Burden of disease due to microcephaly associated with the Zika virus in Colombia Carga de enfermedad debido a la microcefalia asociada al virus del Zika en Colombia A carga de doença por microcefalia associada ao vírus Zika na Colômbia

In 2015, the Zika virus was introduced in Colombia. The emergence of this arbovirus is a public health challenge for the country, considering the association between the infection and congenital disorders such as microcephaly. Thus, we estimated the burden of disease due to microcephaly associated with Zika in Colombia and its administrative subdivisions for the period 20152016. We conducted an exploratory ecological study, using as unit of measurement disability-adjusted life years (DALYs). The cases of microcephaly were obtained from the Zika national and departmental databases built by the National Public Health Surveillance System (SIVIGILA). Deaths attributed to microcephaly were estimated from previous studies. Finally, we calculated mortality rates and incidences, then we performed a sensitivity analysis under three scenarios (conservative, medium, and extreme) to estimate the DALYs. In the 2015-2016 period, 10,609.4 DALYs were caused by microcephaly associated with Zika in Colombia. 71% of the total DALYs were years of life lost and 29% were years lived with disability. Five out of 32 departments (Meta, Córdoba, Tolima, Valle del Cauca, and Norte de Santander) contributed 71% of total DALYs. The burden of microcephaly associated with Zika outweighed the burden of other congenital anomalies such as neural tube defects and Down syndrome in children aged between 0 and 4 years in Colombia. Public health efforts must be made to prevent and monitor these cases.


Introduction
Although the acute Zika virus infection is generally benign, the virus has been associated with severe complications. Firstly, neurological disorders in general population, especially Guillain-Barré syndrome 1,2 . Secondly, congenital anomalies, including congenital microcephaly and fetal losses in women infected during pregnancy 3,4 .
The teratogenic effects of the virus were initially suspected after the increasing number of microcephaly cases reported in Brazil and retrospectively in French Polynesia 5,6 . Given the epidemiological and histopathological evidence collected since the first cases of microcephaly and other neurological malformations, currently, the Zika virus can be considered the causal agent of these alterations 3 . Symptomatic and asymptomatic Zika-infected pregnant women are at risk for vertical transmission of the virus. Zika infection during pregnancy can lead to severe birth defects, however, there is a higher risk of congenital anomalies in the first and second trimester of pregnancy 4,7 .
Between 2015 and the first epidemiological week of 2017, 714,636 autochthonous cases were reported in the Americas, out of which 178,297 (25%) had been confirmed. Brazil and Colombia were the most affected countries, reporting 321,366 and 106,552 cases respectively. During this period, 2,530 cases of microcephaly were reported, out of which 90% occurred in Brazil 8 .
Epidemiological surveillance of Zika virus disease and its complications in Colombia is conducted by SIVIGILA (Colombia's National Public Health Surveillance System). SIVIGILA records morbidity and mortality of diseases or specific health events such as communicable diseases, nutritional surveillance, chronic non-communicable diseases, congenital defects, maternal and perinatal mortality, external cause injuries and mental health, and mortality in children under five years due to acute diarrheal disease, acute respiratory infection, and malnutrition 9 . These events are recorded in mandatory notification forms. SIVIGILA collects and analyzes those forms, and subsequently generates systematic reports, immediately or weekly 10 . This information guides policies and planning in public health; it is the base for decision-making regarding disease prevention and control, and it also optimizes monitoring and assessment of interventions and rational use of resources available 11 .
In Colombia, the first case of Zika reported to SIVIGILA occurred during the epidemiological week 32 of 2015 12 . From the introduction of the virus until the end of 2016, a total of 106,659 cases were notified, out of which 18.5% corresponded to pregnant women (6,363 confirmed cases, 13,383 suspected cases) 13 . Considering the significant number of Zika-infected pregnant women during the 2015-2016 period and the association between infection and the development of microcephaly, we decided to estimate the burden of disease due to microcephaly associated with Zika virus in Colombia, following the methodology proposed by Murray 14 .

Type of study
Ecological exploratory study aimed to estimate the burden of disease due to microcephaly associated with the Zika virus in Colombia, during the years 2015-2016. Disability-adjusted life years (DALYs) were used as unit of measurement and estimated under three scenarios (conservative, medium, and extreme), following the methodology proposed by Murray 14 .

