Is vertebral bone quality an independent predictor of total blood loss in transforaminal lumbar interbody fusion?

Abstract

BACKGROUND

Few studies have analyzed the correlation between magnetic resonance imaging-based vertebral bone quality (VBQ) and blood loss in transforaminal lumbar interbody fusion (TLIF).

AIM

To determine the effect of bone mineral density, as assessed by preoperative VBQ on intraoperative total blood loss (TBL) following TLIF under real world conditions.

METHODS

We retrospectively analyzed patients who received TLIF for lumbar degenerative diseases at our hospital from January 2023 to April 2024. The preoperative VBQ score and T-score were recorded, and data regarding surgical procedures and blood loss were collected. TBL during postoperative the first three days was the primary indicator, and secondary indicators included length of hospitalization, hematocrit and hemoglobin, VBQ score, and T-score.

RESULTS

A total of 247 patients (93 males and 154 females) were enrolled in this study, with a mean age of 59.59 ± 9.09 years. VBQ were measured in all patients’ preoperative lumbar magnetic resonance imaging scans. The value for VBQ (L1-4 median), VBQ (L1-4 average), VBQ (S1) and VBQ (fixation segments) was 3.50 ± 0.70, 3.49 ± 0.67, 3.19 ± 0.72, 3.42 ± 0.71, respectively. Among them, 71 patients had preoperative dual-energy X-ray absorptiometry scan. The average lowest T-score was -2.25 ± 1.29. The T-score was significantly correlated with the VBQ score (r = -0.331 to -0.419). The mean TBL over the first three postoperative days was 796.95 ± 569.94 mL. Controlling for age, number of fixed segments, preoperative platelet count, and coagulation testing, VBQ (S1) was correlated with perioperative TBL in TLIF.

CONCLUSION

Preoperative VBQ (S1) score may be correlated with TBL in TLIF. In multi-segment TLIF cases, greater TBL was associated with considerable perioperative hemorrhage, potentially leading to longer hospital stays.

Key Words: Bone mineral density; Blood loss; Dual-energy X-ray absorptiometry; Osteoporosis; Vertebral bone quality; Transforaminal lumbar interbody fusion

Core Tip: The current study found preoperative magnetic resonance imaging-based vertebral bone quality (VBQ) score, particularly VBQ (S1), is independently correlated with total blood loss in patients undergoing transforaminal lumbar interbody fusion (TLIF). In multi-segment TLIF cases, greater blood loss is associated with increased perioperative hemorrhage and may lead to prolonged hospital stays. These findings suggest that VBQ assessment could serve as a useful preoperative tool for predicting bleeding risk and optimizing perioperative management in TLIF patients.

INTRODUCTION

Posterior fusion surgery for lumbar degenerative disease (LDD) has usually been associated with an increase in perioperative blood loss, leading to insufficient tissue perfusion or even coagulation dysfunction, thereby affecting cardiac, pulmonary and renal status[1]. Intraoperative blood loss exceeding 500 mL was associated with an increased risk of postoperative complications and prolonged hospital stays following lumbar fusion surgery[2]. Major bleeding could be compromised by intra- or postoperative blood transfusion, which increased operative time, the probability of transfusion immune reaction and infection, and the cost-related burden[3,4], so that accurate assessment of perioperative blood loss was of vital importance.

Transforaminal lumbar interbody fusion (TLIF) has been the standard technique used for the treatment of LDD, for its advantages in the complication rate, blood loss, and operation duration over posterior lumbar interbody fusion[5]. It is crucial to identify risk factors for perioperative blood loss after TLIF and ultimately rectify these predisposing factors in patients who require surgery. Previously identified risk factors for blood loss include age, muscle thickness, American Society of Anesthesiologists classification, and hematological parameters such as hematocrit (HCT) and fibrinogen levels[6]. Previous studies have found that low bone mineral density (BMD) was an independent risk factor for major blood loss in adult spinal deformity surgery and hidden blood loss in minimally invasive-TLIF[7,8], few studies have analyzed the correlation of BMD and total blood loss (TBL) following TLIF[9].

