Hispanics are more likely to have biliary pancreatitis than Caucasians or African-Americans

Reference: Ma, M.H., et al., Risk factors associated with biliary pancreatitis in children. J Pediatr Gastroenterol Nutr, 2012. 54(5): p. 651-6.

A study out from Yale on the risk factors associated with biliary pancreatitis in children.

Brief Summary: Acute pancreatitis is an inflammatorty disease of the pancreas and has several leading causes in children including biliary pancreatitis.  There are several causes of biliary induced pancreatitis including obstrucction of the bile duct via gallstones and biliary sludge build-up.  The main objective of this investigation was to idetify the clinical features which differentiate biliary pancreatitis vs. non-biliary pancreatitis.  Another objective was to identify the clinical features which differentiate gallstone from sludge induced biliary pancreatitis.  A total of 271 patients were sampled from the Yale New Haven’s Children Hospital with pancreatitis.  Of these 271 patients, 71 had biliary pancreatitis.  A number of clinical features were collected from the patients (ethnicity, weight, calcium range, etc.). 

Results: The following clinical features were signifiantly associated with the presense of biliary pancreatitis vs. non-biliary pancreatitis: Hispanic ethnicity (vs. white or black), higher levels of medium sereum analyse and lipase, and levels of the biomarker aspartate aminotransferase.  Obesity was found to be the major clinical feature associated with the presense of gallstone induced biliary pancreatitis vs. sludge induced biliary pancreatitis. 

Implications for Practice: If a Hispanic child presents symptoms of pancreatitis, the doctor should be highly wary of biliary pancreatitis as the cause vs. other causes.  Furtermore, if the child is obese and has biliary pancreatitis, the doctor should be wary that the cause is gallstone induced biliary pancreatits vs. sludge induced pancreatitis. 

Discussion: The find that ethnicity (particulary being Hispanic) is found to be significantly associated with the presense of biliary pancreatitis was rather surprising and somewhat suspect in my mind.  From a biological perspective, how can this be?  How can Hispanics be all that different biologically from caucasians or african-americans which would lead to a greater probability of them having biliary pancreatitis?   This is really interesting.   I am curious as if other potential confounding variables were left out of the model - particuarly a diet variable which I will get to in just  a bit in the next section.

            The find that obesity also leads to a greater probability of gallstone vs. sludge induced biliary pancreatits is also another interesting find yet not as suprising.  This would make for a good follow-up study as to why this occurred for obese children vs. non-obsese children.  I am always curious to the under-causing biological mechanisms at play here.    

Commentary on Statistics and Study Design:  Overall, a good study design.  These are just some nit-picky things really which I tend to prefer.  The odds ratios were presented in the first table (Table 1) that differentiated the biliary vs. non-biliary cases, but it would also be helpful to see the odds ratios for the second table (Table 2) which differentiated the gallstone vs. sludge induced biliary pancreatitis as well.  In general, you want to be consistent in your presentation of results throughout the whole paper.  Also, for the first table, it would have been helpful to see the mean scores and/or percentages of the various quantitative and qualitative variables. In general, you want your information from the table to be as descriptive (and succinct) as possible. 

           I liked how the authors presented an interpretation for the reader of the odds ratios in the discussion and results section.  Many authors don't dot this, and this is important for those who are not as statistically inclined. Also, the general layout of the tables is really nice.  

           My biggest concern though is with the first part of the analysis which compared the biliary vs. non-bililary results where Hispanic ethnicity came out to be a significant predictor. I am concerned that there may have been a potential confounding or interacting variable left out of the investigation which may have changed the relationship between ethnicity and the main outcome - even from a signifcant to a non-significant one.  I don’t know enough about the given domain to think of any other variables which would cause the ethnicity variable to change in a full multi-variate regression model, so I will just leave this up to the authors.

            A thanks to our buddies at Yale for doing this!