HTML
-
The different types of cancers and the percentage of men and women in each tumor group are summarized in Figure 1. Stomach cancer, accounting for more than 50% of the cases, was the most common tumor among the studied patients. In addition, breast and stomach cancer ranked first in female and male patients, respectively.
The baseline clinical characteristics are shown in Table 1. Our study consisted of 312 oncology patients, and there were 163 women (median age: 55 years, age range: 28-85) and 149 men (median age: 64 years, age range: 24-90). The overall patient albumin level was lower than the normal value. It is noteworthy that the educational level of 204 (65.38%) of the 312 patients was primary school, and only 5 (1.60%) had attended university. The data indicate that more than half of the patients did not accept nutritional support, which was possibly related to educational status.
Items Male (n = 149) Female (n = 163) Total (n = 312) P Value Age, y 62.77 ± 12.00 55.15 ± 10.40 58.79 ± 11.80 < 0.001 64 (24, 90) 55 (28, 85) 59 (24, 90) < 0.001 BMI, kg/m2 21.0 ± 3.5 23.2 ± 4.1 22.2 ± 4.0 < 0.001 20.6 (14.4, 33.2) 22.9 (14.1, 37.4) 22.0 (14.1, 37.4) < 0.001 Ablumin, g/L 24.10 ± 4.80 26.20 ± 4.41 25.20 ± 4.71 < 0.001 24 (5, 45) 26 (6, 39) 25 (5, 45) < 0.001 TSF, mm 13.7 ± 7.6 18.7 ± 8.4 16.3 ± 8.4 < 0.001 11 (3, 50) 18 (1.7, 45) 15 (1.7, 50) < 0.001 Total protein, g/L 20.5 ± 4.9 21.9 ± 10.9 21.2 ± 8.6 0.155 20.30 (4.10, 35.58) 20.86 (4.70, 149.20) 20.60 (4.10, 149.20) 0.246 Education [n(%)] 0.065 Primary school 88 (59.06) 116 (71.17) 204 (65.38) Junior high school 41 (27.52) 27 (16.56) 68 (21.79) High school 16 (10.74) 19 (11.66) 35 (11.22) University 4 (2.68) 1 (0.61) 5 (1.60) Tumor diagnosis [n(%)] Stomach cancer 76 (51.01) 17 (10.43) 93 (29.81) Esophageal cancer 44 (29.53) 17 (10.43) 61 (19.55) Breast cancer 0 (0) 46 (28.22) 46 (14.74) Cervical cancer 0 (0) 42 (25.77) 42 (13.46) Colorectal cancer 13 (8.72) 24 (14.72) 37 (11.86) Stomach + Nasopharyngeal cancer 7 (4.70) 2 (1.23) 9 (2.88) lung cancer 4 (2.68) 2 (1.23) 6 (1.92) Nasopharyngeal cancer 2 (1.34) 3 (1.84) 5 (1.60) Pancreatic cancer 0 (0) 2 (1.23) 2 (0.64) Lymphoma 2 (1.34) 0 (0) 2 (0.64) Endometrial cancer 0 (0) 1 (0.61) 1 (0.32) Stomach + Cervical cancer 0 (0) 1 (0.61) 1 (0.32) Stomach + Colorectal cancer 1 (0.67) 0 (0) 1 (0.32) Note. BMI: Body mass index; TSF: Triceps skinfold thickness; P < 0.05, determined by Kruskal-Wallis nonparametric test. Table 1. Baseline and Clinical Characteristics of Participants
There was concordance between albumin level and the PG-SGA and NRS-2002 screening tool parameters (Table 2). The PG-SGA score had a higher sensitivity (93.73%) and lower specificity (2.30%) than the NRS-2002 (69.31% and 25.00%, respectively). When the two measurement methods were compared to albumin, there was better agreement between albumin and the NRS-2002 (κ = 0.004, P = 0.0005) than between albumin and the PG-SGA (κ = 0.003, P = 0.0006).