BMD is a measurement of the amount of bone mineral within bone tissue. At present, dual-energy X-ray absorptiometry (DEXA) scans are the gold standard imaging tool used for measuring BMD[10]. However, only 27% patients routinely obtained DEXA before instrumented fusion and only 44% of the 114 surgeons included in a survey would arrange preoperative DEXA for patients[11]. As DEXA was prone to error in patients with vascular calcification[12], obesity[13], spinal degenerative disease or previous spinal surgery[14], vertebral bone quality (VBQ) score based on magnetic resonance examination was proposed as a new method to measure BMD[15]. VBQ has been proven to be a predictor of surgical complications, such as cage subsidence[10], vertebral fracture[16], and a risk factor for reoperation after lumbar fusion surgery[17]. No previous study has analyzed the relationship between VBQ score and perioperative blood loss following TLIF. Thus, we conducted this real-world analysis to determine the effect of magnetic resonance imaging (MRI)-based VBQ on perioperative blood loss of TLIF.

MATERIALS AND METHODS

Study design

A retrospective study was designed to analyze the correlation of BMD and perioperative blood loss of TLIF with real world data. This study was approved by the local Human Ethics Committee of 960^th Hospital of PLA (approval No. 2025KYLL189) and was conducted in accordance with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. Informed consent was obtained from all participants.

Participants

All clinical data was retracted from a single medical center. Inclusion criteria were patients: (1) Diagnosed lumbar disc herniation, lumbar spinal stenosis or lumbar spondylolisthesis with surgical indications; (2) Underwent TLIF surgery; (3) All performed by the same senior surgeon Huang WM; (4) Acquired stable intraoperative and postoperative fluid shift and hemodynamics; and (5) All had preoperative lumbar MRI in our hospital.

Exclusion criteria comprised the following: (1) Presence of spinal tumors or infections; (2) Concurrent hematologic disorders, including severe anemia or coagulopathy; (3) Use of antiplatelet or anticoagulant medications; and (4) Intraoperative dural tear. Consequentially, 247 patients with LDD were included in a period from January 2023 to April 2024. Among them, 71 patients had preoperative DEXA of the left hip in our hospital.

VBQ score calculation

All MRIs included in this study were preoperative T1-weighted MRIs of the lumbar spine. VBQ score was calculated by placing regions of interest within the medullary portions of the L1-S1 vertebral bodies and within the cerebrospinal fluid space at the level of L3, as indicated on the mid-sagittal slice. Each elliptical regions of interest was positioned 3 mm from the vertebral edge in the anterior half of the vertebra to exclude cortical bone[15] (Figure 1).

Figure 1  Illustrative image of the regions-of-interest used to calculate vertebral bone quality score.

VBQ (L1-4 median) = Median SI_L1-L4/SI_{cerebrospinal fluid}

VBQ (L1-4 average) = Average SI_L1-L4/SI_{cerebrospinal fluid}

VBQ (S1) = SI_S1/SI_{cerebrospinal fluid}

VBQ (fixation levels) = Average SI_{fixation levels}/SI_{cerebrospinal fluid}

SI_L1-L4: The signal intensity of L1-L4 vertebral bodies.

SI_CSF: The signal intensity of cerebrospinal fluid.

Data extraction

Demographic information such as gender, age, height, weight, body mass index (BMI), preoperative coagulation function, platelets accounts, and erythrocyte sedimentation rate (ESR); surgery parameters such as the levels of surgical fixations, and parameters including intraoperative blood loss, HCT and hemoglobin (Hb) levels (preoperative and on postoperative days 3 and 7), and the volumes of autologous and allogeneic blood transfusion were assessed.

Primary and secondary outcomes

TBL was regarded as the primary outcome and the secondary outcomes were pre- and postoperative HCT and Hb, preoperative coagulation function and ESR, VBQ score from MRI scan, and T-score from DEXA.

Blood loss management and calculation

Blood loss during surgery was measured by suction apparatus, which was determined by an anesthesiologist, nurse, and surgeon. To prevent deep vein thrombosis, lower extremity pumps and identical anticoagulant drugs were administered postoperatively. In addition, a subfascial drainage tube was routinely placed and removed within 72 hours after surgery.

