Pediatr Radiol
DOI 10.1007/s00247-016-3726-4
ORIGINAL ARTICLE
Modern American scurvy — experience with vitamin C deficiency
at a large children’s hospital
Farahnaz Golriz 1 & Lane F. Donnelly 1 & Sridevi Devaraj 2 & Raj Krishnamurthy 1
Received: 9 June 2016 / Revised: 8 September 2016 / Accepted: 4 October 2016
# Springer-Verlag Berlin Heidelberg 2016
Abstract
Background Until recently scurvy has been viewed in devel-
oped countries as a disease of the past. More recently there
have been reports of case series of children with scurvy who
have had a delayed diagnosis after an extensive diagnostic
workup that included imaging. Most of these children have
had underlying neurologic conditions such as autism.
Objective To review the medical records of children diag-
nosed with vitamin C (ascorbic acid) deficiency based on
serum ascorbic acid levels at a large pediatric health care sys-
tem, to determine imaging findings and utility of imaging in
management, and to identify at-risk pediatric populations.
Materials and methods We retrospectively identified cases of
vitamin C deficiency in children tested for serum ascorbic acid
levels during the last 5 years. We used the criteria of normal
ascorbic acid >23 μmol/L and included children with ascorbic
acid levels <23 μmol/L. We evaluated their clinical history,
underlying medical condition, imaging studies obtained and
imaging findings.
Results We identified 32 children with vitamin C deficiency.
All of these children had underlying medical conditions, most
commonly iron overload from multiple transfusions related to
* Lane F. Donnelly
lfdonnel@texaschildrens.org
1
Department of Radiology, Texas Children’s Hospital,
6701 Fannin St., Suite 470, Houston, TX 77030, USA
2
Department of Pathology and Laboratory Medicine,
Texas Children’s Hospital, Houston, TX, USA
sickle cell anemia or thalassemia (20), neurologic disorders
(4) and bone marrow transplant/chemotherapy (3). No cases
of scurvy from dietary deficiency in otherwise normal chil-
dren were identified. All except two children had multiple
imaging studies, primarily related to their underlying condi-
tions. Three of these children had extensive imaging workups
related to diffuse musculoskeletal pain. Imaging findings in-
cluded ill-defined sclerotic and lucent metaphyseal bands
(mainly at the knee) on radiography and MRI studies that
showed diffuse increased T2-weighted signal in the bilateral
lower-extremity long-bone metaphyses, periosteal reaction
and adjacent soft-tissue edema.
Conclusion Vitamin C deficiency is not uncommon in large
pediatric health care facilities, and it is frequently missed on
clinical evaluation and diagnostic imaging. At-risk popula-
tions include those with iron overload, neurologic conditions
and history of chemotherapy. Scurvy related to dietary defi-
ciency in otherwise normal children was not encountered.
When characteristic MRI findings are seen, particularly in
children with a predisposing condition for vitamin C defi-
ciency, scurvy should be considered and a serum ascorbic
acid level checked to potentially confirm a diagnosis prior
to further invasive tests.
Keywords Children . Magnetic resonance imaging .
Malnutrition . Radiography . Scurvy . Vitamin C deficiency
Introduction
Scurvy has been a known diagnosis for centuries. It is related
to vitamin C (ascorbic acid) deficiency [1–11]. Its treatment