Items PG-SGA NRS-2002 At Risk (B+C) No Risk (A) Total At Risk (NRS-2002 ≥ 3) No Risk (NRS-2002 < 3) Total At risk (albumin < 35 g/L) 284 19 303 210 93 303 No risk (albumin ≥ 35 g/L) 8 0 8 6 2 8 Total 292 19 311 216 95 311 Sensitivity (%) 93.73 (284/303) 69.31 (210/303) Specificity (%) 2.30 (0/8) 25.00 (2/8) κ = 0.003, P = 0.0005 κ = 0.004, P = 0.0006 Note. PG-SGA: Patient Generated Subjective Global Assessment; NRS-2002: Nutritional Risk Screening; P Values were determined by Chi-square test. Table 2. Comparisons of Albumin and Evaluating Tools: PG-SGA and NRS-2002 vs. Albumin
Table 3 shows the low negative correlation between PG-SGA and NRS-2002 scores among the enrolled patients. The PG-SGA better correlated with BMI and TFS than did the NRS-2002 for both male and female patients. However, these results were the opposite of the albumin data. There was a weak positive correlation with albumin for both the PG-SGA and NRS-2002 among women. Furthermore, there was a substantial similarity between the PG-SGA and NRS-2002 for the evaluation of nutritional status.
Items PG-SGA [A/(B+C)] NSR-2002 (< 3/≥ 3) Male Female Male Female Correlation P Value Correlation P Value Correlation P Value Correlation P Value BMI, kg/m2 0.230 0.005 0.259 0.001 0.304 < 0.001 0.301 < 0.001 TSF, mm 0.214 0.009 0.276 < 0.001 0.289 < 0.001 0.316 < 0.001 Total protein, g/L 0.059 0.471 0.172 0.029 0.029 0.724 0.107 0.176 Albumin, g/L 0.070 0.395 0.326 < 0.001 0.065 0.431 0.319 < 0.001 NRS-2002 score -0.240 0.003 -0.410 < 0.001 Note. PG-SGA: Patient Generated Subjective Global Assessment; NRS-2002: Nutritional Risk Screening; BMI: Body mass index; TSF: Triceps skinfold thickness; P value was determined with Spearman correlation analysis. Table 3. P Value and Correlation Co-effcients from Data and Screening Tools according to Sex
We used the risk and no-risk end results (PG-SGA and NRS-2002) to analyze the factors affecting outcome. The factors evaluated included sex, age, BMI, albumin, total protein, and TFS. We utilized a single factor logistic regression analysis (P < 0.05) and a multifactor logistic regression analysis (forward LR method). The results indicated that only BMI and TFS had a significant difference for both the PG-SGA and NRS-2002 (PPG-SGA < 0.001, PNRS-2002 < 0.001, Table 4).
Items Univariate Analysis Multivariate Analysis OR (95% CI) P Value OR (95% CI) P Value PG-SGA Gender (male as reference) 0.189 (0.054-0.662) 0.009 Age, y 1.057 (1.014-1.101) 0.009 BMI, kg/m2 0.777 (0.694-0.871) < 0.001 0.845 (0.741-0.964) 0.012 Albumin, g/L 0.830 (0.751-0.918) < 0.001 TSF, mm 0.894 (0.851-0.939) < 0.001 0.927 (0.876-0.981) 0.009 Total protein, g/L 0.990 (0.954-1.027) 0.591 NRS-2002 Gender (male as reference) 0.380 (0.229-0.632) < 0.001 Age, y 1.044 (1.021-1.068) < 0.001 BMI, kg/m2 0.830 (0.775-0.888) < 0.001 0.904 (0.836-0.978) 0.012 Albumin, g/L 0.888 (0.837-0.942) < 0.001 TSF, mm 0.906 (0.877-0.936) < 0.001 0.926 (0.892-0.961) < 0.001 Total protein, g/L 0.998 (0.971-1.026) 0.910 Note. PG-SGA: Patient Generated Subjective Global Assessment; NRS-2002: Nutritional Risk Screening; BMI: Body mass index; TSF: Triceps skinfold thickness; P value was determined with Logistics regression analysis. Table 4. Analyze Factors to Affect Outcomes According to Logistics Regression Analysis
The PG-SGA score in Figure 2 shows that the prevalence of malnutrition among the studied oncology patient population reached 94% (SGA-B and SGA-C). The incidence of malnutrition in cancer patients is usually directly related to the tumor, causing abnormal body metabolism or adverse physical reactions after receiving a cancer diagnosis. These reactions may include depression, anxiety, and loss of appetite.
