Evaluation of clinical pallor in the identification ... - Wiley Online Library

805_810_TMIH637

9/11/00 9:05 pm

Page 805

TMIH637 Tropical Medicine and International Health volume 5 no 11 pp 805–810 november 2000

Evaluation of clinical pallor in the identification and treatment of children with moderate and severe anaemia Lulu Muhe1, Birhane Oljira1, Hirut Degefu1, Shabbar Jaffar2 and Martin W. Weber3 1 Department of Paediatrics and Child Health, Addis Ababa University, Addis Ababa, Ethiopia 2 Department of Infectious and Tropical Diseases, London School of Hygiene and Tropical Medicine, London, UK 3 Department of Child and Adolescent Health and Development, World Health Organization, Geneva, Switzerland

Summary

background Anaemia from malaria is a common problem in developing countries. Blood transfusion in developing countries is available in few hospitals. Children who are severely anaemic and may require urgent blood transfusion usually present to peripheral first-level health facilities from where they must be referred to hospitals. Since most peripheral facilities do not determine haemoglobin levels, the decision on referral has to be made on clinical grounds. objectives To evaluate the sensitivity and specificity of clinical pallor of the palms, nailbeds, conjunctivae, buccal mucosa or tongue against haemoglobin values and their reproducibility among health workers. methods A total of 2540 children 2 months to 5 years of age presenting to a rural health centre in Ethiopia were enrolled. Clinically detected pallor was compared with measured blood haemoglobin concentrations. results Any anaemia (haemoglobin 11 g/dl) was found in 61% of the children. Severe anaemia (haemoglobin 5 g/dl) was found in 4%. The presence of any pallor clinically correlated with moderate anaemia (haemoglobin level 8 g/dl) could be detected with a sensitivity of 95% and a specificity of 64–68% when the palm and nailbeds were used and a sensitivity of 84% and a specificity of 81% when the conjunctivae were used. Severe anaemia was detected clinically as severe pallor in 50–56% of cases (with a specificity of 95–96%). Agreement between physicians was highest for conjunctivae and nailbed pallor (87%) and lowest for palm pallor (73%). Using multivariate analysis, identification of a systolic ejection murmur or altered sensorium, the presence of splenomegaly or malarial parasitaemia were independently predictive of severe and moderately severe anaemia. conclusions Moderate and severe anaemia can be identified clinically in most cases for treatment and referral purposes. A systolic ejection murmur, altered sensorium, the presence of splenomegaly or malarial parasitaemia may be used as additional tools in considering urgent referral for blood transfusion.

keywords malaria, anaemia, pallor, child, case management, clinical signs correspondence Lulu Muhe, Department of Paediatrics and Child Health, P. O. Box 1768, Addis Ababa University, Addis Ababa, Ethiopia.

Introduction Anaemia resulting from a lack of sufficient iron for synthesis of haemoglobin is by far the most frequent haematological disease of infancy and childhood (Schwartz 1996). Anaemia is not a specific entity but an indication of an underlying pathologic process or disease. Although any reduction in haemoglobin concentration from normal decreases the oxygen-carrying capacity of the blood, the clinical sign of pallor becomes evident in skin and mucous membranes only when the level falls below 7–8 g/dl (Schwartz 1996). During

© 2000 Blackwell Science Ltd

normal growth, children need to expand the iron stores of the body from a total iron content of the newborn infant of about 0.5 g to the content of the adult, which is estimated at 5.0 g. To this growth requirement an additional small amount is necessary to balance the normal losses of iron through excretion. Adequate iron intake is difficult to achieve for children in developing countries, as breast milk by itself is insufficient, and complementary foods tend to be poor in total iron content, and the iron they contain often has a low bioavailability. There is abundant evidence that anaemia is present 805

805_810_TMIH637

9/11/00 9:05 pm

Page 806

Tropical Medicine and International Health L. Muhe et al.

