Abstract
Objective:
Evaluate effects of TB-HIV co-treatment on clinical and growth outcomes in children with HIV (CHIV).
Design:
Longitudinal study among Kenyan hospitalized ART-naive CHIV in the PUSH trial (NCT02063880).
Methods:
CHIV started ART within 2 weeks of enrollment; Anti-TB therapy was initiated based on clinical and TB diagnostics. Children were followed for 6 months with serial viral load, CD4%, and growth assessments (weight-for-age [WAZ], height-for-age [HAZ], and weight-for-height [WHZ]). TB-ART treated and ART-only groups were compared at 6-months post-ART for undetectable viral load [VL] (<40 c/ml), CD4% change, and growth using generalized linear models, linear regression, and linear mixed-effects models, respectively.
Result:
Among 152 CHIV, 40.8 % (62) were TB-ART treated. Pre-ART, median age was 2.0 years and growth was significantly lower, and VL significantly higher in the TB-ART vs. ART-only group. After 6 months on ART, 37.2% of CHIV had undetectable VL and median CD4% increased by 7.2% (IQR 2.0%−11.6%) with no difference between groups. The TB-ART group had lower WAZ and HAZ over 6 month follow-up (WAZ −0.81 [95% CI: −1.23, −0.38], p<0.001; HAZ −0.15 [95% CI: −0.29, −0.01], p=0.030) and greater rate of WAZ increase in analyses unadjusted and adjusted for baseline WAZ (unadj 0.62 [95% CI: 0.18, 1.07, p=0.006] or adj 0.58 [95% CI: 0.12, 1.03, p=0.013]).
Conclusion:
TB-HIV co-treatment did not adversely affect early viral suppression and CD4 recovery post-ART. TB-ART treated CHIV had more rapid growth reconstitution, but growth deficits persisted, suggesting need for continued growth monitoring.
Background
Children living with HIV (CHIV) have a high risk of Mycobacterium tuberculosis (Mtb) infection (1) and TB-related morbidity (2–5) and mortality (6,7). Once TB-exposed, CHIV are eight times more likely to develop TB disease compared to HIV-unexposed uninfected (HUU) children (8). TB disease is the leading cause of death (40%) and admission to the hospital (18%) in people living with HIV (9). Africa accounted for 30.5% of the global pediatric TB burden in 2018 (10) and 37.0% of TB-related deaths among children under 15 years of age. Approximately 36.0% of pediatric TB deaths in this region were among CHIV (11), representing 78.0% of TB deaths in CHIV worldwide. Treatment of both TB and HIV is vital to reducing TB-related mortality (12,13) in CHIV. However, TB-HIV co-treatment poses management challenges, including high pill burden, drug toxicity, drug-drug interactions, and immune reconstitution inflammatory syndrome (9,14).
Rifampicin, a component of standard treatment for TB, is a potent inducer of cytochrome P450 isoenzymes (CYP) and p-glycoprotein (PGP) (15–17) – a primary mechanism of drug elimination that reduces plasma concentrations of concomitantly administered protease inhibitors (PI), and non-nucleoside reverse transcriptase inhibitors (NNRTI) (16). Thus, rifampicin may lower plasma ART concentrations and result in inferior virologic and immunological responses (14–16). CHIV have a higher incidence of virologic failure than adults due to dosing, unpalatable formulations, higher likelihood of drug resistance, adherence, social support, and developmental stage (18,19).
TB increases the metabolic rate or resting energy expenditure and impairs substrate handling, facilitating ingested amino acid oxidation instead of protein anabolism (20,21). Thus, TB contributes to malnutrition – mainly wasting and weight loss (20,21). Children typically regain weight after six months of TB treatment (22). Similarly, CHIV have substantially reduced growth (including wasting or weight loss) (23) as a consequence of reduced energy intake and increased metabolic rates associated with HIV infection and drug side effects (24). ART restores the growth of CHIV (23,25). TB/HIV co-infection may synergistically exacerbate children’s growth deficits (21). While TB-HIV co-treatment improves child survival (12,13), it is unclear how TB-ART treatment influences growth in CHIV receiving ART. Growth trajectories between CHIV with and without TB-HIV co-treatment may differ due to the differential growth effects of TB disease and its’ resolution or due to side effects of TB medications. The few studies that have evaluated HIV-TB co-infected children had inconsistent findings (26,27) regarding the impact of co-treatment on viral suppression or immunologic responses in CHIV. A better understanding of viral suppression and the pattern of growth reconstitution during TB-HIV co-treatment can inform treatment approaches and nutritional monitoring and supplementation strategies.
Methods
Study design:
This analysis used data from the Pediatric Urgent Start of Highly Active Antiretroviral Treatment (PUSH) study. PUSH was a prospective unblinded randomized controlled trial of urgent (within 48 hours of enrollment) versus post-stabilization (7 to 14 days after enrollment) ART between April 2013 and November 2015 (NCT02063880).
Study population:
The PUSH study included ART-naive hospitalized CHIV aged 0–12 years of age. Children were enrolled in four hospitals in Kenya from April 2013 to November 2015 and followed for 6 months. Children with suspected or confirmed central nervous system infections were excluded (28). For this secondary analysis, participants were classified based on TB-HIV co-treatment as TB-ART vs. ART-only.
The PUSH study obtained ethical approval from Kenyatta National Hospital (KNH)/University of Nairobi (UoN) Ethics Research Committee (ERC), Kenya Pharmacy and Poisons Board (PPB), and the University of Washington (UW) Institutional Review Board. Caregivers of participants gave written informed consent before participating in the study.
