A study of over a thousand teens in the UK found text and picture warnings on the backs of packets had little impact on teens' perceptions of the risks of smoking.
Earlier this year, the DH shelved plans to introduce plain packaging for tobacco products, saying it would wait to see the results of a similar scheme in Australia.
Early findings from Australia, published in the BMJ Open in July, found plain packaging appears to make tobacco less appealing and increase the urgency to quit.
This latest research suggests current on-packet warnings may be failing to curb habits among teenage smokers.
In the study, published in the journal Tobacco Control, researchers analysed responses from the Youth Tobacco Policy Survey conducted in 2008 and 2011, when 1,401 and 1,373 teens aged 11-16 years took part respectively.
The UK introduced pictoral health warnings to the back of packets in 2008, although these were on packets during the 2011 survey only.
The warnings include images of diseased lungs, rotten teeth and neck cancer.
The surveys found most teens had never smoked, while 17-22% had experimented with cigarettes and one in 10 were regular smokers.
Half had 'often' or 'very often' noticed the warnings, although the proportion of regular smokers who noticed them fell from 77% to 66% between 2008 and 2011.
Only one in 10 teens said they considered the warnings when the packet was out of sight.
Between 2008 and 2011, a greater proportion of experimental and never-smokers believed the warnings could deter them from smoking, but there was no change in regular smokers.
Recall of the text warnings and images declined among all teens during the three years between surveys.
Study authors said putting warning images only on the reverse of packets 'limits their impact'.
'While recall was high at both waves for pack-front warnings, it was low (below 10%) for the pictorial warnings on the pack reverse, fear-appeal pictures aside,' they said.
They concluded: 'Positioning pictorial warnings only on the back of packs may have had a deterrent effect on never and experimental smokers, but for most measures no significant differences were observed. The impact on regular smokers was negligible.'