[Diagnostic value of tegument pallor in anemia in pregnant women in Benin]

Sante Publique. 2004 Mar;16(1):123-32. doi: 10.3917/spub.041.0123.
[Article in French]


Anemia in pregnant women is a common and severe public health problem encountered in developing countries. Its detection is most often based upon clinical examination and observation of conjunctive pallor; however, little is known or documented on the performance indicators of this clinical method in Africa. This work aims assess the value of performance indicators in detecting moderate and severe anemia by means of tegument pallor examination in a maternity ward in rural Benin. To this end, 480 pregnant women were recruited at random during their prenatal visits to a rural maternity clinic. Two independent observers had noted conjunctive or other forms of teguments in all of the women selected, and a concentration of hemoglobin was confirmed by a blood test. An analysis of the results indicated that 67.4% of pregnant women were biologically anemic (Hb < 11.0 g/dl) while only 49% among them were clinically anemic. The sensitivity of the clinical observation method for conjunctive pallor is 87.5% for severe anemia and 74% for moderate anemia. Its positive predictive value is 62.5% for moderate anemia and very weak for light cases of anemia, at only 3%. The accuracy of this clinical method varies from 51-71%, when the likelihood is determined as 1.8 to 2.3. Its reproducibility appears equally weak and clinical observer variability seems to be poor to fair agreement at k = 0.34 (p = 0.00). For practical reasons, the authors suggest the strengthening of the teaching of this method to obstetrics' care providers and medical students in training in order to improve its more effective use and practice in rural maternity units without laboratories.

Publication types

  • Evaluation Study

MeSH terms

  • Adult
  • Anemia / complications
  • Anemia / diagnosis*
  • Benin
  • Diagnosis, Differential
  • Female
  • Humans
  • Pallor / etiology*
  • Predictive Value of Tests
  • Pregnancy
  • Pregnancy Complications / diagnosis*
  • Reproducibility of Results
  • Rural Population
  • Sensitivity and Specificity
  • Severity of Illness Index