A prospective study on age-specific normative values of the prognostic nutritional index and the effects of malnutrition on spinal alignment using health checkup data of elderly residents
Article information
Abstract
Study Design
A prospective cohort study.
Purpose
To determine the age-specific normative values of the prognostic nutritional index (PNI) among elderly residents in Japan and explore the relationship between malnutrition and spinal alignment.
Overview of Literature
Nutritional status affects postoperative recovery, with malnourished patients often experiencing severe postoperative complications. PNI is a known nutritional indicator based on serological value; however, there is a dearth of age-specific normative values for PNI, with even less research on the impact of malnutrition on spinal alignment.
Methods
We included 237 participants from a 2-yearly resident health checkup conducted in Toei, Aichi, Japan. Participants underwent blood tests and whole-spine standing radiography, and were stratified based on age (60s, 70s, and 80s) and sex to determine age-specific normative PNI values. Additionally, participants were categorized into a lower PNI (PNI <50) or higher PNI (PNI ≥50) group to compare spinal alignment.
Results
The average PNI values for different age groups were: 60s: males (n=13): 50.7, females (n=31): 50.9; 70s: males (n=55): 50.3, females (n=57): 50.1; 80s: males (n=28): 49.1, females (n=53): 48.3. For females, the radiographic spinal alignment parameters were comparable between the lower and higher PNI groups; however, in males, significant differences were noted for pelvic tilt (20° vs. 16°, p=0.020), lumbar lordosis (35° vs. 44°, p<0.001), and pelvic incidence minus lumbar lordosis (10° vs. 4°, p=0.013).
Conclusions
Malnutrition in males negatively impacts their lumbar-pelvic alignment. While the normative PNI value decreases with age, the two variables show a very weak correlation.
Introduction
In recent years, nutritional status has gained considerable attention, resulting in the development of multiple evaluation indices [1,2]. The Global Leadership Initiative on Malnutrition (GLIM) criteria proposed by the GLIM in 2019 is one of the most recognized nutritional assessment tools [3]. However, the GLIM criteria include questionnaire-based items that can be significantly influenced by the patient’s subjective responses [4], resulting in concerns over its validity as a screening tool [1,2]. Alternatively, indices, such as the prognostic nutritional index (PNI) and the Controlling Nutritional Status (CONUT), are based on blood tests and provide an objective assessment of nutritional status [5,6]. However, Evans et al. [7] argued that albumin, a chief component of these indices, decreases because of disease, inflammation, or infection, rendering it an unsuitable nutritional marker. They cited reports which indicated that even in healthy individuals, albumin and prealbumin levels do not decrease unless the body mass index (BMI) is below 12 kg/m2 or the individual has been starving for over 6 weeks [8]. Although albumin levels indeed decrease in patients with chronic diseases, such as rheumatoid arthritis, or infectious diseases, such as pneumonia, we also encountered healthy individuals with varying albumin levels potentially due to age, sex, or other individual-level factors.
The PNI has been commonly used as a prognostic tool in various fields. Oe et al. [9] used the PNI as a nutritional indicator in patients undergoing surgery for adult spinal deformity (ASD) and defined malnutrition as a PNI <50 using receiver operating characteristic curve analysis. In a subsequent study, they also reported that patients with a PNI <50 had a significantly higher incidence of postoperative medical complications (49.2% vs. 22.8%) [10]. Conversely, preoperative nutritional interventions with supplements and counseling helped maintain the total lymphocyte count, PNI, and prealbumin levels in patients with ASD compared with the non-intervention groups, where these markers declined [11].
Nutritional status may decline with age; however, only a few reports have provided the normative values of PNI in healthy elderly populations, with even less evidence on the association between malnutrition and deterioration of spinal alignment. Accordingly, this study aimed to investigate age-specific normative values of the PNI and the relationship between nutritional status and spinal alignment in elderly residents from Japan using prospective data from community health checkups.
