Sickle cell disease (SCD) comprises a group of hematologic disorders of genetic origin, whose main feature is the predominance of hemoglobin S (HbS) in red blood cells. SCD has great clinical and epidemiological importance, being considered a public health problem in Brazil.1–3HbS in homozygous form (HbSS), called sickle cell anemia (SCA), is the most common genotype and the most severe clinical presentation of the disease. The concept of SCD also includes double heterozygosity for HbS with other hemoglobin variants (HbSC, HbSD-Punjab, and others) and the interaction of HbS with beta thalassemia (HbS/β0 and HbS/β+ thalassemias).4,5 The main determinants of the clinical manifestations of the disease are the vaso-occlusive phenomena and chronic hemolysis.4

Hemoglobinopathies are responsible, worldwide, for 3.4% of deaths in children under 5 years.6 Patients with SCD have high rates of morbidity and mortality, mainly in developing countries.7 Even in developed countries, although mortality in children has decreased, SCD is still a significant cause of mortality in adolescents and adults.8

A prevalence of 25,000-30,000 individuals with SCD is estimated in Brazil, with an incidence of 3,500 cases per year.9 In the state of Minas Gerais (MG), the incidence of SCD is 1:1,400 screened newborns, based on data from the Neonatal Screening Program (NSP-MG).10

A study conducted in MG10 characterized the death of 78 children with SCD screened by the NSP-MG in the period of 1998-2005 and concluded that most deaths occurred in children under 2 years and individuals with HbSS, mainly due to infection (septicemia and pneumonia). A predominance of in-hospital deaths was observed; however, the occurrence of at home or in-transit deaths was still significant. An indeterminate cause of death on the death certificate was frequent, which would indicate a lack of knowledge among physicians regarding SCD and acute events that are determinants of death.

This study aimed to characterize the deaths of children with SCD that had been screened by the NSP-MG between 1998 and 2012. Whenever possible, comparative analysis of deaths of children born between March of 1998 and February of 2005 with those born between March of 2005 and December of 2012 was performed to describe possible changes regarding the quality of care provided to these children.

The population investigated in this study consisted of 117 children diagnosed with SCD at the NSP-MG, followed-up in the Blood Banks of Fundação Hemominas, and who died in the period from March 1, 2005 to February 29, 2012. In several analyses, 76 deaths that occurred between March of 1998 and February of 2005 were added to the sample (of the 78 deaths reported, two children were excluded, whose birth was prior to March 1,1998),10 totaling 193 deaths and thus complementing the 14 years of the study.

The NSP-MG performs screening for SCD since March of 1998. All 853 state municipalities are registered. In 2010, 91.44% of the infants born in the state of Minas Gerais were screened at the Center for Diagnostic Support Action and Research (Núcleo de Ações e Pesquisa em Apoio Diagnóstico [NUPAD]), and free medical treatment and follow-up was offered to those born with SCD.

The information on death results from an investigation conducted by the Control and Treatment Department of NUPAD due to non-attendance at scheduled consultation at the Blood Bank. The active searches performed encompass all screened patients with SCD followed-up at the outpatient clinic of Fundação Hemominas throughout the state. Therefore, it can be stated that all deaths that occurred during the study period were recorded.

The data for this research came from the NUPAD database, death certificates, the Brazilian Unified Health System database (DATASUS), the Brazilian Institute of Geography and Statistics (Instituto Brasileiro de Geografia e Estatística [IBGE]), medical records of Fundação Hemominas, and from interviews with the children's family members.

The information in the NUPAD database was used to perform an active search of the 117 families whose children had died between 2005 and 2012. The interview was performed with the child's guardian, using a semistructured form used by Fernandes et al.10 A total of 81 families were interviewed by the researcher and three families were interviewed by nurses from the referral Basic Health Units.

Five families refused to be interviewed, one was unable to talk about the subject, and 27 were not located. The interview covered five aspects: (1) death-related circumstances; (2) follow-up of SCD and the occurrence of other acute events; (3) family's socioeconomic and cultural conditions, which considered per capita income, education, parents’ occupation, and housing conditions; (4) considerations about their experience with the child; (5) applying the scale of knowledge on SCD to the families, consisting of 20 questions with answer options: “right”, “wrong”, or “I don’t know.” The interview was performed only after the informed consent form was signed, which encompassed the objectives, purpose, and benefits of the research.

Regarding the causes of death, they were considered as indeterminate when the physicians showed ambiguity in the medical records of the death certificate, regardless of where the death occurred and whether the child received assistance, thus making it impossible for researchers to determine its cause.

For the statistical analysis, comparisons of frequencies between nominal variables were made using the chi-squared test without continuity correction. Tests with a probability of alpha error ≤ 0.05 were considered significant.

The study was approved by the Ethics Committee of Universidade Federal de Minas Gerais (UFMG) and Fundação Hemominas.

The sample reflects the overall health reality of children with SCD screened during the 14 years of the NSP-MG, as all deaths of children followed at Blood Bank Clinical Centers were informed. This represents the total number of deaths, which configures a population-based study.

The present study, as others,11,12 observed no significant difference between genders in the incidence of deaths in the assessed age range.

