Nutritional scores predict the prognosis of patients with pulmonary tuberculosis

doi:10.21203/rs.3.rs-3739522/v1

Background: Although malnutrition is associated with poor prognosis in Pulmonary Tuberculosis (PTB) patients, there is currently no nutrition-based prediction model for PTB. Herein, we explored the applicability of common nutrition scores on the prognosis of PTB patients, aiming to construct a nutrition-based prediction model related to their prognosis.

Methods: We retrospectively collected clinical baseline data from 167 patients with secondary PTB who had not previously received anti-TB treatment. Subsequently, we determined the CONUT score, PNI index, and NPS score and evaluated the treatment efficacy using changes in lung lesions revealed by the chest CT scan. The Area Under the Receiver Operating Characteristic (AUROC) curve was used to quantify the predictive values of CONUT, PNI, and NPS scores for anti-TB efficacy in new-onset PTB patients, and the critical CONUT, PNI, and NPS values were determined using the Youden Index. We also performed univariate and multivariate analyses of prognostic factors in PTB patients to determine the nutrition scores and other clinical factors associated with the prognosis of patients with the new-onset PTB.

Results: The Youden Index revealed that the critical value of the CONUT score in PTB patients was 4.5, at which the sensitivity (72.2%) and specificity (96.6%) were highest. On the other hand, the critical cut-off values of the PNI index and the NPS score were 39.825 and 3.5, respectively. Univariate analysis of the predictors of poor prognosis in PTB patients showed that patients with diabetes, COPD, pneumonia, and hypoproteinemia (which were risk factors) had a poor prognosis (p<0.05). Multivariate analysis revealed that the CONUT score, PNI, NPS, and NRS-2002 were independent predictors of unfavourable PTB outcomes, with adjusted ORs of 0.016 (95%CI: 0.936-0.993, p<0.0001), 23.667 (95%CI: 9.317-60.115, p<0.0001), 0.119 (95%CI: 0.054-0.262, p<0.0001), and 0.124 (95%CI: 0.055-0.279, p<0.0001), respectively. The Area Under the Curve (AUC) of the CONUT score in predicting poor prognosis of PTB patients was 0.885 (95% CI：0.830-0.940，p<0.0001), which is comparable to that of the PNI index (0.862, 95% CI: 0.805-0.920, p<0.0001), but higher than that of NPS (0.774, 95% CI: 0.702-0.846, p<0.0001), BMI (0.627, 95% CI: 0.541-0.717, p<0.0001), and NRS-2002 (0.763, 95% CI: 0.688-0.838, p<0.0001). We discovered that older patients (p<0.0001), male participants (p< 0.05), and patients with Diabetes Mellitus (DM) (p<0.0001) and Chronic Obstructive Pulmonary Disease (COPD) (p<0.05) were more likely to have a high CONUT score.

Conclusion: The poor prognosis of PTB patients was related to a high CONUT score, low PNI index, and high NPS score, of which the specificity and sensitivity of the CONUT score were higher than those of the PNI index and the NPS score.

Pulmonary Tuberculosis

Malnutrition

Controlling Nutritional Status

Prognostic Nutritional Index

Naples Prognostic Score

Body Mass Index

Nutrition Risk Screening-2002

Prognosis.

Tuberculosis (TB), a chronic disease caused by Mycobacterium Tuberculosis (MTB), will continue to be a major global public health problem. In 2021, TB incidence and mortality were reported to be 10.6 million and 1.6 million people, respectively, with a significantly higher prevalence in developing countries than in developed economies^[1] .The lung is the body organ most susceptible to MTB, with reported lung involvement rates in subjects with active TB ranging between 79% and 87%^{[2, 3]}.

In all pathogen infections, a complex interplay between host response and microbial virulence, which regulates the overall metabolic response as well as the extent of tissue damage, prevails. According to research, TB leads to malnutrition, and malnourished patients are highly susceptible to TB^[4]. Malnutrition is considered an independent risk factor for TB recurrence^[5] as well as an important and potentially reversible risk factor for treatment failure^[6]. Malnutrition can be assessed in various ways, and although BMI and NRS-2002 are readily available and simple to use; hence, they are commonly employed, they mostly focus on macro factors such as patient disease status and body weight^[7]. According to previous research, albumin levels, lymphocyte count, cholesterol content, and other laboratory results are related to the prognosis of PTB patients^[8]. Therefore, based on nutritional assessment, more reliable combined scoring systems such as the CONUT score, the PNI index, and the NPS score were developed to accurately predict patient prognoses. Herein, we hope to verify the applicability of these screening systems in patients who are most likely to benefit from nutritional interventions and better inform anti-TB therapy.

The Controlling Nutritional Status (CONUT) score, an objective and easy-to-apply biomarker for predicting the prognosis of cancer patients^{[9, 10]}, encompasses serum albumin levels, cholesterol content, and lymphocyte counts^[11] (Table 1). On the other hand, the PNI score is determined based on the following formula: (10 × albumin level [g/dL]) + (0.005 × lymphocyte count [number/mm3]) ^[12]. It was originally designed to determine the risk of postoperative complications following Gastrointestinal (GI) surgery. Simple calculations employing experimentally determined formulas can quantify the nutritional status associated with patient outcomes. Finally, Galizia et al. developed a new research method, the Naples Prognostic Score (NPS) (Table 2), a scoring system comprising serum albumin content, Total Cholesterol (TC) level, Neutrophyl-to-Lymphocyte Ratio (NLR), and the Lymphocyte-to-Monocyte Ratio (LMR), that can comprehensively reflect patients' immune and nutritional status^[13]. These three systems have been demonstrated to predict the prognosis of patients with Non-Small Cell Lung Cancer (NSCLC), breast cancer, and GI tumors^[14-16]. However, the correlation of the CONUT score, the PNI index, and the NPS score with outcomes in TB patients has not been reported. As a result, this study aims to explore the predictive value of the above-mentioned screening systems in the prognosis of TB patients, and to compare their predictive ability.

