There is a need for new, clinically relevant interpretation algorithms for genotypic and phenotypic resistance for ritonavir-boosted saquinavir (SQV/r) at the current approved dosage [1000/100 mg twice a day (bid)]. Clinical cut-off levels, which correlate baseline measures of resistance with HIV RNA responses in large cohorts or clinical trials, are the ideal reference for developing such algorithms. Cut-off levels previously developed for unboosted saquinavir may no longer apply, as the plasma drug levels with SQV/r are significantly higher and may be able partially to overcome protease inhibitor-resistant HIV. Clinical cut-off levels for SQV/r, assessed in several cohort studies and clinical trials, also suggest that multiple genotypic mutations are required for complete loss of virological response. For phenotypic analysis of resistance, saquinavir cut-off levels 10-11-fold higher than the wild-type IC50 have best distinguished responders from non-responders in cohort studies. Using Virtual Phenotype, a 12.3-fold upper cut-off level was determined from analysis of large cohort databases. These genotypic and phenotypic algorithms need to be validated in larger prospective studies.