Hemorrhage-adjusted iron requirements, hematinics and hepcidin define hereditary hemorrhagic telangiectasia as a model of hemorrhagic iron deficiency

Abstract

Background: Iron deficiency anemia remains a major global health problem. Higher iron demands provide the potential for a targeted preventative approach before anemia develops. The primary study objective was to develop and validate a metric that stratifies recommended dietary iron intake to compensate for patient-specific non-menstrual hemorrhagic losses. The secondary objective was to examine whether iron deficiency can be attributed to under-replacement of epistaxis (nosebleed) hemorrhagic iron losses in hereditary hemorrhagic telangiectasia (HHT).

Methodology/principal findings: The hemorrhage adjusted iron requirement (HAIR) sums the recommended dietary allowance, and iron required to replace additional quantified hemorrhagic losses, based on the pre-menopausal increment to compensate for menstrual losses (formula provided). In a study population of 50 HHT patients completing concurrent dietary and nosebleed questionnaires, 43/50 (86%) met their recommended dietary allowance, but only 10/50 (20%) met their HAIR. Higher HAIR was a powerful predictor of lower hemoglobin (p = 0.009), lower mean corpuscular hemoglobin content (p<0.001), lower log-transformed serum iron (p = 0.009), and higher log-transformed red cell distribution width (p<0.001). There was no evidence of generalised abnormalities in iron handling Ferritin and ferritin(2) explained 60% of the hepcidin variance (p<0.001), and the mean hepcidinferritin ratio was similar to reported controls. Iron supplement use increased the proportion of individuals meeting their HAIR, and blunted associations between HAIR and hematinic indices. Once adjusted for supplement use however, reciprocal relationships between HAIR and hemoglobin/serum iron persisted. Of 568 individuals using iron tablets, most reported problems completing the course. For patients with hereditary hemorrhagic telangiectasia, persistent anemia was reported three-times more frequently if iron tablets caused diarrhea or needed to be stopped.

Conclusions/significance: HAIR values, providing an indication of individuals' iron requirements, may be a useful tool in prevention, assessment and management of iron deficiency. Iron deficiency in HHT can be explained by under-replacement of nosebleed hemorrhagic iron losses.

Figure 1. Actual and recommended dietary iron intakes for the 50 study participants.

A) Current recommendations using US recommended dietary allowance (RDA [2]) values for iron (blue dropped line circles, at 8 or 18 mg/day): the left hand five datasets, with higher RDA values, represent the five premenopausal women, the remaining 45 datasets with lower RDA values represent males and post menopausal females. Red columns indicate each individual’s iron intake per day from their personalised food frequency questionnaire (FFQ, intake of 130 food items presented in Table S1). Note that the RDA was not met by any of the pre-menopausal females. B) HAIR recommendations: The same intake data as in A) are now illustrated on a natural logarithmic scale to allow presentation of each individual’s personalised HAIR value, calculated according to their personalised iron losses, and US based recommended dietary allowance (RDA) for iron, presented in Table 1. Note that a log(HAIR) of 3 corresponds to a HAIR of 20 mg/day (approximate needs of a male blood donor); a log(HAIR) of 4 to a HAIR of 55 mg/day (approximate needs over 3 months to replace a 3–4 g/dl drop in hemoglobin), and a log(HAIR) of 5 to a HAIR of 148 mg/day. Generally short (0.5–2.5 min) nosebleeds less than once per month resulted in log(HAIR) of approximately 2; several nosebleeds per week of 5 minutes or more in a log(HAIR) of ∼3; daily 10 min nosebleeds a log(HAIR) of ∼4, and several nosebleeds per day, each lasting 2.5–10 minutes, in a log(HAIR) of 5.

Figure 2. Raw data on nosebleed frequency and duration.

Typical number of nosebleeds per month (blue symbols/lines, data from ESS question 1), and typical duration of nosebleeds per month (red symbols/lines, data from ESS question 2) reported by the 50 study participants, ordered by increasing value of HAIR. Nosebleeds reported as “typically gushing or pouring” were significantly longer than nosebleeds reported as “typically not gushing or pouring” (mean [standard deviation] 8.9 [6.4], versus 4.5 [4.9] minutes, Mann Whitney p = 0.0038).

Figure 3. Details of the nosebleeds reported by the online survey respondents.

A) Reported volume (mls) of individual nosebleeds, converted where appropriate from original units of measurement to mls as described in the methods. B) Reported duration (minutes) of individual nosebleeds. Corroborating evidence for specified major bleeds was provided by 16 individuals, and included acute hemodynamic consequences (faints, collapses, n = 5); hematocrit/hemoglobin falls (n = 4 including 3.2 g/dl hemoglobin fall in 8 hours; 8 units of hematocrit over 3 days); and unspecified acute transfusions or hospital admission (n = 8). There was no corroboratory evidence for the two indicated outliers (red crosses) whose values were excluded from calculations for the median, 20^th and 5^th percentile values used in nosebleed rate conversions.

Figure 4. Iron intakes from diet and iron supplements.

The nine individuals using ferrous sulphate, ferrous gluconate, ferrous fumarate or other iron supplements, are illustrated with pairwise comparisons of their dietary iron intake from the FFQ (left red bar) and dietary FFQ intake plus their supplement iron intake (right red bar) in addition to each individual’s log(RDA) and personalised log(HAIR), which were calculated according to their personalised iron losses and RDA for iron. As in Figure 1, note that generally short (0.5–2.5 min) nosebleeds less than once per month resulted in log(HAIR) of approximately 2; several nosebleeds per week of 5 minutes or more in a log(HAIR) of ∼3; daily 10 min nosebleeds a log(HAIR) of ∼4, and several nosebleeds per day, each lasting 2.5–10 minutes, in a log(HAIR) of 5. The two highest intakes were seen in individuals using ferrous sulphate 325 mg bd; the next six in users of once daily ferrous sulphate or ferrous fumarate.

Figure 5. Stratification of blood hematinic and iron indices according to HAIR values and iron supplement use.

A–E: Distributions of all participants, either by conventional recommended dietary allowance (RDA, left two box plots (mens., menses distinguishing males and post menopausal women from pre-menopausal women), or by approximate quintile (Qu) determined by HAIR (right five box plots, exact figures provided in methods). A) HAIR values (mg of iron per day). B) Serum iron (µmol/L). C) Hemoglobin (Hb, g/dL)). D) Mean corpuscular hemoglobin concentration (MCHC, g/dl). E) Red cell distribution width (RDW). F) Quadratic regression plots for the distribution of hemoglobin according to HAIR, in study participants using oral iron supplements (continuous line with 95% confidence interval indicated), and those who did not (dotted line).

Figure 6. Relationships between hepcidin and iron indices.

A and B) Hepcidin levels according to approximate tertile groupings of A) serum iron, and B) serum ferritin. Note that the reference range for hepcidin using this radioimmunoassay is 1.1–55 ng/mL. Details of individual participants are provided in Table 4. C) and D) Best fit quadratic regression relationships (with shaded areas indicating the 95% confidence intervals) for hepcidin with C) serum iron, and D) ferritin.