An evaluation of Brazil's surveillance and prophylaxis of canine rabies between 2008 and 2017

Abstract

An effective surveillance system is critical for the elimination of canine rabies in Latin America. Brazil has made substantial progress towards canine rabies elimination, but outbreaks still occurred in the last decade in two states. Brazil uses a health information system (SINAN) to record patients seeking post-exposure prophylaxis (PEP) following contact with an animal suspected of having rabies. This study evaluated: (i) whether SINAN can be reliably used for rabies surveillance; (ii) if patients in Brazil are receiving appropriate PEP and (iii) the benefits of implementing the latest World Health Organization (WHO) recommendations on PEP. Analysing SINAN records from 2008 to 2017 reveals an average of 506,148 bite-injury patients/year [range: 437k-545k] in the country, equivalent to an incidence of 255 bite-injuries/100,000 people/year [range: 231-280]. The number of reports of bites from suspect rabid dogs generally increased over time. In most states, records from SINAN indicating a suspect rabid dog do not correlate with confirmed dog rabies cases reported to the Regional Information System for Epidemiological Surveillance of Rabies (SIRVERA) maintained by the Pan American Health Organization (PAHO). Analyses showed that in 2017, only 45% of patients received appropriate PEP as indicated by the Brazilian Ministry of Health guidance. Implementation of the latest WHO guidance using an abridged intradermal post-exposure vaccination regimen including one precautionary dose for dog bites prior to observation would reduce the volume of vaccine required by up to 64%, with potential for annual savings of over USD 6 million from reduced vaccine use. Our results highlight the need to improve the implementation of SINAN, including training of health workers responsible for delivering PEP using an Integrated Bite Case Management approach so that SINAN can serve as a reliable surveillance tool for canine rabies elimination.

Fig 1. Algorithm used to evaluate the appropriateness of the PEP administered as recorded in each SINAN form according to the guidelines of the Brazilian MoH.

Rectangles show the Brazilian MoH’s recommendation (see S2 Table) based on the criteria filled on the SINAN form (circles). For patients requiring PEP and/or serum, the full algorithm included a further step discriminating between individuals that previously received complete PEP at least 90 days prior to the current exposure, in which case they did not need either vaccine or serum. Patients requiring serum and previously receiving a pre-exposure prophylaxis were considered as patients without previous PEP because the form does not include whether their titer levels were >0.5IU/ml, in which case serum is not recommended.

Fig 2. Bite incidence in each state of Brazil during 2016.

The main bar plot shows the bite incidence (i.e. number of patients seeking health care after a bite per 100, 000 people) in each state during 2016. The map shows the same distribution in space with states coloured according to their bite incidence. The inner plot shows the country’s average bite incidence per year from 2008 to 2016. Country and state maps were obtained from the GADM (http://www.gadm.org/) database using the getData function from the raster package of R.

Fig 3. Changes in the percentage of dogs assessed as suspect for rabies between 2008 and 2017.

A) Percentage of reports including a dog assessed as ‘suspect’ for rabies (i.e. assessed by a health worker as ‘rabies suspicious’, ‘rabid’ or ‘dead/disappeared’) per year in Brazil. The relative percentage of dogs assessed as ‘non-observable’ for a 10-day observation period (green), ‘observable’ (colour) or ‘no-data’ (brown) are shown within each bar. (B) Map showing the average percentage of dogs reported as ‘suspect’ for rabies annually per state between 2008 and 2017. (C) Map showing changes in this percentage from 2008 to 2017 across states. Positive numbers illustrate an increase in the percentage of suspect dogs. Country and state maps were obtained from the GADM (http://www.gadm.org/) database using the getData function from the raster package of R.

Fig 4. Comparison between rabid dogs reported to SINAN and SIRVERA.

A) Map showing the number of dogs reported as ‘laboratory confirmed’ for rabies in the SINAN database per state between 2008 and 2017. B) Map showing the number of dogs reported as ‘laboratory confirmed’ for rabies in the SIRVERA database per state during the same time period. States coloured in blue did not report a positive case. C) Each plot shows the monthly number of positive cases in both SINAN and SIRVERA databases for selected districts from January 2008 to December 2017. Country and state maps were obtained from the GADM (http://www.gadm.org/) database using the getData function from the raster package of R.

Fig 5. Percentage of appropriate PEP use and number of vaccine doses under different scenarios of PEP regimes.

A) Percentage of cases receiving PEP from the total number of patients attended each year in Brazil (black line) and percentage of cases receiving the ‘appropriate’ recommended PEP according to the Ministry of Health guidelines (blue line). B) Estimated number of vaccine doses required to appropriately attend all patients during 2017 under three hypothetical scenarios: (i) the 2016 protocol in Brazil comprising 5 doses of vaccine according to the intramuscular (IM) Essen regimen (ii) the updated 2017 Brazilian guidelines using the updated IM Essen regimen with 4 doses and (iii) the latest WHO position comprising 3 intradermal (ID) doses. The light grey bar within the latest WHO regime corresponds to the number of vaccines needed in a scenario where vials would be shared completely between patients.