Prognostic value of the prognostic nutritional index in colorectal cancer: a systematic review and meta-analysis

Abstract

Background

This meta-analysis evaluates the prognostic significance of the Prognostic Nutritional Index (PNI) in colorectal cancer (CRC) patients, focusing on overall survival (OS), disease-free survival (DFS), and progression-free survival (PFS).

Methods

We conducted a comprehensive literature search across PubMed, Embase, Web of Science, and CNKI. Observational studies reporting hazard ratios (HRs) with 95% confidence intervals (CIs) for survival outcomes based on PNI were included. Pooled HRs were calculated using random-effects models. Heterogeneity, sensitivity, and publication bias were evaluated, and subgroup analyses were performed by region, follow-up duration, and tumor stage.

Results

A total of 43 studies comprising 19,214 CRC patients were included the meta-analysis. 36 studies with 18,231 patients reported the prognostic value of PNI on the OS of CRC, and the pooled HR was 1.89 (95% CI: 1.70–2.10, P < 0.001). This association remained robust across sensitivity analyses, suggesting PNI as a reliable biomarker for risk stratification. Moderate heterogeneity (I² = 32.9%) was observed, which subgroup analyses attributed to study region, follow-up duration, and inclusion criteria for CRC stages. Non-Asian cohorts, studies with shorter follow-up or partial staging and high cut-off value of PNI exhibited reduced heterogeneity. Eleven studies with 5,181 patients reported the prognostic value for DFS, and the pooled HR was 1.31 (95% CI: 0.84–2.03). Nine studies with 2,856 patients were for PFS, and the pooled HR was 1.15 (95% CI: 0.78–1.72), neither reaching statistical significance. Significant heterogeneity was noted for both DFS and PFS across the studies.

Conclusions

This meta-analysis demonstrates that a low PNI is a robust predictor of poor overall survival in colorectal cancer, particularly in Asian populations and across diverse disease stages. While its prognostic value for DFS and PFS remains uncertain.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12876-026-04702-y.

Introduction

Colorectal cancer (CRC) continues to be one of the leading causes of cancer-related deaths globally, highlighting the critical need for reliable prognostic markers that can inform therapeutic decisions and enhance patient outcomes. Traditionally, prognosis in CRC is determined by tumor stage and conventional survival predictors; however, it is increasingly evident that similarly staged patients often experience diverse outcomes. This variability underscores the necessity to identify additional biomarkers that reflect both tumor biology and host response mechanisms [1]. Among these emerging markers, the Prognostic Nutritional Index (PNI), which evaluates the immune-nutritional status of patients, has attracted considerable attention due to its potential in predicting postoperative complications and long-term survival across various cancers [2]. Several studies have demonstrated that PNI correlates with permanent stoma rates following anterior resection, complications after surgery for locally recurrent rectal cancers, and postoperative complications in CRC [3–5]. However, its association with complete pathological responses and overall survival (OS) in advanced cancers following neoadjuvant chemoradiation remains uncertain [6, 7]. This inconsistency calls for further investigation to clarify the precise impact of PNI on OS in rectal cancer patients.

Recent evidence from multivariate Cox regression analyses indicates that PNI serves as an independent predictor of OS, showing a 1.2% increase in three-year survival for every unit increase in PNI, regardless of preoperative therapy or surgical approach [8]. These findings emphasize the clinical significance of PNI, even though its absolute value may appear modest, suggesting its potential as a modifiable factor through preoperative nutritional rehabilitation or immune-enriched diets [9]. Despite these promising results, several limitations must be acknowledged. The retrospective nature of most studies introduces inherent biases, and the lack of consideration for critical covariates such as extramural vascular invasion and tumor regression grade could influence outcome predictions [10]. Additionally, the generalizability of these findings to different demographic groups requires validation.

