Abstract
Background
Data are limited on the resolution of symptoms and signs in children treated for pulmonary tuberculosis (PTB) and whether this resolution differs from children with other lower respiratory tract infections (LRTIs).
Methods
A prospective study of children ≤ 15 years presenting with features suggestive of PTB was performed. Clinical, microbiological, and radiological investigations were done at enrollment. Symptoms and clinical features were measured 1, 3, and 6 months after enrollment. Participants were categorized into 3 groups based on National Institutes of Health consensus definitions: confirmed PTB, unconfirmed PTB, and unlikely PTB (children with other LRTIs). Univariable and multivariable logistic regression modeling was used to investigate predictors of persistence of symptoms or signs.
Results
Among 2019 participants, there were 427 (21%) confirmed, 810 (40%) unconfirmed, and 782 (39%) with unlikely PTB. Of 1693/2008 (84%) with cough and 1157/1997 (58%) with loss of appetite at baseline, persistence at 3 months was reported in 24/1222 (2%) and 23/886 (3%), respectively. Of 934/1884 (50%) with tachypnoea and 947/1999 (47%) with abnormal auscultatory findings at baseline, persistence at 3 months occurred in 410/723 (57%) and 216/778 (28%), respectively. HIV infection and abnormal baseline chest radiography were associated with persistence of symptoms or signs at month 3 (adjusted odds ration [aOR] 1.6; 95% confidence interval [CI]: [1.1, 2.3] and aOR 2.3; 95% CI: [1.5, 3.3], respectively]. The resolution of symptoms and signs was similar across categories.
Conclusions
Symptoms resolved rapidly in most children with PTB, but signs resolved more slowly. The pattern and resolution of symptoms or signs did not distinguish children with PTB from those with other LRTIs.
Keywords: child, clinical resolution, pulmonary tuberculosis
In children treated for pulmonary tuberculosis, symptoms resolved rapidly in most, but signs resolved more slowly. The pattern and timing of resolution of symptoms and signs cannot distinguish children with pulmonary tuberculosis from those with other lower respiratory tract infections.
Diagnosis of pulmonary tuberculosis (PTB) in children is usually made clinically, based on history of exposure, symptoms, signs, tuberculin skin test (TST) results, and chest radiography (CXR) if available [1]. Challenges in obtaining samples for microbiological confirmation limit the usefulness of microbiology in diagnosis, monitoring disease and classifying treatment outcomes. Clinical response to treatment is used as a classification feature to assist in differentiating children with PTB from those with other lower respiratory tract infections [2] (LRTIs). Clinicians also rely on the resolution of symptoms and weight gain for the assessment of treatment effectiveness [1, 3].
Despite the widespread use of measures of clinical response to TB treatment, data are lacking on the pattern and timing of resolution of symptoms and signs, and the impact of response modulators such as human immunodeficiency virus (HIV) co-infection or malnutrition. Studies of adults with PTB suggest that clinical symptoms improve within 4-8 weeks of treatment [4–6] and can be used in diagnostic algorithms to distinguish people with PTB from those without [4]. Very limited pediatric data, however, suggest symptom resolution in young children takes longer, and that the timing of symptom resolution cannot differentiate these groups [7].
Clinical case definitions of intrathoracic TB in children, developed by a National Institutes of Health (NIH) expert consensus panel, recommend assessment of response 2 months after treatment initiation [2]. However, one study reported that a substantial proportion of children treated for PTB only had symptom resolution beyond 2 months, suggesting that this may be too early to assess response [7]. Improved understanding of the pattern and timing of treatment response in children initiated on PTB treatment is thus needed.
The aim of this study was to longitudinally investigate treatment response in children treated for PTB, to compare this to children with other LRTIs, and to identify factors associated with the persistence of symptoms or signs.
METHODS
Study Design and Participants
This study included children prospectively enrolled in TB diagnostic studies from 1 February, 2009, to 31 December, 2018 [8–11], in Cape Town, South Africa. Study participants were enrolled at Red Cross War Memorial Children’s Hospital, a tertiary referral center, New Somerset Hospital, a secondary level hospital, and Nolungile primary healthcare clinic. Children were eligible for enrollment if they were 15 years of age or younger and had features suggestive of PTB, defined as a cough of any duration, and one of the following: a household contact with TB in the preceding 3 months, weight loss or failure to thrive in preceding 3 months, a positive TST, or a CXR suggestive of PTB. Children were excluded if they had received TB therapy or prophylaxis for 72 hours or more prior to enrollment, were not able to attend follow-up, informed consent was not available, or if a pulmonary specimen could not be obtained. For this study, children without at least one follow-up visit were also excluded.