Table 1. Scoring methods for the CONUT score.

Variables	Degree of nutritional status
	Normal	Mild	Moderate	Serve
Serum albumin concentration (g/dL)	≥3.50	3.0-3.49	2.50-2.99	<2.5
Score	0	2	4	6
Total lymphocyte count (/mm3)	≥1600	1200-1599	800-1199	<800
Score	0	1	2	3
Total cholesterol concentration (mg/dL)	≥180	140-179	100-139	<100
Score	0	1	2	3
Total score	0-1	2-4	5-8	9-12

Abbreviations: CONUT denotes Controlling Nutritional Status.

Table 2. Scoring methods for the NPS score.

Variables	Degree of nutritional status
	score	score
Plasma albumin concentration (g/dL)	≥4.0	<4.0
Score	0	1
Total cholesterol concentration (mg/dL)	>180	≤180
Score	0	1
N/L	≤2.96	>2.96
Score	0	1
L/M	>4.44	≤4.44
Score	0	1
Total score	>0	>2
	Mild malnutrition risk	Severe malnutrition risk

Abbreviations: NPS denotes Naples Prognostic Score.

Patients:

We retrospectively collected clinical data on patients with new-onset secondary PTB who hospitalized in TB-designated hospitals of Beibei district（the Nine People's Hospital of Chongqing) between January 2021 and December 2021.

The inclusion criteria were as follows: ① Patients clearly diagnosed with secondary PTB per the diagnostic criteria for secondary PTB (WS288-2017) ^[17];② Patients positive for MTBC (Mycobacterium Tuberculosis Complex) who returned negative results for RIF resistance per the Xpert MTB/RIF (Rifampicin) test; ③ Patients who underwent a standard first-line treatment regimen of four antibiotics including Isoniazid (H), Rifampicin (R), Pyrazinamide (Z), and Ethambutol (E), in a 2-month initial phase (2HRZE) followed by a 4-month continuation phase (4HR) (2HRZE/4HR); and ④ Patients who underwent chest CT scan again at the end of six months of anti-TB treatment for evaluating PTB lesions.

The exclusion criteria were as follows: ① Patients not on drugs or any regular follow-up assessments; ② Patients who have previously received anti-TB treatment or those with missing data on their TB condition pre-treatment. ③ Patients on long-term use of Glucocorticoids (GCs) or drugs affecting albumin levels and lymphocyte count.

Data collection:

We obtained clinical data on PTB patients before anti-TB treatment from the hospital's Information Management System (IMS) and used it as the baseline. Specifically, we collected data on general demographics (age, gender, occupation, days of hospital stay, and so on), drug and alcohol use history (such as smoking and drinking), comorbidities (such as hypertension, Diabetes Mellitus (DM), renal insufficiency, hyperlipidemia, Chronic Obstructive Pulmonary Disease (COPD), and Acquired Immunodeficiency Syndrome (AIDS)], laboratory tests (such as blood routine tests, liver function tests, kidney function tests, blood lipids, sputum culture, sputum smear, chest CT, and so on) and complications, (such as respiratory failure, pulmonary infection, and so on).

The efficacy of anti-TB drug treatment was assessed through chest Computed Tomography (CT) scans to observe radiological features per the national guidelines.

① Infiltrated lesions

a. Obvious absorption: Lesion absorption ≥1/2 of the original lesion.

b. Absorption: Lesion absorption <1/2 of the original lesion

c. Unchanged: No significant change in the lesion

d. Worsening: Lesion expansion or dissemination

② Cavities

a. Closure: Obstructive closure and scar closure or disappearance

b. Reduction: Cavity reduction ≥ ½ of the original cavity diameter

c. Unchanged: Cavity reduction or increase < ½ of the original cavity diameter

d. Increase: Cavity increase ≥ ½ of the original cavity diameter

This study was approved by the Ethical Review Committee of the Ninth People's Hospital of Chongqing for the use of clinical data, and was conducted per the 1964 Declaration of Helsinki. All patients gave informed consent before participating herein.

All statistical analyses were performed using SPSS v25.0. Continuous variables were expressed as Mean ± Standard Deviation (SD) and compared using two independent sample t-tests or Mann-Whitney U tests. On the other hand, categorical variables were expressed as counts (N) or percentages (%) and compared using the chi-square test or Fisher's exact test. The Area Under the Receiver Operating Characteristic (AUROC) curve of the six-month anti-TB treatment efficacy was used to quantify the predictive value of CONUT, PNI, NPS, BMI and NRS-2002 for the anti-TB efficacy in treatment-naive new onset PTB patients. The maximum Youden Index value was used as the best cut-off value to classify patients into two groups. Spearman's correlation was used to evaluate the correlation of the key five parameters (CONUT, PNI, NPS, BMI, and NRS-2002) with prognosis, as well as among the five individual measures. Univariate and multivariate Logistic regression analyses were performed to determine the independent predictors of anti-TB treatment efficacy, and the Odds Ratio (OR) and 95% Confidence Interval (CI) were calculated. Results with p < 0.05 were considered statistically significant.

