Predictive factors for spontaneous closure of atrial septal defects diagnosed in the first 3 months of life.

OBJECTIVES

To establish the rate of spontaneous closure of atrial septal defects diagnosed before age 3 months, 101 infants (mean age 26 days) with an interatrial shunt confirmed by Doppler echocardiography were followed up for an average of 265 +/- 190 days.

BACKGROUND

Even if interatrial shunts in the newborn are frequently encountered, little is known about their natural history.

METHODS

Defect diameter on two-dimensional echocardiography and width of color flow jet were measured in the subcostal view. Right and left ventricular diameters and atrial septal curvature were also studied. Kaplan-Meier curves were obtained to predict age of spontaneous closure in relation to initial defect diameter.

RESULTS

There was no significant correlation between the diameter of the atrial septal defect and right ventricular/left ventricular ratio or type of septal curvature (vertical or concave toward the left atrium). The classic predominance of girls over boys was observed only for defects > 5 mm. An overall rate of spontaneous closure of 87% was observed. Frequency and timing of closure were inversely correlated to atrial septal defect diameter: closure occurred in 100% (32 of 32) of defects in group 1 (diameter < 3 mm), 87% of defects (39 of 45) in group 2 (diameter 3 to 5 mm), 80% of defects (16 of 20) in group 3 (diameter 5 to 8 mm). Spontaneous closure did not occur in four patients of group 4 (defect > or = 8 mm) during an average follow-up interval of 417 days (range 294 to 597 days).

CONCLUSIONS

These results suggest that infants with an atrial septal defect < 3 mm need not be followed up as 100% of these defects will be closed by age 18 months; those with a defect 3 to 5 or 5 to 8 mm should be evaluated by the end of the 12th and the 15th month, respectively, when > 80% of these defects will be closed. An atrial septal defect with a diameter > or = 8 mm may have little chance of closing spontaneously and the possibility of surgical correction should be considered. Defects < 3 mm probably do not constitute a cardiac malformation in light of their natural evolution and gender distribution.