Original Article
Low Plasma Zinc and Iron in Pica
Sunit Singhi, R. Ravishanker, Pratibha Singhi and R. Nath1
Departments of Pediatrics and 1Biochemistry, Postgraduate Institute of Medical Education and Research,
Chandigarh, India.
Abstract. Objective : To determine role of trace elements in causation of pica with specific reference to zinc and iron we
studied plasma levels of iron (Fe), Zinc (Zn), calcium (Ca) and magnesium (Mg) and blood lead (Pb) levels by atomic absorption
spectrophotometer in 31 children with pica (Pica Group) and 60 controls matched for age, sex and nutrition (Control Group)
in an observational case and control study in the settings of outpatient clinic of a tertiary care, teaching hospital. Methods :
Data from each group were further stratified by hemoglobin level <9 and >9 g/dl into two subgroups pica-1 and pica-2, and
control-1 and control-2 respectively, to control for confounding effect of iron deficiency anemia. Results : The plasma Fe level
(mean ± SD) in children with pica (42.7 ± 9.2) mg/dl) was about 20% lower than that in controls (51.5 ± 10.0 mg/dl, p<0.001).
Plasma Zn levels in the pica group (60 ± 4.4 mg/dl) was about 45% lower than those in controls (110.2±8.5 mg/dl, p<0.001).
Correlation of Zn and Fe levels with pica-related variables such as age at onset, duration and frequency and number of inedible
objects ingested was not significant. Conclusion : These findings suggest that hypozincemia with low iron levels may be the
possible cause of pica and contradict the contention that low levels of plasma Zn and Fe could be an effect of pica.
[Indian J Pediatr 2003; 70 (2) : 139-143]
Key words : Iron; Iron-deficiency; Pica; Zinc; Lead
Pica is defined as persistent or compulsive ingestion of
inedible or non-nutritive substances. 1,2 It has been
reported from all parts of the world2 with prevalence
varying from 103 to 32.5%4 of all children surveyed, and
upto 73% of school children.5 It is more prominent in
younger children.3,5 The adverse effects of pica range from
parasitic infestation 6 anemia 7 and trichobezoar and
intestinal obstruction7,9 to life threatening hypokalemia10
and lead and other poisoning.11, 12 The etiology of pica is
still controversial. 13 Iron (Fe) deficiency anemia, 14-15
hypozincemia, 16,17 psychosocial stress in presence of
anemia,18 and family distress-especially maternal19-have
all been presented as causes. Very little systematic
research has been conducted on the association of pica
with the plasma/blood levels of iron (Fe), zinc (Zn),
copper (Cu), magnesium (Mg), calcium (Ca) and lead
(Pb). We have studied the relation between pica and
plasma levels of Fe, Zn, Ca and Mg and blood Pb levels in
children, while controlling for the confounding effect of
hemoglobin levels.
MATERIALS AND METHODS
The study was conducted in the Department of Pediatrics,
Nehru Hospital, Postgraduate Institute of Medical
Education and Research, Chandigarh, India, over a period
of one year. Informed consent was obtained from the
parents of the subjects prior to enrolment in the study.
Reprint requests : Dr. Pratibha D. Singhi, Professor, Department of
Pediatrics, PGIMER, Chandigarh-160012, India.
Fax : 0091-172-744401, 745078. E-mail : drsinghi@glide.net.in
Indian Journal of Pediatrics, Volume 70—February, 2003
Study Population
Thirty one children with pica (Pica-group) between 18
months to 48 months of age who had been ingesting
inedible substances three or more times per week for 3
months or longer were studied. Sixty children within
same age range and gender distribution without pica
served as control group. The control and pica groups were
drawn from children brought to the outpatient service for
immunization and were free from any acute illness or
symptoms. Children with evident developmental delay,
chronic illness and thalassemia major were excluded from
the study. Children with moderate and severe
malnutrition (weight for age <70%) were also excluded as
malnutrition is associated with significant changes in
plasma Zn and Fe levels.20 Sociodemographic data and
details of pica viz.,-the objects eaten, frequency of
ingestion, amount ingested each time, associated
symptoms and relevant medial history-were recorded on
a predesigned form. A detailed physical examination and
hemoglobin estimation was done in all the cases.