1. Clinical Features

Table 3 summarizes the clinical characteristics of 167 PTB patients before anti-TB treatment. Age ranged between 14 and 85 years, with an average of 53 years. Days of hospital stay ranged between 3 and 122 days, with an average of 23 days. Among the patients enrolled, 88 (52.69%), 37 (22.16%), and 4 (2.40%) had a history of smoking, drinking, and dust work, respectively. Additionally, among the patients, 29 (17.37%), 29 (17.37%), 3 (1.80%), 8 (4.79%), 16 (9.58%), and 11 (6.59%) had hypertension, diabetes, renal insufficiency, hyperlipidemia, COPD, and AIDS, respectively. Major complications included pneumonia and hypoproteinemia in 10 (5.99%) and 11 (6.59%) patients, respectively. Furthermore, among the patients, 42 (25.15%) were underweight (BMI<18.5kg/m2), 89 (53.29%) had NRS-2002 scores ≥3 and were at risk of malnutrition, 60 (35.93%) had CONUT scores >4.5, 69 (41.32%) had PNI ≤39.825, and 81 (48.50%) had NPS scores >3.5.

Table 3. Characteristics of PTB patients.

Characteristics	Value or No. of Patients	Characteristics	Value or No. of Patients
Age (years)		AIDS
Mean (SD)	53.12 (18.51)	Present	11 (6.59%)
Gender		Complications
Male	125 (74.85%)	Pneumonia
Hospital stay (days)		Present	10 (5.99%)
Mean (SD)	23.04 (17.67)	Hypoproteinemia
Smoking status		Present	11 (6.59%)
Present	88 (52.69%)	Nutritional scores
Drinking status		BMI
Present	37 (22.16%)	<18.5kg/m2	42 (25.15%)
Dust-exposed work		≥18.5kg/m2	125 (74.85%)
Present	4 (2.40%)	NRS2002
Basic diseases		≥3	89 (53.29%)
Hypertension		<3	78 (46.71%)
Present	29 (17.37%)	CONUT
Diabetes mellitus		> 4.5	60 (35.93%)
Present	29 (17.37%)	≤ 4.5	107 (64.07%)
Renal insufficiency		PNI
Present	3 (1.80%)	> 39.825	98 (58.68%)
Hyperlipidemia		≤ 39.825	69 (41.32%)
Present	8 (4.79%)	NPS
COPD		> 3.5	81 (48.50%)
Present	16 (9.58%)	≤ 3.5	86 (51.50%)
≤ 3.5	86 (51.50%)

Abbreviations: PTB denotes Pulmonary Tuberculosis, COPD denotes Chronic Obstructive Pulmonary Disease; AIDS denotes Acquired Immunodeficiency Syndrome; BMI denotes Body Mass Index; NRS-2002 denotes Nutrition Risk Screening-2002; CONUT denotes Controlling Nutritional Status; PNI denotes Prognostic Nutrition Index; NPS denotes Naples Prognostic Score; and SD denotes Standard Deviation.

2. Univariate analysis of prognosis

Demographics, clinical features, and laboratory test results were subjected to univariate statistical analysis in relation to PTB prognosis (Table 4). Age (p<0.0001) and days of hospital stay (p<0.05) were found to be risk factors for the prognosis of PTB patients. Patients with DM (p<0.05), COPD (p<0.05), and pneumonia (p<0.05) were more likely to have adverse outcomes. Among the laboratory test results, TC (p<0.0001), albumin (p<0.0001), N/L (p<0.0001), M/L (p<0.05), and P/L (p<0.05) were significantly associated with an unfavorable outcome. On the other hand, nutritional assessment revealed that BMI (p<0.05), NRS-2002 (p<0.0001), CONUT (p<0.0001), PNI (p<0.0001), and NPS (p<0.0001) may be prognostic risk factors. All results were statistically significant.

Table 4. Demographic and clinical parameters in the univariate analysis of participating patients by the six-month prognosis outcome.

Characteristic	Total patients (N= 167)	Favorable outcome (N = 88)	Unfavorable outcome (N= 79)	P-value
Age (years), mean (SD)	53.12 (18.51)	46.48 (18.05)	60.52 (16.14)	<0.0001
Gender (male), n (%)	125 (74.85)	63 (71.59)	62 (78.48)	0.307
Hospital stay (days), mean (SD)	23.04 (17.67)	19.40 (14.46)	27 (20.01)	0.008
Smoking, n (%)	88 (52.69)	45 (51.14)	43 (54.43)	0.670
Drinking, n (%)	37 (22.16)	20 (22.73)	17 (21.52)	0.851
Dust work history, n (%)	4 (2.40)	3 (3.41)	1 (1.27)	0.385
Basic diseases
Hypertension, n (%)	29 (17.37)	14 (15.91)	15 (18.99)	0.600
Diabetes mellitus, n (%)	29 (17.37)	9 (10.23)	20 (25.32)	0.022
COPD, n (%)	16 (9.58)	4 (4.55)	12 (15.19)	0.027
AIDS, n (%)	11 (6.59)	8 (9.10)	3 (3.80)	0.182
Hyperlipidemia, n (%)	8 (4.79)	3 (3.40)	5 (6.33)	0.385
Renal insufficiency, n (%)	3 (1.80)	1 (1.13)	2 (2.53)	0.509
Complications
Pneumonia, n (%)	10 (5.99)	1 (1.13)	9 (11.39)	0.024
Hypoproteinemia, n (%)	11 (6.59)	4 (4.54)	7 (8.86)	0.270
Laboratory feature
Total cholesterol (mg/dL), mean (SD)	160.27 (44.42)	175.84 (41.57)	142.93 (41.16)	<0.0001
Albumin (g/dL), mean (SD)	3.60 (0.66)	3.96 (0.53)	3.19 (0.55)	<0.0001
Blood creatinine (umol/L), mean (SD)	61.36 (20.17)	60.18 (20.84)	62.67 (19.46)	0.427
N/L	6.93 (7.17)	4.50 (4.38)	9.63 (3.87)	<0.0001
M/L	0.54 (0.61)	0.43 (4.70)	0.66 (0.028)	0.008
P/L	305.35 (314.41)	227.84 (27.29)	391.69 (130.93)	0.003
BMI, n (%)				0.012
<18.5kg/m2	42 (25.15)	15 (17.05)	27 (34.18)
≥18.5kg/m2	125 (74.85)	73 (82.95)	52 (65.82)
NRS2002, n (%)				< 0.0001
≥3	89 (53.29)	25 (28.41)	64 (81.01)
<3	78 (46.71)	63 (71.59)	15 (18.99)
CONUT score, n (%)				< 0.0001
Low (≤ 4.5)	107 (64.07)	85 (96.59)	22 (27.85)
High (> 4.5)	60 (35.93)	3 (3.41)	57 (72.15)
PNI, n (%)				< 0.0001
Low (≤ 39.825)	69 (41.32)	9 (10.23)	60 (75.95)
High (> 39.825)	98 (58.68)	79 (89.77)	19 (24.05)
NPS, n (%)				< 0.0001
Low (≤ 3.5)	86 (51.50)	67 (76.14)	19 (24.05)
High (> 3.5)	81 (48.50)	21 (23.86)	60 (75.95)

