Factors influencing malnutrition in critically ill patients: A systematic review and meta-analysis

Abstract

Objective

This study aimed to conduct a systematic review and perform a meta-analysis to identify the factors influencing malnutrition in critically ill patients. It sought to resolve inconsistencies in existing research and address the lack of comprehensive and systematic integration, thereby providing a reference for healthcare professionals to identify the risk of malnutrition in critically ill patients at an early stage and formulate targeted preventive measures.

Methods

Studies were identified through searches of online databases, including PubMed, Embase, Web of Science Core Collection, Cochrane Library, CNKI, and WanFang. The quality of included studies was evaluated using the Newcastle–Ottawa scale and the Agency for Healthcare Research and Quality scale. A meta-analysis was performed using RevMan 5.4 software. This research project was registered in the PROSPERO system, with the registration number CRD420251140462.

Results

The initial search identified 3856 records. After removing duplicates, 3049 records were retained for screening, and 10 studies were ultimately included in the final analysis. The quality assessment showed that eight studies were of high quality and two were of medium quality. The results of the meta-analysis indicated that anemia (odds ratio: 1.48, 95% confidence interval: 1.23–1.78), parenteral nutrition (odds ratio: 4.93, 95% confidence interval: 2.69–9.03), enteral nutrition (odds ratio: 3.2, 95% confidence intervals: 1.04–9.83), and advanced age (odds ratio: 1.24, 95% confidence intervals: 1.01–1.51) were key risk factors for malnutrition in critically ill patients.

Conclusion

Anemia, nutritional support (including enteral nutrition and parenteral nutrition), and advanced age were identified as significant risk factors for malnutrition in critically ill patients and should be prioritized in clinical practice Particular attention should be given to monitoring and correcting anemia in critically ill patients, optimizing nutritional support plans, and conducting continuous nutritional screening and assessment. These measures can help reduce the incidence of malnutrition and improve patient prognosis.

Introduction

Malnutrition is a physiological state resulting from inadequate intake of energy and macronutrients or impaired absorption or utilization. Its primary manifestation is reflected in alterations in body composition, particularly reductions in lean body mass and changes in somatic cell mass. These alterations lead to diminished physical and cognitive function, adversely affecting clinical outcomes. ¹ Critically ill patients typically present with acute onset, severe conditions accompanied by highly unstable vital signs, with the body frequently in a ptstate of high stress. This state not only precipitates a sharp increase in basal metabolic rate, accelerating the consumption of carbohydrates, proteins, and fats, but also increases urinary nitrogen excretion, leading to a negative nitrogen balance. Consequently, the body becomes predisposed to varying degrees of malnutrition. ² Furthermore, due to the high severity of illness, critically ill patients are frequently accompanied by multiple organ dysfunction or failure. The body remains in a state of high catabolism and hypermetabolism, which accelerates weight loss and further exacerbates malnutrition. ³ Studies indicate that intensive care unit (ICU) patients may experience weight loss rates of 0.5–1.0 kg/day. ⁴ Malnutrition is prevalent in ICUs ⁵ and profoundly affects patient outcomes. It prolongs the duration of mechanical ventilation, impairs wound healing, increases complications, extends hospital stays, and elevates mortality rates among critically ill patients. Long-term consequences include diminished quality of life^6,7 and substantial financial burdens on families.