Given the accumulating evidence supporting the prognostic value of PNI in CRC, there is a compelling rationale for conducting a comprehensive meta-analysis to evaluate the relationship between PNI and key oncological endpoints, including OS, disease-free survival (DFS), and progression-free survival (PFS). Such an analysis would provide a more definitive understanding of PNI’s role in predicting outcomes in CRC, potentially guiding personalized treatment strategies and improving patient care. By systematically reviewing and synthesizing available literature on the association between PNI and CRC outcomes, focusing specifically on OS, DFS, and PFS, this meta-analysis aims to establish the magnitude of PNI’s effect on these endpoints. Consequently, our findings will inform clinical practice and future research directions, ultimately contributing to enhanced management and improved outcomes for CRC patients.

Methods

Study design and search strategy

This meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines to ensure methodological rigor, transparency, and reproducibility [11]. A comprehensive literature search was performed across four major electronic databases: PubMed, the Cochrane Library, Web of Science, and the Chinese National Knowledge Infrastructure (CNKI), to identify relevant studies published up to January 31, 2025. The search strategy combined the following key terms: (“Prognostic Nutritional Index” OR PNI) AND (“colorectal cancer” OR CRC) AND (prognosis OR survival). The details of search strategy were list in Supplement file 1. No language restrictions were applied, although only studies involving human participants were included. In addition to database searches, we used the “Related Articles” feature in PubMed and manually screened the reference lists of eligible studies and relevant reviews to identify any potentially missed publications.

Inclusion and exclusion criteria

Studies were eligible for inclusion if they met the following criteria: (1) investigated the prognostic value of PNI in CRC patients; (2) reported sufficient data to estimate hazard ratios (HRs) with corresponding 95% confidence intervals (CIs) for overall survival (OS), disease-free survival (DFS), or progression -free survival (PFS), preferably derived from multivariable Cox regression analyses; and (3) were cohort studies (prospective or retrospective). We only included studies that had already obtained ethical approval from their respective institutional review board.

Data extraction and quality assessment

Two reviewers independently extracted data using a standardized form. Disagreements were resolved through discussion or by consulting a third reviewer when necessary. The following information was extracted from each study: first author, publication year, country, study design, sample size, patient demographics, CRC stage distribution, PNI calculation method, cut-off value, follow-up duration, and adjusted HRs with 95% CIs for OS, DFS, and PFS. The methodological quality of included cohort studies was assessed using the Newcastle-Ottawa Scale (NOS) [12], which evaluates studies based on three domains: selection of study groups (up to 4 points), comparability of groups (up to 2 points), and assessment of outcomes (p to 3 points). A maximum score of 9 indicates high quality, and studies scoring ≥ 7 was considered to be of high methodological quality. Scoring below 7 on the NOS scale were considered low quality and excluded from the analysis.

Statistical analysis

All statistical analyses were performed using the meta package in R software (version 4.3.0). Pooled HRs and their 95% CIs were calculated to evaluate the association between PNI (typically categorized as high vs. low) and survival outcomes (OS, DFS, PFS). An HR < 1 indicates a favorable prognosis associated with higher PNI. Heterogeneity across studies was assessed using the I² statistic, with values of 25%, 50%, and 75% representing low, moderate, and high heterogeneity, respectively. A fixed-effects model (Mantel-Haenszel method) was applied in the presence of low to moderate heterogeneity (I² < 50%), while a random-effects model (DerSimonian and Laird method) was used when heterogeneity was substantial (I² ≥ 50%). Sensitivity analysis was conducted by sequentially removing individual studies to assess the stability of the pooled results. Subgroup analyses were performed based on key variables such as geographic region (Asia vs. non-Asia), median follow-up duration (≥ 36 vs. <36 months), and tumor stage (all stages vs. stage-specific populations), to explore potential sources of heterogeneity. Publication bias was evaluated using Egger’s linear regression test. Funnel plots were visually inspected to supplement these tests. A p-value < 0.05 in either test was considered indicative of statistically significant publication bias.