Outcome variables
Undetectable viral load:
Viral load (VL) was measured at baseline and the 6th month of follow-up. VL suppression was defined as VL of <40 HIV RNA copies/ml at a 6-month follow-up.
CD4 recovery:
CD4% was measured at baseline and 6th month of follow-up. Mean CD4% at 6-months and change in CD4% from baseline were compared.
Growth:
Weight and height (length) were collected 9 times (enrollment, at ART initiation and then, two weeks, and monthly thereafter through 6 months post-ART). Weight-for-age z-score [WAZ], weight-for-height z-score [WHZ], and height/length-for-age z-score [HAZ]) were defined using WHO child growth standards (29). Growth faltering was defined as underweight (WAZ<−2), stunting (HAZ<−2), or wasting (WHZ <−2). WHZ was defined only for children younger than ten years of age and provided the result as a supplement.
Exposure variable
TB-ART treatment:
TB-HIV treated children were defined as the exposed group and ART-only treated children were defined as the control comparator group. The TB-ART treated group received treatment both for TB and HIV. In this study, TB was diagnosed using multiple diagnostic methods, including sputum and gastric aspirate Xpert and culture, stool Xpert, urine LAM, chest X-ray, and clinical signs (28,30). Identified TB cases received TB treatment (rifampicin, isoniazid, pyrazinamide, and ethambutol for a 2-month intensive phase and rifampicin and isoniazid for the 4-month continuation phase per Kenyan and WHO guidelines) (28). ART was initiated with first-line regimens based on Kenyan 2011 guidelines that included abacavir and lamivudine in combination with either efavirenz (children aged three or more than ten kilograms (kg)), nevirapine (children under three years or less than ten kilograms) who had not been exposed to nevirapine for PMTCT), or lopinavir/ritonavir (LPV/r) (children under three years if exposed to nevirapine for PMTCT) (28,31). As of September 2014, all children under three years old were prescribed LPV/r-based regimens regardless of exposure to NVP (28). At the same time, hospital staff initiated the use of cotrimoxazole prophylaxis for all children under three years old. CHIV in the TB-ART group on LPV/r received boosted ritonavir to compensate for potential drug-drug interactions.
Covariates:
Potential confounding factors include baseline maternal characteristics such as maternal age in years and educational status (primary or less, secondary or above) and children’s baseline characteristics including sex, age (<2, 2-<5, and 5+ years of age), baseline CD4%, VL (log10 RNA copies/mL), underweight, stunting, and wasting.
Statistical Analysis:
Mean and standard deviation (SD) were used to describe normally distributed continuous variables. Median and interquartile range (IQR) were used to describe skewed distributions, while frequency and percentage were used to describe categorical variables. We compared baseline caregiver and infant characteristics between TB-ART treated and ART-only treated groups, using a two-sided t-test for continuous variables and the Pearson χ2 test (Fisher’s exact test if assumptions were not met) for categorical variables.
We fitted a multivariable generalized mixed linear model (GLM) using Poisson regression with a log link (to estimate relative risk) to compare VL suppression and multivariable linear regression to compare CD4% at the end of the study between TB-ART treated and ART-only CHIV.
We used linear mixed-effects models with autoregression correlation structure, random intercept for subjects, and random slope for follow-up time to compare growth (WAZ, WHZ, HAZ) between TB-ART and ART-only participants. Since TB diagnosis is associated with lower weight or poorer growth, and lower baseline growth may be associated with greater reconstitution, we conducted analyses both with and without adjustment for baseline growth. We fitted the following three mixed models for all growth outcomes - WAZ, WHZ, HAZ: (Model 1) unadjusted model, (Model 2) adjusted for preselected factors, maternal age and education, children’s sex, children’s age, baseline viral load, and CD4%, and (Model 3) including Model 2 variables and relevant baseline growth marker (WAZ, WHZ, or HAZ) for each outcome. Interaction terms were fitted between TB-ART treatment and follow-up time to examine the rate of change in WAZ and HAZ in Models 2 and 3 for WAZ and HAZ analyses.
Kaplan-Meier survival analysis was used to compare the time from baseline to the first event of increasing WAZ, HAZ, or WHZ by at least 0.5 z-scores from the baseline to the 6 months TB treatment period. Univariate and multivariable Cox proportional hazards regression models were fitted to compare the probability of improving WAZ, HAZ, and WHZ by at least 0.5 z-scores after treatment between the TB-ART treated and ART-only children.
Approach to missing data
Some visits had missing data:
for example, WAZ (6%−16% missing per visit), VL (7% baseline, 15% at 6 months), and 6th month CD4% (12%). Multiple imputations by chained equations (MICE) were used to manage missing data. In addition to covariates of the current study – TB treatment, child’s age, sex, CD4%, ART regimen, baseline WAZ, HAZ, caregiver’s age, and education (32,33). We imputed the data 25 times. Pooled parameter estimates and their standard errors were calculated according to Rubin’s rules to account for the between- and within-imputation variance (32,33). We employed one single imputation model to obtain imputed values for outcomes – missing WAZ and HAZ at each visit during follow-up, baseline WAZ, and baseline viral load. We also ran another imputation model to obtain imputed data for the 6th month CD4% and VL by including only participants who were alive by the end of the study. In the imputation models, we specified the appropriate distributions for each of the variables in the model. In the presence of missing data, the multiple imputation approach should yield unbiased estimates assuming data are missing at random (MAR).
P-values of <0.05 cut-offs and 95% confidence intervals were used to determine statistical significance. R version 3.5.1 and STATA version 16 statistical software were used for analyses.
