TABLE 1.
Characteristics of the studies included and findings on prognostic potential (according to multivariate analysis) and efficacy of the CONUT score in predicting survival1
| Reference | Country, study design | Type of cancer | Patients, n | Age, y | Follow-up time | CONUT score cut offs | Determination of optimal cutoff values for CONUT score/index | CONUT score efficacy for predicting survival | Calculation method of survival rate, HR, and CONUT score2 | Findings | Relation of CONUT score and BMI |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Ahiko et al. (16) | Japan, retrospective cohort | Colorectal cancer (Stage I–IV), underwent surgery | 830 (470 males, 360 females) | Median (range): 78 (75–94) | ≥5 y | 0–1 (n = 508)2–3 (n = 249)≥4 (n = 73) | Not indicated. | Not indicated. | *¶◊ | Higher CONUT group had the lowest 5-y OS rate (P < 0.0001).CONUT score was a significant prognostic factor of OS (≥4 vs. 0–1: HR: 2.24; 95% CI: 1.48, 3.30; P < 0.001) after adjustment for confounders. | BMI did not differ between CONUT groups. |
| Akamine et al. (17) | Japan, retrospective | Lung adenocarcinoma, underwent surgery | 109 (76 males, 33 females) | Mean (range): 72 (45–85) | 4–13 y | High: ≥1 (n = 74)Low: 0 (n = 35) | ROC curve/— | AUC: 0.596 Sensitivity: 0.6711; specificity: 0.4375 | *¶◊ | High-CONUT group had lower 5-y OS (P = 0.04) and DFS (P = 0.01) rates.CONUT score was independently associated with DFS (HR: 2.63; 95% CI: 1.33, 5.68; P = 0.004) and OS (HR: 2.64; 95% CI: 1.06, 7.80; P = 0.04). | High-CONUT group had lower BMI (P = 0.025). |
| Elghiaty et al. (12) | South Korea, retrospective cohort | Nonmetastatic clear cell renal cell carcinoma, underwent radical or partial nephrectomy | 1046 (745 males and 301 females) | Median (range): 56 (46–64) | Median (range): 63 (43–87) mo | High: >2 (n = 115)Low: ≤2 (n = 931) | ROC curve/Youden Index | AUC (based on OS) = 0.633 Sensitivity = 46.4%; specificity = 73.7% (P = 0.001) | *¶◊ | High-CONUT group had lower 3- and 5-y RFS (P < 0.001), CSS (P = 0.006), and OS (P < 0.001) rates.High CONUT score was an independent predictor of RFS (HR: 3.09; 95% CI: 1.45, 6.59; P = 0.003), CSS (HR: 4.66; 95% CI: 1.62, 13.39; P = 0.004), and OS (HR: 2.81; 95% CI: (1.44, 5.50; P = 0.003). | High-CONUT group had lower BMI (P = 0.001). |
| Harimoto et al. (18) | Japan, retrospective cohort | Hepatocellular carcinoma, underwent hepatic resection | 2461 (1785 males, 676 females) | Mean ± SD: Low CONUT: 68.2 ± 10.1; High CONUT: 69.8 ± 9.2 | Not indicated | High: ≥4 (n = 540)Low: ≤3 (n = 1921) | ROC curve/Youden index | AUC (based on OS): 0.580 Sensitivity: 31.3%; specificity: 91.6% (P < 0.01) AUC (based on CSS): 0.563Sensitivity: 30.0%; specificity: 80.1% (P < 0.01) AUC (based on RFS): 0.536 Sensitivity: 63.5%; specificity: 41.9% (P < 0.01). | *¶◊ | High-CONUT group had lower OS and RFS rates (both, P < 0.01).Higher CONUT score was an independent predictor of poor OS (HR: 1.22; 95% CI: 1.06, 1.41; P = 0.006) and RFS (HR: 1.22; 95% CI: 1.06, 1.40; P = 0.006). | High-CONUT group had low BMI (P < 0.01). |