Estimation of Trace Elements and Minerals
Blood (3 ml) was collected from a peripheral vein into
trace element-free, heparinized plastic vials using
disposable plastic hubbed needles. One milliliter of
heparinzed whole blood was set aside in a trace elementfree vial; the remainder was centrifuged at 3000 rpm for
10 minutes. The separated plasma was stored until
analysis in another trace element-free vial at 4oC, as were
the whole blood samples. Samples were digested using
139
Sunit Singhi et al
the standard method adopted in our laboratory. 21
Plasma Zn and Cu levels were determined with the
help of an atomic absorption spectrophotometer (Perkin
Elmer model 4000), while a direct current photometer
(Spectrospan V1, Beckman) was used for estimation of
plasma Fe, Mg, Cu and blood Pb levels. The
spectrophotometer was calibrated with the standards
obtained from Sigma Co. USA. The reference materials
that were used to check the accuracy of analytical
method were obtained from International Atomic Energy
Institute, Vienna, Austria and National Bureau of
Standards, USA. The accuracy for Pb, Zn, Fe and Mg was
above 99%. The precision of the method for Pb, Zn and Fe
was upto 1 mg/dl and for Mg 0.1 mg/dl.
Data Analysis
Data obtained on Pica and Control was compared using
chi-square test for categorical variables and t-test and
analysis of variance (ANOVA) for continuous variable.
Within each group data was further categorized into two
subgroups i.e. Pica-1 and Pica-2 and Control-1 and
Control-2 based on hemoglobin levels. Pica-1 and
Control-1 group had hemoglobin levels <9 g/dl and Pica2 and Control-2 had levels >9 g/dl. Correlation
coefficients (Pearson's least square) were calculated to
look for possible interrelationship between plasma Fe and
Zn levels on one hand and age at onset of pica and
duration of pica on the other. SPSS version 7.5 was used
for data analysis.
Results
The groups and the subgroups were comparable with
respect to age and sex distribution and average weights
(expressed as percentage of the weight expected for age
TABLE 1. Clinical Data in Different Groups
Group
Number
Sex Ratio
(M : F)
Age
(months)
Weight
(%)
Hemoglobin
(g/dl)
Pica
31
21:10*
88.2±7.8†
Control
60
39:21
Pica-1
21
14:7
Pica-2
10
7:3
Control-1
22
14:8
Control-2
38
25:13
27.3 ± 10.3†
(18-46)
29.4 ± 8.0
(18-48)
27.0 ± 10.3
(18-46)
27.9 ± 9.6
(18-44)
27.1 ± 8.3
(18-48)
30.7 ± 8.9
(18-48)
8.5±1.3
(6.5-11.5)
9.4±1.1
(7.2-12.1)
7.8±0.8
(6.5-8.9)
10.0±0.8
(9.1-11.5)
8.2±0.4
(7.2-8.9)
10.0±0.8
(9.0-12.0)
86.3±6.1
88.4±7.9
87.1±8.0
86.1±6.3
86.3±6.0
Figures in parentheis represent range
*P = n.s., by χ2-test, between Pica and Control group
†P=n.s., by t-test, between Pica and Control groups.