Results
Among the 181 participants in the PUSH RCT, 2 children who had confirmed TB disease but did not start TB treatment (due to early death) and 27 children with no follow-up after baseline (5 lost-to-follow-up and 22 deaths) were excluded. Thus, 152 eligible CHIV children were included in the current analyses: 62 TB-ART co-treated and 90 ART-only. Of these 152 children, 15 died before study completion.
Maternal and children’s characteristics
Participants’ median (IQR) age was 2.0 (1.1–6.1) years, 53.3% (81/152) were males, 58.6% (89/152) started NNRTI-based ART regimens, 67.1% (102/152) were WHO stage 3 or 4, and median (IQR) baseline CD4% was 14% (IQR 9.0% – 22.3%). Ten percent (15/152) of CHIV’s mothers were employed (Table 1).
Table 1:
Baseline maternal and child demographic and clinical characteristics
| Variable | Treatment | |
|---|---|---|
| TB-ART cotreated (n=62) | ART-only (n=90) | |
|
| ||
| Maternal baseline characteristics | ||
| Mean age in years (IQR) | 29 (24.0 – 33.0) | 29 (24.0–32.0) |
| Marital status – no. (%) | ||
| Living with partner(s) | 40 (38.8) | 63 (61.2) |
| Separated, Widow, widower | 17 (43.6) | 22 (56.4) |
| Single | 4 (44.4) | 5 (55.6) |
| Maternal education – no. (%) | ||
| Primary or less | 43 (69.4) | 56 (62.2) |
| Secondary or above | 19 (30.1) | 34 (37.8) |
| Maternal employment – no./total no. (%) | ||
| Employed | 3 (20.0) | 12 (80.0) |
| Unemployed | 36 (46.2) | 42 (53.9) |
| Other | 23 (38.9) | 36 (61.0) |
| Children’s baseline characteristics | ||
| Age – no (%) | ||
| Less than two years of age | 28 (38.9) | 44 (61.1) |
| 2-<5 years of age | 14 (36.8) | 24 (63.2) |
| 5+ years of age | 20 (47.6) | 22 (52.4) |
| Female sex – no (%) | 30 (48.4) | 41 (45.6) |
| ART regimen | ||
| NNRTI-based | 35 (38.5) | 56 (61.5) |
| LPV/r-based | 27 (44.3) | 34 (55.7) |
| Mean CD4% (SD)* | 13.6 (9.5) | 17.0 (10.8) |
| Severe immune suppression – no (%)ⴃ | 45 (73.8) | 57 (63.3) |
| Mean log10 viral load (IQR)* | 5.7 (5.3–6.3) | 5.4 (4.9–6.1) |
| Underweight– no (%) * | 45 (76.3) | 45 (52.3) |
| Mean WAZ (SD)* | −3.3 (1.7) | −2.2 (1.5) |
| Stunting (HAZ <−2) – no (%) | 40 (65.6) | 50 (55.6) |
| Median HAZ (IQR) | −2.6 (1.8) | −2.3 (1.6) |
Severe immune suppression – CD4% (<12 months:<25%, 12–35 months:<20%,>36 months: <15%) or, in absence of CD4%, CD4 cell count (<12 months: <1500 cells/ml, 12–35 months: <750 cells/ml, >36 months <350 cells/ml.
These are baseline characteristics that are unevenly distributed between the exposure groups, p-value <0.05.
At baseline, the TB-ART group had higher mean HIV VL (5.7 vs 5.4 log10 c/ml, p=0.023, lower mean CD4% (12.5% vs 15.2%, p=0.043, and a higher proportion of children who were WHO stage 3 or 4 (62.3% vs 51.1%, p<0.001, underweight (72.6% vs 50.0%, p=0.014) than the ART-only group.
Effect of TB-HIV co-treatment on change in CD4 percent and VL suppression
Following 6-months of ART, overall, 37.2% (45/121) CHIV had undetectable VL <40 RNA copies/ml, 42.9% (21/49) in TB-ART group and 33.3% (24/72) in ART-only group (p=0.563). Overall, 63.6% (77/121) and 71.1% (86/121) were virally suppressed to <400 c/ml and <1000 c/ml respectively. The median CD4% increase during 6-month follow-up was 7.2% (2.0–11.6); 7.3% (1.6–11.9) in TB-ART group and 7.0% (2.7, 11.6) in the ART-only group (p=0.823, for difference). In univariate models, female sex and higher baseline HIV VL were associated with significantly lower frequency of undetectable VL at 6 months post-ART. In univariate and multivariate models, adjusted for maternal age, education, child sex, age, baseline VL, and CD4%, the TB-ART group had a comparable proportion with undetectable HIV VL at 6-months follow-up compared to ART-only group (aRR=1.13 [95% CI: 0.96, 1.33], p = 0.153) (Table 2). Similarly, CD4% increase post-ART did not differ significantly between the TB-ART and ART-only groups (β=−0.84 [95% CI: −3.64, 1.97], p=0.555) (Table 3). Because PI-based ART may particularly interact with TB regimens, we examined the subset of CHIV receiving PI-ART and found no significant differences in viral suppression frequency (50% [14/28] vs 40.0% [18/45] p=5.63) or mean change in CD4% (5.4% vs 6.0% p=0.805) between TB-ART vs ART-only groups in CHIV receiving PI-ART.