At enrollment, all participants were evaluated clinically. Respiratory rate was counted over 1 minute, with the child at rest, lying or sitting. Tachypnoea was classified according to World Health Organization (WHO) age-specific criteria (<2 months, respiratory rate [RR] ≥ 60 breaths per minute [bpm]; 2-12 months, RR ≥ 50 bpm; 12-59 months, RR ≥ 40 bpm; and >60 months, RR > 20) [12]. WHO 2007 growth standards were used to calculate weight-for-age z scores (WAZ) [13, 14]. Investigations for microbiological confirmation included 2 induced sputum specimens for Xpert MTB/RIF (Cepheid, USA), or from February 2018 onwards Xpert MTB/RIF Ultra, and liquid culture (Mycobacterium Growth Indicator Tube, BACTEC MGIT. Becton Dickinson, MD, USA). All children had a TST, with positivity defined as transverse induration of >5 mm in HIV-positive and >10 mm in HIV-negative participants [3]. Children living with HIV were classified as having advanced or severe immunodeficiency based on the WHO immunological classification for established HIV infection (<11 months, CD4% < 30%; 12-35 months, CD4% < 25%; 36-59 months, CD4% < 20%; >5 years, CD4 < 350cells/μL or CD4% < 15%) [15].
Baseline CXRs were reported by a radiologist experienced in pediatric radiographs, using a standardized reporting tool and categorized as normal or abnormal. Decisions regarding the initiation of TB therapy were made by the treating clinician, who prescribed TB drug regimens using national TB protocols and based on drug-resistant patterns [1].
Study participants were categorized into 3 groups according to NIH consensus definitions [2]; confirmed PTB (Mycobacterium tuberculosis [Mtb] detected on a pulmonary specimen by Xpert MTB/RIF or Xpert MTB/RIF Ultra or culture), unconfirmed PTB (no bacteriological confirmation with at least 2 of the following: signs and symptoms consistent with PTB, compatible CXR, close TB exposure or immunological evidence of exposure, or a positive response to treatment), and unlikely PTB (no bacteriological confirmation and criteria not met for unconfirmed PTB). The unlikely PTB group was assumed to have LRTIs caused by pathogens other than Mtb; none were treated for PTB.
All children were followed at 1 and 3 months after enrollment; confirmed or unconfirmed PTB cases were also followed at 6 months. At follow-up visits, symptoms and signs of PTB were recorded using a standardized questionnaire, and a clinical examination including anthropometry was done.
This study was approved by the Human Research Ethics Committee, Faculty of Health Sciences, University of Cape Town (HREC 827/2019). A parent provided written informed consent for the participation of their child, and assent was obtained in children older than 7 years of age.
Statistical Analysis
Data were analyzed using STATA version 16.1 (Stata Corp, College Station, TX, USA). Baseline characteristics were summarized as medians (interquartile range [IQR]) or means (standard deviation) for continuous data, and as number (percentages) for categorical data. Differences between groups were explored with chi-squared test for categorical data, and one-way analysis of variance (ANOVA) or Kruskal-Wallis tests for normally or non-normally distributed continuous data, respectively.
Data for follow-up visits were regarded as missing if the timing was not within the specified range of 15-59 days for month 1, 60-149 days for month 3, or 150-240 days for month 6. Persistence of symptoms or signs was calculated for those participants with the specific symptom or sign at baseline and who had been assessed for persistence or resolution at all preceding time points. Symptoms or signs that developed during the follow-up period or that recurred after previous resolution were excluded from the analysis.
Potential predictors of persistence of symptoms or signs were identified with univariable and multivariable logistic regression modeling. Multivariable models were run for individual symptoms (cough and loss of appetite), signs (tachypnoea and abnormal auscultatory findings), as well as for a combined variable including any of the following: persistent cough, loss of appetite, abnormal auscultatory findings, or no weight improvement if underweight (WAZ< −2) at baseline. Variables in all models included sex, HIV status, age category (<1year, 1-5 years, >5 years), PTB category (confirmed, unconfirmed, unlikely), and baseline CXR findings. Variables with a P-value of <.15 in univariable analysis were also included in the multivariable models.
Subset analyses investigated the persistence of symptoms and signs in the unlikely PTB group, in those living with HIV infection, as well as in the whole cohort excluding those participants with extrapulmonary PTB (EPTB). Because our measured outcomes were clinical response to treatment and a positive clinical response to treatment is one of the inclusion criteria for classification in the unconfirmed PTB group, we conducted a sensitivity analysis excluding those 117 participants whose inclusion in the unconfirmed PTB group was contingent on a positive clinical response.
RESULTS
Cohort Description
Of 2389 participants enrolled, 370 (15%) were excluded (Figure 1), predominantly for loss to follow-up (n = 226) or only EPTB (103). Overall, 427 (21%) participants with confirmed PTB, 810 (40%) with unconfirmed PTB, and 782 (39%) with unlikely PTB were included.
Figure 1.
Study profile. aTwo participants included in the confirmed PTB category had Mycobacterium tuberculosis identified only from extrapulmonary samples. PTB, pulmonary tuberculosis.
