Introduction
The Rénarub network, set up in 1976, is the principle source of information on the epidemiology of rubella in France (1, 2). Its objective is to collate, at national level, cases of rubella occurring during pregnancy and congenital rubella, and more widely, to evaluate the impact of vaccination policy and prevention measures put in place in France with the aim of eliminating congenital rubella. The full version of this article is available on the website of the Institut de Veille Sanitaire, including the new case definition, more detailed results (http://www.invs.sante.fr/publications/2003/rubeole_2001/rubbeh_2001.pdf), and a list of laboratories participating in the network (http://www.invs.sante.fr/publications/2003/rubeole_2001/reseau_renarub.htm).

Material and methods
The Rénarub network
The Rénarub network unites the clinical diagnostic laboratories carrying out anti-rubella IgM serology. Information is gathered at two levels:

Inclusion criteria and case definition
Every pregnant woman for whom the first positive IgM serology is identified in 2001 and/or every newborn or fetus with a diagnosis of rubella infection. Cases are notified to the Institut de Veille Sanitaire (InVS), which coordinates this surveillance network.

In order to better take into account the difficulty of confirming the laboratory diagnosis of rubella infection in pregnancy, new case definitions, in particular IgG avidity, were used from 2001 (see full case definition on the InVS website, http://www.invs.sante.fr/publications/2003/rubeole_2001/rubbeh_2001.pdf)(3)). Confirmed and probable cases are included for analysis and possible cases are excluded. When it is unclear whether primary infection or re-infection has occurred, the case is defined as an 'infection'. Women born outside France but whose infection was acquired in France, and whose antenatal care was provided in France are included in the analysis.

Results
Participation of laboratories and doctors
In 2001, the participation level of laboratories that were invited within mainland France was almost 100% (271/272). Compared with 2000, two new laboratories joined the network and one withdrew because it was no longer carrying out anti-rubella IgM serology. The participation level among doctors was also very high (99%).

Cases counted in 2001 and incidence
In 2001, 162 rubella infections were reported by the laboratories. Only 38 met the case definitions. Among these 38 cases, 27 were confirmed primary infections (2 of which occurred around the time of conception, with maternal infection becoming apparent with a history of contact with a rubella case or a rash), 1 confirmed infection, 6 probable primary infections and 4 probable infections. Among these 38 women, 6 gave birth to a child with congenital rubella syndrome (CRS), 2 had spontaneous abortion, and 8 underwent medical termination of pregnancy (table 1). In the case of one of the spontaneous abortions and one of the medical terminations, the fetuses had probable CRS. The other 124 cases were excluded from the analysis for the following reasons: the woman was not pregnant (57), rubella infection was excluded (5), possible infections (5), missing information (9), lost to follow up (8), past rubella infection (29), post vaccination immunity (8), infection prior to conception (1), and case from 2000 (1). Thus the annual incidence of rubella infection in 2001 was 3.6 (confirmed cases) or 4.9 per 100 000 live births (if we add the probable cases), and for CRS, 0.78/100 000 live births (Births in 2000 - provisional data, INSEE, http://www.insee.fr).

Table 1. Clinical status of newborns and fetuses by outcome of pregnancy [primary infections (n = 27 confirmed and 6 probable) and infections (n=1 confirmed and 4 probable)], mainland France, 2001, Rénarub network

Outcome of pregnancy	Presence of malformation			Absence of malformation		Unknown	Total
	CRS	Probable infection	Confirmed infection	Absence of infection	Infection status unknown
Birth	6	-	11	8	3	-	28
Termination of pregnancy	-	1	3	-	-	4	8
Spontaneous abortion	-	1	-	-	-	1	2
Total	6	2	14	8	3	5	38

Characteristics of infected pregnant women
The average age of the 38 women was 23.3 years (range=17 to 38 years), while the average age of maternity in the general population in France is 29.4 years (Source: INSEE 2000). The proportion of women under 20 was 21%, an infection rate of 42.1/100 000 live births (table 2). The country of birth is known for 29 women. For 21 cases, this was mainland France. Among the women for whom previous obstetric history is known (37/38), 10 (27%) had had at least one previous pregnancy. Among the 31 women whose vaccination status was known, none had been vaccinated.

Cases preventable by post partum vaccination
Ten women had been pregnant at least once before and up to three times. At least 6 of them had been living in France during a previous pregnancy. Of these 6 women, 1 woman's serology was known to have been previously negative, and 5 had no known previous negative serology. None of these 6 women had been vaccinated. Had they been vaccinated immediately follow of a previous pregnancy, 1 CRS and one miscarriage of a fetus with probable CRS could have been prevented.

Table 2. Distribution by age of pregnant women infected with rubella, mainland France, 2001, Rénarub Network (n = 38)

Age group	Number of cases	%	Rate of infection/100 000 live births
15-19 years	8	21.1	42.1
20-24 years	16	52.1	10.2
25-29 years	9	23.7	3.4
30-34 years	4	10.5	2.0
35-39 years	1	2.6	5.1
Total	38	100.0	5.0

Discussion
Following a study carried out in 2002 by a group of clinical diagnostic laboratories carrying out anti-rubella IgM serology, Rénarub's size has more than doubled, from 133 to 272 laboratories, making our network almost exhaustive at the present date. The participation of clinicians contacted to document the cases notified by the laboratories has also greatly improved over the years (62% in 1996, 78% in 1997, 92% in 1998, 94% in 1999, and 99% in 2000 and 2001).

One of the consequences of the new case definition, introducing the category of cases of probable infection occurring during pregnancy, has been that infections previously considered as confirmed are now categorised as probable. Twenty six percent of cases which would have been categorised as confirmed under the previous definition were categorised as probable in 2001. Trend analysis was carried out in 2001 for confirmed and probable cases, corresponding to cases included in the old case definition.

Figure 1. Incidence rate of rubella infections in pregnant women and congenital rubella syndrome, mainland France, 1978-2001, Rénarub network.

The incidence of rubella infection during pregnancy measured through Rénarub (figure 1) could be an underestimation since asymptomatic or atypical rubella is common, and not always diagnosed in pregnant women.