Pendred's syndrome (PDS) is an autosomal-recessive disorder characterized by sensorineural hearing loss and goiter. PDS is caused by mutations of the SLC26A4 gene encoding pendrin, a transmembrane exchanger of Cl(-), I(-) and HCO(3)(-), which is expressed in the thyroid and inner ear. SLC26A4 mutations can also be associated with non-syndromic deafness, DFNB4. The goal of our study was to define the identities and frequencies of SLC26A4 mutations in 563 large, consanguineous Pakistani families segregating severe-to-profound recessive deafness. Sequence analyses of SLC26A4 in 46 unreported families segregating deafness linked to DFNB4/PDS revealed 16 probable pathogenic variants, 8 of which are novel. The novel variants include three missense substitutions (p.R24L, p.G139V and p.V231M), two splice site mutations (c.304+2T>C and c.1341+3A>C), one frameshift (p.C565MfsX8) and two different genomic deletions affecting exons 1-2 and 11-18. Each of six pathogenic variants (p.V239D, p.Q446R, p.S90L, p.Y556C, p.R24L and p.K715N) was found in more than one family and haplotype analyses suggest that they are founder mutations. Combined with earlier reported data, SLC26A4 mutations were identified in 56 (7.2%; 95% CI: 5.6-9.2%) of 775 families. Therefore, SLC26A4 mutations are the most common known cause of genetic deafness in this population. As p.V239D (30%), p.S90L (18%) and p.Q446R (18%) account for approximately two-third of the mutant alleles of SLC26A4, hierarchical strategies for mutation detection would be feasible and cost-efficient genetic tests for DFNB4 deafness and PDS in Pakistanis.