Journal of Medical Nutrition and Nutraceuticals

: 2014  |  Volume : 3  |  Issue : 2  |  Page : 89--93

Role of conjunctival imprint cytology in detecting vitamin A deficiency in cancer patients: A case-control study

Chaparala Padmini1, Prayaga Aruna Kumari1, Digumarti Raghunadha Rao2,  
1 Department of Pathology , Nizam's Institute of Medical Sciences, Hyderabad, Andhra Pradesh, India
2 Department of Medical Oncology, Nizam's Institute of Medical Sciences, Hyderabad, Andhra Pradesh, India

Correspondence Address:
Prayaga Aruna Kumari
Department of Pathology, Nizam«SQ»s Institute of Medical Sciences, Punjagutta, Hyderabad 500 082, Andhra Pradesh


Context: Conjunctival Imprint cytology is widely used to detect vitamin A deficiency in field studies. Vitamin A deficiency is known to be associated with malignancies. Aims: To assess the vitamin-A status in cancer patients using conjunctival impression cytology technique (CICT) and to correlate the results with serum levels of the vitamin. Settings and Design: To study CICT in freshly detected cancer patients. To compare with normal controls, and to correlate the results obtained by the cytology technique with a serum retinol by HPLC method. Materials and Methods: Patients and their family members accompanying the patients were taken as subjects for the study and after an informed consent. Conjunctival imprint cytology samples and venous blood for serum retinol were collected from both groups. Statistical Analysis Used: Odds ratio, Pearson Chi-square test, Fisher exact test, analysis of variable, independent and dependent sample t test, mean and standard deviation. Data: Of 1551 subjects analyzed, vitamin A level < 20 mcg/dl was observed in 395 subjects; 322 (81.5%) were patients and 73 (19.5%) were controls (P < 0.001). CICT grades ≥ 2 was found in 357 subjects including 285 (79.8%) patients and 72 (20.2%) controls (P < 0.001). CICT had sensitivity 93.0%, specificity of 95.6%, positive predictive value 86.1%, negative predictive value 92.9%, Cohen«SQ»s kappa value 0.74, and P < 0.001. Conclusions: CICT is a reliable technique to detect vitamin A deficiency in cancer patients.

How to cite this article:
Padmini C, Kumari PA, Rao DR. Role of conjunctival imprint cytology in detecting vitamin A deficiency in cancer patients: A case-control study.J Med Nutr Nutraceut 2014;3:89-93

How to cite this URL:
Padmini C, Kumari PA, Rao DR. Role of conjunctival imprint cytology in detecting vitamin A deficiency in cancer patients: A case-control study. J Med Nutr Nutraceut [serial online] 2014 [cited 2021 Dec 4 ];3:89-93
Available from:

Full Text


Vitamin A (C 20 H 30 O) and related carotenoid compounds are of crucial importance to health. They are important for growth and differentiation of a variety of cells, mainly mucosa-associated. It is essential for vision and has effects on retina, cornea, and conjunctiva. Hypovitaminosis A is an important cause for preventable blindness in children of the developing countries. [1] Population-based studies on vitamin A deficiency in children are done by clinical evaluation and conjunctival impression cytology (CIC). Conjunctival impression cytology is the technique of collection of the most superficial layers of the ocular surface by using cellulose acetate filter paper. Cells adherent to the surface are subsequently transferred from the filter paper on to a glass slide and processed further. It is fast, non-invasive, easy to perform, and economical test. There are not many studies on the utility of conjunctival cytology in adults. To the best of our knowledge, this is the only study in cancer patients.

Vitamin A deficiency is reported in several malignancies. In clinical practice, the use of all-trans retinoic acid (ATRA) as the treatment modality has revolutionized the management of acute promyelocytic leukemia (APML), which used to be a rapidly fatal disease and bexarotene for cutaneous T cell lymphomas.

The method selected was conjunctival impression cytology with transfer (CICT), and the results were correlated with serum levels of the vitamin by high performance liquid chromatography (HPLC) technique as the gold standard. The vitamin status of the patients was compared with appropriate controls both by CICT and serum vitamin levels to find out the differences between the controls and patients and also among the patients with a variety of cancers.

