Impression cytology with transfer in xerophtalmia and conjunctival diseases

International Ophthalmology 16: 445-451, 1992. O 1992 Kluwer Academic Publishers. Printed in the Netherlands.

Impression Cytology with Transfer in xerophtalmia and conjunctival diseases Serge Resnikoff, 1 R6gine Luzeau, 2 Gilles Filliard I & Olivier Amedee-Manesme2 1Institute of African Tropical Ophthalmology (IOTA), BP 248, Bamako, Mali 2INSERM U 56 et D~partement de P~diatrie, HOpital de BicOtre, 93270 Kremlin-BicOtre, France Received 26 March 1992; accepted 30 March 1992

Key words: conjunctivitis, Impression Cytology method, Limbal Vernal Keratoconjunctivitis, retinol, trachoma, xerophthalmia Abstract

During a countrywide survey, we assessed the prevalence of clinical signs of xerophthalmia and of major conjunctival diseases in a randomized sample of 2,445 subjects representative of the population of the Republic of Djibouti. On a part of this sample, conjunctival Impression Cytology with Transfer (ICT) test and a plasma retinol determination were performed. Xerophthalmia as a public health problem was displayed by clinical signs (Bitot's spots, corneal scars among preschool children), low plasma retinol levels and ICT test results: 9.3% with deficient cytology in the rural area and 12.3% in the urban one (agestandardized rates). Results of ICT were related to age (p < 0.00001). Vitamin A deficiency was prevalent not only in preschool children but also up to 15 years. Moreover, ICT results are influenced by conjunctival diseases: compared to age-matched controls, there were more abnormal cytologies among patients with trachomatous inflammation (p = 0.025), conjunctivitis (p = 0.024) or Limbal Vernal Keratoconjunctivitis (p = 0.015). Thus ICT shouldn't be performed among children with conjunctival diseases. In the region under study conjunctival diseases had high rates of prevalence: 16.4% of trachomatous scarrings in the urban area (standardized rate), 8% of conjunctivitis among rural preschool children, and 5% of Limbal Vernal Keratoconjunctivitis among children between 5 and 14 years in both areas.

Introduction

Vitamin A is an essential nutrient for vision, growth, reproduction, epithelial cells differentiation and immuno-competence in man [1]. Vitamin A deficiency is a major problem, affecting 200 million of children in the world and leading to blindness in 500,000 of them every year. The deficiency also carries an increased risk of mortality and morbidity [2, 3]. Finally hypovitaminosis A also occurs in industrialized countries among poor and elderly people [1].

In this population-based study we determined the prevalence of clinical signs of xerophthalmia and of major conjunctival diseases in a randomized sample of 2,445 subjects, representative of the population of the Republic of Djibouti. On a part of the sample, a conjunctival Impression Cytology test with Transfer [4-7] and a plasma retinol determination were performed.

446

S. Resnikoff et al.

Subjects and methods The study was carried out on a representative sample of the Djiboutian population. The Republic of Djibouti is a 23,200 k m 2 country situated in the horn of Eastern Africa. The population is estimated to be 470,000 in 1988. Some 75% of the population live in urban areas, particularly in Djibouti city. The other 25% lead largely nomadic lives in the semi-desert hinterland crisscrossing the borders, in Somalia and Ethiopia. Average annual rainfall is less than 125 ram. Life expectancy at birth was estimated at 45 years in 1985 with an infant mortality rate of 200 per thousand. For administrative and logistic reasons we used a two stages random cluster sampling [8, 9] with two strata (rural and urban). The sampling frame consisted of the 1985 census. From a random start 65 sites ('clusters') were selected: 35 in rural area and 30 in urban one. Within each site we selected at random a starting point ('household'). Examination of all individuals began in the starting household and then continued to the next nearest one, until a total of 35 people had been examined. All individuals living in the last household falling into the sample were included even if the number was over 35. In accordance with this methodology 2,445 people were selected (i.e. one inhabitant in 200 in the urban area and one in 50 in the rural area), providing a representative sample of the Djiboutian population.

Ophthalmologic investigation All ocular examinations were carried out by an ophthalmologist with a portable slit-lamp (Kowa SL5), using standard diagnostic criteria for Trachoma [10, 11], Xerophthalmia [12], Limbal Vernal Keratoconjunctivitis (LVCK) [13] and Conjunctivitis (conjunctival hyperhaemia with purulent discharge in the lower fornix). The prevalence of main tropical conjunctival diseases (Trachoma, Conjunctivitis and LVCK) was determined on the whole sample (N = 2,445). In regard to previous

studies [12] we restricted our search of xerophthalmia to children below 6 (N = 520) and we considered corneal scarrings (stage XS) only in those. cases obviously related to a nutritional problem.

