Daily versus three times weekly dosing for treatment of iron deficiency nonanemia in NCAA Division 1 female athletes

To compare the effectiveness, compliance, and side effect profile between daily or three times weekly (TIW) oral iron supplementation regimens in treating iron deficiency nonanemia (IDNA) in National Collegiate Athletic Association (NCAA) Division 1 female track and field or soccer athletes.


INTRODUCTION
Among its many roles in the body, elemental iron is essential for proper delivery of oxygen to tissue. 1 Aerobic respiration, energy production requiring oxygen, is necessary in nearly all sports.Accordingly, adequate iron levels play a significant role in athletic performance.3][4][5] Performance limitations have been noted with both iron deficiency anemia and iron deficiency nonanemia (IDNA).Having an adequate total body supply of iron is needed to optimize athletic performance and recovery in athletes.
Serum ferritin, a storage protein for iron, is used as a surrogate measure for iron sufficiency. 6Low-normal levels of ferritin ranging from 12 to 40 μg/L, in the context of normal hemoglobin levels, may have a significant effect on athletes, causing fatigue, poor performance, and/or increased recovery times. 7,8Ferritin levels are increasingly used to screen aerobic athletes for iron deficiency and monitor response to supplementation.Correction of IDNA via iron supplementation is theorized to augment performance and recovery. 2,3ron supplementation strategies are multi-faceted.The route of administration varies between intravenous (IV), intramuscular, and oral.Frequency of dosing ranges from daily to monthly.Available iron formulations include oral ferrous salts (ferrous sulfate, ferrous gluconate, and ferrous fumarate), oral iron amino acid chelates (ferrous bisglycine chelate, ferric trisglycine chelate, ferric glycinate, and ferrous bisglycinate hydrochloride), and several formulations of IV iron. 9Each formulation carries a variable risk-benefit profile.For example, one of the original IV iron infusions, iron dextran, is a quick and effective iron supplement, but carries an increased risk for anaphylaxis compared to newer IV iron supplements. 10,11Risks and benefits, side effects, and cost vary with iron supplementation strategy.Oral iron supplementation is the most frequently used strategy due to minimal cost and ease of administration.However, oral iron supplementation commonly causes the side effects of gastrointestinal upset, constipation, and stool discoloration, which may induce anxiety.Stoffel et al. found that, compared to daily oral iron supplementation, pulsed iron supplementation three times weekly (TIW) may maintain or increase absorption of elemental iron while decreasing the severity and frequency of gastrointestinal side effects. 12n our study, we sought to maximize efficacy of iron supplementation while minimizing gastrointestinal side effects.We used oral ferrous bisglycinate since recent research indicates it is as effective as other oral ferrous salts in treating iron deficiency while potentially decreasing side effects. 13Two dosing strategies were used to determine effectiveness and side effect profile of a daily or TIW oral iron supplementation regimen in treating IDNA in National Collegiate Athletic Association (NCAA) Division 1 female athletes.Data gathered included ferritin levels, compliance with oral iron supplementation, and side effects.We hypothesized that TIW dosing of oral iron is as effective at increasing ferritin levels with fewer side effects compared to daily dosing.

METHODS
This single institution prospective cohort study was approved by the university institutional review board.Athletic Department approval was obtained.All university soccer and track and field athletes underwent scheduled screening laboratory analysis per the university's sports medicine protocol.Female athletes who met criteria for IDNA (ferritin below 35 μg/L and hemoglobin >12 g/dL) were included (Figure 1).Informed consent was obtained from all participants.Eighteen female track athletes and 16 female soccer athletes were enrolled.One athlete was excluded because she was undergoing concurrent IV iron therapy for previously diagnosed iron deficiency as directed by a hematologist.Demographics including athlete's age, body mass index (BMI), and sport were collected (Table 1).
Participants self-selected into TIW or daily dosing regimens as a part of a field-based screening protocol.The team's sports nutritionist provided bottles of ferrous