| Harimoto et al. (19) | Japan, retrospective | Hepatocellular carcinoma, underwent hepatic resection | 357 (270 males, 87 females) | Mean ± SD: Low CONUT: 67.3 ± 10.7; High CONUT: 69.8 ± 8.5 | ≥5 y | High: >3 (n = 69)Low: ≤3 (n = 288) | ROC curve/Youden index | AUC (based on OS): 0.621Sensitivity: 56.06%; specificity: 66.56% AUC (based on DSS): 0.651 Sensitivity: 65.7%; specificity: 59.5%. | *¶◊ | High-CONUT group had lower 5-y OS, RFS, and DSS rates (P < 0.01).Higher CONUT score was associated with poor OS (HR: 2.16; 95% CI: 1.25, 3.72; P = 0.03), but not with RFS. | BMI did not differ between CONUT groups. |
| Hirahara et al. (20) | Japan, retrospective | Esophageal cancer, underwent curative thoracoscopic esophagectomy | 148 (132 males, 16 females) | Mean ± SD: CONUT 0: 65.6 ± 8.2; CONUT 1: 67.5 ± 8.4; CONUT 2-3: 65.3 ± 8.9 | Not indicated | Normal nutrition: 0–1 (n = 70)Mild malnutrition: 2–4 (n = 62)Moderate–severe malnutrition: 5–12 (n = 16) | Not indicated. | Not indicated. | *¶◊ | CONUT was independently associated with worse prognosis for CSS (HR: 1.99; 95% CI: 1.07, 3.87), P = 0.03). | Not indicated. |
| Hirahara et al. (21) | Japan, retrospective cohort | Gastric cancer, underwent curative gastrectomy | 368 (254 males, 114 females) | Range: 36–91 | Median: 35.3 mo | High: ≥3 (n = 105)Low: ≤2 (n = 263) | ROC curve/— | AUC (based on 5-y OS): 0.625 Sensitivity: 65.0%; specificity: 57.9% | *¶◊ | High-CONUT group had lower 5-y OS rate (P < 0.001).Among all patients, OS was independently predicted by the CONUT score (HR: 2.25, P = 0.001).CONUT score was an independent prognostic factor for OS among the propensity score–matched subgroup (HR: 2.44; 95% CI: 1.46, 4.07; P < 0.001). | BMI did not differ between CONUT groups. |
| Huang et al. (22) | China, prospective cohort | Gastric cancer, underwent curative gastrectomy | 357 (275 males, 82 females) | Mean ± SD: 73.29 ± 5.24 | 1 y | Normal: 0–1 (n = 153)Light: 2–4 (n = 168)Moderate and severe: ≥5 (n = 36) | Not indicated. | Not indicated. | ¶◊ | CONUT score was an independent predictor of postoperative 1-y survival (OR: 2.91; 95% CI: 0.91, 9.31; P = 0.02). | Moderate–severe CONUT group had lower BMI (P < 0.001). |
| Iseki et al. (23) | Japan, retrospective | Colorectal cancer (Stage II/III), underwent curative surgery | 204 (112 males, 92 females) | Mean ± SD: High CONUT: 66.09 ± 9.23; Low CONUT: 71.13 ± 11.57 | 8 y or until their deaths | High: ≥3 (n = 54)Low: ≤2 (n = 150) | ROC curve/— | AUC (based on 5-y CSS): 0.624 (P = 0.076)Sensitivity: 0.5263; specificity: 0.7622 | *¶◊ | High-CONUT group had lower 5-y CSS (P = 0.002) and RFS (P = 0.002) rates.CONUT score was an independent risk factor for CSS (OR: 4.21; 95% CI: 1.21, 13.35; P = 0.02), but not for RFS. | Not indicated. |
| Ishihara et al. (24) | Japan, retrospective cohort | Localized urothelial carcinoma treated with radical nephroureterectomy | 107 (68 males, 39 females) | Mean ± SD: Low CONUT:72.7 ± 9.98; High CONUT: 76.1 ± 8.65 | Mean ± SD: 46.1 ± 32.8; 25.5 ± 18.4 mo | High: ≥3 (n = 24)Low: <3 (n = 83) | ROC curve/Youden index | AUC (based on RFS): 0.588 | *¶◊ | High-CONUT group had lower 5-y RFS (P = 0.04), CSS (P = 0.004), and OS (P = 0.01) rates. | Not indicated. |
