Age Dependence in COVID-19 Mortality in Germany

There is a clear correlation between COVID-19–related deaths and age. In a position paper, the Leopoldina working group advocated a comparison of COVID-19–related mortality figures with general mortality figures in the respective age groups (1). These data are now publicly available.

Methods

Data on the number of COVID-19–related deaths were retrieved from the Robert Koch Institute (RKI) (https://de.statista.com/statistik/daten/studie/1104173/umfrage/todesfaelle-aufgrund-des-coronavirus-in-deutschland-nach-geschlecht/). These data were correlated with the most up-to-date data for Germany on female and male inhabitants per age group as of 31 December 2018 (https://de.statista.com/statistik/daten/studie/1112611/umfrage/weibliche-bevoelkerung-in-deutschland-nach-altersgruppen/), as well as the corresponding number of deaths in 2018 (www.ec.europa.eu/eurostat/en/web/products-datasets/-/DEMO_MAGEC). Age groups were specified by the sources. Confidence intervals were calculated according to Clopper–Pearson.

Results

In 2018, a total of 953 691 deaths were registered in Germany; this corresponds to 2613 deaths per day. However, the death rate was somewhat higher in 2018 than in 2017 and 2019. The first COVID-19–related deaths were reported on 9 March 2020. By 25 May 2020, 8257 COVID-19–related deaths had been reported to the RKI; this corresponds to 0.87% of all deaths in 2018. However, the period during which COVID-19–related deaths were considered covered only 78 days. The highest daily number of newly reported deaths was reached on 16 April with 315 (corresponding to 12.1% of the average daily rate). The mean number of deaths reported per day was 106 (4.1%) and the median number 98 (3.7%), interquartile range 31–173. The following analyses according to age and sex took only 8252 deaths into account due to lacking information.

In relation to the number of inhabitants in the respective age groups, a strong age dependence was seen for COVID-19–related deaths (Table). In addition, a sex dependence clearly emerged in the 40- to 79-year age group.

Table 1. COVID-19 deaths in relation to the number of inhabitants in 2018.

Age in years		Number of inhabitants	COVID-19 deaths		95% Confidence interval
			Number	Per million
<20	F	7 410 000	1	0.13	[0.00; 0.75]
	M	7 880 000	2	0.25	[0.03; 0.92]
M/F ratio: 1.9
20–39	F	9 900 000	8	0.81	[0.35; 1.59]
	M	10 550 000	20	1.90	[1.58; 2.93]
M/F ratio: 2.3
40–59	F	11 880 000	89	7.5	[6.0; 9.2]
	M	12 020 000	256	21.3	[18.8; 24.1]
M/F ratio: 2.8
60–79	F	9 500 000	804	85	[79; 91]
	M	8 490 000	1814	214	[204; 224]
M/F ratio: 2.5
>80	F	3 360 000	2778	827	[796; 858]
	M	2 030 000	2480	1222	[1174; 1271]
M/F ratio: 1.5
Ratio >80/20–39 years, females: 1023
Ratio >80/20–39 years, males:   644

Recording of COVID-19–related deaths up to 25 May 2020

The Figure presents the COVID-19–related deaths in relation to deaths per age group and sex in Germany in 2018. In the 40- to 89-year group, a clear age dependence for both sexes was evident. However, this was by no means as strong as in the relation of COVID-19 deaths to the number of inhabitants. The differences between the sexes were also far less pronounced.

Figure.

Percentage of COVID-19 deaths in relation to the overall death rate in 2018 in the respective age and sex groups. COVID-19 cases were considered up to 25 May 2020. The error bars show the 95% confidence intervals. The percentage death rates in the 80- to 89-year age group were 3.5-fold (females) and 3.6-fold (males) higher than for the respective 30- to 39-year age group.

Discussion

In a simple assignment by number of inhabitants, the Table shows a high mortality rate among the elderly—a feature that does not exist to the same extent, however, if one considers the baseline risk per age and sex group, as in the Figure. The incidence of COVID-19–related deaths runs largely parallel to normal mortality rates in the respective age groups. Bearing in mind that COVID-19 follows an extremely severe course in some patients, the age-dependence in the higher age groups can come as no surprise. It is self-evident that severe disease poses greater risk in terminally ill, multimorbid, or simply very elderly patients. Conversely, the risk of death due to COVID-19 appears to be minimal in the younger age groups.

The data also reveal a certain anomaly in the highest age group (>90). The risk of death due to COVID-19 in this age group is somewhat lower compared to the 80- to 89-year group—both in men and women. One possible explanation for this might be the fact that less testing has been done to date in the highest age groups. Since efforts are now being made in Germany to carry out more testing in old people’s homes and care homes, this anomaly could disappear again.

The increased risk for men could be due to their known higher susceptibility to cardiovascular disease (2). Furthermore, one must bear in mind that a meta-analysis found chronic obstructive pulmonary disease (COPD) to be associated with a five-fold higher risk for severe COVID-19 disease (3).

The percentages shown in the Figure should be interpreted with caution with regard to excess mortality. At first glance, coronavirus deaths appear to be around 1% of deaths in 2018. However, this does not take into account either the shortened observation period (78 compared to 365 days), seasonality, or offsetting effects over time. Other indications emerge from the above-mentioned estimators in relation to the mean number of deaths for 2018, which were approximately 4% per day on average across all cohorts and 12% at the peak.

The question of excess mortality should be answered primarily on the basis of seasonal comparisons, such as those carried out by, for example, the German Federal Statistical Office (4) and Euromomo (5). According to the German Federal Statistical Office, a slight excess mortality was apparent in April.

As yet, the Euromomo analysis does not point to an excess mortality in Germany (limited to Berlin and the federal state of Hesse), or a number of other countries (for example, Austria and Denmark), in clear contrast to other countries such as Italy, Spain, France, and Belgium. Very strong heterogeneity is seen here. Euromomo also provides analyses according to age group; however, since these are only at the European level, their interpretation would appear unhelpful due to precisely this heterogeneity.