Pneumonia In Yemeni Children [Archives:1998/06/Health]
Summary A one-year study of 529 Yemeni children under 5 years of age hospitalized for severe pneumonia was undertaken to define their clinical characteristics and to identify risk factors associated with death from pneumonia. There were 354 (66.9%) boys, 270 of the 529 (51%) were under 6 months of age and 457 (86.4%) were aged between 6 and 12 months. The clinical characteristics of the group were as follows: Boys constituted 70% of the group and under-1-year-olds 86%, weight-for-age was under 60% in 23%, clinical rickets was present in 50% and anemia in 30% (30.1%). On admission, cyanosis was detected in 56%, heart failure in 21% and isolated hepatomegaly in 14%. Fifty-two children died (CFR 9.8%), of whom 25 (48%) were under 6 months of age and 20 (38.5%) were aged between 6 and 12 months. Only seven children aged over 1 year died from pneumonia. Weight-for-age less than 60%, rickets, hemoglobin < 10 g/dl, cyanosis and heart failure were associated with an increased risk of dying from severe pneumonia.
Introduction In developing countries, more than 4 million children under 5 years of age die each year from acute respiratory infections (ARI), particularly pneumonia. ARI is a leading cause of sickness and death among children under 5 years of age in Yemen. In a recent epidemiological survey, 49% of this age group had cough, and 25% had cough and difficult breathing during the 2 weeks prior to the survey. The same survey found that of those who had died in the preceding 5 years, ‘cough and difficult breathing’ was the probable cause of death in 28.4% of those aged between 1 and 11 months and in 29.8% of those in the 12-59 months age group.
Pneumonia-specific case fatality rates remain unacceptably high in Yemeni hospitals. In Al-Sabeen Hospital for Women and Children, a major Yemeni hospital, pneumonia case fatality rates for the years 1991,1992 and 1993 were 7.9%, 9.4% and 7%, respectively (unpublished data). Hospital-based studies can yield useful information on the impact of ARI on mortality in a community. This study, the first of its kind in Yemen, has two objectives: (1) to document the clinical characteristics of children admitted to the hospital with severe pneumonia, and (2) to identify the risk factors associated with death from severe pneumonia in hospitalized children.
Location of Study & Facilities Al-Sabeen Hospital for Women and Children in Sana’a, Yemen, is one of three major hospitals serving the Sana’a province and the surrounding rural areas with a population of approximately 2 million. The hospital offers both maternal and child health services, including walk-in outpatient clinics and emergency units, and 24-hour X-ray services. Most patients use these facilities as their first medical contact point. Sana’a is in the highlands of Yemen at an altitude of 2350 meters, and malaria is seldom found amongst residents of the main catchment area. The hospital follows the guidelines recommended by the WHO for the diagnosis and management of ARI, and several workshops have been conducted for hospital staff on standard ARI case management.
Patients & Methods Included in the study were all children aged from 2 weeks to 59 months admitted via the emergency or outpatient units with a clinical diagnosis of severe pneumonia (WHO clinical criteria) and radiological evidence of pneumonia, confirmed by the hospital radiologist, who was unaware of the clinical status of the children. Excluded were those with a history of cough for more than 3 weeks, contact with active pulmonary tuberculosis, clinical evidence of meningitis or with pneumonia secondary to measles or whooping cough.
Between 1 April 1995 and 31 March 1996, 529 children under 5 years of age were admitted with a diagnosis of severe pneumonia which met our criteria. The clinical characteristics of each child were recorded by experienced resident medical officers on a standard data collection sheet. Age was recorded in months and weight was measured to the nearest 100g on a pre-set scale in the ward.
