The C-Reactive Protein/Albumin Ratio Is Useful for Predicting Short-Term Survival in Cancer and Noncancer Patients

Abstract

Objective and Background:

Survival predictions by subjective evaluations are important for end-stage patients. However, subjective evaluations based on experience are difficult. Therefore, we investigated whether the Glasgow prognostic score (GPS), prognostic nutritional index (PNI), and C-reactive protein (CRP)/albumin ratio (CRP/Alb ratio) calculated from the laboratory values of objective evaluations are useful for predicting survival times in end-stage patients.

Methods:

We retrospectively investigated the age, sex, death cause, CRP levels, Alb levels, and lymphocyte counts in 363 cancer and noncancer patients who died in the 12-month period between April 2015 and March 2016. A multivariate analysis was performed to calculate GPS, PNI, and the CRP/Alb ratio from laboratory values and adjusted for confounding factors.

Results:

PNI and CRP/Alb ratio exhibited negative and positive correlations with survival days, respectively. All GPS, PNI, and CRP/Alb ratio were useful to predict two to four remaining weeks. Interestingly, CRP/Alb ratio, but not GPS or PNI, was higher in patients with predicted short-term survival of zero to two weeks than in that of two to four weeks (odds ratio 2.32; 95% confidence interval 1.61–3.34).

Discussion:

These results suggest that the CRP/Alb ratio is an independent factor that is beneficial to predict short-term survival of within two weeks.

Introduction

Prognostic predictions are very important for selecting the optimal treatment for both physicians and patients. The palliative prognosis score (PaP) is used to predict survival times.¹ However, subjective evaluations are performed for approximately half of its elements, which makes the use of PaP difficult. Although the palliative prognostic index has fewer subjective elements than PaP, the evaluation of delirium is needed.² Hypoactive delirium is easily overlooked and difficult to diagnose.

The prognostic nutritional index (PNI),³ C-reactive protein/albumin (CRP/Alb) ratio,⁴ and Glasgow prognostic score (GPS)⁵ are calculated from laboratory values, which are objective evaluations and factors used to predict outcomes after solid cancer surgery. However, it currently remains unclear as to whether PNI, the CRP/Alb ratio, and GPS are useful for predicting the short-term survival of terminal patients.^6,7 Further, it has not yet been established as to whether these prognostic factors are effective for hematological malignancies or noncancer patients. Therefore, we herein investigated whether PNI, the CRP/Alb ratio, and GPS calculated from the laboratory values of objective evaluations are useful for predicting short survival times in cancer and noncancer patients.

Patients and Methods

This study investigated the age, gender, and cause of death of 363 patients who died in the 12-month period between April 2015 and March 2016 at the Seichokai Fuchu Hospital. Blood tests, including those to measure CRP levels, Alb levels, and lymphocyte counts, were conducted 6054 times between April 2015 and March 2016. We used 363 patients corresponding to 5718, 5771, and 5553 laboratory data points on Alb, CRP, and the CRP/Alb ratio, respectively, and 355 patients corresponding to 4714 laboratory data points on PNI. Laboratory data were classified and compared for each predicted survival period: 0–2 weeks before death (0–0.5 months), 2–4 weeks (0.5–1 month), 4–12 weeks (1–3 months), and 12–52 weeks (3–12 months). We also used average values for multiple inspection data within the same period.

PNI was calculated according to the following formula as previously described: 10 × serum Alb value (g/dL) +0.005 × peripheral lymphocyte count (/μL).³ GPS was constructed as previously described.^5,8 Briefly, patients with an elevated CRP level (>1.0 mg/dL) and hypoalbuminemia (<3.5 g/dL) were allocated a score of 2. Patients in whom only one of these biochemical abnormalities was present were allocated a score of 1. Patients in whom neither of these abnormalities was present were allocated a score of 0. The CRP/Alb ratio was calculated from CRP and Alb values. This research was approved by the Ethics Committee of Seichokai Fuchu Hospital.

