Study reference

Study characteristics

Patient characteristics

Intervention (I)

Comparison / control (C)

Follow-up

Outcome measures and effect size

Comments

MacKenzie, 2004

Type of study: prospective cohort

Setting and country: multi-institutional, 8 level-1 trauma centers.

Funding and conflicts of interest:

“.. one or

more of the authors received Grant ROI-AR42659 from the National Institutes

of Health/National Institute of Arthritis and Musculoskeletal and

Skin Diseases. None of the authors received payments or other benefits or

a commitment or agreement to provide such benefits from a commercial

entity. Commercial entities (Zimmer and Synthes USA) paid or directed,

or agreed to pay or direct, benefits to a research fund, foundation, educational

institution, or other charitable or non-profit organization with

which one or more of the authors are affiliated or associated.”

Inclusion criteria:

* between 18 and 69 years of age

* treated with an above the ankle amputation within three months after sustaining high-energy trauma below the distal femoral level.

* One or more of the following injuries to the lower extremity:

- A Gustilo Grade IIIB or IIIC fracture

- A Gustilo Grade-IIIA fracture with certain characteristics

- A dysvascular limb (knee dislocation, closed tibial fracture with a pulseless foot, or a penetrating wound with vascular injury)

- A major soft-tissue injur to the tibia (degloving or severe crush/avulsion injury)

- Severe foot or ankle injury (open pilon or Grade-IIIB ankle fracture or a severe hindfoot or midfoot injury).

Exclusion criteria:

* moderate-to-severe brain injury (a score of <15 on the Glasgow Coma Scale at 21 days after injury or at the time of discharge)

* spinal cord deficit

* prior amputation of the leg or foot

* Third-degree burn on the injured lower limb

* transferal to participating center >24 hours after the injury

* not able to speak English or Spanish

* documented psychiatric disorder or mental retardation

* Active military duty

* Living outside hospital’s catchment area and follow-up deemed impossible.

N total at baseline:

I: 18

C-1: 109

C-2: 34

Important prognostic factors:

Mean age ± SD: 35.2 ± 13.3

Not reported per group

aetiology

Not reported per group.

20% motor vehicle

25% motorcycle

17% pedestrian-motor vehicle

17% collision.

Sex:

84% men

Not reported per group.

Groups comparable at baseline?

Unclear.

Amputation through the knee

C1: amputation below the knee

C2: amputation above the knee

Length of follow-up:

24 months, with follow-up measurements before discharge, at 3, 6, 12 and 24 months.

Loss-to-follow-up:

Incomplete outcome data:

Data available, whole group (%)

3 months: n=143 (88.8)

6 months: n=141 (77.0)

12 months: n=139 (86.3)

24 months: n=124 (77.0)

Data available at 24 months, n (%)

I: 16 (89)

C-1:81 (74)

C-2: 27 (79)

No reason for missing data provided.

Complications

Not reported.

Reinterventions/re-operations

Determined as rehospitalization for stump revision, %:

I: 11.1

C-1: 11.9

C-2: 23.5

P=0.23

The authors describe: “one patient had a revision to a higher level subsequent to the three-month evaluation” but the original level of amputation is not described.

Walking distance/speed

As measured with a 100-ft (30.48m) timed test. % of patients with self-selected walking speed of ≥ 4 ft/sec

I: 21.4

C-1: 62.0

C-2: 43.5

P=0.01

In a longitudinal regression model including selected potential confounders, a significant difference between the odds of walking at a speed ≥4ft/sec, there was a significant difference between I and C-1.

Physical functioning

As measured as the degree of independence in transfers, walking and climbing stairs, %

Transfers

I: 100

C-1: 100

C-2: 100

Walking

I: 87.5

C-1: 98.8

C-2: 96.3

P: 0.05

Stair climbing

I: 87.5

C-1: 98.8

C-2: 96.3

P=0.05

As measured with the physical function dimension of the Sickness Impact Profile (SIP), mean (SD)

I: 13.9 (9.0)

C-1: 9.8 (10.6)

C-2: 9.0 (7.2)

P=0.26

In a longitudinal regression model including selected potential confounders, a significant difference was found between I and C-1 in the percentage increase or decrease in the score on the SIP physical function.

Quality of Life, including role functioning

Not reported.

Length of hospitalization

Determined as the mean no. of days of initial hospitalisation:

I: 23.7

C-1: 16.9

C-2: 16.9

P=0.08

Mortality

Not mentioned

Part of the LEAP study, in which the difference in outcome following reconstruction or amputation were examined.