Abbreviations: CONUT denotes Controlling Nutritional Status; PNI denotes Prognostic Nutrition Index; NPS denotes Naples Prognostic Score; BMI denotes Body Mass Index; NRS-2002 denotes Nutrition Risk Screening-2002; AIDS denotes Acquired Immunodeficiency Syndrome; COPD denotes Chronic Obstructive Pulmonary Disease; and SD denotes Standard Deviation.

3. Multivariate analysis of prognosis

Table 5 summarizes our conclusions after subjecting the statistically significant indicators in the univariate analysis to multivariate analysis with CONUT, PNI, NPS, BMI, and NRS-2002. Age and pneumonia were independent risk factors for prognosis in models 1, 2, 3, and 5. Multivariate analysis results (along with the adjusted ORs) revealed that the CONUT score (OR: 0.016; 95%CI: 0.936-0.993; p<0.0001), PNI (OR: 23.667; 95%CI: 9.317-60.115; p<0.0001), NPS (OR: 0.119; 95%CI: 0.054-0.262; p<0.0001), and NRS-2002 (OR:0.124; 95%CI:0.055-0.279; p<0.0001)) were independent predictors of unfavourable PTB outcomes.

Table 5. Multivariate analysis results on the involvement of different variables in PTB patients who experienced poor functional outcomes at six months.

		β	OR	95% CI	P-value
Model 1	Age (years)	-0.037	0.964	0.936-0.993	0.015
	Hospital stay (days)	-0.003	0.997	0.968-1.027	0.839
	COPD	1.023	2.782	0.058-134.594	0.605
	Diabetes mellitus	-0.354	0.702	0.105-5.687	0.715
	Pneumonia	-4.59	0.014	0.001-0.992	0.046
	CONUT	-4.167	0.016	0.444-14.344	<0.0001
Model 2	Age (years)	-0.032	0.968	0.942-0.995	0.022
	Hospital stay (days)	-0.001	1.000	0.973-1.028	0.977
	COPD	0.317	1.373	0.060-31.600	0.843
	Diabetes mellitus	-0.303	0.739	0.132-4.125	0.730
	Pneumonia	-3.617	0.027	0.001-0.914	0.044
	PNI	3.164	23.667	9.317-60.115	<0.0001
Model 3	Age (years)	-0.045	0.956	0.933-0.980	<0.0001
	Hospital stay (days)	-0.009	0.991	0.967-1.015	0.470
	COPD	0.773	2.166	0.170-27.555	0.551
	Diabetes mellitus	-0.319	0.727	0.169-3.130	0.669
	Pneumonia	-3.760	0.023	0.001-0.262	0.024
	NPS	-2.218	0.119	0.054-0.262	<0.0001
Model 4	Age (years)	-0.049	0.952	0.931-0.973	<0.0001
	Hospital stay (days)	-0.009	0.991	0.968-1.014	0.448
	COPD	0.268	1.308	0.133-12.848	0.818
	Diabetes mellitus	-0.509	0.601	0.167-2.169	0.437
	Pneumonia	-2.553	0.078	0.004-1.386	0.082
	BMI	0.679	1.973	0.849-4.581	0.114
Model 5	Age (years)	-0.038	0.963	0.940-0.986	0.002
	Hospital stay (days)	-0.004	0.996	0.972-1.020	0.739
	COPD	-0.222	0.801	0.057-11.179	0.869
	Diabetes mellitus	-0.073	0.930	0.245-3.529	0.915
	Pneumonia	-2.608	0.074	0.003-1.856	0.113
	NRS-2002	-2.090	0.124	0.055-0.279	<0.0001

Abbreviations: CONUT denotes Controlling Nutritional Status; PNI denotes Prognostic Nutrition Index; NPS denotes Naples Prognostic Score; BMI denotes Body Mass Index; NRS-2002 denotes Nutrition Risk Screening-2002; CI denotes Confidence Interval; and OR denotes Odds Ratio.