Statistical analysis

Spearman's rank correlation coefficient was used to study correlations of continuous variables that do not show a normal distribution. The Kruskal-Wallis test and Steel-Dwass method were used to compare three or more continuous variables not showing normality. Using the receiver operating characteristic (ROC) curve and variable of the predicted survival period, we selected optimal cut-off values for Alb, CRP, the CRP/Alb ratio, and PNI that provided the closest point to the left upper hand corner of the graph, and we then divided our patients into two groups. In the univariate analysis, comparisons of categorical variables were performed by using logistic regression. A multivariate logistic regression analysis was performed on factors considered to be important in the univariate analysis. We included variables with a p-value <0.05 in the univariate analysis in the subsequent multivariate analyses. In addition, we included the cause of death in the variables of the multivariate analysis. Collinearity was examined with the variance inflation factor (VIF). The variable we used for multivariate analysis was VIF <5. The results obtained were described as odds ratios (OR) and 95% confidence intervals (CI). A p value <0.05 was defined as being significant. All statistical analyses were performed with EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan),⁹ which is a graphical user interface for R (The R Foundation for Statistical Computing, Vienna, Austria).

Results

The characteristics of 363 patients who died during the study period are shown in Table 1. The causes of death among the patients analyzed were as follows: solid cancer 28.4%, blood cancer 11%, and noncancer 60.6%.

Table 1.

Characteristics of the Subject Patients

	No. of patients or mean age ± SD (n = 363)	%
Age (years)	75 ± 11
Sex
Female	146	40.2
Male	217	59.8
Cause of death
Solid cancer	103	28.4
Stomach	23	6.3
Pancreas	14	3.9
Colon/rectal	13	3.6
Liver	12	3.3
Lung	11	3
Breast	8	2.2
Esophagus	6	1.7
Biliary tract	5	1.4
Renal	3	0.8
Prostate	3	0.8
Bladder	2	0.6
Unknown	2	0.6
Others	1	0.3
Hematologic malignancy	40	11
Leukemia	21	5.8
Lymphoma	13	3.6
Myeloma	6	1.7
Noncancer	220	60.6
Pneumonia	47	12.9
Sepsis	24	6.6
Intracranial hemorrhage	12	3.3
Subarachnoid hemorrhage	11	3
Heart failure	9	2.5
Others	117	32.2

SD, standard deviation.

Correlations were observed between predicted survival periods and Alb, PNI, CRP, and the CRP/Alb ratio (Fig. 1).

FIG. 1.

Correlations between predicted survival periods and Alb (a) (5718 laboratory data points), PNI (b) (4714 laboratory data points), the CRP/Alb ratio (c) (5553 laboratory data points), and CRP (d) (5771 laboratory data points). r_s, Spearman's rank correlation coefficient; Alb, albumin; CRP, C-reactive protein; PNI, prognostic nutritional index.

Figure 2 shows the results of comparisons of laboratory values, PNI, and the CRP/Alb ratio for each predicted survival period. A significant difference was observed in CRP and the CRP/Alb ratio between the predicted short-term survival period of 0.5–1 and 0–0.5 months. There was a significant difference between Alb, PNI, the CRP/Alb ratio, and CRP between the predicted survival periods of 0.5–1, 1–3, and 3–12 months.

FIG. 2.

Comparison of blood test values Alb (a), PNI (b), the CRP/Alb ratio (c), and CRP (d) for each survival period. Data show the mean with 95% CI for 135–343 patients. ***p < 0.001, n.s. versus the 0–0.5 months before death group. ^††p < 0.01, ^†††p < 0.001 versus the 0.5–1 months before death group (Steel-Dwass test).

