Acute pancreatitis (AP) is a common gastrointestinal disease with an annual incidence of 34 per 100,000 person-years (1, 2). Activation of the pancreatic enzymes within the injured pancreatic acini is the most widely accepted pathophysiological theory. After the onset of pancreatitis at the early stage, uncontrolled oxidative stress and large amounts of released inflammatory cytokines induce systemic inflammatory response syndrome (SIRS) and contribute to progressive organ dysfunction (3). Most patients present with a mild clinical course, but organ dysfunction progresses in approximately 20% of the patients. The overall mortality rate remained as high as 10–20% in these patients (4).

Microcirculatory dysfunction plays an important role in the pathogenesis of AP. Uncontrolled oxidative stress and released inflammatory cytokines lead to endothelial injury and the activation of coagulation (5). Abnormal coagulation is manifested by vascular leakage, scattered intravascular thrombosis, and disseminated intravascular coagulation and is associated with increased complications and mortality in AP (6). In general, the platelet is defined as a pivotal factor in homeostasis and thrombosis. It also has been recognized as a member of the immune system and plays a role in inflammation response. Abnormal platelet features are common clinical manifestations associated with poor outcomes in patients who are critically ill (7, 8). However, the role of platelet count in AP pathophysiology has not been elucidated. Here, a retrospective hospital-based cohort study is conducted in patients with AP to evaluate the association between platelet count and clinical outcomes, especially the risk of intra-abdominal infection.

Intra-abdominal infection is a detrimental factor of mortality in patients with acute pancreatitis (13). In the present study, we analyze the platelet count and changes in platelet count within 3 days after admission with the risk of intra-abdominal infection, surgical intervention, in-hospital mortality, and hospital stay. Logistic regression analysis and a generalized additive model are used and adjusted for several potential risk factors at admission. Our study highlights that, in patients with AP, (1) patients with platelet count on admission less than 125 × 10⁹/L or higher than 209 × 10⁹/L are more likely to suffer intra-abdominal infection, receive surgical intervention, and stay longer in the hospital; and (2) changes in platelet count within 3 days after admission have no significant association with intra-abdominal infection, surgical intervention, and in-hospital mortality.

Thrombocytopenia is recognized as an independent risk factor for prognosis. The estimated incidence in critically ill patients is 20–40% during hospitalization (7, 14). Thus, no study reported the incidence of thrombocytopenia in patients with AP. Among the enrolled 1,363 patients in our study, thrombocytopenia occurs in 178 (13%) patients within 3 days after admission. Patients with thrombocytopenia are more likely to deteriorate into SAP compared to patients with non-thrombocytopenia (31.5% vs. 10.7%). Intra-abdominal infection is one of the leading complications that can cause life-threatening organ dysfunction. Early identification of high-risk factors can save medical resources and potentially improve patient outcomes. In previous studies, procalcitonin is the most concerned predictor of intra-abdominal infection (15, 16). However, the small sample size (n < 200) in these studies limited the reliability. Moreover, procalcitonin is not a routine test in all hospitalized patients with AP. Recently, lactate is proven as a reliable predictor of intra-abdominal infection in Shu’s study (17). Even though the sample size is large enough (n = 503), the study design is not comprehensive because only patients with MSAP are included in this study. Prediction models are also constructed to screen the risk factors of intra-abdominal infection. In two multi-center retrospective studies, the platelet count is selected as a risk factor but not included in the final calculated models (18, 19). The reason may be that the linear models are used to select the risk factors. However, the relationship between platelet count and the risk of intra-abdominal infection is a non-linear pattern. In our study, the risk of intra-abdominal infection decreases as the platelet count levels up to 160 × 10⁹/L and then increases. The non-linear relationship is still significant after the adjustment of potential confounders, including age, gender, body mass index, comorbidities, etiologies, creatinine, C-reactive protein, procalcitonin, lactate, D-dimers, and APACHE II score. A two-piecewise linear regression model shows that the risk of intra-abdominal infection decreases by 0.9% per 1 × 10⁹/L increase in platelet when the platelet count is <160 × 10⁹/L, and increases by 0.7% per 1 × 10⁹/L increase in platelet when the platelet count is >160 × 10⁹/L. Interestingly, even the in-hospital mortality shows a declining trend in each quartile separated according to the platelet count (6.5, 2.1, 1.5, and 1.2%, respectively), the platelet count is not significantly associated with in-hospital mortality in patients with AP after adjustment for the confounders. The reason may be that the event (death) is relatively few in our study and, thus, the model is overfitted. In our study, we further analyze the dynamic changes in platelet count within the first 3 days after admission. No significant results are observed. In summary, the platelet count is a reliable and convenient factor for predicting intra-abdominal infection in patients with AP.

