0OPEN
ACCESS
GLOBAL SPINE JOURNAL
Systematic Review 185
Unilateral versus Bilateral Instrumentation in Spinal Surgery: A Systematic Review
Robert W. Molinari1 Ahmed Saleh1 Robert Molinari, Jr.2 Jeff Hermsmeyer3 Joseph R. Dettori3
1 Department of Orthopaedics, University of Rochester, Rochester, New York, United States
2 Brooklyn College BA/MD Program, Brooklyn, New York, United States 3Spectrum Research, Inc., Tacoma, Washington, United States
Global Spine J 2015;5:185-194.
Address for correspondence Robert W. Molinari, MD, Department of Orthopaedics, University of Rochester, 601 Elmwood Avenue, Rochester, NY 14642, United States (e-mail: bobmol@aol.com; kan@rochester.rr.com).
Abstract
Keywords
► unilateral
► bilateral
► fixation
► pedicle screw
► minimally invasive
► lumbar
► fusion
Study Design Systematic review.
Clinical Questions (1 ) What is the comparative efficacy of unilateral instrumentation compared with bilateral instrumentation in spine surgery? (2) What is the safety of unilateral instrumentation compared with bilateral instrumentation in spine surgery? Methods Electronic databases and reference lists of key articles were searched up to September 30, 2014, to identify studies reporting the comparative efficacy and safety of unilateral versus bilateral instrumentation in spine surgery. Studies including recombinant human bone morphogenetic protein 2 as adjunct therapy and those with follow-up of less than 2 years were excluded.
Results Ten randomized controlled trials met the inclusion criteria: five compared unilateral with bilateral instrumentation using open transforaminal or posterior lumbar interbody fusion (TLIF/PLIF), one used open posterolateral fusion, and four used minimally invasive TLIF/PLIF. There were no significant differences between unilateral and bilateral screw instrumentation with respect to nonunion, low back or leg pain scores, Oswestry Disability Index, reoperation, or complications. Conclusions The existing literature does not identify significant differences in clinical outcomes, union rates, and complications when unilateral instrumentation is used for degenerative pathologic conditions in the lumbar spine. The majority of published reports involve single-level lumbar unilateral instrumentation.
Study Rationale and Context
Unilateral instrumentation has been advocated as an alternative to bilateral instrumentation for spine fusion. The advantages touted include avoidance of soft tissue disruption on the contralateral side, reduced operation time, and lower implant costs.1-3 However, the results of some studies suggest that unilateral instrumentation may result in nonunion, metal failure, pseudarthrosis, or cage migration due to the decreased strength or inherent asymmetry of this system.1,4 Whether unilateral instrumentation is as efficacious and safe as bilateral instrumentation for spine fusion is debated.
Clinical Questions
1. What is the comparative efficacy of unilateral instrumentation compared with bilateral instrumentation in spine surgery?
2. What is the safety of unilateral instrumentation compared with bilateral instrumentation in spine surgery?
Materials and Methods
Study design: Systematic review.
received
January 4, 2015 accepted March 23, 2015
DOI http://dx.doi.org/ 10.1055/s-0035-1552986. ISSN 2192-5682.
© 2015 Georg Thieme Verlag KG Stuttgart • New York
License terms
©(D©®
Search: PubMed, Cochrane collaboration database, and National Guideline Clearinghouse databases; bibliographies of key articles. Dates searched: January 1980 to September 30, 2014. Inclusion criteria: (1) Randomized controlled trials in peer-reviewed journals; (2) patients undergoing spinal fusion for any surgical pathology where unilateral instrumentation was compared with bilateral instrumentation; (3) outcomes included at least one of the following: complications, fusion rate, or patient-reported function.
Exclusion criteria: (1) Fusion supplemented with recombinant human bone morphogenetic protein 2; (2) observational studies; (3) follow-up less than 2 years; (4) sample size less than 10 in either treatment arm.
Outcomes: (1) Proportion was nonunion; (2) change in patient-reported and clinical outcomes (baseline to follow-up); (3) complication risk.
Analysis: Meta-analysis was performed using RevMan software (Review Manager version 4, The Nordic Cochrane Centre, Copenhagen, Denmark). Mean differences were calculated for continuous variables and risk differences for dichotomous variables, both with associated 95% confidence intervals. The I2 statistic was used to assess heterogeneity. Details about methods can be found in the online supplementary material. Overall strength of evidence: The overall strength of evidence across studies was based on precepts outlined by the Grades of Recommendation Assessment, Development and Evaluation (GRADE) Working Group.5 Study critical appraisals and the reasons for upgrading and downgrading for each outcome can be found in the online supplementary material.
Results
• We identified 10 randomized controlled trials that met the inclusion criteria, which form the basis for this report (►Fig. 1). All were lumbar fusions for degenerative spinal disorders. A list of excluded studies can be found in the online supplementary material.
• Five studies compared unilateral with bilateral instrumentation using open transforaminal or posterior lumbar
Fig. 1 Flowchart showing results of literature search.
interbody fusion (TLIF/PLIF),6 10 one study used open posterolateral fusion,3 and four used minimally invasive TLIF/PLIF11-14 (►Table 1).
Nonunion
• n = 8 studies.
• There is no statistical difference of nonunion between unilateral and bilateral instrumentation, pooled risk difference, 0.01 (95% confidence interval [CI]: -0.01, 0.04). The results were similar across surgical procedure (open TLIF/PLIF, posterolateral, or minimally invasive [MIS]; ►Fig. 2).
