Safety and efficacy of subcutaneous alpha-tocopherol in healthy adult horses
C. G. Donnelly†,*
, E. Burns‡
, C. A. Easton-Jones§§
, S. Katzman§
, R. Stuart¶
, S. E. Cook††
, C.
J. Finno†,‡‡
†William R. Pritchard Veterinary Medical Teaching Hospital
‡Morris Animal Foundation, Denver, Colorado
§Department of Surgical and Radiological Sciences, University of California: Davis, Davis
¶Stuart Products Inc, Bedford, Texas
††Department of Pathology, Microbiology and Immunology, University of California: Davis, Davis,
USA
‡‡Department of Population Health and Reproduction
§§Rossdales Equine Hospital, Exning, UK
Summary
Vitamin E is essential for neuromuscular function. The primary treatment, oral supplementation
with natural (‘RRR’) α-tocopherol, is not effective in all horses. The objectives of this pilot study
were to evaluate the safety and efficacy of a subcutaneously administered RRR-α-tocopherol
preparation. Horses were randomly assigned in a cross-over design to initially receive RRR-α-
tocopherol (5000 IU/450 kg of 600 IU/mL) subcutaneously (n = 3) or orally (n = 3) or were
untreated sentinels (n = 2). Tissue reactions following injection in Phase I of the study necessitated
adjustment of the preparation with reduction of the RRR-α-tocopherol concentration to 500
IU/mL in Phase 2. Following an 8-week washout period, horses received the reciprocal treatment
route with the new preparation (5000 IU/450 kg of 500 IU/mL). Serum, CSF and muscle α-
tocopherol concentrations were determined by high-performance liquid chromatography over a
14-day period during each phase. Serum and CSF α-tocopherol concentrations increased
significantly postinjection only when the 500 IU/mL product was administered (P<0.0001). There
was no significant difference in the muscle concentration of α-tocopherol following either
*
Corresponding author cgdonnelly@ucdavis.edu.
Authorship
C. Donnelly, C. Finno and R. Stuart contributed to study design, sample collection and data analysis. E. Burns, S. Katzman and C.
Easton-Jones assisted with sample collection. S. Cook assisted with histopathology interpretation. All authors contributed to the
manuscript.
C. G. Donnelly's present address: Department of Population Health & Reproduction University of California: Davis, Davis, California,
USA
Authors’ declaration of interests
R. Stuart is the director and owner of Stuart Products which produces nutritional products.
Ethical animal research
Ethical approval by the University of California-Davis Institutional Animal Care and Use Committee (Protocol number 2009).
Manufacturers’ addressesAuthor Manuscript Author Manuscript Author Manuscript Author Manuscript
Donnelly et al. Page 2
treatment. All eight horses had marked tissue reaction to subcutaneous injection, regardless of
product concentration. Whilst we have demonstrated that this route may be a useful alternative to
oral supplementation, the marked tissue reaction makes use of such products limited at this time to
only the most refractory of cases.
Keywords
horse; vitamin E; alpha-tocopherol
Introduction
Horses appear to be particularly susceptible to vitamin E deficiency mediated neurological
and muscular degeneration (Finno and Valberg 2012). Horses at risk of deficiency, most
commonly due to dry-lot management, or with diagnosed deficiency (serum α-tocopherol
<2 μg/mL) are typically supplemented with RRR- α-tocopherol, the most potent enantiomer
of vitamin E and the natural form found in lush pasture (Finno and Valberg 2012). However,
there are anecdotal reports of horses administered appropriate doses and formulations of oral
α-tocopherol failing to respond clinically, with serum concentrations not increasing as
expected. Recently, malabsorption of orally administered α-tocopherol secondary to
eosinophilic enteritis has been implicated in the aetiology of equine motor neuron disease
(Díez de Castro et al. 2016). Currently, there are no parenteral α-tocopherol products
approved for use in horses. Whilst E-Se® contains alpha-tocopherol, it is in the synthetic
formulation and has been demonstrated not to affect serum alpha-tocopherol concentrations
in foals (Finno et al. 2015). Administration by parenteral routes is routinely performed in
food producing species with apparent efficacy (Hidiroglou and Karpinski 1991). The
objective of this study was to provide preliminary safety and efficacy data for the
subcutaneous administration of α-tocopherol formulated for injection. We hypothesised that
subcutaneous administration of RRR-α-tocopherol would increase serum, cerebrospinal
fluid and muscle concentrations in vitamin E-deficient horses.