The mechanism of platelet count on intra-abdominal infection is poorly understood, but several explanations could be considered. The inflammatory process results in disturbances in pancreatic microcirculation, which has been recognized as an important factor in the development of acute pancreatitis (20). The ultra-structural changes in acute pancreatitis, including the infiltration of leukocytes, the accumulation of platelets intravascularly and extravascularly, and the formation of microthrombi, could be observed (21). However, the linear relationship between platelet count and some inflammatory indicators (C-reactive protein, white blood cell count, and procalcitonin) (Supplementary Figure 2) in this study could not provide a convincing explanation for the results. Another explanation is that the pancreatic vascular endothelial damage will activate the coagulation cascades that increase the consumption of platelet, which will exacerbate the pancreatic necrosis and eventually pancreas become infected. Interestingly, in this study, with the change in platelet count, the change trends of fibrinogen and D-dimers are consistent with the results (Supplementary Figure 3). It could be speculated that complex derangements of hemostasis occurred during acute pancreatitis. As a whole-body inflammatory disease, the damage of endothelial is not limited to the pancreatic tissues in severe pancreatitis. The multi-organ damage will further increase the activation of the platelet. The intestinal tract is the largest microorganism reservoir of the human body and could be easily damaged during systemic inflammation response. During the period of acute pancreatitis, the integrity of intestinal mucosa could be destroyed, which will also increase vascular permeability. Therefore, bacterial translocation could be easy. The decrease or increase in the platelet number may be the reflex of endothelial damage, which is observed in a previous study on sepsis (22). It is interesting that even though the risk of intra-abdominal infection is comparable between the low-level platelet count group (<125 × 10⁹/L) and high-level platelet count group (>209 × 10⁹/L) (10.3% vs. 11.1%) in our study, mortality is relatively low in high-level platelet count group (6.5% vs. 1.2%). The reason may be that a high platelet count reflects the better hematopoietic function of the body. Moreover, platelet possesses immune properties and can inhibit the spread of infection (23).

Our study is subjected to some limitations. First, as an observational study, it could not conclude that the associations between platelet count and clinical outcomes are causal. Second, the data were collected from one center, and the validation should be verified in other data sources. Third, the death event is relatively few in our study, and thus, the model may be overfitted. The association between platelet count and in-hospital mortality should be checked in a large number of patient studies.

This retrospective cohort study among 1,363 patients with AP reveals that a platelet count of approximately 160 × 109/L on admission is associated with the lowest risk of intra-abdominal infection, surgical intervention, and shortest hospital stay in patients with acute pancreatitis. Moderate is the best.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by Institutional Ethics Board of Ruijin Hospital, Shanghai Jiao Tong University School of Medicine. The Ethics Committee waived the requirement of written informed consent for participation.

JH and EM conceived the study. TN, YW, GM, WY, ZY, BZ, and YC helped in data collection. TN, BZ, HS, and LM helped in manuscript revision. WS, JH, TN, and YW were the co-first authors responsible for drafting and data analysis. EM was the responsible author. All authors contributed to the refinement of the study protocol. All authors contributed to the article and approved the submitted version.