Nutritional status was expressed as weight-for-age based on WHO growth charts and the children were grouped as above 80%, 60-80% and below 60% of expected weight-for-age. The presence of subcostal and/or intercostal recessions and cyanosis on inspection of the inner surface of the lips and tongue were recorded. Clinical evidence of rickets was recorded if at least two of the following were found: broad wrists, beading of the ribs (rachitic rosary), cranial bossing, craniotabes and chest deformities. Heart failure was diagnosed when tachycardia was more than 160 beats per minute and hepatomegaly more than 2cm. Hepatomegaly alone was diagnosed when the liver was enlarged by more than 2cm in the absence of tachycardia. Hemoglobin was measured soon after admission by capillary hematocrit using the Hawksley Microhaematocrit Reader (Hawksley, England).
All children received similar management. Intravenous ampicillan and chloramphenicol were given to those of more than 2 months of age and then changed to oral medication when tolerated. Ampicillin and gentamicin were given to infants under 2 months of age. Treatment for staphylococcal pneumonia was considered when a child showed no improvement within 3 days. Oxygen was given by nasal catheter to those with cyanosis and/or severe intercostal recession. Intravenous fluids were given carefully when clinically indicated.
Statistical analysis for risk factors was done by means of Epi Info 6 (CDC, Atlanta, GA, USA). A 95% confidence interval with exact confidence limits was used. P-values were determined according to the Mantel-Haenszel test.
Results During the study period, 529 children aged from 2 weeks to 59 months were admitted with severe pneumonia. There were 354 (66.9%) boys and 457 (86.4%) patients were under 1 year of age.
Table 1 shows the frequency of various clinical features by age. Clinical rickets was found in 50% of the cases. Among those under 60% of expected weight-for-age and those of between 60 and 80% and above 80%, rickets was found in 75%, 51% and 24%, respectively. Clinical rickets was detected at 1 month of age in four and in a further 14 infants at 2 months of age. The pneumonia-specific case fatality rate was 9.8% (52/529). The majority who died (86.5%) were under 1 year of age, and 12 (23%) deaths occurred in the first 2 months of life. There were 14 deaths (27%) within 24 hours of admission and 50% of deaths had occurred by the third day of admission.
Half (50%) of the children who died had a weight-for-age below 60%, and 23 (44%) were between 60 and 80%. Among those who were above 80% of expected weight-for-age, only 3 died. The hemoglobin level was above 10 g/dl in 17 (33%) of the children who died and below this level in the others, falling to below 7 g/dl in 3 children. The male-to-female ratio and age distribution were similar for the deceased and the survivors.
Table 2 shows the significant risk factors for death from severe pneumonia which include: a weight-for-age of less than 60%, clinical evidence of rickets, presence of cyanosis, heart failure and hemoglobin less than 10 g/dl. Hepatomegaly alone was not associated with an increased risk of dying.
Discussion This study shows that over half the children admitted with severe pneumonia were under 6 months of age and that pneumonia was six times more frequent in the first year of life than later in children under 5 years of age. Malnutrition is a major health problem in the under-5s in Yemen. A survey in 1987 revealed that 46% of children under 5 years of age from low income families in Sana’a were malnourished, and that the prevalence of malnutrition in other parts of Yemen was between 40 and 60%.
Malnutrition appeared to have a powerful influence on the occurrence of pneumonia in our children, 75% of whom were below 80% of expected weight-for-age. Herrero found that malnourished children were 12 times more likely than well nourished ones to have an episode of pneumonia. Others have also reported malnutrition to be associated with an increased risk of developing acute lower respiratory infections. Malnutrition was an important factor in mortality in our cases, among whom a weight-for-age of less than 60% was found to be a significant predictor of death from severe pneumonia (OR 4.02; 95% CI 2.13-7.55; p<0.0001).
Escobar et al. documented a seven-fold increase in mortality among moderately malnourished children in the hospital with pneumonia. Central cyanosis is the best clinical sign of hypoxaemia and was evident in 56% of our cases. The detection of such a large number of cases may be explained by late attendance at the hospital, but the high altitude almost certainly increased the risk of hypoxaemia in these vulnerable children. Rachitic changes in the thoracic cage associated with rachitic myopathy may also have impaired pulmonary ventilation in our cases of severe pneumonia.
