Smoking is a Factor in Discordance Between QuantiFERONTB Gold Assay and Tuberculosis Etiology: Especially in Older Patients

Figures & data

Figure 1 The entry rules of study population.

Abbreviations: NTM, nontuberculous mycobacteria; PNB, p-nitrobenzoic acid; HN, high-nil (>8 IU/mL); LM, low mitogen (<0.5 IU/mL) response.

Figure 2 Effects of smoking on IFN-γlevels in different age groups. (A) total patients. (B) paitents’ age < 65 years. (C) patients’ age ≥ 65 years.

Notes: Nil, levels of Interferon-γ were measured in supernatants from QuantiFERON-TB Gold In tube blood collection tubes after incubation with saline; Antigen, (M) tuberculosis-specific antigens; Mitogen, phytohaemagglutinin.

Abbreviations: ns, non-smokers; s, smokers.

Table 1 Comparison of Background Characteristics of Smokers and Non-Smokers Before and After Matching

Characteristics	Before Propensity Matching			After Propensity Matching
	Non-Smokers (n=902)	Smokers (n=311)	P value	Non-Smokers (n=313)	Smokers (n=191)	P-value
Age, years, n, (%)	50.00(28.00–68.00)	55.00(38.00–68.00)	0.002	53.00(30.00–69.00)	54.00(34.50–68.00)	0.770
Gender, male, n, (%)	453(50.20)	300(96.50)	<0.001	299 (95.50)	181 (94.80)	0.861
Alcoholn, n, (%)	45(5.00)	162(52.10)	<0.001	41 (13.10)	42 (22.00)	0.013
Underlying disease, n, (%)
Hypoproteinemian, n, (%)	252(27.90)	92(29.60)	0.579	88 (28.10)	58 (30.40)	0.660
Chronic liver, n,(%)	124(13.70)	49(15.80)	0.382	49 (15.70)	29 (15.20)	0.988
Chronic nephrosis, n, (%)	36(4.00)	16(5.10)	0.386	13 (4.20)	6 (3.10)	0.736
Autoimmune disease, n, (%)	24(2.70)	3(1.00)	0.080	5 (1.60)	3 (1.60)	1.000
Malignancy, n, (%)	21(2.30)	3(1.00)	0.136	5 (1.60)	2 (1.00)	0.905
Diabete, n, (%)	114(12.60)	61(19.60)	0.003	59 (18.80)	39 (20.40)	0.752
Hypertension, n, (%)	151(16.70)	40(12.90)	0.105	35 (11.20)	23 (12.00)	0.881
Cardiovascular, n, (%)	42(4.66)	14(45.02)	0.911	13 (4.20)	11 (5.80)	0.545
Laboratory findings
Albumin, g/L	36.60(32.40–40.73)	35.00(31.10–39.10)	<0.001	35.40(31.20–39.40)	34.80(30.40–39.10)	0.374
Globulin, g/L	32.90(29.60–36.63)	32.80(29.70–36.70)	0.371	32.20(29.40–36.60)	33.00(29.80–36.80)	0.637
CRP (mg/L)	21.70(5.66–51.90)	30.70(9.82–62.10)	0.002	32.50(29.60–36.50)	32.60(29.10–36.70)	0.625
WBC count, *109	5.80(4.70–7.30)	6.90(5.30–8.60)	<0.001	6.50(5.20–8.50)	7.10(5.60–8.95)	0.106
Lymphocyte, %	16.35(11.80–23.13)	15.20(10.70–19.80)	0.001	15.40(10.60–20.20)	15.00(10.60–19.5)	0.788
Total lymphocyte, *10^9	1.00(0.70–1.30)	1.00(0.70–1.35)	0.411	1.00(0.70–1.30)	1.00(0.70–1.40)	0.311
Hemoglobin, g/L	121.00(109.00–133.00)	127.00(113.00–139.00)	<0.001	128.00(114.00–137.00)	126.00(113.00–137.00)	0.516

Note: Values are numbers (%) or medians (interquartile ranges).

Abbreviations: WBC, white blood cells; CRP, C-reactive protein.

Table 2 Logistic Regression Analysis of the Influence of Smoking on the Inconsistency Between QFT Detection and Tuberculosis Etiology

Smoking	Consistent Group (n=1105)	Inconsistent Group (n=108)	Model 1		Model 2
			OR (95% CI)	P- value	OR (95% CI)	P-value
Non-Smokers, n (%)	835 (75.60)	67 (62.00)	Reference		Reference
Smokers, n (%)	270 (24.40)	41 (38.00)	1.89 (1.25–2.86)	0.002	1.69 (1.09–2.63)	0.020

Notes: Model 1 was a crude model. Model 2 was adjusted for Hypoproteinemia, Chronic liver, and variables converted by BOX-COX (lymphocyte percentage, age, globulin).

Table 3 Comparison of Smoking After Propensity Matching by Logistic Regression Analysis

Smoking	Consistent Group (n=343)	Inconsistent Group (n=47)	After Propensity Matching^a
			OR (95% CI)	P-value
Non-Smokers, n (%)	178(51.90)	17(36.20)	Reference
Smokers, n (%)	165(48.10)	30(63.80)	1.95 (1.12–3.41)	0.019

Note: ^aModel was adjusted for Alcohol.

Table 4 Logistic Regression Analysis of the Influence of Smoking on the Inconsistency Between QFT Detection and Tuberculosis Etiology in Patients Aged ≥65 Years

		Consistent Group (n=320)	Inconsistent Group (n=62)	Model 1		Model 2
				OR (95% CI)	P-value	OR (95% CI)	P-value
Age≥65 years	Non-Smokers, n (%)	246 (76.90)	36 (58.10)	Reference		Reference
	Smokers, n (%)	74 (23.10)	26 (41.90)	2.40 (1.36–4.23)	0.002	2.40 (1.30–4.45)	0.005

Notes: Model 1 was a crude model. Model 2 was adjusted for Hypoproteinemia, Chronic liver, Cardiovascular diseases and lymphocyte percentage after box-cox transformation.