| CONUT score was an independent predictor of CSS (HR: 5.44; 95% CI: 1.95, 14.8; P = 0.002), OS (HR: 2.90; 95% CI: 1.18, 6.75; P = 0.02), and RFS (HR: 2.26; 95% CI: 0.97, 4.94; P = 0.058). | |||||||||||
| Kang et al. (25) | Korea, retrospective cohort | Renal cell carcinoma, underwent surgery | 1881 (1361 males, 520 females) | Mean ± SD: Normal: 54.21 ± 12.17; Mild: 58.69 ± 12.80; Moderate to severe: 63.69 ± 12.83 | Median (range): 41 (6–178) mo | High: ≥2 (n = 508)Low: 0–1 (n = 1373) | ROC curve/— | ROC curve | *¶◊ | High-CONUT group had shorter RFS (P = 0.02) and CSM (P < 0.001).High CONUT score was an independent predictor of CSM (HR: 1.89; 95% CI: 1.12, 3.20; P = 0.02), but not for RFS. | High-CONUT group had lower BMI (P < 0.001). |
| Kato et al. (26) | Japan, retrospective | Pancreatic adenocarcinoma, underwent resection | 344 (207 males, 137 females) | Mean ± SD: 64.8 ± 9.9 | Median (range): 29.1 (0.6–178.5) mo | High: ≥4 (n = 79)Low: <4 (n = 265) | ROC curve/— | AUC (based on 2-y survival): 0.614 (95% CI: 0.56, 0.67).Sensitivity: 30.6%; specificity: 88.6% | *¶◊ | High-CONUT group showed lower OS (P = 0.002) but not RFS.High CONUT score was an independent prognostic risk factor for OS (HR: 1.64; 95% CI: 1.19, 2.26; P = 0.003). | Not indicated. |
| Kuroda et al. (27) | Japan, retrospective cohort | Gastric cancer, underwent curative resection | 416 (267 males, 149 females) | Median (range): 67.2 (25–94) | Median (range): 61.2 (1–134) mo | High: ≥4 (n = 62)Low: ≤3 (n = 354) | ROC curve/— | AUC (based on OS): 0.715 (95% CI: 0.68, 0.75)AUC (based on RFS): 0.658 (95% CI: 0.62, 0.70)AUC (based on CSS): 0.662 (95% CI: 0.61, 0.71) | *¶◊ | High-CONUT group had lower 5-y OS (P < 0.001), RFS (P = 0.02), and CSS (P = 0.02) rates.CONUT was an independent prognostic factor for OS (HR: 2.72; 95% CI: 1.74, 4.25; P < 0.001), but not for RFS and CSS. | High-CONUT group had lower BMI (P = 0.02). |
| Liang et al. (28) | China, retrospective cohort | Soft-tissue sarcomas, underwent surgical resection | 658 (393 males, 265 females) | Median (range): 43 (5–85) | Median (range): 103 (61–147) mo | High: ≥2 (n = 223)Low: 0–1 (n = 435) | ROC curve/Youden index | ROC curve | *¶◊ | High-CONUT group had lower 5-y OS (P < 0.001) and DFS (P < 0.001) rates.High CONUT was an independent predictor of OS (HR: 1.86; 95% CI: 1.47, 4.14; P < 0.001) and DFS (HR: 1.63; 95% CI: 1.26, 2.11; P < 0.001). | Not indicated. |
| Lin et al. (29) | China, retrospective cohort | Hepatocellular carcinoma, underwent curative hepatectomy | 380 (333 males, 47 females) | Median (range): 50 (19–80) | Median: 48.5 mo | High: ≥2 (n = 187)Low: <2 (n = 193) | ROC curve/— | AUC (based on OS): 0.618 (95% CI: 0.567, 0.667)Sensitivity: 66.3%; specificity: 56.5% | *¶◊ | High-CONUT group had lower 5-y OS (P < 0.001) and RFS (P = 0.02) rates. High CONUT was an independent prognostic indicator of decreased OS (HR: 2.40; 95% CI: 1.74, 4.25; P = 0.001), but not decreased RFS (HR: 1.36; 95% CI: 1.00, 1.85; P = 0.05). | Not indicated. |