with citrus replacement was first described by James Lind in
1747 [1–4]. For most current radiologists, scurvy is nothing
more than a historical note. The classically described radio-
graphic signs of scurvy such as white line of Frankel, lucent
Trummerfeld zone, Wimberger ring (dense ring epiphysis)
and Pelkan spur (metaphyseal spur) [1–4] are not findings that
most current practicing radiologists in developed countries
have seen during clinical practice.
However there have recently been reports of case
series of children with scurvy with a delay in diagnosis
after an extensive diagnostic workup that included im-
aging [1–14]. Most of these children had underlying
autism and related restricted diets [1–14]. Having en-
countered a number of such cases in our own practice,
we reviewed the medical records of children who were
diagnosed with ascorbic acid (vitamin C) deficiency at a
large children’s health system to determine the use of
imaging studies in these children, the spectrum of im-
aging findings, and the populations at risk for vitamin C
deficiency.
Materials and methods
Our institutional review board approved this retrospective
review. We reviewed laboratory databases from a large
children’s hospital to identify all ascorbic acid (vitamin
C) assays performed over a 5-year period (2011–2015).
All ascorbic acid analyses were performed by quantitative
high-pressure liquid chromatography. We found cases of
ascorbic acid deficiency among the children in whom
ascorbic acid levels were obtained. At our institution,
ascorbic acid levels are categorized as: normal
>23 μmol/L, mild to moderate deficiency 11–23 μmol/L
and severe deficiency <11 μmol/L.
We reviewed cases of moderate and severe vitamin C
deficiency. For each child we reviewed the medical re-
cord for underlying medical conditions, and we obtained
and recorded details of imaging studies. Specifically, all
imaging examinations obtained for musculoskeletal
symptoms were reviewed for clinical indication, imaging
findings and recommendations. We then assessed the
utility of imaging for diagnosis, management and fol-
low-up.
Results
During the period 2011–2015, ascorbic acid levels were or-
dered in 151 children. Of these, 32 were abnormal. This
Pediatr Radiol
included 10 with moderate deficiency and 22 with severe de-
ficiency. There were 19 boys and 13 girls. The mean age was
11 years with a range of 4–19 years.
Underlying conditions
Concerning underlying patient conditions, 20 children
had iron overload related to multiple transfusions for
either sickle cell anemia or thalassemia. Four children
had underlying neurologic disorders; this included three
with autism and one with developmental delay. Three
other children were bone marrow transplant recipients
who had received chemotherapy. One child sustained
neglect and failure to thrive. The remaining children
had decreased oral intake related to hereditary fructose
intolerance (1), Henoch-Schönlein purpura (2), and
epidermolysis bullosa and dysphagia (1). No cases of
scurvy from primary dietary deficiency were identified
in otherwise healthy children.
Imaging findings
All but 2 of the 32 children had multiple imaging studies.
The indications for the imaging studies were related to the
underlying condition in each child, not the evaluation of
potential scurvy. Most of this imaging was of the chest,
abdomen or neurologic system. Only three children had
imaging of the extremities. None of the non-extremity
imaging studies demonstrated any specific findings that
were indicative of scurvy.
The three children who had imaging of the extremi-
ties were imaged specifically to evaluate diffuse lower-
extremity musculoskeletal pain. All three had severe vi-
tamin C deficiency and underlying neurologic disorders.
Two had autism and restricted dietary intake and one
had developmental delay and restricted dietary intake.
All three of these children had radiographs of the
knees and MRI of the lower extremities as part of their
imaging workup. On radiography, all three children had
ill-defined sclerotic and lucent metaphyseal bands that
were most prominent in the long bones at the knee
(Figs. 1, 2 and 3). The area immediately adjacent to
the physis showed a band of increased density consis-
tent with a widened zone of provisional calcification,
also known as the Frankel line. The immediately adja-
cent bone more distal from the physis showed a band of
lucency consistent with what has traditionally been de-
scribed as the lucent Trummerfeld zone [1–4]. The other
historically described classic radiographic findings of
scurvy such as Wimberger ring (dense ring around the
Pediatr Radiol

Fig. 1 Imaging in a 3-year-old

boy with developmental delay,
restricted diet related to food
selectivity, and lower-extremity
musculoskeletal pain. a
Anteroposterior radiograph of the
knee shows minimal changes
including dense band
(arrowheads) in the metaphysis
immediately adjacent to the
physis. These changes are most
prominent in the lateral aspect of
the femur. Immediately adjacent
to the dense band in a more
diaphyseal direction is a lucent
band (arrows). b Coronal T2-
weighted MR image shows
abnormal increased signal within
the metaphyses of the bilateral
distant femurs as well as proximal
and distal tibia. Note also
periosteal reaction (arrowheads)