In recent years, although studies on factors influencing malnutrition in critically ill patients have increased,^8–11 the findings remain inconsistent and lack comprehensive systematic synthesis. For instance, Díaz Chavarro et al. ¹² identified polycythemia as a risk factor for malnutrition, whereas Zhang et al. ¹³ suggested anemia as a determinant. Existing studies predominantly focus on specific populations or geographical regions,^14–16 making it challenging to draw consistent conclusions. Moreover, methodological and outcome variations across these studies contribute to confusion and ambiguity within the field. Previously, Mohialdeen Gubari et al. ¹⁷ conducted a systematic review and meta-analysis of nutritional status in ICU patients. Through this analysis, they provided an overview of malnutrition in ICU patients, highlighting its prevalence and severity, and recommended prioritizing nutritional assessment and intervention in clinical practice. However, they did not conduct an in-depth analysis of the factors influencing malnutrition in critically ill patients. Therefore, the present study aimed to conduct a systematic review and meta-analysis of the existing literature on malnutrition in critically ill patients. It sought to explore the factors influencing malnutrition in this population, thereby providing a reference for healthcare professionals to facilitate early identification and the formulation of preventive measures.

Methods

The primary objective of this systematic review and meta-analysis was to identify and analyze factors influencing the occurrence of malnutrition in critically ill patients and to formulate targeted recommendations to reduce its incidence in this population based on the findings. The review protocol has been registered on the PROSPERO website (CRD420251140462) and adheres to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement.

Search strategy

We employed a comprehensive search strategy across six international databases: PubMed, Embase, Web of Science Core Collection, Cochrane Library, CNKI, and WanFang. We included all full-text articles published in peer-reviewed journals from the databases’ inception to 29 August 2025. Conference proceedings, abstracts, and gray literature were excluded. The search terms included combinations of keywords and medical subject headings (MeSH) terms related to “Critical Illness,” “Malnutrition,” “risk factor,” and their synonyms. Boolean operators “AND” and “OR” were used to combine these terms into complex search expressions. In addition to primary databases, we searched gray literature sources such as conference proceedings, theses, and clinical trial registries to identify unpublished studies. Additionally, we manually searched the reference lists of included studies and relevant review articles to identify additional eligible studies that may have been missed during the initial search. A detailed search plan was developed and documented prior to initiating the search to ensure comprehensiveness and reproducibility. The search was not restricted by language or publication status to minimize the risk of publication bias. All retrieved articles were imported into EndNote reference management software for initial screening based on titles and abstracts. Articles of uncertain eligibility were assessed by reading the full text and further evaluated according to predefined inclusion and exclusion criteria. The detailed search strategy can be found in Appendix 1.

Inclusion and exclusion criteria

The inclusion criteria were as follows: (a) study designs including cross-sectional, case–control, or cohort studies; (b) critically ill patients aged ≥18 years; and (c) studies reporting incidence and influencing factors of malnutrition, with results derived from multivariate analyses, including odds ratios (ORs) and 95% confidence intervals (CIs).

The exclusion criteria were as follows: (a) duplicate publications or studies using the same dataset; (b) incomplete data or unavailable full-text articles; and (c) articles not published in Chinese or English.

Study selection process and data extraction

Two researchers independently conducted comprehensive literature searches across six major electronic databases (PubMed, Embase, Web of Science Core Collection, Cochrane Library, CNKI, and WanFang) to identify studies examining factors influencing malnutrition in critically ill patients. The titles and abstracts of retrieved studies were screened by two independent researchers according to the predefined inclusion and exclusion criteria to determine eligibility. The full texts of potentially eligible studies were assessed for further evaluation and final inclusion. Any discrepancies during the selection process were resolved through discussion or consultation with a third researcher. For included studies, a table summarizing study characteristics was created to extract the following information: authors, year of publication, country, language, study design, sample size and characteristics, and factors influencing malnutrition in critically ill patients. The extracted data were cross-checked by two researchers. This rigorous study selection and data extraction methodology ensured the reliability and validity of the data incorporated into our systematic review and meta-analysis, providing a robust foundation for accurate and comprehensive conclusions.