| Miyata et al. (30) | Japan, retrospective | Intrahepatic cholangiocarcinoma, underwent curative hepatectomy | 71 (45 males, 26 females) | Mean ± SD: Low CONUT: 64.8 ± 1.7; High CONUT: 69.1 ± 1.9 | Mean: 36.9 mo | High: ≥2 (n = 31)Low: <2 (n = 40) | Not indicated. | Not indicated. | *¶◊ | High-CONUT group had lower 1-, 3-, and 5-y OS (P = 0.01), but not RFS.High CONUT was an independent prognostic factor for OS (HR: 3.02; 95% CI: 1.4, 6.8; P = 0.007), but not for RFS. | High-CONUT group had lower BMI (P = 0.009). |
| Ryo et al. (31) | Japan, retrospective cohort | Gastric cancer, underwent gastrectomy | 626 (435 males, 191 females) | Mean ± SD: 67.9 ± 10.9 | Median: 49.2 mo or until death | High: ≥2 (n = 289)Low: <2 (n = 337) | ROC curve/— | AUC (based on DFS): 0.656 Sensitivity: 0.66; specificity: 0.58 | *¶◊ | High-CONUT group had shorter OS (P < 0.0001) and DFS (P = 0.06) times. CONUT score was an independent prognostic factor for OS (HR: 1.74; 95% CI: 1.26, 2.41; P = 0.0007). | High-CONUT group had lower BMI (P < 0.0001). |
| Shoji et al. (32) | Japan, retrospective | Non–small cell lung cancer, underwent surgery | 138 (79 males, 59 females) | Mean (range): 68 (37–86) | Median (range): 58 (0–94) mo | High: ≥1 (n = 79)Low: 0 (n = 59) | ROC curve/— | AUC (based on CSS): 0.703Sensitivity = 91.67%; specificity = 46.07% | *¶◊ | High-CONUT group had lower 5-y RFS (P = 0.046), CSS, (P = 0.01), and OS (P = 0.01) rates. CONUT score was an independent prognostic factor for CSS (RR: 6.06; 95% CI: 1.07, 113.94; P = 0.04). | Not indicated. |
| Song et al. (33) | China, retrospective cohort | Nonmetastatic renal cell carcinoma, underwent surgery | 325 (231 males, 94 females) | Median (IQR): 57 (47–66) | Median (IQR): 64 (56.5–69) mo | High: ≥3 (n = 70)Low: <3 (n = 255) | ROC curve/Youden index | AUC (based on 5-y OS): 0.723, (P < 0.001)Sensitivity: 51.28%; specificity: 82.52% | *¶◊ | High-CONUT group had lower 5-y OS (P < 0.001), CSS (P < 0.001), and DFS (P < 0.001) rates. High CONUT was an independent risk factor for OS (HR: 3.36; 95% CI: 1.73, 6.56; P < 0.001), CSS (HR: 3.34; 95% CI: 1.59, 6.98; P = 0.001), and DFS (HR: 1.85; 95% CI: 1.07, 3.21; P = 0.03). | Not indicated. |
| Suzuki et al. (34) | Japan, retrospective | Gastric cancer, underwent curative resection | 211 (141 males, 70 females) | ≥75 | Median (range): 47 (5–185) mo | Normal nutrition: (n = 75) Light malnutrition: (n = 100) Moderate or severe malnutrition: (n = 36) | Not indicated. | Not indicated. | *¶◊ | Higher-CONUT group had shorter OS (P < 0.001) and CSS (P < 0.001).CONUT score was an independent prognostic factor for OS (HR: 2.12; 95% CI: 1.18, 3.69; P = 0.01) and CSS (HR: 3.75; 95% CI: 1.30, 10.43; P = 0.01). | Higher CONUT group had lower BMI (P = 0.008). |
| Takagi et al. (35) | Japan, retrospective cohort | Hepatocellular carcinoma, underwent hepatectomy | 295 (241 males, 54 females) | Mean ± SD: 65.8 ± 10.4 | Mean: 42.3 mo | High: ≥3 (n = 118)Low: ≤2 (n = 177) | ROC curve/— | AUC = 0.59 | *¶◊ | High-CONUT group had lower 5-y RFS (P = 0.01) and OS (P = 0.006) rates.The CONUT score was an independent predictor of RFS (HR: 1.64; 95% CI: 1.15, 2.30; P = 0.006) and OS (HR: 2.50; 95% CI: 1.47, 4.23; P = 0.001). | BMI did not significantly differ between CONUT groups. |