Materials and methods
Animals
Eight adult mixed breed horses (mares n = 4; geldings n = 4), weighing 547.7 ± 47.7 kg and
aged 3 to 12 years were used in this study. Based on a noninferiority power analysis for a
cross-over design with an α = 0.05, β = 0.8 and δ = 0.4, a sample size of three animals
would be necessary.
All horses were housed at the same facility on dry lots. All horses were fed twice daily with
grass hay at ~ 2% bodyweight per day and underwent annual routine husbandry such as
dental exams and core vaccinations. No animals in this study were provided with an oral α-
tocopherol supplement prior to the study. The Institutional Animal Care and Use Committee
at the University of California, Davis approved the study design.
Equine Vet Educ . Author manuscript; available in PMC 2022 April 01.Donnelly et al. Page 3
Study design
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Initially, a balanced cross-over design was employed with horses randomly assigned to
receive the same α-tocopherol preparation once by subcutaneous injection (n = 3) or orally
(n = 3). Horses were administered a total dose of 5000 IU (i.e. 10 IU/kg) α-tocopherol by
either route. Following an 8-week washout period, animals received the reciprocal treatment.
The washout period was determined from previously published data (Brown et al. 2017). All
subcutaneous injections were performed after aseptic preparation of the skin over the left
pectoral area. The α-tocopherol product was sterilely filtered via a 170- to 250-μm blood
component filter (Y-type blood solution set)1 prior to administration. Due to marked tissue
reactions with the original 600 IU/ml formulation (customised preparation, Phase 1)2, a 500
IU/ml preparation (Vital-E)2 was used for Phase 2 in an attempt to mitigate the swelling
associated with the injection. However, the total dose (5000 IU per horse) remained
unchanged for both phases. Two animals (one mare, one gelding; aged 7 and 11 years,
respectively) served as environmental sentinels for the entire study period, with only serum
samples collected from these animals. The University of California Animal Use and Ethics
Committee approved all procedures.
Cerebrospinal fluid (CSF) was collected on day 0 of the experiment. A jugular catheter was
placed in all horses. Horses were premedicated with xylazine (1.1 mg/kg bwt i.v.), and
general anaesthesia was induced with ketamine hydrochloride (2.2 mg/kg bwt i.v. and
midazolam (0.05 mg/kg bwt i.v.). Horses were placed in right or left lateral recumbency and
CSF fluid collected in sterile fashion by atlanto-occipital (AO) centesis using an 8.9-cm 18
gauge spinal needle. CSF was collected again in a similar fashion on day 7 post α-
tocopherol administration. AO centesis was performed under general anaesthesia in
preference to standing collection techniques in order to obtain a higher volume sample with
less risk of blood contamination. Samples were collected into plain light-protected plastic
vials and kept on ice. Samples were centrifuged at 4°C within 3 h of collection, and the
supernatant stored at −80°C until analysis. The same sampling protocol was followed for the
reciprocal treatment in phase 2 of the experiment.
Serum was collected immediately before anaesthesia on day 0 into plain light-protected
vacutainer tubes. Further serum samples were collected 24, 48, 96 h, 7 and 14 days
following α-tocopherol administration. All samples were centrifuged within 3 h of
collection and stored at −80°C until analysis. The same sampling protocol was followed for
the reciprocal treatment in phase 2 of the experiment.
Muscle was sampled from the gluteus medius whilst the horses were anaesthetised on days 0
and 7. An additional sample was collected on day 14 understanding sedation (xylazine 0.4
mg/kg bwt). Muscle samples were aseptically collected using a Bergström biopsy needle as
previously described (Snow and Guy 1976). The sample site was alternated for each sample.
All samples were flash-frozen in liquid nitrogen immediately at collection and stored in
plastic, light-protected vials at −80°C until analysis. The same sampling protocol was
followed for the reciprocal treatment in phase 2 of the experiment.
1Baxter, Deerfield, Illinois, USA. 2Stuart Products, Bedford, Texas, USA.