TBL is defined as the sum of intraoperative blood loss, postoperative drainage and hidden blood loss. It is calculated using Nadler’s formula for patients’ blood volume (PBV): PBV = k1 × height (m)³ + k2 × weight (kg) + k3, k1 = 0.3669, k2 = 0.03219, k3 = 0.6041 for male and k1 = 0.3561, k2 = 0.03308, k3 = 0.1833 for female. According to the Gross formula, TBL = PBV (HCT_pre - HCT_post)/HCT_ave, where HCT_pre was preoperative HCT, HCT_post was HCT on the third day after surgery, and HCT_ave was the average of HCT_pre and HCT_post[18].

Statistical analysis

Data analysis was performed using SPSS 22.0 software (International Business Machines Corporation, Armonk, New York, United States). Continuous variables were described as means and standard deviations for normally distributed data. Categorical variables were represented as absolute numbers. For comparing continuous variables across groups, one-way analysis of variance was applied, whereas for comparing categorical variables across groups, the χ² test was utilized. The Pearson correlation coefficient or Spearman’s rank correlation coefficient was employed to demonstrate whether a correlation existed between each indicator and the outcome variable. The multiple linear regression model was established to identify the risk factors of TBL. P < 0.05 was considered statistically significant.

RESULTS

A consecutive of 247 patients with LDD were enrolled in this study. TLIF was performed by senior surgeon Huang WM. The mean age and BMI were 59.59 ± 9.09 years old and 25.85 ± 3.31, respectively. Of 172 patients had one level fixation, 60 patients had double levels fixation and 15 patients had ≥ 3 levels fixations (Table 1).

Table 1 Demographic information of all patients, mean ± SD.

	Total	1 level	2 levels	≥ 3 levels	P value
n	247	172	60	15
Male/female	93/154	70/102	17/43	6/9	0.231
Age (year)	59.59 ± 9.09	59.58 ± 9.45	60.03 ± 8.75	57.93 ± 5.82	0.727
Hospital stays (day)	14.23 ± 3.81	14.07 ± 3.07	14.23 ± 4.45	16.13 ± 4.53	0.133
BMI (kg/m2)	25.85 ± 3.31	25.57 ± 3.14	26.61 ± 3.76	27.57 ± 2.86	0.049
Fixation segments	1.37 ± 0.62	1	2	3.13 ± 0.35
Hb (preoperative, g/L)	135.64 ± 14.93	136.39 ± 15.81	133.85 ± 11.91	134.13 ± 15.57	0.486
HCT (preoperative, %)	40.18 ± 3.85	40.34 ± 4.08	39.72 ± 3.08	40.10 ± 3.96	0.556
ESR (mm/hour)	14.27 ± 11.87	14.03 ± 10.69	15.02 ± 15.24	14.00 ± 9.79	0.855
PT (second)	11.34 ± 0.84	11.45 ± 0.89	11.12 ± 0.65	11.09 ± 0.69	0.014
APTT (second)	27.51 ± 4.30	26.99 ± 4.19	28.80 ± 4.17	28.29 ± 5.04	0.014
INR	0.98 ± 0.08	0.98 ± 0.08	0.97 ± 0.07	0.95 ± 0.53	0.352
TT (second)	16.84 ± 2.30	17.12 ± 2.29	16.19 ± 2.16	16.34 ± 2.42	0.017
FIB (g/L)	2.81 ± 0.70	2.78 ± 0.67	2.85 ± 0.77	3.01 ± 0.68	0.422
D-dimer (mg/L)	0.49 ± 0.83	0.53 ± 0.94	0.37 ± 0.33	0.58 ± 0.79	0.405
PLT (109/L)	235.16 ± 59.11	232.73 ± 57.07	243.65 ± 67.26	229.00 ± 45.98	0.431
TBL (mL)	796.95 ± 569.94	761.26 ± 640.40	827.04 ± 323.71	1085.84 ± 384.73	0.095
VBQ (L1-4 median)	3.50 ± 0.70	3.58 ± 0.72	3.39 ± 0.67	3.15 ± 0.44	0.025
VBQ (L1-4 average)	3.49 ± 0.67	3.55 ± 0.69	3.40 ± 0.66	3.14 ± 0.41	0.039
VBQ (S1)	3.19 ± 0.72	3.21 ± 0.73	3.21 ± 0.71	2.87 ± 0.49	0.197
VBQ (fixation levels)	3.42 ± 0.71	3.46 ± 0.73	3.37 ± 0.69	3.08 ± 0.41	0.120