Table 2:
Cofactors of viral load suppression and CD4% increase in hospitalized CHIV initiating ART
| Model | Treatment | Viral load suppressed (<40 copies/mL) | CD4% | ||
|---|---|---|---|---|---|
| cRRσ (95% CI) | aRRꞩβ (95% CI) | cCoefficient¤ | aCoefficient§β | ||
| Overall | TB-ART co-treated (ref: ART-only) | 1.25 (0.79, 1.99) | 1.29 (0.79, 2.11) | −2.31 (−5.59, 0.97) | −0.83 (−3.68, 2.01) |
| Maternal age in years | 1.01 (0.98, 1.05) | 1.01 (0.97, 1.05) | −0.12 (−0.38, 0.14) | −0.14 (−0.37, 0.09) | |
| Maternal education: secondary or above (ref: Primary or less) | 0.73 (0.43, 1.24) | 0.90 (0.52, 1.56) | −3.86 (−7.18, −0.54)** | −0.23 (−3.40, 2.94) | |
| Female children (ref: male) | 0.61 (0.36, 1.01) | 0.63 (0.37, 1.07) | 4.24 (1.10, 7.38)** | 2.44 (−0.34, 5.24) | |
| Age of children | |||||
| <2 years of age | 1 | 1 | 1 | 1 | |
| 2-<5 years of age | 1.92 (1.01, 3.63)* | 0.86 (0.25, 2.90) | −2.03 (−5.98, 1.91) | −1.51 (−6.66, 3.64) | |
| 5+ years of age | 2.17 (1.20, 3.91)* | 0.74 (0.20, 2.72) | −2.82 (−6.65, 1.00) | −0.87 (−7.26, 5.52) | |
| Baseline log10 viral load | 0.90 (0.85, 0.96)** | 0.92 (0.85, 0.98) | |||
| Baseline CD4 percent | 0.55 (0.43, 0.67)*** | −1.02 (−6.48, 4.44) | |||
| ART regimen: LPV/r-based (ref: NNRTI-based) | 0.53 (0.31, 0.92)* | 0.50 (0.16, 1.52) | 1.43 (−2.13, 4.99) | 0.56 (0.41, 0.70)** | |
|
| |||||
| PI regimen | TB-ART co-treated (ref: ART-only) | 1.09 (0.64, 1.88) | 1.08 (0.63, 1.87) | 0.50 (−4.82, 5.82) | 0.27 (−4.98, 5.27) |
cRR – crude relative risk.
aRR – adjusted relative risk.
cCoefficient - crude coefficient.
aCoefficient – adjusted coefficient.
Multivariable models adjusted for all other variables (TB treatment, maternal age, education, baseline viral load/CD4 percent, children sex, and age).
P-value <0.05.
P-value<0.01.
P-value <0.001.
Table 3:
Cofactors of WAZ and HAZ reconstitution in hospitalized CHIV initiating ART
| WAZ¶ | HAZ¥ | |||||
|---|---|---|---|---|---|---|
| Model 1 cCoefficient¤ | Model 2 aCoefficient§ | Model 3 aCoefficientβ | Model 1 cCoefficient¤ | Model 2 aCoefficient§ | Model 3 aCoefficientβ | |
| TB-ART co-treated | −0.89 (−1.31, −0.46)*** | −0.91 (−1.35, −0.47)*** | −0.06 (−0.28, 0.16) | −0.59 (−1.08, −0.10)* | −0.56 (−1.06, −0.07)* | −0.21 (−0.36, −0.06)** |
| Maternal age in years | −0.00 (−0.04, 0.03) | −0.01 (−0.05, 0.03) | 0.00 (−0.02, 0.01) | 0.01 (−0.03, 0.05) | −0.01 (−0.05, 0.04) | 0.00 (−0.01, 0.01) |
| Maternal education: secondary or above (ref: Primary or less) | −0.04 (−0.48, 0.40) | 0.00 (−0.45, 0.46) | 0.02 (−0.19, 0.23) | 0.11 (−0.60, 0.38) | −0.17 (−0.68, 0.34) | 0.04 (−0.11, 0.20) |
| Female children (ref: male) | 0.23 (−0.21, 0.67) | 0.20 (−0.22, 0.62) | 0.22 (0.03, 0.42)* | 0.19 (−0.30, 0.68) | 0.18 (−0.30, 0.66) | 0.00 (−0.15, 0.14) |
| Age of children | ||||||
| <2 years of age | 1 | 1 | 1 | 1 | 1 | |
| 2-<5 years of age | 0.46 (−0.08, 0.99) | 0.36 (−0.16, 0.88) | 0.31 (0.08, 0.54)* | 0.10 (−0.59, 0.61) | −0.07 (−0.66, 0.53) | −0.06 (−0.24, 0.11) |
| 5+ years of age | 0.34 (−0.19, 0.87)* | 0.36 (−0.17, 0.89) | 0.17 (−0.08, 0.41) | 0.49 (−0.09, 1.07) | 0.48 (−0.12, 1.08) | 0.05 (−0.13, 0.23) |
| Baseline log10 viral load | −0.29 (−0.55, −0.04) | −0.04 (−0.14, 0.05) | −0.02 (−0.06, 0.02) | −0.44 (−0.72, −0.15)* | −0.11 (−0.22, −0.01)* | −0.03 (−0.06, 0.01) |
| Baseline WAZ/HAZ | 0.69 (0.62, 0.76)8*** | ⴃ | 0.71 (0.65, 0.77)*** | 0.84 (0.80, 0.89)*** | 0.84 (0.80, 0.88)*** | |
WAZ – weight-for-age.
HAZ – height-for-age.
cCoefficient - crude coefficient.
aCoefficient – adjusted coefficient for all other variables (TB treatment, maternal age, education, baseline viral load, CD4 percent, children sex, age).