Among all participants (median age 31.8 months [IQR: 15.2, 63.0]), 16% (325/2015) were living with HIV, with no differences between the groups, Table 1. Most participants with HIV (88% [269/304]) were receiving antiretroviral therapy. Confirmed PTB cases were older than unconfirmed or unlikely PTB cases and had lower WAZ scores, Table 1. A greater proportion of confirmed than unconfirmed PTB cases had extrapulmonary disease together with PTB (40% vs 16%; P < .001). Features of EPTB were pleural effusion (36% [108/301]), abdominal TB (23%, [70/301]), TB lymphadenitis (23%, [69/301]), or TB meningitis (12%, [37/301]).
Table 1.
Baseline Characteristics of Cohort by TB Categorizationa
| Total (n = 2019) |
Confirmed PTB (n = 427) |
Unconfirmed PTB (n = 810) |
Unlikely PTB (n = 782) |
P-valueb | |
|---|---|---|---|---|---|
| n (%) or median [IQR] | |||||
| Sociodemographics | |||||
| Male sex | 1026 (51) | 229 (54) | 418 (52) | 379 (49) | .194 |
| Age category | |||||
| <1 years | 368 (18) | 64 (15) | 157 (19) | 147 (19) | .000 |
| 1-5 years | 1099 (54) | 209 (49) | 457 (56) | 433 (55) | |
| >5 years | 552 (27) | 154 (36) | 196 (24) | 202 (25) | |
| Age in months, median | 31.8 [15.2, 63.0] | 40.0 [17.5, 85.8] | 30.9 [14.6, 58.9] | 30.4 [15.2, 61.5] | .000 |
| HIV infected | 325 (16) | 73 (17) | 143 (18) | 109 (14) | .110 |
| CD4 count (cells/μL) | 541 [245,1073] | 466 [235, 878] | 543 [253, 1096] | 552 [245, 1005] | .831 |
| Advanced or severe immunodeficiencyc | 163 (63) | 37 (58) | 74 (66) | 52 (61) | .529 |
| Reported symptoms | |||||
| Cough | 1693 (84) | 341 (80) | 669 (83) | 683 (88) | .001 |
| Fever | 1217 (61) | 258 (61) | 495 (62) | 464 (60) | .837 |
| Weight loss | 1371 (70) | 326 (78) | 553 (71) | 492 (65) | .000 |
| Malaise | 817 (42) | 234 (56) | 331 (42) | 252 (33) | .000 |
| Night sweats | 762/1472 (52) | 216/388 (56) | 322/586 (55) | 224/498 (45) | .001 |
| Poor appetite | 1157 (58) | 253 (60) | 472 (59) | 432 (56) | .403 |
| Clinical signs | |||||
| Abnormal respiratory exam | 1013 (50) | 260 (61) | 400 (50) | 353 (45) | .000 |
| Subcostal recession | 300 (15) | 81 (19) | 108 (14) | 111 (15) | .028 |
| Abnormal breath sounds | 357 (19) | 105 (25) | 132 (17) | 120 (17) | .000 |
| Crackles | 578 (29) | 130 (31) | 222 (28) | 226 (29) | .492 |
| Wheeze | 271 (14) | 54 (13) | 105 (13) | 112 (15) | .628 |
| Tachypnoea | 934 (50) | 233 (58) | 366 (48) | 335 (47) | .001 |
| Anthropometry | |||||
| WAZ | −0.9 [−1.8, 0.1] | −1.2 [−2.1, −0.3] | −0.9 [−1.8, 0.0] | −0.6 [−1.6, 0.2] | .000 |
| WAZ < −2 | 428/2003 (21) | 117/425 (28) | 177/808 (22) | 134/770 (17) | .000 |
| Investigations | |||||
| Mantoux-positive | 1022 (55) | 286/372 (77) | 524 (68) | 212 (29) | .000 |
| CXR—abnormal | 1089/1582 (69) | 283/323 (88) | 422/595 (71) | 384/664 (58) | .000 |
| Extrapulmonary disease | |||||
| PTB and EPTB | 301 (15) | 170 (40) | 131 (16) | 0 (0) | .000 |
HIV, human immunodeficiency virus; WAZ, weight-for-age z score; CXR, chest radiography; PTB, pulmonary tuberculosis; EPTB, extrapulmonary tuberculosis.
Denominator shown if >10% of cohort missing data.
P-value relates to differences across the 3 groups.
Defined according to WHO immunological classification for established HIV infection (<11 months, CD4% < 30%; 12-35 months, CD4% < 25%; 36-59 months, CD4% < 20%; >5 years, CD4 < 350cells/μL or CD4% < 15%).
Baseline symptoms differed between confirmed, unconfirmed, and unlikely PTB groups for weight loss (78%, 71%, 65%; P < .001), malaise (56%, 42%, 33%; P < .001), and night sweats (56%, 55%, 45%; P = .001). Similarly, duration of weight loss was longer in confirmed or unconfirmed compared with the unlikely TB group [median (IQR) days 21 (7, 31); 21(10, 31) vs 14 (7, 31); P = .005]. The duration of other baseline symptoms were similar across the groups, with the exception of poor appetite which was longer in the confirmed and unconfirmed groups compared with the unlikely group [median (IQR) days 14 (6, 21); 14 (6, 21) vs 8.5 (5,14); P = .004].