Equine Vet Educ . Author manuscript; available in PMC 2022 April 01.Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Donnelly et al. Page 4
Serum, CSF and muscle α-tocopherol concentrations were measured at the Iowa State
Veterinary Laboratory. Additionally, both injectable products were subjected to independent
α-tocopherol quantification. All samples and both injectable products were analysed by
HPLC as previously described (Finno et al. 2015).
Data analysis
Data were analysed by commercial software (GraphPad Prism 7.4)3. Due to small sample
size, baseline samples for serum, CSF and muscle were evaluated by Kruskal–Wallis test for
each treatment phase. There was no significant difference between baseline measurements;
therefore, the washout period was considered appropriate. Serum, CSF and muscle tissue α-
tocopherol concentrations were then evaluated by two-way repeated measures ANOVA, with
time and experimental group as fixed factors and horse as the random effect. Post hoc testing
was performed using Sidak’s multiple comparison test. Significance was set at P <.05.
Results
Serum α-tocopherol
For oral α-tocopherol administration, no time or treatment phase interaction was detected.
Therefore, all oral α-tocopherol supplementation data were subsequently combined. A
significant time treatment interaction (P <0.0001) was detected for the first versus the second
subcutaneous injection, and therefore, these data were analysed separately. The majority of
variation was derived from treatment (44%). Baseline serum concentrations of α-tocopherol
were not significantly different between treatment groups or between phases (Fig 1). All
animals were considered vitamin E deficient at the beginning of the trial (mean ± s.d.; 1.2 ±
0.07 μg/mL; reference range 2-4 μ/mL) (Finno and Valberg 2012). Serum concentrations
were significantly increased at 24 h (18.1 ± 12.05 P ≤0.0001), 48 h (16.53 ± 2.04 P ≤0.0001)
and 72 h (11.6 ± 3.08 P <0.05), postadministration only for the 500 IU/mL subcutaneous
injection in Phase 2. By days 7 and 14, subcutaneous administration in Phase 2 was not
significantly different compared to oral, Phase 1 injection or baseline concentrations.
CSF α-tocopherol
CSF α-tocopherol baseline concentrations were not significantly different between treatment
groups. Similar to serum, there was a significant time treatment interaction (<0.0001), with
treatments analysed separately. The majority of variation arising from time period (43.2%).
All horses were considered vitamin E deficient in the CSF at the beginning of the trial (4.3 ±
1.6 ng/mL; reference range ≥10 ng/mL) (Finno et al. 2015), injection with the 500 IU/mL
preparation resulted in significantly higher CSF α-tocopherol at day 7 (15.27 ± 0.95 ng/mL,
P ≤0.0001) compared to the 600 IU/mL injection (6.7 ± 0.99 ng/mL), oral administration
(6.66 ± 4.23 ng/mL) and baseline (Fig 2).
3GraphPad, San Diego, California, USA.
Equine Vet Educ . Author manuscript; available in PMC 2022 April 01.Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Donnelly et al. Page 5
Muscle α-tocopherol
Muscle concentrations were highly variable between animals (range; 0.1-4.3 μg/g). Whilst
there was a modest trend towards increased concentrations at 14 days post-subcutaneous
administration in Phase 2, this was not significant.
Tissue reaction
All animals developed marked swelling at the injection site, regardless of concentration of
the α-tocopherol formulation. For seven of eight animals, this tissue reaction was self-
limiting. One horse had persistent swelling and eventual drainage requiring surgical
intervention. Histology of the area was consistent with a sterile granuloma (Fig 3).
Independent assessment of product α-tocopherol concentrations
Each of the products was independently verified using HPLC (Finno et al. 2015). The
product in Phase I contained 336 mg/mL (equivalent to 672 IU/mL), with the product in
Phase 2 containing 290 mg (equivalent to 580 IU/mL).
Discussion
Currently, there are no labelled parenteral vitamin E preparations suitable for use in
treatment or prevention of horses with existing or recurrent deficiencies in vitamin E.
Additionally, with reports of malabsorptive conditions leading vitamin E deficiency, a need
exists for a safe and efficacious parental means of supplementing horses (Finno and Valberg
2012).
