The effect of microprocessor controlled exo-prosthetic knees on limited community ambulators: systematic review and meta-analysis

Figures & data

Figure 1. Flowchart of literature review process and results. Final analysis run on 5 July 2021. Search strings and further information are presented in the Supplementary material.

Table 1. Assessment of internal validity according to AAOP SoSc evidence report guidelines.

Internal validity	Kahle et al. [9]	Hafner and Smith [46]	Theeven et al. [47,48]	Burnfield et al. [50] and Eberly et al. [51]	Hahn et al. [26]	Wong et al. [45]	Hahn et al. [28]	Hasenoehrl et al. [22]	Mileusnic et al. [23]	Kaufman et al. [24]	Lansade et al. [25]	Jayaraman et al. [44]	Davie-Smith and Carse [43]
Type of study (AAOP rating category)	E5	E5	E5	E5	O1/O2	E5	O1/O2	E5	E5	E2	E1	E1	O1
IV-1. Comparison or control group	n/a	n/a	n/a	n/a	n/a	n/a	n/a	n/a	n/a	+	+	+	n/a
IV-2. Groups formed by random assignment	n/a	n/a	n/a	n/a	n/a	n/a	n/a	n/a	n/a	n/a	+	+	n/a
IV-3. Groups comparable at baseline	n/a	n/a	n/a	n/a	n/a	n/a	n/a	n/a	n/a	+	+	+	n/a
IV-4. Groups handled the same way	n/a	n/a	n/a	n/a	n/a	n/a	n/a	n/a	n/a	+	+	+	n/a
IV-5. Control/comparison group appropriate	+	+	+	+	n/a	+	n/a	+	+	+	+	+	n/a
IV-6. Intervention(s) blinded	0	0	0	0	0	0	0	0	0	0	0	0	0
IV-7. Inclusion criteria appropriate	+	+	+	+	+	+	+	+	+	+	+	+	+
IV-8. Exclusion criteria appropriate	+	+	+	+	+	+	+	+	+	+	+	+	+
IV-9. fatigue and learning addressed	0	+	0	0	0	0	0	+	0	+	+	+	0
IV-10. Accommodation and washout addressed	+	+	+	+	+	+	+	+	+	+	+	+	+
IV-11. Attrition explained and <20%	+	+	0	+	+	+	+	0	0	0	+	0	0
IV-12. Attrition equal between groups	+	+	+	+	+	+	+	+	+	+	+	+	n/a
IV-13. Outcome measures reliable	+	+	+	+	+	+	+	+	+	+	+	+	+
IV-14. Statistical analysis appropriate	+	+	+	+	+	+	+	0	+	+	+	+	+
IV-15. Effect size reported	0	0	0	0	0	+	0	0	0	0	0	0	0
IV-16. Statistical significance reported	+	+	+	+	+	+	+	0	+	+	+	+	+
IV-17. Statistical power adequate	+	+	+	+	+	+	+	0	+	+	+	+	+
IV-18. Free from conflicts of interest	+	+	+	+	0	+	0	+	0	+	+	+	+
Total number of threats identified	3	2	4	3	4	2	4	6	5	3	2	3	4
Overall assessment of internal validity	High	High	High	High	Moderate	High	Moderate	Low	Moderate	High	High	High	High

The AAOP rating categories as applicable to this work are: for (quasi) experimental trials: E1 randomized controlled trial, E2 controlled trial, E5 controlled before-and-after trial; for observational trials: O1 cohort study, O2 case control study. The classification determines the applicability and non-applicability of the rating criteria. “+” denotes “no threat”, “0” identifies “potential threat”.

Table 2. Assessment of external validity and overall assessment assignment.

External Validity	Kahle et al. [9]	Hafner and Smith [46]	Theeven et al. [47,48]	Burnfield et al. [50] and Eberly et al. [51]	Hahn et al. [26]	Wong et al. [45]	Hahn et al. [28]	Hasenoehrl et al. [22]	Mileusnic et al. [23]	Kaufman et al. [24]	Lansade et al. [25]	Jayaraman et al. [44]	Davie-Smith and Carse [43]
EV-1. Sample characteristics adequately described	+	+	+	+	+	+	+	+	+	+	+	+
EV-2. Sample representative of the target population	+	+	+	+	+	+	+	+	+	+	+	+	+
EV-3. Outcome measures adequately described	+	+	+	+	+	+	+	+	+	+	+	+	+
EV-4. Outcome measures valid for this study	+	+	+	+	+	+	+	+	+	+	+	+	+
EV-5. Intervention adequately described	+	+	+	+	+	+	+	+	+	+	+	+	+
EV-6. Findings clinically significant/relevant	+	+	+	+	+	+	+	0	+	+	+	+	+
EV-7. Conclusions placed in context of existing literature	+	+	+	+	+	+	+	+	+	+	+	+	+
EV-8. Conclusions supported by findings	+	+	+	+	+	+	+	+	+	+	+	+	+
Total number of threats identified	0	0	0	0	0	0	0	1	0	0	0	0	0
Overall assessment of external validity (circle one)	High	High	High	High	High	High	High	Moderate	High	High	High	High	High
Overall assessment validity	High	High	High	High	Moderate	High	Moderate	Low	Moderate	High	High	High	High

Table 3. Design and demographic data of the MG2 subgroup analyzed.