Statistical analysis
Considering the nonrandomized (observational) design, propensity score matching was employed with the goal to estimate the difference between regimens via the average effect of the treatment on the treated. 14Matching improves parametric statistical models for estimating treatment effects in observational studies by controlling for pretreatment confounding variables. 15,16The present study did not capture why the athletes chose their selected regimen, but the observed pretreatment serum ferritin (preferritin) values, BMI, and proportion of sport exhibited differences between regimens, which were all sources of confounding to mitigate self-selection bias.The propensity scores with optimal full matching were fit using logistic regression on regimen with preferritin, BMI, and sport as the covariates, which matched every daily participant to at least one TIW participant according to these three covariates.To estimate the average effect of the treatment on the treated for increasing serum ferritin between regimens, a linear regression was fit with posttreatment ferritin (postferritin) as the outcome and regimen as the covariate, which included the full matching weights in the estimation.The coefficient on regimen was taken as the average effect of the treatment on the treated estimate, whereas the cluster-robust variance was used to estimate the standard error. 17Precision of estimates was then indicated with 95% confidence intervals (CI) to represent uncertainty in the average effect of the treatment on the treated between regimens.Effect sizes were also calculated to assess relative magnitude of change between regimens, using Hedges' g to adjust for bias with the small sample size.Also considering the limited sample size, we conducted a sensitivity analysis to assess the robustness of our estimates to potential unmeasured confounding variables. 18econd, to assess how effective the two regimens were at increasing ferritin values >35 μg/L, a logistic regression was fit on the postferritin values > and <35 μg/L with regimen and preferritin values as the covariates.The logistic regression model was fit using the propensity score weights from the full matching to mitigate confounding, whereas the cluster-robust variance was used to estimate the standard errors.The predicted probability with 95% CI were then used to estimate the functional relationship of the two regimens conditional on preferritin values.The proportion of athletes who surpassed 35 μg/L postferritin between regimen was analyzed with a χ 2 test.
Finally, a χ 2 analysis was performed to assess the relationship of questionnaire response dimensions between regimens.The Rao-Scott second-order correction and the Satterthwaite degrees of freedom approximation were applied to properly adjust for weighing and stratification from the survey sampling scheme. 19,20ffect size measure (Cramér's V) was used with the adjusted contingency table to assess relative magnitude of effects.For each analysis, significance was determined a priori (α = .05)and effect sizes were qualitatively interpreted following the rules of Cohen. 21

RESULTS
Thirty-three NCAA Division 1 female athletes were included in the intention-to-treat population (18 track and field athletes and 15 female soccer players) (Figure 1).Of the track and field athletes, five self-selected to the daily dosing regimen and the remaining 13 self-selected to the TIW regimen (Table 1).Of the 15 soccer players, 10 self-selected to the daily dosing regimen and the remaining five self-selected to the TIW regimen.Table 2 displays the observed ferritin and hemoglobin values pre and post treatment.After optimal full matching, the standardized mean difference (Hedges' g) between preferritin regimens was well below 0.1, with the variance ratio and Kolmogorov-Smirnov statistic approaching 1 and 0, respectively, indicating adequate balance (Table 3).More details about the matching procedures and balance diagnostics can be found in the Supplemental Content.
The average main effect for the TIW regimen was a significant increase of 5.17 μg/L (95% CI: 0.86-9.47,p = .02)in serum ferritin, to a medium degree (g = 0.86, 95% CI: 0.14-1.58).The average main effect for the daily regimen was a significant increase of 12.88 μg/L (95% CI: 4.84-20.93,p = .003)in serum ferritin, to a large degree (g = 1.14, 95% CI: 0.39-1.88).Thus, the estimated average effect of the treatment on the treated between regimens was a non-significant decrease of À7.17 μg/L (95% CI: À19.02-3.59,p = .17)in serum ferritin, to a medium degree (g = À0.49,95% CI: À1.18-0.21).Meaning, after controlling for the confounding preferritin values, BMI and sport, we expect the TIW regimen to increase serum ferritin 7.17 units less than the daily regimen after 8 weeks of supplementation.Considering the limited sample size, the results suggest that our estimates are only moderately robust to unmeasured confounding, given that an unmeasured confounder would need to be associated with both treatment and postferritin by at least a medium degree (g = 0.62) to negate the observed effect.
Only 17% of the athletes taking iron TIW had posttreatment serum ferritin levels >35 μg/L, whereas 60% of the athletes taking iron daily surpassed this threshold (χ 2 ¼ 6:64, p = .01).For a one-unit increase in preferritin values, the odds of surpassing a ferritin value of 35 μg/L increased by a factor of 1.27 (95% CI: 1.04-1.55,p = .02).The odds of surpassing a ferritin value of 35 μg/L for the TIW regimen compared to the daily regimen decreased by a factor of 0.11 (95% CI: 0.009-1.63,p = .11),meaning that the odds for the TIW regimen to surpass a ferritin value of 35 μg/L were 8.5 times lower than for the daily regimen on average, but this difference lacked statistical significance.