BMI: Body mass index; Hb: Hemoglobin; HCT: Hematocrit; ESR: Erythrocyte sedimentation rate; PT: Prothrombin time; APTT: Activated partial thromboplastin time; INR: International normalized ratio; TT: Thrombin time; FIB: Fibrinogen; PLT: Platelet count; TBL: Total blood loss; VBQ: Vertebral bone quality.

All patients had normal preoperative coagulation function and ESR. The postoperative 3^rd day’s TBL (including intraoperative blood loss, drainage, and hidden blood loss) was 796.95 ± 569.94 mL. TBL correlated with age (P = 0.014), hospital stays (P = 0.007), fixation segments (P < 0.001), activated partial thromboplastin time (P = 0.038), VBQ (L1-4 median) (P = 0.037), VBQ (L1-4 average) (P = 0.037), VBQ (S1) (P = 0.004) and VBQ (fixation levels) (P = 0.013) (Table 2). Variables with P < 0.2 were included in a multiple linear regression analysis using a stepwise regression. The analysis revealed that TBL was correlated with hospital stays (P = 0.008) and VBQ (S1) (P = 0.004) (Table 3).

Table 2 Correlations between total blood loss and all variables.

	r	P value	95%CI
			Lower limit	Upper limit
Gender	-0.061	0.340	-0.251	0.040
Age	-0.157	0.014	-0.256	-0.029
Hospital stays	0.171	0.007	0.097	0.272
BMI	0.066	0.302	-0.036	0.253
Fixation segments	0.235	< 0.001	0.113	0.355
Preoperative Hb	0.120	0.060	0.016	0.353
Preoperative HCT	0.118	0.064	0.017	0.340
ESR	-0.043	0.498	-0.175	0.051
PT	0.003	0.959	-0.177	0.107
INR	-0.073	0.251	-0.260	0.027
APTT	-0.132	0.038	-0.276	-0.057
TT	0.122	0.057	0.023	0.231
FIB	-0.094	0.143	-0.224	0.001
D-dimer	0.044	0.489	-0.067	0.228
PLT	0.025	0.700	-0.095	0.098
VBQ (L1-4 median)	-0.133	0.037	-0.216	-0.030
VBQ (L1-4 average)	-0.133	0.037	-0.227	-0.014
VBQ (S1)	-0.181	0.004	-0.279	-0.099
VBQ (fixation levels)	-0.158	0.013	-0.248	-0.050

BMI: Body mass index; Hb: Hemoglobin; HCT: Hematocrit; ESR: Erythrocyte sedimentation rate; PT: Prothrombin time; APTT: Activated partial thromboplastin time; INR: International normalized ratio; TT: Thrombin time; FIB: Fibrinogen; PLT: Platelet count; TBL: Total blood loss; VBQ: Vertebral bone quality.

Table 3 Multiple linear regression analysis of total blood loss.

	Beta	Standardized beta	t	P value	95%CI
					Lower limit	Upper limit
VBQ (S1)	-146.249	-0.181	-2.905	0.004	-245.433	-47.065
Hospital stays	24.938	0.166	2.662	0.008	6.484	43.392

F = 7.907, P < 0.001, R² = 0.061, VBQ: Vertebral bone quality; CI: Confidence interval.

Of 169 patients had pre- and postoperative 3^rd and 7^th day’s HCT and Hb. The mean age and BMI were 60.77 ± 8.47 years old and 15.03 ± 3.04, the postoperative reductions of Hb and HCT were 22.80 ± 9.33 g/L and 6.79 ± 2.59 g/L respectively at 3 days, 22.05 ± 9.77 g/L and 6.40 ± 2.85 g/L at 7 days. Of 71 patients had preoperative lumbar MRI scan and DEXA of left hip. The total and lowest T-score of the left hip were -1.19 ± 1.22 and -2.25 ± 1.30 (Table 4). The T-score was significantly correlated with the VBQ score (r = -0.331 to -0.419) (Table 5).