		Kahle et al. [9]	Hafner and Smith [46]	Theeven et al. [47,48]	Burnfield et al. [50] and Eberly et al. [51]	Hahn et al. [26]	Wong et al. [45]	Hahn et al. [28]	Hasenoehrl et al. [22]	Mileusnic et al. [23]	Kaufman et al. [24]	Lansade et al. [25]	Jayaraman et al. [44]	Davie-Smith and Carse [43]
Design	# total	Prospective pre-post	Prospective, controlled X-over	Prospective pre-post	Prospective pre-post	Retrospective, pre-post	Prospective cohort	Retrospective, pre-post	Pilot A-B-A prospective	Prospective pre-post	Prospective X-over	Prospective RCT, X-over	Prospective RCT, X-over	Retrospective, cohort
Total (n)	2366	19	17	30	10	1223	8	899	5	29	50	35	10	31
MCFL-2 (n)	704	9	8	30	10	463^a	4	32^a	5	19	48	5	10	31
Etiology	Vascular disease (255)	7	1	6	6	151	4	9	1	6	25	24	10	5
	Trauma (202)	1	5	23	2	128		13	1	1	5	8		15
	Other (225)	1	2	1	0	184		11	3		20	3		11
Sex	Male (542)	n.r.	6	22	5	380	n.r.	27	4	18	28	27	4	21
	Female (161)	n.r.	2	8	5	83	n.r.	5	1	11	22	8	6	10
Age, years (SD)		67.1 (11.1)	57.1 (14.5)	59.1 (13.0)	62 (11.3)	60.33 (12.57)	64.75 (12.7)	54.1 (12.3)	68.2 (7.6)	63.2 (9.5)	69 (9)	65.6 (10.1	69 (9)	59.6 (11.0)
Intervention	C-Leg	9	8	28		382	4						10	14
	C-CPCT			28	10	81	0				x
	Kenevo									29		32		13
	Genium							32
	Other								G prototype		Rh3, Or2 Pl3			2xRh, 1xOr, 1xLinx
Control	Locked			n.r.		9	1	2	1	4		12	1	2
	Brake	4		n.r.	4	167		2	3	6	18%^b	6	1	12
	Polycentric	4	6	n.r.	2	250	1	11	1	8	74%^b	2	5	11
	Hydraulic	1	2	n.r.	4	32	1	5			10%^b	11	2	6
	MPK			0		0		0		4
	Other	0		n.r.		5		12			2%^b	4	1
Accommodation		90 days	2 months	1 week	3 months	1 day	48 weeks	1 week	4–6 weeks	7–8 days	10 weeks	3 months	3 months	6 months

(n.r.) identifies data not reported. Rh3 demarks Rheo 3, Or2 Orion2 and Pl3 Plié3 as investigated by Kaufman et al. Rh demarks Rheo, Or: Orion, C-CPCT identifies C-Leg compact. G-prototype is a non-commercial prosthesis. Age is displayed in years with the standard deviation in years indicated in brackets. Kaufman et al. did not reveal the assignment of MPKs used and hence no number can be extracted.

^a Subgroup of MG2 subjects not having been on an MPK prior.

^b As extracted from publication.

Table 4. Subgroup analysis for subjects with MCFL-2.

Kahle et al. [9]	All analysis extracted from MCFL-2 p Values from two-tailed paired t-test
Outcomes for meta-analysis	# falls, SSWS 75 m, FPWS 6 m, positive change of MCFL
Outcomes favoring MKP	Safety
	#falls (Δabs-1.67, Δrel-79.0%, g: –1.32, p: 0.04)
	Performance
	SSWS 75 m (Δabs 0.11 m/s, Δrel 18.6%, g: 0.52, p: 0.04)
	FPWS 75 m (Δabs 0.11 m/s, Δrel 15.4%, g: 0.53, p: 0.01)
	FPWS UT (Δabs 0.14 m/s, Δrel 25.2%, g: 0.66, p: <0.01)
	MRPP (Δabs 2.7, Δrel 38.8%, g: 0.73, p: 0.03)
	MCFL Change 2 to 3 (Δabs 4, Δrel 44.4%, p = 0.04)
	Perception
	Preference 89%
Outcomes w/o statistical significance	Safety
	# stumbles (Δabs-6, Δrel-76.0%, g: 0.65, p: 0.21)
	FPWS 6 m (Δabs 0.11 m/s, Δrel 13.3%, g: 0.36, p: 0.08)
Outcomes favoring NMPK	None
Other	PEQ cited for total group only, only total score available
	(Δabs 241.8, Δrel 26%, g: 0.92, p: 0.01)

The MRPP (Montreal Rehabilitation Performance Profile) was used to characterize stair descend. Bonferroni’s correction was reported.

Δabs = absolute difference; Δrel = relative difference; ABC = Activity based Balance Confidence Scale; ADAPT = Assessment of Daily Activity Performance in Transfemoral Amputees Test; AS1 = Activity Set 1; AS2 = Activity Set 2; AS3 =  Activity Set 3; AMP = amputee mobility predictor; CL = C-Leg; CPCT = C-Leg compact; FPWS = Fastest possible walking speed; GC =  gait cycle; g = Hedges’ g; GPS = Gait Profile Score; LW = level walking; MCFL = Mobility Classification Functional Level; MG = Mobility Grade; MPK = microprocessor controlled knee; MRPP = Montreal Rehabilitation Performance Profile; NMPK = non-microprocessor controlled knee; PEQ Prosthesis Evaluation Questionnaire; PLUS-M =  Prosthetic Limb Users Survey of Mobility; SSWS = self-selected walking speed; TUG = Timed up and Go test; UT = uneven terrain.

Table 5. Hafner et al. report on the MFCL-2 subgroup separately.