Materials and Methods
Ethical considerations
The study protocol was approved by the Institutional Review Board (IRB) of Hamamatsu University School of Medicine, Shizuoka, Japan (IRB approval no., 23-129). The informed consent in this study was prospectively obtained from all participants.
Health screening study
Since 2012, our research group has conducted a health screening study (TOEI study) every 2 years in the Toei town of Aichi prefecture, Japan; the participants include volunteers recruited through the Toei Clinic and public relations in Toei. In 2022, 267 individuals participated in the musculoskeletal examination of the Toei study. For the present study, we included individuals aged 60–89 years old who had participated in the 2022 study. Participants with missing blood test data values, Cobb angle ≥25°, wedge-shaped deformity of the vertebral body (grade 3 in the semiquantitative method) [12], and those undergoing total joint arthroplasty and instrumented spinal surgery were excluded.
Measured data and radiographic parameters
The following data were recorded for all patients: age, sex, height, weight, BMI, body composition (amount of muscle, body fat percentage, and basal metabolic rate) assessed using a bioelectrical impedance analysis device (Multi-Frequency Segmental Body Composition Analyzer MC-780A-N; TANITA Corp., Tokyo, Japan), smoking habit, bone mineral density (expressed as a percentage of the young adult mean [%YAM] in the total proximal femur using dual-energy X-ray absorptiometry), grip strength, time taken to complete the timed up and go test (TUG), 10-m walking speed, presence of exercise habits, modified frailty index-11 (mFI-11), and patient-reported outcome measures evaluated using the EuroQol 5 Dimensions and Oswestry Disability Index.
The following parameters were recorded from blood samples collected using XS-1000i (Sysmex, Kobe, Japan): blood sugar (mg/dL), hemoglobin A1C (HbA1C, %), aspartate aminotransferase (AST, U/L), alanine aminotransferase (ALT, U/L), gamma-glutamyl transpeptidase (γGTP, U/L), creatinine (Cr, mg/dL), estimated glomerular filtration rate (eGFR, mL/min/1.73 m2), albumin (ALB, g/dL), total cholesterol (TC, mg/dL), red blood cell count (RBC, 10,000 cells/μL), hemoglobin (Hb, g/dL), platelet count (Plt, 100,000 cells/μL), white blood cell count (cell/μL), lymphocyte count (LY, cells/μL), and PNI and CONUT values. The PNI value was calculated as: PNI=10×ALB (g/dL)+0.005×total lymphocyte count (/μL).
For CONUT, the following scores were allotted based on serum ALB, TC, and LY, and a sum of these scores was computed: (1) ALB: ≥3.2 g/dL: 0 points, 3.00–3.49 g/dL: 2 points, 2.50–2.99 g/dL: 4 points, <2.50 g/dL: 6 points. (2) TC: ≥180 mg/dL: 0 points, 140–179 mg/dL:1 point, 100–139 mg/dL: 2 points, <100 mg/dL: 3 points. (3) Total LY: ≥1,600 cells/μL: 0 points, 1,200–1,599 cells/μL: 1 point, 800–1,199 cells/μL: 2 points, <800 cells/μL: 3 points.
Total CONUT scores of 0–1 indicated normal nutritional status, while total scores of 2–4, 5–8, and 9–12 points indicated mildly, moderately, and severely malnourished, respectively.
Radiographic parameters were evaluated using whole-spine standing radiographs. The participants were instructed to place their hands on their clavicles and look at their eyes in a mirror positioned 1 m in front of them. The following radiographic parameters were measured on the radiographs: (1) coronal Cobb angle, (2) sacral slope (SS), (3) pelvic tilt (PT), (4) pelvic incidence (PI), (5) lumbar lordosis (LL; L1–L5), (6) pelvic incidence (PI) minus LL (PI–LL), (7) thoracic kyphosis (TK; T5–T12), (8) T1 slope (TS), (9) cervical lordosis (CL), (10) C2–C7 sagittal vertical axis (SVA), and (11) C7 SVA. These parameters were measured by eight spine surgeons in our hospital using Surgimap Spine software (Nemaris Inc., New York, NY, USA).
