A prospective, randomized, longitudinal, open label, parallel group clinical trial was designed to compare the proportions of failure to reach INR target 2.5-3.5 with either: small dose warfarin (<5 mg/day) or phenindione (<100 mg/day), in high-risk pregnant and non-pregnant patients with bileaflet mechanical heart valve prosthesis.

The patients were chosen from a main population of 3000 females in the child bearing period (18-40 years), with a bileaflet mechanical mitral (+ aortic) valve prosthesis, on regular follow-up at the outpatient clinic of Ain Shams University Hospitals. Patients were primarily screened for their wiliness to participate in the study by expressing their clear wish to become pregnant (Subpopulation A: 310 patients) or otherwise making a clear decision of not becoming pregnant by following a regular contraceptive measure (Subpopulation B: 660 patients).

Study groups A (pregnant) and B (non-pregnant) were then created by assessing the respective subpopulations for the following eligibility criteria (a) regular menstrual cycles and no history of congenital malformation, repeated abortions or still birth; (b) patient in NYHA class I or II c) regular INR records for the last 3 months, with the target INR being achieved with warfarin not exceeding 10 mg/day; d) patient accepting to participate in the study and; (d) getting pregnant for Group A patients, of course Patients were excluded from either groups if they refuse to participate in the study or to sign the informed consent. Pregnant patients were excluded from Group A if reporting their missed menstrual period or positive pregnancy test later than the 6th week of gestation. Patients were excluded from Group B whenever they get pregnant, of course.

A total of 100 patients were intended to be recruited in each group. Out of the 104 patients who met the inclusion criteria of Group A, 4 cases were excluded for reporting pregnancy late at the 9th and 12th week (2 cases; 1.9%) and for refusing to sign the informed consent (2 cases; 1.9%). Out of the 135 patients who met the inclusion criteria of Group B, as much as 35 patients (25.9%) were excluded for refusing to participate in the study.

Randomization and treatment allocation:

In order to become a formal Group A member, patients in subpopulation A were followed up until conception, with clear instructions to report after 3 days of missing a menstrual period and to confirm pregnancy by repeatedly performing pregnancy test every 3 days or until menstruating. Legible patients of both groups had to sign the informed consent before being randomized. We have used a simple 8 elements randomization table, with 6 elements being assigned to warfarin and 2 elements being assigned to phenindione, to create 2 separate randomization lists. Each list was used to randomize the recruited 100 patients of each group to either: low dose warfarin (<5 mg/day) or phenindione (<100 mg/day), in a 3:1 ratio; respectively. In patients assigned to the latter, phenindione replaced warfarin with a starting dose of 50 mg/day (25 mg /day in patients weighing <40 kg.), INR was repeated after 4 days and the dose was adjusted accordingly. All patients targeted an INR between 2.5 and 3.5.

Follow-up:

Patients were followed up till delivery (Group A) or for a whole 9 month's period (Group B). During that time, patients benefited from a monthly cardiac examination, bimonthly checks up for hepatorenal functions and complete blood picture as well as echocardiographic examination. In addition, Group A patients benefited from a monthly obstetrical examination and sonography. All patients were instructed to report any complication immediately and, in Group A, all prescribed drugs were revised by attending obstetrician. Prothrombin time (PT) was done on DADE BEHRING CA 1500 and INR was checked out the day of randomization, after 4 days, 2 weeks and then regularly on a monthly basis. Starting the second trimester (Group A) or the fourth month of follow-up (Group B), 100 mg aspirin was added to all cases. Patients presenting with an INR <20% target were subjected to clinical examination with special emphasis on dietary / bowel habits, recently received medication and INR was repeated 4 days after managing any suspected cause. Patients still failing to reach lower target as well as those initially presenting with an INR below 2 were shifted to the other oral anticoagulant. The starting dose of the latter was calculated on the basis that the anticoagulation effect of 1 mg of warfarin is roughly equivalent to that achieved with 10 mg of phenindione (unpublished experience). In between shifts, the INR gap was bridged with LMWH 1 mg/kg, given twice daily. Patients failing to reach lower INR target with the maximum allowed daily dose of both oral anticoagulants were excluded from the study. Fortnight before delivery, Group A patients were hospitalized and the oral anticoagulant was substituted with continuous infusion of UFH and a planned delivery at 36th week per vagina or by Caesarian section, as indicated.

Inclusion Criteria:

• Regular menstrual cycles and no history of congenital malformation, repeated abortions or still birth.
• Patients in NYHA class I or II c) regular INR records for the last 3 months, with the target INR being achieved with warfarin not exceeding 10 mg/day.
• Patient accepting to participate in the study
• Getting pregnant for Group A patients

Exclusion Criteria:

• Patients were excluded from either groups if they refuse to participate in the study or to sign the informed consent.
• Pregnant patients were excluded from Group A if reporting their missed menstrual period or positive pregnancy test later than the 6th week of gestation.
• Patients were excluded from Group B whenever they get pregnant.