Patient-Reported and Clinical Outcomes
• Low back pain (LBP), 10-point visual analog scale (VAS):
• n = 7 studies.
• In studies where the surgical procedure was an open TLIF/PLIF, the VAS LBP score statistically favored bilateral instrumentation, though the difference was not considered clinically meaningful; mean difference (MD) between changed score was 0.71 (95% CI: 0.06, 1.36; ►Fig. 3). In studies using MIS, there was no statistical difference between unilateral and bilateral instrumentation.
• Leg pain, 10-point VAS:
• n = 4 studies.
• No statistical difference between the procedures was found when doing open TLIF/PLIF or MIS (►Fig. 4).
• Oswestry Disability Index (ODI):
• n = 5 studies.
• There was no statistical difference in mean ODI scores between unilateral and bilateral screw instrumentation (►Fig. 5).
• Japanese Orthopaedic Association (JOA) score:
• n = 3 studies.
• The JOA scores were better in the bilateral screw instrumentation group, with pooled MD of 0.85 (95% CI: 0.08, 1.61; ►Fig. 6).
Complications
• Reoperation:
• n = 2 studies.
• No statistical difference was found between unilateral and bilateral screw instrumentation (►Fig. 7).
• Infection:
• n = 6 studies.
• There was no statistical difference in the risk of infection between unilateral and bilateral screw instrumentation (►Fig. 8).
• Cage migration:
• n = 3 studies.
• No statistical difference was found between unilateral and bilateral screw instrumentation (►Fig. 9).
• Screw failure:
• n = 8 studies.
• There was no statistical difference between unilateral and bilateral screw instrumentation (►Fig. 10).
Table 1 Demographics and characteristic of included studies
Author n (uni:bi) Mean age (% male) Diagnosis F/U (mo) rate (%) Graft used No. of levels
Open TLIF and PLIF
Aoki (2012)6 50 (25:25) 65.9 (40) LS grade I, II 31.1 (94) Uni: 1 cage; bi: 2 cages 1
Duncan (2013)7 116 (57:59) 54.7 (39) LSS, LS, SDDD, LDH 25.1 (87.9) 1 cage 1/2
Kai (2013)15 68 (33:35) 57.5 (35) LSS, LS, SDDD, FBS 25.6 (100) 1 cage 2
Xie (2012)10 108 (56:52) 53.5 (45) LSS, RLDH, SDDD >36 (100) 1 cage 1/2
Xue (2012)8 80 (37:43) 57.7 (44) LSS, LS, LDH, RLDH, DLBP 25.3 (100) 1 cage 1/2
Posterolateral
Fernandez-Fairen (2007)3 82 (40:42) 61.1 (38) LS 36 (98.8) NR 1/2
Choi (2013)11 54 (26:28) 54.8 (40) LSS, LS, LDH, RLDH 28.2 (98.1) 1 cage 1
Dong (2014)13 39 (20:19) 55.2 (31) DLI, LS 36 (100) 1 cage 1
Lin (2013)14'a 85 (43:42) 66.3 (46) LSS, LS, LDH 26 (100) 1 cage 1
Shen (2014)12 65 (31:34) 58.1 (51) LSS, DLBP, LDH 26.6 (100) 1 cage 1
Abbreviations: bi, bilateral segmental fixation; DLBP, discogenic low back pain; DLI, degenerative lumbar instability; DRL, degenerative retrolisthesis; FBS, failed back surgery; F/U, follow-up; IDD, internal disk disruption; LDH, lumbar disk herniation; LS, lumbar spondylolisthesis; LSS, lumbar spinal stenosis; MIS, minimally invasive; NR, not reported; RLDH, recurrent lumbar disk herniation; SDDD, symptomatic degenerative disc disease; TLIF, transforaminal lumbar interbody fusion; PLI, posterior lumbar interbody fusion; uni, unilateral segmental fixation; VEPF, vertebral end plate fracture. aA similar study with same population and surgical procedure by Lin et al was published in Chinese. We chose to abstract data from the English study.
Unilateral Bilateral Risk Difference
Study or Subgroup Events Total Eventi. Total Weigh< M-H. Ran dam, 95%Cj_
Risk Difference M-H, Random. 35% CI
Open TLIF/PLIF
Aofcl 2012 K.li 2013 XI« 2012 Xue2012 Subtotal (35% CO Total events Heterogeneity: Tau' = 0.00; Chi Test for overall effect: Z - 0.5S [P
3 24 1 23 2.9%
3 31 2 35 4.5%
0 56 0 52 55,3%
3 37 J 43 6-0%
ISO 153 <58-6*
I' - 3,06, df - 3 (P = 0,38); 1' = 2%
I O.SS)
Poite relate rat
F« man dflï- Fair« ri 2007 4 40
Subtotal <35% Cl) 40
Total events 4
Heterogeneity; Not applicable Test for overall effect: Z - 0,43 (P - 0,67)
Choi 2013 Dong 2014 Lin 2013 Shen 2014 Subtotal <95K CI) Total events
Tist for overall tHitt: 2 = 0.94 (P - 0,33)
4.7% A,7%
0.08 r-0,07, 0,24] 0.03 (-0,09. 0.16] 0.00 [-0.04. 0,041 0.03 [-0.07, 0,14] 0-01 [-0 02, 0.04)
0.03 [-0,10, 0 15] 0-03 [-010, 0.15]
4 26 1 27 2.9% 0.12 [-0.04, 0.27)
O 20 0 19 7.8% 0.00 r-0,09, 0.09]
3 43 3 42 -0.00 r-0,11, 0,11)
1 31 0 34 10.1% 0.03 [-0,05. 0,12]
120 122 26.7% 0.02 (-0.03, 0.0 8)
1' - 1.98, df = 3 (P = 0.58}; l] - OSÉ
ÏÎ6 100,0%
Total OSS CO
Total events 21 12
Heterogeneity: Tau' - 0.00; Chi' - S.20, df = 6 (P - 0.74}; l! ■ Test for overall effect: Z - 1.01 (P - 0.31)
Test for subgroup differences: Chi' - 0.27. df - 2 <P « 0.87). I'.ffli
0,01 (-0-01, 0.04]
I 03É
I----H-
-O.S -0.25 Favors Unilateral favors Bilateral
0.25 0,5
Fig. 2 Nonunion. Abbreviations: CI, confidence interval; MIS, minimally invasive; M-H, Mantel-Haenszel; TLIF/PLIF, transforaminal or posterior lumbar interbody fusion.