To determine normative values for PNI, data were categorized for different age groups (60s, 70s, and 80s) and compared between males and females. The relationship between nutritional status and spinal alignment was examined by classifying participants into two groups based on the PNI: the lower PNI group (PNI <50) and the higher PNI group (PNI ≥50); further subgroup analyses were conducted separately for males and females.
Statistical analysis
All statistical analyses were performed using IBM SPSS ver. 25.0 (IBM Corp., Armonk, NY, USA); statistical significance was set at p<0.05. Continuous variables between the two PNI-based groups were compared using independent samples t-tests, whereas analysis of variance followed by post hoc analysis with Tukey’s test was used to compare continuous variables among the three age groups. Categorical variables were evaluated using the chi-square test or Fisher’s exact test. Age was adjusted via propensity score matching.
Results
Comparison by age group
Out of the 267 participants, 18 were excluded due to insufficient blood test data, eight were excluded because they were under 60 years of age, and four were excluded because they were over 90 years of age (Fig. 1); accordingly, 237 participants were included in the final analysis.
Flowchart outlining the eligibility criteria and participant registration process.
Table 1 presents the distribution and characteristics of each age group according to sex. There were 13 males and 31 females in the 60s age group, 55 males and 57 females in the 70s age group, and 28 males and 53 females in the 80s age. The three age groups showed statistically significant differences in terms of height, weight, muscle mass, and basal metabolic rate.
The characteristics data each generation
The results of the physical function tests and questionnaires are presented in Table 2. Grip strength decreased with age in both males and females. Regarding the TUG test, males in the 80s age group were significantly slower compared to males in their 60s (p=0.030); similarly, females in their 80s were significantly slower compared with the other age groups (p=0.000). In terms of 10-m walking speed, there was no statistically significant difference among the three age groups for males. Conversely, females in their 80s had a significantly slower walking speed compared to the other age groups (p=0.000).
Motor function and questionnaire results
Table 3 presents a summary of the blood test data. In terms of ALB levels, males exhibited no significant differences between the three age groups, whereas, in females, ALB levels decreased with age, with a significant difference between participants in their 60s versus the 80s (p=0.016). No significant differences were observed between males and females across all age groups. Likewise, TC levels showed no significant differences between the age groups in either males or females. However, within each age group, males had significantly lower TC levels than females. For LY, the cell counts tended to decrease with age in males, however, the difference was not statistically significant. In females, LY counts were comparable for the 60s and 70s age groups but tended to be lower in the 80s age group. No significant differences were noted between males and females across all age groups.
Blood sample data
The PNI value decreased with age in males, i.e., 50.7 in the 60s, 50.3 in the 70s, and 49.1 in the 80s, but the differences were not statistically significant. However, in females, PNI values decreased significantly from 50.9 in the 60s age group to 48.3 in the 80s age group (p=0.018). No statistically significant differences were observed between males and females across all age groups.
For the CONUT score, both males and females showed a statistically significant increase with age (p=0.028 and 0.022, respectively); however, like PNI, there were statistically significant differences between males and females across all age groups.
Table 4 presents a summary of the radiographic spinal alignment parameters for each age group. The intraclass correlation coefficients for intraobserver reliability of measuring PT, PI, SVA, and C2–7 SVA were 0.967, 0.995, 0.996, and 0.918, respectively. In males, statistically significant differences were noted in the C2–7 SVA and SVA values across the three age groups. Likewise, for females, statistically significant differences were observed for the PT, PI–LL, TS, CL, and SVA values across the three age groups.
Radiographic alignment
Fig. 2 presents a scatter plot of the correlation/regression analysis illustrating the relationship between the PNI and age. Accordingly, the regression equations were as follows: for males, PNI=62.7–0.167×age (R=0.212) and for females, PNI=62.3–0.167×age (R=0.263).
