HEMOGLOBIN CONCENTRATION—NORMATIVE DEFINITION COMPARED WITH PHYSIOLOGIC CONSEQUENCE

Defining the normal range of hemoglobin values
The hemoglobin cutoff value commonly used to define anemia is based on the normality definition often applied to other clinical laboratory tests and in anthropometry (10). For anemia, age- and sex-specific hemoglobin distributions are based on healthy reference samples (samples from persons with disease or nutritional deficiencies identified by other laboratory criteria are excluded) (11, 12). By convention, the central 95% of the reference hemoglobin distribution was considered the normal range. The 2.5th percentile hemoglobin value (-2 SDs from the mean) was the cutoff point for low hemoglobin or anemia. Conversely, the 97.5th percentile (+2 SDs from the mean) of the same distribution was the cutoff for high hemoglobin.

Because iron deficiency is often a major cause of anemia in many parts of the world, anemia screening is commonly used as a substitute for screening for iron deficiency. However, the probability of iron deficiency in individuals who are anemic (positive predictive value) depends largely on the actual prevalence of iron deficiency in a specific population. Additionally, because many possible causes of anemia exist, such statistics-based anemia criteria provide only a guideline for establishing the probability of specific health and nutritional causes of anemia for a given population. In essence, the hemoglobin-based definition does not provide sufficient information on the meaning of, or reason for, either anemia or high hemoglobin concentrations. Usually, the further a hemoglobin value is from the central tendency of the distribution, the greater the likelihood of finding pathologic reasons to explain the abnormal value. Because of substantial normal variations in hemoglobin distribution across age, between sexes and races, and at different stages of pregnancy, the cutoff for low or high hemoglobin concentrations should be specific to sex and life cycle (11, 12). The use of inappropriate evaluative hemoglobin criteria during pregnancy can result in misinterpretation of the relation between anemia and resulting health outcomes. An example of such misinterpretation is the disregard of normal hemoglobin concentration variations related to plasma volume changes during pregnancy, which can result in a striking association between preterm births and anemia (13, 14).

Physiologic significance of anemia
Mild anemia is routinely defined as a hemoglobin value within 10 g/L of the anemia cutoff value. The World Health Organization recommends that severe anemia be defined as a hemoglobin concentration <70 href="http://www.ajcn.org/cgi/content/full/72/1/272S#R15">15). Hemoglobin concentrations below that of the mild anemia concentration and >70 g/L can be regarded as indicating moderate anemia. Some investigators have defined very severe anemia as a hemoglobin concentration <50> Hemoglobin concentrations <50> the risk of maternal and fetal mortality because of the effects of hypoxia and anemia on the cardiovascular system, which is known as high-output heart failure (16). Medical evidence shows that very severe anemia is a direct cause of maternal and child mortality. In South Asia and Africa, where severe maternal anemia is common, classifying very severe anemia as a major cause of maternal mortality along with eclampsia, obstructed birth, hemorrhage, and sepsis is appropriate. Ross and Thomas (17) estimated that =" v:shapes="_x0000_i1025" border="0" height="10" width="8">20% of maternal mortality can be attributed to severe anemia. In the less severe range, however, the evidence that anemia is a direct cause of poor reproductive outcomes is not clear. Epidemiologic studies that showed an association between maternal anemia and increased risk of poor birth outcomes did not establish a causal relation (evidence of risk only) (6–9). It is possible that a common factor can cause both anemia and poor birth outcomes.

Although oxygen carrying capacity is proportional to the circulating hemoglobin concentration, an individual with chronic anemia develops a compensatory mechanism to improve oxygen unloading to tissue from hemoglobin during the resting state. This compensatory mechanism can maintain adequate tissue oxygen delivery down to a hemoglobin concentration of 70–80 g/L (18, 19). In an exercise state, however, any loss of hemoglobin or red blood cell mass can be detected as loss in work capacity, even within a hemoglobin range of 120–130 g/L (20). From a physiologic point of view, the evidence is clear that moderate anemia is undesirable, whatever the cause. If the cause of significant anemia is iron deficiency (evidence of deficiency), prevention and correction of iron deficiency anemia are indicated. From the perspective of reproductive health outcomes, however, the evidence is not clear that anemia or iron deficiency are direct risk factors (2). Perhaps a concern more important than the health consequence of anemia is the cause of the anemia. Some major causes of anemia have many other damaging effects or health consequences beyond anemia. For example, malaria is well known to cause severe anemia in many tropical areas, particularly among primigravidae, and it also contributes to the low birth weight of infants.

In summary, sufficient evidence shows that moderate to very severe anemia can produce undesirable health consequences. From the perspective of reproductive health outcomes, only very severe anemia has clearly been shown to result in maternal mortality. The justification for controlling maternal anemia can be based on the undesirable effects of anemia other than poor birth outcomes and harmful health consequences beyond that of anemia alone.

Physiologic significance of a high hemoglobin concentration
As with defining anemia, the normative definition of high hemoglobin can be based on +2 SDs from the reference mean. For nonpregnant women, this is equivalent to a hemoglobin concentration of 155–160 g/L. Analogous with mild anemia being within 10 g/L of the cutoff, a mild-high hemoglobin concentration would be 160–170 g/L. For women, a hemoglobin concentration >170 g/L can perhaps be regarded as a moderately elevated value. During pregnancy, the upper level for defining high hemoglobin would be lower than that in nonpregnant women because of the physiologic changes in the hemoglobin concentration during pregnancy (12). Again, the meaning of the elevated hemoglobin concentration and the probability of association with adverse events depends on the specific individual or population under study.

Although much of the clinical and public health focus has been on anemia or the low end of the hemoglobin distribution, there are also problems at the high end. The most notable consequence is increased blood viscosity. The relation between hemoglobin and blood viscosity is linear when the hematocrit is <0.50 (equivalent to a hemoglobin value of 160 g/L). Above this hemoglobin concentration, the relation becomes exponential—a small increase in hemoglobin or hematocrit results in a large increase in viscosity (18). Once hemoglobin concentrations reach =" v:shapes="_x0000_i1026" border="0" height="10" width="8">180 g/L, the blood viscosity reaches a level that impairs microcirculation and an inadequate amount of oxygen is transported to tissues, similar to the situation with severe anemia. This is often manifested as peripheral cyanosis and impaired mental function resulting from compromised cerebral blood circulation (21). Additionally, because of the poor blood flow, the risk of thromboembolism increases substantially (21). A study of individuals with chronic mountain sickness (or polycythemia), which is due to a combination of high altitude residency and poor pulmonary function, showed that long-term survival with a hemoglobin >200 g/L is not possible (21).
--(Ray Yip, American Journal of Clinical Nutrition)---