Geriatric nutritional risk index as a predictor of complications and long-term outcomes in patients with gastrointestinal malignancy: a systematic review and meta-analysis
老年营养风险指数作为胃肠道恶性肿瘤患者并发症和长期预后的预测因子：一项系统综述和荟萃分析

Abstract 摘要

Background 背景

Gastrointestinal (GI) malignancy is a form of malignancy that occurs in the gastrointestinal tract and its accessory organs, accounting for nearly 30% of cancer incidence and 32% of cancer deaths worldwide [1, 2]. Although significant progress has been made in the individualized treatment of cancer patients, the clinical efficacy of most patients with GI malignancy is still poor. A number of biomarkers have been reported as prognostic factors for GI malignancy. However, these biomarkers are concentrated mainly in histochemistry and molecular biology techniques. Their application is limited to a certain extent because of the high price and need for specific experimental equipment [3, 4]. Therefore, it continues to be important to find convenient, effective, and inexpensive prognostic markers for patients with GI malignancy.
胃肠道（GI）恶性肿瘤是发生在胃肠道及其附属器官的一种恶性肿瘤，占全球癌症发病率的近30%和癌症死亡的32%[ 1，2]。尽管癌症患者的个体化治疗取得了显著进展，但大多数胃肠道恶性肿瘤患者的临床疗效仍然较差。许多生物标志物已被报道为GI恶性肿瘤的预后因素。然而，这些生物标志物主要集中在组织化学和分子生物学技术。由于价格昂贵且需要特定的实验设备，它们的应用在一定程度上受到限制[ 3，4]。因此，为胃肠道恶性肿瘤患者寻找方便、有效且廉价的预后标志物仍然很重要。

Cancer patients, especially those with GI malignancy, are prone to malnutrition. The prevalence of malnutrition in cancer patients ranges from 20 to 70%, while in GI malignancy patients, the prevalence is as high as 80% [5–7]. Many studies have shown that malnutrition in cancer patients may increase postoperative complications and prolong hospitalization, leading to poor treatment outcomes and increased mortality [8, 9]. Our previous studies have also demonstrated that malnutrition is associated not only with increased postoperative complications but also with poor long-term outcomes [10, 11]. In recent years, a comprehensive malnutrition index that integrates height, weight, and serum albumin levels, geriatric nutritional risk index (GNRI), has been reported to be associated with the prognosis of multiple GI malignancies, including colorectal cancer (CRC) [12], gastric cancer (GC) [13], and esophageal squamous cell carcinoma (ESCC) [14]. A previous meta-analysis proved that the GNRI could be used as a predictor of mortality [15]. Two meta-analyses have demonstrated the prognostic role of the GNRI in cancer patients, but both have certain limitations. Lv et al. [16] focused on the relationship between the GNRI and long-term prognosis of cancer patients but did not report the relationship between the GNRI and postoperative complications. Lidoriki et al. [17] comprehensively analyzed the role of the GNRI in the short-term and long-term outcomes of cancer patients but did not independently report the association between GNRI and GI malignancy. In addition, a number of other studies on the prognostic role of the GNRI in GI malignancy have been issued in the past year. Considering that there is no systematic study to determine the relationship between the GNRI and GI malignancy, the aim of this study was to systematically evaluate the value of the GNRI in postoperative complications and long-term outcomes in GI malignancy based on existing evidence.
癌症患者，特别是那些患有胃肠道恶性肿瘤的患者，容易出现营养不良。癌症患者营养不良的患病率为20 - 70%，而GI恶性肿瘤患者的患病率高达80% [ 5- 7]。许多研究表明，癌症患者的营养不良可能会增加术后并发症，延长住院时间，导致治疗效果不佳，死亡率增加[ 8，9]。我们之前的研究也表明，营养不良不仅与术后并发症增加有关，而且与不良的长期结局有关[ 10，11]。近年来，据报道，综合身高、体重和血清白蛋白水平的综合营养不良指数，即老年营养风险指数（GNRI）与多种GI恶性肿瘤的预后相关，包括结直肠癌（CRC）[ 12]、胃癌（GC）[ 13]和食管鳞状细胞癌（ESCC）[ 14]。 先前的荟萃分析证明GNRI可用作死亡率的预测因子[ 15]。两项荟萃分析已经证明了GNRI在癌症患者中的预后作用，但两者都有一定的局限性。Lv等[ 16]关注GNRI与癌症患者长期预后的关系，但未报道GNRI与术后并发症的关系。Lidoriki等人[ 17]全面分析了GNRI在癌症患者短期和长期结局中的作用，但没有独立报告GNRI与GI恶性肿瘤之间的相关性。此外，在过去的一年中，已经发表了一些关于GNRI在GI恶性肿瘤中的预后作用的其他研究。 考虑到目前尚无系统研究确定GNRI与GI恶性肿瘤之间的关系，本研究的目的是基于现有证据系统评价GNRI在GI恶性肿瘤术后并发症和长期结局中的价值。