Certainty of the evidence

The overall certainty of evidence for meta-analysis outcome was assessed using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) framework. Evidence was rated by considering risk of bias, inconsistency, indirectness, imprecision, and publication bias, and was classified into one of four categories: high, moderate, low, or very low.

Results

Search results

The study selection process is summarized in Fig. 1 according to the PRISMA guidelines. An initial database search yielded 219 records. After removing 36 duplicates, 183 records were screened by title and abstract. Of these, 110 were excluded due to irrelevance, specifically, animal studies (n = 12), reviews (n = 57), conference abstracts (n = 21), or unrelated topics (n = 20). The remaining 73 full-text articles were assessed for eligibility. Among them, 29 were excluded: 15 lacked survival data, 11 did not report hazard ratios (HRs) or sufficient data for extraction, and 3 were deemed low quality based on the Newcastle-Ottawa Scale. Ultimately, 43 studies [3–6, 13–52] encompassing 19,214 colorectal cancer patients, were included in the meta-analysis.

Fig. 1.

The flow diagram of identifying eligible studies

Study characteristics and quality assessment

All 44 included studies employed a retrospective cohort design. The majority were conducted in Asian countries, particularly China, Japan, and South Korea. The cutoff values for defining high versus low PNI varied widely across studies, ranging from 38.25 to 50. Follow-up durations ranged from 3 months to 10 years, with median follow-up typically exceeding 36 months. While some studies included patients across all tumor stages (I–IV), others focused on specific subgroups. The methodological quality of the included studies was assessed using the NOS criteria. As shown in Table S1, the majority of studies scored 7 or higher (range: 7–9), indicating low risk of bias and high overall quality. Detailed characteristics of the included studies are presented in Table 1.

Table 1.

Main characteristics of the included studies

Study	Year	Country	Cutoff	Number	Age	Sex	Stage	Follow up (month)
Nozoe	2012	Japan	40	249	69.8	156/93	I-IV	60
Mohri	2013	Japan	45	365	65	223/142	I-IV	72
Shibutani	2015	Japan	47.9	218	69	120/98	I-IV	NA
Tokunaga	2015	Japan	46.94	556	68	330/226	I-IV	31.8
Chen	2016	China	50.2	1318	57.5	774/544	I-IV	60
Cao	2017	China	44.55	228	69	134/94	I-III	60
Noh	2017	Korea	50	3569	61.1	2169/1400	I-IV	54
Peng	2017	China	49.22	274	58	156/118	III	46
Sasaki	2018	Japan	46.7	149	64	93/56	I-IV	59
Luvián-Morales	2019	Mexico	47	3301	58.7	1727/1574	I-IV	12
Maruyama	2020	Japan	44.8	197	71	113/84	I-IV	60
Sato	2020	Japan	35	72	71	41/31	I-IV	24
Tominaga	2020	Japan	49.8	896	65	404/492	I-III	57
Ucar	2020	Turkey	46	308	57.5	192/116	I-IV	21.8
Wang	2020	China	45.85	281	55.9	185/96	I-IV	12
Bardakci	2021	Turkey	38	126	61	75/51	I-III	60
Lin	2021	China	43.7	208	63	122/86	I-III	36
Tamai	2021	Japan	45.6	227	68	124/103	I-IV	52
Wang1	2021	China	45	211	61	140/71	II-III	45
Xu	2021	China	45	413	57.6	260/153	II-III	33.6
Zhu	2021	China	45.61	196	62	118/78	I-II	NA
Kazi	2022	India	46.7	340	48.6	235/105	I-III	36
Mizutani	2022	Japan	50.7	32	68	9月23日	I-IV	NA
Nagashima	2022	Japan	40	83	69	42/41	I-III	3
Qiu	2022	China	46.95	33	52.5	20/13	I-IV	60
Wei	2022	China	51.94	150	63	81/69	I-III	60
Xie	2022	China	43.4	1114	57.3	639/475	I-IV	60
Xu1	2022	China	49.72	142	65	87/55	I-III	39
Erdogan	2023	Turkey	45.7	42	57	20/22	IV	8.3
Kim	2023	Korea	50.9	1112	64	667/445	I-III	60
Kocak	2023	Turkey	39.01	199	65	132/67	I-IV	36
Li	2023	China	48.65	511	61	320/191	I-IV	64
Ni	2023	China	50.45	102	60	62/40	I-III	60
Portale	2023	Italy	36.5	182	69	127/55	I-IV	60
Qiao	2023	China	46.2	105	61	59/46	I-III	60
Taniai	2023	Japan	43	192	68	128/64	I-IV	120
Tuncel	2023	Turkey	55.25	138	62	83/55	III	40
Li1	2024	China	47.82	470	61	279/191	I-III	60
Niu	2024	China	46	67	61	41/26	II-III	36
Nguyen	2024	Vietnam	40.36	87	58.97	50/37	IV	24
Silva	2024	Brazil	41	298	65	172/126	I-III	60
Tatsuta	2024	Japan	40	475	68	266/209	II-III	60
Wang2	2024	China	45.8	106	57.93	70/36	I-IV	36
Yang	2024	China	47.58	90	64	58/32	II-III	36
Zhang	2024	China	2.25	414	66.7	191/223	I-IV	35.9
Zhang1	2024	China	48	120	63.9	72/48	I-III	60
Lee	2025	Korea	48.05	664	65	402/262	I-III	60
Zhu2	2025	China	44.48	201	66	119/82	I-IV	36

