrease in pelvic incidence minus lumbar lordosis, and a 4.7° increase in pelvic tilt (9.14 points). The patient had no major postoperative complications. Preoperative anteroposterior (E) and lateral (F) and postoperative anteroposterior (G) and lateral (H) radiographs of a patient whose procedure had an ASD-SR score of 103. This patient underwent revision arthrodesis (3 points) from T10 to the pelvis (27 points), involving a 3-column osteotomy (14 points), iliac fixation (2 points), and ALIF (8 points). Treatment resulted in a 13-cm decrease in sagittal vertical axis (2.6 points), a 35° increase in thoracic kyphosis (17.5 points), a 31° decrease in pelvic incidence minus lumbar lordosis (15.5 points), and a 6.6° decrease in pelvic tilt (13.2 points). The patient had a major complication (pulmonary embolism) on postoperative day 4, which was treated appropriately and eventually resolved.

Patients above the cutoff had a higher rate of any complication (68%) than did patients below the cutoff (62%) (P < 0.001). Patients above the cutoff also had a higher rate of radiographic proximal junctional kyphosis at any point (48%) versus those below the cutoff (33%) (P < 0.001). Finally, patients above the cutoff had a higher rate of having any complication requiring reoperation (28%) than did patients below the cutoff (22%) (P = 0.008; Table 3).

Mean improvements in ODI value and SRS-22r Total score were not significantly different between the high- and low-invasiveness groups (P = 0.943, P = 0.110, respectively; Table 3).

Major Complications

Patients above this cutoff score had a significantly higher prevalence of having any major complication (36%) versus patients below the cutoff (19%) (P < 0.001). Patients above the cutoff also had a higher prevalence of the following specific major complications: neurologic deficit (19% vs. 11%, P = 0.008), pseudarthrosis with instrumentation failure (9.4% vs. 4.4%, P = 0.018), and excessive blood loss (8.1% vs. 4.0%, P = 0.014). We found no significant differences in the prevalence of deep vein thrombosis, pulmonary embolism, infection requiring irrigation and debridement, or pneumonia above versus below the cutoff (all, P > 0.05).

Minor Complications

Patients above the ASD-SR cutoff score did not have a significantly higher prevalence of any minor complication (P = 0.266). Patients above the cutoff did, however, have a higher prevalence of incidental durotomy (12%) compared with those below the cutoff (6.1%) (P = 0.005). We found no significant differences in the prevalence of neurological deficit, infection, pleural effusion, ileus, implant loosening, or urinary tract infection (all, P > 0.05).

Comparison of Quartile Cutoffs

Patients were divided into quartiles by ASD-SR scores as follows: 0 to 65 points (Q1); 66 to 89 points (Q2); 90 to 119 points (Q3); and ≥120 points (Q4). Q1 consisted of 145 patients (24%); Q2 of 149 patients (24%); Q3 of 162 patients (26%); and Q4 of 154 patients (24%). When comparing the odds of experiencing a major complication by adjacent quartiles, we found a significant increase between Q2 and Q3 (odds ratio, 1.8; 95% confidence interval: 1.0–3.0), but no such increases from Q1 to Q2 or from Q3 to Q4 (P < 0.05). Similarly, patients above the ASD-SR cutoff score of 90 (Q3 or Q4) were 1.9 times as likely to have a major complication than were patients below this threshold (Q1 or Q2) (odds ratio [OR], 1.9; 95% confidence interval [CI]: 1.3–2.9) (Figure 3). We found no significant differences in the odds of having any minor complication between invasiveness groups (all P > 0.05).

Figure 3.

Odds of developing a major complication across invasiveness quartiles. Patients were divided into quartiles according to the Adult Spinal Deformity Surgical and Radiographic (ASD-SR) invasiveness score as follows: 0 to 65 points (Q1); 66 to 89 points (Q2); 90 to 119 points (Q3); and ≥120 points (Q4). When comparing the odds of experiencing a major complication across adjacent quartiles, we found a significant increase between Q2 and Q3 (odds ratio, 1.8; 95% confidence interval: 1.0–3.0), and nonsignificant changes from Q1 to Q2 and from Q3 to Q4 (P<0.05). CI indicates confidence interval; OR, odds ratio; n.s., not significant. *Indicates significance with P<0.05.

We also analyzed the preoperative to 2-year postoperative changes in ODI value and SRS-22r Total score in relation to these quartiles. We found no significant differences in ODI values (P = 0.683) or SRS-22r Total scores (P = 0.906) between quartiles.

Comments

3090D553-9492-4563-8681-AD288FA52ACE

processing....