4. Relationship between the CONUT score, PNI, NPS, BMI, and NRS-2002.

Based on the Youden index, the cut-off values for the CONUT score, the PNI index, and the NPS score were determined to be 4.5 [most appropriate for sensitivity (72.2%) and specificity (96.6%)], 39.825 [most appropriate for sensitivity (89.8%) and specificity (75.9%)], and 3.5 [most appropriate for sensitivity (75.9%) and specificity (76.1%)], respectively. The ROC curve analysis results revealed that the CONUT score had the best diagnostic value with the AUC value of 0.885 (95% CI: 0.830-0.940, p<0.0001), which was comparable to that of the PNI score (AUC: 0.862; 95% CI: 0.805-0.920; p<0.0001) (Fig. 1), but higher than that of the NPS score (AUC: 0.774, 95% CI: 0.702-0.846, p<0.0001), BMI (AUC: 0.627; 95% CI: 0.541-0.717, p<0.0001), and NRS-2002 (AUC: 0.763; 95% CI: 0.688-0.838; p<0.0001).

Among the 167 patients, 88 had a favorable outcome, including 85 with a CONUT score <4.5, 79 with a PNI index > 39.825, 67 with an NPS score ≤ 3.5, 73 with a BMI ≥ 18.5kg/m2, and 63 with an NRS-2002 score <3. According to the Spearman's correlation analysis results, CONUT, PNI, NPS, BMI, and NRS-2002 were correlated with prognosis, with the CONUT score (r =-0.672, p<0.0001), NPS (r =-0.510, p<0.0001), and NRS-2002 (r= -0.526, p<0.0001) having a negative correlation with prognosis, and PNI (r = 0.636, p<0.0001) and BMI (r = 0.219, p<0.05) having a positive correlation with prognosis (Tables 6.1-6.4). Therefore, we concluded that the CONUT score was strongly associated with the prognosis of PTB patients.

Table 6.1: Relationship between PNI and CONUT.

		CONUT
		Low	High
PNI	Low	11	58
	High	96	2

Abbreviations: CONUT denotes Controlling Nutritional Status; PNI denotes Prognostic Nutrition Index.

Table 6.2: Relationship between NPS and CONUT.

		CONUT
		Low	High
NPS	Low	79	7
	High	28	53

Abbreviations: CONUT denotes Controlling Nutritional Status; NPS denotes Naples Prognostic Score.

Table 6.3: Relationship between CONUT and BMI.

		CONUT
		Low	High
BMI	Low	21	21
	High	86	39

Abbreviations: CONUT denotes Controlling Nutritional Status; BMI denotes Body Mass Index.

Table 6.4: Relationship between CONUT and NRS-2002.

		CONUT
		Low	High
NRS-2002	Low	76	2
	High	31	58

Abbreviations: CONUT denotes Controlling Nutritional Status; NRS-2002 denotes Nutrition Risk Screening 2002.

We also performed correlation analysis between the CONUT score and PNI (r -0.854, p<0.0001), NPS (r = 0.685, p<0.0001), BMI (r -0.295, p<0.0001), and NRS-2002 (r 0.647, p<0.0001) in the context of nutritional parameters. A significant correlation was found between the PNI index and CONUT scores. The low CONUT and high PNI groups were largely overlapping (Fig. 2).

5. The difference in clinical features between the high CONUT and low CONUT scores

Based on the cut-off value of the CONUT score, patients were divided into two groups: Low CONUT score (≤4.5) and high CONUT score (>4.5). Patients in the high CONUT score group had a significantly higher average age (62 years) than those in the low CONUT score group (48 years) (p<0.0001), and compared to their female counterparts, male patients were more likely to have a high CONUT score (p< 0.05) (Table 7). Compared to those in the low CONUT score group, patients in the high CONUT score group had a longer hospital stay (28 days on average, p<0.0001) and a significantly higher prevalence of DM (p<0.0001) and COPD (p<0.05), which was more likely to be complicated with hyperlipidemia and hypoproteinemia (p<0.05). Regarding the laboratory test results, compared to the low CONUT score group, the high CONUT score group had significantly lower albumin and cholesterol levels and significantly higher N/L, M/L, and P/L values. Regarding the nutritional assessment aspects, the low CONUT score group had a significantly better BMI than the high CONUT score group. On the other hand, the high CONUT score group showed significantly higher NRS-2002 scores than the low CONUT score group.

Table 7. Baseline clinical characteristics of patients with low CONUT and high CONUT scores.

Characteristic	Low-CONUT score (≤4.5), (N = 107)	High-CONUT score (>4.5), (N= 60)	P-value
Age (years), mean (SD)	47.88 (18.51)	62.47 (18.25)	< 0.0001
Gender (male), n (%)	71 (66.36)	54 (90.00)	< 0.05
Hospital stay (days), mean (SD)	20 (17.72)	28.33 (17.74)	< 0.05
Smoking, n (%)	53 (49.53)	35 (58.33)	0.333
Drinking, n (%)	21 (19.63)	16 (26.67)	0.334
Hypertension, n (%)	15 (14.02)	14 (23.33)	0.140
Diabetes mellitus, n (%)	7 (6.54)	22 (36.67)	< 0.0001
Dust work history, n (%)	2 (1.87)	2 (3.33)	0.619
COPD, n (%)	6 (5.60)	10 (16.67)	0.028
Pneumonia, n (%)	4 (3.74)	6 (10.00)	0.099
AIDS, n (%)	4 (3.74)	7 (11.67)	0.058
Hyperlipidemia, n (%)	2 (1.87)	6 (10.00)	0.026
Renal insufficiency, n (%)	1 (0.93)	2 (3.33)	0.293
Hypoproteinemia, n (%)	3 (2.80)	8 (13.33)	0.018
BMI (< 18.5kg/m2), n (%)	21 (19.63)	21 (35.00)	0.040
NRS2002 (≥ 3), n (%)	31 (28.97)	58 (96.67)	<0.0001

Abbreviations: BMI denotes Body Mass Index; NRS-2002 denotes Nutrition Risk Screening-2002; AIDS denotes Acquired Immunodeficiency Syndrome; COPD denotes Chronic Obstructive Pulmonary Disease; and SD denotes Standard Deviation.

