In this study, the association between serum HMGB1 levels and mortality in canine patients with AP was investigated. The diagnosis of pancreatitis in dogs requires a combination of investigating the clinical symptoms, serum chemistry, and abdominal imaging. However, these diagnostic tools cannot predict if the patient will suffer severely due to pancreatitis, or the risk of the disease when the patient is being diagnosed. Significant differences were observed in the serum levels of HMGB1 at the time of diagnosis of pancreatitis between the normal group and the AP group. Also, statistical significance was shown between the survivors and the non-survivors. These results suggest that high serum HMGB1 levels may be associated with a poor prognosis in patients with AP.
The terms mild, moderate, and severe are used to describe the severity of pancreatitis. They could classify the severity of pancreatitis intuitively and used to determine the survival and prognosis of the patient. In general, the criterion for severe AP is neutrophilic inflammation or necrosis in pancreatic tissues confirmed by histologic examination. However, clinically assessing the severity of the disease through biopsies in every case is difficult. Therefore, patients with multiple organ failures or those needing intensive treatment are considered as severe acute pancreatitis (SAP). When multiple organ failure occurs, pancreatitis is already in the severe state and therapeutic intervention could be delayed, thus, the value of prognostic evaluation is low. Accordingly, there is a need for a method to measure the disease activity in patients with AP, clinically. There are several criteria used in humans and dogs, and they have been summarized in a previous paper [14], but establishment of such criteria based only on the clinical symptoms is needed. In this study, the screening criteria for DAI was established in canine patients and used to analyze the population. The mild DAI group recovered from AP and returned to normal condition, the severe DAI group showed irreversible results, while those with moderate DAI had a 75% survival rate. There was a statistically significant difference in their serum HMGB1 levels at the time of diagnosis. Therefore, it may be helpful to measure serum HMGB1 levels in estimating mortality of AP group whose clinical prognosis is ambiguous.
In this study, the AP group received the necessary symptomatic treatment as needed. The relationship among HMGB1 level, hospitalization cost, and hospitalization duration was evaluated using multiple scatter-plot and regression analysis. The multiple scatter-plot visualized the relationship between the three variables and correlation coefficient values showed that the HMGB1 level, hospitalization cost, and hospitalization period had a positive correlation with each other. Also, the regression analysis was applied to outline a causal view of the increased monetary and temporal effort required to treat patients with AP at high serum HMGB1 levels when serum HMGB1 levels were high (Supplementary Fig. 1). When the S equation was applied, the R squared values were found to be 0.25 and 0.249. These are values for real patients and are considered to be meaningful.
This study included 10 normal dogs and 19 AP patients. Patients with an average age of 11 years and with several concurrent diseases were recruited. About the exclusion criteria and normal group inclusion, some dogs showed serum HMGB1 levels significantly lower than normal dogs, despite a high DAI of AP and their poor prognosis. Their commonality was taking oral steroids against immune-mediated thrombocytopenia (IMT) disease. There has been no clear study that has reported that oral prednisolone administration lowers blood HMGB1 levels in dogs, but E. Mysler [15] showed that prednisolone suppresses HMGB1-induced inflammatory responses in cell-based screening. In this matter, there is a possibility of a correlation between the mechanism of action of HMGB1 and that of prednisolone in vivo.
In AP patients, the severity is typically associated with sepsis or SIRS, which is classified as SAP. When the AP group were screened for SIRS criteria [16], 4 patients (No. 2, 7, 11, 19) met at least two out of four, and one patient (No. 11) showed severe weakness, bradycardia and hypothermia, and was strongly suspected as SIRS. It died despite being hospitalized for 6 days. In particular, papers on AP patients and HMGB1 concentrations published to date have demonstrated an upward trend in SAP patients with SIRS [17]. This result suggests that HMGB1 levels could be used as a prognostic factor in non-SIRS or early SIRS in AP patients, which needs further investigations.
In addition, HMGB1 is also elevated in immune mediated inflammatory diseases including inflammatory bowel disease, rheumatoid arthritis. Although not attached to this paper, the number of patients with immune-mediated diseases other than pancreatitis has been found to have excessively high values. Therefore, HMGB1 is not suitable as a tool for diagnosing pancreatitis, but is rather applicable in the prognostic evaluation in patients who have already been diagnosed with AP.
One study [8] confirmed that HMGB1, a late mediator of inflammation, reached its peak within 48–72 hours in SAP human patients. Patients in this study had their serum collected within 48 hours when diagnosed with AP. However, it may be different from the onset of pancreatitis because the time taken by the caretaker to notice the symptoms of the dog and visit the hospital may vary. In this study, although the trend of HMGB1 level was not confirmed, the clinical symptoms of 2 patients collected serially at 24-hour intervals improved, but there was little change in HMGB1 concentration. However, for CRP [18] or cPL [19], which confirms the presence of an acute inflammatory response, their concentrations could rapidly increase and decrease within 24 hours. They reflect the immediate inflammatory situation, and the tendency to increase or decrease is more important. On the other hand, HMGB1 levels tend to climb slowly [11], indicating that the error due to time of visit does not have a big influence.
Some limitations were encountered in the study. First, most AP cases in the SNU VMTH were already managed for the various underlying diseases. If study participants had only AP, the correlation between serum HMGB1 concentration and prognosis would be more intuitive. Second, the study confirmed the possibility of serum HMGB1 concentration as a prognostic factor of pancreatitis despite the small sample size. The next step should be to conduct a systematic assessment of increasing the evaluation period and the number of cases.
In this study, the serum HMGB1 levels in canine patients diagnosed with AP were significantly higher than normal dogs without physical abnormalities. And statistical significance between patients that died from the disease and those that recovered successfully was also identified. Furthermore, there was a tendency of longer hospital stay and higher cost in AP patients with higher HMGB1 levels. Serum levels at the time of diagnosis could be used to predict the prognosis of patients with AP, especially in patients with moderate DAI.