volume 5 no 11 pp 805–810 november 2000

Treatment of anaemia in children

amongst a large population of children in developing countries (WHO 1959). Aetiologically, dietary deficiencies, i.e. low iron intake or loss of iron through chronic blood loss due to parasitic infestations such as hookworm, are common in developing countries. Malaria affects more than 200 million people in the world (World Bank 1993) and is by far the commonest aetiologic agent of haemolytic anaemia in these children (WHO 1959). In African children, anaemia is a common presenting feature of malaria (Brewster & Greenwood 1993). An acute falciparum malarial attack produces haemolysis and can result in a rapid drop in haemoglobin, leading to a life-threatening situation (WHO 1990). In many developing countries, blood transfusion is available only in selected hospitals. Severely anaemic children who mostly present to peripheral first-level health facilities have to be recognized and must be referred. Because most such peripheral facilities cannot carry out accurate haemoglobin determinations, the decision on referral must be made on clinical grounds. The Integrated Management of Childhood Illness (Anonymous 1995) strategy of the World Health Organization thus recommends the use of clinical pallor. Children presenting at a first-level health facility with any clinical pallor should be prescribed iron, and those with clinically severe pallor should be referred urgently for blood transfusion. Clinically, anaemia can be detected on the basis of pallor of the palms, nailbeds, conjunctiva, mouth mucosa or tongue. This paper presents an evaluation of the sensitivity, specificity and reproducibility of pallor at different sites and other clinical signs used to diagnose moderate and severe anaemia in children from 2 months up to 5 years of age.

in a logbook where a research nurse entered their weight, height, residence and diagnoses. A systematic sample of these children, i.e. every other child, was enrolled. Enrolled children had a detailed history and clinical examination by a study physician (first physician). The first physician documented the child’s age in months, sex, difficulty in breathing, respiratory rate, pulse rate, pallor (whether severe or moderate, and if the conjunctiva, palm, nailbed or tongue were involved), grunting, nasal flaring, chest indrawing, neck vein congestion, pulsations over the neck, raised jugular venous pressure, gallop rhythm, ejection systolic murmur, hepatomegaly, total liver span, and tenderness of the liver. After clinical examinations were completed by the first physician, a second physician re-examined the children who had been enrolled and documented his decisions on pallor, whether severe or moderate and whether the conjunctiva, palm, nailbed or tongue were involved, without any information from the first examination. After the second examination, the child underwent a blood examination for haemoglobin. The second physician examined all the children enrolled in the study, whilst the initial examination was carried out by five different physicians over the study period. To maintain the quality of the clinical data of the first physicians, in addition to the intensive initial training and pilot study, the senior investigators saw the same patients during their weekly visits and filled in the same forms blindly. Discrepancies between the findings were discussed, but not formally analysed. Repeated presentations of the same patients were excluded. Laboratory tests

Methods Study setting The study site is located in the densely populated Butajira district, 130 km south of Addis Ababa, Ethiopia. The topography of Butajira comprises highland, lowland and inbetween characteristics with an altitude ranging between 1500 and 2300 m above sea level. The town of Butajira, where the study health centre is situated, lies at an altitude of 2000 m, and therefore at the border of the malaria-endemic area. There is usually very little transmission of malaria during the first 4 months of the year, January to April. Both Plasmodium falciparum and P. vivax are prevalent in the district. Patient enrolment Children between 2 and 59 months of age who presented at the outpatient department in the morning working hours of the day were eligible for the study. All children were registered 806

A fingerprick blood sample for haemoglobin, malarial parasites (thick and thin film examination) and Borrelia recurrentis was obtained. Haemoglobin was measured using the ‘HemoCue’ method (von Schenck et al. 1986). Children were classified as anaemic if their haemoglobin was 11 g/dl, moderately anaemic if the haemoglobin was 8 g/dl and severely anaemic if the haemoglobin level was 5 g/dl. For the analysis of sensitivity and specificity, the cut-off points of 8 and 5 were used, as they are clinically meaningful for the initiation of iron supplementation and referral for transfusion, respectively. The investigators ensured that appropriate measures were taken for sterilization of equipment in order to avoid transmission of microorganisms on materials used to obtain fingerprick samples. Treatment was according to standard practice. Very sick children were referred to the nearest hospital. Informed consent was obtained from guardians. The study was approved by the ethical committee of Addis Ababa University and the Secretariat Committee on Research in Human Subjects of the World Health Organization.