Clinical signs also differed between groups, with abnormal respiratory examination found in 61% (260/426), 50% (400/807), and 45% (353/776) (P < .001), and tachypnoea in 58% (233/402), 48% (366/765), and 47% (335/717), respectively (P = .001). Baseline tachypnoea and abnormal auscultatory findings were present in 66% (199/301) and 64% (207/324) of children living with HIV vs 46% (732/1579) and 44% (739/1671) of HIV-negative participants (P < .001).
A baseline CXR, available in 78% (1582/2019) of participants, was abnormal in 88% (283/323), 71% (422/595), and 58% (384/664) of the confirmed, unconfirmed, and unlikely PTB groups, respectively (P < .001). Abnormal CXR findings included lymphadenopathy in 23% (367/1564), consolidation in 51% (792/1564), cavitation in 5% (75/1564), and pleural effusion in 12% (181/1564). Specific CXR findings that were more common in confirmed compared with unconfirmed or unlikely TB groups included lymphadenopathy (38%, 27%, 14%, respectively; P < .001), consolidation (69%, 50%, 43%; P < .001), cavitation (13%, 2%, 3%; P < .001), or pleural effusion (22%, 13%, 5%; P < .001). An abnormal CXR was more common in children living with HIV than in the HIV-negative children (82% [212/257] vs. 66% [875/1323]; P < .001).
Time to follow-up was similar between the groups; the median time to month 1 follow-up was 30 (IQR 28, 35), 31 (IQR 28, 35), and 32 days (IQR 29, 36) in the confirmed, unconfirmed, and unlikely groups, respectively. The median time to month 3 follow-up visit was 91 days (IQR 87, 97), and in those treated for PTB time to month 6 visit was 181 days (IQR 174, 186).
Resolution of Symptoms and Signs
Most of the participants had rapid resolution of cough or loss of appetite, Table 2. At 1 month, only 9% (129/1402) of participants who had a cough at baseline reported no improvement. This was lower in the confirmed or unconfirmed groups than in the unlikely group: 9% (24/278) and 7% (39/564) vs 12% (66/560), respectively (P = .017). Children with persistent cough declined to 2% (24/1222) at 3 months and 1% (4/697) at 6 months, with no significant difference between groups. Similarly, 11% (111/999) reported persistence of loss of appetite at 1 month, 3% (23/886) at 3 months, and 1% (3/471) at 6 months, with no differences between groups.
Table 2.
Persistence of Symptoms or Signs at 1, 3, and 6 months of Follow-up by TB Categorization
| Total (n = 2019) |
Confirmed PTB (n = 427) |
Unconfirmed PTB (n = 810) |
Unlikely PTB (782) |
P-value | ||
|---|---|---|---|---|---|---|
| n (%) or median (IQR) | ||||||
| Symptoms | ||||||
| Cough | Baseline | 1693/2008 (84) | 341/425 (80) | 669/806 (83) | 683/777 (88) | .001 |
| Cough non-improvement | M1 | 129/1402 (9) | 24/278 (9) | 39/564 (7) | 66/560 (12) | .017 |
| M3 | 24/1222 (2) | 4/240 (2) | 5/495 (1) | 15/487 (3) | .061 | |
| M6 | 4/697 (1) | 2/219 (1) | 2/478 (0) | .422 | ||
| Loss of appetite | Baseline | 1157/1997 (58) | 253/424 (60) | 472/803 (59) | 432/770 (56) | .403 |
| M1 | 111/999 (11) | 21/220 (10) | 50/412 (12) | 40/367 (11) | .606 | |
| M3 | 23/886 (3) | 4/195 (2) | 12/371 (3) | 7/320 (2) | .595 | |
| M6 | 3/471 (1) | 2/164 (1) | 1/307 (0) | .245 | ||
| Signs | ||||||
| Tachypnoea | Baseline | 934/1884 (50) | 233/402 (58) | 366/765 (48) | 335/717 (47) | .001 |
| M3 | 410/723 (57) | 116/182 (64) | 163/298 (55) | 131/243 (54) | .085 | |
| M6 | 171/380 (45) | 72/140 (51) | 99/240 (41) | .054 | ||
| Subcostal Recession | Baseline | 300/1973 (15) | 81/420 (19) | 108/792 (14) | 111/761 (15) | .028 |
| M3 | 0/229 (0) | 0/61 (0) | 0/88 (0) | 0/80 (0) | ||
| Abnormal Auscultation | Baseline | 947/1999 (47) | 246/424 (58) | 366/803 (46) | 335/772 (43) | .000 |
| M3 | 216/778 (28) | 75/202 (37) | 71/309 (23) | 70/267 (26) | .002 | |
| M6 | 44/424 (10) | 25/161 (16) | 19/263 (7) | .007 | ||
| Combination scorea | ||||||
| Baseline | 1944/2012 (97) | 407/426 (96) | 780/807 (97) | 7575/779 (97) | .323 | |
| M3 | 295/1750 (17) | 90/364 (25) | 100/708 (14) | 105/678 (16) | .000 | |
| Anthropometry | ||||||
| WAZ | Baseline | −0.9 [−1.8, 0.1] | −1.2 [−2.1, −0.3] | −0.9 [−1.8, 0.0] | −0.6 [−1.6, 0.2] | .000 |
| M1 | −0.5 [−1.4, 0.3] | −0.7 [−1.7, −0.0] | −0.5 [−1.4, 0.3] | −0.4 [−1.2, 0.5] | .000 | |
| M3 | −0.4 [−1.3, 0.4] | −0.6 [−1.5, 0.2] | −0.4 [−1.3, 0.4] | −0.3 [−1.2, 0.5] | .001 | |
| M6 | −0.3 [−1.1, 0.5] | −0.4 [−1.2, 0.2] | −03 [−1.0, 0.6] | .030 | ||
WAZ, weight-for-age z score; PTB, pulmonary tuberculosis; M1, month 1 follow-up visit; M3, month 3 follow-up visit; M6, month 6 follow-up visit.
