And 70 per cent of family doctors who expressed an opinion using an official National Health Service (NHS) drug feedback form felt the drug wasn't effective.
Researchers at the UK's Drug Safety Research Unit and University of Portsmouth analysed prescribing data for 11,185 patients seen by NHS family doctors.
The data was gathered from official Green Form Questionnaires that ask doctors to record any significant events recorded in a patient's notes after prescribing newly marketed medicines.
Key findings included:
•65 per cent of doctors said their patient stopped taking Uprima because they felt it wasn't effective and 14 per cent said the patient didn't request a further prescription. (68 per cent of GPs responded to the question).
•Further analysis showed that 59 per cent of patients who didn't find it effective stopped taking Uprima after a month and a further 23 had joined them by month two. (Timescale information was recorded for 79 per cent of patients who judged the drug ineffective).
•70 per cent of doctors said they didn't think the drug was effective and 30 per cent said it was. (65 per cent of GPs responded to the question)
•28.5 per cent of the total patient sample had a history of diabetes, 18.8 per cent had heart disease and 37.2 per cent had tried Viagra. 7.5 per cent were receiving both Uprima and nitrate therapy for angina.
•The most common adverse drug reactions were headache and nausea, in line with the clinical trials and as listed in the Summary of Product Characteristics by the manufacturers.
•The average age of the subjects was 61 and all but eight were male. Three of the women were prescribed Uprima for decreased libido, pain and sexual arousal problems. Prescribing information was not provided for the other five.
•Most patients (99 per cent) received the manufacturer's recommended dose of 2-3mg, with the remainder receiving between 4-12mg. (Starting dose information was completed on 79 per cent of forms).
•99 deaths (just under one per cent of the total sample) were recorded on the forms. The most frequently reported cause of death was heart attack (27 patients), followed by heart disease (nine) and lung cancer (five). In 21 cases the cause of death could not be established by the researchers.
The 11,185 forms covered patients who were first prescribed Uprima between October 2001 - three months after the drug was launched in the UK - and December 2002. Just over 21,000 Questionnaires were issued and 57 per cent of GPs responded.
At least six months after the first prescription for Uprima was issued for each patient, their family doctor was sent a Green Form Questionnaire that requested details about the patient's age and sex and their use of Uprima.
They were also asked to record any significant events that had occurred since the patient was prescribed Uprima, such as suspected drug reactions, unexpected deterioration or improvement in the patient's condition, referral to a specialist or any clinically important changes in laboratory tests.
The form also included questions on diseases commonly associated with erectile dysfunction - such as diabetes and heart disease - together with previous use of similar drugs and co-prescribing of medication for angina.
"Prescription-Event Monitoring provides surveillance on a national scale after new drugs are launched" explains lead researcher Professor Saad Shakir.
"Because family doctors are not approached before the decision to treat a patient has been made, they are not subject to detailed inclusion or exclusion criteria. The decision to prescribe is made purely as a result of their clinical knowledge of the patient and the drug involved.
"The result is real world clinical data that provides information of illness and death in patients treated with newly marketed drugs - a valuable tool in drug safety assessment."