A univariate analysis was performed by dividing patients into two groups with predicted short-term survival periods of 0–0.5 and 0.5–1 months to verify factors useful for identifying patients with a predicted short-term survival of within 1 month that was actually expected to be within 0.5 months (Table 2). The optimal cut-off values for continuous variables were as follows from the ROC analysis: age: 73, Alb: 2.5, CRP: 6.5, the CRP/Alb ratio: 2.9, and PNI: 31. GPS was previously reported to have a cut-off value of 2.⁶ The areas under ROC curves were as follows: age: 0.554, Alb: 0.558, CRP: 0.596, the CRP/Alb ratio: 0.601, PNI: 0.528, and GPS: 0.523. The univariate analysis revealed significant differences in age, CRP, the CRP/Alb ratio, and the cause of death. A multivariate logistic regression analysis was performed to control confounding factors. CRP was not selected because of strong collinearity. The results of the multivariate logistic regression analysis showed that a higher CRP/Alb ratio correlated with short-term survival of two weeks (adjusted OR: 2.32, 95% CI: 1.61–3.34, p < 0.001). A univariate analysis was then performed by dividing patients into two groups with short-term survival periods of 0.5–1 and 1–3 months to verify factors useful for identifying patients with short-term survival of within 3 months that was actually expected to be within 1 month (Table 3). The optimal cut-off values for continuous variables from the ROC analysis were as follows: age: 73, Alb: 2.6, CRP: 4.1, the CRP/Alb ratio: 1.6, and PNI: 32. GPS was previously reported to have a cut-off value of 2.⁶ The areas under the ROC curves were as follows: age: 0.495, Alb: 0.628, CRP: 0.588, the CRP/Alb ratio: 0.595, PNI: 0.619, and GPS: 0.588. The univariate analysis revealed significant differences in Alb, CRP, the CRP/Alb ratio, PNI, and GPS. A multivariate logistic regression analysis was performed to control for confounding factors. CRP was not selected because of strong collinearity. The results of the multivariate logistic regression analysis showed that lower Alb levels correlated with short-term survival of one month (adjusted OR: 0.53, 95% CI: 0.29–0.98, p = 0.044).

Table 2.

Univariate Analysis and Multivariate Logistic Regression Analysis of Distinct Parameters in Patients with 0–0.5 and 0.5–1 Remaining Months

Parameters	Remaining months 0–0.5 of 1
	Univariate analysis			Multivariate analysis			Total, n
	OR	95% CI	p	OR	95% CI	p	Total, n
Age
<73	1	Reference		1	Reference		563
≥73	1.51	1.07–2.14	0.019	1.39	0.95–2.03	0.087
Albumin
<2.5	1	Reference					559
≥2.5	0.77	0.55–1.08	0.13
CRP
<6.5	1	Reference					550
≥6.5	2.24	1.57–3.21	<0.001
CRP/Alb ratio
<2.9	1	Reference		1	Reference		548
≥2.9	2.21	1.55–3.17	<0.001	2.32	1.61–3.34	<0.001
PNI
<31	1	Reference					514
≥31	0.74	0.52–1.06	0.1
mGPS
0, 1	1	Reference					549
2	1.34	0.87–2.06	0.18
Cause of death
Hematologic malignancy	1	Reference		1	Reference		563
Solid cancer	0.94	0.54–1.64	0.82	0.82	0.46–1.46	0.5
Noncancer	1.56	0.93–2.63	0.091	1.49	0.86–2.58	0.16

Alb, albumin, CI, confidence interval; CRP, C-reactive protein; GPS, Glasgow prognostic score; PNI, prognostic nutritional index; OR, odds ratio.

Table 3.

Univariate Analysis and Multivariate Logistic Regression Analysis of Distinct Parameters in Patients with 0.5–1 and 1–3 Remaining Months

Parameters	Remaining months 0.5–1 of 3
	Univariate analysis			Multivariate analysis			Total, n
	OR	95% CI	p	OR	95% CI	p	Total, n
Age
<73	1	Reference					430
≥73	0.94	0.64–1.37	0.74
Albumin
<2.6	1	Reference		1	Reference		425
≥2.6	0.43	0.29–0.64	<0.001	0.53	0.29–0.98	0.044
CRP
<4.1	1	Reference					416
≥4.1	1.86	1.26–2.74	0.0018
CRP/Alb ratio
<1.6	1	Reference		1	Reference		416
≥1.6	1.91	1.29–2.82	0.0012	1.42	0.86–2.36	0.17
PNI
<32	1	Reference		1	Reference		387
≥32	0.46	0.31–0.70	<0.001	0.79	0.44–1.41	0.42
mGPS
0, 1	1	Reference		1	Reference		414
2	1.71	1.11–2.65	0.016	0.96	0.54–1.70	0.89
Cause of death
Hematologic malignancy	1	Reference		1	Reference		430
Solid cancer	0.92	0.51–1.65	0.78	0.74	0.40–1.37	0.34
Noncancer	1.05	0.60–1.85	0.85	0.8	0.44–1.46	0.47

Discussion

The results of this study indicate that the CRP/Alb ratio is useful for predicting the short-term survival of patients with solid cancer, hematological malignancies, and noncancer of within two weeks. However, it was not useful for predicting longer outcomes. Therefore, the CRP/Alb ratio may be suitable to predict survival times within two weeks in both cancer and noncancer patients.