805_810_TMIH637

9/11/00 9:05 pm

Page 807




Table 1 Prevalence of anaemia defined by cut-off values for haemoglobin of 11, 8 and 5 g/dl by age groups. The numbers in parentheses under ‘Total number’ are column percentages, the other numbers in parentheses are row percentages

Age groups

2–11 month 12–23 month 24–35 month 36–47 month 48–59 month Total

Total number(%)

Anaemia cut-off levels Hb 11 g/dl Hb 8 g/dl

Hb 5 g/dl

0744 (29) 0565 (22) 0520 (21) 0339 (13) 0372 (15) 2540

0450 (60) 0357 (63) 0324 (62) 0190 (56) 0223 (60) 1544 (61)

011 (1.5) 030 (5.3) 020 (3.8) 015 (4.4) 025 (6.7) 101 (4.0)

095 (13) 116 (21) 107 (21) 072 (21) 081 (22) 471 (19)

Statistical analysis

Results

Data were entered into the Epi Info version 6 program. Frequencies and cross-tabulations were computed using EpiInfo software. Concordance and kappa values were computed using the STATA software package. Discrete clinical signs were compared between anaemic and non-anaemic children using the 2-squared test with Yates’ correction or by Fisher’s exact test if the frequencies were small. Continuous variables were compared using Student’s t-test or the Wilcoxon rank sum test, as appropriate. Signs and symptoms that showed a significant association with anaemia in the univariate analyses at a significance level of 0.1 or below, plus those that had been reported as important predictors in previously published studies, were considered for multivariate logistic regression analyses to assess their independent utility in predicting anaemia. Modelling was repeated using variables in the sections of pallor, heart failure variables and other signs. In each section, models were built up consecutively, starting with the most significant variables until no further variables reached significance. Variables were then dropped from the multivariate model one at a time to see if they remained significant in the presence of all other variables. Significant and independent predictors of anaemia from each of these sections were then investigated together and a final set of predictors was found. Different models were compared with one another by change in deviance. The SAS System for windows was used for the multivariate analysis.

The study was conducted over a period of 12 months during the calendar year 1996. A total of 2540 children aged between 2 months and 5 years were enrolled (Table 1). Of these, 744 (29%) were younger than 12 months, and 565 (22%) were between 12 and 25 months of age. A total of 101 children (4%) required admission for urgent transfusion; nine of these needed referral to a regional hospital for further investigation. One child died during the follow-up period. All except one child were re-examined for pallor by a second physician. Of the total 2540 children enrolled, 120 (5%) had a complete physical examination performed by senior clinicians to countercheck reproducibility of clinical findings. Anaemia as defined by a haemoglobin level of less than 11 g/dl was found in 61% of all children enrolled in the study (Table 1). The anaemia prevalence was lowest in the age group 12 months. Males were affected more than females. Anaemia of moderate degree with a haemoglobin level 8 g/dl was found in 19%. Severe anaemia (requiring referral for urgent blood transfusion) as defined by a haemoglobin level of 5 g/dl was found in 4.0% of all children. Severe anaemia was found least frequently in the age group 12 months. The correlation between anaemia level and clinical pallor at different sites is shown in Table 2. Moderate anaemia with a haemoglobin level 8 g/dl was detected best by palm pallor (sensitivity of 84% and specificity 81%) and pallor of the

Table 2 Sensitivity and specificity of moderate and severe pallor at different sites as detected by the physician against haemoglobin values of less than 8 g/dl and 5 g/dl determined by the Hemocue method Conjunctival pallor Palm pallor Mucosal (tongue) pallor Nailbed pallor