NA Not available

Pooled analysis of prognostic value of PNI on the OS of CRC patients

A total of 36 studies [3–5, 13–22, 24–27, 29, 31–38, 40, 41, 44–47, 49–52] (n = 18,231 patients) reported adjusted HRs for OS stratified by PNI status. The pooled analysis demonstrated that a low PNI was significantly associated with poorer OS (random-effects model: HR = 1.88, 95% CI: 1.69–2.09, p < 0.001; Fig. 2), indicating that patients with lower PNI had an 89% higher risk of death compared to those with higher PNI. Moderate heterogeneity was observed across these studies (I² = 32.8%, p = 0.032). Sensitivity analysis, conducted by sequentially omitting each study, showed that no single study significantly influenced the overall effect estimate, confirming the robustness of the results (Figure S1). To explore potential sources of heterogeneity, subgroup analyses were performed based on geographic region (Asian vs. non-Asian), follow-up duration (> 36 vs. ≤36 months), tumor stage inclusion (all stages vs. stage-limited) and cut-off of PNI value (> 45 vs. ≤45 months). As shown in Table 2, heterogeneity was reduced in subgroups of non-Asian populations, shorter follow-up periods, studies including only specific stages and high PNI value, suggesting that these factors may contribute to inter-study variability. Notably, the association between low PNI and worse OS remained statistically significant in all major subgroups. No evidence of publication bias was detected based on Egger’s test (p = 0.306) among the studies, and the funnel plot was visually symmetrical (Figure S2).

Fig. 2.

Forest plot of the prognostic value of PNI on the overall survival of CRC patients. HR: hazard ratio; CL: confidence interval; PNI: Prognostic Nutritional Index; OS: overall survival; CRC: colorectal cancer

Table 2.