Previous models to predict survival times were developed by using variables from a single time point (e.g., discharge or surgery date).^10–13 In contrast, this study was based on robust data, utilizing laboratory variables monitored within one year before death events. A few recent studies have identified some candidate biomarkers that may be indicative of a common biological process to dying in cancer patients,^14,15 whereas little such information is available for noncancer patients. The CRP/Alb ratio, PNI, and GPS are used as prognostic tools for estimating survival time in patients with gastric cancer,^16–18 pancreatic cancer,^19,20 colon cancer,^21–24 liver cancer,^25,26 lung cancer,^6,27,28 and esophageal cancer.^4,29–30 The present data from both cancer and noncancer patients were, in general, consistent with the previous studies on solid cancer patients. Although the CRP/Alb ratio is used as a tool to predict outcomes, as described earlier, few studies have evaluated short-term survival. Alb is also important for predicting outcomes.^31–34

PNI,^35,36 GPS,³⁷ and CRP³⁸ are used to predict the outcomes of patients with hematological malignancies. PNI, Alb, and CRP were previously demonstrated to be effective as a tool to predict the outcomes of noncancer patients (e.g., heart failure).^10,39 A neutrophil/lymphocyte ratio, known to be useful for predicting the outcome of cancer patients, was not measured in this study. However, it is noteworthy that the neutrophil/lymphocyte ratio is well correlated with the CRP/Alb ratio.⁴⁰ PNI,¹⁸ GPS⁵, and CRP/Alb ratio⁴ have been reported to be useful for predicting survival time in a month-year range. Our study suggests that the CRP/Alb ratio, but not PNI or GPS, is also applicable for predicting relatively short-term (within two weeks) survival.

Accurate predictions facilitate the provision of optimal treatments to patients. The prediction of outcomes according to objective criteria is very useful in clinical practice, to avoid the administration of unnecessary parenteral hydration, since subjective outcome prediction requires skilled staff and may not be reliable.⁴¹ Actually, parenteral hydration is not recommended for end-stage cancer patients with predicted short-term survival of less than two weeks.^42–44 It is often emphasized that parenteral hydration for patients with predicted short-term survival is really disadvantageous. Potential limitations need to be considered when interpreting the results of this study. This was an observational study with a retrospective design and was performed at a single general hospital. Therefore, the available sample size in our study was small, resulting in a low predicted value and the potential necessity to combine these data.

In summary, this study suggests that the CRP/Alb ratio of objective data is effective for predicting short-term survival. The results obtained will contribute to improving treatments for end-of-life patients.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

References

Pirovano

, Maltoni

, Nanni

, et al. A new palliative prognostic score: A first step for the staging of terminally ill cancer patients. J Pain Symptom Manage, 1999; 17:231–239.

Morita

, Tsunoda

, Inoue

, et al. The Palliative Prognostic Index: A scoring system for survival prediction of terminally ill cancer patients. Support Care Cancer, 1999; 7:128–133.

Onodera

, Goseki

, Kosaki

. [Prognostic nutritional index in gastrointestinal surgery of malnourished cancer patients]. Nihon Geka Gakkai Zasshi. 1984; 85:1001–1005. [In Japanese.]

X-L

, Yu

H-Q

, Hu

, et al. A novel inflammation-based prognostic score, the C-reactive protein/albumin ratio predicts the prognosis of patients with operable esophageal squamous cell carcinoma. PLoS One, 2015; 10:e0138657.

Forrest

, McMillan

, McArdle

, et al. Comparison of an inflammation-based prognostic score (GPS) with performance status (ECOG) in patients receiving platinum-based chemotherapy for inoperable non-small-cell lung cancer. Br J Cancer, 2004; 90:1704–1706.

Minami

, Ogata

, Ihara

, et al. Pretreatment Glasgow prognostic score and prognostic nutritional index predict overall survival of patients with advanced small cell lung cancer. Lung Cancer, 2017; 8:249–257.