Combination score includes any of the following: cough, loss of appetite, abnormal auscultation, or no weight gain if underweight at baseline.
In contrast, abnormal signs persisted in a greater proportion. While subcostal recession resolved in all participants by 3 months, tachypnoea or abnormal auscultatory findings (wheeze, crackles, reduced breath sounds, or abnormal breath sounds) persisted in a high proportion. Among 723/934 (77%) with baseline tachypnoea, 57% remained tachypnoeic at 3 months and 45% (171/380) at 6 months, similar between the groups. Similarly, among 778/947 (82%) with baseline abnormal auscultatory findings, 28% had persistence at month 3, with a greater proportion in the confirmed (37%, 75/202) than unconfirmed (23%, 71/309) or unlikely (26%, 70/267) groups (P = .002). Persistent tachypnoea and abnormal auscultatory findings were present in 70% (104/148) and 36% (56/154) of children living with HIV vs 53% (305/574) and 20% (119/610) of HIV-negative participants (P < .001).
The median WAZ score was the lowest in the confirmed and the highest in the unlikely group at all time points. In those who were underweight at baseline, the median increase in WAZ at 3 months was 0.9 (IQR 0.5, 1.8), 0.9 (IQR 0.5, 1.5), and 0.7 (IQR 0.4, 1.2) in the confirmed, unconfirmed, and unlikely groups, respectively (P = .037).
Factors Associated With Persistence of Symptoms or Signs
Table 3 presents factors associated with the persistence of any symptom or signs at 3 months in univariable and multivariable logistic regression models. After adjustment for other variables, only HIV infection (adjusted odds ration [aOR] 1.6, 95% confidence interval [CI]: 1.1, 2.3; P = .011) and abnormal baseline CXR (aOR 2.3, 95% CI: 1.5, 3.3; P < .001) were significantly associated with persistence of any symptoms or signs. This association was unchanged in the sensitivity analysis (Supplementary Data).
Table 3.
Prevalence and Predictors of Symptom or Sign Persistence at 3 Monthsa
| Prevalence | Univariable Predictors | Multivariable Predictors | ||||
|---|---|---|---|---|---|---|
| % (n/total) | OR (95%CI) | P-value | aOR (95%CI) | P-value | ||
| Sociodemographics | ||||||
| HIV status | Infected | 25 (72/286) | 1.9 (1.4, 2.5) | .000 | 1.6 (1.1, 2.3) | .011 |
| Gender | Female | 16 (137/852) | 0.9 (0.7, 1.2) | .398 | 0.9 (0.7, 1.2) | .392 |
| Age category | <1 year | 19 (62/324) | 1.0 | 1.0 | ||
| 1-5 | 15 (139/947) | 0.7 (0.5, 1.0) | .058 | 0.7 (0.5, 1.0) | .061 | |
| >5 years | 20 (94/479) | 1.0 (0.7, 1.5) | .864 | 0.9 (0.6, 1.4) | .798 | |
| PTB disease classification | ||||||
| Unlikely | 16 (105/678) | 1.0 | 1.0 | |||
| Confirmed | 25 (90/364) | 1.8 (1.3, 2.5) | .000 | 1.3 (0.9, 1.9) | .156 | |
| Unconfirmed | 14 (100/708) | 0.9 (0.7, 1.2) | .475 | 0.7 (0.5, 1.0) | .086 | |
| Chest radiography findings | ||||||
| CXR abnormal | Yes | 20 (189/962) | 2.5 (1.7, 3.6) | .000 | 2.3 (1.5, 3.3) | .000 |
OR, odds ration; aOR, adjusted odds ration; HIV, human immunodeficiency virus; PTB, pulmonary tuberculosis; CXR, chest radiography.