Hafner and Smith [46]	All analysis extracted from MFCL-2 subgroup p Values from two-tailed paired t-test
Outcomes for meta-analysis	# falls, fear of falling, positive change of MCFL, attention speed as SSWS, PEQ ambulation, PEQ appearance, PEQ residual limb, PEQ sounds, PEQ utility, PEQ well being
Outcomes favoring MKP	Safety
	#falls (Δabs-0.5, Δrel-100.0%, g: –1.31, p: 0.01)
	Fear of falling (Δabs-4.2, Δrel-4.5%, g: 1.47, p: 0.01)
	Fear of stumbles (Δabs: 11.6, Δrel: 15.7%, g: 0.90, p: 0.05)
	Performance
	Stair mobility (Δabs-5.7, Δrel 172.7%., g: 1.89, p: 0.01)
	Hill mobility (Δabs-2.1, Δrel 38.9%, g: 0.60, p: 0.01)
	Hill speed (Δabs-0.46 m/s, Δrel 27.1%, g: 1.22, p: 0.002)
	Obstacle course (Δabs-0.09 m/s, Δrel 11.3%, g: 0.33, p: 0.02)
	Attention speed (Δabs-0.1 m/s, Δrel 12.1%, g: 0.54, p: 0.02
	Multitasking (Δabs-15.0s, Δrel 21.2%, g: 1.00, p: 0.04
	MFCL change: (Δabs-4, Δrel 50%, p = 0.03)
Outcomes w/o statistical significance	Safety
	# stumbles
	# semi-controlled falls
	Performance
	Activity avoidance (VAS)
	Attention accuracy (VAS)
	Mental energy expenditure, confidence while walking, difficulty with concentration
	Perception
	Frustration with falls
	Embarrassment with falls
	Satisfaction VAS
	PEQ ambulation, appearance, frustration, perceived response, residual limb, social burden, sounds, utility, well-being.
Outcomes favoring NMPK	None
Comments	AMP changes correlate strongly with MG changes

Δabs: absolute difference; Δrel: relative difference; ABC: Activity based Balance Confidence Scale; ADAPT [53]: Assessment of Daily Activity Performance in Transfemoral Amputees Test; AS1: Activity Set 1; AS2: Activity Set 2; AS3: Activity Set 3; AMP: Amputee Mobility Predictor; CL: C-Leg; CPCT: C-Leg compact; FPWS: fastest possible walking speed; GC: gait cycle; g: Hedges’ g; GPS: Gait Profile Score; LW: level walking; MCFL: Mobility Classification Functional Level; MG: mobility grade; MPK: microprocessor controlled knee; MRPP: Montreal Rehabilitation Performance Profile; NMPK: non-microprocessor controlled knee; PEQ: Prosthesis Evaluation Questionnaire; PLUS-M: Prosthetic Limb Users Survey of Mobility; SSWS: self-selected walking speed; TUG: timed up and go test; UT: uneven terrain.

Table 6. Theeven et al. investigated the impact of C-Leg (CL) and C-Leg Compact (CPCT) on MCFL-2 subgroups with high, intermediate (int) and low activity level.

Theeven et al./Seelen et al.	MFCL-2 stratified to subgroups low, int. and high. p Values from Wilcoxon signed rank test for paired observations
Outcomes for meta-analysis	PEQ ambulation, PEQ appearance, PEQ residual limb, PEQ sounds, PEQ utility, PEQ well-being, patients perception safety on stumbles
Outcomes favoring MKP	Safety Patients perception safety on Stumbles (Δabs 12.2, Δrel 17.6%, g: 0.79, p < 0.001) Stops (Δabs 7.8, Δrel 10.3%, g: 0.52, p: 0.005) Sidesteps (Δabs 10.3, Δrel 14.2%, g: 0.70, p: 0.002 Patients perception stability on Stops (Δabs 10.8, Δrel 14.7%, g: 0.73, p < 0.001) Sidesteps (Δabs 7.6, Δrel 10%, g: 0.54, p: 0.007) Obstacle clearance (Δabs 13.3, Δrel 18.9%, g: 0.81, p: 0.001) Performance ADAPT performance scores AS1 total (p = 0.0001 CL; 0.002 CPCT) AS1 high (p = 0.01 CL; 0.019 CPCT) AS3 high (p = 0.023) AS1 int (p = 0.004 CL; p = 0.008 CPCT) AS2 int (p = 0.016 CL) SAI CL int. (Δabs 5.5, Δrel 171.9%, , p < 0.025) SAI CL high (Δabs 5.3, Δrel 147.2%, , p < 0.025) SAI CPCT int. (Δabs 4,6, Δrel 143.8%, , p < 0.025) SAI CPCT high (Δabs 5,2, Δrel 144.4%, p < 0.025) Perception ADAPT perceived difficulty (VAS): Total group AS2 (p = 0.023) and AS3 (p = 0.008) (CL) Preference: 28 out of 29 PEQ Ambulation high (Δabs 11.8, Δrel 17.2%, g: 0.79, p: 0.06) Resid. limb high (Δabs 13.8, Δrel 27.0%, g: 0.66, p: 0.003) Utility high (Δabs 8.7, Δrel 12.9%, g: 0.60, p: 0.04) Satisfact. w. walking high (Δabs 10.6, Δrel 15.0%, g: 0.55, p: 0.04) Ambulation int (Δabs 8, Δrel 11.3%, g: 0.47, p: 0.009) Utility int (Δabs 11.6, Δrel 17.1%, g: 0.76, p: 0.02) Satisfact. w. walking int (Δabs 28.1, Δrel 49.8%, g: 1.37, p: 0.007)
Outcomes w/o statistical significance	Performance ADAPT AS2 high ADAPT AS3 high CL ADAPT AS2 int ADAPT AS3 int ADAPT all AS in low Perception ADAPT perceived difficulty total group for CPCT ADAPT perceived difficulty in subgroup stratification PEQ subscales as not cited above. SAI and HAI in “Low” and all not cited above.
Outcomes favoring NMPK	None
Comments	Absolute changes on ADAPT are presented in graphs only

The ADAPT [70] instrument investigates three different activity sets (AS 1–3). Significant outcomes were observed in the groups high and intermediate while no significant change was to be detected in group “low”. Results for SAI and HAI were extracted from Seelen et al.