Subgroup analysis for the prognostic value of PNI on the OS of CRC patients

Categories	No. study	HR (95%CI)	P value	I2	P heterogeneity
Country
Asian country	29	1.94 (1.71–2.19)	0.038	36.7%	0.019
Non-Asian	7	1.67 (1.43–1.94)	< 0.001	18.4%	0.289
Follow-up duration
> 36 months	19	2.02 (1.70–2.40)	< 0.001	52.5%	0.003
≤ 36 months	14	1.73 (1.53–1.95)	< 0.001	0%	0.733
Tumor stage
All stage	19	1.85 (1.59–2.16)	< 0.001	48.9%	0.007
Part stage	17	1.91 (1.68–2.18)	< 0.001	0%	0.603
Cut-off
> 45	15	1.76 (1.59–1.95)	< 0.001	21.2%	0.187
≤ 45	18	1.91 (1.61–2.27)	< 0.001	47.7%	0.022

Pooled analysis of prognostic value of PNI on the DFS of CRC patients

Eleven studies [4, 5, 21, 28–30, 32, 34, 40, 43, 48] (n = 5,181 patients) provided data on DFS. The pooled HR for DFS was 1.31 (95% CI: 0.84–2.03, p = 0.229) using a random-effects model (Fig. 3), indicating no statistically significant association between PNI and DFS. However, substantial heterogeneity was present (I² = 89.6%, p < 0.001). Sensitivity analysis confirmed the stability of the pooled estimate, as exclusion of any single study did not alter the overall conclusion (Figure S3). Subgroup analyses by country, follow-up duration, and tumor stage and cut-off of PNI value (> 45 vs. ≤45 months) are detailed in Table 3. Heterogeneity was lower in non-Asian studies and those with shorter follow-up, suggesting these factors may underlie the observed variability. However, the lack of statistical significance persisted across subgroups. Egger’s test revealed no significant publication bias (p = 0.352; Figure S4).

Fig. 3.

Forest plot of prognostic value of PNI on the DFS of CRC patients. HR: hazard ratio; CL: confidence interval. DFS: disease-free survival

Table 3.

Subgroup analysis for the prognostic value of PNI on the DFS of CRC patients

Categories	No. study	HR (95%CI)	P value	I2	P heterogeneity
Country
Asian country	9	1.21 (0.71–2.05)	0.142	89.8%	< 0.001
Non-Asian	2	1.88 (1.36–2.58)	< 0.001	0%	0.840
Follow-up duration
> 36 months	7	1.46 (1.01–2.11)	0.002	91.0%	< 0.001
≤ 36 months	4	0.99 (0.32–3.07)	0.238	88.9%	< 0.001
Tumor stage
All stage	3	1.86 (1.37–2.53)	0.003	36.6%	0.207
Part stage	8	1.07(0.64–1.79)	0.229	90.5%	< 0.001
Cut-off
> 45	7	1.15 (0.62–2.11)	0.229	86.4%	< 0.001
≤ 45	4	1.63 (0.85–3.13)	0.436	92.2%	< 0.001

Pooled analysis of prognostic value of PNI on the PFS of CRC patients

Nine studies [4, 5, 21, 25, 29, 30, 32, 34, 40] (n = 2,856 patients) reported HRs for PFS. The combined analysis showed a non-significant trend toward worse PFS in patients with low PNI (HR = 1.15, 95% CI: 0.78–1.72, p = 0.483; Fig. 4), with high heterogeneity (I² = 84.0%, p < 0.001). Sensitivity analysis indicated that the overall result was not driven by any single study (Figure S5). Subgroup analyses (Table 4) suggested that geographic region and follow-up duration contributed to heterogeneity, but tumor stage and cut-off value did not. Despite reduced heterogeneity in certain subgroups, the association remained non-significant. Publication bias was not detected across the studies (Egger’s test: p = 0.435; Figure S6).

Fig. 4.

Forest plot of prognostic value of PNI on the PFS of CRC patients. HR: hazard ratio; CL: confidence interval; PFS: progression-free survival

Table 4.