Study or Subgroup
Unilateral Mean SO Total
Bilateral Mean SO TotaE
Weight
Meart Difference IV, Random, 9SX G
Me an Difference IV, Rjndqwi, 954 Cl
г íp =
Open TLIF/PLIF
Aokl2012 -1.6 1.99 24 -3,7
Kai 2013 -5,7 0.51 33 -6.1
Xue 2012 -6,1 0,48 37 -6,4
Subtotal (95% CO 94 Heterogeneity. Tan' = 0,25. Ctil' m 11.30. df Test fiar overall -äfftet: I - 1Д4 (P » 0,03)
Choi 2013 -5.8 0.97
Dong 2014 -8.25 0,97
Lin 2013 -4,1 0,9B
Sben гон -4. S 0.96 Su hiatal (9S% tl>
Heterogeneity. Tau1 - 0,00; Chi* в Test for overall effect:. Z = 0.74 (P = 0,46)
Total (35-?, CI) 214
Heterogeneity; Tan' = 0,13, Chi1 - 23.31. df = Test Tor overall effect; г » 133 (P = 0,ie) Тел for subgroup A if leren« 5 : Chi' = 5,13, df
1.31 0,24
35 43 101
Û.0O4); lJ
6.936 15.2K 20.6K 42,7%
г.io [i-06r3 m]
0.40 [-0.07,0,87) 0.30 [0.13, 0.479 0,71 [0-06. 1.36)
26 -S.9 0,99 27 14. OK 0.10 (-0.43, 0,631
20 7.65 0,99 19 12.4K -0.40 ¡-1.02. 0,22)
43 -3,9 0,7 42 17.3S6 -0.20 [-0.56, 0,161
31 -5 1 Î9 34 13-6* 0.20 [-0.3Sj 0.75]
120 122 S 7.3% -0.09 [-0,13,fi.IS]
2.90, df - 3 If - 0.41); 1" - 0%
223 100.0%
■ 0,0007); Г - 74%
0.22 [-0.11, 0.S6]
1 IP m 0.02), Г = 60.5%
—t— -4
Favors Un ¡late rai Favors Bilateral
Fig. 3 Low back pain. Abbreviations: CI, confidence interval; IV, inverse variance; MIS, minimally invasive; TLIF/PLIF, transforaminal or posterior lumbar interbody fusion; SD, standard deviation.
Study or Subgroup
Unilateral Mean SD Total
Bilateral Mean SO Total
Mean Difference Weight IV. Handom. 95% CI
Oper TLIF/PUF
Aoki 2012 Kai 2013
Subtotal (95%CI)
Heterogeneity. Tau' = Test for overall effect;
-2.2 -5.5
2,09 1.88
5.23; Chi' > Z -0.74 IP
24 -33 -57 25.IS, df » 0.46}
S.l 1.69 23 24.№. 2 90 ¡1.62, 3.98] 5,1 0,82 35 26.4* -0.40 ¡-1.10, 0.30] 58 50.4% 1.22 [-2,01,4,«) » 1 (P < 0.00001); I" » 96%
Choi 2013 Pong 2014
-5, S 1.98 -7,04 1.38 Subtotal (95Э( CI)
Heterogeneity; Tau' - 0.00; Chi2 -Test for overall effect: Z - 0,17 {P ■
26 -5,7 20 -7.04 46
0.02, df - 1 <P 0.87)
1,25 27 Z5.2SÍ -0.101-1.00,0.60]
1,25 )9 24.4% 0.00 !-1-03, 1-03]
46 49.64 -0.06 [-0.73, 0 62] - 0.B9}; I1 - 0%
Total (95% CI) 103
Heterogeneity; Tau' = 1.72; Chi2 = 27,05, df Test for overall effect: Z ■ D.B1 f з 0.42Ï
Test for subgroup differences; Chi* - 0.56. df - HP - 0.45], l" - OK
104 100.0%
3 <P < 0,00001); l¡ = 89S
0.S7 [-0.60, 1.93]
Mean Difference IV, Randen, 95% CI
—t— -2
Favor} unilateral Favarj Bilateral
ra с 3
Fig. 4 Leg pain. Abbreviations: CI, confidence interval; IV, inverse variance; MIS, minimally invasive; TLIF/PLIF, transforaminal or posterior lumbar interbody fusion; SD, standard deviation.
