Vaccines | Free Full-Text | Mortality of Invasive Pneumococcal Disease following Introduction of the 13-Valent Pneumococcal Conjugate Vaccine in Greenland
[ad_1]
1. Introduction
The Inuit populations of the Arctic regions of Alaska, Canada, and Greenland have historically suffered from high incidence rates of invasive pneumococcal disease (IPD) [1,2,3,4].
Surveillance of invasive pneumococcal disease (IPD) in 1986 showed that the Inuit population of Alaska had the highest reported average overall IPD rate in the world, which was four times higher than the rate of the non-native population [5]. Given the shared genetic, social, and environmental risk factors among Inuit populations in the North American Arctic, including common living conditions, socio-economic challenges, and levels of comorbidity, the International Circumpolar Surveillance System (ICS) for IPD was established. This surveillance system has confirmed high rates of IPD in other Inuit populations [5]. Yet, few studies have evaluated the burden of IPD in the Inuit population of Greenland [3,6,7,8].
The 13-valent pneumococcal conjugate vaccine Prevnar 13® (PCV13; Pfizer, New York, NY, USA), covering 13 Streptococcus pneumoniae serotypes was introduced in the childhood vaccination program (CVP) by 1 September 2010, in Greenland [9]. The CVP is free and is eligible to all children with a permanent address in Greenland [10]. Before the introduction of the PCV13, IPD rates among Inuit in Greenland were up to four times higher than among non-Inuit in the country [6,8].
In our previous study [8], it was shown, that after the introduction of the PCV13, the overall incidence of IPD declined with 30% and IPD caused by vaccine serotype (VT) declined with 55%. Moreover, VT-IPD incidence rates among children ≤1 years declined with 30%, and no VT-IPD cases among children 2–4 years were observed post-PCV13 introduction. Besides this direct effect, the vaccine has also shown to have an indirect (herd) effect.
The indirect vaccine protection of persons older than the targeted group for vaccination has shown to be due to reduced carriage of S. pneumoniae in the vaccinated persons, thereby reducing the transmission of the infection [2]. However, as in many other countries, the introduction of the PCV13 in Greenland has been associated with shifts in nasopharyngeal pneumococcal serotype distribution with increased carriage of non-vaccine serotypes (NVT), increasing NVT-IPD incidence rates among adults ≥60 years [8].
Worldwide, it is estimated that 1.6 million people of all ages die each year from IPD, including about 814,000 deaths among children under the age of five [11]. IPD case-fatality rates (CFR) have been shown to be about four times higher in Inuit populations than in non-Inuit [4,6].
Outcomes of IPD in terms of severity and mortality are influenced by host characteristics such as age, socioeconomic status, ethnicity, alcoholism, immunodeficiency, and comorbidities [3,12,13]. Moreover, studies have shown an association between the outcome of IPD and particular serotypes [12,14,15].
While our previous study demonstrated the effect of PCV13 on serotype distribution in Greenland [8], little is known regarding the effect of pneumococcal conjugate vaccines (PCVs) on IPD-related mortality, particularly in Greenland.
The aim of this study was to describe changes in IPD-related mortality and fatality after the introduction of the PCV13 in 2010 in Greenland, and how this varies by age, sex, region, ethnicity, and pneumococcal serotype.
3. Results
Of the entire cohort of 295 IPD patients, 66 died within 31 days after IPD hospitalization representing a CFR of 22.4%. During the pre-PCV13 period (January 1995–September 2010), 49 out of 206 (CFR 23.8%) IPD patients died within 31 days after IPD hospitalization, while 17 out of 89 (CFR 19.1%) IPD patients died within 31 days after IPD hospitalization during the post-PCV13 period (September 2010–October 2020). This difference was not significant (p = 0.463).
The average length of hospitalization for all IPD cases was 14 days. This includes 15 cases who had a length of hospitalization more than 40 days. The average length of hospitalization was 14 days (range 0 to 121 days) for the pre-PCV13 period and 15 days (range 0 to 138 days) for the post-PCV13 period.