epiphysis) and Pelkan spur (metaphyseal spur) were not
visualized. In all three children, wide-field-of-view MR
images were obtained of the lower extremities. The pro-
tocol included both short tau inversion recovery (STIR)
and T1-weighted coronal images from the level of the
hips to the angles. In all three children there was ab-
normally increased T2 signal in the marrow cavities of
the metaphyses, particularly of the bones adjacent to the
knee joint (Figs. 1, 2 and 3) as well as periosteal ele-
vation and edema. In two children there was also soft-
tissue edema adjacent to the involved bones. In these
three children the diagnosis of scurvy was not made
until after the imaging studies were obtained.
Discussion
Humans are one of the few animals that cannot synthesize
vitamin C and require exogenous sources for proper colla-
gen biosynthesis [1–4]. Serum ascorbic acid levels become
abnormal approximately 41 days after vitamin C is absent
from the diet. Clinical manifestations of scurvy develop in
approximately 1–3 months [1–4]. Clinical symptoms of
vitamin C deficiency include fatigue, malaise, loss of ap-
petite, follicular hyperkeratosis, cork screw hairs, petechial
hemorrhages, purpura and swollen or bleeding gums [2, 3].
Musculoskeletal symptoms are common and include bone
pain, limp and refusal to walk [2, 3].
The occurrence of scurvy related to poor dietary intake
of vitamin C was first described in sailors during the 14th
to 18th centuries [2, 3]. The description by James Lind of
citrus replacement to prevent the development of scurvy
was slowly adopted over centuries and led to a marked
reduction in the prevalence of scurvy [1–4]. Until recent-
ly, in developed nations the occurrence of scurvy in chil-
dren had become a historical footnote, with most radiolo-
gists having never actively encountered a case. More re-
cently a number of case reports and case series have de-
scribed the occurrence of scurvy in children with autism
and restricted diets related to food selectivity [1–14].
Findings in this study suggest that vitamin C deficien-
cy and scurvy are now not uncommonly encountered in
quaternary pediatric health systems. In the last 5 years at a
single pediatric health care system, 32 cases of vitamin C
deficiency were documented. In modern health care, there
are now several groups of children at risk. These include
children with iron overload related to multiple transfu-
sions, such as is seen with sickle cell anemia or thalasse-
mia; children with neurologic disorders, such as autism or
developmental delay; and children who are bone marrow
transplant and chemotherapy recipients. In this study, 20/
32 (62.5%) children with vitamin C deficiency had iron
 Pediatr Radiol

Fig. 2 Imaging in a 9-year-old

boy with autism and restricted
diet and lower-extremity pain. a
Anteroposterior radiograph of the
knee shows a dense band
(arrowheads) in the metaphysis
immediately adjacent to the
physis. Immediately adjacent to
the dense band in a more
diaphyseal direction is a lucent
band (arrows). b–d Coronal T2-
weighted MR images of the pelvis
(b), femurs (c) and tibias (d) show
abnormal increased signal within
the metaphyses of the long bones
of the lower extremity. There is
also increased signal in the
subtrochanteric regions of the
femurs. Note the periosteal
reaction (arrowheads). Also note
the abnormal increased signal in
soft tissues adjacent to the distal
femurs and tibias bilaterally