Δabs: absolute difference; Δrel: relative difference; ABC: Activity based Balance Confidence Scale; ADAPT [53]: Assessment of Daily Activity Performance in Transfemoral Amputees Test; AS1: Activity Set 1; AS2: Activity Set 2; AS3: Activity Set 3; AMP: Amputee Mobility Predictor; CL: C-Leg; CPCT: C-Leg compact; FPWS: fastest possible walking speed; GC: gait cycle; g: Hedges’ g; GPS: Gait Profile Score; LW: level walking; MCFL: Mobility Classification Functional Level; MG: mobility grade; MPK: microprocessor controlled knee; MRPP: Montreal Rehabilitation Performance Profile; NMPK: non-microprocessor controlled knee; PEQ: Prosthesis Evaluation Questionnaire; PLUS-M: Prosthetic Limb Users Survey of Mobility; SSWS: self-selected walking speed; TUG: timed up and go test; UT: uneven terrain.

Table 7. As only relative differences were reported, TUG, ABC, and Houghton scores could not be utilized for the meta-analysis.

Burnfield/Eberly et al.	MFCL-2 amputees on C-Leg Compact p Values from two-tailed paired t-test
Outcomes for meta-analysis	SSWS (free walking); FPWS (fast walking)
Outcomes favoring MKP	Safety
	TUG (Δabs 6.8 s, Δrel 38.4%, g: n.a., p < 0.02)
	ABC (Δabs 15.6, Δrel 26.0%, g: n.a., p < 0.001)
	Performance
	Level walking
	SSWS (Δabs 0.11 m/s, Δrel 20.0%, g: 0.54, p < 0.001)
	Cadence (Δabs 7.0, Δrel 9.2%, g: 0.39, p: 0.001)
	Stride length (Δabs 0.1 m, Δrel 11.9%, g: 0.55, p: 0.003)
	FPWS (Δabs 0.17 m/s, Δrel 23.7%, g: 0.69, p < 0.001)
	Cadence (Δabs 8.8 m, Δrel 10.1%, g: 0.50, p: 0.002)
	Stride length (Δabs 0.13 m, Δrel 13.6%, g: 0.55, p < 0.001)
	Single limb support SSWS (%GC), p = 0.048
	Heel rise prosthetic limb, SSWS and FPWS (p: 0.33/0.019)
	Slope ascend
	Speed (Δabs 0.13 m/s, Δrel 27.7%, g: 0.56, p: 0.01)
	Cadence (Δabs 9.7 ,Δrel 14.3%, g: 0.59, p: 0.01)
	Stride length (Δabs 0.10 m, Δrel 11.9%, g 0.52, p: 0.02)
	Slope descend
	Speed (Δabs 0.14 m, Δrel 36.2%, g: 0.64, p < 0.001)
	Cadence (Δabs 10.9, Δrel 16.0%, g: 0.57, p < 0.02)
	Stride length (Δabs 0.13 m, Δrel 13.7%, g: 0.57, p: 0.003)
	Perception
	PEQ ambulation (Δabs 1.6, Δrel 25.4%, g: n.a., p: 0.04)
Outcomes w/o statistical significance	Houghton’s Scale (p: 0.058)
	Single limb support FPWS (%GC) prosthetic side
	EMG signals (various)
Outcomes favoring NMPK	None-
Other	Both publications report in detail on further biomechanical parameters (kinematic and kinetic) as well as EMG. The data suggest in general either an increase in symmetry or remain not significant. We refrain from detailed reporting here

Δabs: absolute difference; Δrel: relative difference; ABC: Activity based Balance Confidence Scale; ADAPT [53]: Assessment of Daily Activity Performance in Transfemoral Amputees Test; AS1: Activity Set 1; AS2: Activity Set 2; AS3: Activity Set 3; AMP: Amputee Mobility Predictor; CL: C-Leg; CPCT: C-Leg compact; FPWS: fastest possible walking speed; GC: gait cycle; g: Hedges’ g; GPS: Gait Profile Score; LW: level walking; MCFL: Mobility Classification Functional Level; MG: mobility grade; MPK: microprocessor controlled knee; MRPP: Montreal Rehabilitation Performance Profile; NMPK: non-microprocessor controlled knee; PEQ: Prosthesis Evaluation Questionnaire; PLUS-M: Prosthetic Limb Users Survey of Mobility; SSWS: self-selected walking speed; TUG: timed up and go test; UT: uneven terrain.

Table 8. Subgroup analysis on original data.

Hahn et al. [26]	MOBIS 2 data extracted from 463 subjects
Outcomes for meta-analysis	Fear of falling, change of mobility grade
Outcomes favoring MKP	# subjects reporting
	Safety
	Reduced fear of falling 87%; 95%CI [84%; 90%]
	Perception of safety 85%; 95%CI [82%; 88%]
	Performance
	# subjects rated by CPO to have improved
	Relieve of sound leg 95%; 95%CI [93%; 97%]
	Attention 90%; 95%CI [87%; 93%]
	Gait harmonization 93%; 95%CI [90%; 95%]
	Gait speed variation 86%; 95%CI [83%; 89%]
	Overall effort 84%; 95%CI [81%; 87%]
	Reduction walking aids 37%; 95%CI [27%; 41%]
	Change MG2 to MG3 50%; 95%CI [45%; 54%]
Outcomes w/o statistical significance	None
Outcomes favoring NMPK	None
Other	Confidence intervals of subgroup MOBIS 2 extracted in post hoc analysis

Δabs: absolute difference; Δrel: relative difference; ABC: Activity based Balance Confidence Scale; ADAPT [53]: Assessment of Daily Activity Performance in Transfemoral Amputees Test; AS1: Activity Set 1; AS2: Activity Set 2; AS3: Activity Set 3; AMP: Amputee Mobility Predictor; CL: C-Leg; CPCT: C-Leg compact; FPWS: fastest possible walking speed; GC: gait cycle; g: Hedges’ g; GPS: Gait Profile Score; LW: level walking; MCFL: Mobility Classification Functional Level; MG: mobility grade; MPK: microprocessor controlled knee; MRPP: Montreal Rehabilitation Performance Profile; NMPK: non-microprocessor controlled knee; PEQ: Prosthesis Evaluation Questionnaire; PLUS-M: Prosthetic Limb Users Survey of Mobility; SSWS: self-selected walking speed; TUG: timed up and go test; UT: uneven terrain.