3.1. Mortality Rates
Table 1 shows crude all-cause mortality rates and age-standardized all-cause mortality rates for the entire population in Greenland and crude IPD-related mortality rates and age-standardized IPD-related mortality in Greenland.
Although there was a statistically significant change in crude all-cause mortality in Greenland from the pre- to the post-PCV13 period (9.1 per 1000 PYRS to 9.7 per 1000 PYRS) (p < 0.05), the mortality rates were similar. However, after age-standardization, all-cause mortality rates in Greenland decreased from pre- to the post-PCV13 period (16.4 per 1000 PYRS to 12.8 per 1000 PYRS), which was statistically significant (p < 0.05).
For the whole period, crude IPD-related mortality rate was up to 350-times higher than the crude all-cause mortality rate for the population in Greenland. This rate decreased from the pre- to the post-PCV13 period (3589 per 1000 PYRS to 2658 per 1000 PYRS), but after age-standardization, IPD-related mortality rates increased from the pre- to the post-PCV13 period (3347 per 1000 PYRS to 4441 per 1000 PYRS).
Table 2 shows SMRs in patients IPD. Overall, SMR in the pre-PCV13 period was significantly higher than in the post-PCV13 period (267.6 vs. 114.2) (p = 0.003). SMR in patients with IPD was in general higher during the pre-PCV13 period compared to the post-PCV13 period when IPD-related mortality cases were stratified by sex, S. pneumoniae serotypes, age group, region of Greenland, ethnicity, and level of Charlson Comorbidity Score. There was no IPD-related mortality in patients up to 39 years of age during the post-PCV13 period.
IPD patients had a non-significant higher 31-day survival probability during the post-PCV13 period, compared to the pre-PCV13 period (p = 0.31) (Figure 1a). The difference between the pre-PCV13 period and the post-PCV13 period for 31-day IPD survival probability was greater when adjusting for sex, age, and CCI score (p = 0.08) (Figure 1b).
Figure 2 shows unadjusted IPD-related mortality odds ratios (OR) for the pre- and the post-PCV13 period. Except for IPD patients from the east and the west region of Greenland during the pre-PCV13 period, IPD patients outside of the capital Nuuk during both the pre- and the post-PCV13 period were associated with higher 30-day IPD-related mortality compared to IPD patients in Nuuk. However, only the association of 31-day IPD-related mortality between the west and east regions of Greenland during the post-PCV13 period and Nuuk was statistically significant.
IPD patients with meningitis, compared to IPD patients with non-meningitis, had a higher 31-day IPD-related mortality both during the pre- and the post-PCV13 period, but this was only statistically significant for the post-PCV13 period.
Patients with VT-IPD had a lower 31-day IPD-related mortality during both periods compared to patients with NVT-IPD.
There were no IPD-related deaths in non-Inuits during the post-PCV13 period. Inuits had almost the same 31-day IPD-related mortality during the pre-PCV13 period as non-Inuits. Females had a lower 31-day IPD-related mortality compared to males during both the pre- and the post-PCV13 periods.
3.2. Mortality by Serotypes
Of the total 295 IPD patients in the entire study period, only 185 IPD patients had their isolates serotyped. Of these 185 isolates, 115 (55.8%) were serotyped during the pre-PCV13 period and 70 (78.6%) were serotyped during the post-PCV13 period.
Figure 3 shows the proportion of deaths by serotype. Of the 185 serotyped isolates, VT serotypes 3, 4, 6A, 7F, 9V, 14, and 18C all caused IPD-related deaths during the pre-PCV13 period. Only VT serotype 4 and 19F caused IPD-related deaths during the post-PCV13 period (Figure 3: left panel). During the pre-PCV13 period, the NVT serotypes that caused IPD-related deaths, were serotypes 8, 12F, 15B, and 22F. During the post-PCV13 period, NVT serotypes 9N, 10A, 10B, 16F, 20, and 22F caused IPD-related deaths (Figure 3: right panel).
