Skip to main content

Table 2 Variables associated with (1) bacterial load (16S rRNA copies) and (2) Shannon diversity indices in multivariate linear mixed effect models

From: Sputum bacterial load and bacterial composition correlate with lung function and are altered by long-term azithromycin treatment in children with HIV-associated chronic lung disease

 

Bacterial load (16S rRNA copies)

Shannon diversity index

*Variable

Levels

aNumber of observations (n = 875)

Participants (n = 346)

1Adjusted co-efficient (95% CI)

1p value

2Adjusted co-efficient (95% CI)

2p value

Visit

Placebo at week 48

150 (17.1%)

150 (43.4%)

Reference

 

Reference

 

AZM at week 48

154 (17.6%)

154 (44.5%)

− 0.46 [− 0.63; − 0.29]

< 0.0001

0.25 [0.07; 0.42]

0.01

Placebo at week 72

117 (13.4%)

117 (33.8%)

Reference

 

Reference

 

AZM at week 72

123 (14.1%)

123 (35.5%)

− 0.19 [− 0.38; 0.0]

0.051

0.2 [0.01; 0.40]

0.04

Site

Malawi

233 (26.6%)

106 (30.6%)

Reference

 

Reference

 

Zimbabwe

642 (73.4%)

240 (69.4%)

0.3 [0.11; 0.49]

0.003

0.27 [0.06; 0.47]

0.01

Medical Research Council dyspnoea score at baseline

1

479 (54.7%)

184 (53.2%)

  

Reference

 

2

316 (36.1%)

126 (36.4%)

  

0.26 [0.1; 0.42]

0.04

3

53 (6.1%)

23 (6.6%)

  

0.16 [− 0.18; 0.49]

 

4

23 (2.6%)

11 (3.2%)

  

0.25 [− 0.21; 0.71]

 

5

4 (0.5%)

2 (0.6%)

  

0.52 [− 1.05; 2.09]

 

Forced expiratory volume in 1 s (FEV1) z-score (FEV1z)

 

862 (98.5%)

346 (100%)

− 0.09 [− 0.16; − 0.02]

0.02

0.19 [0.12; 0.27]

< 0.001

Ever treated for tuberculosis before enrolment

No

609 (69.8%)

248 (71.9%)

Reference

 

Reference

 

Yes

263 (30.2%)

97 (28.1%)

0.17 [0.03; 0.32]

0.02

− 0.19 [− 0.34; − 0.04]

0.02

  1. aThe difference between the number of observations and the total (875) represent number of missing observations for that variable
  2. 1The estimate of coefficient with 95% confidence intervals and p values were obtained from multivariate linear mixed effect model with participant included as a random effect, trial arm, visit and trial arm: visit interaction term and selected variables ( visit, site, Medical Research Council dyspnoea score at baseline, forced expiratory volume in 1 s (FEV1) z-score (FEV1z), previous tuberculosis treatment prior to enrolment, age group at baseline, adherence, site, sex, season of sampling, viral load suppression at baseline, height-for-age z-score, cotrimoxazole prophylaxis at baseline, previous admission to chest problems in the year preceding enrolment, duration of ART at baseline) as explanatory variables and 16S rRNA copies of the sputum samples as dependent variable
  3. 2The estimate of coefficient with 95% confidence intervals and p values were obtained from multivariate linear mixed effect model with participant included as a random effect, trial arm, visit and trial arm: visit interaction term and selected variables ( visit, Medical Research Council dyspnoea score at baseline, forced expiratory volume in 1 s (FEV1) z-score (FEV1z), previous tuberculosis treatment prior to enrolment, age group at baseline, adherence, site, sex, season of sampling, viral load suppression at baseline, height-for-age z-score, any events during intervention, cotrimoxazole prophylaxis at baseline, previous admission to chest problems in the year preceding enrolment, duration of ART at baseline) as explanatory variables and Shannon indices of the sputum samples as dependent variable. Any event refers to either acute respiratory exacerbation; additional antibiotics other than interventional drug or cotrimoxazole; or hospitalisation during intervention. Forced expiratory volume in 1 s (FEV1)/forced vital capacity z-score, forced vital capacity z-score (FVCz), FEV1 percentage predicted (FEVpcpred), FVC percentage predicted (FVCpcpred) were excluded from the final models because of collinearity with FEV1 z-score. Viral load and CD4 counts were excluded from the final models because data was not collected at 72 weeks, the values at baseline were used instead. Weight-for-age z-score, body mass index for age z-score were excluded from the final models because of collinearity with height-for-age-z-score
  4. 3Any event refers to either acute respiratory exacerbation; additional antibiotics other than interventional drug or cotrimoxazole; or hospitalisation during intervention
  5. *Full table with coefficients and 95% confidence intervals from univariate linear mixed effect models are included supplementary materials
Back to article page

Microbiome

ISSN: 2049-2618

Contact us