Information sources
We obtained data on microcephaly cases reported to the SIVIGILA (https://www.ins.gov.co/Direc ciones/Vigilancia/sivigila/) during the 2015-2016 period. Based on the study of Nembhard et al. 15 , we estimated the number of deaths caused by microcephaly. According to Nembhard et al. 15 , the firstyear survival rate of microcephaly is 79.7% (95%CI: 73.6-84.6).
Demographic data on infants (< 1-year-old) both at national and departmental levels for 2015-2016 were obtained from the National Administrative Department of Statistics (DANE) 16

Morbidity and mortality
Cases of acute Zika virus infection were registered in the SIVIGILA databases under mandatory notification forms 17 . In addition to the acute cases of the disease, these databases also report microcephaly cases related to the infection. Microcephaly is defined as a head circumference more than two standard deviations below the mean for a given sex, age, and ethnicity 18 . Reference values for Colombia are based on the WHO Child Growth Standards and the Resolution n. 2,465 of June 14, 2016 19 .
We counted the microcephaly cases from the SIVIGILA databases according to the department of origin. The result represents the number of cases alive after the first year of life. Incidences per 1,000 infants were calculated at the national and departmental levels. The number of deaths attributed to microcephaly was determined by subtracting the number of cases alive after the first year of life to the number of microcephaly cases reported to SIVIGILA. Mortality rates per 1,000 infants were calculated at the national and departmental levels.
According to the Global Burden of Disease, Zika and congenital Zika syndrome deaths are classified in group I: communicable diseases, maternal and perinatal disorders, and nutritional deficiencies 20 .

Burden of disease (DALYs)
The burden of disease was expressed in terms of DALYs, following the methodology proposed by Murray 14 . DALYs are the sum of the years of life lost (YLLs) and years lived with disability (YLDs). YLLs were estimated based on microcephaly mortality rates by year and department. YLDs were estimated according to parameter values defined in previous studies (Table 1) 15,21,22 .
There is no disability weight for microcephaly available. We use the disability weight for severe intellectual disability (disability weight: 0.16) according to Salomon et al. 22 and Alfaro-Murillo et al. 23 .
Therefore, we used the duration and survival rates by age (after the first year of life) for intellectual disability (mental retardation) measured by Honeycutt et al. 21 .

Analysis of information
Statistical analyses and DALYs estimates were performed using Microsoft Excel software (https:// products.office.com/). We use the World Health Organization (WHO) burden of disease templates (https://www.who.int/healthinfo/global_burden_disease/tools_national/en/), following the methodology proposed by Murray 14 .

Sensitivity analysis
DALYs for microcephaly were estimated under three scenarios: conservative, medium, and extreme. The distributions for the parameter values vary by scenario as follows: (1) mortality rate by microcephaly after the first year; and (2) the duration of the disease. The disability weight and the first-year survival rate were constant for the three scenarios (Table 1).  (Tables 2 and 3). The mortality rate and incidence of microcephaly associated with Zika nationwide increased 26,2 times from 2015 to 2016. By department, Córdoba had the highest mortality rate and incidence during 2015, while Meta, Tolima, and Córdoba had the highest mortality rate and incidence during 2016 (Tables 2 and 3).

Figure 1
Disability-adjusted life years (DALYs) due to microcephaly associated with Zika at the departmental level. Colombia, 2015-

2016.
Note: five out of 32 departments (Meta, Córdoba, Tolima, Valle del Cauca, and Norte de Santander) contributed 71% of Regarding sensitivity analysis results, in the conservative scenario at least 10,173.16 DALYs were lost due to microcephaly associated with the Zika virus in the period 2015-2016 at the national level. On the other hand, in the extreme scenario at most 10,860.57 DALYs were lost. The difference between these scenarios is 687.41 DALYs (Tables 2 and 3).

