Summary of findings 1. Tonsillectomy compared to tonsillotomy for obstructive sleep‐disordered breathing in children.
| Tonsillectomy compared to tonsillotomy for obstructive sleep‐disordered breathing in children | ||||||
| Patient or population: children aged 2 years up to the age of 16 years with obstructive sleep‐disordered breathing Setting: secondary or tertiary care Intervention: tonsillectomy Comparison: tonsillotomy | ||||||
| Outcomes | № of participants (studies) | Relative effect (95% CI) | Anticipated absolute effects* (95% CI) | Certainty of the evidence (GRADE) | Comments | |
| Risk with tonsillotomy | Risk with tonsillectomy | |||||
| Clinical effectiveness expressed as disease‐specific quality of life (Measured using a validated instrument) Follow‐up: 0 to 6 months, 7 to 12 months and 13 to 24 months |
410 (3 RCTs) | — | Three studies reported no evidence of a difference between treatment groups at 0 to 6 months | ⊕⊝⊝⊝ very low1 | It is very uncertain whether there is any difference in disease‐specific quality of life in the short (0 to 6 months), medium (7 to 12 months) or long term (13 to 24 months) | |
| 79 (1 RCT) |
— | One study reported no evidence of a difference between treatment groups at 7 to 12 months | ⊕⊝⊝⊝ very low2 | |||
| 161 (2 RCTs) |
— | Two studies reported no evidence of a difference between treatment groups at 13 to 24 months | ⊕⊝⊝⊝ very low3 | |||
| Peri‐operative blood loss (Volume measured in mL) |
610 (8 RCTs) | — | Peri‐operative blood loss volume ranged from 11 mL to 45 mL | MD 14.06 mL higher
(1.91 higher to 26.21 higher) |
⊕⊝⊝⊝ very low4 | Although tonsillotomy might reduce peri‐operative blood loss, the reduction was not clinically meaningful and the evidence is very uncertain. A further 2 studies did not provide crude data; 1 reported no difference in blood loss and the other reported less bleeding in the children who underwent tonsillotomy. |
| Postoperative complications requiring medical intervention (with or without hospitalisation) Follow‐up: 7 days |
1416 (16 RCTs) | RR 1.75 (1.06 to 2.91) | Study population (0 to 7 days) | ⊕⊕⊕⊝ moderate5 | The risk of postoperative complications in the first week after surgery was probably lower in children who underwent tonsillotomy. One further study reported no complications requiring intervention. |
|
| 26 per 1000 | 46 per 1000 (28 to 76) | |||||
| Severity of postoperative pain (Rated by parents using a 10‐point visual analogue scale) Follow‐up: 24 hours, 2 to 3 days and 4 to 7 days |
368 (4 RCTs) | — | The mean severity of postoperative pain at 24 hours ranged from 3.2 to 5.6 | MD 1.09 higher (0.88 higher to 1.29 higher) | ⊕⊝⊝⊝ very low6 | When considering pain at 24 hours postoperatively we found only a small difference between tonsillectomy and tonsillotomy but the evidence is very uncertain. A further 7 trials did not provide crude data; 6 trials reported less pain in the children who underwent tonsillotomy and 1 trial reported no difference in pain between the groups. |
| 301 (3 RCTs) | — | The mean severity of postoperative pain at 2 to 3 days ranged from 2.7 to 5.3 | MD 0.93 higher (0.14 lower to 2.00 higher) | ⊕⊝⊝⊝ very low7 | When considering pain at two to three days postoperatively we found no evidence of a difference between tonsillectomy and tonsillotomy but the evidence is very uncertain. A further 2 trials did not provide crude data; all reported less pain in the children who underwent tonsillotomy. |
|
| 370 (4 RCTs) | — | The mean severity of postoperative pain at 4 to 7 days ranged from 1.9 to 3.7 | MD 1.07 higher (0.40 lower to 2.53 higher) | ⊕⊝⊝⊝ very low7 | When considering pain at four to seven days postoperatively we found no evidence of a difference between tonsillectomy and tonsillotomy but the evidence is very uncertain. A further 3 trials did not provide crude data; all reported less pain in children who underwent tonsillotomy. |
|
| Return to normal activity Follow‐up: 14 days |
284 (3 RCTs) | — | The mean return to normal activity ranged from 2.4 to 12.3 days | MD 3.84 higher (0.23 higher to 7.44 higher) | ⊕⊕⊝⊝ moderate8 | Tonsillotomy probably results in a faster return to normal activity (4 days). A further 4 trials did not provide crude data; 3 reported that the median number of days was shorter in children who underwent tonsillotomy; 1 trial reported no difference between the groups. |
