Bisphosphonates May Reduce Intraoperative Blood Loss in Surgery for Metastatic Spinal Disease: A Retrospective Cohort Study

Figures & data

Table 1 Baseline Characteristics of Participants

Variables	No Use	Ever Use	P value
Age (years, mean ± sd)	58.7 ± 10.8	57.7 ± 10.5	0.391
Operative time (minutes, mean ± sd)	230.3 ± 86.3	206.8 ± 73.1	0.013
Blood loss (mL, mean ± sd)	1391.8 ± 1140.9	907.2 ± 717.5	<0.001
Sex (n, %)			0.436
Male	175 (53.8%)	82 (57.7%)
Female	150 (46.2%)	60 (42.3%)
Charlson comorbidity index			0.909
6	51 (15.7%)	22 (15.5%)
7	92 (28.3%)	43 (30.3%)
≥8	182 (56.0%)	77 (54.2%)
Type of tumor ^a			0.130
Highly vascularized metastases	60 (18.5%)	18 (12.7%)
Moderately vascularized metastases	232 (71.4%)	114 (80.3%)
Hematologic metastases	33 (10.2%)	10 (7.0%)
Tumor location			0.038
Cervical	32 (9.8%)	26 (18.3%)
Thoracic	165 (50.8%)	64 (45.1%)
Lumbar	128 (39.4%)	52 (36.6%)
Number of spinal metastases			0.782
Single	128 (39.4%)	54 (38.0%)
Multiple	197 (60.6%)	88 (62.0%)
Preoperative Frankel score			0.403
A–C	73 (22.5%)	27 (19.0%)
D–E	252 (77.5%)	115 (81.0%)
Surgical approach			0.371
Anterior	37 (11.4%)	20 (14.1%)
Posterior	277 (85.2%)	120 (84.5%)
Combined	11 (3.4%)	2 (1.4%)
Surgical procedure ^b			0.258
Type 1	57 (17.5%)	24 (16.9%)
Type 2	247 (76.0%)	114 (80.3%)
Type 3	21 (6.5%)	4 (2.8%)
Level of decompression			0.494
1	199 (61.2%)	80 (56.3%)
2	69 (21.2%)	37 (26.1%)
≥3	57 (17.5%)	25 (17.6%)
Level of instrumentation			0.870
191 (58.8%)	81 (57.0%)
4–5	102 (31.4%)	48 (33.8%)
>5	32 (9.8%)	13 (9.2%)

Notes: ^aPrimary tumors were divided into 3 groups according to vascularization: highly vascularized metastases (renal, hepatocellular, and thyroid), moderately vascularized metastases (lung, breast, prostate, gastrointestinal, and other epithelial tumors), and hematologic metastases (myeloma and lymphoma). ^bSurgical procedures were divided into 3 groups: type 1 (palliative instrumentation and decompression), type 2 (subtotal corpectomy), and type 3 (total en bloc spondylectomy).

Table 2 Effects of Risk Factors on Blood Loss by Univariate Analysis

Variables	Statistics	β (CI 95%)	P value
Sex
Male	257 (55.0%)	0
Female	210 (45.0%)	87.7 (−104.5, 279.9)	0.372
Age	58.4 ± 10.7	−2.3 (−11.3, 6.6)	0.608
Age
Bottom tertile	148 (31.7%)	0
Middle tertile	140 (30.0%)	116.8 (−126.8, 360.3)	0.348
Top tertile	179 (38.3%)	−48.9 (−278.4, 180.6)	0.676
Charlson comorbidity index
6	73 (15.6%)	0
7	135 (28.9%)	−50.0 (−350.2, 250.3)	0.744
≥8	259 (55.5%)	−150.1 (−424.0, 123.8)	0.283
Type of tumor ^a
Highly vascularized	78 (16.7%)	0
Moderately vascularized	346 (74.1%)	−567.1 (−820.0, −314.2)	<0.001
Hematologic metastases	43 (9.2%)	−95.6 (−478.8, 287.7)	0.625
Tumor location
Cervical	58 (12.4%)	0
Thoracic	229 (49.0%)	656.4 (358.0, 954.8)	<0.001
Lumbar	180 (38.5%)	502.4 (196.0, 808.9)	0.001
Number of spinal metastases
Single	182 (39.0%)	0
Multiple	285 (61.0%)	−26.8 (−223.0, 169.4)	0.789
Preoperative Frankel score
A–C	100 (21.4%)	0
D–E	367 (78.6%)	−45.2 (−278.4, 188.1)	0.704
Operative time	223.3 ± 83.2	5.5 (4.3, 6.6)	<0.001
Operative time
Bottom tertile	103 (28.0%)	0
Middle tertile	129 (35.1%)	269.6 (29.0, 510.1)	0.029
Top tertile	136 (37.0%)	1015.8 (778.0, 1253.5)	<0.001
Surgical approach
Anterior	57 (12.2%)	0
Posterior	397 (85.0%)	422.6 (132.1, 713.1)	0.005
Combined	13 (2.8%)	616.5 (−13.7, 1246.8)	0.056
Surgical procedure ^b
Type 1	81 (17.3%)	0
Type 2	361 (77.3%)	123.6 (−126.7, 373.8)	0.334
Type 3	25 (5.4%)	926.6 (460.9, 1392.3)	<0.001
Level of decompression
1	279 (59.7%)	0
2	106 (22.7%)	33.4 (−202.7, 269.6)	0.782
≥3	82 (17.6%)	48.4 (−211.6, 308.3)	0.716
Level of instrumentation
<4	272 (58.2%)	0
4–5	150 (32.1%)	35.8 (−174.2, 245.7)	0.739
>5	45 (9.6%)	243.1 (−89.1, 575.3)	0.152
Bisphosphonate use
No use	325 (69.6%)	0
Ever use	142 (30.4%)	−484.6 (−687.9, −281.3)	<0.001

Notes: ^aPrimary tumors were divided into 3 groups according to vascularization: highly vascularized metastases (renal, hepatocellular, and thyroid), moderately vascularized metastases (lung, breast, prostate, gastrointestinal, and other epithelial tumors), and hematologic metastases (myeloma and lymphoma). ^bSurgical procedures were divided into 3 groups: type 1 (palliative instrumentation and decompression), type 2 (subtotal corpectomy), and type 3 (total en bloc spondylectomy).