overload, 4/32 (12.5%) had neurologic disorders and 3/32
(9.4%) were bone marrow transplant recipients.
The mechanism for development of scurvy has been
well described in children with autism and other neurologic
disorders [2, 3]. Development of vitamin C deficiency is
related to longstanding food selectivity resulting in mark-
edly restricted diets [2, 3]. In children with bone marrow
transplants, development of vitamin C deficiency is related
to decreased oral intake related to mucusitis [15] and nau-
sea and vomiting. We also encountered vitamin C deficien-
cy associated with various causes of restricted dietary in-
take including fructose intolerance, Henoch-Schönlein
purpura and gastrointestinal wall hemorrhage, and dyspha-
gia secondary to epidermolysis bullosa. It is important to
note that no otherwise healthy children developed vitamin
C deficiency purely related to diet without an underlying
medical condition.
The etiology of vitamin C deficiency in children with iron
overload secondary to multiple transfusions is more complex.
Scurvy has been common among South African Bantus with
dietary hemochromatosis caused by consumption of large
amounts of homemade beer brewed in iron pots [16]. The
relationship between ascorbic acid deficiency and iron over-
load was further clarified by an investigation in which low
Pediatr Radiol
Fig. 3 Imaging in a 9-year-old boy with autism and restricted diet of only
oatmeal and refusal to walk for 1 month. a, b Coronal T1-weighted MR
images of the lower extremities show abnormal low T1-weighted signal
in the metaphyses of the long bones of the lower extremity. c, d Coronal
leukocyte ascorbic acid was found in subjects with idiopathic
hemochromatosis or transfusional siderosis [16]. In addition,
ascorbic acid deficiency was present in 64% of patients with
beta-thalassemia major, particularly in the older and more
transfused patients [17]. Most patients with iron overload
who develop clinical findings of scurvy also have an unbal-
anced diet [18–20]. The mechanism by which iron overload
depletes vitamin C is that massive ferric deposits accelerate
the catabolism of ascorbic acid [16, 21].
Most children recently diagnosed with scurvy had imaging
evaluation with both radiography and MRI [1–14]. This might
lead some to consider imaging a first-line tool in the evalua-
tion of a child with suspected scurvy. However most of the
children with ascorbic acid deficiency diagnosed by serum
laboratory tests in this study had multiple imaging studies
related to their underlying medical conditions. Only 3/32
(9.4%) had an imaging workup related to diffuse musculoskel-
etal pain. No other children had imaging of the extremities. In
the diagnostic sense, imaging does not play a central role in
making the diagnosis of scurvy or vitamin C deficiency. In a
child with diffuse musculoskeletal pain and gum or cutaneous
findings suggestive of scurvy, a serum laboratory test evalu-
ating for low ascorbic acid should be the first test of choice.
Imaging studies such as radiography or MRI might not
be needed.
However, imaging studies are often obtained in children
who present with diffuse musculoskeletal pain and either no
or unrecognized other findings of scurvy. In many of the
T2-weighted MR images show abnormal increased signal within the
metaphyses of the long bones of the lower extremity. Note periosteal
reaction (arrowheads). Also note the abnormal increased signal in soft
tissues adjacent to the distal femurs and tibias bilaterally
recently described cases in the literature as well as in our
personal experience, there has often been an extensive workup
and delay in diagnosis of scurvy [1–3]. In this sense, it is
important to recognize the imaging findings of scurvy and
raise the possibility of this diagnosis, particularly when imag-
ing a child in an at-risk group for vitamin C deficiency.
Concerning imaging findings on radiography, the classic
radiographic findings of scurvy were not encountered in any
of the cases of scurvy in autistic children recently described in
the literature [1–14] or in the cases encountered in this series.
Classically described radiographic findings of scurvy include
Pelkan spur, which represents a healing metaphyseal patho-
logical fracture, and Wimberger ring sign, which denotes a
thin sclerotic cortex surrounding a lucent epiphysis [1–4].
Periosteal new bone formation secondary to subperiosteal
hemorrhage is described as a classic radiographic finding
[1–4], and although this was not seen on radiography in this
series, periosteal new bone formation was depicted on MRI in
some of these cases. However in the metaphyses around the
knee, a dense band of provisional calcification (Frankel line)
immediately adjacent to the physis and an adjacent lucent
band more diaphyseal in location (lucent Trummerfeld line)
were identified both in the cases described here as well as in
others described in the literature [1–4].
In the children in this series, MRI demonstrated increased
T2-W signal in the marrow cavities of the metaphyses, partic-
ularly of the bones adjacent to the knee, increased periosteal
reaction, and significant soft-tissue edema adjacent to the

