Search results
A total of 503 studies were initially retrieved from the five databases. 234 repeated studies were excluded. A further 128 studies were excluded based on information in the title and abstract. Of the remaining 141 studies, 21 met the inclusion and exclusion criteria, and four were excluded because of data problems. Finally, 17 studies were selected for the analysis. The flow chart for study inclusion and exclusion is shown in Fig. 1.
Study characteristics
There was a total of 936 patients (506 in the LPF group and 430 in the HAP group) enrolled in the included studies. The characteristics of the final 17 included studies are presented in Table 1.
Table 1: Characteristics of included studies
Methodological quality and publication bias
Of the 17 included studies, 16 were retrospective studies and one was a RCT. The methodological quality and risk of bias for each study were evaluated as described in Methods (Fig. 2; Table 2). MINORS scores greater than 12 were included in the study [27], and the quality index scores of the non-RCTs were 13–18 points (13 for one study, 16 for eleven studies, 17 for one study, and 18 for three studies), indicating a moderate risk of bias. A funnel plot for method-related complications (Fig. 3) showed no obvious bias. However, the analysis did include a majority of Chinese studies and so publication bias cannot be excluded.
Table 2
MINORS appraisal scores for the included non-RCTs studies.
Study | Methodological items | Total |
A | B | C | D | E | F | G | H | I | J | K | L | |
Dietrich 2008 Solberg 2009 Wang 2009 Wild 2011 Spross 2011 Xu 2012 You 2013 Zhao 2015 Celik 2016 Huang 2016 Shen 2017 Xue 2017 Wei 2018 Hao 2018 Qu 2020 Ye 2020 | 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 | 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 | 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 | 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 | 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 | 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 | 0 0 0 0 2 1 0 0 0 0 2 0 2 0 0 0 | 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 | 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 | 0 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 | 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 | 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 | 13 16 16 16 18 17 16 16 16 16 18 16 18 16 16 16 |
MINORS Methodological index for nonrandomized studies. |
A: A clearly stated aim; B: inclusion of consecutive patients; C: prospective collection of data; D: endpoints appropriate to the aim of the study; E: unbiased assessment of the study endpoint; F: follow-up period appropriate to the aim of the study; G: loss to follow-up that is < 5%; H: prospective calculation of the study size; I: an adequate control group; J: contemporary groups; K: baseline equivalence of groups; L: adequate statistical analyses. The items were scored as “0” (not reported), “1” (reported but inadequate), or “2” (reported and adequate)
Results of the meta-analysis
Efficacy (Constant–Murley score)
The Constant–Murley score was recorded in the 10 retrospective studies [8–12,15−17,20,24], and included 263 cases in the LPF group and 235 cases in the HAP group. There was significant heterogeneity (P < 0.00001, I2 = 84%) and so a random effect model was used. In this analysis, the Constant-Murley score was significantly greater in the LPF treated group than in the HAP treated group [SMD = 0.73, 95%CI: (0.23, 1.22), Fig. 4] suggesting that LPF resulted in a better clinical outcome in treatment of PHFs. However, in a subgroup analysis there was no significant difference [SMD = 0.54, 95%CI: (-0.45, 1.52), Fig. 5] in patients more than 60 years old. A subgroup analysis on the fracture type showed a significantly greater Constant–Murley score with LPF compared with HAP in 3-part fractures [SMD = 1.52, 95%CI (1.07, 1.98)] but not in 4-part fractures [SMD = 0.35, 95%CI (-0.07, 0.77), Fig. 6].
Method-related complications
Method-related complications were described in 15 studies [8–9,11−23], which included 457 cases in the LPF group and 397 cases in the HAP group. There was no significant heterogeneity (P = 0.39, I2 = 6%) and a fixed-effect model was used for analysis. There was significant difference between the LPF and HAP groups in method-related complications [OR = 1.80, 95%CI (1.24, 2.61), Fig. 7]. We got the same results [OR = 2.12, 95%CI (1.18, 3.79), Fig. 8] when we separately analyzed patients older than 60 years old based on age range.
Revision
Revision surgery was described in 11 studies [8–9,11−17,19,21], which included 329 cases in the LPF group and 274 cases in the HAP group. There was significant heterogeneity (P = 0.61, I2 = 0%) and a random effect model was used for analysis. There was a significantly higher rate of reoperation after LPF than after HAP [OR = 3.61, 95%CI (1.99, 6.56), Fig. 9] A significantly higher rate was also seen for patients more than 60 years old [OR = 5.66, 95%CI (1.97, 16.24), Fig. 10].