Symptom and sign persistence includes any of the following: cough, loss of appetite, abnormal auscultatory findings, or no weight improvement if underweight at baseline.
In models of each individual symptom or sign, HIV infection was associated with persistence at 3 months of tachypnoea (aOR 2.2, 95% CI: 1.2, 4.0; P = .009) or abnormal auscultatory findings (aOR 1.6, 95% CI: 1.1, 2.5; P = .025). The association between HIV and persistence of symptoms or signs was evident in all subset analyses with the exception of those with unlikely PTB, where HIV infection was not associated with persistence at 3 months (aOR 1.3, 95% CI: .7, 2.4; P = .461).
Abnormal baseline CXR was the strongest predictor of persistent symptoms or signs. Persistent symptoms (compared with those whose symptoms resolved) were associated with radiological consolidation (74% vs 47;, P < .001), cavitation (14% vs 3%; P < .001), or pleural effusion (22% vs 10%; P < .001). Abnormal CXR also increased the adjusted odds for persistent abnormal auscultatory findings (aOR 2.7, 95% CI: 1.3, 5.4; P = .006). There was no association between baseline CXR abnormality and persistence of either cough or loss of appetite at month 3.
Persistence of tachypnoea at 3 months was strongly associated with age greater than 5 years (aOR 13.6, 95% CI: 6.9, 27.0; P < .001). This association was evident in all subset analyses. There was no association between baseline CXR abnormality or specific CXR findings and persistence of tachypnoea (aOR 1.2, 95% CI: 0.6, 2.1; P = .623).
DISCUSSION
In this large prospective study of over 2000 children with symptoms of PTB, including 21% with microbiologically confirmed disease, symptoms resolved rapidly in most, but clinical signs persisted in more than half at 3 months. The timing of resolution of symptoms and signs was similar for those with confirmed PTB, unconfirmed PTB, or other LRTIs, but a greater proportion of confirmed cases had persistence of abnormal auscultatory findings and lower weight for age at 3 months. Living with HIV infection and abnormal baseline CXR were the strongest predictors of persistent symptoms or signs at 3 months.
Rapid symptom resolution is consistent with data from adult studies [4, 5] but not with the limited available pediatric data. In a study of 191 young children treated for PTB, the median time to resolution of cough was 63 days, with only 42% (10/24) of children with “definite” or “probable” TB [16] having a resolution of baseline cough 60 days after treatment initiation [7]. However, the study population was younger than our cohort (median age 12 months), excluded children with HIV, included only 9 confirmed PTB cases, and was a much smaller sample size. Differences in timing of symptom resolution may reflect our larger sample size, shorter median time to first follow-up visit (31 vs 61 days), and reporting of improvement of symptoms, not only complete resolution. These differences likely provide greater sensitivity to detect earlier symptom improvement.
While symptoms resolved rapidly for most of the children, a very small percentage (2%) had a persistent cough at 3 months, reducing to 1% at 6 months. In a clinical trial of almost 400 adults with confirmed PTB, a much higher proportion, 13% reported a persistent cough at 6 months, after completion of therapy [6]. While the exact mechanism of cough in PTB has not been definitively elucidated, it may be due to the presence of mycobacteria, substances released by the organism, or host immune responses [17]. Mtb bacillary load declines rapidly following initiation of effective anti-tuberculosis therapy [18–20], and thus more rapid improvement may occur in children who usually have a lower mycobacterial load. Persistent cough may be due to slow resolution of disease, drug resistance, or complications. In this study, only 7 cases of drug-resistant PTB occurred; 1 had a persistent cough at 6 months.
Clinical signs may reflect the results of the complex host immune response to the presence of Mtb and the resultant tissue injury and lung impairment [21]. This tissue injury may at least partially explain the slower resolution of abnormal auscultatory findings. The higher proportion of persistent abnormal auscultatory findings in the confirmed PTB group, as well as the strong association between abnormal baseline CXR and persistence of abnormal auscultatory findings, indicates that extent of pulmonary involvement influences clinical resolution. Tachypnoea was also found to persist in an unexpectedly high proportion of participants, especially in children older than 5 years. This, however, may be due to the use of WHO-recommended age-specific cutoff values which were used to define tachypnoea, with a single recommended value for those over 5 years, which may lack sensitivity to distinguish tachypnoea in children across a spectrum of ages older than 5 years.
Participants living with HIV infection were more likely to have abnormal baseline CXR findings and persistent symptoms or signs at 3 months, despite the majority receiving antiretroviral therapy. Our findings are consistent with HIV frequently associated with chronic lung disease and the association of HIV co-infection with poorer outcomes of TB treatment or slower resolution, which have been previously reported in children [22, 23].