Quality appraisal

In this systematic review and meta-analysis, the Newcastle–Ottawa scale (NOS) ¹⁸ was employed to assess the quality of included cohort studies, whereas the Agency for Healthcare Research and Quality (AHRQ) ¹⁹ was used to evaluate the quality of included cross-sectional studies. These tools were selected because of their comprehensive and effective methodologies for assessing the rigor of research approaches and potential biases within studies examining malnutrition in critically ill patients. NOS evaluates three domains: selection of subjects (4 items, 4 points), comparability of groups (1 item, 2 points), and measurement of outcomes or exposures (3 items, 3 points). With a total of 9 items, the maximum score was 9 points. Scores ≤3 indicated low-quality literature, 4–6 denoted moderate-quality literature, and ≥7 signified high-quality literature. The AHRQ scale comprises 11 items that were answered as “Yes,” “No,” or “Unclear.” A response of “Yes” scored 1 point, whereas “Unclear” or “No” scored 0 points. A total score ≤3 indicated low-quality literature, 4–7 indicated moderate-quality literature, and ≥8 indicated high-quality literature. Each study was independently assessed by two authors to ensure objectivity and consistency in the evaluation process. In case of discrepancy, discussions were held with a third reviewer to reach a final determination.

Statistical analysis

Data analysis was conducted using RevMan 5.4 software. Data were pooled using the OR values and 95% CI reported in the original studies. Heterogeneity was assessed using the I² statistic. An I² value ≤50% indicated acceptable heterogeneity among studies, allowing data pooling using a fixed-effects model. An I² value >50% suggested substantial heterogeneity, necessitating the use of a random-effects model for effect size pooling. Sensitivity analysis was conducted by sequentially excluding individual studies. A p value <0.05 was considered statistically significant.

Results

Selection of included studies

A literature search across six electronic databases identified 3856 research papers. After removing duplicate records, 3049 studies were subjected to rigorous assessment and screening according to predefined criteria. Ultimately, 10 research papers met the inclusion criteria and were included in the quantitative synthesis. The literature selection flowchart is presented in Figure 1. The basic characteristics of all included studies are summarized in Table 1.^{12,13,20–27}

Figure 1.

Flow chart of study selection.

Table 1.

Summary of study characteristics.

Author and publication year	Country/region	Language	Study design	Sample size	Influencing factors
Kang, 2020 20	China	Chinese	Cross-sectional study	178	①②③④⑤⑥⑦
Li et al., 2023 21	China	Chinese	Cohort study	782	①⑧⑩⑪⑫⑬⑭㉚
Liu et al.,2025 22	China	Chinese	Cohort study	183	⑪⑲
Zhang, 2022 23	China	Chinese	Cross-sectional study	80	⑧⑳㉑㉒㉔
Zhang et al., 2024 13	China	Chinese	Cross-sectional study	112	⑧⑬㉕㉖㉗㉘㉚
Zhu et al., 2025 24	China	Chinese	Cross-sectional study	326	㉓㉙
Zhu et al., 2025 25	China	Chinese	Cross-sectional study	195	⑧⑰⑳㉒㉕
Barreto et al., 2019 26	United States	English	Cohort study	398	⑧⑯
Díaz Chavarro et al., 2024 12	Colombia	English	Cohort study	630	①⑨⑭⑮⑱㉚
Shabanpur et al., 2022 27	Iran	English	Cross-sectional study	400	⑧⑱

①: parenteral nutrition; ②: gastrointestinal bleeding; ③: use of sedative drugs; ④: outpatient examination; ⑤: abdominal hypertension; ⑥: vomiting and diarrhea; ⑦: acute physiology and chronic health evaluation II (APACHE II) score; ⑧: age; ⑨: erythrocyte increase; ⑩: Charlson comorbidity index (CCI) score; ⑪: Glasgow coma scale (GCS) score; ⑫: neutrophil count; ⑬: total egg count; ⑭: enteral nutrition; ⑮: insufficient intake; ⑯: male gender; ⑰: drinking alcohol; ⑱: loss of appetite; ⑲: early nutritional support; ⑳: complicated with heart failure; ㉑: fever; ㉒: negative emotions; ㉓: pulmonary disease; ㉔: activity impairment; ㉕: body mass index (BMI); ㉖: diabetes; ㉗: low cholesterol; ㉘: low albumin; ㉙: tracheotomy; ㉚: anemia.

