At least 40 motorists convicted of drug-driving offences have been cleared after evidence of manipulation was found in the forensic testing process.
The motorists were banned from driving and in some cases fined, but their convictions have since been overturned.
About 10,500 test results are being reviewed after data was allegedly manipulated at Randox Testing Services.
The National Police Chiefs’ Council (NPCC) described it as a “most serious breach” of forensic science standards.
A further 50 drug-driving cases have been dropped as a result of the alleged data manipulation at the firm’s Manchester laboratory.
Police suspended all contracts with the company, used by 42 of the UK’s 43 forces, in November last year.
Greater Manchester Police has arrested two men, aged 47 and 31, who worked at one of its laboratories.
They were arrested on suspicion of perverting the course of justice.
Six others have been interviewed under caution, with one remaining under investigation as part of what Justice Minister, Lucy Frazier, described last month as an “expansive” criminal inquiry.
Danny Shaw, BBC home affairs correspondent
The scale of the operation to review the 10,500 Randox cases is unprecedented.
It began in January 2017 and is likely to continue until the end of 2019.
Randox, which is paying for samples to be re-tested by other laboratories, estimates it will cost the firm £2.5m.
Police are also likely to incur costs because of delays to other cases, while motorists whose careers and livelihoods have been affected after being wrongly banned from driving may sue for compensation.
But the most significant impact of this disturbing affair may be on public confidence in forensic science: can we be sure that the test results we almost take for granted are accurate?
In total, up to 2,700 cases have been re-analysed so far and two drug-driving cases that resulted in road deaths were referred to the Court of Appeal.
Both convictions were upheld but one motorist’s sentence was reduced.
Two other non-fatal cases involving drug-driving have also been sent to the Court of Appeal. One conviction was quashed and the other has yet to be decided.
Ch Con James Vaughan, NPCC lead on forensics, said he could not remember a forensic science failure “of this magnitude”.
He said re-testing was taking longer than expected because there was a “chronic shortage” of scientific expertise and accredited laboratories, leading to delays in providing toxicology analysis in unrelated cases of sexual offence and rape.