Malnutrition may impair the body's immunity and increase susceptibility to infections by affecting the body's production and activation of cytokines^[18]. Severe malnutrition can impair various immune functions in the human body, including phagocytosis, cellular immunity, antibody concentration, and cytokine production, increasing susceptibility to MTB^[19]. This study aimed to analyze the correlation between the nutritional status scores and outcomes following regular anti-TB treatment in patients diagnosed with new-onset PTB. We found that CONUT, PNI, NPS, and NRS-2002 are all independent predictors of prognosis in PTB patients. Among them, CONUT had a significantly strong correlation with prognosis as well as the highest predictive value. As a result, we believe that high CONUT scores at initial diagnosis are related to a poor prognosis in anti-TB treatment and that elderly males, as well as patients with diabetes and COPD, are more likely to have high CONUT scores.

A vast body of previous research evidence has demonstrated a two-way relationship between TB and malnutrition^[20-22]. On the one hand, TB patients have GI dysfunction and decreased appetite, leading to insufficient intake of nutrients, resulting in reduced anabolism^[20]. On the other hand, MTB uses body proteins for self-metabolism, resulting in physiological changes in individuals^[22]. The changes include low fever, night sweats, weight loss, and so on, leading to increased body catabolism, reduced fat storage, and loss of lean body tissue, further exposing patients to varying degrees of malnutrition. A nationwide multi-center, large-sample study reported that per the NRS-2002 classification, 55.86% of PTB patients are at risk of malnutrition, with 26.56% having a low BMI (<18.5 kg/m²)^[23]. These results are correspondingly consistent with our findings which showed rates of 53.29% and 25.15%, respectively. Previous research has shown that cellular immunity is an important defense mechanism for the host to resist TB^[24]. Prolonged malnutrition could impair Cellular Immune Function (CMI), particularly the function of helper T lymphocyte CD4 cells, resulting in increased susceptibility to infections. Furthermore, the CD8 ratio of T lymphocytes decreases, affecting the body's normal immune function, resulting in a decreased MTB clearance ability, causing repeated TB procrastination, and further complicating PTB treatment. In this regard, malnutrition may affect TB prognosis. Consistent with most previous research findings^{[8, 25]}, we found that NRS-2002 and BMI were related to prognosis but with low correlation and predictive value. At the same time, some studies indicate that patients did not experience significantly weight loss or low BMI when malnourished or immunocompromised^{[26, 27]}.Therefore, besides the general screening factors, TB patients have their specific clinical characteristics, necessitating a more comprehensive and appropriate evaluation model specifically designed for TB.

Many new nutritional scores, including immune status assessments, such as the CONUT score, which encompasses nutritional indicators (such as albumin and cholesterol) and immune indicators (such as lymphocytes) are currently available. The PNI index includes albumin (nutritional indicator) and lymphocytes (immune indicator). The NPS score, on the other hand, comprises nutritional indicators, including lymphocytes, monocytes, and neutrophils. When the above-mentioned single nutritional indicators and immune indicators are combined, their prognostic value for acute diseases such as severe traumatic brain injury^[28] could be verified more quickly, objectively, and accurately. However, further research into their application value in chronic diseases is still required. Herein, we used a novel nutritional score for the prognostic assessment of PTB patients. Our findings revealed that CONUT, PNI, and NPS are all independent predictors of PTB prognosis, with the CONUT score having a higher specificity and sensitivity in predicting outcomes.

Serum albumin is widely considered a nutritional status indicator, and reduced serum albumin levels are often associated with inflammatory responses^[29]. Furthermore, albumin is one of the most sensitive nutritional status indicators in TB patients^[30]. A multitude of animal experiments have shown that hypoalbuminemia can alter the total number of T lymphocytes as well as the absolute and relative numbers of immune system cell subsets^[31], reducing the host's immunity to MTB^[32]. Low serum albumin levels, low total lymphocyte counts, and a low PNI index are highly associated with impaired immune function^[8]. Consistent with previous research, we found that albumin is positively associated with TB prognosis.

Cholesterol on the macrophage membrane is directly involved in MTB phagocytosis^[33]. By altering the intracellular environment of macrophages, MTB infection dysregulates host lipid biosynthesis, uptake, and sequestration^[34]. In vivo studies revealed that cholesterol is required for MTB persistence during the latent phase of infection^[35]. The presence of cholesterol is essential for efficient MTB internalization by macrophages^[36] and is a vital energy source for MTB to maintain persistent infection^[35]. On the other hand, lipid metabolism is critical for virulence. Furthermore, epidemiology studies revealed a link between patients' cholesterol levels and TB prognosis. Higher serum cholesterol levels can downregulate TB bacteria in sputum, enhance the sterilization rate of sputum cultures, and accelerate the conversion of sputum bacteria to negative, implying that high cholesterol levels are associated with faster sterilization^[37]. Similarly, our findings revealed a strong correlation between cholesterol and TB treatment efficacy, with high pre-treatment cholesterol levels potentially predicting a better efficacy.

In summary, the new nutritional score encompassing the serum albumin levels, cholesterol content, and lymphocyte count is more useful in assessing PTB prognosis. After further analyzing the clinical characteristics of the two patient groups (the high and low CONUT score groups), we discovered that the high-scoring group mainly comprised elderly patients, people with underlying illnesses such as diabetes and COPD, and patients in this group may have longer hospital stays. As a result, more aggressive nutritional measures should be taken to improve the prognosis and shorten the hospitalization duration for patients in the high-scoring group.