The sex, age, PG-SGA score, and BMI for SGA classifications are shown in Table 5. The SGA-C group contained the majority of the total population, and the number of men was greater than that of women, which is consistent with Figure 2. The SGA-C group is defined as severely malnourished patients with the highest scores, and there was a significant difference in the median PG-SGA score for each of the SGA classifications (P < 0.001). In addition, the SGA-C group had the lowest BMI and the median age was 63 years. These results led us to focus on malnutrition specifically in the elderly population. These four indicators were all statistically significant (P < 0.001).
PG-SGA SGA-A SGA-B SGA-C P Value N (%) 19 (6.1) 126 (40.4) 167 (53.5) Gender (M/F) 3/16 45/81 101/66 < 0.001 Age, y (x̅ ± s; median, IQR) 51.84 ± 10.28 56.46 ± 11.92 61.33 ± 11.29 < 0.001 52 (36, 69) 55 (28, 112) 63 (24, 90) < 0.001 SGA score (median, IQR) 1 (1, 1) 6 (2, 8) 12 (9, 22) < 0.001 BMI, kg/m2 (x̅ ± s; median, IQR) 26.40 ± 3.02 23.47 ± 4.01 20.70 ± 3.29 < 0.001 25.8 (22.90, 33.20) 23.0 (15.62, 37.40) 20.4 (14.10, 31.30) < 0.001 Note. PG-SGA: Patient Generated Subjective Global Assessment; IQR: Interquartile range; BMI: Body mass index; SGA-A rating in well-nourished patients, SGA-B rating in moderately malnourished patients, SGA-C rating in severely malnourished patients; P value was determined by ANOVA variance analysis. Table 5. Comparisons of Clinical Indicators According to PG-SGA
We also comprehensively assessed patient malnutrition using the EORTC QLQ-C30 to determine quality of life. As shown in Table 6, our study concentrated on various life changes due to cancer, such as memory, economic status, depression, and fatigue. We investigated 11 separate items and found that the majority of the target population had 2 points for each item. This result suggests that the quality of life of cancer patients is affected by their disease. These findings are consistent with the PG-SGA scores.
Items Number of Patients [n (%)] Total patients 310 (100.00) Tired 1 77 (24.84) 2 186 (60.00) 3 43 (13.87) 4 4 (1.29) Affect daily life 1 75 (24.19) 2 177 (57.10) 3 53 (17.10) 4 5 (1.61) Have difficulty concentrating on doing things 1 71 (22.90) 2 198 (63.87) 3 38 (12.26) 4 2 (0.65) Nervous 1 84 (27.10) 2 199 (64.19) 3 25 (8.06) 4 2 (0.65) Worried 1 80 (25.81) 2 199 (64.19) 3 28 (9.03) 4 3 (0.97) Bad temper 1 90 (29.03) 2 181 (58.39) 3 36 (11.61) 4 3 (0.97) Repressed 1 67 (21.61) 2 206 (66.45) 3 33 (10.65) 4 4 (1.29) Memory difficult 1 76 (24.51) 2 190 (61.29) 3 42 (13.55) 4 2 (0.65) Family life affected by physical condition 1 35 (11.29) 2 189 (60.97) 3 77 (24.84) 4 9 (2.90) Social activity affected by Physical condition 1 33 (10.65) 2 184 (59.36) 3 84 (27.10) 4 9 (2.90) Economy affected by Physical condition 1 22 (7.10) 2 147 (47.42) 3 105 (33.87) 4 31 (10.00) 5 5 (1.61) Note. EORTC QLQ-C30: European organization for research and treatment of cancer quality of life core questionnaire 30. Table 6. Life Quality of Patients by Assessment Tool