Despite the importance of vitamin E as a potent antioxidant, the exact mechanisms of its
biokinetics in horses are poorly understood and extrapolated from studies in other species
(Finno et al. 2011, 2015). The primary route of absorption of vitamin E is alimentary and is
closely associated with fat absorption requiring appropriate pancreatic, biliary and small
intestinal function (Desmarchelier and Borel 2018). A specific mechanism for malabsorption
in horses has not been evaluated. Speculatively, it is likely that perturbation of small
intestinal function plays a role, and however, this is yet to be substantiated. Horses require a
constant dietary supply to maintain appropriate serum concentrations of vitamin E, with
serum levels falling quickly in the absence of supply (Stuart et al. 2010; Brown et al. 2017).
Horses in the current experiment demonstrate this, with all animals considered deficient at
the beginning of the trial, presumably subsequent to dry-lot management (Stuart et al. 2010).
Despite this deficiency, no horses demonstrated any clinical signs of neuromuscular disease
throughout this trial. Further, horses in this experiment appeared to have normal intestinal
absorption of α-tocopherol, with an expected approximate doubling of serum concentrations
24 h after oral administration (Lodge et al. 2004; Stuart et al. 2010). Whilst this increase was
not statistically significant, it was a biologically appropriate response and indicates both a
normal ability to absorb and also to distribute α-tocopherol (Finno and Valberg 2012).
Vitamin E is fat-soluble and as such may also be absorbed by nonenteric routes. Following
subcutaneous, intramuscular or intraperitoneal injection in other species, α-tocopherol is
primarily taken up by lymphatics, before rapidly equilibrating between plasma and the cell
Equine Vet Educ . Author manuscript; available in PMC 2022 April 01.Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Donnelly et al. Page 6
membrane of erythrocytes (Lodge et al. 2004). Distribution to tissues rarely results in
accumulation of vitamin E to toxic levels, reflecting a highly regulated and active process
(Lodge et al. 2004). This process may also take place in the absence of major transport
proteins found in the liver (Irías-Mata et al. 2018). As such, the tissue-specific absorption
remains tightly regulated and is able to utilise nonalimentary α-tocopherol.
Injectable formulations of α-tocopherol have been investigated in a number of domestic
species (Knight and Roberts 1985; Hidiroglou and Karpinski 1987,1988). Administration by
this route appears to effectively and efficiently increase serum α-tocopherol concentrations.
In contrast, parenteral administration in horses has not been well studied and there are
currently no approved products for use in this species. As mentioned previously, products
containing synthetic vitamin E and selenium (such as E-Se®) are commonly administered
parenterally to neonatal foals, and however, these products provide insufficient amounts and
minimally bio-potent vitamin E to be used to treat deficiency mediated diseases (Finno et al.
2015).
Following subcutaneous administration with the 500 IU/mL preparation, supra-physiological
α-tocopherol concentrations in the serum and CSF were attained. Very low levels of α-
tocopherol administered parenterally are required to maintain erythrocyte stability in
deficient horses (Stowe 1968). Administration by this route circumvents both alimentary
losses as well as liver-mediated regulation and may account for the large difference in serum
and CSF concentrations. Subcutaneous administration demonstrates an opportunity to
rapidly increase nervous tissue concentrations, the main target of treatment for vitamin E-
mediated neurological diseases. Oral administration of suitable formulations may take up to
14 days to increase CSF concentrations, and therefore, parenteral preparations may be more
suitable for early treatment of vitamin E deficient neurological disease (Hidiroglou and
Karpinski 1991).
Administration of the 600 IU/mL preparation did not result in the same robust increase in α-
tocopherol serum and CSF concentrations. The primary difference between the preparations
was concentration, with equipotent doses administered. Local inflammation has previously
been recognised as a source of variation for absorbance of nonaqueous preparations
administered subcutaneously to horses (Alvinerie et al. 1998). Given the variability in
individual local inflammatory responses and the small number of animals available in this
study, the disparity of absorbance between products may represent the spectrum in local
reactions of individual horses. That is to say that horses with more marked tissue reaction
(like those in the Phase 1) would have reduced absorbance as reflected by lower serum and
CSF concentrations.