Only the percentages of patients who were re-admitted to the hospital due to complications were reported, not the number of complications during the initial hospital stay/complete study period.

A score of >10 points on the SIP represent substantial disability and a difference of 2 or 3 points reflect meaningful differences in function.

Three questions were added to the outcome measures to capture the ability to perform tasks requiring advanced functional capabilities. The results are not reported: no validated questionnaire.

The authors tried to find an explanation for the results found:

“The through the knee amputation was in the zone of injury in all but one

patient. Twelve of the eighteen patients did not have gastrocnemius

muscle coverage over the femoral condyles, as this tissue was not available. Management of the distal part of the femur varied. The condyles were contoured in ten of the patients. The patella was removed in nine of the patients, and it was fused to the distal end of the femur in one.”

And

“In the performance of an elective through-the-knee amputation, the gastrocnemius muscles are

employed to form a well-padded layer over the end of the femur. In the case of a trauma-related through-the-knee amputation, long posterior flaps containing the gastrocnemius muscles are often not available, as the amputation is being performed in or near the zone of injury. Of our eighteen patients with a through-the-knee amputation, seventeen underwent the amputation in the zone of injury and twelve had no remaining

gastrocnemius muscle. We believe that our data bring into question the wisdom of performing a through-the-knee amputation in patients who have sustained high-energy trauma, especially when the knee is included in the zone of injury.”

Tintle, 2014

Type of study: Retrospective study, consecutive sample.

Setting and country: Tertiary military Medical Center

Funding and conflicts of interest: No funding or conflicts of interest to disclose

Inclusion criteria:

* Combat-wounded personnel sustaining a major lower extremity amputation from operations in Iraqi and Enduring Freedom.

Exclusion criteria:

none

N total at baseline:

I: 22

C-1: 155

C-2: 115

In 226 patients. There were 74 bilateral lower extremity amputees, the limbs were considered separately.

Important prognostic factors²:

Not reported per group.

Groups comparable at baseline? Unclear.

Knee disarticulation

C-1: transtibial amputations

C-2: transfemoral amputations

Length of follow-up:

Mean 23 months (IQR: 16-32)

Loss-to-follow-up/incomplete outcome data:

Not reported

Complications

Defined as a complication that led to a (initial) reoperation/indication for re-operation.

I: Wound infection (n=2, 9%),

Neuroma excision (n=1, 5%), myodesis failure revision (n=1, 5%). No sterile dehiscence/wound breakdown, HO excision, scar revision and bone-bridge complications (latter not available).

C-1: Wound infection (n=51, 33%), sterile dehiscence/wound breakdown (n=6, 4%), HO excision (n=23, 15%),

Neuroma excision (n=26, 17%), scar revision (n=12, 8%), myodesis failure revision (n=6, 4%) bone-bridge complications (n=19, 12%).

C-2: Wound infection (n=23, 20, sterile dehiscence/wound breakdown (n=5, 4%), HO excision (n=47, 41%),

Neuroma excision (n=6, 5%), scar revision (n=13, 11%), myodesis failure revision (n=11, 10%) bone-bridge complications (Not available).

There was a significant decreased rate of reoperations in the knee disarticulation cohort compared with all other amputation levels (p<0.01, OR=7.6, 95%CI=2.2;21.4)

Reinterventions/re-operations

See complications

Walking distance/speed

Not reported

Physical functioning

Not reported

Quality of Life, including role functioning

Not reported

Length of hospitalization

Not reported

Mortality

Not reported.

Data were obtained from computerized inpatient and outpatient medical records.

“… potential differences between rates and proportions of occurrences were assessed with the use of X² test and the Fisher exact test, as appropriate, with the level of significance set at a 2-tailed α<0,05.”

80.4% of the length-preserving amputations were performed in theater, but the definite amputation closure was performed a median 14 days from the time of the initial injury at the research center.

The overall operation rate was 53%, 156 limbs required reoperation. There were 261 complications.

The authors state in the discussion:

“Finally, in this retrospective

study, we found a significantly lower rate of reoperation for knee disarticulations, suggesting that this much-maligned

amputation level may have benefits in terms of complications

and reoperation rates in well-selected patients.”