T3 (Я
Unilateral
Bilateral
Mean Difference
Study or Subgroup Mean SD Total Mean SQ Total Weight IV, Bandom. 95% CI
Open TLIF/PLIF
KaiZOIÎ -23.6 13.58
Xue20)2 -2S 1-2
Subtotal (95% CO Heterogeneity: Tau' - 0,00: Chi' ■ Test for overall effect: 2 - 3.70 (P
33 -26.4 12.95 35 1D.7X
37 -29 3 1-S6 43 35-8SÍ
70 78 i6,S* 0,21, df - 1 IP - 0,64), I'- 0SÍ ■ 0.0002)
Choi 2013 Dong 2014 Lin 2013 Shen 2014 SuhtoUl t95* CO Heterogeneity: Tau' = Test for overall effect:
-21.2 )0.78 25 -13-4 10.11 27 12.5»
-29.5 10.78 20 -29-5 10.11 19 10.1SÍ
-26.5 10.Й6 43 -28 10.04 42 17.4%
-35.2 ll.S 31 -28.6 10.17 34 13.5%
120 122 53-5* 6-3Î: Chi' - 5-5!. df = 3 [P = 0.13); l' - 46* 2 =0.98 ÍP-0.33)
Total (9 5% Cl) 190
Heterogeneity: Tiu' «4.23. Chi' - 10.27, df t 5 [? теп for overall effect: г =■ о.is {P = o.be) Test for subqroup differences: Oil' - 2-74, d( = 1 <P = 0-Ю1.
2.80 [-3.51, 9.11) 1 3010-60, 2-00] 1.3? [0.62. 2.02]
-2.80 [-8.43, 2.831 0.00 [-6.56. 6.56t 1.S0 [-2.77, 5.771 -6.40 [-11.70,-1.101 -1Л2 [-5-47, 1 83)
Mean Difference IV. Random, 9SX CI
ZÛO lOO.O* ■ 0.07). I' - 51*
-0.19 [-2.64, 2.26]
z¡ о о
TD (Я
■ 63.59t
-10 -5 Favori unilateral Favors Bilateral
Fig. 5 Oswestry Disability Score. Abbreviations: CI, confidence interval; IV, inverse variance; MIS, minimally invasive; TLIF/PLIF, transforaminal or posterior lumbar interbody fusion; SD, standard deviation
UnplaiEra Bilateral Mean Différente
Study or Subgroup Mean SD Total Mear SD Total Weight IV. Random, 9SM Cl
Mean Difference IV. Random, 9SK CI
Open TUF/PU F
Aokl Î01Î 1.6 0-S4 24 1-î 0,74 23 40.03É
№№ 11.6 2.03 56 Ï0.1 1,55 52 30.5%
Subtotal (95% Cl) £0 25 70,9«
Heterogeneity ; Tau! - 0,63; Chi1 = 7.85. dl = 1 (P = 0.005); l! = в 7% Test for ove-ral effect: Z - 1,43 (P » 0.1 S)
12.2 1.29
Dorg 2014 Subtotal 0SK Cl) Heterogeneity. Not applicable Test for overall effect: Z - 1.13 (P - 0.03)
11.3 1,34
29. IK 29.1*
Total [95* CI) 100 94 100 0%
Heterogeneity: Tau3 = 0.34: Chi7 - 6.52, df 2 (P = 0.01): I* » 775É Test for overall effect: Z - 2.18 IP ■ 0.03)
Test for subgroup différentes: Chi1 = 0.00, df = 1 tP = 0,95!, I* = 0Ж
O.Jo J45-07, 0 67] 1.50 10.75, 2.25] 0.85 [-0,32, 2.03]
0.50 [0.07, 1,73] 0.90 [0.07, 1.73]
0.35 [0.QÍ. Ml]
-4 Í" 6 2 4 Favors Unilateral Fairors Bilateral
Fig. 6 Japanese Orthopaedic Association Score. Abbreviations: CI, confidence interval; IV, inverse variance; MIS, minimally invasive; TLIF/PLIF, transforaminal or posterior lumbar interbody fusion; SD, standard deviation.
unilateral Bilateral Risk difference
Study or Subgroup Even^ Tout Events Total Weig h t M-H, Rand опт, 9 5% C1
Risk Difference M-H, Random. 95* CI
Poïte relaterai
Fe mandez-Fairen 2007 0
Subtotal <95% Cl) Total events 0
Heterogeneity. Not applicable Test for overall effect: 1 = L.S9 (P =
26 31 57
Choi 2013 Shen 2014 Subtotal (95% CI) Total events 2
Heterogeneity. Tau' - 0.00, Chi* = 0.95, df ■ Test for overall effect: Z = 0.27 1? = 0,79)
Total (95% CI) 97
Total events 2
Heterogeneity: Tau' = 0.00; Chi' = 2.10, df ■ Test for overall effect. Z = 1.13 № - 0.26) Test for subgroup differences: Chi' = 1,24. df
27 34 61
35.8% 35.8%
19.1% 45.1% 64.2%
1 IP - 0,3 3>; Ia - 05É
102 100,0%
-0.07 [-0,16. 0,021 -0.07 [-0,16, 0.02]
0.04 [-0,08. 0,16] -0.03 [-0,11, 0,05] -0.01 [-0.08, 0.06]
-0.03 [-0.09. 0.02]
0.35); I' " 53Í
1 (P- 0.27), I1 * 19,3*
-0 5 -0 25 0 0.25 0 5 Favors unilateral Favors Bilateral
Fig. 7 Reoperation. Abbreviations: CI, confidence interval; MIS, minimally invasive; M-H, Mantel-Haenszel; TLIF/PLIF, transforaminal or posterior lumbar interbody fusion.
• Other complications:
• n = 8 studies.
