Prescrire Int. 2004 Dec;13(74):227-32.
(1) Since the 1940s, a large number of comparative randomised placebo-controlled trials have evaluated antibiotic therapy for pharyngitis, initially parenteral benzathine benzylpenicillin, then oral phenoxymethylpenicillin (penicillin V). Our literature search identified a Cochrane meta-analysis of all these trials, with the exception of one published in 2003. (2) When group A betahemolytic streptococci (group A streptococci) are present in the throat, antibiotic therapy accelerates symptom relief (particularly fever and pain) by a day or two. This has been shown with 7-day treatments but not with 3-day treatments. There is no convincing evidence that antibiotics relieve symptoms in children. (3) According to the Cochrane meta-analysis, signs of progression to locoregional suppuration were noted in 1% of patients receiving placebo, compared to 0.09% of patients receiving antibiotics in the most recent trials (statistically significant difference). (4) Comparative trials done in the 1950s showed that benzathine benzylpenicillin helped prevent acute rheumatic fever, reducing the risk by about 75%. Since 1985 nearly 1000 patients with pharyngitis have been given a placebo in clinical trials, and none have developed acute rheumatic fever. (5) There is no firm evidence that antibiotics reduce the risk of acute glomerulonephritis. (6) The adverse effects associated with most antibiotics are mild. This is especially true for penicillin. However, there is a risk of rare but serious adverse effects: anaphylaxis is estimated to occur in 5 per 10 000 patients treated with injectable penicillin, while the risk associated with oral penicillin used to treat pharyngitis has not been quantified. Moreover, antibiotics affect the bacterial ecology, encouraging resistance among some bacterial species other than group A streptococci. (7) A strategy based on the use of a clinical diagnostic score, followed by a rapid test if the score is intermediate, seems to be the best way of restricting antibiotics to patients with pharyngitis due to group A streptococci. (8) In patients with group A streptococcal pharyngitis, a strategy of starting antibiotics only after 48 hours of symptoms delays symptom control but seems to reduce the risk of relapse. According to a clinical trial in patients with pharyngitis from all causes, advising patients to postpone antibiotic therapy reduces antibiotic use by about 85%, without increasing the risk of serious clinical complications. (9) In practice, immediate antibiotic therapy is justified for patients with severe symptoms or signs of progression to locoregional suppuration, and when the local incidence of acute rheumatic fever is high. In other situations, whether or not group A streptococci are involved, antibiotic therapy should be started only if symptoms do not begin to improve after 48 hours of symptomatic treatments.
（1）自20世纪40年代以来，大量比较性随机安慰剂对照试验评估了咽炎的抗生素治疗，最初是胃肠外注射苄星青霉素，然后是口服苯氧甲基青霉素（青霉素V）。我们的文献检索确定了除2003年发表的一项试验外所有这些试验的Cochrane荟萃分析。（2）当咽喉部存在A组β溶血性链球菌（A组链球菌）时，抗生素治疗可使症状缓解（尤其是发热和疼痛）加速一两天。这在7天疗程中得到了证实，但在3天疗程中未得到证实。没有令人信服的证据表明抗生素能缓解儿童的症状。（3）根据Cochrane荟萃分析，接受安慰剂的患者中有1%出现局部化脓进展的迹象，而在最近的试验中接受抗生素治疗的患者这一比例为0.09%（具有统计学显著差异）。（4）20世纪50年代进行的比较试验表明，苄星青霉素有助于预防急性风湿热，将风险降低约75%。自1985年以来，近1000例咽炎患者在临床试验中接受了安慰剂治疗，无一例发生急性风湿热。（5）没有确凿证据表明抗生素能降低急性肾小球肾炎的风险。（6）大多数抗生素相关的不良反应较轻。青霉素尤其如此。然而，存在罕见但严重的不良反应风险：估计每10000例接受注射用青霉素治疗的患者中有5例发生过敏反应，而用于治疗咽炎的口服青霉素相关风险尚未量化。此外，抗生素会影响细菌生态，促使A组链球菌以外的某些细菌产生耐药性。（7）基于临床诊断评分，然后在评分处于中等水平时进行快速检测的策略，似乎是将抗生素限制用于A组链球菌所致咽炎患者的最佳方法。（8）在A组链球菌性咽炎患者中，仅在症状出现48小时后才开始使用抗生素的策略会延迟症状控制，但似乎能降低复发风险。根据一项针对所有病因咽炎患者的临床试验，建议患者推迟抗生素治疗可使抗生素使用减少约85%，而不会增加严重临床并发症的风险。（9）在实践中，对于有严重症状或有局部化脓进展迹象的患者，以及当地急性风湿热发病率较高时，立即使用抗生素是合理的。在其他情况下，无论是否涉及A组链球菌，只有在对症治疗48小时后症状仍未开始改善时才应开始使用抗生素治疗。