Coefficient – adjusted coefficient including baseline WAZ for WAZ modes/ baseline HAZ for HAZ models.
P-value <0.05.
P-value<0.01.
P-value <0.001.
Variable not adjusted
In sensitivity analysis, we compared VL outcomes in the TB-ART and ART only groups using varied VL cut-offs (≥ 400 copies/mL vs <400 copies/mL) and also (≥ 1000 copies/mL vs <1000 copies/mL) and found the similar results as when VL suppression was defined as <40 copies/mL (Supplemental Table 1).
Growth between pre-ART and 6-months post-ART
Overall, participants’ unadjusted average WAZ and HAZ scores increased substantially and significantly by 1.05 (95% CI: 0.81, 1.30) and 0.30 (95% CI: 0.14, 0.45) z-scores, respectively, from baseline to 6-month follow-up. In the TB-ART group, the average WAZ score increased significantly by 1.42 (95% CI: 1.04, 1.80), while HAZ did not (0.07 (95% CI: −0.17, 0.31). In the ART-only group, the average WAZ and HAZ scores increased significantly by 0.72 (95% CI: 0.47, 0.98) and 0.44 (95% CI: 0.24, 0.64), respectively.
Effect of TB-HIV co-treatment on change in WAZ
After adjustment for maternal age, education, and children’s sex, age, and baseline VL, but not baseline WAZ, the TB-ART group had WAZ scores that were 0.91 less than ART-only group (β=−91 [95% CI: −1.35, −0.47], p<0.001) (Table 3). TB-ART treated children had a significantly more rapid increase in WAZ during the treatment period than the ART-only children. TB-ART treated children gained 0.07 more change in WAZ score per month during the treatment period than ART-only treated children (β=0.07 [95% CI: 0.02, 0.13], p=0.006) (Figure 1).
Figure 1:
The left two curves are scatter plot of change in WAZ and HAZ over time by TB treatment
(The Lowess curves represent growth reconstitution following TB and /or ART treatment. The middle and right figures show adjusted change in WAZ and HAZ without and with baseline growth adjusted, respectively.)
Because the TB-ART group started from lower growth at baseline, their more rapid growth reconstitution could have been due to baseline growth differences. To understand how much of the growth reconstitution differences were due to baseline differences in growth, we adjusted for both confounding variables and the baseline pre-ART growth. In these adjusted analyses, there was no statistically significant difference in WAZ between the TB-ART and the ART groups during the treatment period (β=−0.06 [95% CI: −0.28, 0.16], p=0.575) (Table 3). However, the rate of growth recovery remained significantly faster in TB-ART treated children than in ART-only children. After adjustment for baseline WAZ, TB-ART treated children gained 0.08 higher change in WAZ scores per month during the treatment period than ART-only children (β=0.08 [95% CI: 0.03, 0.13], p=0.003) (Figure 1).
Effect of TB-HIV co-treatment on change in HAZ
After adjustment for maternal age, education, child sex, age, and baseline VL, but not baseline HAZ, CHIV receiving TB-ART had a significant lower HAZ during the treatment period (β=−0.56 [95% CI: −1.06, −0.07], p =0.027) (Table 3). The TB-ART group had a statistically significantly slower HAZ increase per month than ART-only children (β=−0.04 [95% CI: −0.08, −0.01]), p =0.026) (Figure 1).
Adjusting for confounding variables and baseline pre-ART HAZ, there was a statistically significant association between TB-HIV co-treatment and HAZ during the treatment period. TB-ART treated children had on average 0.21 lower HAZ scores during the study period than ART-only children (β=−0.21 [95% CI: −0.36, −0.06], p=0.003) (Table 3). The TB-ART group had a statistically significantly slower HAZ increase per month than ART-only children (β=−0.05 [95% CI: −0.08, −0.01], p=0.007) (Figure 1).
Effect of TB-HIV co-treatment on change in WHZ
After adjustment for maternal age, education, and children’s sex, age, and baseline VL, but not baseline WHZ, the TB-ART group had WHZ scores that were 0.93 less than ART-only group (β=−93 [95% CI: −1.39, −0.47], p<0.001) (Supplemental Table 2). TB-ART treated children had a significantly more rapid increase in WHZ during the treatment period than the ART-only children. TB-ART treated children gained 0.18 greater change in WHZ score per month during the treatment period than ART-only treated children (β=0.18 [95% CI: 0.11, 0.26], p<0.001).
After adjusting for both confounding variables and the baseline pre-ART growth, there was no statistically significant difference in WHZ between the TB-ART and the ART groups during the treatment period (β=−0.08 [95% CI: −0.41, 0.24], p=0.616) (Supplemental Table 2). However, the rate of growth recovery remained significantly faster in TB-ART treated children than in ART-only children. After adjustment for baseline WHZ, TB-ART treated children had a 0.19 higher change in WHZ scores per month during the treatment period than ART-only children (β=0.19 [95% CI: 0.11, 0.27], p<0.001).
The probability of improving WAZ, HAZ, and WHZ by at least 0.5 Z-scores among CHIV who were underweight or stunted at baseline.
We examined time to growth recovery during 6-month treatment. Figure 2 shows the cumulative probability of improving WAZ, HAZ, and WHZ by at least 0.5 Z-score among CHIV who were <2 z-scores for the relevant growth parameter at baseline (Supplemental Table 3). Children treated with TB-HIV had no significant difference in their chances of experiencing a minimum of 0.5 improvements in WAZ (aHR=1.21 [95% CI: 0.78, 1.89]) and HAZ (aHR=0.65 [95% CI: 0.38, 1.10]) scores compared to children treated with ART alone.
