This meta-analysis was performed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement (24). All analyses were based on previously published studies; therefore, ethical approval and informed consent were waived for this study.

We performed a comprehensive search of the PubMed, Embase, the Web of Science, Cochrane Library, China National Knowledge Infrastructure, and Wanfang electronic databases from inception to April 2021. Based on relevant studies (25, 26), we used the following search terms: gynecological cancer OR gynecological carcinoma OR gynecological neoplasm OR endometrial neoplasm OR endometrial carcinoma OR endometrial cancer OR endometrium cancer OR endometrium carcinoma OR cervical cancer OR cervical carcinoma OR ovarian carcinoma OR ovarian cancer OR uterine neoplasm and C-reactive protein OR albumin OT CAR OR CRP/Alb OR C-reactive protein to albumin ratio, C-reactive protein/albumin ratio and prognosis OR prognostic OR survival OR outcome OR mortality. Both free text and medical subject heading terms were used for the literature search. We selected articles in English and Chinese. References from the identified publications were also retrieved for potential inclusion.

Following were the inclusion criteria: (1) histopathologically diagnosed GC; (2) studies that investigated the prognostic role of CRP/Alb ratio for survival outcomes including but not limited to OS, cancer-specific survival (CSS), disease-free survival (DFS), PFS, and recurrence-free survival; (3) availability of hazard ratios (HRs) and 95% confidence intervals (CIs) or sufficient data to calculate these values; and (4) availability of cutoff values of the CRP/Alb ratio. Following were the exclusion criteria: (1) reviews, meeting abstracts, case reports, and letters; (2) unavailability of data to calculate HRs and corresponding 95% CIs; (3) studies in languages other than English or Chinese; and (4) overlapping or duplication of studies.

Two investigators (YF and TZ) independently screened all retrieved articles and extracted information using a predetermined form. All disagreements were resolved through discussion with a third investigator (CZ). The following data were recorded: first author’s name, year of publication, country, age, cancer type, the International Federation of Gynecology and Obstetrics (FIGO) stage, treatment, study period, cutoff value of the CRP/Alb ratio, method used to determine the cutoff value, duration of follow-up or the last date of follow-up, survival endpoints, and HRs and 95% CIs for OS, PFS, and DFS. HRs and 95% CIs were extracted from multivariable analyses depending on availability; HRs and 95% CIs were extracted from univariate analyses in the remaining cases. Two independent reviewers assessed the methodological quality of the included studies based on the Newcastle–Ottawa Quality Assessment Scale (NOS) (http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp). The NOS consists of the following three dimensions: selection (four stars), comparability (two stars), and outcome assessment (three points). NOS scores range from 0 to 9, and scores ≥6 indicate high-quality studies.

Combined HRs and 95% CIs were used to estimate the prognostic effect of the CRP/Alb ratio in GC. HR >1 without a 95% CI overlapping 1 indicates that a high CRP/Alb ratio predicts poor prognosis. Heterogeneity among studies was assessed using the Cochrane Q test and the I² statistic. p < 0.10 and/or I² > 50% was interpreted as indicative of significant heterogeneity, and a random-effects model was used in such cases; a fixed-effects model was used in the remaining cases. Subgroup analysis stratified by various factors was performed to determine the source of heterogeneity. Pooled odds ratios (ORs) and 95% CIs were used to determine the association between the CRP/Alb ratio and clinicopathological features. Publication bias was evaluated using Begg’s funnel plots. All statistical analyses were performed using the Stata software, version 12.0 (Stata Corporation, College Station, TX, USA). A p < 0.05 was considered statistically significant.

The initial literature search identified 292 studies from the aforementioned databases; 126 studies remained after exclusion of duplicate studies ( Figure 1 ). After screening titles and abstracts, 112 studies were eliminated, and we reviewed the full text in 14 articles, of which 7 studies with insufficient data were excluded. Finally, seven studies with 1,847 patients (17–23) were included in this meta-analysis.

Table 1 summarizes the characteristics of studies included in this meta-analysis. The studies were published between 2017 and 2021. Five studies were performed in China (17, 19, 20, 22, 23) and two in Japan (18, 21). Six studies were published in English (17–19, 21–23) and one in Chinese (20). Four studies enrolled patients with CC (17, 20–22) and three recruited patients with OC (18, 19, 23). The sample size ranged from 200 to 407 (median, 235). The cutoff values of the CRP/Alb ratio ranged from 0.022 to 0.68 (median, 0.15). Therefore, we used 0.15 as the cutoff value of the CRP/Alb ratio for subgroup analysis. Five studies used receiver operating characteristic curve analysis to determine the cutoff value (17–20, 22), one study used the median value (21), and one study used cutoff finder software (23). Six studies (17, 19–23) reported the prognostic value of the CRP/Alb ratio for OS, two studies (22, 23) reported its prognostic value for PFS, and one study (18) for DFS. Five studies (17–19, 21, 23) enrolled patients with FIGO stages I–IV, and two studies (20, 22) included patients with FIGO stages I–II. NOS scores of the included studies ranged from 6 to 9 (median, 7), which indicates that all included studies were of high quality. Table 2 shows NOS score details.

A total of six studies that included 1,539 patients (17, 19–23) investigated the association between the CRP/Alb ratio and OS in GC. Pooled data showed HR of 1.84, 95% CI of 1.41–2.40, and p < 0.001 ( Figure 2 ; Table 3 ). Owing to significant heterogeneity (I² = 71.2%, p = 0.004), we used a random-effects model. Subgroup analyses based on various factors showed that a high CRP/Alb ratio remained a prognostic tool to determine poor survival in subgroups across different countries, cancer types, FIGO stage, cutoff values, cutoff value determination methods, and treatment modalities ( Table 3 ). Three studies that included 780 patients (18, 22, 23) reported the role of the CRP/Alb ratio for prognosis of PFS/DFS (HR, 2.58; 95% CI, 1.42–4.68; p = 0.002) based on a random-effects model ( Figure 3 ; Table 3 ). Similar to the findings associated with OS, subgroup analysis revealed that an elevated CRP/Alb ratio was an indicator of poor PFS/DFS in various subgroups of patients with GC.

Using data from five studies, we investigated the association between the CRP/Alb ratio and clinicopathological characteristics (17–20, 22). A high CRP/Alb ratio was significantly associated with FIGO stages III–IV (OR, 2.98; 95% CI, 1.45–6.14; p = 0.003) ( Figure 4 ; Table 4 ). However, we observed a non-significant correlation between the CRP/Alb ratio and lymph node metastasis (yes vs. no) (OR, 2.54; 95% CI, 0.59–10.90; p = 0.209), tumor size (≥4 vs. <4 cm) (OR, 2.54; 95% CI, 0.84–7.72; p = 0.100), and histopathological grade (G3 vs. G1/G2) (OR, 1.07; 95% CI, 0.75–1.53; p = 0.176) ( Figure 4 ; Table 4 ).

Begg’s funnel plots were used to assess the potential publication bias for OS and PFS/DFS. We observed no significant publication bias for OS (p = 0.135) or PFS/DFS (p = 0.296) in this meta-analysis ( Figure 5 ). Therefore, the results of our meta-analysis are reliable.