Table 5.
On-treatment outcomes (treatment success, case fatality and default on-treatment) among smear, Xpert and/or culture positive TB patients reported in n=7 observational studies and n=1 CRT, and, all-cause mortality reported in n=2 CRT
| Observational studies | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| First author, country and population, screening tool | Group | Treatment success |
PR | Case fatality |
PR | LTFU on treatment |
Pre-treatment LTFU* |
Comments | ||||
| n/N | % (95%CI) | n/N | % (95%CI) | n/N | % (95%CI) | n/Nʃ | (%) | |||||
| General population | ||||||||||||
| den Boon 2008 South Africa smear & culture |
Screen | 16/20 | 80% (56-94%) | 1.00 | 2/27 | 7% (1-24%) | 1.95 | - | - | 7/27 | 26% | Denominator for case fatality - screened group includes those LTFU pre-treatment; PCF those starting treatment only. Baseline characteristics of diagnosed in screened and on treatment in PCF groups - no difference in age, gender, smear grade between groups. |
| PCF | 379/473 | 80% (76-84%) | 18/473 | 4% (2-6%) | - | - | - | - | ||||
| Santha 2003 India CXR and symptoms |
Screen | 45/65 | 69% (57-80%) | 1.01 | 4/65 | 6% (2-15%) | 0.88 | 13/65 | 20% (11-32%) | 31/96 | 32% | Baseline characteristics of all (smear +ve and -ve) diagnosed in screened and on treatment in PCF groups - screened group more likely to be older, male, illiterate, sole earner, have poor quality house, 1 room house, lower smear grade and new smear -ve disease. |
| PCF | 225/330 | 68% (63-73%) | 23/330 | 7% (4-10%) | 63/330 | 19% (15-24%) | - | - | ||||
| Harper 1996 Nepal Symptoms |
Screen | 50/64 | 78% (66-87%) | 1.00 | 5/64 | 8% (3-17%) | 0.96 | 4/64 | 6% (2-15%) | - | - | Baseline characteristics of diagnosed TB patients (screened vs PCF) – screened more likely to be female (and age among women tended to be older). |
| PCF | 997/1272 | 78% (76-81%) | 104/1272 | 8% (7-10%) | 96/1272 | 8% (6-9%) | - | - | ||||
| Cassel 1982 Nepal Symptoms |
Screen | - | - | 9/111 | 8% (4-15%) | 0.76 | - | - | 11/111 | 10% | Denominator for case fatality - screened group includes those LTFU pre-treatment; PCF group are those starting treatment. Baseline characteristics of diagnosed TB patients (screened vs PCF) – screened group were older and the male to female ratio was lower. | |
| PCF | - | - | 17/159 | 11% (6-17%) | - | - | - | - | ||||
| Risk groups | ||||||||||||
| Shewade 2019 India; Marginalised and vulnerable† Symptoms |
Screen | 247/274 | 90% (86-93%) | 1.03 | 7/274 | 3% (1-5%) | 0.69 | 16/274 | 6% (3-9%) | - | - | Baseline characteristics of on treatment TB patients (screened vs PCF)- screened group more likely to be older, from rural areas and live further from microscopy units. No association between screening and treatment success after adjusting for age, gender and distance from microscopy unit. |
| PCF | 260/296 | 88% (83-91%) | 11/296 | 4% (2-7%) | 22/296 | 7% (5-11%) | - | - | ||||
| Verver 2001 Netherlands: Migrants CXR |
Screen | 384/454 | 85% (81-88%) | 1.06 | 1/454 | 0.2% (0-1%) | 0.07 | 47/454 | 10% (8-14%) | - | - | Baseline characteristics of on treatment TB patients (screened vs PCF) - screen detection varied by country of origin, decreased with increasing length of stay and was less likely among illegal migrants. |
| PCF | 293/368 | 80% (75-84%) | 12/368 | 3% (2-6%) | 36/368 | 10% (7-13%) | - | - | ||||
| Churchyard 2000 South Africa: Miners CXR |
Screen | - | - | 12/1225 | 1% (0.5-2%) | 0.14 | - | - | - | - | Baseline characteristics of on treatment TB patients (screened vs PCF) - screened less likely to be HIV infected. After adjusting for HIV status, sputum status, treatment category, age, disease extent on CXR, silicosis and drug resistance, association between PCF and case fatality maintained (PCF versus screened aOR 5.6; 95%CI 2.6-12.2) |
