Drug/metabolite ratios (MRs) are used as in vivo markers of enzyme activity. The ratios are potentially confounded by the renal clearance of the drug (urine-based MRs) or metabolite (plasma-based MRs). The authors have investigated the relative sensitivity of urinary MR of 3 in vivo probe substrates of CYP2D6 debrisoquine (DB), dextromethorphan (DM), and metoprolol (MP) to changes in urine pH. Three groups of healthy volunteers each comprising 12 individuals were given DB (10 mg), DM (25 mg), or MP (100 mg) on 3 occasions. In 1 study arm, urine was acidified by the oral intake of ammonium chloride; in another, it was alkalinized by intake of sodium bicarbonate; and in the third, urine pH was uncontrolled. Urinary MP/alpha-hydroxy-MP, DM/dextrorphan, and DB/4-hydroxy-DB ratios were calculated. The mean(geo) MR for DB was not significantly different in any of the study arms, whereas those for MP and DM were significantly different under acidified and alkalinized urine conditions compared to uncontrolled urine pH (P < .01) and were correlated with urine pH (P < .001). Without control of urine pH, in vivo estimates of CYP2D6 metabolic activity are likely to be less precise using DM or MP as probe substrates compared to DB. Although this is unlikely to cause any problem in distinguishing the large functional differences in CYP2D6 in poor metabolizer (PM) and extensive metabolizer (EM) phenotypes, this may contribute to difficulties in differentiating in vivo metabolic activity among allelic variants within the overall CYP2D6 EM phenotype using MP or DM. However, because DB is not available in many countries (eg, United States), alternative in vivo markers of CYP2D6 with low sensitivity to urine pH should be sought.