Figure 2:
The figure shows cumulative probability of weight-for-age (WAZ), weight-for-length (WLZ), and length-for-age (LAZ) improved by at least 0.5 by TB-treatment
Discussion
In this study, among hospitalized CHIV newly initiating ART, TB-ART co-treatment resulted in similar viral suppression and immune reconstitution during the 6-month follow-up as ART treatment alone. CHIV receiving TB-HIV co-treatment gained weight (WAZ) more rapidly but still had significantly lower weight and height for age than ART-only children by the end of TB treatment period. TB-ART treated children lagged height reconstitution despite robust weight reconstitution compared to CHIV receiving ART only. Overall, our data suggest good clinical outcomes but persistent growth deficits in TB-ART treated children that will require continued growth monitoring and potential nutritional interventions.
We found that TB-ART co-treatment did not compromise immune reconstitution and viral load suppression compared to ART-only group. There are conflicting data on the impact of TB co-treatment on viral suppression and immune recovery from other studies, some using different VL cut-offs (26,27). One study from South Africa, which used a <50 c/ml cut-off (similar to our study) found no difference in the proportion of CHIV with undetectable VL receiving TB-ART and those on ART-only (26), findings consistent with our results. Although some studies which used a higher VL cut-off (400 c/ml) found that TB-ART co-treatment decreased frequency of viral suppression (27), we failed to find an association in sensitivity analyses using this higher cut-off. Both studies, however, have noted that the impact of TB co-treatment varies by ART regimen, with the limited impact of TB co-treatment on VL suppression among CHIV receiving NNRTI-based regimens, in contrast to CHIV on PI-based regimens which may have inferior virologic and immunologic responses with TB co-treatment (26,27). In our study, CHIV receiving PI-ART had a lower frequency of viral suppression than those receiving NNRTI-based regimens in univariate but not in multivariate analyses and was associated with decreased CD4 recovery in multivariate analyses. However, in the subset of CHIV receiving PI-based regimens, there was no difference in viral suppression and immune recovery in those receiving TB-ART versus ART. However, we had limited statistical power in this sub-group analysis. CHIV in the TB-ART group on LPV/r, received boosted ritonavir to compensate for potential drug-drug interactions, which may have attenuated the impact of TB co-treatment on viral suppression or immune reconstitution.
As expected, ART resulted in rapid growth reconstitution in all CHIV, including those receiving TB-ART (34). TB-ART treated children had lower WAZ and HAZ at baseline and despite more rapid growth reconstitution, had residual deficits throughout the treatment period. Children treated for TB and HIV have a double burden of disease that contributes to their weight loss (35). TB-ART treated children gained weight more rapidly than ART-only children, on average about 0.08 WAZ scores more per month than the ART-only group. In children treated for TB, weight gain is an important indicator of prognosis (36). Rapid growth reconstitution is encouraging but residual growth deficits persisted at 6 months, underscoring the importance of continued growth and nutritional monitoring and support.
The TB-ART treated group had more rapid weight reconstitution but lower height reconstitution for their ages than ART-only treated group. This may be due to some persistent metabolic costs of TB and HIV during the first 6 months of treatment (20,21,24). Height deficits might eventually result in stunting, with long-term implications on late childhood and adult obesity and related metabolic disorders (37).
Our study has strengths and limitations. In this study, about 70% of CHIV patients who were on TB-ART, started ART and TB treatment within 8 days, which provides information for clinicians and policymakers about the effects of ART and TB co-treatment at ART initiation. We were able to assess important HIV outcomes, viral suppression, CD4 reconstitution, and growth parameters during the critical first 6 months post-ART but not longer-term outcomes. This cohort of CHIV who were hospitalized had high mortality, which decreased the sample size and may have led to selection bias in estimating the impact of TB-HIV co-treatment on viral suppression, immune recovery, and growth. Follow-up was limited to 6 months which although encompasses sufficient time for viral suppression and completion of TB treatment but does not span the entire period of potential growth recovery. If there was a meaningful difference below what this study was powered to identify, a type II error (the severity of adverse effect of TB-HIV cotreatment could be underestimated) might have occurred.
Conclusion
TB co-treatment did not affect immune reconstitution (CD4% recovery) and viral load suppression at 6-months post-ART among CHIV newly initiating ART. TB-ART treated children had more rapid weight gain during the follow-up period but still had lower weight and height than ART-only treated children by the end of the 6-month follow-up period. TB-ART treated children had robust weight reconstitution, but lagged height reconstitution compared to CHIV receiving ART-only. Despite good clinical outcomes and rapid growth reconstitution, growth deficits are persistent in TB co-treated children, which will require continued growth monitoring and nutritional interventions.
Supplementary Material
Acknowledgments
Our sincere gratitude goes out to the study participants and their families. We are also grateful to the funders. The parent study was supported by the National Institute of Child Health and Human Development (NICHD) at the National Institutes of Health (NIH) (R01 HD023412 and K24 HD054314-06 to GJS, K12 HD000850 to LMC). ASC had a diversity supplement grant from NIH (NIH/NIAID 1R01AI142647), and SML was supported by NIH/NIAD K23 grant (NIH/NIAID 1K23AI120793-01A1).
Footnotes
Declaration of interests:
None declared.
