The Prognostic Role of Prognostic Nutritional Index and Controlling Nutritional Status in Predicting Survival in Older Adults with Oncological Disease: A Systematic Review

Simple Summary: Most new cancers worldwide are diagnosed in older adults, highlighting the need for accessible and easy-to-use prognostic tools that contribute to lowering the burden of the disease in this age group. The aim of this systematic review is to understand whether the Prognostic Nutritional Index (PNI) and the Controlling Nutritional Status (CONUT) can predict survival in older adult cancer patients. The 38 studies included in this review vary substantially in terms of patients, cancer types, survival outcomes, and what is considered high and low PNI or CONUT. Overall, PNI showed an association with overall survival in most studies, indicating that it is an inexpensive biomarker that could be used as a prognostic tool in older adults diagnosed with cancer. Abstract: The increase in new cancer diagnoses in the elderly calls for new, accessible, and easy-to-use prognostic tools that contribute to lowering the burden of the disease. Recognising the importance of inflammation and nutritional status in the progression of the disease, the purpose of this systematic review was to synthesise the evidence on the prognostic role of Prognostic Nutritional Index (PNI) and Controlling Nutritional Status (CONUT) in predicting survival of older adult cancer patients. A comprehensive search was conducted in PubMed and Web of Science Core Collection databases until 22 February 2024. The articles included in this review ( n = 38) examined the relationships of PNI and CONUT with survival outcomes in elderly cancer patients. Despite high heterogeneity between the studies, most concluded that low PNI values are associated with poor overall survival (OS), particularly in gastric cancer patients. Most studies did not find an association between PNI and cancer-specific survival, progression-free survival, disease-free survival, recurrence-free survival, and mortality. Results regarding the prognostic role of CONUT in predicting survival were inconclusive. This study suggests that PNI could be used to predict OS in elderly cancer patients, while more studies are needed to assess the prognostic role of CONUT.


Introduction
In 2022, more than 53% (over 10.5 million cases) of new cancers worldwide were diagnosed in the elderly, a number expected to double by 2045 [1].The five most common new diagnoses in individuals over 65 years old in 2022 were lung, colorectum, prostate, breast, and stomach cancers [1].Regarding mortality in this age group, 52% were attributed to lung, colorectum, stomach, liver, and prostate cancers combined [1].Therefore, easy-touse and accessible tools are necessary to rapidly evaluate prognosis, hence contributing to reducing the burden of the disease.
Onco 2024, 4 102 Systemic inflammation is involved in angiogenesis, tumour development, and metastasis [2].Several non-invasive models based on routine peripheral blood parameters have been developed to assess the prognostic of cancer-related outcomes [2][3][4].Examples that have been shown to predict cancer prognosis are the platelet-to-lymphocyte ratio (PLR), neutrophil-to-lymphocyte ratio (NLR), lymphocyte-to-monocyte ratio (LMR), and the systemic immune inflammation index (SII) [5][6][7][8].Besides inflammation status, malnutrition is a significant factor in cancer, for it is a common hallmark in cancer patients and can be a determinant of poor prognosis [9,10].Malnutrition is responsible for 10-20% of the mortality in cancer patients [9], and it has been shown that cachexia and anorexia are correlated with poor prognosis in these patients [10,11].Thus, combining inflammation and nutritional factors in a single, easy-to-obtain model that can predict cancer prognosis in older adults is a necessary step in choosing effective treatment approaches that could alleviate the burden of cancer in the elderly.
Two prominent models that include inflammation and nutritional factors are the PNI and CONUT [12,13].The PNI was initially proposed to evaluate the nutritional status of patients undergoing gastrointestinal surgery, but recent studies have attested its prognostic value in cancer patients, including older adults [14,15].Similarly, CONUT was proposed as a simple, low-cost tool to assess malnutrition in hospitalised patients and has recently been shown to predict OS in older adults with several types of cancer [16,17].However, evidence of PNI and CONUT's predictive value for survival in elderly cancer patients is still insufficient and conflicting.Therefore, this systematic review aims to synthesise the evidence on the ability of PNI and CONUT to predict survival in older adult cancer patients.

Materials and Methods
This study follows the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) [18].

