Table 3.
Meta-analysis of the association of ART with cervical lesion incidence, progression and regression, and invasive cervical cancer incidence among women living with HIV
|
Crude analysis* |
Adjusted analysis† |
|||||||
|---|---|---|---|---|---|---|---|---|
| n studies | HR (95%CI)‡ | I2 | p value for heterogeneity | n studies | HR (95%CI)‡ | I2 | p value for heterogeneity | |
| SIL incidence | ||||||||
| All | 10 | 0·75 (0·56–1·00) | 40·9% | 0·09 | 5 | 0·64 (0·47–0·86) | 19·4% | 0·29 |
| Africa | 2 | 0·59 (0·44–0·80) | 0% | 0·71 | 1 | 0·62 (0·42–0·91) | ·· | ·· |
| Latin America | 1 | 1·90 (0·90–4·01) | ·· | ·· | ·· | ·· | ·· | ·· |
| Europe or North America | 7 | 0·73 (0·52–1·03) | 14·0% | 0·32 | 4 | 0·64 (0·40–1·02) | 39·0% | 0·18 |
| SIL progression | ||||||||
| All | 6 | 0·64 (0·56–0·74) | 0% | 0·42 | 4 | 0·64 (0·54–0·75) | 17·8% | 0·30 |
| Africa | 3 | 0·67 (0·56–0·79) | 0% | 0·68 | 2 | 0·68 (0·57–0·80) | 0% | 0·65 |
| Europe or North America | 3 | 0·62 (0·43–0·90) | 46·4% | 0·16 | 2 | 0·57 (0·39–0·85) | 58·0% | 0·12 |
| SIL-CIN regression | ||||||||
| All | 6 | 1·61 (1·31–1·97) | 18·3% | 0·30 | 5 | 1·54 (1·30–1·82) | 0% | 0·42 |
| Africa | ·· | ·· | ·· | ·· | 1 | 1·71 (1·29–2·27) | ·· | ·· |
| Europe or North America | 5 | 1·62 (1·21–2·16) | 28·4% | 0·23 | 4 | 1·45 (1·17–1·81) | 1·8% | 0·38 |
| Invasive cervical cancer incidence | ||||||||
| All | 2 | 0·40 (0·18–0·87) | 32·7% | 0·22 | 1 | 0·50 (0·29–0·87) | ·· | ·· |
*
Includes studies with no adjustment potential confounders and studies that adjust for sociodemographic factors only but no adjustment for HIV related factors.
†
Includes studies that adjusted for time-varying ART or time-varying CD4 cell count.
‡
Only studies that reported HR from time-to-event analysis included in the meta-analysis. HR=hazard ratio. SIL=squamous intraepithelial lesions. CIN=cervical intraepithelial neoplasia.