Correlation and regression analysis between age and the prognostic nutritional index (PNI). A very weak correlation is noted between age and PNI. Furthermore, there is minimal difference between males and females. Female: R=0.263, p=0.002, PNI=62.3–0.167×age. Male: R=0.212, p=0.038, PNI=62.7–0.167×age.
These equations indicate a very weak correlation between PNI and age, with minimal difference between males and females. Additionally, a PNI value of 50 corresponds to 75 years of age.
Comparison based on nutritional status (lower PNI versus higher PNI)
The lower PNI group comprised 49 males and 78 females, whereas the higher PNI group comprised 47 males and 63 females. The prevalence and breakdown of different PNI values are illustrated in Fig. 3. Individuals with a higher PNI were significantly less likely to have comorbidities compared with those with lower PNI. Among females with a lower PNI, the prevalence of hypertension and hyperlipidemia was notably higher. However, the prevalence of osteoporosis was high among all females regardless of their PNI status. Among males, participants in the lower PNI group tended to be of older age (p=0.056); likewise, in females, a significant difference was observed between the two groups in terms of age (p=0.014).
The prevalence and breakdown of comorbidity. Individuals with prognostic nutritional index (PNI) ≥50 were significantly less likely to have comorbidities compared to those with PNI <50. Among females with PNI <50, the prevalence of hypertension and hyperlipidemia was notably higher. The prevalence of osteoporosis was high among women regardless of their PNI status. d., desease. *p<0.05.
Regarding radiographic parameters of spinal alignment, there were no statistically significant differences between females with lower versus higher PNI for any spinal parameter. However, in males, the lower PNI group demonstrated significantly higher values of SS (p=0.000), PT (p=0.020), and PI–LL (p=0.013), and significantly lower values of LL (p=0.000) (Table 5).
Comparison between lower PNI and higher PNI group
To exclude the influence of age, propensity score matching was used (Table 6). According to this, there were 37 males and 55 females in each group. For spinal alignment in males, the propensity-score-matched lower PNI group had significantly higher values for SS (p=0.000), PT (p=0.015), and PI–LL (p=0.004), and significantly lower values for LL (p=0.000).
Comparison between lower PNI and higher PNI group after age adjustment
Discussion
While questionnaire-based tools, such as the GLIM criteria, are generally preferred among the numerous evaluation methods for nutritional status, survey-based assessment is prone to being influenced by the subjectivity of both the patient and the evaluator [4]. This subjectivity is a significant drawback, particularly for retrospective evaluations. Conversely, blood test-based evaluation methods, such as the PNI, offer greater objectivity and can be utilized for retrospective assessments as long as blood test data are available. However, blood test parameters like albumin or prealbumin levels tend to decrease in chronic inflammatory diseases, infections, or trauma, resulting in decreased validity of these evaluation measures [7]. Thus, these methods are not suitable for evaluating the nutritional status of patients with chronic conditions. As demonstrated in this study, albumin levels even vary among healthy individuals; therefore, in the absence of chronic or acute inflammatory diseases or trauma, it is reasonable to assume that albumin levels reflect the nutritional status of healthy individuals, which is consistent with previous concepts [13]. Previous reports indicating that a decrease in PNI is effective in predicting postoperative complications are not limited to the field of gastrointestinal surgery but are also recognized in numerous other medical disciplines [9,13–16]. Thus, clarifying the normal values of the PNI is important because it is a useful prognostic evaluation indicator.
In our study, there were no statistically significant differences in the ALB levels or LY counts between males and females across all age groups (Table 3). The average PNI for both sexes was 51 in their 60s, 50 for those in their 70s, and 49 for males and 48 for females in their 80s. Although the difference was not statistically significant, the PNI declined gradually in males across the age groups. A similar decrease was seen for females, with a significant difference between the PNI for the 60s versus 80s age group. It is generally believed that nutritional status declines with age; however, this study revealed that while the PNI tends to decrease with age, the correlation between age and PNI was very weak, (Fig. 2), with no significant difference between the two sexes.
