"Patients given amoxicillin don't recover much quicker or have significantly fewer symptoms," explains Paul Little, Professor of Primary Care Research at the University of Southampton.

"Indeed, using amoxicillin to treat respiratory infections in patients not suspected of having pneumonia is not likely to help and could be harmful. Overuse of antibiotics, which is dominated by primary care prescribing, particularly when they are ineffective, can lead to side effects such as diarrhea, rash, vomiting and the development of resistance."

LRTI (chest infections) are one of the most common acute illnesses treated in primary care in developed countries. Although viruses are believed to cause most of these infections, whether or not antibiotics are beneficial in the treatment of LRTI, particularly in older patients, is still hotly debated. Research so far has produced conflicting results.

In the study, from the GRACE (Genomics to Combat Resistance against Antibiotics in Community-acquired LRTI in Europe) consortium and funded by the European Community's Sixth Framework Programme, 2061 adults with acute uncomplicated LRTI from primary care practices in 12 European countries (England, Wales, Netherlands, Belgium, Germany, Sweden, France, Italy, Spain, Poland, Slovenia, and Slovakia) were randomly assigned to receive either amoxicillin or a placebo three times a day for seven days. Doctors assessed symptoms at the start of the study and participants completed a daily symptom diary.

Little difference in severity or duration of symptoms was reported between the two groups. This was true even for older patients aged 60 or more who were generally healthy, in whom antibiotics appeared to have a very limited effect.

Although significantly more patients in the placebo group experienced new or worsening symptoms (19.3% vs 15.9%), the number needed to treat was high (30), and just two patients in the placebo group and one in the antibiotic group required hospitalisation.

What is more, patients taking antibiotics reported significantly more side effects including nausea, rash, and diarrhea, than those given placebo (28..7% vs 24%).

Professor Little adds: "Our results show that most people get better on their own. But, given that a small number of patients will benefit from antibiotics the challenge remains to identify these individuals."

Writing in a linked comment, Philipp Schuetz from the University of Basel in Switzerland says: "Little and colleagues have generated convincing data that should encourage physicians in primary care to refrain from antibiotic treatment in low-risk patients in whom pneumonia is not suspected. Whether this one-size-fits-all approach can be further improved remains to be seen. Guidance from measurements of specific blood biomarkers of bacterial infection might help to identify the few individuals who will benefit from antibiotics despite the apparent absence of pneumonia and avoid the toxic effects and costs of those drugs and the development of resistance in the other patients."