Methods 方法

Results 结果

Literature search 文献检索

Figure 1 shows the literature screening process of this study. A total of 268 studies were retrieved from the databases according to the search strategy. There were 116 duplicate studies that were removed, leaving 152 studies for further screening. After carefully reading the titles and abstracts of the studies, three reviews, 24 conference abstracts, and 105 studies not about GNRI with GI malignancy were deleted. Then, the full texts of the 20 included studies were evaluated. Two of these studies were from the same center and the same period; therefore, the one with incomplete data was excluded. Three studies were unable to extract HR due to the GNRI group ≥ 3. Seven of the studies lacked relevant outcomes. Therefore, our meta-analysis included nine studies involving 2,153 patients [10, 12, 14, 19–24].
图1显示了本研究的文献筛选过程。根据检索策略，从数据库中共检索到268项研究。删除了116项重复研究，留下152项研究进行进一步筛选。在仔细阅读研究的标题和摘要后，删除了3篇评论、24篇会议摘要和105项与GNRI伴GI恶性肿瘤无关的研究。然后，对纳入的20篇研究进行全文评价。其中两项研究来自同一中心和同一时期;因此，排除了数据不完整的研究。由于GNRI组≥ 3，3项研究无法提取HR。其中7项研究缺乏相关结果。因此，我们的荟萃分析纳入了9项研究，涉及2，153例患者[ 10，12，14，19- 24]。

Open in a separate window
在单独的窗口中打开

Fig. 1  图1

Flowchart of included studies
纳入研究的流程图

GNRI and overall survival
GNRI和总生存期

A total of eight studies, with 1,805 patients, reported a relationship between the GNRI and OS. The combined forest plot showed that a low GNRI was associated with poor OS in patients with GI malignancy (HR = 1.94, 95% CI 1.65–2.28, p < 0.001) (Fig. 2). Since there was no significant heterogeneity between the studies (I² = 0.0%, P_Q = 0.851), a fixed-effects model was used. We also conducted a subgroup meta-analysis by correcting for the influence of publishing time, country, sample, cutoff value, cancer site, and primary therapy. The results indicated that the GNRI was an independent prognostic factor affecting the OS of patients in all subgroups. The results are presented in Table ​Table2.2. In addition, we performed a sensitivity analysis by removing each enrolled study each time. The results showed that ignoring any of the enrolled studies did not significantly change the effect of the GNRI on the combined meta-analysis for OS. In other words, our results were robust and reliable (Fig. 3). In the meta-analysis for OS, the Begg’s test (p = 0.035) and Egger’s test (p = 0.012) determined significant publication bias (Fig. 4a, b). We further used the trim-and-fill method to correct for potential publication bias. The results revealed that three imputed studies were added to form a symmetric funnel plot (Fig. 4c), and the corrected HR was still significant (HR = 1.831, 95% CI 1.577–2.126, p < 0.001), indicating that our results were reliable.
共有8项研究（1，805例患者）报告了GNRI与OS之间的关系。联合森林图显示，低GNRI与GI恶性肿瘤患者的OS较差相关（HR = 1.94，95%CI 1.65-2.28，p < 0.001）（图2）。由于研究之间无显著异质性（I ² = 0.0%，P _Q = 0.851），因此使用固定效应模型。我们还进行了亚组荟萃分析，校正了发表时间、国家、样本、截止值、癌症部位和主要治疗的影响。结果表明，GNRI是影响各亚组患者OS的独立预后因素。结果见表2.2。此外，我们通过每次删除每项入组研究进行了敏感性分析。结果表明，忽略任何入组研究均不会显著改变GNRI对OS合并荟萃分析的影响。 换句话说，我们的结果是稳健和可靠的（图3）。在OS的荟萃分析中，Begg检验（p = 0.035）和Egger检验（p = 0.012）确定了显著的发表偏倚（图4a，B）。我们进一步使用修剪和填充方法来纠正潜在的发表偏倚。结果显示，添加三项插补研究形成对称漏斗图（图4c），校正后的HR仍然显著（HR = 1.831，95%CI 1.577-2.126，p < 0.001），表明我们的结果可靠。