Inclusion and Exclusion Criteria
All published research was included if it met the following criteria: (1) patients with cancer diagnosis and age equal to or above 60 years old [19]; (2) presented comparisons between high and low PNI or high and low CONUT scores; (3) assessed the relationship between PNI or CONUT and overall survival (OS), cancer-specific survival (CSS), progression-free survival (PFS), disease-free survival (DFS), recurrence-free survival (RFS), or mortality, assessed through hazard ratios (HRs); (4) with an observational design, including cohort, case-control, or cross-sectional.Studies were excluded if they were (1) not primary research, (2) animal or cell studies, (3) written in languages other than English, and (4) unavailable for retrieval.

Search Strategy
Comprehensive literature searches were conducted in PubMed and Web of Science Core Collection up until 22 February 2024.Terms used in the searches were related to the population under study (older adults), the outcomes of interest (mortality, survival), the prognostic tools (PNI, CONUT), and cancer diagnoses (cancer).Table A1 in Appendix A provides the detailed search strategy for each database.

Study Selection and Data Collection
All records were downloaded from PubMed and Web of Science into Rayyan [20], where duplicates were automatically identified and manually removed after confirmation by two researchers.The remaining records were screened independently by two researchers based on their titles and abstracts to assess their fit against the inclusion criteria.Relevant studies were then retrieved, and full texts were assessed by two researchers.Disagreements were resolved through consensus or by the opinion of a third researcher.Lastly, one researcher extracted data from the included studies using a custom form that included the research design, country, duration and follow-up periods, sample size, elderly age definition, cancer characteristics (location, stage, treatment), PNI and/or CONUT cut-offs, cut-off determination, timing of blood tests, outcomes, and results.

Risk-of-Bias Assessment
Potential bias assessment in the included studies was performed independently by two researchers using a standardised quality evaluation tool, namely the Newcastle-Ottawa quality assessment scale (NOS) for cohort studies [21].The NOS evaluates three components in each study: representativeness of selected participants (0-4 points), comparability between the exposed and control groups (0-2 points), and the outcome of the study (0-3 points) [22].The NOS scale ranges from 0 to 9, and studies attaining ≥7 points were considered of good quality.

Selection and Characteristics of Included Studies
As Figure 1 shows, 210 records were identified through searches in PubMed (n = 94) and Web of Science Core Collection (n = 116) databases.After duplicate removal (n = 77), 15 records were excluded based on title and abstract screening.Of the 110 full texts retrieved (not retrieved = 8), 72 studies were excluded for using the wrong predictor (n = 27), the wrong population (n = 20), the wrong design (n = 12), and the wrong outcome (n = 13).Examples of excluded studies in this step are, for instance, that of Okada et al. [23], which assessed the prognostic value of PNI in postoperative complications but not in survival.In another study [24], even though the PNI was calculated, the authors combined it and other nutritional markers with the TNM stage to construct a new staging system that was posteriorly analysed for its prognostic validity.
Relevant studies were then retrieved, and full texts were assessed by two researchers.Disagreements were resolved through consensus or by the opinion of a third researcher.Lastly, one researcher extracted data from the included studies using a custom form that included the research design, country, duration and follow-up periods, sample size, elderly age definition, cancer characteristics (location, stage, treatment), PNI and/or CO-NUT cut-offs, cut-off determination, timing of blood tests, outcomes, and results.

Risk-of-Bias Assessment
Potential bias assessment in the included studies was performed independently by two researchers using a standardised quality evaluation tool, namely the Newcastle-Ottawa quality assessment scale (NOS) for cohort studies [21].The NOS evaluates three components in each study: representativeness of selected participants (0-4 points), comparability between the exposed and control groups (0-2 points), and the outcome of the study (0-3 points) [22].The NOS scale ranges from 0 to 9, and studies attaining ≥7 points were considered of good quality.

