Pompe disease (PD) is a progressive neuromuscular disorder caused by a lysosomal acid α-glucosidase (GAA) deficiency. Enzymatic replacement therapy is available, but early diagnosis by newborn screening (NBS) is essential for early treatment and better outcomes, especially with more severe forms. We present results from 7 years of NBS for PD and the management of infantile-onset (IOPD) and late-onset (LOPD) patients, during which we sought candidate predictive parameters of phenotype severity at baseline and during follow-up. We used a tandem mass spectrometry assay for α-glucosidase activity to screen 206,741 newborns and identified 39 positive neonates (0.019%). Eleven had two pathogenic variants of the GAA gene (3 IOPD, 8 LOPD); six carried variants of uncertain significance (VUS). IOPD patients were treated promptly and had good outcomes. LOPD and infants with VUS were followed; all were asymptomatic at the last visit (mean age 3.4 years, range 0.5–5.5). Urinary glucose tetrasaccharide was a useful and biomarker for rapidly differentiating IOPD from LOPD and monitoring response to therapy during follow-up. Our study, the largest reported to date in Europe, presents data from longstanding NBS for PD, revealing an incidence in North East Italy of 1/18,795 (IOPD 1/68,914; LOPD 1/25,843), and the absence of mortality in IOPD treated from birth. In LOPD, rigorous long-term follow-up is needed to evaluate the best time to start therapy. The high pseudodeficiency frequency, ethical issues with early LOPD diagnosis, and difficulty predicting phenotypes based on biochemical parameters and genotypes, especially in LOPD, need further study.