Table 4 Information on demographic characteristics of patients who had preoperative dual-energy X-ray absorptiometry scan, mean ± SD.

	Subgroup	1 level	2 levels	≥ 3 levels	P value
n	71	42	24	5
Male/female	23/48	15/27	8/16	1/4	0.927
Age (year)	60.52 ± 8.46	61.40 ± 8.40	59.67 ± 8.62	57.20 ± 4.66	0.485
Hospital stays (day)	12.76 ± 3.32	12.50 ± 3.45	12.96 ± 2.77	14.00 ± 4.85	0.601
BMI (kg/m2)	26.06 ± 3.51	25.76 ± 3.06	26.43 ± 4.30	26.75 ± 3.39	0.687
Fixation segments	1.48 ± 0.63	1	2	3
Hb (preoperative, g/L)	132.30 ± 13.22	132.56 ± 15.67	132.13 ± 9.41	130.80 ± 5.45	0.959
HCT (preoperative, %)	39.56 ± 3.55	39.70 ± 4.26	39.45 ± 2.34	38.88 ± 1.48	0.874
ESR (mm/hour)	15.01 ± 10.38	16.24 ± 11.96	13.21 ± 7.67	13.40 ± 6.50	0.495
PT (second)	11.03 ± 0.65	11.05 ± 0.71	11.04 ± 0.59	10.86 ± 0.43	0.827
APTT (second)	31.68 ± 2.63	31.76 ± 2.78	31.30 ± 2.42	32.76 ± 2.33	0.506
INR	0.99 ± 0.06	1.00 ± 0.07	1.00 ± 0.07	0.97 ± 0.04	0.618
TT (second)	14.67 ± 1.00	14.57 ± 0.99	14.89 ± 0.88	14.48 ± 1.64	0.429
FIB (g/L)	3.11 ± 0.67	3.23 ± 0.72	2.88 ± 0.53	3.26 ± 0.69	0.113
D-dimer (mg/L)	0.32 ± 0.28	0.33 ± 0.31	0.31 ± 0.26	0.29 ± 0.17	0.938
PLT (109/L)	225.43 ± 47.01	224.69 ± 43.79	229.29 ± 55.04	213.00 ± 36.27	0.775
T value	-1.19 ± 1.22	-1.34 ± 1.20	-0.99 ± 1.35	-0.86 ± 0.69	0.446
Minimal T value	-2.25 ± 1.30	-2.36 ± 1.29	-2.06 ± 1.43	-2.27 ± 0.68	0.676
TBL (mL)	667.78 ± 293.78	589.09 ± 276.05	780.95 ± 295.84	785.51 ± 25.55	0.023
VBQ (L1-4median)	3.50 ± 0.50	3.57 ± 0.43	3.40 ± 0.58	3.33 ± 0.52	0.278
VBQ (L1-4 average)	3.51 ± 0.50	3.59 ± 0.46	3.41 ± 0.57	3.28 ± 0.48	0.212
VBQ (S1)	3.37 ± 0.63	3.43 ± 0.63	3.28 ± 0.67	3.25 ± 0.52	0.585
VBQ (fixation levels)	3.52 ± 0.56	3.59 ± 0.52	3.44 ± 0.63	3.34 ± 0.54	0.431

BMI: Body mass index; Hb: Hemoglobin; HCT: Hematocrit; ESR: Erythrocyte sedimentation rate; PT: Prothrombin time; APTT: Activated partial thromboplastin time; INR: International normalized ratio; TT: Thrombin time; FIB: Fibrinogen; PLT: Platelet count; TBL: Total blood loss; VBQ: Vertebral bone quality.

Table 5 Correlations between T score and vertebral bone quality score and blood loss.

	T-total		T-lowest		TBL
	Rho	P value	Rho	P value	Rho	P value
VBQ (L1-4median)	-0.394	0.001	-0.331	0.005	-0.095	0.429
VBQ (L1-4 average)	-0.408	0.000	-0.334	0.004	-0.065	0.590
VBQ (S1)	-0.380	0.001	-0.342	0.004	-0.165	0.170
VBQ (fixation levels)	-0.419	0.000	-0.347	0.003	-0.090	0.455
TBL	0.190	0.113	0.165	0.168	-	-

VBQ: Vertebral bone quality; TBL: Total blood loss.