Limitations of this study include analysis of selected symptoms and signs (cough, poor appetite, abnormal auscultatory findings, and weight gain); however, these are key factors used in the diagnosis and assessing treatment response. Secondly, signs and symptoms were assessed at specific time points; thus, it is not possible to differentiate persistence from recurrence. However, follow-up was within a month and 3 months of diagnosis. Although the NIH definitions suggest treatment response be assessed at 2 months, our follow-up schedule provided more time points, optimized cohort retention, and enabled earlier assessment of rapid resolution of symptoms. It is also possible that despite using current consensus guidelines, some participants in the unconfirmed or unlikely TB groups may have been incorrectly classified, thus limiting the ability to differentiate between these 2 categories. Nevertheless, this study, comprising 427 children with confirmed TB, comprises one of the largest cohort studies of children with microbiologically confirmed TB. A further limitation is the absence of microbiological testing for other pathogens, so assigning etiology in the unlikely group was not possible. However, exclusion of PTB in this group was done based on microbiologic testing for PTB and careful follow-up in the absence of TB treatment. Strengths are the large study population, careful follow-up, high cohort retention, standardized assessment, and comprehensive microbiological investigations for TB as well as the inclusion of children across ages and those living with HIV, enabling assessment of the impact of these on resolution.
CONCLUSION
Clinically differentiating children with PTB from those with other LRTIs remains a challenge, but the resolution of symptoms or signs does not reliably distinguish these groups. Most symptoms resolved rapidly, while clinical signs persisted in a greater proportion of participants. Abnormal CXR and HIV were the strongest predictors of the persistence of symptoms or signs. Clinical response to treatment may not be a reliable indicator of PTB.
Supplementary Data
Supplementary materials are available at the Journal of The Pediatric Infectious Diseases Society online (http://jpids.oxfordjournals.org).
Acknowledgments
We would like to thank the children and their caregivers for participating in this research as well as all the clinical and study staff involved in patient care and data collection.
Notes
Financial support. This work was supported by the Regional Prospective Observational Research in Tuberculosis (RePORT TB) Consortium, which is co-funded by the Medical Research Council of South Africa (SAMRC) and the US Office of AIDS Research of the National Institutes of Health of the United States. Additional funding was received from the SAMRC from the National Institutes of Health, USA (grant number RO1HD058971), and the European and Developing Countries Clinical Trials Partnership (EDCTP) (TB-NEAT; IP.2009.32040.009). K. D. and the work presented here were supported by the SAMRC (grant number RFA-EMU-02-2017) and the EDCTP (grant numbers TMA-2015SF-1043, TMA-1051-TESAII). H. J. Z. is supported by the SAMRC. M. P. N. is supported by the Australian National Health and Medical Research Council (grant number APP1174455).
Potential conflicts of interest . All authors declare that they have no potential conflicts of interest.
All authors have submitted the ICMJE Form for Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.
Contributor Information
Julie Copelyn, Paediatric Infectious Diseases Unit, Red Cross War Memorial Children’s Hospital, Cape Town, South Africa; Department of Paediatrics and Child Health, Faculty of Health Sciences, University of Cape Town, Cape Town, South Africa.
Brian Eley, Paediatric Infectious Diseases Unit, Red Cross War Memorial Children’s Hospital, Cape Town, South Africa; Department of Paediatrics and Child Health, Faculty of Health Sciences, University of Cape Town, Cape Town, South Africa.
Helen Cox, Division of Medical Microbiology, Wellcome Centre for Infectious Disease Research and Institute of Infectious Disease and Molecular Medicine, University of Cape Town, Cape Town, South Africa.
Lesley Workman, Department of Paediatrics and Child Health, Faculty of Health Sciences, University of Cape Town, Cape Town, South Africa; South African Medical Research Council Unit on Child & Adolescent Health, University of Cape Town, Cape Town, South Africa.
Keertan Dheda, Centre for Lung Infection and Immunity, Division of Pulmonology, Department of Medicine and UCT Lung Institute & South African MRC/UCT Centre for the Study of Antimicrobial Resistance, University of Cape Town, Cape Town, South Africa; Department of Immunology and Infection, Faculty of Infectious and Tropical Diseases, London School of Hygiene and Tropical Medicine, London, UK.
Mark P Nicol, Division of Medical Microbiology, University of Cape Town, Cape Town, South Africa; Department of Biomedical Sciences, Division of Infection and Immunity, University of Western Australia, Perth, Australia.
Heather J Zar, Department of Paediatrics and Child Health, Faculty of Health Sciences, University of Cape Town, Cape Town, South Africa; South African Medical Research Council Unit on Child & Adolescent Health, University of Cape Town, Cape Town, South Africa.
References
- 1. South African National Department of Health. Guidelines for the Management of Tuberculosis in Children. 2013. http://www.kznhealth.gov.za/family/National-Childhood-TB-Guidelines-2013-ZA.pdf. Accessed 5 July 2021.