TABLE 2. Age of Onset and Duration of Pica, Number and Frequency of Non-nutritive Substances Ingested in Children with Pica
Pica characteristics
A. Age of onset (months)
<12
12-24
>24
B. Duration (months)
3-12
13-24
>24
C. No of substances eaten
1
2-3
4 and more
D. Frequency of ingestion
< 3/day
3-5/day
<5/day
All
n=31
Pica-1
n=21
Pica-2
n=10
3 (9.6)
26 (84)
2 (6.4)
3 (14)
17 (81)
1 (5)
–
9 (90)
1 (10)
23 (74)
6 (19)
2 (6.5)
15 (71)
5 (25)
1 (5)
8 (80)
1 (10)
1 (10)
13 (42)
17 (55)
1 (3)
9 (43)
11 (52)
1 (5)
4 (40)
6 (60)
4 (13)
21 (68)
6 (19)
1 (5)
15 (71)
5 (24)
3 (30)
6 (60)
1 (10)
*Percentages in brackets
140
Indian Journal of Pediatrics, Volume 70—February, 2003
Low Plasma Zinc and Iron in Causation of Pica
and sex). On physical examination no abnormal sign
was detected in any of the study subjects except the
presence of pallor in children with hemoglobin <9 g/dl.
Age at onset, duration of pica, number of substances
eaten and frequency of ingestion in children with pica is
shown in Table 2. The mean ± SD age at onset of pica was
16.4±7.2 months and the duration 10.8±7.7 months. Both
pica subgroups were comparable in this respect. Eightfour percent were 1-2 years old at the onset of pica, 74%
had pica for 3-12 months and 81% ingested inedible
substances up to five times per week. All the children
with pica had geophagy; other inedible substances
ingested were bricks, cement, ash, plaster, coal and soap.
The mean plasma Fe level in children with pica was
about 20% lower than that in the controls (P<0.001) Table
3; however, mean Fe levels were similar in Pica and
control subgroups made on the basis of hemoglobin level.
The mean plasma Zn level in children with pica was
significantly lower than that of controls (p<0.001, t-test)
Table 3; the difference persisted between subgroups
pica-1 and control-1 and pica-2 and control-2 (p<0.001;
t-test). The Zn levels were similar in subgroup pica-1
and pica-2 and control-1 and control-2 Table 3. Plasma
CU, Mgn, Ca, and blood Pb levels were similar in both
the gropus and their subgroups (Table 3).
Plasma Fe and Zn levels in Pica group did not correlate
with the age at onset (r=0.99 and 0.126 with Fe and Zn
respectively) or duration of pica. (r=0.009) with Fe and Zn.
Plasma Zn and Fe levels did not differ significantly with
the number of inedible substances ingested or frequency
of ingestion (Table 4).
DISCUSSION
Mean hemoglobin and plasma Fe levels were significantly
lower in children with pica compared to controls.
Others14,14,22-27 have reported similar findings. In view of
the overwhelming consensus in the literature on low
TABLE 3. Plasma Iron, Zinc, Copper, Magnesium, Calcium and Blood Lead in Different Groups (Mean ± SD)
Iron
mg/dl
Zinc
mg/dl
Copper
mg/dl
Magnesium
mg/dl
Calcium
mg/dl
Lead
mg/dl
Pica
(n-31)
Controls
(n-60)
Pica-1
(n-21)
Control-1
(n-22)
Pica-2
(n-10)
Control-2
(n-38)
42..7±9.2
(30-67)
60.0±4.4
(53-70)
123.2±8.4
(110-140)
2.4±0.5
(1.9-3.9)
10.0±0.2
(9.2-11.2)
26.9±5.0
(17.6=34)
51.5±10
(33-76)
110.2±8.5
(96-130)
124.0±6.8
(105-140)
2.3±0.3
(1.7-3.0)
10.2±0.5
(9.1-11.1)
27.8±3.2
(18.4-32)
37.6±5.2*
(30-50)
60.2±4.5†
(53-70)
122.5±9.4
(110-140)
2.4±0.3
(1.9-3.0)
10.0±.6
(9.1-11.2)
27.5±5.0
(17.6-3.4)
39.7±4.3
(33-45)
107.6±7.6
(98-118)
123.5±6.8
(112-140)
2.4±0.3
(1.7-2.8)
10.1±0.4
(9.3-11.0)
28.6±2.1
(24-32)
53.5±5.5*
(47-67)
59.5±4.6†
(54-79)
124.6±7.8
(113-140)
2.5±0.5
(2.0-3.9)
9.9±0.3
(9.4-11.0)
25.5±5.0
(18.9-32)
57.8±6.2
(50-76)
111.7±8.7
(96-130)
124.3±6.8
(105-136)
2.3±0.3
(1.8-3.0)
10.3±0.6
(9.1-11.1)
27.6±3.8
(18.4-32)
Figures in parenthesis represent the range
*P<0.05; by t-test, as compared to corresponding subgroup within the same group[.