• There was no statistical difference between unilateral and bilateral screw instrumentation (►Fig. 11). Pooled results in open TLIF/PLIF suggest a slightly lower but nonstatistically significant risk of other complications favoring the unilateral instrumentation. Those other complications include pulmonary embolism (n = 1), deep vein thrombosis (n = 1), dural sac laceration (n = 7), postoperative proximal scoliosis (n = 1), transient motor weakness (n = 2), and cerebral spinal fluid leak (n = 2).
Evidence Summary
There was no difference in nonunion, low back or leg pain, ODI, reoperation, infection, cage migration, screw failure, or other complications comparing unilateral with bilateral screw instrumentation. The overall strength for these findings are considered low or very low (►Table 2).
Clinical Guidelines
None found.
Illustrative Case
A 69-year-old man had chronic back and bilateral leg pain. The patient's pain was refractory to conservative measures over a 2-year period. He had physical therapy, medications, and epidural steroid injections without significant long-term relief of symptoms.
The preoperative radiographs showed L4-L5 grade 1 degenerative spondylolisthesis (►Fig. 12A, B). The preoperative magnetic resonance imaging demonstrated L4-L5 low-grade degenerative spondylolisthesis with severe spinal stenosis (►Fig. 12C, D).
The patient had L4-L5 decompression and fusion surgery. His surgical procedure was complicated by the inability to successfully place his left L5 pedicle screw. The left L4 screw
Study or Subgroup
Unilateral Events Total
Bilateral Events Total
Weight
Risk Difference M-H. Random, 95*CI
Risk Difference M-H, Random. 95% CI
Open TLIF/PLIF
Aoki 2012 tai 2013 Xie 2012 Hue 2012 Subtotal (9S% CO Total events Heterogeneity: Tau' =■ Test for overall effect
0 24 0 23 10.9K
1 33 2 35 73%
2 S6 2 52 13.5X
3 37 2 43 5.9%
150 153 37,6%
0.00; Chi* -Z = 0.04 (P
Posterolateral
Fernande7-Fairen 2O07 0
Subtotal OS* CO Total events o
Htwcgentlty: Not applicable Test for overall effect: Z = 0-00 (P =
Lin 2013 5 lien 2014 Subtotal (95* CO Total events Heterogeneity: Tiu' = Test for overall effect
0.71, df p ■ 0.97)
3 fP = 0.87); l!
■ 0«
31.1% 3L1%
1 43 2 42 11.1%
0 31 û 34 20.2%
74 76 IL«
1'- 0.29, df - 1 (P = 0.59); Ia ■= QW
Z - 0.36 (P = 0.72)
0.00 [-O.OS, 0,08] -0.03 [-0,12, 0,071 -0.00 [-0,07, 0,07] 0.03 [-0,07, 0,14] -0 (H) [-0-04. 0-04]
0.00 [-0,05, 0 05] 0-00 [-0-05 0 05]
-0.02 1-0,10, 0,051 0.00 [-0.06, 0.06] -0.01 [-0 06. 0-04]
Total <95% CI) 264
Total events 7
Heterogeneity Tau' = o.oo; Chi' = l.oi, df = Teil for overall effect: Z = 0-23 <P = 0-82) Test for subgroup différences: Chi' - 0,06. df
270 100.0% -0.00 [-0.03, 0.02]
0-99); I1 = 05Í
'-as -cús
0.25 O.S
■ 2 (P = 0.96). I1 -0*
Favors Unilateral Favors Bilateral
Fig. 8 Infection. Abbreviations: CI, confidence interval; MIS, minimally invasive; M-H, Mantel-Haenszel; TLIF/PLIF, transforaminal or posterior lumbar interbody fusion.
was intentionally left in place, and right-sided unilateral fixation was performed along with bilateral posterolateral iliac crest bone grafting.
The 3-year postoperative anteroposterior and lateral radiographs (►Fig. 12E, F) demonstrated solid bilateral fusion without loosening of unilateral fixation. There was a slight progression of the spondylolisthesis when compared with the preoperative standing radiographs. The patient reported minimal back pain and improved function at 3-year follow-up.
Discussion
• Strengths:
• Several randomized controlled trials assessing the treatment options allowed for meta-analysis stratified by type of surgery.
• Limitations:
• Important outcomes were included inconsistently among studies resulting in small sample sizes for
Favors Unilateral
Bilateral
Risk Difference
Study Or Subgroup
Events Total Events Total weight M-H, Random, 95% Cl
Î 24 1 23 9.9% 0.04 [-0. Í0, 0163
11 46 6 Í6 9.0% 0,B[-0.0î, 0,26]
0 33 0 35 243JÍ O OO [-0 06, 006]
0 56 0 S2 29.25Í O.OO [-0.04, 0.04]
159 166 72.2% 0.03 [-0.05, 0.10]
Risk Difference M-H, Random, 95% Ci
Open TLIF/PLIF
Aoki 2012 Duncan 2013 Kal 2013 Xle 2012 Subtotal (95% CI) Total events Heterogeneity. Tau* i Test for overall effect: Z - 0.71 (P = 0.46)
O.OO, Chi' = 10.73, df = 3 (P = 0.01); P » 72*
Choi 2013 2
Dong 2014 0
Subtotal (9S% CO Total events 2 0
Heterogeneity. Tau' - 0.00; Chi' = 1.13, df = 1 (P - 0.29); I' Test for overall effect: Z ■ 0.75 (P = 0.45)
26 20 4fi
27 19 46
12.0% IS S* 2 7.6%
212 100.0%
Total (9S% CI) Total events IS 7
Heterogeneity. Tau' = 0.00; Chi1 - ILM, df - 5 0.04): I* -Test for overall effect: Z = 0,95 (P = 0.34) Test for subgroup differences: Chi' . 0.01, df « 1 <P » 0,93), I*
O.08 [-0.04. 0.20] O.OO [-0 09, 0 09] 0 03 [-0.05, 0,11]
0.02 [-0.03, 0.08]
-0.5 -0.25 0,25
Favors Unilateral Favors Bilateral
Fig. 9 Cage migration. Abbreviations: CI, confidence interval; MIS, minimally invasive; M-H, Mantel-Haenszel; TLIF/PLIF, transforaminal or posterior lumbar interbody fusion.