DISCUSSION
The primary aim of this study was to compare the effectiveness of TIW dosing to daily dosing of ferrous bisglycinate for correcting iron deficiency in IDNA while minimizing side effects in NCAA Division 1 female athletes.Each regimen exhibited a significant average main effect of ferritin increase over the 8-week supplementation period.The daily regimen produced a larger improvement in serum ferritin compared to the TIW regimen, although the average effect of the treatment on the treated between regimens did not achieve statistical significance (p = .17).This is most likely explained by the total dosing difference between the groups.The TIW group received three doses per week whereas the daily group received seven doses per week.However, these findings provide evidence that either strategy of oral supplementation will significantly increase the serum ferritin in IDNA athletes without a significant difference between the TIW strategy.Figure 3 illustrates the functional relationship of the logistic regression model for surpassing a postferritin value of 35 μg/L.Female athletes with preferritin <15 μg/L were unlikely to achieve the clinical threshold within 8 weeks, regardless of regimen.Preferritin values between 20 and 30 μg/L exhibit a discernable discrepancy in the point estimate between regimens, but with wide confidence intervals due to the small sample sizes.The TIW point estimate for postferritin ≥35 μg/L does not exceed 50% until a preferritin value of 30 μg/L, which coincides with the average main effect of 5.17 unit increase over 8 weeks.These results suggest that preferritin values may be an important factor to consider when prescribing iron supplementation regimens.The failure of several athletes to achieve a serum ferritin of 35 μg/L despite oral iron supplementation may be related to the relatively short duration (8 weeks) of therapy.Failure to achieve the threshold lends evidence that an athlete with a preferritin <15 μg/L would likely require higher dosages or longer duration of treatment to achieve a ferritin level above the clinical threshold for iron deficiency.
The subjectively reported side effects of nausea (p = .04)and constipation (p = .002)were significantly more common in the daily regimen compared to the TIW regimen, which is consistent with prior findings, whereas dark stools and stomach pain did show significant difference. 22The differences in the side effects could be caused by several mechanisms including improved absorption, decreased total dose, or small sample size, to name a few.With less nausea and constipation, signs of gastrointestinal distress, and as most of iron absorption occurs in the duodenum, it could be evidence the TIW dosing was more efficiently absorbed compared to the daily dose.Therefore, if the primary mechanism is indeed improved absorption, this would provide clinical relevance in supporting the use of TIW dosing.TIW dosing with ferritin goals of 35-50 μg/L have been recommended in the literature previously, with supporting evidence demonstrating improved iron stores and reduced side effects. 23The significantly reduced side effect profile in the TIW dosing compared to the daily dosing is highly relevant in the clinical setting because minimizing side effects is a large part in optimizing supplementation strategies.
With reduced side effects seen in TIW regimens, future research may investigate a TIW dosing regimen that is dosed twice on supplementation days, as this may further increase ferritin levels but continue to avoid side effects when compared to daily regimens. 23espite a difference in side effects, there was no statistical difference in compliance between the regimens ( p = .14).We suspect this could be due to the short duration of iron supplementation.If the supplementation period was extended to ferritin levels within normal values we would expect increased risk of side effects.Given the findings in this study, it provides support that using a TIW regimen has fewer side effects and, therefore, it may be a potential strategy for increasing the dose without having increased side effects one would expect to see in daily dosing strategies.It should also be mentioned that the formulation used in this study has been shown to decrease these side effects when compared to the ferrous sulfate, one of the most common supplement formulations.Therefore, although there was no direct comparison in this study, the results of this study do provide evidence the bisglycinate formulation likely is more favorable in regard to side effects than we would expect with ferrous sulfate.Future studies could compare the efficacy and time course of oral and IV iron supplementation regimens.IV iron infusions are gaining in popularity and have shown potential for rapid correction of iron deficiency and increased athlete performance although they are more invasive. 24ecognizing that our participants were competitive athletes actively engaged in their sports season, we adopted a self-selection model to ensure that the regimen would not disrupt their established routines.By giving athletes the autonomy to choose a regimen that aligned with their personal preferences and performance strategies, we anticipated higher compliance and a more accurate reflection of the regimen's integration into their competitive routines.The self-selection model also serves to mimic real-world conditions under which athletic supplementation strategies are employed, thus we believe this enhances the external validity and ecological validity of our findings.
This study provides insight into practical iron supplementation for correction of IDNA in female athletes.Female athletes have a higher risk for iron deficiency than male athletes, likely related to blood loss with menstruation. 25We found that correction of IDNA can be achieved using either the TIW or daily dosing regimens, theoretically leading to improvement in aerobic markers of maximum muscle contraction, VO 2 max, and distance-based time trials.This potential improvement in athletic performance comes at little cost because oral supplementation of iron is well tolerated.Our research suggests that supplementation with oral ferrous bisglycinate is effective, noninvasive, relatively inexpensive, and is marked by few gastrointestinal symptoms.
This study has several limitations.It was a nonblinded, nonrandomized prospective study with a limited sample size due to the field-based screening protocol.We attempted to preserve the validity of our findings by (1) mitigating potential confounding factors by achieving covariate balance through propensity score matching, (2) applying bias correction strategies on all estimates, and (3) performing a sensitivity analysis to quantify the minimum confounder association needed to negate the observed estimates.Nonetheless, the limited sample size may still affect the generalizability of the findings.Our study used ferrous bisglycinate as the iron formulation, so our findings may not be comparable to studies with other oral iron formulations such as ferrous sulfate, particularly regarding side effects.Survey questions gathered throughout the study may have created a bias in perception of side effects.The survey also assessed compliance as a binary yes or no question compared to a more detailed question regarding compliance with timing of supplement and if it was taken with food.Future research should use a double-blinded placebo methodology to minimize subjectivity in participant responses as well as expanding the duration with a more detailed compliance survey.

