Valor del tacto en el diagnóstico de fiebre y de la palidez en el diagnóstico de anemia. Estudio observacional en adultos

Estimation of fever by touch and of pallor as a sign of anemia are common practices during physical examination. The diagnostic value of these procedures has been extensively studied in children, but there are fewer studies in adults. Objetive: To evaluate the diagnostic value of physician's estimation of fever and anemia by touch and pallor, respectivelly. Patients and methods: The study included 600 patients (mean age 66 years, range 14-94 years, 58% males) admitted to a hospital (n=300) or seen at its Emergency Department (n=300). Two physicians independently estimated the presence of fever by touch, as well as the presence of anemia by pallor (in the skin, nail or conjunctiva). Axillary temperature and peripheral blood hemoglobin were employed as standards for comparison. Results: Specificity of touch was higher than 90%, both for the diagnosis of axillary temperatures 37.2oC and 38 oC, with a lower sensiti vity. Interobserver concordance was fair (Kappa index, 0.47). Negative predictive value of touch was also high (99%) for the observed prevalen ce of fever (5%), with a very much lower positive predictive value. Similarly, specificity of pallor was high, particularly for hemoglobin values lower than 10 g/dL, with a substantially lower sensitivity. The best diagnostic values were obtained with conjunctival pallor. The Kappa index of concordance for the various types of pallor (skin, nail or conjunctiva) ranged between 0,39 and 0,44. Negative predictive value of pallor was high (98%) for the observed prevalence (9%) of significant anemia (hemoglobin <10 g/dL), with a lower positive predictive value. Conclusions: In adult patients of a hospital, physi-cian's estimation of fever by touch and anemia by pallor have a relatively low diagnostic value. Nevertheless, the estimation that the patient is afebrile or not pale make very improbable the presence of fever and significant anemia, respectively.