Moderate pallor ——————————— Sensitivity Specificity

Severe pallor ——————————— Sensitivity Specificity

84 95

81 64

50 54

95 96

92 95

75 68

56 52

95 96

807

805_810_TMIH637

9/11/00 9:05 pm

Page 808




Table 3 Concordance of detection of pallor at different anatomic sites by first and second physicians Sign

Presence of pallor

Detected by first physician

Detected by second physician

Percentage agreement

Kappa

Conjunctival pallor

Yes No Yes No yes No yes No yes No yes No yes No yes No

0572 1967 0081 0321 0892 1647 0124 0534 0745 1794 0127 0472 0768 1771 0105 0463

0705 1834 0060 0342 1105 1434 0078 0580 0845 1694 0073 0421 1012 1527 0072 0496

81.4

0.51

87.3

0.56

73.2

0.44

86.9

0.50

79.3

0.52

85.0

0.52

74.6

0.45

87.5

0.53

Severe conjunctival pallor Palm pallor Severe palm pallor Mucosal pallor Severe mucosal pallor Nailbed pallor Severe nailbed pallor

nailbeds (sensitivity of 95% and specificity of 68%) (Table 2). Severe anaemia with a haemoglobin level of 5 g/dl was detected by clinical examination as severe pallor of any of the sites examined in 50–56% of the cases. The specificity values for severe pallor ranged from 95 to 96%. The concordance of detecting any pallor (mild, moderate, or severe) and severe pallor between the first physicians and the second physician is shown in Table 3. For detecting any pallor, simple agreement was highest for conjunctivae (81%) and lowest for palm pallor (73%). The concordance for severe pallor was better than for any pallor, as all sites had an agreement of over 85%, the highest being for nailbed and conjunctivae with 87.5 and 87.3% agreement, respectively. The kappa values ranged from 0.44 to 0.56. The z-scores were statistically significant in all cases. In multivariate analysis, the following signs were independently predictive of moderate anaemia (Hb 8 g/dl): conjunctiva pallor, mucosal pallor, pallor of the nailbeds, a systolic heart murmur, altered state of consciousness, splenomegaly and a positive blood film for malaria (both for P. falciparum and P. vivax). The odds ratios were generally higher for these signs when they were used in the model for severe anaemia than when they were used in the model for moderate anaemia (Table 4).

Discussion This study shows that anaemia is a major health problem in Ethiopia, similar to other studies from sub-Saharan Africa (Zucker et al. 1996). In addition to the high case-fatality rate that anaemia brings on children (Brewster & Greenwood 808

1993), it affects their development (Lozoff et al. 1991). In subSaharan Africa, where there is poor access to good health services and where malaria in children is rampant, testing standardized clinical approaches for identifying and treating anaemia is essential (Genton et al. 1994). Our data revealed a high prevalence of both severe and mild anaemia in Ethiopia, demonstrating the importance of identifying and appropriately managing children with anaemia. We were able to show that anaemia is less common among infants aged below 12 months, but then reaches a persistent level over all age groups, and that males are affected more often than females. Our study suggests that moderate anaemia is best detected by using palm pallor or pallor of the nailbeds (sensitivity 95%), while conjunctival pallor was least useful (sensitivity 84%). This may be due to the fact that conjunctivitis is common in these children in Ethiopia and that it may prevent detection of pallor. A similar trend is also shown for severe pallor, even though the sensitivity values are low in this case. In Bangladesh, palm pallor had a lower sensitivity than conjunctival pallor for both severe and moderate anaemia probably because of the increased palm pigment (Kalter et al. 1997). Similarly, in a study from Pakistan it was found that conjunctival pallor had the highest sensitivity of all sites for detecting anaemia with haemoglobin 11 g/dl in young children (Thaver & Baig 1994). Other studies in Africa (Luby et al. 1995; Zucker et al. 1997) and of whites in the USA (Nardone et al. 1990) have shown that the nailbeds and palm are the best sites for assessing pallor in these settings. Our data indicate that a significant proportion of children with mild anaemia can be missed if clinical pallor alone is used. However, compared to mild anaemia, the sensitivity of