CONCLUSIONS
Thirty-three female NCAA Division 1 athletes with IDNA self-selected into one of two oral iron dosing regimens (daily or TIW) over 8 weeks.Although both regimens exhibited a significant average main effect of ferritin increase, the daily regimen produced a larger increase in serum ferritin.Athletes in TIW regimen experienced less nausea and constipation compared to the daily regimen.Our results suggest that baseline ferritin values are important to consider when prescribing iron supplementation regimens.

3
Predicted probability (± 95% confidence interval) of surpassing clinical threshold of ≥35 μg/L per regimen, conditional on the preferritin values.Probability estimates were obtained from logistic regression using the propensity score weights and cluster-robust variance for the SEs.TIW, three times weekly.
T A B L E 2 Mean ± SD with adjusted estimates (95% confidence interval) of ferritin and hemoglobin values, with the number improved to 35 μg/L normal clinical values prior to and following 8-week supplementation.Numeric summaries of balance between observed data before and after full matching.Categorical variable of sport are proportions.
a Adjusted estimates after covariate balancing.bAveragetreatmenteffecton the treated.cOddsratio(95%confidence interval).T A B L E 3Abbreviations: BMI, body mass index; KS, Kolmogorov-Smirnov statistic; TIW, three times weekly.T A B L E 4 χ 2 results of questionnaire response dimensions between regimens.Note: * indicates p-values < .05.