Methodological quality of included studies

The methodological quality of the included studies was evaluated using different tools based on study design. Cohort studies were assessed using the NOS scale, which has a maximum score of 9 points. Scores ≤3 indicated low-quality literature, 4–6 denoted moderate quality, and ≥7 signified high quality. Cross-sectional studies were evaluated using the AHRQ questionnaire, with a maximum score of 11 points. Scores ≤3 denoted low-quality literature, 4–7 indicated moderate quality, and ≥8 signified high quality. All studies appropriately targeted their intended populations and measured participants in a reliable manner. The specific scoring details are presented in Tables 2 and 3.^{12,13,20–27}

Table 2.

Quality appraisal of included cohort studies.

	Representativeness of the exposed cohort				Comparability	Outcome
Studies	Q1	Q2	Q3	Q4	Q5	Q6	Q7	Q8	Total score	Grade
Li et al., 2023 21	1	1	1	2	1	1	1	1	9	High quality
Liu et al.,2025 22	1	1	1	2	1	1	1	1	9	High quality
Barreto et al., 2019 26	1	1	1	2	1	1	1	1	9	High quality
Díaz Chavarro et al., 2024 12	1	1	1	1	1	1	1	1	8	High quality

Q1: selection of the nonexposed cohort; Q2: selection of the exposed cohort; Q3: ascertainment of exposure; Q4: demonstration that the outcome of interest was not present at the start of the study; Q5: comparability of cohorts on the basis of design or analysis; Q6: assessment of outcome; Q7: follow-up long enough for outcomes to occur; Q8: adequacy of follow-up of cohorts.

Table 3.

Quality appraisal of included cross-sectional studies.

Studies	Q1	Q2	Q3	Q4	Q5	Q6	Q7	Q8	Q9	Q10	Q11	Total score	Grade
Kang, 2020 20	Y	Y	Y	Y	N	N	Y	Y	Y	Y	N	8	High quality
Zhang and Zhang, 2022 23	Y	Y	Y	Y	N	N	Y	Y	Y	Y	N	8	High quality
Zhang et al., 2024 13	Y	Y	Y	U	N	N	Y	Y	Y	Y	N	7	Moderate quality
Zhu et al., 2025 24	Y	Y	Y	Y	N	N	Y	Y	Y	Y	N	8	High quality
Zhu et al., 2025 25	Y	Y	Y	Y	N	N	N	Y	Y	Y	N	6	Moderate quality
Shabanpur et al., 2022 27	Y	Y	Y	Y	N	N	Y	Y	Y	Y	Y	9	High quality

Q1: define the source of information (survey and record review); Q2: list inclusion and exclusion criteria for exposed and unexposed subjects (cases and controls) or refer to previous publications; Q3: indicate the time period used for identifying patients; Q4: indicate whether subjects were consecutive if not population-based; Q5: indicate whether evaluators of subjective components of the study were masked to other aspects of participants’ status; Q6: describe any assessments undertaken for quality assurance purposes (e.g. test/retest of primary outcome measurements); Q7: explain any patient exclusions from the analysis; Q8: describe how confounding was assessed and/or controlled; Q9: if applicable, explain how missing data were handled in the analysis; Q10: summarize patient response rates and completeness of data collection; Q11: clarify what follow-up, if any, was expected and the percentage of patients for whom incomplete data or follow-up was obtained.

Factors associated with malnutrition in critically ill patients

Based on a review of 10 studies, several factors associated with the development of malnutrition in critically ill patients were identified.

Anemia

Three studies^12,13,21 reported the impact of anemia on malnutrition in critically ill patients. The meta-analysis results demonstrated homogeneity among the included studies (I² = 0%, p = 0.73). The fixed-effects model indicated that anemia was a risk factor for malnutrition in critically ill patients (OR: 1.48, 95% CI: 1.23–1.78, p < 0.0001), as shown in Appendix A in the Supplementary material.