805_810_TMIH637

9/11/00 9:05 pm

Page 809




Table 4 Independent predictors of moderate and severe anaemia as ascertained by multiple logistic regression

Moderate anaemia (Hb 8) Severe anaemia(Hb 5) Variables OR (95%CI) Age group (months) 2–11 12–23 24–35 36–47 48–59 Conjunctival pallor None Tongue pallor None Nailbed pallor None Systolic murmur None Consciousness Obtunded/coma Alert Falciparum malaria Positive Negative Vivax malaria Positive Negative Splenomegaly None Temperature*

0.8 (0.4, 1.5) 0.7 (0.3, 1.3) 0.5 (0.3, 1.0) 0.9 (0.4, 1.8) 1.0 3.6 (2.4, 5.2) 1.0 4.0 (2.5, 6.5) 1.0 1.8 (1.1, 2.8) 1.0 6.5 (1.0, 44.0) 1.0

P-value

OR (95%CI)

P-value

0.02

1.4 (0.5–3.9) 1.0 (0.4–2.6) 1.1 (0.5–3.2) 0.8 (0.3–2.4) 1.0 5.9 (2.2–16.1) 1.0 7.0 (1.2–42.6) 1.0 3.3 (0.7–15.4) 1.0 22.7 (5.4–95.9) 1.0

1.9 (1.1, 3.3) 1.0

0.02

8.5 (1.4–5.7) 1.0

0.004

6.2 (1.3, 9.3) 1.0

0.0001

5.4(2.9–10.1) 1.0

0.0001

3.1 (2.1, 4.5) 1.0 1.43 (1.0, 2.0) 1.0 –

0.15 0.0001 0.0001 0.02

0.0001

2.7 (1.4–5.1) 1.0 –

0.7 0.0001 0.015 0.09 0.0001

0.002

0.04 1.37 (1.1, 1.7) 0.002

*Fitted as a continuous variable. The odds of anaemia for each degree centigrade increase in temperature is shown.

clinical pallor improves when moderate anaemia with a haemoglobin level 8 g/dl is used as the outcome. This study demonstrates that clinical criteria can be used to identify children with anaemia of moderate degree who need treatment with iron. Moderate anaemia is associated with morbidity that may progress to severe anaemia and death unless prevented by treatment with iron (Lozoff et al. 1991; Schwartz 1996; Zucker et al. 1996). The analysis of the sensitivity of clinically severe pallor does not give as high a sensitivity but can still pick up as many as 50–56% of the children with haemoglobin levels 5 g/dl, similar to findings in Uganda and Bangladesh (Kalter et al. 1997). Identification of clinically severe pallor, thus, can detect more than half the children that require urgent transfusion and therefore need to be referred to a first-level referral hospital. Our study has shown that detection of clinical pallor is reproducible, as shown by the high simple agreement results and the reasonably high kappa values. Unlike the sensitivity results in the comparison with the haemoglobin measurement, conjunctival pallor had the highest agreement for any pallor. In the study from the Gambia, pallor of the conjunc-


tivae was detected more reproducibly than pallor of the nailbeds or palms (Weber et al. 1997). When only severe pallor is considered, agreement values exceed 85%. Adjusting for age in the multivariate analysis, we looked for signs that may indicate early congestive heart failure due to severe anaemia. The presence of an ejection systolic murmur was found to be an important independent predictor of severe anaemia as well as moderate anaemia. This probably indicates a state of pending or full-blown heart failure due to severe anaemia, which requires urgent transfusion. Obtundation or coma, another finding which is an important predictor of severe anaemia, may also be a sign of hypoxaemia due to heart failure. The presence of tachypnoea, chest indrawing, nasal flaring and hepatomegaly were not found to be significantly associated with anaemia in this study. However, the presence of splenomegaly or a positive blood film for malaria parasites was significantly associated with severe anaemia. In summary, our study confirms the high prevalence of anaemia in a developing country setting and that clinical criteria can pick up as many as 95% of cases with moderate anaemia. Detection of clinical pallor is reproducible and thus 809

805_810_TMIH637

9/11/00 9:05 pm

Page 810



can be taught to lower level health workers. In addition to severe pallor, signs suggestive of pending heart failure such as a systolic ejection murmur, change in sensorium, splenomegaly and detection of malaria parasites may be used to decide upon urgent referral of children with anaemia.

