1) Clinical assessment: in moderate to severe anemia (11,13) -i) significant pallor of eyelids (difficult in children due to frequent conjunctivitis), tongue, nail beds and palms (pale palm creases suggest severe anemia; ii) fatigue, low exercise capacity (mild anemia can produce decreased exercise tolerance); iii) fissures at the corner of the mouth suggest anemia; iv) nails, show pallor, flatness, softness to feel and later become spoon shaped (koilonychia). It is rare in children < 6 yr of age, as hemoglobin is sacrificed to maintain tissue growth; v) children develop irritability, pica for ice/mud/coal/substances containing lead; vi) findings of congestive failure indicate severe anemia, hemoglobin below 5.0g/dl; vii) iron deficiency affects mental functions (10) i.e. attention span, alertness and learning, and viii) deficiencies of vitamin B12 and folic acid are associated with psycho-neurological changes and pigmentation of mucous membranes and distal parts of the body. These clinical features are of diagnostic assistance for the individual patient with severe anemia to seek advice from a medical expert. However, clinical features fail to assess degree of anemia and therapeutic response. Thus hemoglobin estimation is the criterion. Anemia (as measured by hemoglobin or hematocrit estimations) is not specific of iron deficiency; however it is the commonest deficiency in India, thus a presumptive diagnosis can be made. Normal hemoglobin and hematocrit distribution varies with age, gender, at different stages of pregnancy and with altitude and smoking. Given the wide range of hemoglobin concentrations in the population as a whole, an individual can have a substantial decrease in hemoglobin with the value still in the normal range.

2) Hemoglobin and Hematocrit: can be used to assess anemia in community where prevalence is of public health significance. The problem arises in capillary blood collection, the technician should be very careful. To obtain the best possible skin puncture sample, it is important to warm the extremity in order to facilitate a free flow of blood and to avoid any squeezing of the finger. In community surveys discrepancy up to 0.5 g/dl between capillary and venous blood is acceptable. Collections of venous blood give accuracy and allow study of other parameters, if needed.

Hemoglobin estimation (14-19)

• a) Copper sulphate method - used in blood banks, discriminates donor's hemoglobin content > 10g/dl.
  
  
• b) Sahli's method (acid-hematin) - developed color is unstable, begins to fade almost immediately after it reaches its peak. It does not estimate fetal hemoglobin in infants.
  
  
• c) Oxyhemoglobin method - Though simplest and quickest, does not have a stable HbO2 standard. It is also not satisfactory in presence of methemoglobin, sulphemoglobin, etc.
  
  
• d) Alkali-hematin method - It gives true estimate of total hemoglobin, except the resistant hemoglobins i. e. fetal and Bart's hemoglobins require heating in a water bath for 4 min. Method is cumbersome and less accurate than the oxyhemoglobin and cyanmethemoglobin methods.
  
  
• e) Cyanmethemoglobin method - is the best time tested method for estimating the hemoglobin concentration quantitatively. Blood is diluted in a solution containing potassium cyanide and potassium ferricyanide. Hemoglobin, in blood is found in variety of forms including oxyhemoglobin, carboxyhemoglobin, methemoglobin, fetal hemoglobin and other minor components are converted to cyanmethemoglobin (HiCN). Sulphemoglobin is not converted, but is rarely present in significant amount. Standard is internationally developed and is stable for several yr. In field surveys blood is delivered on to No 1 Whatman filter paper (1.5 x 1.5 cm square), dried and transported. Drabkin's solution in screw-capped tubes can be taken to the field; once the blood is mixed some loss of fluid will not alter the result. This method is reproducible and suited for field hemoglobin estimation. It is also possible to use battery operated colorimeters for hemoglobin estimation in field. Main errors in measurement arise from i) dilution, ii) sample turbidity (incomplete lysis of blood cells), lipids, and raised leucocyte count. In automated method the precision is less than 1%.
  
  
• f) HemoCue system- hemoglobin estimations: India had large field survey experience in NFHS II (1998-1999) (18), it is suited for rapid field surveys, no dilution required, equipment is portable; results are available within 45 sec and read directly. Comparison with cyanmethemoglobin method showed variability at different hemoglobin levels. Thus, no simple conversion factor is applicable. It is not possible to compare data on prevalence and severity of anemia as earlier studies used cyanmethemoglobin method. Taking cyanmethemoglobin as gold standard, the sensitivity of hemoCue method was 0.75 and specificity 1.0, corresponding hemoglobin values with both the methods fell within ± 2 SD with correlation coefficient of r = 0.922. These two methods showed the magnitude of difference as - 1.19 g/dl (C I: - 1.40- 0.98), thus suggesting an overestimate of hemoCue values by 10-15 % (correction factor being 0.389 + 0.831 Hb- hemoCue) (20). There is need to validate hemoCue method in 2-3 laboratories of our country, as it has field applicability, and overcomes dilution, collection, transportation and sample turbidity, etc. This method is costlier as compared to cyanmethemoglobin method (for equipment as well as cuvettes).
  
  
• g) Hematocrit (microhematocrit)/ packed cell volume : It is an acceptable and recommended method for anemia determination. Widely used for simplicity and availability. Equipment can be transported, needs constant voltage for operation. Blood is collected in anticoagulant treated capillary tube and spun in a portable microhematocrit centrifuge; plasma trapping is around 1-3% (15). Trapping of plasma is more if red cells are abnormal i.e. sickle cell, microcytosis, macrocytosis and spherocytes. Individuals with shock or dehydration may have risen or normal hematocrit due to hemoconcentration. Automated analyzer do not depend on centrifugation technique but calculate by direct measurement of RBC number and red cell volume (precision cv <1%).