Table 3.
Evidence Profile: Comparative Effectiveness and Safety of Allograft Versus Autograft in ACLR
| Certainty Assessment |
No. Analyzed Patients∗,† |
Effect | Certainty | Importance | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| No. of Studies | Study Design | LoE | Risk of Bias | Inconsistency | Indirectness | Imprecision | Other Considerations | Allograft | Autograft | |||
| Effectiveness | ||||||||||||
| Patient-reported function, activity level, and symptoms (follow-up: mean ≥5 years; assessed with: Lysholm score) | ||||||||||||
| 5 19,21, 22, 23, 24 |
RCT | II | Serious‡ | Serious§ | Not serious‖ | Not serious | None | 303 | 292 | MD 0.39 (95% CI –1.22 to 2.00; P = .63; I2 = 51%) | ⨁⨁⊕⊕◯◯ Low |
Critical |
| Patient-reported function, activity level and symptoms (follow-up: mean ≥5 years; assessed with: Tegner score) | ||||||||||||
| 5 19,20,22, 23, 24 | RCT | II | Serious¶ | Not serious# | Not serious‖ | Not serious | None | 299 | 287 | MD –0.26 (95% CI –0.53 to 0.01; P = 0.06 I2 = 0%) | ⨁⨁⨁⊕⊕⊕◯ Moderate |
Critical |
| Patient-reported function, activity level and symptoms (follow-up: mean ≥5 years; assessed with: Cincinnati Knee score) | ||||||||||||
| 3 19,23,24 | RCT | II | Serious∗∗ | Not serious†† | Not serious‖ | Not serious | None | 218 | 207 | MD 0.19 (95% CI –2.02 to 2.39; P = 0.87; I2 = 9%) | ⨁⨁⨁⊕⊕⊕◯ Moderate |
Critical |
| Patient-reported function, activity level and symptoms (follow-up: mean 10.5 years; assessed with: SANE score) | ||||||||||||
| 120 | RCT | II | Serious‡‡ | Not serious | Serious§§ | Not serious | None | 49 | 48 | The study did not find a statistically significant difference: The postoperative mean score was 2.7 points lower in the allograft group when compared to the autograft group. Postoperative mean SANE score: 78.8 ± 18.8 vs 81.5 ± 16.4 |
⨁⨁⊕⊕◯◯ Low |
Critical |
| Patient-reported function, activity level and symptoms (follow-up: mean ≥5 years; assessed with: subjective IKDC score) | ||||||||||||
| 6 19, 20, 21, 22, 23, 24 |
RCT | II | Serious‡‡ | Not serious | Not serious‖ | Not serious | None | 352 | 340 | MD –2.25 (95% CI –3.02 to –1.47; P < .00001; I2 = 0%). | ⨁⨁⨁⊕⊕⊕◯ Moderate |
Critical |
| Clinical knee stability (follow-up ≥5 years; assessed with: Lachman test) | ||||||||||||
| 4 19,22, 23, 24 |
RCT | II | Very serious‖‖ | Serious¶¶ | Not serious | Not serious | None | 250 | 239 | s.s. difference in Lachman scores (grade 0-1) in 1 study19: 31/43 (72%) vs 37/40 (93%). No statistically significant differences in Lachman scores in 3 studies22, 23, 24 Lachman test (grade 0-1; ranges across studies): 31/43 (72%) to 74/80 (92.5%) vs 84/91 (92.3%) to 30/32 (93.8%) |
⨁⊕◯◯◯ very low |
Important |
| Clinical knee stability (follow-up: mean ≥5 years; assessed with: pivot shift test) | ||||||||||||
| 4 19,22, 23, 24 |
RCT | II | Very serious‖‖ | Serious¶¶ | Not serious | Not serious | None | 250 | 239 | s.s. difference in Pivot shift test (Grade 0-1) in 1 study19: 38/43 (88.4%) vs 40/40 (100%) No statistically significant differences in pivot shift test in 3 studies22, 23, 24 Pivot shift (grade 0-1; ranges across studies): 38/43 (88.4%) to 95/95 (100%) vs 32/32 (100%) to 91/91 (100%) |
⨁⊕◯◯◯ Very low |
Important |
| Clinical knee stability (follow-up: mean ≥5 years; assessed with: KT arthrometer; better indicated by lower values) | ||||||||||||
| 4 19,22, 23, 24 |
RCT | II | Serious‡‡ | Serious¶¶ | Not serious | Not serious | None | 250 | 239 | 2/4 studies19,22 found a statistically significant difference in instrumented knee laxity favoring autografts, while the other 2/4 studies23,24 did not find any statistically significant difference in side-to-side differences measured with the KT arthrometer between treatment groups. Mean side-to-side differences (in mm; ranges across studies): 2.5 ± 0.9 to 5.5 ± 1 vs 2.1 ± 1.6 to 2.5 ± 0.7 |
⨁⨁⊕⊕◯◯ Low |
Important |