This and several other studies11,13–17 have demonstrated that the highest frequency of deaths occurred in patients with HbSS (79.3%), the most common and most clinically severe genotype. The relatively high frequency of deaths in children with Sβ+ thalassemia (6.5%) does not correspond to reality, as only from 2010 onwards all suspected cases of SCD in NSP-MG were confirmed by molecular biology techniques. The diagnosis of this subtype of SCD may be overestimated in the neonatal period due to blood transfusions not reported by family members or errors in the interpretation of small concentrations of HbA.

Three quarters of the children diagnosed between 2005 and 2012 underwent neonatal screening for SCD in up to 8 days of life, less time than the 18 days in the initial period between 1998 and 2005,10 which shows improvement of the NSP-MG.

The distribution of children per age range at the time of death shows increased occurrence of deaths in those younger than 5 years, similar to that observed by other authors.11,18–22 Nonetheless, some authors4,8,12,16,23,24 have reported that infant mortality rate due to SCD is decreasing and the median age of death is increasing. In the present study, the high mortality in children younger than 5 years is explained by the higher incidence of potentially fatal events in this age group, such as infections and ASS.

Although in-hospital deaths were the most prevalent, it is noteworthy that one-fifth of them occurred at home or in-transit. This can be explained by several factors, such as family residence located far from the health service, difficulty getting to the service, low socioeconomic and educational levels, in addition to the family's incapacity to identify risk situations for the child. The statistically non-significant increase of 11% in the number of in-hospital deaths shows improvement, albeit small, in the health care of children with SCD in the last six years. A study conducted in Angola25 concluded that patients with SCD who lived in the countryside or in rural communities were three times more likely to die during hospitalization, due to socioeconomic factors and lack of immediate access to medical care.

In the present study, as in many others,11,14,15,18,25–27 infection appears as the main cause of death, considering the age groups studied. The high number of indeterminate causes suggests a difficulty for physicians and health care professionals to recognize SCD and its severe acute events determinants of death, as in the study by Alves,28 given that most of the records came from in-hospital deaths. ASS was the third most common cause, which reflects the difficulty by health teams and families to identify this complication. The proportion of deaths caused by ASS in SS children was similar to that reported in a previous study, restricted to SS children.29 Other authors13,15,19 indicated ASS as the main cause of death in children younger than 3 years.

A significant number (46%) of death certificates did not mention the term “sickle cell” as the cause of death. However, the number of certificates that mention this term has increased significantly over the past seven years, indicating greater visibility and awareness of the disease by physicians.

Most children received medical care before death, but 25% of them were unassisted. An increase, albeit non-significant, of 13% in the frequency of medical care provided to children with SCD before death was observed, which once again indicates a probable increase of the knowledge about SCD by physicians.

Medical care occurred within the first 6hours in almost half the cases and two-thirds in the period of up to 24hours. Death occurred within the first 24hours of symptom onset in approximately 40% of cases. Manci et al.15 demonstrated that the first 24hours after the onset of the pathological process represent the period of greatest risk of death for patients with SCD. This emphasizes how fast the clinical condition worsens, culminating in death.

High adherence to antibiotic prophylaxis and folic acid, in addition to the regular use of special immunobiologic agents indicates that these data may be overestimated, as they were obtained a posteriori and in the presence of an “authority figure”, i.e., the interviewer. A study conducted at the Blood Bank Clinical Center of MG showed lower levels of adherence to antibiotic prophylaxis.30 The incorporation of “special vaccines” to the basic vaccination schedule occurred only recently, and may have contributed to the low adherence of families. This is probably due to the lack of awareness by health professionals regarding the special immunization schedule and the lack of priority when requesting special immunobiologic agents by the municipalities.

At the time of the child's death, only little more than half of the families regularly received visits from the FHP team. Considering that SCD is a chronic disease with the possibility of severe acute events, in addition to the fact that it affects low-income families, better monitoring by the FHP through home visits was expected, in order to ensure the inclusion of these patients in the health care system.

Lack of priority and/or delay in treatment, lack of knowledge by the team and poor health service infrastructure were difficulties reported by the families that may have contributed to death. Low parental educational level and low income constitute additional barriers. A study conducted in Gabon, Africa,21 also reported that the vast majority (91.3%) of children with SCD who died were from families of low socioeconomic status. Inadequate transport conditions were also reported by some families, such as oxygen cylinder with insufficient or absent volume for the entire transfer, inadequate oxygen mask size for the child, among others.

The cognitive questionnaire administered to 41 families indicates that knowledge about SCD is still precarious, even after the occurrence of such an important event as the death of a child.

The main limitation of the study was its retrospective design, as locating the families becomes more difficult over the years, as well as their incapacity to accurately report some of the data. The interviews added a more accurate qualitative character to the study, rather than a simple assessment of death certificates. In the authors’ opinion, the absence of 33 interviews, in 117, did not impair the desired goals.

It can be concluded that low socioeconomic and educational levels of families, difficulties in access to health services, lack of transportation, lack of priority in health care, and lack of knowledge about SCD by health professionals were aggravating factors for the clinical status of the child.

Screening for SCD, even when performed in a comprehensive and effective way, was still not enough to reduce the mortality rate observed at the NSP-MG. Social and economic development in the state are necessary to change this scenario. Care of acute events, consultations with specialists, adequate transportation of patients, and continuing education for families and health care teams should be recognized as a priority in the care of patients with SCD.

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) and Núcleo de Ações e Pesquisa em Apoio Diagnóstico (Nupad).

The authors declare no conflicts of interest.