involved bones. These findings are consistent with what has
been described in case reports and case series in the literature
[1–14]. The MR imaging appearance of scurvy in children is
consistent and characteristic. The exact cause of the marrow
changes is not known [1–14]. Although multiple metaphyseal
signal abnormalities are nonspecific and can be seen with
hematological malignancies, other metastatic disease, osteo-
myelitis, the presence of bilateral increased lower-extremity
metaphyseal signal changes, periosteal reaction and adjacent
soft-tissue edema should raise the possibility of scurvy. This is
particularly true if the child has any of the described predis-
posing conditions. In the vast majority of recently described
cases of children diagnosed with scurvy, invasive procedures
such as bone marrow aspiration or bone biopsy was performed
[1]. In children with MRI findings consistent with scurvy,
and in particular those with an underlying predisposing
condition, a serum ascorbic acid level can be drawn to con-
firm a diagnosis of scurvy and more invasive diagnostic
testing can be avoided.
Limitations of this study include that it was retrospective in
nature and the experience seen at a single children’s health
care system. In addition, there were only a few children who
actually had extremity pain and associated imaging. Finally,
the reason why these children had serum ascorbic acid levels
drawn was not evaluated.
Conclusion
Vitamin C deficiency and scurvy are not uncommon at large
pediatric health care systems dealing with chronic disease. At-
risk populations include children with iron overload (such as
from multiple blood transfusions in sickle cell anemia or thal-
assemia); neurologic conditions (such as autism or develop-
mental delay), which are associated with food selectivity and
restricted diet; and decreased oral intake in children, such as
those who have had bone marrow transplant. We did not
encounter scurvy related to dietary deficiency in otherwise
normal children.
In the absence of extremity symptoms, imaging does not
play a central role in the workup of most of these populations
of children. In this series, only 9.4% had imaging studies to
evaluate musculoskeletal symptoms. However imaging stud-
ies such as radiography or MRI might be obtained in children
who present with diffuse musculoskeletal pain regardless of
whether there are findings of scurvy. Radiography most com-
monly shows a dense metaphyseal band immediately adjacent
to the physis with an immediately adjacent lucent metaphyseal
band. This is most common at the knee. MRI shows bilateral
increased T2-weighted signal in the metaphyses with associ-
ated periosteal reaction, and increased signal in the immedi-
ately adjacent soft tissues. When such imaging findings are
seen, particularly in children with a predisposing condition for
Pediatr Radiol
vitamin C deficiency, scurvy should be considered and a se-
rum ascorbic acid level checked prior to further invasive test-
ing such as bone marrow aspiration or bone biopsy.
Compliance with ethical standards
Conflicts of interest None
References
1. Gulko E, Collins LK, Murphy RC et al (2015) MRI findings in
pediatric patients with scurvy. Skelet Radiol 44:291–297
2. Ma NS, Thompson C, Weston S (2015) Brief report: scurvy as a
manifestation of food selectivity in children with autism. J Autism
Dev Disord 46:1464–1470
3. Weinstein M, Babyn P, Zlotkin S (2001) An orange a day keeps the
doctor away: scurvy in the year 2000. Pediatrics 108:e55
4. Polat AV, Bekci T, Say F et al (2015) Osteoskeletal manifestations
of scurvy: MRI and ultrasound findings. Skelet Radiol 44:1161–
1164
5. Brennan CM, Atkins KA, Druzgal CH et al (2012) Magnetic reso-
nance imaging appearance of scurvy with gelatinous bone marrow
transformation. Skelet Radiol 41:357–360
6. Gongidi P, Johnson C, Dinan D (2013) Scurvy in an autistic child:
MRI findings. Pediatr Radiol 43:1396–1399
7. Khan N, Furlong-Dillard JM, Buchman RF (2015) Scurvy in an
autistic child: early disease on MRI and bone scintigraphy can
mimic an infiltrative process. BJR Case Rep. doi:10.1259
/bjrcr.20150148
8. Niwa T, Aida N, Tanaka Y et al (2012) Scurvy in a child with
autism: magnetic resonance imaging and pathological findings. J
Pediatr Hematol Oncol 34:484–487
9. Choi SW, Park SW, Kwon YS et al (2007) MR imaging in a child
with scurvy: a case report. Korean J Radiol 8:443–447
10. Alqanatish JT, Alqahtani F, Alsewairi WM et al (2015) Childhood
scurvy: an unusual cause of refusal to walk in a child. Pediatr
Rheumatol 13:23
11. Estienne M, Bugiani M, Bizzi A et al (2011) Scurvy hidden behind
neuropsychiatric symptoms. Neurol Sci 32:1091–1093
12. Harknett KM, Hussain SK, Rogers MK et al (2014) Scurvy mim-
icking osteomyelitis: case report and review of the literature. Clin
Pediatr 53:995–999
13. Popovich D, McAlhany A, Adewumi AO et al (2009) Scurvy:
forgotten but definitely not gone. J Pediatr Health Care 23:405–415
14. Kitcharoensakkul M, Schulz CG, Kassel R et al (2014) Scurvy
revealed by difficulty walking. Three cases in young children. J
Clin Rheumatol 20:224–228
15. Kletzel M, Powers K, Hayes M (2014) Scurvy: a new problem for
patients with chronic GVHD involving mucous membranes; an
easy problem to resolve. Pediatr Transplant 18:524–526
16. Wapnick AA, Lynch SR, Krawitz P et al (1968) Effects of iron
overload on ascorbic acid metabolism. Br Med J 3:704–707
17. Chapman RW, Hussain MA, Gorman A et al (1982) Effect of ascor-
bic acid deficiency on serum ferritin concentration in patients with
beta-thalassaemia major and iron overload. J Clin Pathol 35:487–
491
18. Cohen A, Cohen IJ, Schwartz E (1981) Scurvy and altered iron
stores in thalassemia major. New Engl J Med 304:158–160
Pediatr Radiol

19. Prakash A, Pandey AK (2015) Joint effusions and purpura in
multiply-transfused adult beta-thalassemia-clinical pointers to diag-
nosis of scurvy. Kathmandu Univ Med J 11:360–362
20. Ray M, Marwaha RK, Sethuraman G et al (1999) Scurvy in trans-
fusion dependent beta-thalassemia. Indian Pediatr 36:504
21. Lynch SR, Seftel HC, Torrance JD et al (1967) Accelerated oxida-
tive catabolism of ascorbic acid in siderotic Bantu. Am J Clin Nutr
20:641–647