Nakamura

, Nagao

, Saida

, et al. Use of the Prognostic Nutritional Index to predict clinical outcomes of patients with terminal stage cancer. Palliat Care Res, 2013; 8:199–202.

McMillan

, Crozier

JEM

, Canna

, et al. Evaluation of an inflammation-based prognostic score (GPS) in patients undergoing resection for colon and rectal cancer. Int J Colorectal Dis, 2007; 22:881–886.

Kanda

. Investigation of the freely available easy-to-use software “EZR” for medical statistics. Bone Marrow Transplant, 2013; 48:452–458.

10.

Yoshihisa

, Kanno

, Watanabe

, et al. Impact of nutritional indices on mortality in patients with heart failure. Open Heart, 2018; 5:e000730.

11.

Guo

, He

, Chen

, et al. The C-reactive protein/albumin ratio, a validated prognostic score, predicts outcome of surgical renal cell carcinoma patients. BMC Cancer, 2017; 17:171.

12.

Namiuchi

, Sugie

, Saji

, et al. The systemic inflammation-based Glasgow Prognostic Score as a prognostic factor in patients with acute heart failure. J Cardiovasc Med, 2015; 16:409–415.

13.

Okuno

, Wakabayashi

, Kato

, et al. Esophageal stenosis and the Glasgow Prognostic Score as independent factors of poor prognosis for patients with locally advanced unresectable esophageal cancer treated with chemoradiotherapy (exploratory analysis of JCOG0303). Int J Clin Oncol, 2017; 22:1042–1049.

14.

Uneno

, Taneishi

, Kanai

, et al. Development and validation of a set of six adaptable prognosis prediction (SAP) models based on time-series real-world big data analysis for patients with cancer receiving chemotherapy: A multicenter case crossover study. PLoS One, 2017; 12:e0183291.

15.

Reid

, McDonald

, Nwosu

, et al. A systematically structured review of biomarkers of dying in cancer patients in the last months of life; An exploration of the biology of dying. PLoS One, 2017; 12:e0175123.

16.

Takeno

, Hashimoto

, Shibata

, et al. The high-sensitivity modified Glasgow prognostic score is superior to the modified Glasgow prognostic score as a prognostic predictor in patients with resectable gastric cancer. Oncology, 2014; 87:205–214.

17.

Liu

, Sun

, Liu

, et al. Preoperative C-reactive protein/albumin ratio predicts prognosis of patients after curative resection for gastric cancer. Transl Oncol, 2015; 8:339–345.

18.

Sakurai

, Ohira

, Tamura

, et al. Predictive potential of preoperative nutritional status in long-term outcome projections for patients with gastric cancer. Ann Surg Oncol, 2016; 23:525–533.

19.

Imaoka

, Mizuno

, Hara

, et al. Evaluation of modified glasgow prognostic score for pancreatic cancer: A retrospective cohort study. Pancreas, 2016; 45:211–217.

20.

Kanda

, Fujii

, Kodera

, et al. Nutritional predictors of postoperative outcome in pancreatic cancer. Br J Surg, 2011; 98:268–274.

21.

Nakagawa

, Tanaka

, Nojiri

, et al. The modified Glasgow prognostic score as a predictor of survival after hepatectomy for colorectal liver metastases. Ann Surg Oncol, 2014; 21:1711–1718.

22.

Maeda

, Shibutani

, Otani

, et al. Low nutritional prognostic index correlates with poor survival in patients with stage IV colorectal cancer following palliative resection of the primary tumor. World J Surg, 2014; 38:1217–1222.

23.

Ishizuka

, Nagata

, Takagi

, et al. Clinical Significance of the C-reactive protein to albumin ratio for survival after surgery for colorectal cancer. Ann Surg Oncol, 2016; 23:900–907.

24.

Kishiki

, Masaki

, Matsuoka

, et al. Modified Glasgow prognostic score in patients with incurable stage IV colorectal cancer. Am J Surg, 2013; 206:234–240.

25.

Kinoshita

, Onoda

, Imai

, et al. The C-reactive protein/albumin ratio, a novel inflammation-based prognostic score, predicts outcomes in patients with hepatocellular carcinoma. Ann Surg Oncol, 2015; 22:803–810.

