Pathways to injury in chronic pancreatitis: decoding the role of the high-risk SPINK1 N34S haplotype using meta-analysis

Abstract

Background: The complex interactions between recurrent trypsin-mediated pancreatic injury, alcohol-associated pancreatic injury and SPINK1 polymorphisms in chronic pancreatitis (CP) are undefined. We hypothesize that CP occurs as a result of multiple pathological mechanisms (pathways) that are initiated by different metabolic or environmental factors (etiologies) and may be influenced differentially by downstream genetic risk factors. We tested this hypothesis by evaluating the differences in effect size of the high risk SPINK1 N34S haplotype on CP from multiple etiologies after combining clinical reports of SPINK1 N34S frequency using meta-analysis.

Methods and findings: The Pubmed and the Embase databases were reviewed. We studied 24 reports of SPINK1 N34S in CP (2,421 cases, 4,857 controls) using reported etiological factors as surrogates for pathways and multiple meta-analyses to determine the differential effects of SPINK1 N34S between alcoholic and non-alcoholic etiologies. Using estimates of between-study heterogeneity, we sub-classified our 24 studies into four specific clusters. We found that SPINK1 N34S is strongly associated with CP overall (OR 11.00; 95% CI: 7.59-15.93), but the effect of SPINK1 N34S in alcoholic CP (OR 4.98, 95% CI: 3.16-7.85) was significantly smaller than in idiopathic CP (OR 14.97, 95% C.I. = 9.09-24.67) or tropical CP (OR 19.15, 95% C.I. = 8.83-41.56). Studies analyzing familial CP showed very high heterogeneity suggestive of a complex etiology with an I(2) = 80.95%.

Conclusion: The small effect of SPINK1 N34S in alcoholic subjects suggests that CP is driven through a different pathway that is largely trypsin-independent. The results also suggest that large effect sizes of SPINK1 N34S in small candidate gene studies in CP may be related to a mixture of multiple etiologic pathways leading to the same clinical endpoint.

Figure 1. Hypothesis of etiology-defined pathways to pancreatic fibrosis.

Hypothetical influence diagram illustrating pathologic pathways linking proximal factor (Factor A and B) to PSC (pancreatic stellate cell) and fibrosis through multiple steps (e.g. a₁, a₂, a₃). Etiological factors of type B activate trypsinogen to trypsin, and therefore their pathologic pathway to the PSC can be interrupted by SPINK1. Etiological factors of type A are independent of trypsin, and therefore will not be influenced by variations in SPINK1 expression or function.

Figure 2. Flow diagram of the studies included in the meta-analysis.

* The report by Bernardino et al was excluded from the meta-analysis because it did not detect the N34S haplotype in neither the cases nor the controls and was therefore assigned a weight of zero.

Figure 3. Meta-analysis results for chronic pancreatitis all etiologies combined based on allelic frequency.

Heterogeneity testing: Q-value = 41.05, df = 23, p = 0.01, I² = 43.97 (95% CI: 10.56–64.90).

Figure 4. Meta-analysis results for alcoholic chronic pancreatitis based on allelic frequency.

Heterogeneity testing: Q-value = 7.36, df = 8, p = 0.5, I² = 0 (95% CI: 0.00–62.01).

Figure 5. Meta-analysis results for tropical pancreatitis based on allelic frequency.

Heterogeneity testing: Q-value = 5.72, df = 3, p = 0.13, I² = 47.56 (95% CI: 20.96–78.99).

Figure 6. Meta-analysis results for idiopathic chronic pancreatitis based on allelic frequency.

Heterogeneity testing: Q-value = 20.86, df = 13, p = 0.08, I² = 37.98 (95% CI: 0.00–65.00).