†P< 0.001; by t-test, as compared to correspond controls
TABLE 4. Plasma Iron and Zinc in Children with Pica Related to Frequency and Number of Non-nutritive Substances Eaten
Frequency of ingestion
Iron in
µg/dl
Zinc in
µg/dl
No. of non-nutritive substances eaten
< 3/day
3-5/day
>5/day
1
2–3
>3
P1
30.0
(n-1)
38.0±5.0
(n-15)
38.4±5.8
(n-5)
38.0±3.9
(n-9)
36.6±4.3
(n-11)
45.0
(n-1)
P2
51.7±7.6
(n-3)
53.0±3.9
(n-6)
62.0
(n-1)
52.0±7.7
(n-4)
52.8±5.9
(n-6)
-
All
46.2±12.5
(n-4)
42.2±8.5
(n-21)
42.3±10.9
(n-6)
42.3±9.1
(n-13)
42.9±9.8
(n-17)
45.0
(n-1)
P1
56.0
59.2±8.2
59.4±5.0
60.3±5.5
60.1±3.8
60.0
P2
All
61.0±7.0
59.8±6.2
59.5±3.4
59.3±7.1
56.0
58.8±4.7
56.5±2.5
59.1±5.1
60.5±5.6
60.9±4.1
60.0
*P= not significant, by ANOVA, for plasma iron or zinc level with respect to the frequency or the number of substances eaten within Pica group
or each subgroup
Indian Journal of Pediatrics, Volume 70—February, 2003
141
Sunit Singhi et al
levels of hemoglobin and Fe in pica, we decided to
control for the confounding effect of these variables
while studying plasma levels of Zn, Cu, Mg, Ca and
blood Pb levels by stratifying data on the basis of
hemoglobin level.
The most significant finding of the study was that the
mean plasma Zn level in children with pica was almost
half that of controls, suggesting a strong association
between hypozincemia and pica. The first report of
hypozincemia in pica came from Prasad et al.16 They
described syndrome of geophagy, Fe deficiency anemia,
hepatosplenomegaly, growth retardation and hyogonadism in association with hypozincemia, which was
later confirmed by Okcuolgu et al 24. Since, then, more
authors have reported low levels of Zn in pica 17,28-30 but
little attention had been paid to Zn after systematically
controlling for hemoglobin and Fe levels. We found that
plasma zinc levels were lower in children with pica
compared to controls with similar hemoglobin and iron
levels.
Do low plasma levels of Zn and/or Fe cause pica or is
it an effect of pica? Controversy surrounds the answer to
this question. Some investigators feel that pica results in
malabsorption of Fe and/or Zn from the diet16, 24,25,31,32
while others believe that Fe deficiency 13,14,27,33 and Zn
deficiency17,28,29 can cause pica. If indeed pica causes Fe
and/or Zn deficiency plasma Fe and Zn levels should
correlate positively with age at onset of pica and
negatively with duration of pica. Further, levels of these
trace elements would be lower in children who ingested
more non-nutrient substances more frequently. We did
not find such as association. The correlation between
plasma Fe and Zn and age at onset and duration of pica
was not significant; as also of Fe and Zn. In a recent study,
induction of zinc deficiency in sheep led to pica, then pica
so induced relieved Zn-deficiency.34 In our opinion, these
observations support the hypotheis that low levels of Fe
and Zn are the cause of pica, and not its effect.