| Takagi et al. (36) | Japan, retrospective cohort | Hepatocellular carcinoma, underwent hepatectomy | 331 (269 males, 62 females) | Median (range): 67 (60–74) | 1 mo | High: ≥5 (n = 30)Low: ≤4 (n = 301) | Not indicated. | Not indicated. | ¶◊ | High-CONUT group had higher incidence of 30-d mortality (P < 0.001).High CONUT score was an independent predictor of in-hospital mortality after hepatectomy (HR: 9.41; 95% CI: 1.15, 77.4; P = 0.04). | BMI did not differ between CONUT groups (P > 0.05). |
| Takamori et al. (37) | Japan, retrospective | Malignant pleural mesothelioma | 83 (66 males, 17 females) | Median (range): 59 (31–81) | — | High: ≥3 (n = 31)Low: ≤2 (n = 52) | ROC curve/— | AUC (based on 1-y survival): 0.772Sensitivity: 73.1%; specificity: 64.0% | *¶◊ | High-CONUT group had lower OS and DFS rates (both, P < 0.001).High CONUT score was an independent predictive factor for OS (HR: 1.92; 95% CI: 1.17, 3.11; P = 0.01) and DFS (HR: 1.88; 95% CI: 1.14, 3.06; P = 0.01).High CONUT score was a prognostic factor of OS in patients who underwent surgery (HR: 4.86; 95% CI: 1.16, 19.14; P = 0.03). | Not indicated. |
| Tokunaga et al. (38) | Japan, retrospective | Colorectal cancer, underwent curative resection | 417 (247 males, 170 females) | Median (range): 68 (19–93) | Mean (range): 38.0 (1–115) mo | Normal: (n = 246)Light: (n = 127) Moderate: (n = 33) Severe: (n = 11) | Not indicated. | Not indicated. | *¶◊ | High-CONUT group (moderate/severe) had lower 5-y OS and RFS (both, P < 0.001).CONUT score was an independent prognostic factor for OS (moderate/severe vs. normal: HR: 5.92; 95% CI: 2.30, 14.92; P < 0.001; light vs. normal: HR: 2.74; 95% CI: 1.30, 5.87; P = 0.008), but not for RFS. | Moderate/severe CONUT group had lower BMI (P = 0.005). |
| Toyokawa et al. (39) | Japan, retrospective | Lung squamous cell carcinoma, underwent surgery | 108 (96 males, 12 females) | Median (range): 71 (45–89) | ≥5 y | High: ≥2 (n = 32)Low: 0–1 (n = 76) | ROC curve/— | AUC (based on OS): 0.590 Sensitivity: 0.786; specificity: 0.385 | *¶◊ | High-CONUT group had lower 5-y DFS (P = 0.02) and OS (P = 0.006) rates. High CONUT score was an independent prognostic factor for DFS (HR: 1.90; 95% CI: 1.04, 3.37; P = 0.04) and OS (HR: 1.91; 95% CI: 0.92, 3.86; P = 0.08). | BMI did not differ between CONUT groups (P > 0.05). |
| Wang et al. (40) | China, retrospective | Malignant peritoneal mesothelioma | 125 (46 males, 79 females) | Mean ± SD: 61.2 ± 9.3 | Median (range): 8 (0.6–53) mo | High: ≥3 (n = 81)Low: ≤2 (n = 44) | ROC curve/— | AUC (based on OS): 0.861 (P < 0.001)Sensitivity: 0.784; specificity: 0.821 | ¶◊ | CONUT score was an independent predictive factor for OS (RR: 1.26; 95% CI: 1.16, 1.38; P < 0.001). | BMI did not differ between CONUT groups (P > 0.05). |
| Yamamoto et al. (41) | Japan, retrospective | Colorectal cancer (stage I–IV) underwent surgery | 522 (291 males, 231 females) | — | ≥5 y | High: ≥3 (n = 158)Low: <3 (n = 364) | ROC curve/— | AUC (based on OS): 0.627 (P < 0.0001) | *¶◊ | High-CONUT group had a lower 5-y OS rate (P < 0.0001). | Not indicated. |