Discussion
Acute Zika virus infection is usually mild and self-limiting; however, chronic complications represent a public health challenge for Zika-endemic countries. Our results highlight the importance of congenital anomalies related to Zika in Colombia. Out of the 19,746 cases of acute Zika virus disease reported in pregnant women 13 , 5.7% of newborns developed microcephaly.
Colombia experienced, especially during 2016, an increase in microcephaly cases compared to previous years. During 2014 and 2015, the number of microcephaly cases per epidemiological week was similar, but in 2016, the cases increased, reaching a peak in the epidemiological week 28 (23 weeks after the peak of acute cases) 24 . In 2016 and the first three epidemiological weeks of 2017 the number of microcephaly cases increased compared to the same periods in 2014 and 2015. Therefore, it is possible to argue that the incidence of microcephaly reported during 2016 was associated with the Zika epidemic happening between the epidemiological week 32 of 2015 and the epidemiological week 28 of 2016 (according to SIVIGILA, the epidemiological week 28 of 2016 was the beginning of the endemic phase in Colombia 13 ).
In our literature review, we only found one published article on the burden of disease due to microcephaly associated with Zika. Alfaro-Murillo et al. 23 estimated 29.95 DALYs lost per every microcephaly case. This value is higher than our estimate (9.48 DALYs). The difference between Alfaro-Murillo et al. 23 estimate and ours is explained by distinct parameter values used for estimating the DALYs (except disability weight) in both studies. Also, their methodology assumes attack rates range between 23.5% and 77% and only first-trimester Zika infections can cause microcephaly 23 . On the other hand, their expected microcephaly cases attributable to Zika were based on previous studies, while we used the SIVIGILA database to determine the number of microcephaly cases associated with the Zika virus in Colombia.
In comparison to other diseases, we found that our 2016 microcephaly estimate (10,220.64 DALYs) exceeded the DALYs caused by other congenital anomalies, according to the Global health estimates 2016 published by the WHO 25 . For instance, 8,900 DALYs were caused by neural tube defects, 2,700 DALYs by Down syndrome, and 300 DALYs by cleft lip and cleft palate. The burden of Zika-associated microcephaly is comparable to the burden of disease encompassing "other chromosomal anomalies"  (10,700 DALYs) 25 . In summary, microcephaly associated with Zika was the fourth leading cause of DALYs lost by congenital anomalies in Colombia during 2016. Regarding Brazil, the country with the highest number of microcephaly cases reported in the world, it is even more difficult to estimate DALYs attributable to this condition. The definition of microcephaly in Brazil changed three times in the 2015-2016 period affecting the number of reported cases 26,27 . In Colombia, the prevalence of microcephaly increased fourfold between 2015 and 2016, while Brazil's prevalence increased ninefold compared with the previous 14 years 28 . Consequently, the estimation of the burden of disease caused by Zika-associated microcephaly in Brazil is even more complex than in Colombia.
According to the 57th Epidemiological Report, 2,366 confirmed cases were reported in Brazil between November 8, 2015, and December 31, 2016 29 . If we assume that 9.48 DALYs were caused by Cad. Saúde Pública 2020; 36(11):e00215319 each microcephaly case in Brazil too, the total number of DALYs lost due to microcephaly in Brazil was 22,429.68 DALYs during that period. In comparison with WHO global health estimates 2016, microcephaly surpassed the burden of disease attributable to cleft lip and cleft palate (3,200 DALY) in Brazil 25 . However, the number of cases could be even higher, given that 3,183 cases were under investigation at the time the report was published 29 . In the worst-case scenario, DALYs caused by microcephaly could be comparable to DALYs attributable to Down syndrome (33,400 DALY) in Brazil 25 .
There are some limitations in this study. Firstly, underreporting affects estimates of the DALYs caused by microcephaly associated with the Zika virus. Surveillance systems are not able to detect every case of Zika virus infection and possibly neither every microcephaly case. Thus, this underreporting translates into an underestimation of the real burden of disease. Generally, Expansion factors are used to correct the underreporting 30  After the first Zika-associated microcephaly case was identified in Brazil, the criteria (head circumference below 2 or 3 standard deviations) used to define microcephaly has been the subject of discussion. We applied the Colombian reference growth curve for head circumference (head circumference below 2 standard deviations) 19 . However, abnormal skull shapes and inter-rater differences between clinicians and ultrasonographers may lead to an erroneous diagnosis of microcephaly 32,33 .
