Iran had an unusual rate of death in fall 2019 before the COVID-19 outbreak.
Iran was among the first countries to report COVID-19 infections. With 97 exported cases to 11 countries by Feb 28, it has since been the epicentre of the outbreak in the Middle East. The country has been widely accused by the media and general public for underreporting and manipulating its COVID-19 records.
Underdetection and the resulting underreporting of COVID-19 cases due to limited testing capacity is a major barrier to the effective management of the Coronavirus crisis around the world. Iran is no exception and the lack of testing capacity has limited the country’s ability to monitor and detect active COVID-19 cases, especially for those with no or mild symptoms. Like in many other countries, the true number of active and recovered cases in Iran is believed to be many times bigger than the reported numbers.
How about the death records?
Because of all the complications with testing active cases, experts have tried to gauge the burden of the outbreak using death records. Their general assumption is that the reported COVID-19 death data are reliable as most countries are capable of effectively tracking the COVID-19-driven deaths. Yet, the story is different when it comes to Iran.
The validity of Iran’s death records have been also put into question. In early March, Iran’s COVID-19 national emergency response committee decided to stop the release of deaths data registered by the National Organization for Civil Registration (NOCR). This decision raised further suspicions that Iran is trying to hide the truth.
No one knows what the truth is but curious minds still try to explore what has been happening in Iran. In the absence of reliable data, studies focusing on the COVID-19 outbreak in Iran have to rely on major assumptions and speculations. That is why some of their findings are sometimes significantly different. At the end of the day, it is impossible, at least at this point in history, to verify which study is providing the best estimation of reality.
While the number of registered deaths in the past few months is still unknown, Iran’s NOCR has recently publicized the registered death records in winter (22 December 2019 to 19 March 2020). One can have different reasons to suspect the accuracy of this data. But examining the available data in search for new insights is not always a totally worthless exercise.
Where was impacted by COVID-19 in winter?
NOCR only reports seasonal deaths. So, we are only left with four time points per year. The data does not tell us about the death causes. So, we do not know if people have died due to cancer, heart attack, accident, suicide, COVID-19, another virus, or other reasons. To explore the COVID-19 impact, we can only examine if death patterns have behaved differently from the past. In other words, we need to search for patterns of excess mortality across the country.
By comparing the deaths during last winter to the 5-year average death in winter for each province, we detect a clear COVID-19 outbreak signal. Qom, Gilan (Guilan), Mazandaran, Golestan, and Alborz provinces that were reportedly hit very hard by COVID-19 showed a sharp rise in excess mortality.
The winter excess mortality was over 20% in Alborz, 30% in Golestan and Mazandaran, 40% in Gilan, and 50% in Qom, with Gilan having the highest rise in terms of the absolute number of excess deaths (1,475 excess deaths compared to last year). Tehran did not have a significant percentage increase in mortality due to its large population even though the absolute number of excess mortality in this province (1,032 people) was not insignificant.
Iran reported 1,433 COVID-19-related deaths till the end of its winter season (before Nowruz, the Persian new year) whereas total excess mortality across the country during winter compared to last year was 4,565 people with 3,846 of that belonging to Alborz, Golestan, Mazandaran, Gilan, and Qom provinces. The winter excess deaths cannot be entirely attributed to COVID-19. But, if valid, the NOCR data clearly suggests that these provinces have severely suffered from the COVID-19 outbreak during winter while there was no significant COVID-19 impact on other provinces.
What else does the NOCR data tell us?
Looking at the registered deaths in the last five years, we see a trend towards more deaths during the cold seasons (fall and winter). So like in many other countries, generally, we expect more people dying in the cold seasons in Iran when mortalities due to seasonal influenza are common. It is expected to see the number of seasonal deaths vary between years but winter deaths are usually higher than fall deaths in Iran.
Upon first glance, Iran’s recent death records have the normal trend with deaths in winter being higher than fall and deaths in colder seasons being higher than warmer seasons. However, the excess mortality rates of 7.5% and 5% made last fall and winter the deadliest seasons for Iranians in the past five years. While an increase in winter mortality was expected, one might wonder why the fall deviation was so high given that the COVID-19 virus is believed to have reached Iran in winter.
What caused the whooping rise in excess mortality during fall 2019?
Many Iranian provinces had unusually high levels of excess mortality in fall 2019. This season registered 9,488 more deaths than fall 2018.
Could this be a sign of early emergence of COVID-19?
Probably not. If the outbreak had started back in fall, the excess mortality rate should have only gone up during winter as new cases were getting infected and no social distancing measures were in place at the time. An undetected outbreak of COVID-19 in fall must have resulted in elevated levels of mortality in the same provinces both in fall and winter. But we see a good number of provinces with high excess mortality in fall and no or negative excess mortality in winter. This rejects the idea that Iran was suffering from a national-scale outbreak of COVID-19 in fall 2019, which complies with the reported observations.
So something other than the COVID-19 must have driven up the deaths during fall. The actual number of Iranians who were killed in the brutal suppression of the November protests has not been officially disclosed. But this bloody incident cannot be the single driver of the fall excess mortality in Iran.
In search of possible causes, we can compare the WHO annual influenza situation reports for the Eastern Mediterranean Region during fall 2019 to previous years. The data suggests a 20% increase in the number of positive flu cases since the start of fall (with the co-circulation of both H1N1 and flu type B in Iran).
There were also reports from doctors inside Iran who detected an unusually high rate of infection and death due to flu-like illnesses which may, at least partially, explain the 7.5% national excess mortality during fall 2019. China and the US also reported a surge in flu-like infections during the same period but only a small fraction of individuals in these countries showed traces of SARS-CoV-2, the virus responsible for the COVID-19.
In the absence of sufficient data and evidence, what caused the drastic rise in deaths in Iran during fall 2019 remains a mystery. But the excess mortality might have been due to an unusually high peak of seasonal flu (such as H1N1 or any other strain). This could have interfered with the proper diagnosis of COVID-19 patients in winter, who showed similar respiratory disease symptoms, delaying appropriate measures to contain the outbreak across the country.
About the authors
Mahan Ghafari (Twitter: @Mahan_Ghafari) is an expert on viral evolution, DPhil student, and an Oxford-Radcliffe scholar at the University of Oxford. He has previously worked on genetic variations during person-to-person transmission of influenza A viruses, and complex adaptations at the University of Cambridge and Emory University.
Kaveh Madani (Twitter: @KavehMadani) is a Henry Hart Rice Senior Fellow at the Yale MacMillan Center for International and Area Studies and a Visiting Professor at the Centre for Environmental Policy at Imperial College London.