| Recurrence of oSDB as a result of tonsil regrowth Follow‐up: 0 to 6 months, 7 to 12 months and 13 to 24 months |
186 (3 RCTs) | RR 0.26 (0.03 to 2.22) | Study population (0 to 6 months) | ⊕⊝⊝⊝ very low9 | We found no evidence of a difference in the risk of recurrence of oSDB between the 2 groups in the short term (0 to 6 months), medium (7 to 12 months) or long term (13 to 24 months). One further trial did not provide crude data and reported no recurrence in either group at 13 to 24 months. |
|
| 33 per 1000 | 8 per 1000 (1 to 72) | |||||
| 206 (4 RCTs) | RR 0.35 (0.04 to 4.23) | Study population (7 to 12 months) | ⊕⊝⊝⊝ very low9 | |||
| 48 per 1000 | 9 per 1000 (1 to 60) | |||||
| 65 (1 RCT) | RR 0.21 (0.01 to 4.13) | Study population (13 to 24 months) | ⊕⊝⊝⊝ very low9 | |||
| 61 per 1000 | 13 per 1000 (1 to 250) | |||||
| Reoperation rates Follow‐up: 7 to 12 months and 13 to 24 months |
166 (2 RCTs) | RR 0.32 (0.08 to 1.28) | Study population (7 to 12 months) | ⊕⊝⊝⊝ very low10 | We found no evidence of a difference in reoperation rates between the 2 groups in the medium term (7 to 12 months) or long term (13 to 24 months). |
|
| 92 per 1000 | 29 per 1000 (7 to 118) | |||||
| 41 (1 RCT) | RR 0.35 (0.02 to 8.10) | Study population (13 to 24 months) | ⊕⊝⊝⊝ very low10 | |||
| 48 per 1000 | 17 per 1000 (1 to 386) | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; MD: mean difference; oSDB: obstructive sleep‐disordered breathing; RCT: randomised controlled trial; RR: risk ratio | ||||||
| GRADE Working Group grades of evidence High quality: We are very confident that the true effect lies close to that of the estimate of the effect Moderate quality: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low quality: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low quality: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect | ||||||
1Our confidence in this estimate is 'very low' because of very serious limitations in study methodology, serious imprecision and suspected publication bias, with only three studies reporting on this outcome (in a manner that precluded meta‐analysis).
2Our confidence in the estimate is 'very low' because of very serious limitations in study methodology, serious imprecision and suspected publication bias, with only one study reporting on this outcome.
3Our confidence in the estimate is 'very low' because of very serious limitations in study methodology, serious imprecision and suspected publication bias, with only two studies reporting on this outcome.
4Our confidence in the estimate is 'very low' due to inconsistency of effect estimates between main and sensitivity analyses as well as across individual trials (statistical heterogeneity) and imprecision of the evidence based on the wide confidence intervals.
5Our confidence in the effect estimate is 'moderate' due to imprecision of the evidence based on the wide confidence intervals.
6Our confidence in the effect is 'very low' due to inconsistency of effect estimates between main and sensitivity analyses as well as across individual trials (statistical heterogeneity), imprecision of the evidence based on the wide confidence intervals and publication bias based on the small proportion of studies that reported data in a manner that permitted meta‐analysis.
7Our confidence in the effect is 'very low' due to inconsistency of effect estimates across individual trials (statistical heterogeneity), imprecision of the evidence based on the wide confidence intervals and publication bias based on the small proportion of studies that reported data in a manner that permitted meta‐analysis.
8Our confidence in the effect is 'moderate' due to imprecision of the evidence based on the wide confidence intervals and publication bias based on the small proportion of studies that reported data in a manner that permitted meta‐analysis.
9Our confidence in the effect is 'very low' due to very serious limitations in study methodology, inconsistency of effect estimates across individual trials (statistical heterogeneity), imprecision of the evidence and publication bias, with only a small number of studies reporting on this outcome.
10Our confidence in the effect is 'very low' due to very serious limitations in study methodology, imprecision of the evidence and publication bias, with only one small study reporting on this outcome.