Open in a separate window

Fig. 2  图2

Forest plot for the association between GNRI and overall survival
GNRI与总生存期之间关联的森林图

Table 2 表2

Stratification analysis of the meta-analysis for overall survival in patients with gastrointestinal malignancy
胃肠道恶性肿瘤患者总生存期荟萃分析的分层分析

Subgroup 亚组	No. study 号研究	No. patients 号患者	Pooled HR (95% CI) 合并HR（95% CI）	p	p for subgroup difference p表示亚组差异	Heterogeneity 异质性
						I2(%) I 2 （%）	PQ
Altogether 完全	8	1805	1.94(1.65–2.28) 1.94（1.65-2.28）	< 0.001  <0.001		0.0	0.851
Publishing time 出版时间					0.753
< 2019  <2019年	2	353	2.04(1.44–2.88) 2.04（1.44-2.88）	< 0.001  <0.001		0.0	0.898
≥ 2019	6	1452	1.91(1.59–2.30) 1.91（1.59-2.30）	< 0.001  <0.001		0.0	0.665
Country 国家					0.351
China 中国	2	512	1.79(1.41–2.26) 1.79（1.41-2.26）	< 0.001  <0.001		0.0	0.836
Japan 日本	6	1293	2.09(1.66–2.62) 2.09（1.66-2.62）	< 0.001  <0.001		0.0	0.787
Sample 样品					0.175
< 300  <300	5	1295	1.84(1.53–2.20) 1.84（1.53-2.20）	< 0.001  <0.001		0.0	0.967
≥ 300	3	510	2.46(1.68–3.61) 2.46（1.68-3.61）	< 0.001  <0.001		0.0	0.623
Cutoff value 截止值					0.873
< 98  <98	3	763	1.90(1.42–2.54) 1.90（1.42-2.54）	< 0.001  <0.001		0.0	0.831
≥ 98	5	1042	1.96(1.60–2.38) 1.96（1.60-2.38）	< 0.001  <0.001		0.0	0.567
Cancer site 癌症部位					0.673
ESCC	3	597	1.91(1.44–2.53) 1.91（1.44-2.53）	< 0.001  <0.001		0.0	0.821
GC	1	303	2.09(1.11–3.96) 2.09（1.11-3.96）	0.023		—	—
PDAC	1	282	1.76(1.32–2.34) 1.76（1.32-2.34）	< 0.001  <0.001		—	—
CRLM	1	80	3.72(1.41–9.85) 3.72（1.41-9.85）	0.008		—	—
CRC	2	543	2.04(1.47–2.83) 2.04（1.47-2.83）	< 0.001  <0.001		0.0	0.437
Primary therapy 主要治疗					0.181
With-surgery 伴随手术	7	1725	1.90(1.61–2.24) 1.90（1.61-2.24）	< 0.001  <0.001		0.0	0.956
Mixed 混合	1	80	3.72(1.41–9.85) 3.72（1.41-9.85）

Open in a separate window
在单独的窗口中打开

Open in a separate window

Fig. 3  图3

Sensitivity analysis for the association between GNRI and overall survival
GNRI与总生存期相关性的敏感性分析

Open in a separate window

Fig. 4  图4

Plots for publication bias test in meta-analysis for overall survival. a Begg’s funnel plot; b Egger’s publication bias plot; c The trim-and-fill methods;
总生存期荟萃分析中的发表偏倚检验图。a Begg漏斗图; B Egger发表偏倚图; c修剪填充法;