Nutritional support

Three studies^12,20,21 reported the effects of parenteral nutrition on malnutrition in critically ill patients. The meta-analysis results indicated homogeneity among the included studies (I² = 0%, p = 0.49). The fixed-effects model demonstrated that parenteral nutrition was a risk factor for malnutrition in critically ill patients (OR: 4.93, 95% CI: 2.69–9.03, p < 0.00001), as shown in Appendix B in the Supplementary material.

Two studies^12,21 reported the effects of enteral nutrition on malnutrition in critically ill patients. The meta-analysis results revealed significant heterogeneity among the included studies (I² = 86%, p = 0.007). Under the random-effects model, enteral nutrition was identified as a risk factor for malnutrition in critically ill patients (OR: 3.2, 95% CI: 1.04–9.83, p = 0.04), as shown in Appendix C in the Supplementary material.

Age

Six studies^{13,21,23,25–27} reported the impact of age on malnutrition in critically ill patients, with significant statistical heterogeneity (I² = 88%, p < 0.00001). Using a random-effects model, the meta-analysis indicated that older critically ill patients were at a higher risk of malnutrition (OR: 1.24, 95% CI: 1.01–1.51, p = 0.04), as shown in Appendix D in the Supplementary material.

Source of heterogeneity and sensitivity analysis

In this study, two influencing factors exhibited high heterogeneity: enteral nutrition (two included studies, I² = 86%, p = 0.007) and age (six included studies, I² = 90%, p < 0.00001). As only two studies on enteral nutrition were available, subgroup analysis could not be performed. For the age factor, we extracted information on study design, data collection methods, and geographic regions of the included studies. Subsequently, we conducted a source-of-heterogeneity analysis for both factors, with the results as follows.

Age

We conducted three subgroup analyses for age, none of which reduced study heterogeneity. Detailed results are presented in Appendix E. Sensitivity analysis was performed by sequentially excluding studies. After removing four studies,^13,23,26,27 no statistical heterogeneity between the remaining two studies was observed (I² = 0%, p = 0.86). Results from the random-effects model and fixed-effects model showed minimal divergence, indicating robust stability (good sensitivity), as illustrated in Appendix F in the Supplementary material.

The significant heterogeneity observed across these four studies can be attributed to multiple differences in methodology and study populations. First, there were marked variations in the nutritional assessment tools employed. These studies used distinct evaluation criteria, including the NS2002 Nutritional Assessment Scale, the Global Leadership Initiative on Malnutrition (GLIM) criteria, the Sarcopenia Index (SGA), and the Nutritional Risk Screening 2002 (NRS-2002). These tools differed in their emphasis on phenotypic indicators (e.g. weight loss and body mass index) versus etiological indicators (e.g. inflammatory response and reduced food intake), leading to inconsistent malnutrition classification outcomes. Second, the disease spectrum within the study populations varied considerably. Two studies particularly targeted distinct critical care subgroups (patients with severe coronavirus disease 2019 (COVID-19) and older patients with severe stroke),^13,23 whereas another two studies included mixed critical care cohorts (e.g. patients with sepsis, shock, and major trauma),^26,27 exhibiting variations in baseline characteristics such as the prevalence of comorbidities (e.g. chronic heart failure, diabetes mellitus, and malignancies) and disease severity (e.g. Acute Physiology and Chronic Health Evaluation III score, Sequential Organ Failure Score, and Nutrition Risk Screening Scale score). Third, study 6 conducted nutritional assessments 3–5 days postadmission, whereas Zhang et al. ¹³ performed assessments within 48 h of ICU admission. Two studies^23,26 adjusted for age, sex, and disease severity scores, whereas Shabanpur et al. ²⁷ did not explicitly report adjustment for all key confounders. These methodological discrepancies may have influenced the estimated effect sizes for risk factors. Furthermore, Zhang and Zhang ²³ included 80 patients, whereas Shabanpur et al. ²⁷ enrolled 400 patients. This disparity in sample size may have further exacerbated the observed heterogeneity, as effect estimates from smaller cohorts are more susceptible to random variation than those from larger samples. Collectively, these factors contribute to the substantial heterogeneity in outcomes and conclusions across these four studies.