Table 9. Subgroup extracted from publication.

Wong et al.	4 subjects with MFCL 2 p Values from two-tailed paired t-test
Outcomes for meta-analysis	# falls; fear of falling, risk of falling (TUG)
Outcomes favoring MKP
Outcomes w/o statistical significance	# falls (Δabs −0.75, Δrel −60.0%, g: −0.67, p: 0.21)
	Fear of falling (Δabs −3.0, Δrel −80.0%, g: −0.84, p: 0.24)
	TUG (Δabs −26.3, Δrel −45.1%, g: −0.39, p: 0.35)
	ABC (Δabs −41.1, Δrel −118.7%, g: −1.05., p: 0.09)
	BBS (Δabs −2.25, Δrel −7.3%, g: −0.09., p: 0.56)
	Houghton (Δabs 2.75 ,Δrel 100.0%, g: 0.58, p: 0.19)
Outcomes favoring NMPK	None
Other

Δabs: absolute difference; Δrel: relative difference; ABC: Activity based Balance Confidence Scale; ADAPT [53]: Assessment of Daily Activity Performance in Transfemoral Amputees Test; AS1: Activity Set 1; AS2: Activity Set 2; AS3: Activity Set 3; AMP: Amputee Mobility Predictor; CL: C-Leg; CPCT: C-Leg compact; FPWS: fastest possible walking speed; GC: gait cycle; g: Hedges’ g; GPS: Gait Profile Score; LW: level walking; MCFL: Mobility Classification Functional Level; MG: mobility grade; MPK: microprocessor controlled knee; MRPP: Montreal Rehabilitation Performance Profile; NMPK: non-microprocessor controlled knee; PEQ: Prosthesis Evaluation Questionnaire; PLUS-M: Prosthetic Limb Users Survey of Mobility; SSWS: self-selected walking speed; TUG: timed up and go test; UT: uneven terrain.

Table 10. Subgroup extracted from original data.

Hahn et al. [28]	All analysis extracted from MG2 prior on NMPK
Outcomes for meta-analysis	Change mobility grade
Outcomes favoring MKP	Safety
	# subjects reporting improved
	Safety perception 100%; 95%CI [91.2%; 100%]
	Performance
	# subjects rated improved by CPOs (functional benefits)
	Symmetry 95.5%; 95%CI [89.6%; 100%]
	Relief sound side 89.3%; 95%CI [77.6%; 100%]
	Attention 89.7%; 95%CI [78.2%; 100%]
	Variability speed 78.6%; 95%CI [63.1%; 94.1%]
	Overall effort 85.7%; 95%CI [72.5%; 98.9%]
	# subjects rated improved by prosthetist
	(Advanced maneuvers)
	Stair ambulation (alt. desc.) 41.3%; 95%CI [22.8%; 60%]
	Ramps (desc.) 86.8%; 95%CI [74.0%; 99.6%]
	Walking backwards 75.0%; 95%CI [58.7%; 91.3%]
	Small steps 65.0%; 95%CI [47.0%; 83.0%]
	Obstacles 84.0%; 95%CI [70.2%; 97.8%]
	Heavy loads 78.2%; 95%CI [62.6%; 93.8%]
	“door test” 92.3%; 95%CI [82.6%; 100%]
	Perception
	# subjects reporting improvement in
	Toileting 56.0%; 95%CI [37.3%; 74.7%]
	Stair ambulation (alt. desc.) 88.9%; 95%CI [77.0%; 100%]
	Standing ramps 92.0%; 95%CI [81.8%; 100%]
	Variability gait speed 92.6%; 95%CI [82.7%; 100%]
	Carrying w visual obstruction 88.0%; 95%CI [75.7%; 100%]
Outcomes w/o statistical significance	None
Outcomes favoring NMPK	None
Other

Δabs: absolute difference; Δrel: relative difference; ABC: Activity based Balance Confidence Scale; ADAPT [53]: Assessment of Daily Activity Performance in Transfemoral Amputees Test; AS1: Activity Set 1; AS2: Activity Set 2; AS3: Activity Set 3; AMP: Amputee Mobility Predictor; CL: C-Leg; CPCT: C-Leg compact; FPWS: fastest possible walking speed; GC: gait cycle; g: Hedges’ g; GPS: Gait Profile Score; LW: level walking; MCFL: Mobility Classification Functional Level; MG: mobility grade; MPK: microprocessor controlled knee; MRPP: Montreal Rehabilitation Performance Profile; NMPK: non-microprocessor controlled knee; PEQ: Prosthesis Evaluation Questionnaire; PLUS-M: Prosthetic Limb Users Survey of Mobility; SSWS: self-selected walking speed; TUG: timed up and go test; UT: uneven terrain.

Table 11. The exploratory pilot study by Hasenoehrl et al. showed high heterogeneity of subjects’ baseline status and subsequent performance.

