The Lancet later suggested the conflict had caused the deaths of 650,000
Government uncertainty about how to reply to a medical journal's article on the numbers killed in Iraq has been revealed in official documents.
In October 2004 the Lancet published estimates that 98,000 people died in the conflict's first 18 months.
Then-Prime Minister Tony Blair rejected the figure, but one official warned the government would be "ripped apart" if it questioned the Lancet's methodology.
The government tried to stop publication, but was over-ruled.
A subsequent Lancet article claimed that the number of deaths caused by the conflict up to 2006 was more than 650,000.
The documents - released after the Information Commissioner dismissed government objections - reveal officials' confusion as they tried to work out how to cast doubt on the 98,000 figure.
The Foreign Office said the Ministry of Defence should take the lead on the issue, but the MoD said it was the Foreign Office's job.
While then-Foreign Secretary Jack Straw said the government was looking into the statistics, Mr Blair told Parliament: "We do not accept these figures at all."
The bureaucratic tangles reflected the contradiction that while government experts declared the Lancet's methodology to be sound, officials did not want to accept the 98,000 estimate.
One wrote that they would be "ripped apart" if they criticised the methodology.
Another commented: "I think we need a bit more explanation why our soldiers can't do the counting: if they are able to kill people they are able to count them."
And one memo, presciently, urged caution on the grounds that it "would not be at all surprising if we one day have to release the documents on which our assessment of the Lancet is based".