The literature search and include studies
A flowchart of the literature search was shown in Figure 1. Initially, in the primary search from the major databases, a total of 741 studies were included. After removing duplicates and screening titles and abstracts, a total of 154 papers remained, but 138 of them did not meet our purpose. The remaining 16 articles were assessed for eligibility based on full-text review, 11 were deemed ineligible. After qualitative and quantitative analysis, according to the inclusion criteria, only 5 studies published from 2016 to 2019 were selected for our meta-analysis [18, 21-24].
Basic characteristics of the included studies were listed in Table 1. A total of 4343 patients were included. These studies were all observation researchers and one conducted in Netherlands [18], one conducted in Turkey [21], three conducted in China [22-24]. The mean age of the patients ranges from 60.77 to 65.12 years old. Two studies in this meta-analysis enrolled STEMI patients [21, 24], two studies enrolled NSTEMI patients [22, 23], and the remaining one study enrolled ACS patients [18]. Two of studies explicitly stated that the patients underwent PCI [21, 22], while others did not specify if enrolled patients underwent PCI [18, 23, 24]. Two studies reported the mortality [18, 21], and three studies reported MACE [22-24]. All the studies have reported adjusted HR values. Adjusted confounding factors of each study were shown in Table 2. According to the Newcastle-Ottawa scale (NOS) [20], all the studies were of high quality and had scores of seven or more.
LMR and mortality/MACE
The short-term was defined as within 30 days after admission to hospital. The combined analysis of 2 studies covering 1281 patients described the relationship between LMR and short-term mortality/MACE [21, 23]. The result showed that LMR predicted short-term mortality/MACE (HR = 3.44, 95% CI: 1.46–8.14, P < 0.05, Figure 2A), with low levels of heterogeneity among studies (I2 = 0%). The combined analysis of 5 studies covering 4343 patients described the relationship between LMR and long-term mortality/MACE [18, 21-24]. The pooled outcome for low LMR value compared with high LMR value group was found to be 1.70 (95% CI: 1.36– 2.13, P < 0.05, Figure 2B), with moderate levels of heterogeneity among studies (I2 = 46.8%).
Subgroup analysis
There were moderate levels of heterogeneity (I2 = 46.8%) in the analysis of LMR predicting long-term mortality/MACE. We performed subgroup analysis according to mean age (≥62 and <62), LMR cut-off value (≥2 and <2), sample size (≥1000 and <1000) and diseases of patients (ACS, STEMI and NSTEMI). Low LMR predicted long-term mortality/MACE showed a statistical significance in any subgroup. Based on the change of I2, the sources of heterogeneity might be mean age of enrolled patients and defined cut-off value (Table 3). In the subgroup of older (≥62) ACS patients, I2 increased to 61.8%. In the subgroup of higher (≥2) LMR cut-off value, I2 increased to 64.7%.