GNRI and postoperative complication
GNRI与术后并发症

A total of four studies involving 1,194 patients reported a relationship between the GNRI and postoperative complications in patients with GI malignancy. As shown in Fig. 5, the combined results showed that patients with a low GNRI had a higher risk of complications than those with a high GNRI (OR = 2.19, 95% CI 1.57–3.05, p < 0.001). Because there was no heterogeneity, a fixed model (I² = 0.0%, P_Q = 0.940) was used. We performed a further subgroup analysis based on publishing time, country, sample, cutoff value, and cancer site subgroups. The results revealed that a low GNRI was an independent risk factor affecting postoperative complications in all subgroups (Table ​(Table3).3). We also performed a sensitivity analysis by deleting one study at a time and recalculating the combined effect. The results showed that the combined effect was not significantly changed with the deletion of any study, indicating that the results of the meta-analysis for complications were reliable (Fig. 6). In addition, Begg’s test (p = 0.308) and Egger’s test (p = 0.262) both suggested that there was no potential publication bias in the meta-analysis for complications (Fig. 7).
共有4项研究（涉及1，194例患者）报告了胃肠道恶性肿瘤患者GNRI与术后并发症之间的关系。如图5所示，综合结果显示，GNRI低的患者比GNRI高的患者具有更高的并发症风险（OR = 2.19，95%CI 1.57-3.05，p < 0.001）。由于不存在异质性，因此使用固定模型（I ² = 0.0%，P _Q = 0.940）。我们根据发表时间、国家、样本、截止值和癌症部位亚组进行了进一步的亚组分析。结果显示，低GNRI是影响所有亚组术后并发症的独立风险因素（表3）.3）。我们还通过每次删除一项研究并重新计算综合效应进行了敏感性分析。结果显示，删除任何研究均未显著改变联合效应，表明并发症荟萃分析的结果可靠（图1）。 6）。此外，Begg检验（p = 0.308）和Egger检验（p = 0.262）均表明并发症荟萃分析中不存在潜在的发表偏倚（图7）。

Open in a separate window

Fig. 5  图5

Forest plot for the association between GNRI and complications
GNRI与并发症之间关联的森林图

Table 3 表3

Stratification analysis of the meta-analysis for complications in patients with gastrointestinal malignancy
消化道恶性肿瘤并发症Meta分析的分层分析

Subgroup 亚组	No. study 号研究	No. patients 号患者	Pooled HR (95% CI) 合并HR（95% CI）	p	p for subgroup difference p表示亚组差异	Heterogeneity 异质性
						I2(%) I 2 （%）	PQ
Altogether 完全	4	6035	2.19(1.57–3.05) 2.19（1.57-3.05）	< 0.001  <0.001		0.0	0.94
Publishing time 出版时间					0.921
< 2019  <2019年	2	2171	2.22(1.37–3.60) 2.22（1.37-3.60）	0.001		0.0	0.573
≥ 2019	2	3864	2.15(1.35–3.42) 2.15（1.35-3.42）	0.001		0.0	0.783
Country 国家					0.875
China 中国	1	5788	2.28(1.22–4.25) 2.28（1.22-4.25）	0.009		—	—
Japan 日本	3	83	2.15(1.45–3.19) 2.15（1.45-3.19）	< 0.001  <0.001		0.0	0.828
Sample 样品					0.875
< 300  <300	1	1295	2.15(1.45–3.19) 2.15（1.45-3.19）	0.009		—	—
≥ 300	3	4740	1.81(1.58–2.07) 1.81（1.58-2.07）	< 0.001  <0.001		0.0	0.828
Cutoff value 截止值					0.921
< 98  <98	2	3274	2.22(1.37–3.60) 2.22（1.37-3.60）	0.001		0.0	0.573
≥ 98	2	2761	2.15(1.35–3.42) 2.15（1.35-3.42）	0.001		0.0	0.783
Cancer site 癌症部位					0.921
GC	2	888	2.22(1.37–3.60) 2.22（1.37-3.60）	0.001		0.0	0.573
CRC	2	6035	2.15(1.35–3.42) 2.15（1.35-3.42）	0.001		0.0	0.783

Open in a separate window
在单独的窗口中打开

Open in a separate window

Fig. 6  图6

Sensitivity analysis for the association between GNRI and complications
GNRI与并发症相关性的敏感性分析

Open in a separate window

Fig. 7  图7

Plots for publication bias test in meta-analysis for complications. a Begg’s funnel plot; b Egger’s publication bias plot
并发症荟萃分析中的发表偏倚检验图。a Begg漏斗图; B Egger发表偏倚图