Enteral nutrition

The two included studies^12,21 exhibited substantial heterogeneity. Upon comparing these two studies, we identified significant heterogeneity across multiple dimensions. First, in terms of study participants, Li et al. ²¹ focused on older patients with severe stroke, whereas Díaz Chavarro et al. ¹² included critically ill patients with multiple disease etiologies. This resulted in differences in the disease background and physiological status of the study populations. Second, regarding research tools, Li et al. ²¹ employed a nomogram prediction model to assess malnutrition risk, whereas Díaz Chavarro et al. ¹² used the malnutrition universal screening tool (MUST) scale. These instruments differ in their assessment methodologies and areas of emphasis. Moreover, the two studies differ in design: Li et al. ²¹ is a single-center retrospective study, whereas Díaz Chavarro et al. ¹² is a multicenter prospective study. This may affect the generalizability and extrapolation of the findings. Finally, regarding the findings, the nomogram model constructed in Li et al. ²¹ demonstrated favorable predictive efficacy, whereas Díaz Chavarro et al. ¹² identified multiple factors associated with nutritional risk, including gastrointestinal symptoms and nutritional support modalities. These differences collectively constitute the heterogeneity of this factor.

Publication bias analysis

This study employed funnel plot analysis based on age. The results indicate that the funnel plot exhibits incomplete symmetry between the left and right sides, suggesting the potential presence of publication bias, as shown in Appendix G in the Supplementary material.

Discussion

This systematic review and meta-analysis was conducted to comprehensively examine the factors influencing malnutrition in critically ill patients. The findings revealed that anemia, mode of nutritional support, and advanced age are key risk factors for this condition.

This study identified anemia as a significant risk factor for malnutrition in critically ill patients, consistent with previous studies.^28–30 Reduced hemoglobin levels not only reflect impaired iron metabolism and suboptimal protein status but also serve as an indicator of underlying nutritional deficiency. ³¹ Specifically, anemia-induced inadequate tissue oxygenation can disrupt cellular metabolic processes, impairing the body’s ability to utilize and metabolize nutrients, ultimately contributing to the development of malnutrition. ³² Furthermore, patients with anemia often experience diminished appetite, leading to reduced food intake that further exacerbates the severity of malnutrition. ³³ Consequently, in clinical practice, priority should be given to monitoring the blood test results of critically ill patients, promptly correcting anemia while sustaining ongoing nutritional screening and assessment. Additionally, timely nutritional interventions aimed at improving appetite and nutrient intake, particularly through the supplementation of essential protein and energy, ³⁴ can help reduce the risk of malnutrition.

Parenteral and enteral nutrition are common methods of nutritional support for critically ill patients. However, this study found that patients receiving these two forms of intervention were at a higher risk of malnutrition than those consuming food orally. Similarly, Chen et al. ³⁵ reported that tube feeding (a form of enteral nutrition) is a risk factor for malnutrition in patients with stroke, a phenomenon that may be attributed to issues related to the method, dosage, and timing of nutritional support. Specifically, parenteral nutrition may lead to incomplete absorption and utilization of nutrients, which in turn contributes to malnutrition. ³⁶ It may also damage the intestinal mucosa, causing gastrointestinal dysfunction that further elevates the risk of malnutrition. Although enteral nutrition better aligns with the body’s physiological requirements, nasogastric feeding (a common type of enteral nutrition) may reduce patients’ interest in food and increase the risk of reflux, aspiration, and pulmonary infections, all of which indirectly contribute to malnutrition. Therefore, clinicians should select appropriate nutritional support methods based on individual patient circumstances, optimize the dosage and timing of support, and prioritize the protection of the intestinal mucosa and maintenance of patients’ dietary interest to minimize the incidence of malnutrition.

