REF-1553 · Ishibashi et al., 2022 · Nutrients · Paper · PubMed · Free full text

Does a double keep hepcidin high in women?

In 13 university runners it stayed above a rested, fasted baseline to the next morning, on about 6 g/kg carbohydrate. No single-session day to compare, so the double is not shown as the cause.

Main caveat: No control. No rest day and no single session; the authors state direct comparison is still required.
limited, mixed evidenceMeasured on highly trained athletes

Thirteen women from one top university team, with no training on the day before, ran on the track for 1 h from 07:00 and for 2 h from 15:00, averaging 22.4 ± 0.9 km for the day. Blood was drawn at 06:00 fasted (P0), 14:00 (P8), 20:00 (P14) and 06:00 fasted next day (P24). Hepcidin showed a main effect of time (F = 4.2, p = 0.01) and was higher than P0 at each later sample.

Serum iron was higher at P24 than at P0. Haptoglobin and IL-6 showed no significant change over time, although no sample was taken immediately after exercise. Haemoglobin and ferritin between the two mornings showed no clear change. Food diaries gave 2,153 ± 98 kcal and 6.0 ± 0.4 g/kg of carbohydrate on the exercise day.

The authors read this as evidence that twice-a-day training raises hepcidin into the next morning and may raise iron-deficiency risk, likely with low carbohydrate as a contributor. They say themselves that a control day and a single-session day were not included.

Our reading: the only same-day hepcidin data in women that the search found, and a hypothesis, not a result on the double.

Protocol

Design: one group, one day; no rest-day or single-session condition, which the authors list as their first limitation. The rationale was to keep the sessions real. Ethics approval from the Japan Institute of Sports Sciences (IRB-2016-033).
Who: as above; training twice daily, 1 h early and 2 h in the afternoon, 6 days a week.
Day before: no training, and the diet that day was recorded.
Sessions: 07:00-08:00 and 15:00-17:00 on the track. Pace and intensity are not given in the text; only total distance is.
Blood: Hb, iron, ferritin, haptoglobin by clinical laboratory; IL-6 and hepcidin by ELISA in duplicate (intra-assay CV 5.0% and 4.4%).
Diet: food diaries with photos and plate weighing, checked by a dietitian.
Analysis: one-way repeated-measures ANOVA with Tukey post hoc; paired t tests for diet; data as mean ± SE; p < 0.05.
Funding and interests: no external funding; the authors declare no conflict of interest.

Key Results

MeasureP0 (06:00 fasted)P8 (14:00)P14 (20:00)P24 (06:00 fasted)Time effect
Hepcidinbaselinehigher than P0higher than P0higher than P0F = 4.2, p = 0.01, η2 = 0.26
Serum ironbaselinehigher than P0F = 0.471, p = 0.016, η2 = 0.29
IL-6no clear changeF = 0.57, p = 0.47, η2 = 0.01
Haptoglobinno clear changeF = 0.08, p = 0.97, η2 = 0.01
Haemoglobin (g/dL)12.7 ± 0.112.4 ± 0.4p = 0.16
Ferritin (as printed, "g/dL")24.3 ± 5.425.4 ± 5.8p = 0.89
Diet (mean ± SE)Day beforeExercise dayp
Energy (kcal)2,177 ± 552,153 ± 98> 0.05
Carbohydrate (g)250.5 ± 7.9283.4 ± 17.3> 0.05
Carbohydrate (g/kg)5.3 ± 0.26.0 ± 0.4> 0.05
Carbohydrate (% energy)46.8 ± 0.754.4 ± 1.7< 0.01
Fat (g)83.6 ± 2.061.2 ± 4.0< 0.001
Protein (g)101.2 ± 3.3106.8 ± 5.1> 0.05
Iron (mg)15.6 ± 0.415.7 ± 0.6> 0.05
Hepcidin values are in a figure. The text prints no concentrations and the figure was not read, so the size of the rise is not known here.
"Under 6 g/kg" is the abstract's wording. The table's mean for the exercise day is 6.0 ± 0.4 (SE); the discussion gives a range of 5.3 to 6.0, which are in fact the two day means.
Clock check (our arithmetic). The next-morning sample is 13 h after the afternoon session ended and 22 h after the morning one, not "24 h after completion" as in the abstract; the 24 h counts from the baseline sample. And 14:00 is 6 h after the morning session ended, where the methods write 8 h.

Mechanisms

Measured: hepcidin above fasted baseline at all later samples; no clear change in IL-6 at the sampled times; higher next-morning serum iron.
Authors' interpretation (speculation by their own word): low muscle glycogen, low carbohydrate and low energy availability drove the prolonged hepcidin; energy expenditure and glycogen were not measured. They also suggest the higher serum iron may reflect red-cell breakdown; haptoglobin did not change, which they put down to chronically low levels.
Time of day (authors' caveat): hepcidin has a daily rhythm, low early and higher in the afternoon, so the P8 and P14 rises are mixed with the clock. This does not apply to P24, drawn at the same hour as P0.
Rest day (inference): the baseline followed a day off, while these runners usually train six days a week. A next-morning value above a rested baseline may be their ordinary state after any training day; this design cannot tell.

Practical Application

Does a double keep hepcidin high in women? In 13 university runners it stayed above a rested, fasted baseline to the next morning, on about 6 g/kg carbohydrate. No single-session day to compare, so the double is not shown as the cause.
Good for: a reason to think about carbohydrate, and by our inference iron timing (the paper does not discuss it), in women who train twice a day, as a hypothesis to test in the individual, not a measured cost of doubling.
Good for: a reminder that a hepcidin value in the morning may depend on the training day before it (inference; the base's blood-test guidance is in SYNTH-97).
Not good for: the authors' line that at-risk athletes should mind training frequency. Frequency was not varied, so this goes beyond the data.
Not good for: diagnosing iron deficiency or claiming lost stores. Ferritin showed no clear change over 24 h, which is too short to expect one (inference), and iron absorption was not measured.
Not good for: deciding carbohydrate targets. Intake was recorded, not manipulated.
Not from this paper: carbohydrate targets for double days; see the base's fuelling guides.

Limitations

No control. No rest day and no single session; the authors state direct comparison is still required.
Second session cannot be credited. Hepcidin was already raised at 14:00, an hour before the second session started.
Rest day before baseline, so the comparison is with a state these athletes are rarely in.
Four samples only. IL-6 was not sampled right after exercise, so the "no change" says little about the exercise response. Only the two 06:00 samples are stated as fasted.
Statistics look inconsistent. An F of 0.471 for serum iron is printed with p = 0.016, and the ferritin unit is printed as g/dL. Values are reproduced here as printed; the first looks like an error (our reading).
Diet is self-report, with energy expenditure unmeasured; low energy availability is a guess by the authors.
Session intensity is not given, and one team on one day is one set of conditions.
Data are not public, by the authors' statement, because of ethical restrictions.
Menstrual phase and contraceptive use are not mentioned.
Not measured: iron absorption, muscle glycogen, performance, any longer period.
How this entry was checked and changed
Why this level

an uncontrolled pre-post earns B by design, lowered one step because the only outcome is a blood marker over a day, with no comparison condition, and because the printed statistics are not all coherent (see Limitations). Read from the full text (CC BY)