Staying on an anti-HIV combo until it fails increases the risk of cross-resistance to future therapies, so early indicators of drug failure are urgently needed. Recent findings from a study by Agouron (maker of nelfinavir [Viracept]) found three variables that predicted who would do well—both initially and over the long haul—on a nelfinavir/AZT/3TC combination: baseline viral load, the change in virus at four weeks and the blood concentration of nelfinavir (but not AZT; 3TC data is still being analyzed). While each had significant predictive value when used alone, their combined power was more striking: 93 percent of people in whom all three predictors were positive (a baseline viral load under 162,000, at least a 2.35-log drop in viral load, and a nefinavir drug level of at least 1.9 mg/liter two hours after dosing) responded well to the drug therapy. People with no positive predictors responded very poorly to the meds; those with one or two positives fell somewhere in between. If future trials confirm these preliminary findings (not yet ready for prime time), PWAs may have a powerful new tool to help to design treatment strategies. By identifying early those at high risk of combo failure, changes could be quickly implemented, greatly improving the chances of long-term success.