OBJECTIVES The sweat test remains the current diagnostic gold standard for CF disease. Many CF testing centres have switched from the Gibson and Cooke to the Macroduct. Since the validity and sensitivity of Macroduct has not been tested in patients with intermediate sweat chloride concentrations, we compared both methods simultaneously including subjects expected to have intermediate results. DESIGN AND METHODS We prospectively evaluated controls, obligate heterozygotes, patients with CF and with an uncertain diagnosis of CF (congenital absence of the vas deferens, pancreatitis and sinopulmonary disease). RESULTS We assessed 82 subjects (3.7-60.1 years); 14 healthy controls, 7 obligate heterozygotes, 20 CF (15 pancreatic insufficient, 5 pancreatic sufficient), and 41 with unproven diagnosis. Mean test difference was close to 0 (95% CI+/-20 mmol/L) and test values were highly correlated (r=0.93, p < or =0.0001). Discrepancies between the two testing methods occurred in 22% of subjects. CONCLUSION Sweat chloride measured by Macroduct highly correlates with Gibson and Cooke for concentrations in all ranges, including the intermediate range. This study reveals the limitations of sweat testing for excluding a diagnosis of CF since 38% of subjects had intermediate range results.