References
- 1.Marquez C, Chamie G, Achan J, Luetkemeyer AF, Kyohere M, Okiring J, et al. Tuberculosis Infection in Early Childhood and the Association with HIV-exposure in HIV-uninfected Children in Rural Uganda. Pediatric Infectious Disease Journal. 2016. May 1;35(5):524–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Rupérez M, González R, Maculuve S, Quintó L, López-Varela E, Augusto O, et al. Maternal HIV infection is an important health determinant in non-HIV-infected infants. AIDS [Internet]. 2017. Jul 17 [cited 2020 Nov 25];31(11):1545–53. Available from: http://journals.lww.com/00002030-201707170-00006 [DOI] [PubMed] [Google Scholar]
- 3.Marais BJ, Schaaf HS. Tuberculosis in children. Cold Spring Harb Perspect Med. 2014;4(9):4:a017855. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Afran L, Garcia Knight M, Nduati E, Urban BC, Heyderman RS, Rowland-Jones SL. HIV-exposed uninfected children: a growing population with a vulnerable immune system? Clin Exp Immunol [Internet]. 2014. Apr [cited 2020 Nov 25];176(1):11–22. Available from: 10.1111/cei.12251 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Koyanagi A, Humphrey JH, Ntozini R, Nathoo K, Moulton LH, Iliff P, et al. Morbidity among human immunodeficiency virus-exposed but uninfected, human immunodeficiency virus-infected, and human immunodeficiency virus-unexposed infants in zimbabwe before availability of highly active antiretroviral therapy. Pediatric Infectious Disease Journal. 2011. Jan;30(1):45–51. [DOI] [PubMed] [Google Scholar]
- 6.Shapiro RL, Lockman S, Kim S, Smeaton L, Rahkola JT, Thior I, et al. Infant morbidity, mortality, and breast milk immunologic profiles among breast-feeding HIV-infected and HIV-uninfected women in Botswana. Journal of Infectious Diseases [Internet]. 2007. [cited 2020 Nov 25];196(4):562–9. Available from: https://academic.oup.com/jid/article/196/4/562/833570 [DOI] [PubMed] [Google Scholar]
- 7.Marinda E, Humphrey JH, Iliff PJ, Mutasa K, Nathoo KJ, Piwoz EG, et al. Child mortality according to maternal and infant HIV status in Zimbabwe. Pediatric Infectious Disease Journal. 2007. Jun;26(6):519–26. [DOI] [PubMed] [Google Scholar]
- 8.Dodd PJ, Prendergast AJ, Beecroft C, Kampmann B, Seddon JA. The impact of HIV and antiretroviral therapy on TB risk in children: A systematic review and meta-analysis. Thorax [Internet]. 2017. [cited 2020 Nov 25];72(6):559–75. Available from: http://www.r-project.org [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Meintjes G, Brust JCM, Nuttall J, Maartens G. Management of active tuberculosis in adults with HIV. Vol. 6, The Lancet HIV. Elsevier Ltd; 2019. p. e463–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.WHO. Global tuberculosis report [Internet]. 2019. [cited 2020 Feb 22]. Available from: http://apps.who.int/bookorders.
- 11.Dodd PJ, Yuen CM, Sismanidis C, Seddon JA, Jenkins HE. The global burden of tuberculosis mortality in children: a mathematical modelling study. Lancet Glob Health [Internet]. 2017. [cited 2020 Mar 2];5(9):e898–906. Available from: www.thelancet.com/lancetgh [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Zar HJ, Cotton MF, Strauss S, Karpakis J, Hussey G, Schaaf S, et al. Effect of isoniazid prophylaxis on mortality and incidence of tuberculosis in children with HIV: Randomised controlled trial. Br Med J. 2007;334(7585):136–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Ayieko J, Abuogi L, Simchowitz B, Bukusi EA, Smith AH, Reingold A. Efficacy of isoniazid prophylactic therapy in prevention of tuberculosis in children: A meta-analysis. BMC Infect Dis [Internet]. 2014. [cited 2020 Feb 23];14(1):1–10. Available from: http://www.biomedcentral.com/1471-2334/14/91 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Cohen K, Meintjes G. Management of individuals requiring antiretroviral therapy and TB treatment. Vol. 5, Current Opinion in HIV and AIDS. 2010. p. 61–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Svensson EM, Murray S, Karlsson MO, Dooley KE. Rifampicin and rifapentine significantly reduce concentrations of bedaquiline, a new anti-TB drug. Journal of Antimicrobial Chemotherapy [Internet]. 2014. [cited 2020 Feb 21];70(4):1106–14. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4356204/ [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Stolbach A, Paziana K, Heverling H, Pham P. A Review of the Toxicity of HIV Medications II: Interactions with Drugs and Complementary and Alternative Medicine Products. Journal of Medical Toxicology. 2015. Sep 27;11(3):326–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Chan GNY, Patel R, Cummins CL, Bendayan R. Induction of P-glycoprotein by antiretroviral drugs in human brain microvessel endothelial cells. Antimicrob Agents Chemother [Internet]. 2013. Sep [cited 2020 Feb 22];57(9):4481–8. Available from: http://www.ncbi.nlm.nih.gov/pubmed/23836171 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Mutwa PR, Boer KR, Rusine J, Muganga N, Tuyishimire D, Schuurman R, et al. Long-term effectiveness of combination antiretroviral therapy and prevalence of HIV drug resistance in HIV-1-infected children and adolescents in Rwanda. Pediatric Infectious Disease Journal [Internet]. 2014. Jan [cited 2020 Mar 25];33(1):63–9. Available from: http://www.ncbi.nlm.nih.gov/pubmed/24352189 [DOI] [PubMed] [Google Scholar]