This is a descriptive, prospective, correlative, case-control, diagnostic, non-therapeutic study whereby conjunctival impression cytology was used to assess vitamin A status in cancer patients and their family members.

 Materials and Methods


Blood samples and conjunctival impression cytology with transfer (CICT) samples were collected from patients who were freshly diagnosed with cancer and accompanying healthy kin as controls. An informed consent was obtained from the study subjects. Smears were collected from both eyes. Sample was considered adequate if cells were present in either of the samples.

Technique for obtaining CICT

Cellulose acetate filter paper with pore size 0.2-0.4 μ was cut into approximately 20 Χ 4 mm size strips. Filter paper was pressed gently on bulbar conjunctiva of the subjects for 4-6 seconds. [2] Filter paper removed with a gentle peeling motion, and the material was transferred on to the clean labeled glass slide with gentle pressure on the reverse of the filter paper. Procedure was repeated for the second eye.

Glass slides were fixed immediately in alcohol-based fixative. [3] The fixative was prepared with 75 ml of 95% ethyl alcohol, 25 ml distilled water, 5 ml glacial acetic acid, and 5 ml of 37% formaldehyde. Smears were stained with periodic acid Schiff (PAS) stain. Normal conjunctival smears show sheets of polygonal cells with a nuclear cytoplasmic ratio of 1:2 and a basophilic cytoplasm. Interspersed with these cells are numerous goblet cells, and the background shows droplets of mucin. PAS-stained smears were graded by Nelson's grading based on the presence of goblet cells, mucin, and presence or absence of squamous cells from grade 0 to 3. Grades 0 and 1 are considered normal; grades 2 and 3 are considered xerosis [Figure 1]. [4]{Figure 1}

Serum vitamin A estimation

A 2 ml of blood sample was collected by venepuncture, and the sample was allowed to clot at room temperature. Serum separated was aliquoted into plastic storage (Eppendorf) vials at -20 o C for further analysis. Estimation was done by HPLC method using Thermo Fennigan instrumentation. [5]

Vitamin A levels <20 μg/dl were considered hypovitaminosis and ≥20 μg/dl as normal. [6]

Statistical analysis

Data on epidemiological and vitamin A, CICT parameters from the patients and controls were coded and tabulated in separated Microsoft excel work sheets for further analysis. Thereafter, appropriate statistical tools were used to analyze the data using SPSS 15 software. The percentage mentioned in the parenthesis.


The study has clearance from the institutional ethics committee.


Statistical methods used were Odds ratio, Pearson Chi-square test, Fisher exact test, analysis of variable, independent and dependent sample t test, mean and standard deviation.


There were 1728 subjects with 980 recently diagnosed, untreated cancer patients and 748 controls, the kin who accompanied the patients. The types of malignancies included hematologic and also solid tumors. The youngest patient was 3-years-old and the oldest 85-years-old. The clinical diagnosis included a wide variety of cancers. Youngest control was 15 years of age and oldest 66 years. The mean age in patient group was 42.1 ΁ 13.9 years and 36.84 ΁ 10.26 years in controls. There were 44 patients and 8 controls below the age of 18 years. Majority of the subjects were between 20-59 years old. A total of 238 (13.8%) subjects were >60 years, of these, 23 (3.1%) were controls. Amongst the patients, 578 were males and 402 were females. Amongst the healthy controls, 536 were male and 212 were females. Male:Female ratio was 1 4:1 in patients and 2.5:1 in controls.

Serum vitamin A levels in patients and controls were estimated in 959 (97.9%) and 736 (98.4%), respectively [Table 1]. Vitamin A levels <20 mcg/dl were detected in 395 samples (23.3%); of them, 322 (81.5%) were from patient group and 73 (18.5%) from controls [Table 2]. Samples could not be collected in 33 subjects (1.9%); of them, patients were 21 and controls 12. Patients with vitamin A levels <20 and ≥20 mcg/dl were grouped into two categories [Table 3]. Vitamin A levels were found to be <20 mcg/dl in 33.5% of patients and 9.9% of controls with P < 0.001.{Table 1}{Table 2}{Table 3}