Lim, 2018

Type of study: retrospective analysis of prospective collected data. propensity score matching

Setting and country: USA, American College of Surgeons National Surgical Quality Improvement Program (data derived from patient’s chart)

Funding and conflicts of interest: none

Inclusion criteria:

* current procedural terminology code for AKA (27590, 27591), TKA (27598)

Exclusion criteria:

* <18 years of age

* patients with open amputation or redo amputation

* trauma victims

* musculoskeletal oncologic disease

N total at baseline:

Intervention: 251

Control: 7469 / 251 after propensity score matching)

“As TKA is more often applied under emergency circumstances with systematic infection, a 1:1 propensity score match was necessary between the AKA and TKA groups to minimize the confounding effect of baseline comorbidities.”

Baseline characteristics after propensity score matching:

Age (SD)

I: 64.7 (16.2)

C: 64.1 (15.1)

Gender

I: 61% male

C: 61.8% male

Functional status, independent, %

I: 45.8

C: 46.6

Emergency operation, %

I: 28.7

C: 28.7

Groups comparable at baseline? Yes, after propensity matching.

Through-knee amputation (TKA)

Above-knee amputation (AKA)

Length of follow-up:

30 days postoperative follow-up

Loss-to-follow-up / incomplete outcome data:

Not reported

Complications

Soft tissue infection, %

I: 11.2

C: 7.2

P=0.16

Wound disruption, %

I: 0.8

C: 1.2

P=1.0

There was also no difference between the groups concerning the percentage of patients with pneumonia, unplanned reintubation, intubated>48 hours, pulmonary embolism, renal failure, urinary tract infection, stroke, cardiac arrest and sepsis or septic shock.

Reinterventions/re-operations

Defined as ‘Return to the operating room’, %:

I: 27.9

C: 12.4

P=<0.01

“This parameter in the NSQIP database does not identify the reason for reoperation, nor does it distinguish between planned and unplanned return to the operating room. Thus, when TKA is performed as a temporizing operation before definitive staged AKA in patients with overwhelming infection, the rate of TKA reoperation misleadingly increases. Moreover, the Current Procedural Terminology code system does not specify open (guillotine) TKA versus definitive TKA, whereas open (guillotine) AKA was coded separately and was excluded from the present study.”

Walking distance/speed

Not reported

Physical functioning

Not reported

Quality of Life, including role functioning

Not reported

Length of hospitalization

Defined as the length of stay postoperatively, mean (SD)

I: 10.75 (10.5)

C: 8.64 (13.6)

P=0.052

Mortality

Defined as 30-day mortality, %

I: 11.2

C: 9.6

P=0.66

“Additional multivariate regression analyses using these five variables* plus the level of amputation confirmed that TKA had no impact on mortality (HR. 1.15; CI: 0.61-2.19: p=.667).”

* predictors of mortality.

“Propensity score matching was performed using all baseline characteristics variables to balance comorbidities between AKA and TKA. Operative variables and postoperative complications were compared between the groups using the same statistical tests with baseline characteristics. Multivariate logistic regression analyse was again used to identify risk factors for reoperation and mortality.”

The authors state:

“The study demonstrated that contemporary surgical outcomes of TKA are largely comparable with those of AKA: 30-day mortality and intraoperative and postoperative complications were similar in both groups. The rate of surgical site infection and wound disruption of both procedures also demonstrated similar outcomes.”

Met, 2008

Type of study:

Retrospective (patient case histories and electronic files), consecutive cohort.

Setting and country: St. Antonius Ziekenhuis, Nieuwegein, The Netherlands

Funding and conflicts of interest: not reported

Inclusion criteria:

* underwent either TKA or AKA because of end-stage peripheral arterial disease

Exclusion criteria:

* amputation because of osteomyelitis or trauma

N total at baseline:

Intervention: 39

Control:34

Important prognostic factors²:

For example

Age (SD):

I :71.5 (11.0)

C:71.0 (9.7)

Sex:

I: 64% M

C: 71% M

History of stroke or transient ischemic attack

I: 41%

C: 15%

Indication: progressive ischemia after previous amputation:

I: 15%

C: 44%

Indication: gangrene

I: 64%

C: 47%

Indication: rest pain

I: 21%

C: 9%

Groups comparable at baseline? No, see the differences in medical history.

Through-knee-amputation (TKA)

Above-knee-amputation (AKA)

Length of follow-up:

End-stage rehabilitation level.

Loss-to-follow-up, n(%):

I: 13 (33): missing data (n=2), dead (n=3), converted TKA (n=8)

C: 10 (29): dead (n=10).