4. Discussion
In this register-based study of all IPD cases in Greenland 1995–2020, we found that CFR for IPD decreased with 24.5% from the pre- (1995–2010) to the post-PCV13 (2010–2020) period. The SMR decreased significantly in IPD patients from the pre- to the post-PCV13 period (0.43, 95% CI 0.25–0.75), regardless of sex, ethnicity, comorbidity, and age. More so, there was no IPD-related death in patients up to 39 years of age in the post-PCV13 period.
For IPD patients in the post-PCV13 period, 31-day survival probability was higher, but not significant, compared to IPD patients in the pre-PCV13 period, also after adjusting for sex, age, and CCI score. In addition, there was no difference in length of hospital stay between the two periods.
During the years of 1999–2020, life expectancy in Greenland increased from 63 years to 69 years for men and from 69 years to 73 years for women [21]. This may explain the increase in crude all-cause mortality rate in Greenland from the pre- to the post-PCV13 period. However, despite the increasing life expectancy in Greenland, it is still below the estimated global average life expectancy of around 73 years [22]. As the population in Greenland changes towards becoming older [16], we used age-standardization adjustment to see how the rate would change if the age structure of the Greenlandic population was the same in both periods and as that of the estimated world population. After age-standardization, we observed a decrease in crude all-cause mortality in the post-PCV13 period in Greenland.
IPD is a disease that carries a significant mortality in Greenland, which is reflected by the 350-times higher risk of IPD-related mortality when compared with all-cause mortality. However, crude IPD-related mortality rates decreased from the pre- to the post-PCV13 period with 26% (p = 0.291). On that note, the reason for the 33% (p = 0.232) increase in IPD-related mortality after age-standardization from the pre- to the post-PCV13 period is because the elderly IPD cases in the post-PCV13 period are very few but have a high mortality rate, affecting the world-weighted deaths.
It is well known that age influences the outcome of IPD. In this study, SMR in patients with IPD decreased in all age groups from the pre- to the post-PCV13 period. In fact, no children with IPD under the age of 5 years or persons in the age group 5–39 died within 31 days after IPD diagnosis in the post-PCV13 period. These observations are in line with studies from other countries, where PCV13 introduction has been associated with a reduction in IPD-related mortality rates in children aged 23,24,25,26] in young adults, and in adults [24,27].
Previous studies from Greenland and other countries have shown that mortality from IPD is greater in the elderly ≥65 years than in any other age group [3,6,12,28,29,30]. In our study, SMR was lowest in patients with IPD older than 60 years during the entire study period with almost similar SMRs for this age group in the two periods. Similar results were observed in a study from Spain, where no significant difference in IPD-related mortality rates in patients ≥65 years was observed [27]. This is, however, not the general case, as some studies have observed a reduction in IPD-related mortality rates and CFRs in the elderly [24,25], although the decrease in IPD-related mortality rate in persons aged ≥ 65 years observed in one study was minimal [25].
CFRs of both VT- and NVT-IPD decreased among adults aged 40–59 years from the pre- to the post-PCV13 period (19.4–16.6% and 30–18.8%, respectively). However, in adults ≥60 years CFRs of VT-IPD remained the same (20% in both periods), and CFR of NVT-IPD in the same age group increased from 12.5% in the pre-PCV13 period to 25.9% (p = 0.753) in the post-PCV13 period. Increases in death from NVT-IPD among the elderly has also been observed in the study from Spain, where a trend towards an increase in NVT-IPD-related mortality rates was observed in patients ≥65 years following PCV13 introduction [27].
The above findings suggest that while PCV13 has a direct effect on mortality of the vaccinated persons, and a herd protection of the adults, there is no effect of PCV13 introduction for children on mortality of the elderly above 60 years of age.