GNRI and RFS/DFS/CSS GNRI和RFS/DFS/CSS

We also explored the effects of the GNRI on RFS/DFS/CSS in patients with GI malignancy. Two studies involving 217 patients reported a relationship between the GNRI and RFS. The combined results showed that patients with a low GNRI had shorter RFS (HR = 2.45, 95% CI 1.50–4.00, p < 0.001) than patients with a high GNRI, and the heterogeneity between studies was not significant (I² = 0.0%, P_Q = 0.950), as shown in Fig. 8a. Similarly, a study involving 230 patients confirmed that a low GNRI was an independent risk factor for DFS in patients with GI malignancy (HR = 1.84, 95% CI 1.23–2.76, p = 0.003) (Fig. 8b). However, a study involving 244 patients showed that the GNRI was not an independent predictor of CSS for patients with GI malignancy (HR = 1.60, 95% CI 0.91–2.82, p = 0.101) (Fig. 8c).
我们还探讨了GNRI对胃肠道恶性肿瘤患者RFS/DFS/CSS的影响。两项涉及217例患者的研究报告了GNRI和RFS之间的关系。组合结果显示，具有低GNRI的患者比具有高GNRI的患者具有更短的RFS（HR = 2.45，95%CI 1.50-4.00，p < 0.001），并且研究之间的异质性不显著（I ² = 0.0%，P _Q = 0.950），如图8a所示。类似地，涉及230名患者的研究证实，低GNRI是患有GI恶性肿瘤的患者中DFS的独立风险因素（HR = 1.84，95%CI 1.23-2.76，p = 0.003）（图8b）。然而，涉及244名患者的研究显示，对于GI恶性肿瘤患者，GNRI不是CSS的独立预测因子（HR = 1.60，95%CI 0.91-2.82，p = 0.101）（图8 c）。

Open in a separate window
在单独的窗口中打开

Fig. 8  图8

Forest plot for the association between GNRI and relapse-free survival (a)/disease-free survival (b)/cancer-specific survival (c)
GNRI与无复发生存期（a）/无疾病生存期（B）/癌症特异性生存期（c）之间关联的森林图

Discussion 讨论

In 2005, Bouillanne et al. [25] established the GNRI based on commonly used clinical parameters as a nutrition-related prognostic assessment tool for inpatients. Since then, the GNRI has been widely used to evaluate the prognosis of elderly patients with chronic liver failure, chronic obstructive pulmonary disease, cardiovascular disease, and other chronic diseases. Over the past 5 years with the development of the tumor nutrition theory, the GNRI, as a simple, inexpensive, and readily available biomarker, has been gradually applied to evaluate the prognosis of cancer patients. It was not until 2018 that many studies emerged to explore the relationship between the GNRI and clinical outcome of GI malignancy. As shown in Table ​Table1,1, most studies included in our meta-analysis were published between 2018 and 2020. In other words, the GNRI is still a very novel index in the field of GI malignancy. Hirahara et al. [22] suggested that perioperative nutritional support based on a preoperative GNRI assessment could reduce the incidence of postoperative complications in patients with gastric cancer, thus improving their long-term prognosis. A study of 240 patients reported that the GNRI is a more independent prognostic factor for ESCC patients than body mass index (BMI) and albumin alone [14]. Sasaki et al. [12] also proposed that the GNRI is an independent prognostic factor for the short-term and long-term prognosis of CRC independent of TNM stage. In our previous studies, the authors believed that the GNRI could help identify whether patients were truly malnourished and that the GNRI was more suitable than other nutrition-related indicators (prognostic nutritional index, BMI, and albumin) to assess the prognostic value in elderly CRC patients. In addition, studies have reported that the GNRI is a simple and useful predictor of sarcopenia, which is often regarded as an important adverse prognostic factor in GI malignancy. The GNRI may be more clinically useful than sarcopenia because it is easier to measure and obtain [26].
2005年，Bouillanne等人[ 25]根据常用的临床参数建立了GNRI，作为住院患者的营养相关预后评估工具。此后，GNRI被广泛用于评估慢性肝功能衰竭、慢性阻塞性肺疾病、心血管疾病和其他慢性疾病的老年患者的预后。近5年来，随着肿瘤营养学说的发展，GNRI作为一种简单、廉价、易得的生物标志物，逐渐被应用于肿瘤患者的预后评估。直到2018年，才出现了许多研究来探索GNRI与GI恶性肿瘤临床结局之间的关系。如表1，1所示，我们荟萃分析中纳入的大多数研究都是在2018年至2020年之间发表的。换句话说，GNRI在GI恶性肿瘤领域仍然是一个非常新颖的指标。Hirahara等人 [ 22]提出，基于术前GNRI评估的围手术期营养支持可降低胃癌患者术后并发症的发生率，从而改善其长期预后。一项240例患者的研究报告称，GNRI是ESCC患者比体重指数（BMI）和白蛋白单独更独立的预后因素[ 14]。Sasaki等[ 12]还提出GNRI是CRC短期和长期预后的独立预后因素，与TNM分期无关。在我们之前的研究中，作者认为GNRI可以帮助确定患者是否真正营养不良，并且GNRI比其他营养相关指标（预后营养指数，BMI和白蛋白）更适合评估老年CRC患者的预后价值。 此外，研究报告称，GNRI是肌肉减少症的一个简单而有用的预测因子，肌肉减少症通常被认为是GI恶性肿瘤的一个重要不良预后因素。GNRI在临床上可能比肌肉减少症更有用，因为它更容易测量和获得[ 26]。

