Table 5. Summary of results of individual studies eligible for meta-analysis.
| Outcome (Analysis) | Study ID | Event /Total | Univariate analysis: Unadjusted HR (95% CI) | Multivariate analysis: Adjusted HR (95% CI) | Covariates adjusted for* |
|---|---|---|---|---|---|
| RFS (1) | Amankwah, 2013 [31] | 28/65 (43%) | (Cut-off = median; log-rank P<0.0001) KM plot favouring high miR-21 Estimated HR (95% CI)1: = 4.83 (2.26–10.35), P=0.00005 Inverse2: = 0.21 (0.10–0.44), P=0.00005 |
1.99 (0.70–5.64), P=0.20 Inverse2: = 0.50 (0.18–1.42), P=0.20 |
Age GS cT |
| Leite, 2015 [60] | 50/127 (39%) | (Cut-off = median; log-rank P=0.003) KM plot favoring low miR-21 Estimated HR (95% CI)1: = 2.32 (1.33–4.03), P=0.003 |
2.505 (1.356–4.629), P=0.003 |
GG PSA pT |
|
| Li, 2012 [61] | 116/168 (69%) | (Cut-off = median; log-rank P<0.001) KM plot favoring low miR-21 Estimated HR (95% CI)1: = 1.91 (1.33–2.75), P=0.0005 |
2.059 (1.075–3.944), P=0.029 | Age Capsular invasion GS pN PSA pT Surgical margin |
|
| Melbø-Jørgensen, 2014 [68] | 170/478 (36%) | (Cut-off = 4th quartile; log-rank P=0.006) KM plot favoring low miR-21 Estimated HR (95% CI)1: = 1.65 (1.15–2.36), P=0.006 |
1.4 (1.0–1.9), P=0.089 | Apical PSM GG Non-apical PSM Perineural infiltration PSA pT Vascular infiltration |
|
| RFS (2) | Zedan, 2017 [85] | 19/49 (39%) | (Continuous miR-21) 1.231 (0.697–2.177), P=0.474 |
(No multivariate analysis data) | (N/A) |
| Zhao, 2019a [89] | 98/206 (48%) | (Continuous miR-21) 1.12 (1.01–1.24), P=0.049 |
(Continuous miR-21) 1.35 (0.86–2.12), P=0.188 |
15 other miRNAs of interest | |
| OS (3) | Lin, 2014 [64] | 55/97 (57%) | (High vs. low miR-21, cut-off = median) 2.3 (1.3–3.9), log-rank P=0.004 |
(No multivariate analysis data) | (N/A) |
| Lin, 2017 [65] | 53/87 (61%) | (High vs. low miR-21, cut-off = median) 1.2204 (0.7028–2.1192), P=0.477 |
(Continuous miR-21) 1.1488 (0.8849–1.4916), P=0.303 |
Alkaline phosphatase Hemoglobin PSA |
|
| Sharova, 2021 [78] | 13/31 (42%) | (Cut-off = 2.69; log-rank p = 0.0067) KM plot favoring high miR-21 5.2 (1.7–15.7), P=0.0191 Inverse2: = 0.192 (0.064–0.588), P=0.0191 |
5.8 (1.0–33.1), P=0.049 Inverse2: = 0.172 (0.03–1.0), P=0.049 |
Hemoglobin Time to CRPC |
|
| Yang, 2016 [84] | 42/92 (46%) | (Cut-off not stated; log-rank P<0.05) KM plot favoring low miR-21 Estimated HR (95% CI)1: = 2.02 (1.09–3.73), P=0.025 |
3.567 (1.287–9.882), P=0.014 | Age BCR Bone metastasis cT GS PSA |
|
| PFS (4) | Guan, 2016 [42] | 47/85 (55%) | (Cut-off = mean; log-rank P=0.006) KM plot favoring low miR-21 2.381 (1.250–4.537), P=0.008 |
1.985 (1.032–3.817), P=0.040 | cT |
| Sharova, 2021 [78] | 26/31 (84%) | (Cut-off = 2.69; log-rank P= =0.0002) KM plot favoring high miR-21 7.4 (2.6–21.2), P=0.0021 Inverse2: = 0.135 (0.047–0.385), P=0.0021 |
4.8 (1.3–17.8), P=0.019 Inverse2: = 0.208 (0.056–0.769), P=0.019 |
Hemoglobin Time to CRPC |
Abbreviations: BCR, biochemical recurrence; CI, confidence interval; CRPC, castration-resistant prostate cancer; cT, clinical tumor stage; GG, Gleason grade; GS, Gleason score; HR, hazard ratio; KM, Kaplan–Meier; N/A, not applicable; OS, overall survival; PFS, progression-free survival; pN, lymph node metastasis; PSA, prostate-specific antigen; PSM, positive surgical margins; pT, pathological tumor stage; RFS, recurrence-free survival.
*
GS/GG and pT/cT were predefined as important prognostic factors that should be adjusted for in multivariate analysis.
1Unadjusted HR (95% CI) was not reported; hence it was estimated using an Excel calculator [94].