Table 2.

Summary of GRADE evidence profile of antibiotics vs placebo or standard care

Outcome/timeframe	Study results and measurements	Absolute effect estimates		Certainty in effect estimates (quality of evidence)	Plain text summary
		No antibiotics	Antibiotics
Treatment failure/1 month	OR 0.58 (95% CI 0.37 to 0.90) Based on data from 2517 patients in eight studies Follow-up 7–21 days	93 per 1000	56 per 1000	Low Due to serious risk of bias and serious inconsistency*	Antibiotics probably reduce the risk of treatment failure.
		Difference: 37 fewer per 1000 (95% CI 56 fewer—9 fewer)
Treatment failure (antibiotics with activity against MRSA)/1 month	OR 0.45 (95% CI 0.33 to 0.62) Based on data from 2305 patients in six studies Follow-up 7–21 days	128 per 1000	62 per 1000	High	Antibiotics with activity against MRSA reduce the risk of treatment failure.
		Difference: 66 fewer per 1000 (95% CI 82 fewer—45 fewer)
Treatment failure (antibiotics without activity against MRSA)/1 month	OR 1.82 (95% CI 0.68 to 4.85) Based on data from 212 patients in two studies Follow-up 7–21 days	58 per 1000	101 per 1000	Moderate Due to serious imprecision†	Antibiotics without activity against MRSA may not reduce the risk of treatment failure.
		Difference: 43 more per 1000 (95% CI 18 fewer—172 more)
Recurrence within/1 month	OR 0.48 (95% CI 0.30 to 0.77) Based on data from 2134 patients in six studies Follow-up 7–30 days	129 per 1000	66 per 1000	Moderate Due to serious risk of bias and borderline inconsistency‡	Antibiotics probably reduce the risk of early abscess recurrence.
		Difference: 63 fewer per 1000 (95% CI 86 fewer—27 fewer)
Late recurrence/1−3 months	OR 0.64 (95% CI 0.48 to 0.85) Based on data from 1111 patients in two studies Follow-up 63–90 days	267 per 1000	189 per 1000	Moderate Due to serious risk of bias, borderline imprecision§	Antibiotics probably reduce the risk of late abscess recurrence.
		Difference: 78 fewer per 1000 (95% CI 118 fewer—31 fewer)
Hospitalisation/3 months	OR 0.55 (95% CI 0.32 to 0.94) Based on data from 1206 patients in two studies Follow-up 40–90 days	39 per 1000	22 per 1000	Moderate Due to serious imprecision¶	Antibiotics probably reduce the risk of hospitalisation.
		Difference: 17 fewer per 1000 (95% CI 26 fewer—2 fewer)
Pain (tenderness)/(3–4 days)	OR 0.76 (95% CI 0.60 to 0.97) Based on data from 1057 patients in one study Follow-up 3–4 days	559 per 1000	491 per 1000	Moderate Due to serious imprecision**	Antibiotics probably increase the risk of pain at 3–4 days.
		Difference: 68 fewer per 1000 (95% CI 126 fewer—8 fewer)
Pain (tenderness)/(8–10 days)	OR 0.56 (95% CI 0.35 to 0.88) Based on data from 1057 patients in one study Follow-up 8–10 days	101 per 1000	59 per 1000	Moderate Due to serious imprecision††	Antibiotics may not increase the risk of pain at 8–10 days.
		Difference: 42 fewer per 1000 (95% CI 63 fewer—11 fewer)
Additional surgical procedures within/1–3 months	OR 0.58 (95% CI 0.39 to 0.87) Based on data from 1013 patients in one study Follow-up 43–63 days	136 per 1000	84 per 1000	Moderate Due to serious imprecision‡‡	Antibiotics probably increase the risk of additional surgical procedures.
		Difference: 52 fewer per 1000 (95% CI 78 fewer—16 fewer)
Infections in family members within/1 month	OR 0.58 (95% CI 0.34 to 1.01) Based on data from 1013 patients in one study Follow-up 7–14 days	67 per 1000	40 per 1000	Moderate Due to serious imprecision§§	Antibiotics probably do not increase the risk of infection in family members.
		Difference: 27 fewer per 1000 (95% CI 43 fewer—1 more)
Invasive infections/1 month	OR 1.02 (95% CI 0.14 to 7.24) Based on data from 1057 patients in one study Follow-up 7–14 days	4 per 1000	4 per 1000	Moderate Due to serious imprecision¶¶	Antibiotics probably do not reduce the risk of serious complications at 7–14 days.
		Difference: 0 more per 1000 (95% CI 3 fewer—24 more)
Invasive infections/3 months	OR 7.46 (95% CI 0.15 to 376.12) Based on data from 1013 patients in one study Follow-up 42–56 days	0 per 1000	1 per 1000	Moderate Due to serious imprecision***	Antibiotics probably do not reduce the risk of serious complications at 42–56 days.
		Difference: 2 more per 1000 (95% CI 4 fewer—8more)

Evidence have summarised at Magic App (www.magicapp.org/public/guideline/jlRvQn).

*Risk of bias: serious. There was substantial missing data/lost to follow-up: the results are not robust to worth plausible sensitivity analysis (assuming that missing patients from the control arms have the same rate of treatment failure as those with complete follow-up, and five times the rate of treatment failure in the patients who were lost to follow-up in the antibiotic arm); inconsistency: serious. Effects might differ in different type of antibiotics.

†Imprecision: serious. CI approaches no effect.

‡Risk of bias: serious. There was substantial missing data/lost to follow-up: the results are not robust to worth plausible sensitivity analysis; inconsistency: no serious. The magnitude of statistical heterogeneity was high, with I²=45%, but the direction of effect was similar in almost all trials, favouring antibiotics over no antibiotics.

§Risk of bias: serious. Incomplete data and/or large loss to follow-up: results are not sensitive to worst plausible sensitivity analysis: OR 1.48, 95% CI 0.55 to 3.96; imprecision: no serious. A single large study, and one small study contributed data to this outcome.

¶Imprecision: serious. CI approaches no effect.

**Imprecision: serious. Only data from one study, CI approaches no effect.

††Imprecision: serious. Only data from one study.

‡‡Imprecision: serious. Data from one study only.

§§Imprecision: serious. Only data from one study; CI include no effect.

¶¶Imprecision: serious. Only data from one study.

***Imprecision: serious. Only data from one study; CI include no effect.

GRADE, Grading of Recommendations Assessment, Development and Evaluation; MRSA, methicillin-resistant Staphylococcus aureus.