In conclusion, a high CONUT score, a low PNI index, and a high NPS score were associated with poor outcomes in PTB patients. The CONUT score, NPS score, and PNI index, which are low-cost and widely available biomarkers, could be helpful in identifying patients with poor prognoses who could benefit from early nutritional therapy. Further prospective research should be conducted to evaluate these findings.

Finally, there were some potential limitations to this study. First, we only conducted a nutritional assessment on newly diagnosed patients pre-treatment and did not follow up on their nutritional status. Second, there was limited data on detailed dietary intake during hospitalization as well as other body composition parameters (including body fat or lean body mass). Additionally, this was a retrospective, single-center design study with a relatively small sample size. Therefore, we intend to conduct additional prospective controlled trials to verify whether nutritional support therapy can improve the prognosis of PTB patients.

PTB Pulmonary Tuberculosis

AUROC The Area Under the Receiver Operating Characteristic

COPD Chronic Obstructive Pulmonary Disease

DM Diabetes Mellitus

CONUT Controlling Nutritional Status

PNI Prognostic Nutritional Index

NPS Naples Prognostic Score

BMI Body Mass Index

NRS-2002 Nutrition Risk Screening-2002

TB Tuberculosis

MTB Mycobacterium Tuberculosis

GI Gastrointestinal

TC Total Cholesterol

NLR Neutrophyl-to-Lymphocyte Ratio

LMR Lymphocyte-to-Monocyte Ratio

NSCLC Non-Small Cell Lung Cancer

RIF Rifampicin

H Isoniazid

R Rifampicin

Z Pyrazinamide

E Ethambutol

2HRZE 2-month initial phase

4HR 4-month continuation phase

GCs Glucocorticoids

IMS Information Management System

AIDS Acquired Immunodeficiency Syndrome

OR Odds Ratio

CI Confidence Interval

CMI Cellular Immune

Ethics approval and consent to participate:

All procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1964 and later versions.

Consent for publication: Not applicable.

Availability of data and materials: The datasets used and analysed during the current study are available from the corresponding author on reasonable request.

Competing interests: The authors declare that they have no competing interests.

Funding: Chongqing Natural Science Foundation project (No.csts2021jcyj-msmxmX1219).

Authors' contributions: TJY and LTJ is responsible for providing research direction, TJY is responsible for experimental design, statistical analysis, PYH is responsible for collecting data, TJY and LJY is responsible for writing articles, XMQ and SXF are responsible for article proofreading. All authors reviewed the manuscript.

Acknowledgements：Not applicable.

Author information:

Tan Junyi, Infection Department, the Second Affiliated Hospital of Chongqing Medical University, Chongqing 400010, China.

LI Tianju, Infection Department, the Nine People's Hospital of Chongqing, Chongqing 400070, China.

Pi Yinhuan, Infection Department, the Second Affiliated Hospital of Chongqing Medical University, Chongqing 400010, China.

Xiang Mingque, Infection Department, the Nine People's Hospital of Chongqing, Chongqing 400070, China.

Shi Xiaofeng, Infection Department, the Second Affiliated Hospital of Chongqing Medical University, Chongqing 400010, China.