Complications

Divided in general complications and wound related complications

Wound related, n (%)

I: wound infection 9 (23), delayed wound healing 10 (26), haemorrhage 4 (10).

C: Wound infection 6 (18), delayed wound healing 3 (9), haemorrhage 6 (18).

The differences between the groups were not significant.

General complications

I: pneumonia 1 (3), urinary tract infection 4 (10), myocardial infarction 1 (3)

C: pneumonia 5 (15), urinary tract infection 6 (18).

The differences between the groups were not significant.

Reinterventions/re-operations, n(%)

I: 8 (21), due to consisted wound infection and ischemia.

C: 0 (0)

Walking distance/speed

See physical functioning. Otherwise not reported.

Physical functioning

Assessed using the Special Interest Group in Amputee Medicine (SIGAM) mobility grades (see right column), n (%)

I: A 17 (65), B 1 (4), C 3 (12), D 4 (15), E 1 (4)

C: A 16 (67), B 1 (4), C 4 (17), D 3 (13).

In the discussion the authors mention: “No significant difference could be found in functional outcome, with most patients only reaching level A according to the SIGAM score).”

Quality of Life, including role functioning

Not reported

Length of hospitalization, median, days

I: 19

C: 23

P= Not significant

Mortality, n (%)

I: 3 (8)

C: 10 (29)

P=Not significant.

The level of amputation was merely chosen on clinical factors (normally the lowest possible level was chosen).

“Reasons for a primary AKA included iliac occlusions

classified as Trans Atlantic Inter-Society Consensus D5 without revascularization possibilities, in combination with ischemic-induced sensibility disorders or skin circulation disorders around the knee joint as determined clinically, eventually completed with transcutaneous oxygen (TcPO2) measurements, or ischemic ulceration at the level of the skin that would be required for skin flaps to perform a TKA.”

The Special Interest Group in Amputee Medicine (SIGAM) mobility grades were used to evaluate functional outcome. The levels A-F indicate:

“A: Limb wearing abandoned or use of cosmetic limb only; B: Therapeutic wearer wears prosthesis only for transfers, to assist nursing, walking with the physical aid of another, or during therapy; C: Walks on level ground only, ≤50 m with or without use of walking aids; D: Walks outdoors on level ground only and in good weather ≥50 m, with or without use of walking aids; E: Walks ≥50 m independent of walking aids except occasionally for confidence or to improve confidence in adverse terrain or weather; F: Normal or near normal gait)”.

Statistical analyses: X2 test and fisher exact tests.

It is questionable if the result of the statistical test for mortality is correct.

Study reference

(first author, year of publication)

Bias due to a non-representative or ill-defined sample of patients?

(unlikely/likely/unclear)

Bias due to insufficiently long, or incomplete follow-up, or differences in follow-up between treatment groups?

(unlikely/likely/unclear)

Bias due to ill-defined or inadequately measured outcome?

(unlikely/likely/unclear)

Bias due to inadequate adjustment for all important prognostic factors?

(unlikely/likely/unclear)

MacKenzie, 2004

Unclear, demographic data as well as information concerning the recruitment of patients is not provided (per group).

Unclear. Loss to follow-up is higher in the control groups. No reasons provided

Unlikely for ‘hard’ outcomes, likely for ‘soft’ outcomes.

The number of ´stump revisions´ might be underestimated, as only stump revisions performed during a rehospitalization were included.

Likely - Not performed for the outcome measures reinterventions and most of the physical functioning outcomes

Unclear – longitudinal regression results provided for the SIP and walking speed, however, all variables with p<0.2 included. Furthermore, the reference group was the below-the knee amputation.

Tintle, 2014

Unclear, demographic data is not provided per group.

Unclear, follow-up time not mentioned per group.

Unlikely

Likely - no adjustment

Lim, 2018

Unclear/unlikely: propensity matching, but demographic information was limited due to the administrative data.

Unlikely (for the outcomes reported)

Unlikely for most outcomes, except for the number reoperations (see evidence table)

Unclear-unlikely – propensity matching performed, but limited patient data available.

Met, 2008

Likely the, medical history of the patients in the groups differed.

Likely for follow-up data, loss-to follow-up is quite significant. Plus, the reasons for loss-to-follow-up differ between the groups.

Likely for ‘soft’ outcomes (no blinding). Statistical analyses for mortality questionable.

Likely – no adjustment performed.