Hasenoehrl et al.	Analysis extracted from publication
Outcomes for meta-analysis	TUG, SSWS, FPWS
Outcomes favoring MKP
Outcomes w/o statistical significance	TUG, 2 min walk test, SSWS, FPWS, AMP, BBS, MG change, observed fall events, Houghton’s Scale, OPUS, LCI, ABC, SF-36, 3D gait analysis
Outcomes favoring NMPK	None
Other	Due to the low number of patients and the heterogeneity of the group the authors used descriptive analysis only. The explicit reporting of the individual scores allowed for post hoc analysis and use in the meta-analysis

Δabs: absolute difference; Δrel: relative difference; ABC: Activity based Balance Confidence Scale; ADAPT [53]: Assessment of Daily Activity Performance in Transfemoral Amputees Test; AS1: Activity Set 1; AS2: Activity Set 2; AS3: Activity Set 3; AMP: Amputee Mobility Predictor; CL: C-Leg; CPCT: C-Leg compact; FPWS: fastest possible walking speed; GC: gait cycle; g: Hedges’ g; GPS: Gait Profile Score; LW: level walking; MCFL: Mobility Classification Functional Level; MG: mobility grade; MPK: microprocessor controlled knee; MRPP: Montreal Rehabilitation Performance Profile; NMPK: non-microprocessor controlled knee; PEQ: Prosthesis Evaluation Questionnaire; PLUS-M: Prosthetic Limb Users Survey of Mobility; SSWS: self-selected walking speed; TUG: timed up and go test; UT: uneven terrain.

Table 12. Analysis from publication.

Mileusnic et al.	All group analysis
Outcomes for meta-analysis	#falls, fear of falling
Outcomes favoring MKP	Safety
	# non-stumblers (Δrel 42%, p: 0.044)
	Performance
	Wheelchair dependence -(Δrel 43%, p: 0.005)
	Perception
	Responders for
	Concentration 79%; 95%CI [61%; 97%]
	Perceived exertion 84%; 95%CI [67%; 100%]
	Level walking 79%; 95%CI [42%; 86%]
	Uneven ground 64%; 95%CI [42%; 86%]
	Improved standing 69%; 95%CI [46%; 92%]
	Preference 89%; 95%CI [78%; 100%]
Outcomes w/o statistical significance	Fear of falling, LCI, PLUS-M, Houghton’s Scale, number of non-fallers, stairs, satisfaction with prosthesis
Outcomes favoring NMPK	None
Other	Confidence interval analysis was conducted retrospectively based on the information given in the publication

Δabs: absolute difference; Δrel: relative difference; ABC: Activity based Balance Confidence Scale; ADAPT [53]: Assessment of Daily Activity Performance in Transfemoral Amputees Test; AS1: Activity Set 1; AS2: Activity Set 2; AS3: Activity Set 3; AMP: Amputee Mobility Predictor; CL: C-Leg; CPCT: C-Leg compact; FPWS: fastest possible walking speed; GC: gait cycle; g: Hedges’ g; GPS: Gait Profile Score; LW: level walking; MCFL: Mobility Classification Functional Level; MG: mobility grade; MPK: microprocessor controlled knee; MRPP: Montreal Rehabilitation Performance Profile; NMPK: non-microprocessor controlled knee; PEQ: Prosthesis Evaluation Questionnaire; PLUS-M: Prosthetic Limb Users Survey of Mobility; SSWS: self-selected walking speed; TUG: timed up and go test; UT: uneven terrain.

Table 13. The authors kindly provided their de-personalized data set for meta-analyses.

Kaufman et al.	Analysis from original data
Outcomes for meta-analysis	# falls, PEQ ambulation, PEQ appearance, PEQ residual limb, PEQ sounds, PEQ utility, PEQ well being
Outcomes favoring MKP	Safety
	# falls (Δabs −2.91, Δrel −62.1%, g: –0.50, p = 0.05)
	Performance
	Activity time (Δabs 4%, Δrel 20.%, g: 2.5, p: 0.02)
	Sitting time (Δabs 9%, Δrel 17%, g: 2.96, p: 0.01)
	Perception
	PEQ dimensions
	Appearance (Δabs 10.9, Δrel 17.0%, g: 0.55, p < 0.01)
	Utility (Δabs 9.9, Δrel 18.4%, g: 0.52, p < 0.01)
Outcomes w/o statistical significance	Gait complexity
	PEQ dimensions
	Ambulation
	Frustration
	Perceived response
	Residual limb health
	Social burden sounds
	Well being
Outcomes favoring NMPK	None
Other	All analysis extracted from MG2. p Values from two-tailed paired t-test

The individual MPK assignment remained blinded.

Δabs: absolute difference; Δrel: relative difference; ABC: Activity based Balance Confidence Scale; ADAPT [53]: Assessment of Daily Activity Performance in Transfemoral Amputees Test; AS1: Activity Set 1; AS2: Activity Set 2; AS3: Activity Set 3; AMP: Amputee Mobility Predictor; CL: C-Leg; CPCT: C-Leg compact; FPWS: fastest possible walking speed; GC: gait cycle; g: Hedges’ g; GPS: Gait Profile Score; LW: level walking; MCFL: Mobility Classification Functional Level; MG: mobility grade; MPK: microprocessor controlled knee; MRPP: Montreal Rehabilitation Performance Profile; NMPK: non-microprocessor controlled knee; PEQ: Prosthesis Evaluation Questionnaire; PLUS-M: Prosthetic Limb Users Survey of Mobility; SSWS: self-selected walking speed; TUG: timed up and go test; UT: uneven terrain.

Table 14. Data extracted from original publication.