DISCUSSION

VBQ is a novel MRI-based measure that assesses bone quality without radiation. It can be routinely obtained during the evaluation of patients with spinal pathologies and has been proven to have excellent interrater and interrater reliability[19,20]. VBQ has been proven to be a risk factor of some surgical complications, such as cage subsidence[10], vertebral fracture[16], reoperation after lumbar fusion surgery[17]. In our study, we found a correlation between VBQ (S1) and blood loss in TLIF, suggesting that VBQ (S1) may serve as a predictor of blood loss in TLIF.

Previous studies have confirmed that posterior fusion surgery was closely related to intraoperative hemorrhage and postoperative blood loss[21,22]. TBL was reported to be 543 ± 353 in 1-2 levels and 667 ± 327 in ≥ 3 levels in Xu et al’s report[23] of posterior lumbar interbody fusion, 742 ± 275 in Zhang et al’s report[24] of TLIF, and 786 ± 107 in Sulaiman and Singh’s report[25] of TLIF. We confirmed that perioperative TBL was 796.95 ± 569.94 mL, which was similar with previous reports.

Surgeons always pay more attention to intraoperative blood loss, which was reported to be 375.24 ± 31.52 mL in TLIF[26]. The total blood and hidden blood loss might be overlooked in daily clinical work for they are invisible, however, TBL reflects the true loss of body fluid. In order to identify the accurate TBL, routine blood tests were performed at postoperative 3^rd and 7^th day, and patients who lost more than 800 mL of blood or whose Hb was less than 90 g/L required blood transfusions. In this study, the postoperative reductions of Hb and HCT were 22.80 ± 9.33 g/L and 6.79 ± 2.59 g/L respectively at postoperative 3 days, 22.05 ± 9.77 g/L and 6.40 ± 2.85 g/L at postoperative 7 days.

Hidden blood loss results from tissue interstitial extravasation, blood accumulated in the surgical site, and hemolysis, which has a large volume[27]. Previous studies have reported that hidden blood loss accounts for nearly half of TBL in spine surgeries[21,22], and even twice of the visible blood loss[24], which was seriously underestimated in TLIF. HBL might be a risk factor for postoperative anemia and increased the incidence of postoperative complications. Previous studies found that age was the risk factor for HBL in posterior lumbar fusion[6], which suggests that age might be a risk factor for TBL. Our study also found a correlation between age and TBL (P = 0.014). The reason that age was correlated with HBL was that muscle weakness and hypercoagulability in senile patients might cause increasing bleeding that infiltrates and agglutinates into interstitial space[28] and that older patients might have a decreased recovery capacity, poor compensatory capacity, and reduced self-regulatory ability due to angiosclerosis.

The number of surgical levels is a recognized indicator for predicting intraoperative blood loss and transfusion requirements[29]. In our study, mean TBL exceeded 1000 mL in cases with three or more fixation levels. Univariate analysis also revealed a correlation between TBL and the number of surgical segments (P < 0.001). Additionally, postoperative hidden blood loss primarily stems from muscle and soft tissue oozing, which is influenced by operative duration and the number of fixed segments[30].

The traditional method for BMD is DEXA examination[31]. However, it is prone to error in patients with vascular calcification[12], obesity[13], spinal degenerative disease or previous spinal surgery[14]. Quantitative computed tomography examination is an alternative method, which can provide more accurate results. However, quantitative computed tomography is expensive and involves a higher amount of radiation[12]. VBQ is a novel MRI-based measure that assesses bone quality without radiation[19]. The previous studies found that the VBQ score (L1-4 median) was weakly or moderately correlated with DEXA T-scores of the total hip or the lumbar spine[16]. In our study, we measured VBQ scores in all methods, including VBQ (L1-4 median), VBQ (L1-4 average), VBQ (S1) and VBQ (surgical level), and found that all VBQ scores moderately correlated with DEXA T-score of the total hip and lowest T-score of the hip.