CICT samples were collected in 1664 subjects; 961 (98.1%) patients and 733 (98%) controls. CICT grades ≥2 were found in a total of 238 subjects; 185 patients (21.4%) and 53 controls (7.4%) with P < 0.001 [Table 4]. In 118 subjects, the smears were acellular [Table 1]. CICT was not available in 34 subjects (2.0%); 19 were patients and 15 were controls.{Table 4}

Serum vitamin A levels <20 mcg/dL were seen in 201 (84.5%) subjects with CICT grades ≥2 Statistical significance of CICT results were compared to serum vitamin A levels [Table 5].{Table 5}


Animal experiments, cell culture studies, and population-based studies showed association hypovitaminosis A with cancers. Initial studies hinted at the role of the vitamin in chemoprevention. However, CARET a phase III clinical trials failed to prove the role of the vitamin A in cancer prevention. [7] In clinical practice, majority of these patients require supplementation. Excessive consumption of vitamin is not devoid of side-effects. Hence, it is important to identify the population that will benefit from the supplementation.

Conjunctival cytology (CIC) is non-invasive, relatively easy to perform, and yields information about the area sampled with minimal discomfort to the subject. In view of these advantages, CIC is recommended by several workers for evaluation of dry eye states and many other conjunctival disorders. However, there are no studies providing its utility for detecting dry eye state in cancer patients.

The present study comprised of subjects including cancer patients attending the outpatient department of Medical Oncology division for the first time without any treatment and their accompanying kin as controls. The kin were selected as controls to ensure uniformity in dietary patterns with the presumption that members of the same family get similar nourishment. Compliance was good as blood samples for serum for vitamin A levels and smears for CICT could be collected in of the 98% subjects each.

Vitamin A levels were found to be <20 mcg/dl in 33.5% of patients and 9.9% of controls with P < 0.001. CICT grades ≥2 were found in 239 subjects including 186 (77.8%) patients and 53 (22.2%) controls with P < 0.001. The results of CICT compared with serum vitamin A levels had a sensitivity 93.0%, specificity of 95.6%, positive predictive value 86.1%, negative predictive value 92.9%, Cohen's' kappa value 0.74, and P < 0.001. Paired t test comparing the patients and their kins showed vitamin A deficiency and xerosis proved by CICT are significantly higher in patients. This proves that cooking practices and economic stratum did not play a role.

Hypovitaminosis A and malignancies

Hypovitaminosis A is seen in cancer patients noticed by Abels et al. [8] as early as 1941, and their work was mainly on tumors of gastrointestinal tract. Kuwata et al. [9] reported that vitamin A deficiency in mice caused systemic expansion of myeloid cells and a rapid recovery after retinoic acid therapy. The authors expressed that in the absence of mutational events, mere deficiency of vitamin A may not contribute to malignancies. They observed that it is possible that tumor cells may have been derived from expansion of vitamin A-depleted cells in the presence of mutational event or had acquired vitamin A deficiency status as a result of a genetic mutation in one of the key retinoid signaling genes. In addition, localized deficiency may not be reflected in systemic levels under normal conditions of vitamin A nutriture.

Kuwata et al. [10] observed dramatic increase in myeloid cells in bone marrow, spleen, and peripheral blood in vitamin A deficient mice. The abnormal expansion of myeloid cells was detected from an early stage of vitamin A deficiency and contrasted with essentially normal profiles of T and B lymphocytes. This abnormality was reversed on addition of retinoic acid to the vitamin A deficient diet. The authors opined that it indicates that the myeloid cell expansion is a direct result of retinoic acid deficiency.

Shivakumaran [11] estimated vitamin A levels in 34 patients of various types of myeloproliferative disorders (MPD). None of the patients in this group had vitamin A deficiency. Hence, it is not clear whether vitamin A deficiency is directly involved in causation of myeloid malignancies in general population in the absence of specific gene defects.

The largest prospective study with respect to plasma carotenoids, tocopherol, and retinol was done by Tamimi et al. [12] with respect to breast cancer risk in 121,700 registered nurses followed them up for a long period. During the follow-up, 974 patients developed breast carcinoma. They have taken almost equal number of controls and performed plasma carotenoid, retinol, and tocopherol and tested with HPLC. They have observed inverse association of carotenoids and breast cancer in post-menopausal women. In addition, α- carotene had pronounced inverse correlation with nodal metastasis. In their study on alcohol consumption and risk of breast cancer, Zhang et al. [13] observed that intake of carotenoids reduced the risk of breast cancer in pre-menopausal but not post-menopausal women.