Study or Subgroup
Unilateral Events Total
Bilateral Events Total
Weight
Risk Difference M-H. Random, 95%Ci
Risk Difference M-H, Random. 95% CI
Open TLIF/PLIF
Xue 2012 3 82
Xie 2012 1 126
Kal 2013 0 99
AOW20I2 3 4$
Subtotal (95% CO 3 55
Total events ?
Heterogeneity; Tau' = 0,00; Chi' = 3,43, df Test for overall effect: Z - 0.64 [P - 0,52)
Posterolateral
Fe man de;-Fairen 2007 0 S 7
Subtotal <95% Cl) 97
Total events 0
Heterogeneity: Not applicable Test for overall effect: Z - 1.55 [P - 0,12)
1 92 6.2%
0 120 27,0%
1 105 1S.S%
1 46 2.0%
363 54,1%
■ 3 (P = 0.33); I' = 13»
7.9% 7.9%
40 188
Sheii 2014 Lin 2013 Dong 2014 Subtotal (95% CI) Total events Heterogeneity: Tau' - 0.00; Chi1 - 0.00, df ■ Test for overall effect: 1 » 0.03 (P »0,98)
Total (95% Cl) 640
Total events ft
Heterogeneity; Tau' = 0.00; Chi' - 5,66, df■ Test for overall effect: Z - 0.06 [P - 0,95) Test for subgroup differences: Chi' ■ 2.SI. df
1 6« 7.3%
0 84 25.3%
0 3« 5.5»
190 38,1%
2 (P - l.OOJ; I1 - 0%
646 100-0%
0.58); I' - 0«
0.03 [-0,02, 0.07J 0.01 i-0,01, 0,031 -0.01 |-0,04, 0,021 0.04 [-0 04, 0 121 0-01 [-0 01, ÎM)2]
-0.03 [-0.07, 0 011 -0.03 [-0.07. 0.01]
0.00 [-0.04, 0.04] 0.00 [-0,02, 0,021 0.00 [-0,05. 0,0S] 0.00 [-0.02, 0.02]
0-00 [-0.01, OJOI]
-0.2 -0,1
0.1 0,2
■ 2 (P -0.25), I' - 23.8%
Favors Unilateral Favors Bilateral
Fig. 10 Screw failure. Abbreviations: CI, confidence interval; MIS, minimally invasive; M-H, Mantel-Haenszel; TLIF/PLIF, transforaminal or posterior lumbar interbody fusion.
Study or Subgroup
Unilateral Events Total
Bilateral Events Total
Risk Difference Weight M-H, Random, 95% CI
Risk Difference M-H, Random, 95% CI
Open TLIFiPLIF
Aoki 2012 Kai 2013 We 2012 Xue 2012 Subtotal {95% CI) Total events Heterogeneity; Tin' - 0,00; Chi' ■ Test for overall effect: 2 - 1.45 (P
Choi 2013 Dong 2014 Un 2013 Shen 2014 Subtotal [95% CI)
Total events Heterogeneity: Tau1 =
Test for overall effect;
0 Î4 3 23 5.9%
2 33 4 as 7.6%
4 56 i 52 12.3%
2 37 3 43 12.1%
150 153 37,9%
-0.13 [-0.2S, 0.021 -0.05 1-0.19,0,08) -0.02 1-0,13,0,081 -0.02 [-0.12,0,051 -0.04 [-0.10, 0.02]
1,68, df ■ ■ 0.15)
3 (P - 0.64); ll - 0S(
2 26 2 27 6.7% 0.00 t-0,14,0,15)
0 20 0 19 15.1» 0-00 [-0,09,0,091
1 43 0 42 34.5% 0.02 [-0.04, 0.05)
4 31 1 34 5.9% 0.04 [-0.11, 0,191
120 122 62.1% 0.02 [-0.03, 0.06]
Chi' = 0-31, df = 3 (P = 0.96): I1 = 0*
Z = 0.72 (P - 0.47}
275 100,0%
Total (95% CI) Total events ÎS 20
Heterogeneity: Tati1 - O.OO, Chi3 - 4.7S, df - 7 (P - 0.69): I3 -Test for overall effect; Z = 0.32 <P = 0.75) Test for subflroup differences: Chi' = 2.51, df - 1 (P- 0.11), f!
-O.Ol J-0,04, 0.03]
-0.5 -o'iS
0,25 0-5
Favors Unilateral Favors Bilateral
Fig. 11 Other complications. Abbreviations: CI, confidence interval; MIS, minimally invasive; M-H, Mantel-Haenszel; TLIF/PLIF, transforaminal or posterior lumbar interbody fusion.
some outcomes. The outcomes that occurred infrequently resulted in low power to detect statistical differences (see the online supplementary material).