Selection and Characteristics of Included Studies
As Figure 1 shows, 210 records were identified through searches in PubMed (n = 94) and Web of Science Core Collection (n = 116) databases.After duplicate removal (n = 77), 15 records were excluded based on title and abstract screening.Of the 110 full texts retrieved (not retrieved = 8), 72 studies were excluded for using the wrong predictor (n = 27), the wrong population (n = 20), the wrong design (n = 12), and the wrong outcome (n = 13).Examples of excluded studies in this step are, for instance, that of Okada et al. [23], which assessed the prognostic value of PNI in postoperative complications but not in survival.In another study [24], even though the PNI was calculated, the authors combined it and other nutritional markers with the TNM stage to construct a new staging system that was posteriorly analysed for its prognostic validity.A total of 38 studies, published between 2012 and 2023, were included in this systematic review [16,17,, representing 8715 participants diagnosed with cancer and aged 60 or above (Table 1).Sample sizes varied substantially between studies, ranging from 34 A total of 38 studies, published between 2012 and 2023, were included in this systematic review [16,17,, representing 8715 participants diagnosed with cancer and aged 60 or above (Table 1).Sample sizes varied substantially between studies, ranging from 34 [55] to 1949 [17] participants (median: 155 [P 25 : 89, P 75 : 278]).There was no consensus as to the definition of older individuals among the studies, with most including people above or equal to 75 years old (39%) or 65 years old (24%).The majority of the studies were retrospective cohort studies, with only two being prospective cohort studies [17,34].Twenty-three studies (61%) were conducted in Japan, ten in China (26%), two in the Republic of Korea (5%), one in Australia, one in France, and one in Türkiye.Nineteen studies (50%) focused on gastric cancer, five (13%) on colorectal cancer, four (11%) on oesophageal squamous-cell carcinoma, four (11%) on lung cancer, and the remaining dealt with diffuse large B cell lymphoma (n = 2), glioblastoma (n = 1), hepatic cancer (n = 1), and osteosarcoma (n = 1).Notably, the study by Zhang et al. [17] analysed the prognostic value of CONUT and PNI in a pooled sample, irrespective of cancer location, and independently for lung, digestive, and other locations.

Assessment of Study Quality
According to the NOS scale (Table 2), all cohort studies but one were considered of good quality (7-9 points), with the study by Xishan et al. [51] scoring a moderate quality (6 points).

Prognostic Nutritional Index
Thirty-six studies reported the hazard ratios for OS based on PNI scores of older adults with cancer.The independent association of PNI with OS was found to be significant in 23 studies, while 13 did not found a significant relationship in multivariate analysis.Thirteen studies supported the prognostic value of PNI in gastric cancer patients [25,28,29,[37][38][39]41,45,48,50,54,56,58], one in colorectal cancer [60], one in oesophageal squamouscell carcinoma (ESCC) [53], three in lung cancer [17,35,49], one in diffuse large B cell lymphoma [52], one in hepatocellular carcinoma [26], one in osteosarcoma [31], one in digestive cancer [17], and one in cancer patients in general [17].The relationship between PNI and CSS was assessed by five studies, and none supported the relationship.Only one of the seven studies assessing the relationship between PNI and PFS/DFS/RFS found a significant association [26].The only study using PNI to predict postoperative 90-day mortality failed to establish a significant relationship [44].

Controlling Nutritional Status
The hazard ratios for OS based on CONUT scores were reported by ten studies, of which six found it to be an independent predictor of OS.One of those studies was in elderly gastric cancer patients [42], one in ESCC [36], one in digestive cancer [17], two in lung cancer [16,17], and one in cancer patients in general [17].One study found that CONUT was an independent predictor of CSS in gastric cancer patients [42].Lastly, one of the two studies that assessed the relationship between CONUT and PFS/DFS/RFS found it to be a significant association, specifically in ESCC patients [36].
Table 3 summarises the number of studies that did or did not support the association of the nutritional indicators with survival outcomes.