It is essential to mention that we focused on one of the most important clinical manifestations of congenital Zika syndrome (microcephaly), however, it does not reflect the entire burden of disease due to congenital anomalies associated with the Zika virus infection. The spectrum of congenital manifestations associated with the Zika virus is broad, and new clinical conditions related to Zika are being described. On the other hand, the risk of occurrence of congenital anomalies caused by the Zika virus infection has not accurately been estimated yet. Previous risk estimates range between 6% and 42% 4,34,35 . Although microcephaly was the most alarming congenital condition identified by the scientific community, the risk of developing microcephaly not as high as it was believed at the beginning of the epidemic. According to Cauchemez et al. 6 , the risk of microcephaly increases to about 1% in pregnant women who get infected during the first trimester. Subsequent analyses indicate that this risk could reach up to 13.2% 6,36 . Other studies have found prevalences of 3.4%-4% 4,34 . In summary, although microcephaly is a distinctive feature of the congenital Zika syndrome, its presence should raise awareness on the diagnosis and treatment of other congenital anomalies associated with Zika virus infection affecting an endemic community. Similarly, the burden of disease caused by Zika-associated microcephaly represents a fraction of the total burden of disease attributable to the congenital Zika syndrome.
Likewise, cohort studies have revealed that the disability weight for microcephaly must include more clinical manifestations than intellectual disability. The ZODIAC Study (the outcomes of Zika and development in infants and children) is collecting information related to the health and development of children older than 12 months who were born with microcephaly 37 . According to their first preliminary report, 74% to 79% of children develop cerebral palsy and severe motor disorders, based on the analysis of a subgroup of 19 children aged 19-24 months. Other less frequent findings are epilepsy, auditory and visual disturbances, and sleep disorders 37 . In conclusion, the real burden of disease due to Zika-associated microcephaly is higher than our results.
On the other hand, van der Linden et al. 38 , conducted a case series of 13 infants who were infected with Zika before birth and normal head size. During the follow-up period, 11 children develop postnatal microcephaly as early as five months of age. Postnatal microcephaly was not included in our estimates since these cases are not reported in mandatory notification forms. SIVIGILA databases were built with forms filled at the moment of diagnosis, either in the emergency room, a prenatal consultation or immediately after childbirth.
The main strengths of the study are as follow: firstly, we had access to primary data (SIVIGILA database). Every microcephaly case should be reported in SIVIGILA; hence, the database accurately reflects the microcephaly situation in Colombia. Secondly, this is one of the first studies to estimate Cad. Saúde Pública 2020; 36(11):e00215319 the burden of disease due to microcephaly related to the Zika virus infection in Colombia. Lastly, the burden of disease was estimated in three possible scenarios. The range of results obtained will facilitate comparisons with future studies. The entire spectrum of clinical manifestations, especially congenital anomalies caused by the Zika virus infection, have not been identified, so it is not possible to estimate the total burden of disease attributable to Zika. However, we expect that our results will raise awareness among control agencies, not only at the national level but also in all countries affected by the Zika epidemic. Children born to mothers with confirmed or suspected Zika virus infection during pregnancy need long-term monitoring and extensive epidemiological surveillance should be implemented in endemic areas. Likewise, routine screening during pregnancy should be considered in pregnant women at high risk of Zika virus infection.

Conclusion
This study is the first attempt to estimate the burden of disease due to microcephaly associated with the Zika virus in Colombia. The distribution of the burden of disease was not homogeneous; five out of 32 departments contributed 71% of the total burden caused by microcephaly. On the other hand, more DALYss were lost due to premature deaths than to chronic cases. This analysis prioritizes the departments requiring further public health action, including prevention, surveillance, and medical care for congenital Zika syndrome in Colombia. Considering the limitations of this study, new research questions emerged to accurately estimate the burden of disease caused by the Zika virus in the future: What is the disability weight of the Zika virus-associated microcephaly? What are the life expectancy and quality of life in patients with Zika virus-associated microcephaly? The Global Burden of Disease and the WHO Global Health Estimates groups should add, in their future burden of disease estimates, the Zika virus infection and its related neurological and congenital anomalies, given its worldwide distribution and its associated morbidity and mortality.

Contributors
A. F. Mora-Salamanca searched literature, designed the study, performed data analysis, and elaborated the discussion. A. Porras-Ramírez designed the study, performed data analysis, and elaborated the discussion. F. P. H. Restrepo performed data analysis, elaborated the discussion, and reviewed the manuscript.