Muscle α-tocopherol concentrations were not significantly different pre- and post-
supplementation. Muscle concentrations of α-tocopherol are affected by the amount of lipid
within the muscle (Ronéus et al. 1986). Adipose tissue is a main storage site for α-
tocopherol, and therefore, muscle with increased amounts inter-fascicle fat will have
increased concentrations of α-tocopherol. The amount of adipose tissue was not
standardised in the current experiment, with whole tissue submitted for evaluation.
Equine Vet Educ . Author manuscript; available in PMC 2022 April 01.Donnelly et al. Page 7
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Additionally, as these animals were deficient prior to the study, there was likely preferential
distribution to adipose tissue over muscle (Lodge et al. 2004).
The tissue reaction observed in the current experiment was marked and affected all animals.
The rapidity by which reactions occurred may indicate large scale degranulation of tissue-
resident mast cells, leading to marked oedema in the region of injection (Krystel-
Whittemore et al. 2016; Jørgensen et al. 2018). Whilst the tissue reaction was self-limiting in
7 of 8 horses and required no further intervention, the degree of tissue reaction makes use of
this product in horses limited only to situations where the benefit of treatment outweighs the
development of complications.
In conclusion, parenteral administration of α-tocopherol via the subcutaneous route
effectively increases serum and CSF α-tocopherol concentrations. Preparations that
circumvent alimentary absorption provide a novel area of investigation for more efficacious
and long-term preparations of α-tocopherol. However, the current trial demonstrated that the
tested product is not safe for use in horses due to local tissue reaction and as such cannot be
recommended for clinical use at this time.
Acknowledgments
Source of funding
Stuart Products, Bedford, Texas, USA.
References
Alvinerie M, Escudero E, Sutra JF, Eeckhoutte C and Galtier P (1998) The pharmacokinetics of
moxidectin after oral and subcutaneous administration to sheep. Vet. Res 29, 113–118. [PubMed:
9601143]
Brown JC, Valberg SJ, Hogg M and Finno CJ (2017) Effects of feeding two RRR-α-tocopherol
formulations on serum, cerebrospinal fluid and muscle α-tocopherol concentrations in horses with
subclinical vitamin E deficiency. Equine Vet. J 49, 753–758. 10.1111/evj.12692. [PubMed:
28432750]
Desmarchelier C and Borel P (2018) Bioavailability of Vitamin E. Physiology of the Gastrointestinal
Tract: Sixth Edition (Sixth Edition, Vol. 2–2). Elsevier Inc. 10.1016/B978-0-12-809954-4.00052-9.
Díez de Castro E, Zafra R, Acevedo LM, Pérez J, Acosta I, Rivero JLL and Aguilera-Tejero E (2016)
Eosinophilic enteritis in horses with motor neuron disease. J. Vet. Intern. Med 30, 873–879.
10.1111/jvim.13944. [PubMed: 27015868]
Finno CJ and Valberg SJ (2012) A Comparative review of vitamin E and associated equine disorders. J.
Vet. Intern. Med 26, 1251–1266. 10.1111/j.1939-1676.2012.00994.x. [PubMed: 22925200]
Finno CJ, Higgins RJ, Aleman M, Bannasch DL, Ofri R, Hollingsworth SR, Reilly CM and Madigan
JE (2011) Equine Degenerative Myeloencephalopathy in Lusitano Horses. J. Vet. Intern. Med 25,
1439–1446. 10.1111/j.1939-1676.2011.00817.x. [PubMed: 22092640]
Finno CJ, Estell KE, Katzman S, Winfield L, Rendahl A, Textor J, Bannasch DL and Puschner B
(2015) Blood and cerebrospinal fluid α-tocopherol and selenium concentrations in neonatal foals
with neuroaxonal dystrophy. J. Vet. Intern. Med 29, 1667–1675. 10.1111/jvim.13618. [PubMed:
26391904]
Hidiroglou M and Karpinski K (1987) Vitamin E kinetics in sheep. Br. J. Nutr 58, 113–125. 10.1079/
bjn19870075. [PubMed: 3620433]
Hidiroglou M and Karpinski K (1988) Pharmacokinetic disposition in sheep of various vitamin E
preparations given orally or intravenously. Br. J. Nutr 59, 509–518. 10.1079/bjn19880060.