|
| PCF | - | - | 69/1011 | 7% (5-9%) | - | - | - | - | ||||
| Cluster randomised controlled trials | ||||||||||||
| First author, country and population, screening tool | Community, number and baseline data | Results | ||||||||||
| General population | ||||||||||||
| Shargie 2006∆ Ethiopia: Symptoms |
87 contiguous administrative units clustered into 32 communities 32 communities randomised – 12 to screening and 20 to PCF NŦ smear +ve TB patients - screen=159; PCF=221 Follow-up during treatment Communities and TB patients - similar baseline characteristics between groups |
Treatment success: screen vs PCF | n=128 (81%) vs n=165 (75%); difference (95%CI) 6 (-4 to 15); p=0.12 | |||||||||
| Death: screen vs PCF | n=5 (3.1%) vs n=7 (3.2%); difference (95%CI) -0.1 (-4 to 4); p=0.49 | |||||||||||
| LTFU on treatment: screen vs PCF | n=26 (16%) vs n=48 (22%); difference (95%CI) -6 (-14 to 3); p=0.11 | |||||||||||
| Risk groups | ||||||||||||
| Jenum 2018 India: neonates Symptoms |
Cluster – villages or subsection of towns 592 clusters randomised (8 strata) – 297 to screening and 295 to PCF NŦ in each group - screen=2215; PCF=2167 Follow-up 2 years Study groups – PCF group had more Hindus, lower paternal literacy and higher use of wood/agricultural residues for fuel. No difference in other characteristics |
All-cause mortality: screen vs PCF | n=49 (2.2%) vs n=71 (3.3%); aOR¶ (95%CI) 0.68 (0.47-0.98) | |||||||||
| Cause of death: screen vs PCF | Reduction in deaths due to pneumonia/respiratory infections (aORƳ 0.34; 95%CI 0.14-0.80). | |||||||||||
| LTFU: screen vs PCF | n=38 (1.7%) vs n=60 (2.8%); aOR¶ (95%CI) 0.62 (0.41-0.94) | |||||||||||
| Fox 2018 Vietnam: household contacts CXR and symptoms |
70 of 112 districts in 8 Vietnamese provinces selected with probability proportional to population. 70 districts randomised – 36 to screened and 34 to PCF NŦ in each group - screen=10,069; PCF=15,638 Follow-up 2 years Study groups – PCF group household size higher and lower proportion reported prior history of TB. |
All-cause mortality: screen vs PCF | n=60 (0.6%) vs 265 (1.7%); RR (95%CI) 0.60 (0.50-0.80) | |||||||||
CRT=cluster randomised controlled trial; PR=prevalence ratio (screened/passive case finding population); LTFU=loss to follow-up.
pre-treatment LTFU =lost to follow-up between diagnosis and treatment start; n/N=number with outcome/total number started on TB treatment (unless otherwise indicated); 95%CI = 95% confidence interval.
n/N=number lost to follow-up pre-treatment/total number diagnosed with TB.
included slums, tribal areas, scheduled caste communities, areas where occupational lung diseases is high, areas where individuals with high risk of acquiring TB reside including stone crushing/mining/weaving industry/unorganized labour (construction workers etc)/homeless, high HIV/AIDS burden areas, areas or communities with high TB incidence (including prisons) and among household contacts of sputum smear positive TB patients; PCF=passive case-finding; CXR=chest radiographs; +ve=positive; -ve=negative; aOR=adjusted odds ratio; ŦDenominator in each study group.
adjusted for clustering, gender, religion, father's education and fuel type used.
adjusted for clustering, gender, religion and father's education; RR=relative risk; ∆Data not shown in table - weighted mean of median pre-treatment symptom duration 89 days in screened vs 136 days in control group (difference [95%CI] -47 [-76 to -19]; p=0.001).