26.

Pinato

, North

, Sharma

. A novel, externally validated inflammation-based prognostic algorithm in hepatocellular carcinoma: The prognostic nutritional index (PNI). Br J Cancer, 2012; 106:1439–1445.

27.

Tomita

, Ayabe

, Chosa

, et al. Prognostic significance of pre- and postoperative glasgow prognostic score for patients with non-small cell lung cancer. Anticancer Res, 2014; 34:3137–3140.

28.

Zhou

, Zhan

, Hong

, et al. Ratio of C-reactive protein/albumin is an inflammatory prognostic score for predicting overall survival of patients with small-cell lung cancer. Sci Rep, 2015; 5:10481.

29.

Zhang

, Xi

, Li

Q-Q

, et al. The modified glasgow prognostic score is an independent prognostic factor in patients with inoperable thoracic esophageal squamous cell carcinoma undergoing chemoradiotherapy. J Cancer, 2014; 5:689–695.

30.

Kubo

, Ohira

, Tamura

, et al. Prognostic significance of baseline nutritional index for patients with esophageal squamous cell carcinoma after radical esophagectomy. Esophagus, 2017; 14:84–90.

31.

Viganó. A, Bruera E, Jhangri GS, et al. Clinical survival predictors in patients with advanced cancer. Arch Intern Med, 2000; 160:861–868.

32.

Barbot

A-C

, Mussault

, Ingrand

, et al. Assessing 2-month clinical prognosis in hospitalized patients with advanced solid tumors. J Clin Oncol, 2008; 26:2538–2543.

33.

Penel

, Negrier

, Ray-Coquard

, et al. Development and validation of a bedside score to predict early death in cancer of unknown primary patients. PLoS One, 2009; 4:e6483.

34.

González-Trejo

, Carrillo

, Carmona-Herrera

, et al. Baseline serum albumin and other common clinical markers are prognostic factors in colorectal carcinoma: A retrospective cohort study. Medicine (Baltimore), 2017; 96:e6610.

35.

Zhou

, Wei

, Huang

, et al. Low prognostic nutritional index predicts poor outcome in diffuse large B-cell lymphoma treated with R-CHOP. Int J Hematol, 2016; 104:485–490.

36.

Bairey

, Shacham-Abulafia

, Shpilberg

, et al. Serum albumin level at diagnosis of diffuse large B-cell lymphoma: An important simple prognostic factor. Hematol Oncol, 2016; 34:184–192.

37.

Chou

W-C

, Kao

C-Y

, Wang

P-N

, et al. The application of the palliative prognostic index, charlson comorbidity index, and Glasgow prognostic score in predicting the life expectancy of patients with hematologic malignancies under palliative care. BMC Palliat Care, 2015; 14:18.

38.

Combariza

, Lombana

, Pino

, et al. C-reactive protein and the MASCC risk index identify high-risk patients with febrile neutropenia and hematologic neoplasms. Support Care Cancer, 2015; 23:1009–1013.

39.

Narumi

, Arimoto

, Funayama

, et al. The prognostic importance of objective nutritional indexes in patients with chronic heart failure. J Cardiol, 2013; 62:307–313.

40.

, Li

, Liu

, et al. Prognostic value of C-reactive protein/albumin ratio in predicting overall survival of Chinese cervical cancer patients overall survival: Comparison among various inflammation based factors. J Cancer, 2018; 9:1877–1884.

41.

Stone

, Lund

. Predicting prognosis in patients with advanced cancer. Ann Oncol, 2007; 18:971–976.

42.

Bruera

, Hui

, Dalal

, et al. Parenteral hydration in patients with advanced cancer: A multicenter, double-blind, placebo-controlled randomized trial. J Clin Oncol, 2013; 31:111–118.

43.

Bruera

, Sala

, Rico

, et al. Effects of parenteral hydration in terminally ill cancer patients: A preliminary study. J Clin Oncol, 2005; 23:2366–2371.

44.

McCann

, Hall

, Groth-Juncker

. Comfort care for terminally ill patients: The appropriate use of nutrition and hydration. J Am Med Assoc, 1994; 272:1263–1266.