[PubMed: 3395610]
Equine Vet Educ . Author manuscript; available in PMC 2022 April 01.Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Donnelly et al. Page 8
Hidiroglou M and Karpinski K (1991) Disposition kinetics and dosage regimen of vitamin E
administered intramuscularly to sheep. Br. J. Nutr 65, 465–473. 10.1079/bjn19910105. [PubMed:
1878360]
Irías-Mata A, Sus N, Flory S, Stock D, Woerner D, Podszun M and Frank J (2018) α-Tocopherol
transfer protein does not regulate the cellular uptake and intracellular distribution of α- and γ-
tocopherols and -tocotrienols in cultured liver cells. Redox. Biol 19, 28–36. 10.1016/
j.redox.2018.07.027. [PubMed: 30098456]
Jørgensen E, Lazzarini G, Pirone A, Jacobsen S and Miragliotta V (2018) Normal microscopic
anatomy of equine body and limb skin: A morphological and immunohistochemical study. Ann.
Anat 218, 205–212. 10.1016/j.aanat.2018.03.010. [PubMed: 29730469]
Knight ME and Roberts RJ (1985) Tissue vitamin E levels in newborn rabbits after pharmacologic
dosing. Influence of dose, dosage form, and route of administration. Dev. Pharmacol. Ther 8, 96–
106. [PubMed: 3996166]
Krystel-Whittemore M, Dileepan KN and Wood JG (2016) Mast cell: A multi-functional master cell.
Front. Immunol 6, 1–12. 10.3389/fimmu.2015.00620.
Lodge JK, Hall WL, Jeanes YM and Proteggente AR (2004) Physiological factors influencing vitamin
E biokinetics. Ann. N. Y. Acad. Sci 1031, 60–73. 10.1196/annals.1331.006. [PubMed: 15753134]
Ronéus BO, Hakkarainen RVJ, Lindholm CA and Työppönen JT (1986) Vitamin E requirements of
adult Standardbred horses evaluated by tissue depletion and repletion. Equine Vet. J 18, 50–58.
10.1111/j.2042-3306.1986.tb03537.x. [PubMed: 3948831]
Snow DH and Guy PS (1976) Percutaneous Needle Muscle Biopsy in the Horse. Equine Vet. J 8, 150–
155. 10.1111/j.2042-3306.1976.tb03327.x. [PubMed: 976229]
Stowe HD (1968) A-Tocopherol Requirements for Equine Erythrocyte Stability. The American Journal
of Clinical Nutrition 21, 135–142.
Stuart RL, Kane E, Puschner B, Pusterla N, Collier J and Steidl S (2010) α-tocopherol concentrations
in equine serum and cerebrospinal fluid after vitamin E supplementation. Vet. Rec 166, 366–368.
10.1136/vr.b4802. [PubMed: 20305293]
Equine Vet Educ . Author manuscript; available in PMC 2022 April 01.Donnelly et al. Page 9
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Fig 1:
Serum α-tocopherol concentrations with individual animals plotted by treatment group.
Significant increase in concentration in 500 IU/mL injection compared to sentinel, oral and
600 IU injections at 24 h (P<0.0001), 48 h (P<0.0001) and 72 h (P<0.05). Red line denotes
the serum normal threshold 2 μg/mL
Equine Vet Educ . Author manuscript; available in PMC 2022 April 01.Donnelly et al. Page 10
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Fig 2:
CSF α-tocopherol concentrations pre- and 7-day postadministration with individual animals
plotted by treatment group. Significant increase in concentration in 500 IU/mL injection
compared to oral and 600 IU injections at 7 days (P<0.0001). Redline denotes the CSF
normal threshold 4 ng/mL
Equine Vet Educ . Author manuscript; available in PMC 2022 April 01.Donnelly et al. Page 11
Author Manuscript Author Manuscript Author Manuscript Author Manuscript
Fig 3:
Histological section at 200× magnification showing granulomatous inflammation
surrounding a mineralised core and surrounded by a rim of lymphoplasmacytic
inflammation and fibrosis. Numerous macrophages are multinucleated and often contain
clear, distinct vacuoles consistent with lipid. Clinical image of gross tissue reaction 24 h
postinjection (inset)
Equine Vet Educ . Author manuscript; available in PMC 2022 April 01.