Subgroup analysis for the prognostic value of PNI on the PFS of CRC patients

Categories	No. study	HR (95%CI)	P value	I2	P heterogeneity
Country
China	7	1.80 (0.31–3.78)	0.256	85.7%	< 0.001
Non-China	2	1.23 (0.71–2.13)	0.374	88.2%	< 0.001
Follow-up duration
> 36 months	5	1.33 (0.71–2.50)	0.309	91.0%	< 0.001
≤ 36 months	4	1.19 (0.75–1.89)	0.469	88.9%	< 0.001
Cut-off
> 45	5	1.50 (0.71–3.20)	0.367	85.5%	< 0.001
≤ 45	4	0.97 (0.53–1.75)	0.439	89.4%	< 0.001

Certainty of evidence

All the included studies were observed design, therefore were rated as having a serious risk of bias. Substantial heterogeneity (I² > 50%) was present for the outcomes of DFS and PFS, leading to a rating of serious inconsistency for these measures. All outcomes were based on direct evidence, so indirectness was not a concern. Imprecision was rated as serious for DFS and PFS due to the small total sample sizes and wide confidence intervals. Publication bias was also judged to be unserious based on the Egger’s test results and funnel plot. Therefore, the certainty of evidence for OS outcome was rated as low, while DFS and PFS were rated as very low (Table S2).

Discussion

This meta-analysis evaluates the prognostic role of the PNI in CRC patients, synthesizing data from 44 studies involving 18,552 individuals. The pooled HR for OS was 1.89 (95% CI: 1.70–2.10,), indicating that low PNI is strongly associated with poorer OS. This association remained consistent across sensitivity analyses, confirming its robustness. Moderate heterogeneity was observed, which subgroup analyses attributed to geographic region, follow-up duration, and inclusion of all CRC stages versus selective stages. Non-Asian cohorts and studies with shorter follow-up or partial staging showed reduced heterogeneity, suggesting population-specific variations. In contrast, PNI’s prognostic utility for DFS and PFS was inconclusive. For DFS, the pooled HR was 1.31 (95% CI: 0.84–2.03), and for PFS, it was 1.15 (95% CI: 0.78–1.7), neither reaching statistical significance. Significant heterogeneity was noted, but subgroup analyses did not identify a single influential study. Geographic origin and follow-up duration were suggested as sources of heterogeneity. These findings confirm that PNI is a reliable predictor of OS in CRC, particularly in Asian populations and across diverse disease stages. Its simplicity, using serum albumin and lymphocyte count, makes it a practical tool for risk stratification. However, the lack of significant associations with DFS and PFS suggests that while PNI reflects host immunity and nutritional status critical for long-term survival, it may be less sensitive to early recurrence. Future research should explore how nutritional and immune interventions might modulate PNI to improve survival outcomes in CRC patients.

The present meta-analysis underscores the significant prognostic value of the PNI in CRC, particularly for OS. This finding aligns with previous studies demonstrating the importance of systemic inflammation and nutritional status as critical determinants of cancer prognosis [53]. The consistency observed across sensitivity analyses reinforces the robustness of this association, suggesting that PNI could serve as a reliable biomarker for risk stratification in CRC patients. However, the heterogeneity observed necessitates further exploration. Subgroup analyses revealed that geographic region, follow-up duration, and inclusion criteria for CRC stages contributed to this variability. Non-Asian cohorts and studies with shorter follow-up or partial staging exhibited reduced heterogeneity, implying potential differences in patient characteristics, treatment protocols, or nutritional profiles across populations [54]. These findings are consistent with earlier reports highlighting regional variations in PNI thresholds and their implications on clinical outcomes.

In contrast, the prognostic utility of PNI for DFS and PFS was less conclusive. Significant heterogeneity was noted for both endpoints. Despite the lack of significant associations, subgroup analyses suggested that geographic origin and follow-up duration might influence these results [55]. This discrepancy may reflect differences in tumor biology, adjuvant treatment efficacy, or surveillance practices affecting recurrence patterns more than overall mortality [56]. In addition, the non-significant pooled HRs for DFS and PFS may reflect insufficient statistical power due to the smaller sample sizes in these subgroups, rather than a true lack of association. Furthermore, the substantial heterogeneity observed in DFS/PFS analyses may be attributed to variations in adjuvant therapy regimens, tumor molecular subtypes (microsatellite instability), or differences in postoperative surveillance protocols, which were not uniformly reported in the included studies.