Figure 1 The intraoperative blood loss in patients with bisphosphonate treatments and patients without bisphosphonate treatments. There was a significant difference in the intraoperative blood loss between the two groups (p < 0.001).

Table 3 Association of Any Bisphosphonate Use with the Amount of Blood Loss

Variable	Crude Model (β, 95% CI, P value) ^a	Minimally Adjusted Model (β, 95% CI, P value) ^b	Fully Adjusted Model (β, 95% CI, P value) ^c
Bisphosphonate use
No use	0	0	0
Ever use	−484.6 (−687.9, −281.3) <0.001	−244.4 (−448.9, −39.9) 0.020	−246.4 (−447.0, −45.8) 0.017
Timing
No use	0	0	0
Past use	−337.5 (−723.6, 48.6) 0.087	−287.3 (−577.3, 2.7) 0.055	−311.1 (−599.8, −22.3) 0.035
Recent use	−486.4 (−781.4, −191.4) 0.001	−299.9 (−594.9, −4.9) 0.047	−231.1 (−655.3, 193.1) 0.532
Current use	−560.3 (−850.8, −269.7) <0.001	−325.9 (−613.8, −37.9) 0.027	−311.7 (−594.6, −28.9) 0.031
P for trend	<0.001	<0.001	0.231
Duration
No use	0	0	0
<1 week	−468.6 (−797.2, −139.9) 0.005	−142.9 (−467.7, 181.9) 0.389	−135.8 (−454.3, 182.7) 0.404
1 week-1 month	−479.3 (−767.9, −190.6) 0.001	−249.6 (−578.1, 78.9) 0.137	−239.6 (−562.8, 83.5) 0.147
>1 month	−508.9 (−844.6, −173.3) 0.003	−319.4 (−613.4, −25.4) 0.034	−340.3 (−630.2, −50.3) 0.022
P for trend	<0.001	<0.001	<0.001

Notes: ^aCrude model: we did not adjust other covariants. ^bMinimally adjusted model: we adjusted type of tumor and tumor location. ^cFully adjusted model: we adjusted type of tumor, tumor location, operative time, surgical approach and surgical procedure.

Table 4 Effect Size of Bisphosphonate on Blood Loss in Exploratory Subgroups

Subgroup	Statistics	Effect Size (95% CI)	P for Interaction
Sex			0.932
Male	257	−497.3 (−755.2, −239.3) <0.001
Female	210	−461.7 (−788.0, −135.3) 0.006
Age			0.765
Bottom tertile	148	−464.6 (−792.5, −136.7) 0.006
Middle tertile	140	−666.3 (−1085.5, −247.1) 0.002
Top tertile	179	−362.6 (−682.2, −43.1) 0.027
Charlson comorbidity index			0.789
6	73	−491.1 (−985.4, 3.2) 0.055
7	135	−792.2 (−1218.6, −365.8) <0.001
≥8	259	−319.6 (−574.0, −65.1) 0.015
Type of tumor			0.554
Highly vascularized	78	−915.2 (−1637.2, −193.2) 0.015
Moderately vascularized	346	−350.5 (−551.9, −149.2) <0.001
Hematologic metastases	43	−485.9 (−1257.2, 285.3) 0.224
Tumor location			0.633
Cervical	58	−280.5 (−611.2, 50.3) 0.102
Thoracic	229	−486.1 (−806.2, −165.9) 0.003
Lumbar	180	−443.0 (−763.7, −122.4) 0.007
Number of spinal metastases			0.898
Single	182	−515.8 (−858.6, −172.9) 0.004
Multiple	285	−464.7 (−716.9, −212.4) <0.001
Preoperative Frankel score			0.679
A–C	100	−647.4 (−1134.0, −160.7) 0.011
D–E	367	−442.9 (−666.4, −219.5) <0.001
Operative time			0.053
Bottom tertile	103	−48.2 (−282.4, 186.0) 0.688
Middle tertile	129	−483.4 (−814.6, −152.2) 0.005
Top tertile	136	−374.8 (−821.3, 71.8) 0.102
Surgical approach			0.352
Anterior	57	−102.6 (−512.1, 307.0) 0.626
Posterior	397	−532.5 (−757.7, −307.3) <0.001
Combined	13	−277.3 (−2102.7, 1548.2) 0.772
Surgical procedure			0.285
Type 1	81	−706.7 (−1164.9, −248.5) 0.003
Type 2	361	−369.7 (−591.6, −147.7) 0.001
Type 3	25	−1280.9 (−2756.2, 194.3) 0.102
Level of decompression			0.846
1	279	−473.1 (−745.5, −200.7) <0.001
2	106	−814.8 (−1242.3, −387.4) <0.001
≥3	82	−633.6 (−984.7, −282.5.5) 0.103
Level of instrumentation			0.428
<4	272	−500.6 (−773.4, −227.7) <0.001
4–5	150	−526.8 (−883.9, −169.8) 0.004
>5	45	−267.5 (−831.0, 295.9) 0.357