The high incidence of cyanosis in those who died is similar to that reported by others who also identify cyanosis as a major risk factor for death from pneumonia. Oxygen is given to seriously ill children, usually with no benefit, except when given early in the course of the disease when it is more effective. Clinical rickets was a striking feature in our study, affecting half of the patients, of whom 82% were under 1 year of age. Several children manifested rickets at 1 month of age, indicating that the rachitic process probably begins in utero.
Yemeni women lack exposure to sunlight because they are confined to their homes and are totally covered in dark clothes when outdoors. Similarly, newborns and young infants are traditionally well covered because of fear of ‘evil eyes’ and changeable weather and to preserve a pale skin color. These traditions block ultraviolet radiation which, combined with lack of dietary vitamin D, causes severe vitamin D deficiency in Yemeni females and their children. Calcidiol blood levels in pregnant women and their babies living under the purdah system were found to be significantly lower than those not observing purdah. Pregnant women and their newborn infants in Libya and Saudia Arabia have been reported to be severely vitamin D deficient. Breastfed infants in these situations develop clinical rickets before showing general malnutrition.
Both rickets and malnutrition were found in our study patients, although rickets was not directly associated with malnutrition but was a separate clinical feature. Biochemical characteristics of rickets in infants tend to be disguised by malnutrition.
The frequency of pneumonia in rachitic children may be the consequence of general muscle weakness, soft under-mineralized ribs and a variety of chest deformities, but there is also evidence that rickets impairs the immune system, leading to impaired phagocytosis by neutrophils. Even children with sub clinical rickets tend to be susceptible to respiratory infections, and clinical rickets has been reported to increase susceptibility to infection in general among children. It has long been recognized that respiratory infections are more dangerous and frequently lethal in the presence of severe rachitic chest deformities.
Anemia (Hb <10g/dl) was found in 30% of our cases, mainly in the first year of life. Several factors contribute to this high incidence of anemia, including short birth interval among women who are themselves anemic and produce infants with low iron stores, and the common habit, mainly in traditional home deliveries, of allowing the umbilical cord to bleed for some moments before ligation.
The prevalence of breast feeding exceeds 94% in the catchment area of our study but cereals are introduced early, frequently before 3 months of age. The advantage of good absorption of iron from breast-milk is impaired when solids are introduced. The prevalence of nutritional anemia has been reported to be greater in children on a mixed diet of breast-milk and solids (73%) than those who are fully weaned (30%).
All this may explain the increased prevalence of anemia among our cases. It is unlikely that the anemia in our study was due partly to malaria infection since malaria seldom occurs in the catchment area and during the study period only 62 cases of proven falciparum malaria were admitted to the pediatric ward.
Anemia and iron deficiency impair cell-mediated immunity and decrease bactericidal activity of neutrophils. Respiratory infections have been reported to be more frequent in children with iron deficiency anemia than in non-anemic children. Although not included in the WHO criteria for severe pneumonia, heart failure was a significant clinical feature in our study. The association of heart failure with severe pneumonia has been found in other studies.
Severe rickets may be a factor in the development of heart failure in our cases since it grossly impairs pulmonary ventilation, leading to terminal right-sided heart failure, as has been reported in Ethiopia and Mexico. It is worthy of note that the altitude where these studies were conducted is similar to that of Sana’a.
The specific risk factors identified in this study – weight-for-age less than 60%, clinical rickets, hemoglobin less than 10 g/dl, cyanosis and heart failure – are preventable. Correction of nutritional deficiencies in pregnancy including vitamin D and iron, and education in health and nutrition including promotion of breast feeding as well as early introduction of vitamin D and iron supplements to newborns could dramatically reduce pneumonia-related mortality. In clinical practice, early and adequate oxygen administration to infants with pneumonia may improve prognosis.
Najla N. Al-Sunbali Salem M. Banajeh Saleh H. Al-Sanahani Department of Pediatrics, Al-Sabeen Hospital for Women and Children, Sana’a