Plasma levels of Cu and Mg in otherwise healthy and
mentally normal children with pica were studied perhaps
for the first time. These were within normal limits in both
the groups and their subgroups. Earlier studies on
mentally retarded adults 33 and children with iron
deficiency anemia, geophagy, hepatosplenomegaly,
dwarfism and hypogonadism 24 have yielded variable
results. Ca levels in children with pica have not been
reported in the literature. In a previous study we found
that rickets was significantly more prevalent in children
with pica for plaster or chalk than for other substances.35
Clinically overt rickets was not seen in any of our
children, and plasma Ca levels were similar and within
normal range in all children. High Pb levels in children
with pica have been traced to painted toys, discarded
battery castings, paint peelings from ceilings and window
sills, house dust and soil contaminated with Pb.6,12,36
However, none of the children in this study had high Pb
levels, probably because they did not ingest any
142
substances rich in Pb.
To conclude, we found that children with pica had low
level of plasma Zn and Fe, which in our opinion was the
likely cause of their pica. These data should from the basis
for an intervention trial comparing efficacy of Fe and Zn
therapy for treatment of pica.
REFERENCES
1. Scot C, Datton R. Vegetative disorders. In Behrman RE,
Kliegman RM, Jenson HB, eds. Nelson Textbook of Pediatrics,
16th edn. Philadelphia : WB Saunder 2000; 72.
2. Cooper M. Pica Springfield, Charles C Thomas, 1957.
3. Barltrop D. Prevalence of pica. Am J Dis Cild 1966; 112 : 116123.
4. Millican FK, Layman EM, Lowrie RS, Takashashi L, Dublin
CC. Prevalence of ingestion and mouthing of non-edible
substances by children. Clin Proc Child Hosp Wash 1962; 18 :
207-214.
5. Geissler PB, Mwaniki DL, Thiong’o F, Friis H. Geophagy
among school children in western Kenya. Trop Med Int Health
1997; 7: 624-630.
6. Glickman TL, Chaudry IV, Constantino J, Clark FB, Cypess
RH, Winslow L. Pica patterns, toxocariasis and elevated blood
lead levels in children. Am J Trop Med Hyg 1981; 30: 77-80.
7. Crosby WH Pica : compulsion caused by iron deficiency. Br J
Haematol 1976; 34 : 341-342.
8. McGehee FT Jr, Buchanan GR. Trichoghagia and trichobezoar:
etiologic role of iron deficiency. J Pediatr 1980; 97 : 946-948.
9. Phillips MR, Zaheer S, Drugas GT. Gastric trichobezoar: case
report and literature review. Mayo Clin Proc 1998; 7 : 653-656.
10. Gonzaelez JJ, Owens W, Ungaro PC, Werk EE, Hurtz PH. Clay
ingestion : a rare cause of hypokalemia. Ann Intern Med 1982;
97: 65-66.
11. Menge H, Lang A, Cuntze H. Pica in Germany-amylophagia as
the etiology of iron deficiency anemia. Z Gastroenterol 1998; 8:
635-640.
12. Jacobziner H. Lead poisoning in childhood : epidemiology,
manifestation and prevention. Clin Pediatr 1966; 5: 277-286.
13. Federman DG, Kirsner RS, Federman GS. Pica : are you
hungry for the facts ? Conn Med 1997; 4: 207-209.
14. Lanzkowsky P. Investigation into the aetiology and treatment
of pica. Arch Dis Child 1959; 34 : 140-148.
15. Munoz JA, Marcos J, Risueno CE, de Cos C, Lopez R, Capote
FJ et al. Iron deficiency and pical. Sangre (Barc) 1988; 43 : 31-34.
16. Prasad AS, Miale A, Farid Z, Sansted H, Schuler AR. Zinc
metabolism in patients with the syndrome of iron deficiency
anemia, hepatosplenomegaly, dwarfism and hypogonadism. J
Lab Clin Med 1963; 61 : 537-549.
