Table 2 Characteristics of included retrospective cohort studies

Dicker 1988 [15](Israel)

59 IDDM women attended a pre-conception clinic compared to 35 pregnant women who did not attend

PCC included: insulin and dietary glycemic control, advice on contraception and screening for diabetes complications

PCC group had significantly lower HA1C at the first trimester (p < 0.001) and significantly lower rate of spontaneous abortion (p < 0.001) compared to the NPCC group.

Low (Clear description of participants and intervention, noted confounding factors and well presented results. There was significant difference between the two groups in the diabetes complications before intervention)

Dunne 1999 [16](UK)

47 women with IDDM 12 of them attended preconception care clinic and 35 women did not.

PCC included assessment of diabetes complications and glycemic control

The PCC group had significantly lower level of HA1C level compared to the NPCC group (p < 0.008). There were no congenital malformations in both groups. The cesarean section rate, the macrosomia rate and the small for gestation age were similar between the two groups

Medium (Due to the audit nature of the report there is no clear description of the intervention, some important confounders were not addressed such as White's classification and the outcome assessment was not blinded )

Damm 1989 [14](Denmark)

197 attended PCC and 61 didn't attend

PCC included: contraception and glycemic control.

The rate of congenital malformations was significantly lower in the PPC group 1.0% than the NPPC group 8.2%, (p < 0.01). No significant difference in the level of HA1C during the first trimester between the two group

High (unclear description of the participants, the intervention and the outcome, the data of the preconception care were a subset of from different periods of the study)

Goldman 1986 [21](Israel)

44 women with type I diabetes attended the preconception clinic and 31 women with type I diabetes did not attend

PCC included assessment of diabetic complications, Contraception advice, Glycemic control and dietary advice

The NPCC group had significantly shorter duration of pregnancy (p < 0.05) significantly heavier mean birth weight (p < 0.05) than the PCC group. The two groups were similar in neonatal hypoglycemia, hypocalcaemia and respiratory distress syndrome

Low (Clear description of participants and intervention, noted confounding factors and well presented results. There was significant difference between the two groups in the diabetes complications before intervention)

Fuhrmann 1986 [19]& 1984 [18]& 1983 [17](Germany)

620 pregnant women with insulin dependent diabetes,183 received pre-pregnancy care 437 women did not

PCC included: short hospitalization every 3 month until conception, education, self monitoring of blood glucose, assessment and treatment of diabetes complications and glycemic control

PCC group had significantly lower rate of congenital malformations 1.1% compared to the NPCC group 7.0% (p < 0.01)

Medium (Well described intervention, no blinding for the outcome, no description of the possible confounding factors)

Rowe 1987 [27](UK)

21 IDDM 14 received preconception care and 7 did not

PCC included Glycemic control, counseling and blood glucose self monitoring

The PCC group had significantly better initial HA1C level (p < 0.0001), and lower mean birth weight (p < 0.05)

High (Unclear description of the participants, no description of possible confounding factors, no blinding in assessment of the outcome, small group, high target of HbA1C 5-9%)

Steel 1990 [28]& 1982 [31](UK- Scotland)

143 IDDM women attended the preconception care clinic and 96 IDDM women did not attend

PCC included: education, glycemic controlled and contraception

PCC group had lower initial HbA1C as compared to NPCC group (p < 0.0001) and lower rate of congenital mal formations (p < 0. 005) , maternal hypoglycemia was significantly common in the PCC group than the NPCC (p < 0. 001)

Medium (Good description of interventions, contamination of the control who might know about the usefulness of the and the outcome assessment was not blinded )