Epplein et al. [14] conducted a nested case - control, multi-ethnic cohort study on 215, 000 subjects to examine the association of retinoids with post-menopausal breast cancer risk. This resulted in a study population of 286 cases and 55 controls. Women with breast cancer tended to have lower levels of carotenoids compared to the controls, but the differences were not statistically significant, and trends were not monotonic.

Williams et al. [15] assessed CRBP1 expression by immunohistochemistry in prophylactically removed ovaries from women with a genetic risk of ovarian cancer. They have also tested the ability of normal, immortalized but non-tumorigenic and tumorigenic ovarian epithelial cells to synthesize retinol and retinaldehyde when challenged with physiologic doses of retinol and determined the expression levels of retinoid related genes, RARα, RXRα, CRABP1, CRABP2, RALDH1, and RALDH2. The authors concluded that impaired conversion of retinol to RA and decreased CRBP1 protein expression in prophylactic oophorectomies support the hypothesis that concomitant losses of vitamin A metabolism and CRBP1 expression contribute to ovarian oncogenesis.

Natadisastra et al. [16] studied CIC in 149 children with clinical evidence of vitamin A deficiency and their age-matched controls. CIC was abnormal in some of the controls as well. CIC results correlated monotonically through all stages with a P < 0.001 both in all the subjects. CIC returned to normal with vitamin A supplementation in these children. The authors concluded that CIC can detect early, physiologically significant vitamin A deficiency.

Amedee-Manesme et al. [6] studied the utility of CIC in sub-clinical vitamin A deficiency in 16 children presenting with liver diseases. CIC results were compared with plasma retinol concentrations by HPLC, vitamin A content in liver biopsy, and relative dose response. The authors observed that vitamin A levels were low in 4/16 of the children and CIC correlated with biochemically determined vitamin A status. They recommend RDR over plasma retinol levels as they were observed to be better indicators of liver stores. In another study, the authors have validated vitamin A concentrations in needle biopsies with macro samples at autopsy with an r value of 0.91. They have recommended needle biopsies in suitable clinical conditions. [17] Sommer and Muhilal [18] have studied general nutrition in corneal xerophthalmia and keratomalacia and concluded that both protein and vitamin A levels affect vitamin A metabolism and latter is more important than the former.

Hatchell and Sommer [19] studied ocular surfaces abnormalities in vitamin A-deficient rabbits and found CIC to useful in detection of sub-clinical xerophthalmia. Electron microscopy was also done on the samples. Rahman et al. [20] studied CIC in children <3 years with sub-clinical vitamin A deficiency. They found changes in CIC to be highly specific (91%) but with low sensitivity (9%). They concluded that the conjunctival changes take long time to develop and hence not suitable for sub-clinical deficiencies. Chowdary et al. [21] found a sensitivity of 90.6% and specificity of 100%, which is closer to the present study. Reddy et al. [22] studied CIC in 246 school children before and after vitamin A therapy and observed reversal of abnormal cytology to normal following therapy and concluded that the CIC changes are specific to vitamin A deficiency. CIC was abnormal in 25% of subjects who has no clinical evidence of xerosis; all of them had low serum vitamin A levels. In the present study, there was no obvious clinical evidence of vitamin A deficiency in any of the subjects.

To conclude, hypovitaminosis A and conjunctival xerosis are seen more frequently in patients with cancer than in controls. CICT is a reliable indicator of hypovitaminosis A with high sensitivity and specificity.


