A case-control study of 340 finishing pigs aged 10 to 25 weeks in 15 commercial Danish pig herds was carried out to investigate serum haptoglobin concentration as an objective marker of clinical signs of disease. Pigs with different clinical signs were matched to control pigs without clinical signs with respect to herd, pen, estimated weight and gender, and each pig was subjected to a standard clinical examination. In 86 of the case-control pairs, the rectal temperature was also recorded. There was a significantly higher mean haptoglobin concentration in the serum of lame pigs (P<0.0001), pigs with respiratory disease (P=0.0004), pigs with tail or ear bites (P=0.0004) and pigs with diarrhoea (P=0.02). Similarly, a higher mean rectal temperature was recorded in lame pigs (P<0.0001), pigs with respiratory disease (P=0.002) and pigs with tail or ear bites (P=0.0003). There was a significant but low correlation between rectal temperature and haptoglobin concentration in serum (P=0.003, r=0.20). The area under the receiver-operating-characteristic curve was between 0.67 and 0.78 for the different clinical signs. The maximum simultaneous levels of sensitivity (0.61 to 0.71) and specificity (0.61 to 0.77) of serum haptoglobin for the different clinical signs were obtained at a cut-off value of 1.1 mg/mi. At a cut-off value of 1.8 mg/ml, the sensitivity decreased to 0.31 to 0.60, and the specificity increased to 0.82 to 0.86. It was not possible to define a cut-off value which classified individual pigs according to their clinical signs.