• A serious risk of bias was present in all included studies. The indication of concealed allocation was not reported in any trial; blinding of evaluators occurred rarely; and
6 of 10 studies did not compare patients at baseline to ensure similar distribution of prognostic factors.
• This systematic review highlights the paucity of decent literature involving the efficacy of unilateral instrumentation in lumbar spinal surgery. A total of 10 studies met the inclusion criteria for this report. All the studies involved
192 Unilateral versus Bilateral Instrumentation in Spinal Surgery Molinari et al. Table 2 Strength of evidence summary
Outcome Studies (N) Strength of evidence MD or RD (95% CI) Favors
Nonunion 9 RCTs (626) Low RD: 0.01 (-0.01, 0.04) Neither
Low back pain 7 RCTs (437) Very low MD: 0.22 (-0.11, 0.56) Neither
Leg pain 4 RCTs (207) Low MD: 0.57 (-0.80, 1.93) Neither
ODI 6 RCTs (390) Very low MD: -0.19 (-2.64, 2.26) Neither
JOA 3 RCTs (194) Low MD: 0.85 (0.08, 1.61) Bilateral
Reoperation 5 RCTs (348) Low RD -0.03 (-0.09, 0.02) Neither
Infection 7 RCTs (534) Low RD -0.00 (-0.03, 0.02) Neither
Cage migration 6 RCTs (417) Low RD 0.02 (-0.03, 0.08) Neither
Screw failure 8 RCTs (573) Low RD 0.00 (-0.01, 0.01) Neither
Other complications 8 RCTs (545) Very low RD -0.01 (-0.04, 0.03) Neither
Abbreviations: CI, confidence interval; JOA, Japanese Orthopaedic Association; MD, mean difference; ODI, Oswestry Disability Index; RCT, randomized controlled trial; RD, risk difference.
Fig. 12 Preoperative radiographs show L4-5 grade 1 degenerative spondylolisthesis (A, B). Preoperative magnetic resonance imaging demonstrates L4-5 low-grade degenerative spondylolisthesis with severe spinal stenosis (C, D). Three-year postoperative anteroposterior and lateral radiographs demonstrating solid bilateral fusion without loosening of unilateral fixation (E, F).
degenerative pathology in the lumbar spine. The vast majority of procedures were single-level fusions for degenerative disk disorder or degenerative spondylolisthesis.
• Examination of the existing literature does not reveal significant differences in the patient outcomes between unilateral and bilateral fixation when performed for lumbar spinal pathology. A serious risk of bias exists in all the included studies resulting in an overall strength for these findings as either low or very low.
• The reported fusion rates with either unilateral or bilateral fixation are high, without a significant difference. The rates of instrumentation failure and nonunion were similarly low.
• The reported functional outcomes including ODI, VAS, and leg pain scores were not different between unilateral and bilateral fixation cases.
• The existing literature does not demonstrate outcome differences for MIS patients having either unilateral or bilateral lumbar fixation.
• The complication rates remain low with both unilateral and bilateral fixation for lumbar degenerative pathology.
• The theoretical advantages of unilateral fixation are many and include shorter operative times and reduced blood loss. The operative costs for implants are also reduced. From a technical standpoint, unilateral fixation cases do involve the utilization of less surgical resources. The degree of diminished segmental stability achieved with unilateral fixation did not appear to lead to worse outcomes or higher complication rates for single-level degenerative cases.
• We have demonstrated that there appears to be equivalency between unilateral and bilateral fixation when performed in adult patients who do not have significantly unstable lumbar conditions. It is important to emphasize that successful clinical and radiographic outcomes for unilateral fixation in patients with highly unstable lumbar conditions have not been described in this review. It is our opinion that additional prospective comparative studies are needed to better define the role of unilateral instrumentation in the treatment of lumbar spinal disorders.
Conclusions
The existing literature does not identify significant differences in clinical outcomes, union rates, and complications when unilateral instrumentation is used for degenerative pathologic conditions in the lumbar spine. The majority of published reports involve single-level lumbar unilateral instrumentation.
Disclosures
Robert W. Molinari, none Ahmed Saleh, none Robert Molinari, Jr., none Jeff Hermsmeyer, none Joseph R. Dettori, none
Acknowledgments
Analytic support for this work was provided by Spectrum Research, Inc. with funding from AOSpine. Special thanks to Mark Junge for his assistance in producing the figures for the meta-analysis.