Lansade et al.	Data extracted from intent to-treat analysis
Outcomes for meta-analysis	Risk of falling (TUG), # falls
Outcomes favoring MKP	Safety
	TUG (pp analysis) (Δabs 3.7 s, Δrel 16.0%, g: 0.70, p = 0.001)
	Perception
	LCI-5 (Δabs 2.4, Δrel 5.6%, g: 0.34, p = 0.02)
	QUEST 2.0 (Δabs 0.8, Δrel 17%, d: n.a., p = 0.001)
	SF-36 (MCS) (Δabs 3.86, Δrel 7.4%, g: 0.45, p = 0.04)
Outcomes w/o statistical significance	# falls, # fallers, SF-36 PCS
Outcomes favoring NMPK	None
Other	p Values from two-tailed paired t-test

Δabs: absolute difference; Δrel: relative difference; ABC: Activity based Balance Confidence Scale; ADAPT [53]: Assessment of Daily Activity Performance in Transfemoral Amputees Test; AS1: Activity Set 1; AS2: Activity Set 2; AS3: Activity Set 3; AMP: Amputee Mobility Predictor; CL: C-Leg; CPCT: C-Leg compact; FPWS: fastest possible walking speed; GC: gait cycle; g: Hedges’ g; GPS: Gait Profile Score; LW: level walking; MCFL: Mobility Classification Functional Level; MG: mobility grade; MPK: microprocessor controlled knee; MRPP: Montreal Rehabilitation Performance Profile; NMPK: non-microprocessor controlled knee; PEQ: Prosthesis Evaluation Questionnaire; PLUS-M: Prosthetic Limb Users Survey of Mobility; SSWS: self-selected walking speed; TUG: timed up and go test; UT: uneven terrain.

Table 15. Jayaraman et al. provide both groups with the same foot component.

Jayaraman et al.	Data extracted from original publication
	p Values relate to ANOVA analysis
Outcomes for meta-analysis	Fear of falling (mFES), risk of falling (TUG), MG change
	Gait speed, PEQ-A
Outcomes favoring MKP	Safety
	mFES (Δabs 1.42, Δrel 16.7%, g: 1.55, p = 0.004)
	Performance
	10 mWT: (Δabs −2.6 s, Δrel −14.4%, g: −0.33, p = 0.045)
	Speed 10 mWT (Δabs 0.1 m/s, Δrel 15.2%, g: 0.34, p = 0.009)
	Change MG2 to MG3 50%; 95%CI [12%; 88%]
	Perception
	PEQ-A (Δabs 22.77, Δrel 38.5%, g: 1.15, p = 0.003)
Outcomes w/o statistical significance	6 minWT, BBS, AMPPRO, FSST, TUG
Outcomes favoring NMPK	None
Other	MG change cited in publication based on changes in the 10 mWT gait speed

Δabs: absolute difference; Δrel: relative difference; ABC: Activity based Balance Confidence Scale; ADAPT [53]: Assessment of Daily Activity Performance in Transfemoral Amputees Test; AS1: Activity Set 1; AS2: Activity Set 2; AS3: Activity Set 3; AMP: Amputee Mobility Predictor; CL: C-Leg; CPCT: C-Leg compact; FPWS: fastest possible walking speed; GC: gait cycle; g: Hedges’ g; GPS: Gait Profile Score; LW: level walking; MCFL: Mobility Classification Functional Level; MG: mobility grade; MPK: microprocessor controlled knee; MRPP: Montreal Rehabilitation Performance Profile; NMPK: non-microprocessor controlled knee; PEQ: Prosthesis Evaluation Questionnaire; PLUS-M: Prosthetic Limb Users Survey of Mobility; SSWS: self-selected walking speed; TUG: timed up and go test; UT: uneven terrain.

Table 16. Davie-Smith et al. applied 18 times Bonferroni’s correction to significance level.

Davie-Smith et al.	Authors kindly provided re-analyzed data on request for meta-analysis and effect size determination
Outcomes for meta-analysis	#falls, gait speed, ABC
Outcomes favoring MKP	Safety
	# falls (Δabs −1.41, Δrel −73.1%, g: −0.56, p = 0.007)
	Performance
	2 minWT (Δabs 16.84 m, Δrel 18.1%, g: 0.54, p = 0.001)
	Gait speed (Δabs 0.14 m/s, Δrel 17.9%, g: 0.53, p = 0.001)
	AMP (Δabs 3.49, Δrel 9.6%, g: 1.01, p < 0.001)
	L-TEST (Δabs −8.01 s, Δrel −21.1%, g: −0.56, p < 0.001)
	Perception
	ABC (Δabs 9.66, Δrel 14.7%, g: 0.56, p < 0.001)
	PLUS-M (Δabs 4.67, Δrel 10.6%, g: 0.47, p = 0.004)
Outcomes w/o statistical significance	EQ-5D-5L (Δabs 0.08, Δrel 12.1%, g: 0.53, p = 0.01), GPS, SCS
Outcomes favoring NMPK	None
Other	Low mobility classification based on AMP ≤40. p Values from paired t-test, gait speed derived from 2 minWT. For #falls outlier correction was applied and p values from group comparison

Δabs: absolute difference; Δrel: relative difference; ABC: Activity based Balance Confidence Scale; ADAPT [53]: Assessment of Daily Activity Performance in Transfemoral Amputees Test; AS1: Activity Set 1; AS2: Activity Set 2; AS3: Activity Set 3; AMP: Amputee Mobility Predictor; CL: C-Leg; CPCT: C-Leg compact; FPWS: fastest possible walking speed; GC: gait cycle; g: Hedges’ g; GPS: Gait Profile Score; LW: level walking; MCFL: Mobility Classification Functional Level; MG: mobility grade; MPK: microprocessor controlled knee; MRPP: Montreal Rehabilitation Performance Profile; NMPK: non-microprocessor controlled knee; PEQ: Prosthesis Evaluation Questionnaire; PLUS-M: Prosthetic Limb Users Survey of Mobility; SSWS: self-selected walking speed; TUG: timed up and go test; UT: uneven terrain.