Regarding the comparison between the low PNI and high PNI groups, the age-wise difference was not statistically significant in males but was significantly higher in females in the low PNI group. For spinal alignment, no significant differences were observed in females, but males in the low PNI group showed significant deterioration in the lumbopelvic parameters, namely SS, PT, LL, and PI–LL (Tables 5, 6). It is well-known that age strongly influences physical function and spinal alignment [17,18]; therefore, even after adjusting for age, the low PNI group showed significantly worse outcomes in terms of walking speed and lumbopelvic alignment compared to the high PNI group. The relationship between nutritional status and physical function has been explored in large-scale randomized controlled trials [19]; however, only a few reports have elucidated the relationship between spinal alignment and physical functioning. For patients with ASD, Wang et al. [20] reported that those with malnutrition exhibited poorer spinal alignment parameters, such as TS and SVA. Additionally, the study reported that frailty, which is closely associated with undernutrition, significantly influenced the deterioration of global spinal alignment parameters, such as SVA; this effect was not limited to patients with ASD and was observed even among healthy participants undergoing regular health checkups [21]. Similarly, Takasawa et al. [22] reported that among patients with cervical spondylotic myelopathy, those with malnutrition were more prone to postoperative cervical kyphosis. However, the present study did not find any association between nutritional status and spinal alignment in females. The rationale for this difference remains unclear; presumably, the mechanisms underlying spinal alignment deterioration may differ between males and females.
This study had some limitations. First, the sample size was relatively small, particularly for males in their 60s, which may have affected the statistical power of the analysis, especially when comparing different age groups. Second, although this study targeted healthy individuals undergoing community health checkups, we did not exclude confounding factors, such as diabetes, gastrointestinal disorders, mental health conditions, respiratory diseases, and autoimmune disorders, which could potentially influence nutritional status. These factors are extremely relevant and may have impacted the statistical relationship between spinal alignment and PNI. Third, the reference values for PNI used in this study may not be universally applicable to all facilities and may vary among countries, regions, and ethnicities; therefore, further investigation is required to determine normative PNI values for different populations. Finally, some reports suggest that blood sample data may vary depending on the measurement equipment used [23], warranting further investigation in future studies.
Conclusions
As a nutritional indicator, PNI demonstrates an age-related decline, a trend observed consistently in both males and females. The average PNI values were 51 for individuals in their 60s, 50 for those in their 70s, and 48–49 for participants in their 80s. A PNI value of 50 corresponds to an age of approximately 75 years. It is noteworthy that the correlation between PNI and age was very weak for both sexes. Furthermore, while there was no clear association between spinal alignment and nutritional status in females, males with a PNI <50 exhibited poorer lumbar-pelvic alignment, suggesting a potential relationship between spinal alignment and nutritional status in males.
Key Points
The normative values of prognostic nutritional index (PNI) were approximately 51 for residents in their 60s, 50 in. their 70s, and 48–49 in their 80s.
There is little difference in the normal range of PNI between males and females.
A PNI value of 50 corresponds to an approximate age of 75 years.
Male residents with malnutrition (PNI <50) experience deterioration in radiographic sagittal spinal alignment compared to those with a PNI ≥50.
Notes
Conflict of Interest
Shin Oe and Tomohiko Hasegawa are members of the Division of Geriatric Musculoskeletal Health which is funded by a donor. Otherwise, no potential conflict of interest relevant to this article was reported.
Acknowledgments
We appreciate the help of Yoko Suzuki, Naomi Uchiyama, Nao Kuwahara, and Chieko Suzuki for case collections or manuscript discussions. The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.
Funding
Source of funding as follows: Medtronic Sofamor Danek Inc., Japan Medical Dynamic Marketing Inc., Meitoku Medical Institution Jyuzen Memorial Hospital. We have not received funding from the NIH or HHMI.
Author Contributions
SO was responsible for the study’s conception and design. KI, TH, GY, TB, HA, TY, and KK acquired, analysed, and interpreted data, drafted the article. YM and YY approved the final version on behalf of all authors. All authors have critically revised the article and reviewed the submitted version.