- 19.WHO. HIV and adolescents: guidance for HIV testing and counselling and care for adolescents living with HIV Specific inclusion criteria for each review 1. Should HIV testing and counseling (HTC) be offered to adolescents? 2013. [Google Scholar]
- 20.Macallan DC, McNurlan MA, Kurpad AV, De Souza G, Shetty PS, Calder AG, et al. Whole body protein metabolism in human pulmonary tuberculosis and undernutrition: Evidence for anabolic block in tuberculosis. Clin Sci. 1998. Mar 1;94(3):321–31. [DOI] [PubMed] [Google Scholar]
- 21.Macallan DC. Malnutrition in Tuberculosis. Diagn Microbiol Infect Dis. 1999;34:153–7. [DOI] [PubMed] [Google Scholar]
- 22.Al Araf PM, Airlangga E. The body weights’ follow up before and after 6 months therapy of oral anti-tuberculosis therapy in children in Medan, Sumatra Utara. Open Access Maced J Med Sci. 2019;7(20):3469–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Golucci APBS Marson FAL, Valente MFF Branco MM, Prado CC, Nogueira RJN. Influence of AIDS antiretroviral therapy on the growth pattern. Vol. 95, Jornal de Pediatria. Elsevier Editora Ltda; 2019. p. 7–17. [DOI] [PubMed] [Google Scholar]
- 24.Macallan DC, Noble C, Baldwin C, Jebb SA, Prentice AM, Coward WA, et al. Energy Expenditure and Wasting in Human Immunodeficiency Virus Infection. New England Journal of Medicine. 1995;333(2):83–8. [DOI] [PubMed] [Google Scholar]
- 25.Williams PL, Jesson J. Growth and pubertal development in HIV-infected adolescents. Vol. 13, Current Opinion in HIV and AIDS. Lippincott Williams and Wilkins; 2018. p. 179–86. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Soeters HM, Sawry S, Moultrie H, Van Rie A. The effect of tuberculosis treatment on virologic and immunologic response to combination antiretroviral therapy among South African Children. J Acquir Immune Defic Syndr (1988). 2014;67(2):136–44. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Zanoni BC, Phungula T, Zanoni HM, France H, Feeney ME. Impact of tuberculosis cotreatment on viral suppression rates among HIV-positive children initiating HAART. AIDS. 2011. Jan 2;25(1):49–55. [DOI] [PubMed] [Google Scholar]
- 28.Njuguna IN, Cranmer LM, Otieno VO, Mugo C, Okinyi HM, Benki-Nugent S, et al. Urgent versus post-stabilisation antiretroviral treatment in hospitalised HIV-infected children in Kenya (PUSH): a randomised controlled trial. Lancet HIV [Internet]. 2018. [cited 2020 Feb 23];5(1):e12–22. Available from: http://depts.washington.edu/gwach/study-protocols-3/. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.World Health Organization [Internet]. [cited 2022 Feb 18]. World Health Organization, United Nations Children’s Fund (UNICEF) & World Bank. Levels and trends in child malnutrition: UNICEF / WHO / The World Bank Group joint child malnutrition estimates: key findings of the 2021 edition. Available from: https://apps.who.int/iris/handle/10665/341135
- 30.Graham SM, Cuevas LE, Jean-Philippe P, Browning R, Casenghi M, Detjen AK, et al. Clinical Case Definitions for Classification of Intrathoracic Tuberculosis in Children: An Update. Clinical Infectious Diseases. 2015;61:S179–87. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Ogutu O, al et. 4. Guidelines for Anti Retroviral therapy in Kenya Chapter on Gynaecological treatment. MOH. 2001; [Google Scholar]
- 32.Yuan YC. Multiple Imputation for Missing Data: Concepts and New Development [Internet]. [cited 2019 Jun 24]. Available from: https://stats.idre.ucla.edu/wp-content/uploads/2016/02/multipleimputation.pdf
- 33.Sterne JAC, White IR, Carlin JB, Spratt M, Royston P, Kenward MG, et al. Multiple imputation for missing data in epidemiological and clinical research: potential and pitfalls. BMJ [Internet]. 2009. Sep 1 [cited 2019 May 15];338(jun29 1):b2393–b2393. Available from: http://www.ncbi.nlm.nih.gov/pubmed/19564179 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Sharland M, Watkins AM, Dalgleish AG, Cammack N, Westby M. Immune reconstitution in HAART-treated children with AIDS. The Lancet [Internet]. 1998. Aug 15 [cited 2022 Apr 13];352(9127):577–8. Available from: http://www.thelancet.com/article/S0140673605792898/fulltext [DOI] [PubMed] [Google Scholar]
- 35.WHO. Tuberculosis: Key facts [Internet]. 2021. [cited 2022 Apr 4]. Available from: https://www.who.int/news-room/fact-sheets/detail/tuberculosis
- 36.Chiang SS, Park S, White EI, Friedman JF, Cruz AT, Del Castillo H, et al. Using Changes in Weight-for-Age z Score to Predict Effectiveness of Childhood Tuberculosis Therapy. J Pediatric Infect Dis Soc [Internet]. 2020. [cited 2022 Apr 4];9(2):150–8. Available from: https://pubmed.ncbi.nlm.nih.gov/30715497/ [DOI] [PubMed] [Google Scholar]
- 37.Muhammad HFL. Obesity as the Sequel of Childhood Stunting: Ghrelin and GHSR Gene Polymorphism Explained - PubMed. Acta Med Indonesiana [Internet]. 2018. Apr [cited 2022 Apr 13];50(2):159–64. Available from: https://pubmed.ncbi.nlm.nih.gov/29950536/ [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.