1Ghai OP, Gupta P, Paul VK. Vitamins. Micronutrients in Health and Disease. In: Text book of Essential Pediatrics. 6 th ed. New Delhi: CBS Publishers; 2005. p. 119-22.
2Gadkari SS, Adrianwala SD, Prayag AS, Khilnani P, Mehta NJ, Shaha NA. Conjunctival impression cytology: A study of normal conjunctiva. J Postgrad Med 1992;38:21-3.
3Wittpenn JR, West KP, Keenum D. ICEPO Training Manual. Assessments of Vitamin A status by impression cytology. The John Hopkins University, Baltimore; 1988. p. 7-16.
4Nelson JD, Wright JD. Conjunctival goblet cell densities in ocular surface disease. Arch Ophthalmol 1984;102:1049-51.
5Bieri JG, Tolliver TJ, Catignani GL. Simultaneous determination of alpha-tocopherol and retinol in plasma or red cells by high pressure liquid chromatography. Am J Clin Nutr 1979;32:2143-9.
6Amedee-Manesme O, Luzeau R, Wittepen JR, Hanck A, Sommer A. Impression cytology detects subclinical vitamin A deficiency. Am J Clin Nutr 1988;47:875-8.
7Goodman GE, Alberts DS, Meyskens FL. Retinol, vitamins and cancer prevention: 25 years of learning and relearning. J Clin Oncol 2008;26:5495-6.
8Abels JC, Gorham AT, Pack GT, Rhoads CP. Metabolic studies in patients with cancer of the gastro-intestinal tract. I. Plasma vitamin A levels in patients with malignant neoplastic disease, particularly of the gastro-intestinal tract. J Clin Invest 1941;20:749-64.
9Kuwata T, Wang IM, Tamura T, Ponnamperuma RM, Levine R, Holmes KL, et al. Vitamin A deficiency in mice causes a systemic expansion of myeloid cells. Blood 2000;95:3349-56.
10Kuwata T, Wang I-M, Tamura T, Ponnamperuma RM, Levine R, Holmes KL, et al. Vitamin A deficiency in mice causes a systemic expansion of myeloid cells. Blood 2000;95:3349-56.
11Sivakumaran M. Role of vitamin A deficiency in the pathogenesis of myeloproliferative disorders. Blood 2001;98:1636-7.
12Tamimi RM, Hankinson SE, Campos H, Spiegelman D, Zhang S, Colditz GA, et al. Plasma carotenoids, retinol, and tocopherols and risk of breast cancer. Am J Epidemiol 2005;161:153-60.
13Zhang S, Hunter DJ, Forman MR, Rosner BA, Speizer FE, Colditz GA, et al. Dietary carotenoids and vitamins A, C, and E and risk of breast cancer. J Natl Cancer Inst 1999;91:547-56.
14Epplein M, Shvetsov YB, Wilkens LR, Franke AA, Cooney RV, Le Marchand L, et al. Plasma carotenoids, retinol, and tocopherols and postmenopausal breast cancer risk in the Multiethnic Cohort Study: A nested case-controls study. Breast Cancer Res 2009;11:R49.
15Williams SJ, Cvetkovic D, Hamilton TC. Vitamin A metabolism is impaired in human ovarian cancer. Gynecol Oncol 2009;112:637-45.
16Natadisastra G, Wittpenn JR, West KP Jr, Muhilal, Sommer A. Impression cytology for detection of vitamin A deficiency. Arch Ophthalmol 1987;105:1224-8.
17Amédée-Manesme O, Furr HC, Olson JA. The correlation between liver vitamin A concentrations in micro-(needle biopsy) and macrosamples of human liver specimens obtained at autopsy. Am J Clin Nutr 1984;39:315-9.
18Sommer A. Nutritional factors in corneal xerophthalmia and keratomalacia. Arch Ophthalmol 1982;100:399-403.
19Hatchell DL, Sommer A. Detection of ocular surface abnormalities in experimental vitamin A deficiency. Arch Ophthalmol 1984;102:1389-93.
20Rahman MM, Mahalanabis D, Wahed MA, Islam M, Habte D, Khaled MA, et al. Conjunctival impression cytology fails to detect subclinical vitamin A deficiency in young children. J Nutr 1995;125:1869-74.
21Chowdhury S, Kumar R, Ganguly NK, Kumar L, Nain CK, Walia BN. Conjunctival Impression cytology with transfer (CICT) to detect pre-clinical vitamin A deficiency among slum children in India. Br J Nutr 1996;75:785-90.
22Reddy M, Reddy PR, Reddy SC. Conjunctival impression cytology in dry eye states. Indian J Ophthalmol 1991;39:22-4.