- 2. Graham SM, Cuevas LE, Jean-Philippe P, et al. Clinical case definitions for classification of intrathoracic tuberculosis in children: an update. Clin Infect Dis 2015; 61:S179–87. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. World Health Organization (WHO). Guidance for National Tuberculosis Programmes on the Management of Tuberculosis in Children. 2nd ed. World Health Organization; 2014. [Google Scholar]
- 4. Wilson D, Nachega J, Morroni C, Chaisson R, Maartens G.. Diagnosing smear-negative tuberculosis using case definitions and treatment response in HIV-infected adults. Int J Tuberc Lung Dis 2006; 10:31–8. [PubMed] [Google Scholar]
- 5. Feleke BE, Alene GD, Feleke TE, Motebaynore Y, Biadglegne F.. Clinical response of tuberculosis patients, a prospective cohort study. PLoS One 2018; 13:e01902071–11. https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0190207. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Bark CM, Dietze R, Okwera A, Quelapio MI, Thiel BA, Johnson JL.. Clinical symptoms and microbiological outcomes in tuberculosis treatment trials. Tuberculosis. 2011; 91:601–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Mpofu N, Moyo S, Mulenga H, et al. Time to symptom resolution in young children treated for pulmonary tuberculosis. Pediatr Infect Dis J 2014; 33:1226–30. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Nicol MP, Workman L, Isaacs W, et al. Accuracy of the Xpert MTB/RIF test for the diagnosis of pulmonary tuberculosis in children admitted to hospital in Cape Town, South Africa: a descriptive study. Lancet Infect Dis 2011; 11:819–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Zar HJ, Workman LJ, Prins M, et al. Tuberculosis diagnosis in children using Xpert Ultra on different respiratory specimens. Am J Respir Crit Care Med 2019; 200:1531–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Zar HJ, Workman L, Isaacs W, Dheda K, Zemanay W, Nicol MP.. Rapid diagnosis of pulmonary tuberculosis in African children in a primary care setting by use of Xpert MTB/RIF on respiratory specimens: a prospective study. Lancet Glob Heal 2013; 1:e97–e104. [DOI] [PubMed] [Google Scholar]
- 11. Nicol MP, Workman L, Prins M, et al. Accuracy of Xpert MTB/RIF ultra for the diagnosis of pulmonary tuberculosis in children. Pediatr Infect Dis J 2018; 37(10):e261–e263. [DOI] [PubMed] [Google Scholar]
- 12. World Health Organization (WHO). Pocket Book of Hospital Care for Children: Guidelines for the Management of Common Childhood Illnesses. 2nd ed. Geneva: World Health Organisation; 2013. [PubMed] [Google Scholar]
- 13. World Health Organization (WHO). WHO Growth Reference Data for 5-19 years. Geneva: World Health Organisation. Accessed November 11, 2020. https://www.who.int/growthref/en/ [Google Scholar]
- 14. World Health Organization (WHO). WHO Child Growth Standards. Geneva: World Health Organisation. [Google Scholar]
- 15. World Health Organization. WHO case definitions of HIV for surveillance and revised clinical staging and immun ological classification of HIV-related disease in adults and children. World Health Organization. 2007; 52. https://apps.who.int/iris/handle/10665/43699. Accessed 3 July 2021. [Google Scholar]
- 16. Moyo S, Verver S, Hawkridge A, et al. Tuberculosis case finding for vaccine trials in young children in high-incidence settings: a randomised trial. Int J Tuberc Lung Dis 2012; 16:185–91. [DOI] [PubMed] [Google Scholar]
- 17. Turner RD, Bothamley GH.. Cough and the transmission of tuberculosis. J Infect Dis 2015; 211:1367–72. [DOI] [PubMed] [Google Scholar]
- 18. Heslop R, Bojang AL, Jarju S, et al. Changes in host cytokine patterns of TB patients with different bacterial loads detected using16 S rRNA analysis. PLoS One 2016; 11:e01682721–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Hai HT, Vinh DN, Thu DDA, et al. Comparison of the Mycobacterium tuberculosis molecular bacterial load assay, microscopy and GeneXpert versus liquid culture for viable bacterial load quantification before and after starting pulmonary tuberculosis treatment. Tuberculosis. 2019; 119: 101864. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Sabiiti W, Azam K, Farmer ECW, et al. Tuberculosis bacillary load, an early marker of disease severity: the utility of tuberculosis Molecular Bacterial Load Assay. Thorax 2020; 60:6–608. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Ravimohan S, Kornfeld H, Weissman D, Bisson GP.. Tuberculosis and lung damage: from epidemiology to pathophysiology. Eur Respir Rev 2018; 27:170077. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Madhi SA, Huebner RE, Doedens L, Aduc T, Wesley D, Cooper PA.. HIV-1 co-infection in children hospitalised with tuberculosis in South Africa. Int J Tuberc Lung Dis 2000; 4:448–54. [PubMed] [Google Scholar]
- 23. Belay GM, Wubneh CA.. Childhood tuberculosis treatment outcome and its association with HIV co-infection in Ethiopia: a systematic review and meta-analysis. Trop Med Health 2020; 48:1–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.