17. Hambidge MK, Silverman A. Pica with rapid improvement
after dietary zinc supplementation. Am J Clin Nutr 1971; 24 :
1021-1022.
18. Singhi S, Singhi P, Adwani CB. Role of psychosocial stress in
the etiology of pica. Clin Pediatr 1981; 20 : 783-785.
19. Bithoney WG, Synder J, Michalek J, Newberger EH.
Childhood ingestions as a symptom of family distress. Am J
Dis Child 1985; 139: 456-459.
20. Kumar S, Rao KSJ. Plasma and erythrocytic zinc levels in
protein caloric malnutrition. Nutr Metabol 1973; 15: 364-371.
21. Nath R. Proceedings of integrated programme on heavy
metals. Training camp-cum-workshop on analysis and quality
control of heavy metals in biological samples. Chandigarh,
PGIMER 1986; 18-20.
22. Gutelius MF, Millican FK, Layman EM, Cohen GJ, Dubin CC.
Nutritional studies of children with pica : I. Controlled study
Indian Journal of Pediatrics, Volume 70—February, 2003
Low Plasma Zinc and Iron in Causation of Pica
23.
24.
25.
26.
27.
28.
29.
evaluating nutritional status II. Treatment of pica with iron
given intramusculary. Pediatrics 1962; 29: 1012-1023.
Halstead JA. Geophagia in man : Its nature and nutritional
effects. Am J Clin Nutr 1968; 21: 1384-1393.
Okcuoglu CA, Arcasoy A, Minnich V, Tarcon Y, Clin S,
Yorukoglu O et al. Pica in Turkey I. The incidence and
association with anemia. Am J Clin Nutr 1966; 19 : 125-131.
Minnich V, Okcuoglu CA, Tarcon Y, Arcasoy A, Clin S,
Yorukoglu O et al. Pica in Turkey II. Effect of clay on iron
absorption. Am J Clin Nutr 1968; 21 : 78-86.
Johnson NE, Stephens Dl. Geomelophagia. Unusual pica in
iron deficiency anemia. Am J Med 1982; 73 : 931-932.
Shapiro MD, Linas SL. Sodium chloride pica secondary to iron
deficiency anemia. Am J Kidney Dis 1985 ; 5:67-68.
Bhalla JN, Khanna PK, Srivastava JR, Sur BK, Bhalla M. Serum
zinc levels in pica. Indian Pediatr 1983; 19: 615-618.
Chen XC, Y In TA, He HS, Ma QY, Han Zm, Li LX. Low levels
of zinc in hair and blood, pica, anorexia and poor growth in
Indian Journal of Pediatrics, Volume 70—February, 2003
Chinese pre-school children. Am J Cli Nutr 1985; 45: 694-700.
30. Sipahi T, Akar N et al. Plasma zinc levels in patients with irondeficiency anemia. J Trop Pediatr 2001; 2: 122-123.
31. Thomas FB, Falko JM, Zuckerman K. Inhibition of intestinal
iron absorption by laundry starch. Gastoenterology 1976; 71 :
1028-1032.
32. Arcasoy A, Cavadar AO, Babacan E. Decreased iron and zinc
absorption in Turkish children wih iron deficiency and
geophagia. Acta Haematol 1978; 60: 76-84.
33. Danford DE, Huber AM. Pica among mentally retarded adults.
Am J Ment Defic 1982; 87 : 141-146.
34. Kendall NR, Telfer SB. Induction of zinc deficiency in sheep
and its correction with a soluble glass bolus containing zinc.
Vet Rec 2000; 22: 634-637.
35. Singhi P, Singhi S. Pica type of non-food articles eaten by
Ajmer children and their significance. Indian J Pediatr 1982; 49:
681-684.
36. Kumar A, Dey PK, Singla PN, Ambasht RS, Upadhyay SK.
Blood lead levels in children with neurological disorders. J
Trop Paediatr 1998; 44 : 320-322.
143
Sunit Singhi