Search results
As shown in Fig.1, among 1787 articles were obtained from the databases via the search strategy. After removing duplicates, 767 articles were screened. From among them 722 articles were removed after reading the title and abstract based on the inclusion criteria. Then, 19 studies were excluded on the basis of exclusion criteria (the reasons were offered in the supplementary material). Finally, 11 RCTs [2,13-22]and 15 non-RCTs [23-37] were included in this study after meeting the inclusion and exclusion criteria.
Baseline information and quality assessment
26 studies involving 2369 patients in cementless group and 2654 patients in cemented group were divided into three subgroups according to the length of follow-up. The length of follow-up ≤5.5 years including 17 studies [14-18,20,21,23-25,27,29,31,32,35-37], which ranged from 2 to 5.3 years. The length of follow-up ≤10.5 years including 6 studies [13,19,26,28,30,33], which ranged from 5.7 to 10.2 years. The length of follow-up >10.5 years including 3 studies [2,22,34], which ranged from 12.1 to 16.6 years. The baseline information including study design, demographical data, revision rate, prosthesis and length of follow-up were clearly listed in Table 1. Especially, the duration of follow-up was same for cemented and cementless replacements in studies included.
The JADAD score of 11 RCTs were listed in Table 2, both of them were ≧5, four of them [2,15,21,22] got 7 points. The MINORS scores of 15 non-RCTs were listed in Table 3, both of them were ≧20, only one of them [37] got 24 points.
Rate of revision and reasons of revision
A total of 22 studies involving 2178 patients in cementless group and 2442 patients in cemented group reported the rate of revision during the follow-up. There were 101(4.6%) and 131 (5.3%) patients in two groups underwent the revision surgery for all kinds of reasons. Among of them, aseptic loosing was the most common reason of revision, next was the periprosthetic joint infection (PJI). The specific number of revisions and all kinds of reasons were listed in Table 4 in detail. In addition, we made a heat-map based on the risk ratio of revision caused by all kinds of reasons in every study (Fig.2). We divided 22 studies into three subgroups according to the length of follow-up and reported results as following:
First subgroup: follow-up≤5.5 years
Among 14 studies involving 1251 patients in cementless group and 1450 patients in cemented group followed≤5.5 years reported the rate of revision between cementless and cemented fixation in TKA. There were 32 (2.6%) and 72 (5%) patients in two groups revised, patients in cemented group significantly increased the rate of revision compared with cementless group (RR=0.47, 95% CI: [0.32, 0.70], P=0.0002; Fig.3). It was important to note that two non-RCTs showed high rates of revision in the cemented group, the others didn't. Although the I2 was 35% with low heterogeneity, we also looked forward to more RCTs could be included in this topic in future research to prove the presence of difference.
8 of 14 studies involving 878 patients in cementlesss group and 1092 patients in cemented group reported the rate of revision caused by aseptic loosing. There were 8 (0.6%) and 35 (2.4%) patients in two groups revised for aseptic loosing, patients in cemented group significantly increased the rate of revision caused by aseptic loosing compared with cementless group (RR=0.28, 95% CI: [0.14, 0.54], P=0.0002; Fig.4).
11 of 14 studies involving 1080 patients in cementlesss group and 1270 patients in cemented group reported the rate of revision caused by periprosthetic joint infection (PJI). There were 11 (0.9%) and 11 (0.8%) patients in two groups revised for PJI, and no significant difference between two groups (RR=1.07, 95% CI: [0.51, 2.22], P=0.86; Fig.5).
Second subgroup: follow-up≤10.5 years
Among 5 studies involving 697 in cementless group and 762 in cemented group followed≤10.5 years reported the rate of revision between cementless and cemented fixation in TKA. There were 62 (8.9%) and 49 (6.4%) patients in two groups revised, which was not significantly different between two groups (RR=1.22, 95% CI: [0.85, 1.77], P=0.28; Fig.3).
All 5 studies involving 697 patients in cementlesss group and 762 patients in cemented group reported the rate of revision caused by aseptic loosing. There were 38 (5.5%) and 22 (2.9%) patients in two groups revised for aseptic loosing, no significant difference between groups (RR=1.61, 95% CI: [0.96, 2.72], P=0.07; Fig.4).
All 5 studies involving 697 patients in cementlesss group and 762 patients in cemented group reported the rate of revision caused by periprosthetic joint infection (PJI). There were9 (1.3%) and 16 (2.1%) patients in two groups revised for PJI, and no significant difference between two groups (RR=0.59, 95% CI: [0.26, 1.34], P=0.20; Fig.5).
Third subgroup: follow-up>10.5 years
Among 3 studies involving 230 patients in cementless group and 230 patients in cemented group followed>10.5 years reported the rate of revision between cementless and cemented fixation in TKA. There were 7 (3%) and 10 (4.3%) patients in two groups revised, which was not significantly different between two groups (RR=0.73, 95% CI: [0.30, 1.76 ], P=0.48; Fig.3).
All 3 studies involving 230 patients in cementlesss group and 230 patients in cemented group reported the rate of revision caused by aseptic loosing. There were 4(1.7%) and 6 (2.6%) patients in two groups revised for aseptic loosing, no significant difference between two groups (RR=0.71, 95% CI: [0.23, 2.22], P=0.56; Fig.4).
Only 1 study involving 100 patients in cementlesss group and 100 patients in cemented group reported the rate of revision caused by periprosthetic joint infection (PJI). There were 0 (0%) and 1 (1%) patients in two groups revised for PJI.
Functional recovery
12 studies involving 652 patients in cementlesss group and 656 patients in cemented group recorded the Knee Society knee score. They were divided into two subgroups according to if the duration of follow-up was >5.5 years. Pooled results revealed there was no significant difference regarding Knee Society knee score between two groups in all subgroups (Fig.6A).
9 studies involving 827 patients in cementlesss group and 819 patients in cemented group recorded the Knee Society function score. There was no significant difference in short duration subgroup. While patients in cementless group had a better Knee Society function score when followed longer than 8.5 years (MD=1.77, 95% CI: [0.36, 3.17], P=0.01; Fig.6B).
5 studies involving 230 patients in cementlesss group and 221 patients in cemented group recorded the Oxford knee score, all of them followed shorter than 5.5 years. Pooled results revealed that fixation method didn’t make a difference on Oxford knee score in short duration (Fig.6C).
7 studies involving 626 patients in cementlesss group and 679 patients in cemented group reported the range of motion (ROM) following TKA. Pooled results revealed the ROM was not significantly different even when the patients followed longer (Fig.7A). However, 7 studies involving 566 patients in cementlesss group and 588 patients in cemented group revealed that the rate of manipulation under anesthesia was significantly more in cemented group when the duration longer than 5.5 years (RR=0.41, 95% CI: [0.19, 0.87], P=0.02; Fig.7B).
Complications
5 studies involving 369 patients in cementlesss group and 390 patients in cemented group recorded the rate of deep vein thrombosis (DVT) following TKA. All of duration of follow-up were shorter than 5.5 years. There were 17 (4.6%) in cementless group and 27 (6.9%) in cemented group diagnosed as DVT, while it was not significantly different (Fig.8A).
20 studies involving 2048 patients in cementlesss group and 2337 patients in cemented group recorded the rate of all infection following TKA. There were 29 (1.4%) in cementless group and 32 (1.4%) in cemented group diagnosed as superficial wound infection or PJI. In all subgroups at three different time periods, the results were not significantly different (Fig.8B).