Pneumococcal serotypes differ in invasiveness [31,32]. It has been suggested that serotypes that have a high risk of causing invasive disease infect healthy individuals and therefore behave as ‘primary pathogens’, whereas serotypes with a lower potential to cause invasive disease cause disease in patients with underlying conditions, and in those patients cause more severe disease and mortality thus behaving as ‘opportunistic pathogens’ [31]. In particular, serotypes 3, 6A, and/or 19F are among the serotypes described as less invasive and thus associated with the highest mortality rates in persons with comorbidities [13,30,33,34]. In our study, persons who died from a suggested less invasive serotype had low levels of comorbidity, except for one person who had a moderate level of comorbidity as the highest level. Moreover, persons with “high-invasive serotypes” (4, 7F, 14, and 18C) had overall low levels of comorbidity, and there were more deaths from these serotypes than from the suggested less-invasive serotypes. Thus, our study could not support an association between potential suggested serotype-invasiveness and mortality, which is in line with one other study [12].
Compared with non-Inuits, Inuits had a higher IPD-related SMR. This is in accordance with previous studies of IPD in Greenland [3,6]. Deaths among non-Inuits were only observed in the pre-PCV13 period, but as the number of IPD-related deaths among this group was very small (n
We found a decrease in SMR in patients with IPD from the pre-PCV13 to the post-PCV13 period in all levels of comorbidity defined by CCI-score level groups. The decrease in SMR was less pronounced in IPD patients with high CCI-scores confirming that underlying disease are effect modifiers of IPD-related mortality as observed in other studies [12,13,23,30]. Nevertheless, our observations support that PCV13 has protective effects against IPD-related mortality in all levels of comorbidity.
There were clear differences in SMR overall, and over time, between the regions of Greenland; North and South regions of Greenland had the highest SMR of IPD in the pre-PCV13 period which decreased in the post-PCV13 period, while IPD SMR in the East and West regions were lowest during the pre-PCV13 period but increased in the post-PCV13 period. The highest decrease in IPD SMR from the pre- to the post-PCV13 period was, however, observed in Nuuk (0.17, 95% CI: 0.06–0.44). Reasons for these opposite results between regions are unclear, but as diagnostic and treatment possibilities differ markedly between the regions (mainly between Nuuk and coastal hospitals) we believe that such differences over time and place may explain the observations.
IPD patients with meningitis compared to IPD patients with non-meningitis were associated with the highest mortality rates in both periods. Studies from other countries have shown decreases in fatal cases of non-meningitis IPD presentations, such as bacteremia and/or septicemia, after PCV introduction [23,25], yet most of the IPD-related deaths remained due to meningitis [23].
Thus, our study shows that IPD-related mortality has decreased in Greenland since the introduction of the PCV13 in 2010 in terms of numbers who die from the disease. Decreases in IPD-related mortality after PCV13 implementation have generally been observed in studies from other countries [24,25,27]. Although we did not have access to information of vaccine status in IPD cases, we believe from our previous [8] and current figures that it is likely that the reduction in IPD-related mortality following vaccine introduction is caused both by a reduction in IPD incidence and a lower risk of death.
Obviously, a temporal association between PCV13 introduction and reduced mortality may not incur causality. Other factors may also account for the reduced mortality, such as better access to health facilities and to laboratory diagnostic testing, and more timely hospital treatment. The finding of a markedly reduced mortality from IPD in the capital Nuuk, where the central hospital in Greenland is located, compared to other regions, could suggest that improved medical treatment could play a role in the reduced mortality following PCV13 introduction.
The strengths of this study are that our population was well-defined, and the study period was long, enabling us to identify long-term association between PCV13 introduction and mortality. We were able to obtain exact demographic information of the study population as we could uniquely identify case patients on a personal identifiable level. Moreover, we were able to obtain a full background population size.
There are a number of limitations in this study. First, the number of observed deaths is small and therefore estimates of, e.g., rates may be interpreted with care. Second, a temporal association between PCV13 introduction and reduction in mortality may not reflect causality as discussed above. Third, IPD-related mortality was defined as deaths within 31 days after hospital admission for IPD, irrespectively of the actual cause of death as we did not have access to information of causes of death for the majority of patients. Fourth, by censoring 31 days after IPD hospitalization we exclude potential cases that might have died from IPD beyond 31 days after hospitalization. Finally, information on serotypes was incomplete for the majority of our study population, and the sample sizes for each specific serotype was small.
[ad_2]
