NHS Employers' guidance for quality framework clinical assessors says: 'Excessive clustering at and just under 150/90 and 145/84mmHg for people with diabetes' could be evidence of 'inappropriate recording', or fraud.
Large numbers of results recorded on the same day could also be cause for concern, it says.
GPC prescribing subcommittee member Dr Chaand Nagpaul said: 'If you set a target within a time frame, you will get clusters.
'There is no evidence of widespread inappropriate recording of data by practices. Many recall patients to have checks and run extra clinics. That may lead to a cluster of results on the same day.'
He added that the recording of many parameters bypassed GPs. 'All results for bloods are automatically entered through e-links. Large amounts of data relate to that, including diabetes, kidney conditions and cholesterol,' said Dr Nagpaul.
Primary Care Diabetes Society chairman Dr Colin Kenny said there was no basis for such a heavy handed approach: 'The universality of the quality framework results across the UK makes it unlikely there's been a vast amount of fraud going on.'
He accepted that BP measurement was open to operator error, but added: 'It is not unusual for results to be clustered around one date as many practices have nurse-run clinics where patients are recalled to have their BP measured.'
Bradford GP and North Bradford PCT diabetes lead Dr Brian Karet said: 'You get huge variables with BP measures, day-to-day variability, minute-to-minute variability, inter-operator variability.'
He said it was incredible to suggest that fraud was one possibility for hitting the quality framework targets.
NHS Employers lead negotiator Dr Adrian Jacobs said PCTs had a duty to investigate where they suspected fraud, but he admitted that procedures in some PCTs fell short of best practice.