Impression Cytology with Transfer (ICT) ICT was performed on a large part of the sample (N = 2223). The test was done on both eyes of each patient, as previously described [5-7]. Celluloseacetate filter paper (Millipore Cat. HAWP 304) is cut into strips (5 x 25 mm). The distal end of the strip is applied on the temporal portion of the bulbar conjunctiva for a few seconds. The cells collected on the paper are immediately transferred to a glass slide by simple application and light finger pressure. These cells are fixed in 95% alcohol during 15 minutes. The slide is dried, then immersed in a staining solution (Ziehl/Alcain blue) for 15 minutes. After water washing, the slide is examined under light microscopy. Nuclei and cytoplasm of epithelial cells and nuclei of goblet-cells are stained in pink by acid fuschin. The acid mucous substance is stained in blue with alcian blue. Vitamin A status is determined according to the presence (sufficient status) or absence (deficient status) of goblet cells and according to the morphology of the epithelial cells. A normal impression cytology is characterized by small epithelial cells without distinct borders and by the presence of goblet cells or mucine spots with nuclei (Fig. 1). This corresponds to a liver vitamin A concentration above 20 microg/g of liver [4]. In the deficient one, epithelial cells are enlarged and isolated with picnotic nuclei and no goblet cells: liver vitamin A concentration is under 20 microg/g of liver (Fig. 2). All stages between normal and deficient cytology may be seen: a decrease in the number of goblet cells and an enlargement of epithelial cells. They are called marginal (borderline). Although some distinctions are possible within this marginal group, we have not used such a subclassification in this study. When a difference was noted between the two eyes, the most normal result was taken into consideration.

ICT in xerophtalmia and conjunctival diseases

Fig. 1.

Fig. 2.

Determination of retinol concentration

Results

2 ml of blood were collected from each individual 4 years and above, using vacutainer tubes. After centrifugation the serum was divided in two aliquots and stored at - 2 0 ~ until analysis. Retinol was determined in all children aged between 4 and 10 years (83 in the urban area and 114 in the rural one). Serum vitamin A concentration was estimated by reversed-phase high pressure liquid chromatography (HPLC) with UV detection at 325 nm [14]. Statistical analyses employed the Chi-square test (Yates corrected) and Fisher's exact test. Confidence intervals (CI) are given at 95%.

Clinical signs of xerophthalmia

447

Results of xerophthalmia assessment in children under 6 years are reported in Table 1. Prevalence of Nightblindness (XN) was low. On the other hand, many preschool children had Bitot's spots (X1B), especially in the rural area (1.04%) where this rate was probably above the W H O public health criterion (p = 0.13) [14]. Although no keratomalacia was observed during the survey, we found an important number of children with corneal scarring in both areas: 0.8% (rural) and 0.4% (urban).

Table 1. Prevalence of clinical signs of xerophthalmia among preschool children in the rural area (N = 385) and the urban area (N = 235). Comparison with WHO's cut off criteria (Republic of Djibouti, 1988). Prevalence

CP

WHO criteria b

Nightblindness (XN) Rural area Urban area

0.26% 0.0%

0.01-1.4% 0.00-1.5%

1.0% 1.0%

Bitot's spots (X1B) Rural area Urban area

1.04% 0.43%

0.3-2.5% 0.01-2.2%

0.5% 0.5%

p = 0.12

Corneal scar (XS) Rural area Urban area

0.80% 0.43%

0.2-2.3% 0.01-2.2%

0.05% 0.05%

p < 0.00001 p = 0.01

a CI: Confidence Interval. bWHO's prevalence criteria for a public health problem of xerophthalmia. cp value when observed prevalence is above cut-off criteria.

p value c

448

S. Resnikoff et al.

Table 2. Standardized (age and sex adjusted) prevalence rates of conjunctival diseases (per 100) in rural and urban areas (Republic of Djibouti, 1988). Age

0--4

5-14

15-44

45+

All ages

Area No

R 133

U 111

R 266

U 222

R 729

U 616

R 198

U 168

R 1326

U 1117

Tr. Inf. ~.Scar. Conjunct. LVKC

1.2 0.0 8.0 1.3

0.6 0.0 1.7 1.1

5.7 1.2 5.7 5.3

0.0 0.4 1.6 4.5

0.3 11.7 2.5 1.1

0.6 2L3 0.6 1.9

0.0 10.3 1.3 0.0

0.0 30.4 0.0 0.0

1.4 8.2 4.1 1.8

0,4 16,4 0.8 1.7

R: Rural area; U: Urban area; Tr. Inf.: Trachomatous inflammation; Tr. Scar.: Trachomatous scarring; Conjunct.: Conjunctivitis; LVKC: Limbal Vernal Keratoconjunctivitis.

Prevalence of conjunctival diseases In order to take into account the differences in the age distribution of the samples, we used a computational process known as standardization (age and sex-adjustement on WHO's African standard population) (Table 2). Although trachomatous inflammation was relatively rare, it was significantly more frequent in the rural area (1.9%) than in the urban one (0.4%, p = 0.01). Conversely, trachomatous scarrings were markedly more frequent in the urban area (16.4%) than in the rural one (8.2%, p < 0.00001). On the other hand, Limbal Vernal Keratoconjunctivitis (LVKC) was particularly widespread in both areas, affecting up to 5% of children between Table 3. Impression Cytology with Transfer (ICT). Distribution of results in 3 groups (normal, borderline, deficient) and mean value of age for each group, ICT results

No.

Percentage rate

CI a

Mean value CI a of age (yrs)

Rural area (N = 799) Normal 620 77.6 Borderline 105 13.1 Deficient 74 9.3

+2.9 29.8 +2.3 20.1 ___2.0 12.9

+1.1 +2.7 + 2.3

Urban area (N = 844) Normal 618 73.2 Borderline 122 14.5 Deficient 104 12.3

+3.0 29.1 +2.4 24.8 + 2.2 18.5

+ 1.3 +2.7 + 2.6

a CI: Confidence intervals.

5 and 14 years, posing a more acute public health problem than inflammatory trachoma. In the rural area, prevalence of conjunctivitis was high among preschool children (8%) but decreased steadily with age. In the urban area, agerelated variations were the same but the rates were significantly lower (p