| Clinical knee stability (follow-up: mean ≥5 years; assessed with: objective IKDC score) | ||||||||||||
| 4 19,22, 23, 24 |
RCT | II | Serious‡‡ | Not serious | Not serious | Not serious | None | 250 | 239 | None of the studies found a statistically significant difference in the objective IKDC score between treatment groups. Objective IKDC score (normal or nearly normal scores; ranges across studies): 38/43 (88.4%) to 75/80 (93.8%) vs 29/32 patients (90.6%) to 38/40 (95%) |
⨁⨁⨁⊕⊕⊕◯ Moderate |
Important |
| Patient satisfaction (assessed with: NR) | ||||||||||||
| 1 21 | RCT | II | Very serious## | Not serious | Not serious | Not serious | None | 53 | 53 | Patient satisfaction was analyzed in 106 patients from 1 study. The study found no statistically significant difference between patients undergoing allograft ACLR (n = 53) or autograft ACLR (n = 53). The instrument used to measure patient satisfaction was not reported. Satisfied: 46/53 (86.8%) vs 47/53 (88.7%) Nearly satisfied: 7 (13.2%) vs 5 (9.4%) Diff. n.s.; P > .05 |
⨁⨁⊕⊕◯◯ Low |
Important |
| Safety | ||||||||||||
| Graft failure (follow-up: mean ≥5 years) | ||||||||||||
| 2 19,20 |
RCT | II | Not serious∗∗∗ | Not serious | Serious‖,§§ | Not serious | None | 92 | 88 | Bottoni et al.20: 13/49 (26.5%) vs 4/48 (8.3%), diff. s.s. with P < .05 Tian et al.19: 13/43 (30.2%) vs 3/40 (7.5%), diff. s.s. with P < .001††† |
⨁⨁⨁⊕⊕⊕◯ Moderate |
Critical |
| Revisions (follow-up: mean ≥5 years) | ||||||||||||
| 220,22 | RCT | II | Not serious∗∗∗ | Serious‡‡‡ | Serious‖,§§ | Not serious | None | 81 | 80 | Bottoni et al.20: 13/49 (26.5%) vs 4/48 (8.3%), diff. s.s. with P < .05 Li et al.22: no patient needed additional surgery because of recurrent or residual symptoms (0/32 vs 0/32) |
⨁⨁⊕⊕◯◯ Low |
Critical |
| Complications (follow-up: mean ≥5 years) | ||||||||||||
| 6 19, 20, 21, 22, 23, 24 |
RCT | II | Serious§§§ | Not serious | Not serious‖ | Serious‖‖‖ | None | 352 | 340 | Overall complication rate: NR Arthrofibrosis (reported in 2/6 studies19,24; 269 patients): 0/138 (0%) vs 0/131 (0%) Effusion (0/6 studies): NR Tenderness (reported in 1 study24; 186 patients): 0/95 (0%) vs 2/91 (2.1%) Infections (reported in 4 studies19,22, 23, 24; 489 patients): 5/250 (2%) vs 0/239 (0%), range: 0-4.6% vs 0% Hypoesthesia (reported in 2 studies19,24; 269 patients): 0/138 (0%) vs 6/131 (4.6%), range: 0% vs 3.3-7.5%. Synovitis was not reported in any of the included studies. Deep venous thrombosis (reported in 3 studies19,23, 24; 425 patients): 2/218 (0.9%) vs 1/207 (0.5%), range: 0%-2.5% vs 0%-1.3% Further reported complications: Postoperative mean fever time in days (reported in 1 study23; 156 patients): 6.8 vs 4.4 diff. s.s., with P < .05) Arthritic progression (reported in 1 study19; 83 patients): 14/43 (32.6%) vs 4/40 (10%), diff. s.s. with P < .05. Tibial and femoral tunnel widening in millimeters (reported in 2 studies; 203 patients): Jia et al.21: Tibial (in mm), mean ± SD: 7.8 ± 0.4 vs 7.61 ± 0.22, diff. s.s. with P < .05 Femoral (in mm), mean ± SD: 7.64 ± 0.35 vs 7.51 ± 0.42, diff. s.s. with P < .05 Bottoni et al.20: Tibial (in mm), mean (range): 9.2 (7-10) vs 8.9 (7-10); diff. n.s. with P = .651 Femoral (in mm), mean (range): 8.8 (7-10) vs 8.3 (7-10), diff n. s. with P = .453 Furthermore, some of the included studies specifically stated that there were no cases of pain when kneeling, anterior knee pain, etc. |
⨁⨁⊕⊕◯◯ Low |
Critical |
NOTE. Source: Goetz and de Villiers12; results of crucial outcomes (for which at least 3 studies reported on) are depicted quantitatively using MDs between allograft groups and autograft groups across studies. Results of the remaining outcomes for which no meta-analysis was conducted are depicted qualitatively. Furthermore, the LoE of individual studies was further added based on a guidance document.25
