Aging and Comorbidities in Acute Pancreatitis I: A Meta-Analysis and Systematic Review Based on 194,702 Patients

Abstract

Background: Acute pancreatitis (AP) is one of the most common cause of hospitalization among gastrointestinal diseases worldwide. Although most of the cases are mild, approximately 10-20% of patients develop a severe course of disease with higher mortality rate. Scoring systems consider age as a risk factor of mortality and severity (BISAP; >60 years, JPN>70 years, RANSON; >55 years, APACHE II >45 years). If there is a correlation between aging and the clinical features of AP, how does age influence mortality and severity? Aim: This study aimed to systematically review the effects of aging on AP. Methods: A comprehensive systematic literature search was conducted in the Embase, Cochrane, and Pubmed databases. A meta-analysis was performed using the preferred reporting items for systematic review and meta-analysis statement (PRISMA). A total of 1,100 articles were found. After removing duplicates and articles containing insufficient or irrelevant data, 33 publications involving 194,702 AP patients were analyzed. Seven age categories were determined and several mathematical models, including conventional mathematical methods (linear regression), meta-analyses (random effect model and heterogeneity tests), meta-regression, funnel plot and Egger's test for publication bias were performed. Quality assessment was conducted using the modified Newcastle-Ottawa scale. The meta-analysis was registered in the PROSPERO database (CRD42017079253). Results: Aging greatly influences the outcome of AP. There was a low severe AP incidence in patients under 30 (1.6%); however, the incidence of severe AP showed a continuous, linear increase between 20 and 70 (0.193%/year) of up to 9.6%. The mortality rate was 0.9% in patients under 20 and demonstrated a continuous linear elevation until 59, however from this age the mortality rate started elevating with 9 times higher rate until the age of 70. The mortality rate between 20 and 59 grew 0.086%/year and 0.765%/year between 59 and 70. Overall, patients above 70 had a 19 times higher mortality rate than patients under 20. The mortality rate rising with age was confirmed by meta-regression (coefficient: 0.037 CI: 0.006-0.068, p = 0.022; adjusted r²: 13.8%), and severity also (coefficient: 0.035 CI: 0.019-0.052, p < 0.001; adjusted r²: 31.6%). Conclusion: Our analysis shows a likelihood of severe pancreatitis, as well as, pancreatitis-associated mortality is more common with advanced age. Importantly, the rapid elevation of mortality above the age of 59 suggests the involvement of additional deteriorating factors such as co-morbidity in elderly.

Keywords: acute pancreatitis; aging; co-morbidity; mortality; severity.

Figure 1

Life expectancy at birth. There is a steadily rising average life expectancy at birth. It has dramatically risen by 16 years (from 55.4 to 71.4 y) in the last half century. Data sources: between 1960 and 1999, World Bank; between 2000–2015, WHO.

Figure 2

PRISMA flow diagram. The diagram for the study selection for this meta-analysis is based on the PRISMA-recommended flow chart (Moher et al., 2009).

Figure 3

Meta-regression of severity. The figure shows 29 data from 23 reports where x = age (mean), y = logit event rate: ln[p/(1-p)], and circle diameters show the weight of each study based on the random effect model. The meta-regression shows a significant (p < 0.001) relationship between age and severity (r² = 31.6), therefore the risk for developing severe cases is elevated by aging.

Figure 4

Conventional regression of severity. The conventional regression, which is independent of distortion from diverse numbers of patients, shows a linear rise (0.193%/year) in severity from young to old age.

Figure 5

Forest plot of studies evaluating severity in acute pancreatitis in age groups. Full diamonds show the weighted event rates for studies, respectively, line represents the 95% confidence interval (CI), and empty diamonds show the pooled results of severe cases with a steadily rising frequency from young to older age. Wideness of the empty diamond represents the confidence limits. Under 40 there is a slight elevation concerning severe cases, from 40 to 60 severity rates differs in the studies, then A60 remains stable.

Figure 6

Forest plot results for cut-off values for severity. Summary table of pooled effect with CI and significance levels to detect cut off value. Concerning mortality all comparisons were significant, however examining severity only three. Explanation might be that in young ages there is a low event rate, in middle age groups there is a higher proportion therefore the difference is equalized leading to a non-significant difference. The same occur in the aged vs. middle aged groups.

Figure 7

Conventional regression of mortality. The conventional regression shows a linear elevation until 59, however from this age the mortality rate started elevating with 9 times higher rate until the age of 70.

Figure 8

Forest plot of studies evaluating mortality in acute pancreatitis. Full diamonds show the weighted event rates for studies, respectively, line represents the 95% confidence interval (CI), and empty diamonds show the pooled results of mortality with a steadily rising frequency from young to older age. Wideness of the empty diamond represents the confidence limits. The diamonds show a steadily rising frequency in mortality from youth to old age.

Figure 9

Meta-regression of mortality. The figure shows 43 data from 30 reports where x = age (mean), y = logit event rate: ln[p/(1-p)], and circle diameters show the random size of each study. The meta-regression shows a significant relationship (p = 0.022) between age and mortality.

Figure 10

Forest plot results for cut-off values for mortality. Forest plot results from studies evaluating the cut-off values for mortality in acute pancreatitis with significant results in each of four groups. All comparisons showed a significant difference.

Figure 11

Factors that may prepossess mortality and severity in AP. Our data show that age linearly correlates to higher risk of developing severe AP. Concerning mortality other factors may elevate the risk of decease cases above 59 years of age.