Discussion
Older adults with cancer are at a particularly high risk of malnutrition, which derives not only from cancer-induced factors and anti-cancer treatment, but also from age-related changes and associated comorbidities [9,61].Recommendations stress the importance of identifying the risk of malnutrition in cancer patients as soon as the cancer diagnosis is established [9,62].However, malnutrition is still underestimated and undertreated in these patients [62,63].If left unchecked, a poor nutritional status can lead to functional decline, poorer treatment responses, increased perioperative complications, longer recovery, increased length of hospital stay, poorer quality of life, and increased mortality [15,61,64].Even though nutritional screening tools like the PG-SGA or the NRS2002 have been used to identify nutritional risk in cancer patients [65,66], the results are often disparate due to their inherent subjectivity and complexity.Thus, scholars and practitioners strive for a simple, accessible, and affordable screening tool that can, simultaneously, be sensitive and objective in identifying malnutrition risk in cancer patients [64,67].In this context, the PNI and CONUT have been proposed as nutritional screening tools with prognostic value for older adults diagnosed with cancer [15][16][17].
The results of this study show that the relationship of PNI with OS of older adults with cancer was the most studied in the corpus, being supported by 21 out of 34 studies.This is consistent with the results of previous systematic reviews and meta-analyses of individuals of all age groups with gastric cancer [68,69], lung cancer [70], and lymphoma [71], which concluded that a low PNI adversely affected OS.Conversely, results are less conclusive about the prognostic value of CONUT in OS, where the number of studies supporting the relationship is the same as that of those failing to find an association.Given the predominance of studies conducted in Japan and China, and the variability in PNI and CONUT cut-off values, subgroup analyses were performed, revealing that neither the country nor the PNI or CONUT cut-off value affected the first conclusions.Thus, while PNI seems to have an ability to predict OS in elderly cancer patients, CONUT is less likely to have a prognostic role in this age group.
The relationship of PNI with other survival outcomes (CSS, PFS/DFS/RFS, 90-day mortality) was supported only by one out of thirteen studies.This is a contradictory finding when compared with several meta-analyses with cancer patients from all age groups [68,70,71] that found low PNI values to be significantly associated with poor CSS and PFS/DFS/RFS.Unfortunately, the small number of studies analysing the association between CONUT and these survival outcomes prevents meaningful subgroup analyses.Overall, the results do not support a prognostic role of PNI in CSS, PFS/DFS/RFS, or 90-day mortality in older adults with cancer, and more research is needed on the ability of CONUT to predict these outcomes.
This systematic review has several limitations that should be acknowledged when interpreting its results.First, all studies were observational and only two were not retrospective cohorts, which could have introduced selection and recall biases [72].Only adjusted HRs were analysed to partially overcome these biases, yet control variables differed between studies.Second, there was a high between-studies variability regarding the number of participants, PNI and CONUT cut-off values, and outcomes.Sub-group analyses were performed to try to understand the source of variability.However, the low number of studies inhibited drawing a solid conclusion.Third, most of the included studies were conducted in Asian countries, namely Japan, China, and the Republic of Korea.Consequently, the results cannot be generalised to populations that are not of Asian descent.This highlights the need for further well-designed studies in other regions where the elderly are a significant part of the population, particularly Europe and North America.

Conclusions
This study revealed the usefulness of the PNI, an inexpensive and easy-to-use biomarker, to predict OS in elderly cancer patients, particularly those with gastric cancer.The results could not, however, attest the prognostic value of PNI in other measures of survival and that of CONUT in any survival outcome.

Figure 1 .
Figure 1.Selection of studies for inclusion in the review.

Figure 1 .
Figure 1.Selection of studies for inclusion in the review.

Table 1 .
Characteristics of included studies.

Table 2 .
NOS quality assessment of cohort studies.

Table 2 .
Cont.Representativeness of the exposed cohort; selection of the non-exposed cohort; ascertainment of exposure; demonstration that outcome of interest was not present at start of study (maximum of 4 points).b Comparability of cohorts on the basis of the design or analysis (maximum of 2 points).c Assessment of outcome; was follow-up long enough for outcomes to occur; adequacy of follow up of cohorts (maximum of 3 points). a

Table 3 .
Summary of relationships between nutritional indicators and survival outcomes.

Table 3 .
Cont.Values show the number of studies that "support/do not support" the relationships between nutritional indicators and the outcomes.ESCC: oesophageal squamous cell carcinoma.PNI: Prognostic Nutritional Index.CONUT: Controlling Nutritional Status.OS: overall survival.CSS: cancer-specific survival.PFS: progression-free survival.DFS: disease-free survival.RFS: recurrence-free survival.