It is well-known that malnutrition may increase the incidence of postoperative complications in cancers and adversely affect long-term survival. Timely nutritional support can reduce postoperative morbidity and hospitalization rates [27, 28]. It has been reported that decreased BMI is associated with an increased risk of disease progression and death in patients with GI malignancy [29]. The GNRI integrates weight-related information, which may more objectively reflect the weight change in tumor patients due to tumor consumption. Hypoalbuminemia has been proven to be a poor prognostic factor for a variety of malignancies [30]. It may be associated with decreased albumin, which inhibits the body's ability to cope with stresses such as malignancy and surgery. This would lead to an inadequate anti-tumor immune response and reduced wound healing ability, thereby increasing the risk of postoperative complications and a poor prognosis [31]. The GNRI, combined with these factors, is a useful comprehensive indicator of nutritional and immune status.
众所周知，营养不良可能会增加癌症术后并发症的发生率，并对长期生存产生不利影响。及时的营养支持可以降低术后发病率和住院率[ 27，28]。据报告，BMI降低与GI恶性肿瘤患者疾病进展和死亡风险增加相关[ 29]。GNRI整合了体重相关信息，可以更客观地反映肿瘤患者由于肿瘤消耗而导致的体重变化。低白蛋白血症已被证明是多种恶性肿瘤的不良预后因素[ 30]。它可能与白蛋白减少有关，白蛋白减少会抑制身体科普恶性肿瘤和手术等压力的能力。这将导致抗肿瘤免疫反应不足，伤口愈合能力降低，从而增加术后并发症的风险和预后不良[ 31]。 GNRI与这些因素相结合，是营养和免疫状况的有用综合指标。

In this study, we included nine studies with a total of 2153 patients with GI malignancy. We found that the GNRI is an independent factor influencing the long-term outcome of patients with GI malignancy. The consistent results of sensitivity and subgroup analysis showed that our results were reliable and robust. Although publication bias was detected, we utilized the trim-and-fill method to correct the bias. The results showed that it did not change the significant correlation between a low GNRI and poor OS, indicating that our conclusion was robust. We have further explored the relationship between GNRI and postoperative complications. The combined results showed that the GNRI was an independent influencing factor for postoperative complications in patients with GI malignancy. At the same time, a stratified meta-analysis showed that despite differences in publication year, country, sample size, cutoff value, and cancer site among different groups, a low GNRI was still significantly associated with postoperative complications. Moreover, the sensitivity analysis suggested that the results of mate-analysis for complications were reliable, and there was no publication bias in the mate-analysis for complications. In addition, we also found a significant association between a low GNRI and unfavorable RFS/DFS in GI malignancy, while no statistical association was found between a low GNRI and adverse CSS. Based on these results, the GNRI may be considered an effective and practical clinical indicator for predicting the short-term and long-term outcomes of patients with GI malignancy.
在这项研究中，我们纳入了9项研究，共2153例胃肠道恶性肿瘤患者。我们发现GNRI是影响胃肠道恶性肿瘤患者长期预后的独立因素。敏感性分析和亚组分析的结果一致，表明我们的结果是可靠和稳健的。虽然检测到发表偏倚，但我们使用修剪和填充方法来纠正偏倚。结果表明，这并没有改变低GNRI和低OS之间的显著相关性，表明我们的结论是稳健的。我们进一步探讨了GNRI与术后并发症之间的关系。综合结果表明，GNRI是胃肠道恶性肿瘤患者术后并发症的独立影响因素。 与此同时，分层荟萃分析显示，尽管不同组之间的发表年份、国家、样本量、截止值和癌症部位存在差异，但低GNRI仍与术后并发症显著相关。敏感性分析提示并发症配对分析结果可靠，并发症配对分析无发表偏倚。此外，我们还发现低GNRI与GI恶性肿瘤的不良RFS/DFS之间存在显著相关性，而低GNRI与不良CSS之间没有统计学相关性。基于这些结果，GNRI可以被认为是一个有效和实用的临床指标，用于预测胃肠道恶性肿瘤患者的短期和长期的结果。