Acknowledgements We would like to acknowledge that this project was funded by the World Health Organization, Special Programme for Research and Training in Tropical Diseases (WHO/TDR) grant no. 940890. Shabbar Jaffar is funded by the United Kingdom Medical Research Council. We are grateful to the physicians, nurses, laboratory technicians, and other staff of the Butajira Health Centre for their assistance during the study, in particular Drs Mohammed Ahmed, Mulugeta Wondewossen, Kiflu Bereda, Anteneh Kebede, Gera Baruda and Wubnew Eshetu. We extend our appreciation to Drs Jackie Cattani and Sandy Gove from TDR/WHO and CHD/WHO, respectively, for their advice and support in initiating the study. We thank Mr Kidanemariam Woldeyesus for entering and ‘tidying up’ the data. We are indebted to the parents of the many children who consented to be enrolled in the study.

References Anonymous (1995) Integrated management of the sick child. Bulletin of the World Health Organization 73, 735–740. Brewster DR & Greenwood BM (1993) Seasonal variation of paediatric diseases in The Gambia, west Africa. Annals of Tropical Paediatrics 13, 133–146. Genton B, Smith T, Baea K et al. (1994) Malaria: how useful are clinical criteria for improving the diagnosis in a highly endemic area. Transactions of the Royal Society of Tropical Medicine and Hygiene 88, 537–541. Kalter HD, Burnham G, Kolstad PR et al. (1997) Evaluation of clini-

810


cal signs to diagnose anaemia in Uganda and Bangladesh, in areas with and without malaria. Bulletin of the World Health Organization 75 (Suppl. 1), 103–111. Lozoff B, Jimenez E & Wolf AW (1991) Long-term developmental outcome of infants with iron deficiency. New England Journal of Medicine 325, 687–694. Luby SP, Kazembe PN, Redd SC et al. (1995) Using clinical signs to diagnose anaemia in African children. Bulletin of the World Health Organization 73, 477–482. Nardone DA, Roth KM, Mazur DJ & McAfee JH (1990) Usefulness of physical examination in detecting the presence or absence of anemia. Archives of Internal Medicine 150, 201–204. Schwartz E (1996) Iron deficiency anemia. In: Nelson’s Textbook of Pediatrics. 15 Edn (ed. RE Behrman et al.) W. B. Saunders, Philadelphia, pp. 1387–1389. Thaver IH & Baig L (1994) Anaemia in children: Part I. Can simple observations by primary care provider help in diagnosis? Journal of the Pakistan Medical Association 44, 282–284. von Schenck H, Falkensson M & Lundberg B (1986) Evaluation of ‘HemoCue’, a new device for determining hemoglobin. Clinical Chemistry 32, 526–529. Weber MW, Kellingray SD, Palmer A et al. (1997) Pallor as a clinical sign of severe anaemia in children: an investigation in the Gambia. Bulletin of the World Health Organization 75 (Suppl. 1), 113–118. WHO (1959) Iron deficiency anaemia. Report of a study group. WHO Technical Series 182. WHO (1990) Severe and complicated malaria. Transactions of the Royal Society of Tropical Medicine and Hygiene 84 (Suppl. 2), 1–65. World Bank (1993) World Development Report 1993. Investing in Health. Oxford University Press, New York. Zucker JR, Lackritz EM, Ruebush TK et al. (1996) Childhood mortality during and after hospitalization in western Kenya: effect of malaria treatment regimens. American Journal of Tropical Medicine and Hygiene 55, 655–660. Zucker JR, Perkins BA, Jafari H et al. (1997) Clinical signs for the recognition of children with moderate or severe anaemia in western Kenya. Bulletin of the World Health Organization 75 (Suppl. 1), 97–102.