ACLR, anterior cruciate ligament reconstruction; CI, confidence interval; Diff., difference; IKDC, International Knee Documentation Committee; LoE, Level of Evidence, MD, mean difference; n.s., not statistically significant; RCT, randomized controlled trial; SANE, Single Assessment Numerical Evaluation; s.s., statistically significant.
Grading of Recommendations Assessment, Development and Evaluation (GRADE) Working Group grades of evidence17:
High certainty: We are very confident that the true effect lies close to that of the estimate of the effect.
Moderate certainty: We are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different.
Low certainty: Our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect.
Very low certainty: We have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.
The reader is reminded that in Bottoni et al. the number of patients actually refers to the number of knees.
Excluding patients with hybrid grafts.
In 4/5 studies, the risk of bias for blinding the outcome assessors was judged to be high. Therefore, we judged that this may have seriously affected the certainty.
Heterogeneity: I2 = 51%.
Differences in interventions were present across studies (e.g., irradiated vs nonirradiated grafts, single-bundle vs double-bundle, etc.).
In 5/5 studies, the risk of bias for blinding the outcome assessors was judged to be high. Therefore, we judged that this may have seriously affected the certainty.
Heterogeneity: I2 = 0%.
In 3/3 studies, the lack of blinding significantly increases the risk of bias. Therefore, we judged that this may have seriously affected the certainty.
None of the studies showed any statistically significant differences in Cincinnati Knee scores between treatment groups. The nonstatistical findings showed slightly greater scores in allograft patients in 2/3 studies and lower scores in allograft patients in 1/3 study/studies when compared with the autograft groups, respectively. The difference of the mean scores ranged from 1 to –3.
The lack of blinding in the study/studies may seriously affect the certainty to believe in the evidence of this outcome measure.
The overall applicability for the broad population selected in these assessment results may suffer due to the fact that numerous different graft types were used and that some studies used a subpopulation of the population of interest. Bottoni et al., for instance, only included highly active military (mostly) men, and Tian et al. used irradiated allografts. It is unclear in how far the generalizability suffers due to the aforementioned factors.
It was judged that the lack of blinding may have very seriously affected the certainty to believe this specific outcome.
Heterogeneity was suspected within the included studies. It appears that the studies do not consistently show any difference/difference favoring a treatment group.
There were 2 substantial factors that increased the risk of bias: lack of blinding and selective outcome reporting; the latter was present insofar as it was insufficiently described how patient satisfaction was measured. In addition, no scores were reported, but it was stated that no statistically significant differences between treatment groups was found.
Lack of blinding for outcome assessors was judged to be less likely to affect this outcome.
Graft failure, however, was defined differently in the studies. Tian et al. defined it as knee laxity >5 mm measured with a KT-2000, and Bottoni et al. did not clearly mention how graft failure was defined.
Bottoni et al. found a considerably large difference in the revision rate, whereas Li et al. stated that no additional surgeries were needed in either of the treatment groups.
The risk of bias for selective outcome reporting was judged high in 2/6 studies, and unclear in the remaining 4/6 studies. Most of the studies, however, did not report on an overall complication rate. Instead, they were presented narratively in the studies.
The optimal information size may have not been reached for most of the specific complications.