| Yang C et al. (42) | China, retrospective | Colorectal cancer, underwent curative resection | 160 (90 males, 70 females) | Mean ± SD: 58.4 ± 11.8 | Median (range): 30 (6–42) mo | High: ≥3 (n = 74)Low: <3 (n = 86) | ROC curve/Youden index | AUC (based on CSS): 0.759 (P < 0.001)Specificity: 0.821; sensitivity: 0.625 | *¶◊ | High CONUT score was correlated with poor RFS (P < 0.001) and CSS (P < 0.001).CONUT score was an independent prognostic factor for RFS (HR: 2.02; 95% CI: 1.19, 3.43; P = 0.01) and CSS (HR: 3.45; 95% CI: 1.68, 7.10; P = 0.001). | Not indicated. |
| W Yang, C Shou, J Yu, Q Zhang, X Liu, H Yu, X Lin, unpublished results, 2019 | —, retrospective | Gastrointestinal stromal tumors, underwent resection | 455 (222 males, 233 females) | Median (range): 57 (20–80) | Median (range): 110 (7–232) mo | Normal: 0–1 (n = 219)Light undernutrition: 2–4 (n = 196)Moderate–severe undernutrition: ≥5 (n = 40) | Not indicated. | Not indicated. | *¶◊ | Higher-CONUT group had a lower RFS rate (P = 0.001).CONUT score was an independent prognostic factor for RFS (COUNT ≥5 vs. CONUT = 0–1: HR: 2.83; 95% CI: 1.46, 5.50; P = 0.002). | Not indicated. |
| Zhang et al. (43) | China, retrospective cohort | Metastatic prostate cancer, underwent surgery | 94 males | Median (range): 71 (53–84) | Median (range): 16.31 (4.6–55.10) mo | High: ≥3 (n = 42)Low: 0–2 (n = 52) | X-tile program | Using the X-tile software | *◊ | High-CONUT group had shorter PFS before surgery (P < 0.05).High CONUT score was an independent prognostic factor for PFS (HR: 3.97; 95% CI: 1.05, 11.43; P = 0.004). | BMI did not differ between CONUT groups (P > 0.05). |
| Zheng Z-F et al. (44) | China, retrospective | Gastric cancer, underwent radical gastrectomy | 532 (403 males, 129 females) | Mean ± SD: 61.1 ± 11.5 | Median (range): 60 (2–76) mo | Normal nutrition: n = 291Light malnutrition: n = 183Moderate or severe malnutrition: n = 58 | Not indicated. | Not indicated. | *¶◊ | Higher-CONUT group had the lowest 5-y OS (P = 0.006) and RFS (P = 0.02) rates.CONUT score was not associated with 5-y OS and RFS. | Higher-CONUT group had low BMI (P = 0.01). |
| Zheng Y et al. (45) | China, retrospective | Renal cell carcinoma, underwent nephrectomy | 635 (400 males, 235 females) | Mean ± SD: 61.71 ± 12.51 | Median (range): 48.4 (29.3–80.1) mo | High: ≥2 (n = 349)Low: <2 (n = 286) | X-tile program | X-tile program | *¶◊ | High-CONUT group had shorter OS and CSS (both, P < 0.0001).CONUT score was an independent risk predictor of OS (HR: 3.01; 95% CI: 1.52, 5.95; P = 0.001) and CSS (HR: 3.00; 95% CI: 1.29, 6.98; P = 0.01). | High-CONUT group had low BMI (P < 0.001). |
1
CONUT, controlling nutritional status; CSS, cancer-specific survival; DFS, disease-free survival; OS, overall survival; PFS, progression-free survival; RFS, recurrence/relapse-free survival; ROC, receiver operating characteristic.
2
*Survival rate, ¶HR, ◊CONUT score. The Kaplan–Meier method and log-rank test were used to estimate survival rates. The Cox proportional hazards regression model was used to calculate HRs and 95% CIs. CONUT score was calculated from serum albumin and total cholesterol concentrations and total peripheral lymphocyte counts.