Figure 2. Results of meta-analysis of safety-related outcomes. (A) The number of falls presented as standardized mean difference between MPK and NMPK. (B) The fear of falling presented as standardized mean difference between MPK and NMPK. (C) The “timed up and go” completion time presented as standardized mean difference between MPK and NMPK.

Figure 3. Results of meta-analysis of outcomes related to function, performance and mobility. (A) Change of mobility grade from 2 to 3 with MPK fitting. (B) Change in self-selected walking speed presented as standardized mean difference between MPK and NMPK. (C) Change in fastest possible walking speed presented as standardized mean difference between MPK and NMPK.

Kahle JT, Highsmith MJ, Hubbard SL. Comparison of non-microprocessor knee mechanism versus C-Leg on Prosthesis Evaluation Questionnaire, stumbles, falls, walking tests, stair descent, and knee preference. J Rehabil Res Dev. 2008;45(1):1–14.

 PubMedGoogle Scholar

Hafner BJ, Smith DG. Differences in function and safety between Medicare Functional Classification Level-2 and -3 transfemoral amputees and influence of prosthetic knee joint control. J Rehabil Res Dev. 2009;46(3):417–434.

 PubMedGoogle Scholar

Theeven P, Hemmen B, Rings F, et al. Functional added value of microprocessor-controlled knee joints in daily life performance of Medicare Functional Classification Level-2 amputees. J Rehabil Med. 2011;43(10):906–915.

 PubMed Web of Science ®Google Scholar

Theeven PJ, Hemmen B, Geers RP, et al. Influence of advanced prosthetic knee joints on perceived performance and everyday life activity level of low-functional persons with a transfemoral amputation or knee disarticulation. J Rehabil Med. 2012;44(5):454–461.

 PubMed Web of Science ®Google Scholar

Burnfield JM, Eberly VJ, Gronely JK, et al. Impact of stance phase microprocessor-controlled knee prosthesis on ramp negotiation and community walking function in K2 level transfemoral amputees. Prosthet Orthot Int. 2012;36(1):95–104.

 PubMed Web of Science ®Google Scholar

Eberly VJ, Mulroy SJ, Gronley JK, et al. Impact of a stance phase microprocessor-controlled knee prosthesis on level walking in lower functioning individuals with a transfemoral amputation. Prosthet Orthot Int. 2014;38(6):447–455.

 PubMed Web of Science ®Google Scholar

Hahn A, Lang M. Effects of mobility grade, age, and etiology on functional benefit and safety of subjects evaluated in more than 1200 C-Leg trial fittings in Germany. J Prosthet Orthot. 2015;27(3):86–95.

 Google Scholar

Wong K, Rheinstein J, Stern MA. Benefits for adults with transfemoral amputation and peripheral artery disease using microprocessor compared with non-microprocessor prosthetic knees. Arch Phys Med Rehabil. 2015;94(10):804–810.

 Google Scholar

Hahn A, Lang M, Stuckart C. Analysis of clinically important factors on the performance of advanced hydraulic, microprocessor-controlled exo-prosthetic knee joints based on 899 trial fittings. Medicine. 2016;95(45):e5386.

 PubMed Web of Science ®Google Scholar

Hasenoehrl T, Schmalz T, Windhager R, et al. Safety and function of a prototype microprocessor-controlled knee prosthesis for low active transfemoral amputees switching from a mechanic knee prosthesis: a pilot study. Disabil Rehabil Assist Technol. 2017;95:1–9.

 Google Scholar

Mileusnic MP, Hahn A, Reiter S. Effects of a novel microprocessor-controlled knee, kenevo, on the safety, mobility, and satisfaction of lower-activity patients with transfemoral amputation. J Prosthet Orthot. 2017;29(4):198–205.

 Google Scholar

Kaufman KR, Bernhardt KA, Symms K. Functional assessment and satisfaction of transfemoral amputees with low mobility (FASTK2): a clinical trial of microprocessor-controlled vs. non-microprocessor-controlled knees. Clin Biomech. 2018;58:116–122.

 PubMed Web of Science ®Google Scholar

Lansade C, Vicaut E, Paysant J, et al. Mobility and satisfaction with a microprocessor-controlled knee in moderately active amputees: a multi-centric randomized crossover trial. Ann Phys Rehabil Med. 2018;61(5):278–285.

 PubMed Web of Science ®Google Scholar

Jayaraman C, Mummidisetty CK, Albert MV, et al. Using a microprocessor knee (C-Leg) with appropriate foot transitioned individuals with dysvascular transfemoral amputations to higher performance levels: a longitudinal randomized clinical trial. J Neuroeng Rehabil. 2021;18(1):88.

 PubMed Web of Science ®Google Scholar

Davie-Smith F, Carse B. Comparison of patient-reported and functional outcomes following transition from mechanical to microprocessor knee in the low-activity user with a unilateral transfemoral amputation. Prosthet Orthot Int. 2021;45(3):198–204.

 PubMed Web of Science ®Google Scholar

Theeven P, Hemmen B, Stevens C, et al. Feasibility of a new concept for measuring actual functional performance in daily life of transfemoral amputees. J Rehabil Med. 2010;42(8):744–751.

 PubMed Web of Science ®Google Scholar

The Rehabilitation of Individuals with Lower Limb Amputation Work Group. 2017. VA/DoD clinical practice guideline for rehabilitation of individuals with lower limb amputation. Version 2.0 – 2017. Available from: https://www.healthquality.va.gov/guidelines/Rehab/amp/VADoDLLACPG092817.pdf

 Google Scholar