Our findings indicate that critically ill older patients are at a higher risk of malnutrition, consistent with the findings of Weng et al. ³⁷ This increased vulnerability may result from age-related physiological characteristics, including advancing age, diminished body’s physiological functions, reduced bodily reserves, and weakened digestive and masticatory capacities,^38,39 all of which reduce the ability to absorb energy and protein. ⁴⁰ Compared with younger patients, older patients are therefore more susceptible to malnutrition due to these age-related impairments (e.g. diminished physiological function, impaired digestion, and swallowing difficulties). ⁴¹ Furthermore, older patients have a reduced basal metabolic rate, diminished capacity for nutrient absorption and utilization, and frequently suffer from multiple chronic conditions, all of which interfere with nutrient metabolism and utilization, further elevating the risk of malnutrition. Decreased appetite and reduced food intake are additional key contributors to malnutrition. Therefore, particular emphasis should be placed on enhancing nutritional assessment for critically ill older patients by providing personalized nutritional support programs, improving dietary composition and feeding methods, and concurrently managing chronic conditions to effectively reduce the incidence of malnutrition.

This study has several limitations that should be acknowledged. First, significant heterogeneity existed among the included studies in terms of “age,” which, to some extent, reduced the reliability of the meta-analysis results. Although we conducted subgroup analyses based on study characteristics, data collection methods, and survey regions, heterogeneity persisted. Sensitivity analyses and in-depth examination of the included studies revealed that this heterogeneity primarily resulted from differences in study populations, assessment methods, research designs, and statistical approaches across four studies. For instance, Zhang and Zhang ²³ conducted a cross-sectional survey using the NS2002 Nutrition Assessment Scale among critically ill patients with COVID-19. Zhang et al. ¹³ focused on older patients with severe stroke, employing the GLIM criteria in a cross-sectional design. Barreto et al. ²⁶ included a cohort of critically ill patients across various ICU settings and used the Subjective Global Assessment as a reference and the Sarcopenia Index to determine malnutrition status. Shabanpur et al. ²⁷ did not explicitly restrict inclusion to a single critical illness type and instead enrolled a mixed cohort of critically ill hospitalized patients, utilizing the NRS-2002 to assess nutritional status and failing to clearly report a comprehensive adjustment strategy for age-related confounding factors. These differences collectively resulted in high heterogeneity among the four studies regarding the association between “age” and malnutrition. Regarding “enteral nutrition,” the two studies incorporating this factor also exhibited significant discrepancies. Li et al. ²¹ employed a nomogram prediction model to assess critically ill older patients with stroke in a single-center retrospective study, whereas Díaz Chavarro et al. ¹² used the MUST scale to evaluate critically ill patients with multiple conditions in a multicenter prospective study. These differences in study populations and designs constitute the primary sources of significant heterogeneity between the two studies. Second, the inclusion of only Chinese and English articles may have introduced language bias. Additionally, our literature review identified factors such as alcohol consumption, negative emotions, and diabetes as contributors to malnutrition in critically ill patients; however, insufficient data precluded meta-analysis. Finally, asymmetric funnel plots suggest potential publication bias. Therefore, further cross-sectional studies are warranted with adequate sample sizes.

Conclusions

A comprehensive analysis of 3284 participants across 10 studies concluded that anemia, parenteral nutrition, enteral nutrition, and age are key factors influencing malnutrition in critically ill patients. However, given the current scarcity of research in this field and limitations in evidence quality, further high-quality, large-scale studies are warranted.

Supplemental material

Supplemental material for this article is available online.