References
1 Suk KS, Lee HM, Kim NH, Ha JW. Unilateral versus bilateral pedicle screw fixation in lumbar spinal fusion. Spine (Phila Pa 1976) 2000; 25(14):1843-1847
2 Kabins MB, Weinstein JN, Spratt KF, et al. Isolated L4-L5 fusions using the variable screw placement system: unilateral versus bilateral. J Spinal Disord 1992;5(1):39-49
3 Fernández-Fairen M, Sala P, Ramírez H, Gil J. A prospective randomized study of unilateral versus bilateral instrumented posterolateral lumbar fusion in degenerative spondylolisthesis. Spine (Phila Pa 1976) 2007;32(4):395-401
4 Zhao J, Zhang F, Chen X, Yao Y. Posterior interbody fusion using a diagonal cage with unilateral transpedicular screw fixation for lumbar stenosis. J Clin Neurosci 2011;18(3): 324-328
5 Balshem H, Helfand M, Schünemann HJ, et al. GRADE guidelines: 3. Rating the quality of evidence. J Clin Epidemiol 2011;64(4): 401 -406
6 Aoki Y, Yamagata M, Ikeda Y, et al. A prospective randomized controlled study comparing transforaminal lumbar interbody fusion techniques for degenerative spondylolisthesis: unilateral pedicle screw and 1 cage versus bilateral pedicle screws and 2 cages. J Neurosurg Spine 2012;17(2):153-159
7 Duncan JW, Bailey RA. An analysis of fusion cage migration in unilateral and bilateral fixation with transforaminal lumbar inter-body fusion. Eur Spine J 2013;22(2):439-445
8 Xue H, Tu Y, Cai M. Comparison of unilateral versus bilateral instrumented transforaminal lumbar interbody fusion in degenerative lumbar diseases. Spine J 2012;12(3):209-215
9 Zhang K, Sun W, Zhao CQ, et al. Unilateral versus bilateral instrumented transforaminal lumbar interbody fusion in two-level degenerative lumbar disorders: a prospective randomised study. Int Orthop 2014;38(1):111-116
10 Xie Y, Ma H, Li H, et al. Comparative study of unilateral and bilateral pedicle screw fixation in posterior lumbar interbody fusion. Orthopedics 2012;35(10):e1517-e1523
11 Choi UY, Park JY, Kim KH, et al. Unilateral versus bilateral percutaneous pedicle screw fixation in minimally invasive transfora-minal lumbar interbody fusion. Neurosurg Focus 2013;35(2):E11
12 Shen X, Zhang H, Gu X, Gu G, Zhou X, He S. Unilateral versus bilateral pedicle screw instrumentation for single-level minimally invasive transforaminal lumbar interbody fusion. J Clin Neurosci 2014;21(9):1612-1616
13 Dong J, Rong L, Feng F, et al. Unilateral pedicle screw fixation through a tubular retractor via the Wiltse approach compared with conventional bilateral pedicle screw fixation for single-segment degenerative lumbar instability: a prospective randomized study. J Neurosurg Spine 2014;20(1):53-59
14 Lin B, Xu Y, He Y, Zhang B, Lin Q, He M. Minimally invasive unilateral pedicle screw fixation and lumbar interbody fusion for the treatment of lumbar degenerative disease. Orthopedics 2013;36(8):e1071-e1076
15 Kai Z, Sun W, Zhao CQ, et al. Unilateral versus bilateral instrumented transforaminal lumbar interbody fusion in two-level degenerative lumbar disorders: a prospective randomised study. Int Orthop 2014;38(1):111-116
Editorial Perspective
All reviewers unanimously congratulate the authors in selecting a topic that questions conventional wisdom—more specifically, the need to use bilateral versus unilateral segmental fixation in form of pedicle screw/rod constructs to achieve fusion for single-level (monosegmental) arthrodesis procedures. The question of how much pedicle screw fixation is really needed to achieve a solid fusion in a stable single segment lumbar spine has been associated with divisive arguments from both sides, along with poignant arguments to bolster either stance. Shorter surgery, lower hardware insertion-related complication rates, and reduced implant costs are certainly strong arguments in favor of unilateral fixation for lumbar fusion. But where does one draw the line to the counterargument that unilateral instrumentation may be favored out of an economic gains motivation?
This enclosed systematic review offers an unprecedented overview on this topic and the findings—not surprisingly —are less than conclusive. The authors stressed two important factors: (1) the sample sizes were very small, thus opening the door for a type II error, and (2) due to the data reporting, Molinari et al were unable to provide a power calculation. The other consideration lies in the apparent and substantial selection bias commonly practiced in the source studies. In addition to the factors mentioned in the article, lumbar segmental stability and deformity correction, iatrogenic destabilization or variations in bone density were variables not really addressed by the source publications, thus adding more confounding factors to attempts at a definitive comparison. The use of recombinant bone substitutes in MIS constructs may add further confusion as the dosing and application strategy can influence healing results. That said, the available comparison data for the most part does not suggest any substantial difference in patient safety or reported outcomes, thus opening the door for further discussion. This leaves us as clinicians with an important question that remains to be
answered. For the clinical setting of a fusion to be performed in an inherently stable single degenerative lumbar spine segment, is a unilateral instrumentation an acceptable primary stabilization strategy or should we use the information gleaned from the preceding systematic review as an assurance that unilateral segmental instrumentation is sufficient in patients where bilateral instrumentation has for some reason not worked out? This important differentiation at this time is not resolvable. EBSJ invites further commentary from its global readership.
On a side note, the example chosen as an illustrative case was critiqued by our reviewers for several reasons. The case denotes an "accidental unilateral fixation"—one where a planned bilateral instrumentation was abandoned due to technical difficulties. The underlying pathology (an unstable-appearing degenerative spondylolisthesis) would clearly not be an ideal situation for unilateral fixation by any of the inclusion criteria of the studies used for this comparison. The report that the patient fused despite unilateral fixation also underscores the potential for bias in results reporting. It can be very challenging to establish a firm fusion in the lumbar spine in the presence of posterolateral fusions without interbody grafts and in the absence of hardware failure. In this patient, the spondylolisthesis clearly slipped more compared with the preoperative images, thus calling into question the assurance that an uneventful fusion resulted from the unilateral instrumentation. The symptom relief of the patient may have been mainly influenced by an effective stenosis decompression, and the back pain relief may have been secondary to that circumstance alone and may have nothing to do with a fusion or instrumentation. EBSJ would also like to point out that it does not endorse leaving an isolated screw without fixation purpose behind.
Finally, the reviewers recommended the readership take a look at the Web-based supplemental materials due to their depth and quality. EBSJ thanks the authors for their hard work on this topic.
