A University of Hull team concluded the drugs actively help only a small group of the most severely depressed.
Marjorie Wallace, head of the mental health charity Sane, said that if these results were confirmed they could be "very disturbing".
But the makers of Prozac and Seroxat, two of the commonest anti-depressants, said they disagreed with the findings.
A spokesman for GlaxoSmithKline, which makes Seroxat, said the study only looked at a "small subset of the total data available".
And Eli Lilly, which makes Prozac, said that "extensive scientific and medical experience has demonstrated it is an effective anti-depressant".
Alan Johnson, the Health Secretary, has announced that 3,600 therapists are to be trained during the next three years in England to increase patient access to talking therapies, which ministers see as a better alternative to drugs.
Patients are strongly advised not to stop taking their medication without first consulting a doctor.
The researchers accept many people believe the drugs do work for them, but argue that could be a placebo effect - people feel better simply because they are taking a medication which they think will help them.
In total, the Hull team, who published their findings in the journal PLoS Medicine, reviewed data on 47 clinical trials.
They reviewed published clinical trial data, and unpublished data secured under Freedom of Information legislation.
They focused on drugs which work by increasing levels of the mood controlling chemical serotonin in the brain.
These included fluoxetine (Prozac) and paroxetine (Seroxat), from the class known as Selective Serotonin Reuptake Inhibitors (SSRIs), alongside another similar drug called venlafaxine (Efexor) - all commonly prescribed in the UK.
The number of prescriptions for anti-depressants hit a record high of more than 31 million in England in 2006 - even though official guidance stresses they should not be a first line treatment for mild depression.
There were 16.2m prescriptions for SSRIs alone.
The researchers found that the drugs did have a positive impact on people with mild depression - but the effect was no bigger than that achieved by giving patients a sugar-coated "dummy" pill.
People with severe symptoms appeared to gain more clear-cut benefit - but this might be more down to the fact that they were less likely to respond to the placebo pill, rather than to respond positively to the drugs.
Lead researcher Professor Irving Kirsch said: "The difference in improvement between patients taking placebos and patients taking anti-depressants is not very great.
"This means that depressed people can improve without chemical treatments.
"Given these results, there seems little reason to prescribe anti-depressant medication to any but the most severely depressed patients, unless alternative treatments have failed to provide a benefit."
Professor Kirsch said the findings called into question the current system of reporting drug trials.
Dr Tim Kendall, deputy director of the Royal College of Psychiatrists Research Unit, has published research concluding that drug companies tend only to publish research which shows their products in a good light.
He said the Hull findings undermined confidence in the ability to draw meaningful conclusions about the merit of drugs based on published data alone.
He called for drug companies to be forced to publish all their data.
The National Institute for Health and Clinical Excellence (NICE) is currently reviewing its guidance on the use of antidepressants.
Marjorie Wallace of Sane commented: "If these results were upheld in further studies, they would be very disturbing.
"The newer anti-depressants were the great hope for the future.... These findings could remove what has been seen as a vital choice for thousands in treating what can be a life-threatening condition."
Dr Andrew McCulloch, of the Mental Health Foundation, said: "We have become vastly over-reliant on antidepressants when there is a range of alternatives.
"Talking therapies, exercise referral and other treatments are effective for depression.
"It is a problem that needs a variety of approaches matched to the individual patient."
Dr Richard Tiner, of the Association of the British Pharmaceutical Industry, said there was no doubt that was a "considerable placebo effect" from anti-depressants when treating people with mild to moderate symptoms.
But he said no medicine would get a licence without demonstrating it was better than a placebo.
Dr Tiner said: "These medicines have been licensed by a number of regulatory authorities around the world, who looking at all the evidence, have determined that they do work better than placebo."