It was brought to the attention of the authors of the article listed above that the values for pancreas weights used to generate Fig. 4B and listed in the Network for Pancreatic Organ Donors with Diabetes (nPOD) database were incorrect for some of the donors included in the study. This just affected Fig. 4B (pancreas weight). Although unfortunate, this event does not change the message of the article or modify in any way the validity of the conclusions emanating from this particular figure. The authors have redone the analysis using the new pancreas weight data provided by nPOD as of October 2015 and the only change is in the magnitude of the significance (see figure below).

For clarity, the authors have also included the statistical analysis for both the original and the new data. This additional data can be found in the Supplementary Data online ( and includes:

A) Statistical data corresponding to all the groups, including the number of samples, mean, standard deviation, standard error of the mean, and distribution of the data.

B) Results from statistical analysis (Kruskal-Wallis test).

C) Results from multiple comparisons test (Dunn correction for multiple comparisons).

The authors would like to emphasize their and nPOD’s commitment to high-quality science as both parties have been in constant communication to ensure the accuracy of the data, and they are both fully aligned in this regard.

Lastly, the authors apologize for any inconvenience this might have caused the readers of the journal.

A: Graph as shown in the original article using the data contained in the nPOD database as of June 2014. B: New graph originated from the new data listed in the nPOD database as of October 2015. Note that the only change is in the magnitude of the significance from P < 0.001 to P < 0.01 for type 1 diabetes with short duration (<5 years) and from P < 0.01 to P < 0.001 for type 1 diabetes with longer duration (>5 years). T2D, type 2 diabetes; T1D, type 1 diabetes.

Supplementary data