BAGCCHI S. WHO's Global Tuberculosis Report 2022 [J]. Lancet Microbe, 2023, 4(1): e20.
FARER L S, LOWELL A M, MEADOR M P. Extrapulmonary tuberculosis in the United States [J]. Am J Epidemiol, 1979, 109(2): 205-17.
ARANGO L, BREWIN A W, MURRAY J F. The spectrum of tuberculosis as currently seen in a metropolitan hospital [J]. American Review of Respiratory Disease, 1973, 108(4): 805.
MACALLAN D C. Malnutrition in tuberculosis [J]. Diagn Microbiol Infect Dis, 1999, 34(2): 153-7.
DHEDA K, BARRY C E, MAARTENS G. Tuberculosis [J]. Lancet, 2016, 387(10024): 1211-26.
CHOI R, JEONG B H, KOH W J, et al. Recommendations for Optimizing Tuberculosis Treatment: Therapeutic Drug Monitoring, Pharmacogenetics, and Nutritional Status Considerations [J]. Ann Lab Med, 2017, 37(2).
KO Y, KIM C, PARK Y B, et al. Changes in Nutritional Status in Pulmonary Tuberculosis: Longitudinal Changes in BMI According to Acid-Fast Bacilli Smear Positivity [J]. J Clin Med, 2020, 9(12).
LIU Q X, TANG D Y, XIANG X, et al. Associations between nutritional and immune status and clinicopathologic factors in patients with tuberculosis: A comprehensive analysis [J]. Front Cell Infect Microbiol, 2022, 12: 1013751.
CASAER M P, MESOTTEN D, HERMANS G, et al. Early versus late parenteral nutrition in critically ill adults [J]. N Engl J Med, 2011, 365(6): 506-17.
HEIDEGGER C P, BERGER M M, GRAF S, et al. Optimisation of energy provision with supplemental parenteral nutrition in critically ill patients: a randomised controlled clinical trial [J]. Lancet, 2013, 381(9864): 385-93.
KURODA D, SAWAYAMA H, KURASHIGE J, et al. Controlling Nutritional Status (CONUT) score is a prognostic marker for gastric cancer patients after curative resection [J]. Gastric Cancer, 2018, 21(2): 204-12.
ONODERA T, GOSEKI N, KOSAKI G. [Prognostic nutritional index in gastrointestinal surgery of malnourished cancer patients] [J]. Nihon Geka Gakkai Zasshi, 1984, 85(9): 1001-5.
GALIZIA G, LIETO E, AURICCHIO A, et al. Naples Prognostic Score, Based on Nutritional and Inflammatory Status, is an Independent Predictor of Long-term Outcome in Patients Undergoing Surgery for Colorectal Cancer [J]. Dis Colon Rectum, 2017, 60(12): 1273-84.
CHEN S, LIU S, XU S, et al. Naples Prognostic Score is an Independent Prognostic Factor in Patients with Small Cell Lung Cancer and Nomogram Predictive Model Established [J]. J Inflamm Res, 2022, 15: 3719-31.
CHEN Y, GUAN Z, SHEN G. Naples prognostic score: a novel predictor of survival in patients with HER2-positive breast cancer [J]. Future Oncol, 2022, 18(21): 2655-65.
CHEN F, XIE C, REN K, et al. Prognostic Value of the Naples Prognostic Score in Patients with Gastrointestinal Cancers: A Meta-Analysis [J]. Nutr Cancer, 2023, 75(7): 1520-30.
肺结核诊断标准(WS 288—2017) [J]. 新发传染病电子杂志, 2018, 3(01): 59-61.
KATONA P, KATONA-APTE J. The interaction between nutrition and infection [J]. Clin Infect Dis, 2008, 46(10): 1582-8.
KANT S, GUPTA H, AHLUWALIA S. Significance of nutrition in pulmonary tuberculosis [J]. Crit Rev Food Sci Nutr, 2015, 55(7): 955-63.
CHANG S W, PAN W S, LOZANO BELTRAN D, et al. Gut hormones, appetite suppression and cachexia in patients with pulmonary TB [J]. PLoS One, 2013, 8(1): e54564.
MARTINS N, MORRIS P, KELLY P M. Food incentives to improve completion of tuberculosis treatment: randomised controlled trial in Dili, Timor-Leste [J]. Bmj, 2009, 339: b4248.
DELOGU G, SALI M, FADDA G. The biology of mycobacterium tuberculosis infection [J]. Mediterr J Hematol Infect Dis, 2013, 5(1): e2013070.
CHEN W, DING Q, ZHANG S K, et al. Nutritional status in patients with active pulmonary tuberculosis and new nutritional risk screening model for active tuberculosis: a national, multicenter, cross-sectional study in China [J]. J Thorac Dis, 2023, 15(5): 2779-99.
O'GARRA A, REDFORD P S, MCNAB F W, et al. The immune response in tuberculosis [J]. Annu Rev Immunol, 2013, 31: 475-527.
LI Y, YANG F, ZHOU H, et al. Clinical application of NRS-2002 in nutritional risk screening of tuberculosis inpatients [J]. Ann Palliat Med, 2021, 10(5): 5322-8.
PRESSOIR M, DESNé S, BERCHERY D, et al. Prevalence, risk factors and clinical implications of malnutrition in French Comprehensive Cancer Centres [J]. Br J Cancer, 2010, 102(6): 966-71.
NORMAN K, SMOLINER C, KILBERT A, et al. Disease-related malnutrition but not underweight by BMI is reflected by disturbed electric tissue properties in the bioelectrical impedance vector analysis [J]. Br J Nutr, 2008, 100(3): 590-5.
LI Y, LIU C, LUO X, et al. Controlling nutritional status score and prognostic nutrition index predict the outcome after severe traumatic brain injury [J]. Nutr Neurosci, 2022, 25(4): 690-7.
HORWICH T B, KALANTAR-ZADEH K, MACLELLAN R W, et al. Albumin levels predict survival in patients with systolic heart failure [J]. Am Heart J, 2008, 155(5): 883-9.
KARYADI E, SCHULTINK W, NELWAN R H, et al. Poor micronutrient status of active pulmonary tuberculosis patients in Indonesia [J]. J Nutr, 2000, 130(12): 2953-8.
CEGIELSKI J P, MCMURRAY D N. The relationship between malnutrition and tuberculosis: evidence from studies in humans and experimental animals [J]. Int J Tuberc Lung Dis, 2004, 8(3): 286-98.
DAI G, MCMURRAY D N. Altered cytokine production and impaired antimycobacterial immunity in protein-malnourished guinea pigs [J]. Infect Immun, 1998, 66(8): 3562-8.
DI PIETRO N, FORMOSO G, PANDOLFI A. Physiology and pathophysiology of oxLDL uptake by vascular wall cells in atherosclerosis [J]. Vascul Pharmacol, 2016, 84: 1-7.
WIPPERMAN M F, SAMPSON N S, THOMAS S T. Pathogen roid rage: cholesterol utilization by Mycobacterium tuberculosis [J]. Crit Rev Biochem Mol Biol, 2014, 49(4): 269-93.
PANDEY A K, SASSETTI C M. Mycobacterial persistence requires the utilization of host cholesterol [J]. Proc Natl Acad Sci U S A, 2008, 105(11): 4376-80.
GATFIELD J, PIETERS J. Essential role for cholesterol in entry of mycobacteria into macrophages [J]. Science, 2000, 288(5471): 1647-50.
DENIZ O, GUMUS S, YAMAN H, et al. Serum total cholesterol, HDL-C and LDL-C concentrations significantly correlate with the radiological extent of disease and the degree of smear positivity in patients with pulmonary tuberculosis [J]. Clin Biochem, 2007, 40(3-4): 162-6.

No competing interests reported.

Nutritional scores predict the prognosis of patients with pulmonary tuberculosis

Status:

Version 1

Abstract

Figures

Introduction

Methods

Results

Discussion

Abbreviations

Declarations

Ethics approval and consent to participate:

Consent for publication: Not applicable.

Acknowledgements：Not applicable.

References

Additional Declarations

Status:

Version 1