Nine patients with metastatic or recurrent superficial tumors of varying size and histology were administered 100 g oral glucose to investigate whether hyperglycemia can selectively lower tumor pH. pH was measured by a 21 ga modified glass needle electrode inserted through an 18 ga open-ended Angiocath. Serum glucose was monitored every 7.5 min by finger stick and a blood glucose analyzer. Tumor pH was measured over 50-80 min concomitantly with determination of blood glucose. In five nondiabetic patients (eight measurement points) tumor pH decreased 0.05-0.5 units from a pre-glucose range of 6.8-7.4 (7.14 +/- 0.08) to 6.4-7.3 (6.90 +/- 0.10) as blood glucose increased from a baseline of 80-120 mg/dl to 165-215 mg/dl. There was considerable heterogeneity from patient to patient regardless whether blood glucose increased to a peak at 40-60 min post-ingestion and then decreased, or whether it remained elevated up to the end of the 80 min observation period. In general, tumor pH decreased as blood glucose increased and then continued to fall throughout the period of observation. In one patient, tumor pH did not change although blood glucose increased to 175 mg/dl. Normal tissue pH was 7.36 +/- 0.02 when determined on four occasions in three patients (subcutaneous and intramuscular sites), and was unaffected by glucose administration. As a further control for tumor tissue and pH probe stability, pH probes in two patients were left in place for 30 min before glucose ingestion. The tumor pH was stable for the entire interval. Interestingly, three patients had an abnormal glucose response: two of those patients (one patient on two separate occasions) had an increase in blood glucose to 230-260 mg/dl in 40-60 min and pH actually increased 0.1-0.3 units. The third patient had a transient increase in blood glucose to 290 mg/dl along with a corresponding increase and subsequent decrease in tumor pH. In summary, whether glucose was given pre- or post-hyperthermia, independently of position in the tumor, and independently of whether pH increased or decreased, the slope of the curve of pH = f(time) was similar in a given patient tumor on all measurement occasions. These preliminary results suggest that hyperglycemia may be useful in non-diabetic patients, and perhaps in diabetic patients given insulin, to selectively reduce tumor pH and sensitize tumors to hyperthermia.