To our knowledge, our study is the first meta-analysis to comprehensively explore the value of the GNRI in postoperative complications and long-term outcomes of patients with GI malignancy. We synthesized the existing evidence to prove that the GNRI has a good value in predicting the prognosis of patients with GI malignancy. Our study provides evidence-based support for the clinical application of the GNRI in the prognosis evaluation of patients with GI malignancy. In addition, the critical value of the GNRI in most of the included studies was 98, which may provide some reference value for determining the critical value of the GNRI in clinical applications. We believe that our meta-analysis can provide some inspiration for further research on the relationship between the GNRI and GI malignancy. However, this study still has some limitations. First, all the recruited studies were single center retrospective studies, and the total number of samples and studies was relatively small. More prospective randomized controlled trials are required to investigate the value of the GNRI in GI malignancy. Second, due to the limited number of enrolled studies, the value of the GNRI in RFS/DFS/CSS in GI malignancy still needs to be explored. Finally, this study only included research on all complications. The relationship between the GNRI and specific complications, such as respiratory complications, infectious complications, and anastomotic leakage, still needs to be explored.
据我们所知，我们的研究是第一个荟萃分析，以全面探讨胃肠道恶性肿瘤患者术后并发症和长期结局的GNRI的价值。我们综合现有的证据，证明GNRI对GI恶性肿瘤患者的预后有很好的预测价值。本研究为GNRI在胃肠道恶性肿瘤患者预后评估中的临床应用提供了循证支持。此外，纳入研究中GNRI的临界值大多为98，这可能为临床应用中GNRI临界值的确定提供一定的参考价值。我们相信我们的荟萃分析可以为进一步研究GNRI与胃肠道恶性肿瘤之间的关系提供一些启示。然而，这项研究仍然存在一些局限性。首先，所有招募的研究均为单中心回顾性研究，样本和研究总数相对较少。 需要更多的前瞻性随机对照试验来研究GNRI在胃肠道恶性肿瘤中的价值。其次，由于入组研究数量有限，GNRI在胃肠道恶性肿瘤RFS/DFS/CSS中的价值仍需探索。最后，本研究仅包括所有并发症的研究。GNRI与特定并发症（如呼吸系统并发症、感染并发症和吻合口瘘）之间的关系仍需探讨。

Conclusion 结论

Based on existing evidence, this meta-analysis confirmed that the GNRI was a valuable predictor of complications and long-term outcomes in patients with GI malignancy. Of course, large, multicenter prospective queues are still needed to validate our findings.
基于现有证据，该荟萃分析证实GNRI是GI恶性肿瘤患者并发症和长期结局的有价值的预测因子。当然，仍需要大规模、多中心的前瞻性队列来验证我们的发现。

Acknowledgements 确认

Abbreviations 缩写

Authors’ contributions 作者的贡献

HX and ST designed this research; LW and JG performed the statistical analysis; HX and ST performed the data extraction, and drafted and revised the manuscript. All authors reviewed and approved the final manuscript. All authors read and approved the final manuscript.
HX和ST设计了本研究; LW和JG进行了统计分析; HX和ST进行了数据提取，并起草和修改了手稿。所有作者均审查并批准了最终手稿。所有作者阅读并批准了最终手稿。

Funding 资金

None. 没有。

Availability of data and materials
数据和材料的提供

Please contact author for data requests.
请联系作者获取数据请求。

Ethics approval and consent to participate
伦理学批准和参与同意书

Not applicable. 不适用因

Consent for publication 发表同意书

Not applicable. 不适用因

Competing interests 相互竞争的利益

The authors declare that there is no conflict of interest.
作者声明不存在利益冲突。

Footnotes 脚注

References 引用