The absence of significant associations for DFS and PFS suggests that while PNI captures aspects of host immunity and nutritional status crucial for long-term survival, it may be less sensitive to early disease recurrence. This observation is supported by studies showing that PNI primarily reflects systemic resilience rather than early tumor behavior [57]. Additionally, the dynamic nature of PNI during treatment, such as chemotherapy, may play a role in its predictive capacity for different endpoints [58].

To be noted, while the pooled analysis did not demonstrate a statistically significant association between PNI and DFS, PFS, the point estimate (DFS: 1.31, PFS: 1.15) did not rule out a potential clinically meaningful effect. Given the moderate statistical power and substantial heterogeneity among included studies, this finding should be regarded as hypothesis-generating rather than conclusive. Future large-scale, prospective studies with standardized PNI measurement are needed to clarify this relationship.

Several limitations must be acknowledged. First, although we retrospectively registered the protocol, the lack of prospective registration may introduce reporting bias. Second, despite performing subgroup analyses based on PNI cut-offs, the lack of standardized thresholds (ranging from 38.25 to 50) limits the clinical applicability. Future studies should utilize population-specific or ROC-derived cut-offs to harmonize risk stratification. Third, all included studies were retrospective, introducing risks of selection bias and unmeasured confounding (sch as socioeconomic status, detailed treatment variations, and comorbidities). Second, important pathological factors, including extramural vascular invasion, tumor regression grade, and molecular subtypes (e.g., MSI status), were rarely adjusted for, potentially affecting outcome interpretation. Third, PNI cutoff values varied widely across studies, reflecting the lack of standardized thresholds and limiting direct comparability. Although most studies used ROC-derived cutoffs, population-specific differences in baseline nutrition and immunity may influence optimal values. Despite these limitations, our analysis highlights the clinical utility of PNI as a simple, cost-effective, and readily available biomarker. Derived from serum albumin and absolute lymphocyte count—both routinely measured in clinical practice—PNI offers a practical tool for risk stratification at diagnosis or preoperatively. Its strong association with OS supports its integration into prognostic models, particularly in resource-limited settings where complex biomarkers are inaccessible. Importantly, the modifiable nature of PNI opens avenues for therapeutic intervention. Nutritional support, immunonutrition, and exercise programs may improve PNI and potentially enhance treatment tolerance and survival outcomes. Future research should explore whether targeted interventions to optimize PNI before and during treatment can translate into measurable clinical benefits.

Conclusions

In conclusion, this meta-analysis provides compelling evidence that PNI is a reliable predictor of OS in CRC, particularly in Asian populations and across various disease stages, while its prognostic utility for DFS and PFS remains uncertain and needs large sample size of study to validate. Future prospective studies are needed to standardize PNI thresholds, evaluate its dynamic changes during treatment, and investigate whether interventions aimed at improving PNI can lead to improved oncological outcomes.

Authors’ contributions

Study concept and design: XYB, and ZF; Collection and assembly of data: ZF, HYS and HXX; Performed the experiment: XYB, HXX, HYS, NSF and ZF; Data analysis and interpretation: ZF and XYB; Manuscript writing and review: All authors. All authors have read and approved the manuscript in its current state.

Funding

This study was supported by the Science and Technology Projects of Qinzhou (No. 201612).

Data availability

All data generated or analyzed during this study are included in this article.

Declarations

Ethics approval and consent to participate

This study was approval by the Ethics Committee of the Tenth Affiliated Hospital of Guangxi Medical University (No. KY-2025071). All the procedures were carried out in accordance with institutional guidelines. Since this is a meta-analysis based on aggregate data from published literature, it was determined that no additional informed consent or ethical review of individual patient records was required.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.