In clinical practice, VBQ may be a better measurement to analyze BMD than T-score, and VBQ-score (S1) is recommended for its simple measurement and its comparable diagnosis accuracy of osteoporosis to VBQ (L1-4)[32]. The histological reason might be that osteoporotic bone is characterized by trabecular atrophy and local adipocyte replacement[33]. The other possible reason is that the VBQ score reflects the fat replacement in the medullary portion of the lumbar vertebral body, whereas the BMD may take the sclerotic osteophytes and posterior elements of the spine into account[10]. We hypothesized that VBQ might predict TBL, and indeed found a correlation between VBQ (S1) and TBL. However, no correlation was observed between T-score and TBL in TLIF in subgroup. This is likely attributable to the limited sample size within the subgroup. Previous study found that T-score was a predictor of major blood loss in adult spinal deformity surgery of > 5 levels fusion[7]. In such cases, three-column osteotomies - a known independent risk factor for blood loss (odds ratio = 4.1) - are often performed, and osteoporosis may increase cancellous bone bleeding[34]. In contrast, TLIF does not require a 3-column osteotomy, and cancellous bleeding is relatively minimal. Muscle seepage and venous plexus bleeding were the main cause of intraoperative bleeding and tissue interstitial extravasation was the main cause of HBL, all these factors was not significantly affected by osteoporosis. VBQ was operated on the principle of indirectly reflecting bone integrity by measuring fat content within vertebral bodies[15], it has been found to be significant correlated with fat infiltration of paraspinal muscle[35]. Intramuscular adipocyte accumulation parallels bone marrow fat expansion by sharing pathophysiological mechanisms[35], and the paraspinal muscles and vertebrae act as a cohesive biomechanical unit, a decrease in mechanical stress alters the biomechanical environment of the vertebrae and paraspinal muscle. Thus, VBQ can reflect both bone and paraspinal muscles degeneration, which can be considered a comprehensive and reliable indicator which can evaluate the overall condition of the lumbar musculoskeletal system to identify the TBL. This might explain why TBL in TLIF was correlated with VBQ and not correlated with T-score. Up to the author’s knowledge, this is the first study that analyzed the correlation between VBQ and TBL.

The TBL of TLIF was accompanied by mounts of perioperative hemorrhage, which might lead to longer hospital stays. The mean hospital stay in our study was 14.23 ± 3.81 days, similar to previous reports in China[26,36]. Previous studies found that hospital stays were influenced by the surgeon’s experience, surgical method, number of TLIF levels, and the American Society of Anesthesiologists physical status classification scores[37]. Endoscopic TLIF and minimally invasive TLIF could decrease blood loss and length of hospital stays while achieving similar clinical outcomes[38]. However, no difference in hospital stays was found between different fixation levels in our study (P = 0.133), we found that hospital stays correlated with TBL (P = 0.008). A detailed preoperative evaluation, preoperative intravenous tranexamic acid, routine postoperative blood test, and proper blood transfusion were beneficial for improving surgical outcomes and shortening the duration of hospital stays.

The current study has a few limitations. First, the current study merely included Chinese patients receiving lumbar spinal surgery from one single hospital, which could affect the generalizability of our results to other populations. Second, this is a retrospective study, as the DEXA scan was not examined by all patients since it was not regarded as a preoperative routine measurement in our institution before February 2024, not all patients included in this study had DEXA scan, a further study that includes more patients was needed. Third, as a preliminary exploratory study, the limited sample size may have constrained the robustness of our findings. The predictive model demonstrated modest explanatory power (R² = 0.061). Future studies with larger cohorts are warranted to validate these findings. Last, the purpose of this study was to analyze the correlation between VBQ and TBL, the intraoperative blood loss and hidden blood loss were not analyzed for there might be manual errors in recording these data.

CONCLUSION

This is the first study to correlate the novel MRI-based VBQ score with TBL in TLIF. VBQ (S1) score may serve as a predictor of TBL of TLIF. In multi-segment TLIF cases, increased TBL is associated with significant perioperative hemorrhage, potentially leading to longer hospital stays.

ACKNOWLEDGEMENTS

We thank Ning-Ning Wang for her help with the English writing of our manuscript.