Abstract
Objective. Pancreatitis is a rare but potentially life-threatening complication of systemic lupus erythematosus (SLE). Vasculitis of the gastrointestinal tract is the most commonly proposed mechanism. We determined the frequency of SLE-related pancreatitis in the Hopkins Lupus Cohort.
Methods. A large prospective cohort of 1811 patients with SLE was reviewed and clinical and laboratory measures of SLE patients who developed pancreatitis were compared to patients who did not develop pancreatitis.
Results. Four percent of patients with SLE had pancreatitis due to SLE. The best multivariate model of clinical and laboratory associations included hypertriglyceridemia, psychosis, pleurisy, gastritis, and anemia.
Conclusion. Hypertriglyceridemia appears to be a strong associate of pancreatitis in SLE, but antiphospholipid antibodies are not. SLE patients with psychosis and pleurisy are at increased risk for pancreatitis.
Pancreatitis in adult-onset systemic lupus erythematosus (SLE) is a well recognized yet rare complication1–10. It was first reported by Reifenstein, et al in 1939 as a symptom complex of unknown etiology and fatal outcome11. It has also been reported in pediatric lupus, where it follows a more aggressive course and is often noted to be fatal at presentation12. The diagnosis of SLE-induced pancreatitis can only be made after excluding all other mechanical and toxic-metabolic etiologies of pancreatitis (including but not limited to cholelithiasis, alcohol, medications, hypertriglyceridemia, hypercalcemia, and infections/sepsis).
The etiopathogenesis of pancreatitis in lupus is not clear10, with several suggested pathogenic mechanisms. SLE-induced vasculopathy of the gastrointestinal tract is one of the most commonly proposed mechanisms13, with vasculitis leading to ischemic necrosis of the pancreas4,12. Immunosuppressive drugs such as azathioprine and cyclosporine14,15, as well as corticosteroids6,9, have also been implicated in several case reports. In a case series of 4 patients, an association with high levels of anticardiolipin antibodies (aCL) was shown, attributing pancreatitis to be a manifestation of antiphospholipid syndrome16. Anticardiolipin antibodies have also been found with chronic calcifying pancreatitis in a patient with SLE17.
We determined the frequency of SLE-related pancreatitis in the Hopkins Lupus Cohort. We then compared the clinical and laboratory measures in patients with SLE who had pancreatitis versus those who did not.
MATERIALS AND METHODS
The Hopkins Lupus Cohort
The Hopkins Lupus Cohort Study has been approved by the Johns Hopkins University School of Medicine Institutional Review Board. All patients gave written informed consent. The Hopkins Lupus Cohort was begun in 1987 and consisted of 1811 patients with SLE at the time of this study. Patients diagnosed with SLE by one faculty member were invited to enroll. After cohort entry, routine visits were scheduled quarterly, or more often if warranted by disease activity or complications. At each visit, clinical assessment of disease activity was ascertained using the physician’s global assessment (PGA) and Safety of Estrogens in Lupus Erythematosus National Assessment-Systemic Lupus Erythematosus Disease Activity Index (SELENA SLEDAI)18,19. Laboratory tests included a complete blood count, erythrocyte sedimentation rate, serum cholesterol, serum creatinine, urinalysis (and since 2006, urine protein/creatinine ratio), complement C3 and C4 levels, anti-dsDNA, dilute Russell viper venom time for the lupus anticoagulant, and aCL measures.
Cumulative cohort history and damage index database
A cumulative history was recorded for each patient at cohort entry including demographic variables, clinical and laboratory manifestations of SLE, and treatment (prednisone, hydroxychloroquine, and immunosuppressive drugs) and then updated at least every 3 months at followup visits. The Systemic Lupus International Collaborating Clinics/American College of Rheumatology (SLICC/ACR) Damage Index20 was also similarly recorded at the first visit and updated at least every 3 months.
Data extraction
The Hopkins Lupus Cohort database was reviewed to identify all patients who developed episodes of acute pancreatitis since the diagnosis of SLE. A patient was classified as having acute pancreatitis in the presence of clinical features of abdominal pain or tenderness with at least a 3-fold elevation of pancreatic enzymes (amylase or lipase) and confirmation of the diagnosis by an imaging study [computed tomography (CT) scan/magnetic resonance imaging scan or ultrasound of the abdomen]. Patients with isolated elevation of enzymes without evidence of pancreatitis on imaging studies were excluded from the analysis.
The number of episodes of pancreatitis, the laboratory evaluation, therapeutic management, and outcome were determined through outpatient or inpatient records of the hospital stay. For patients whose records were incomplete, the patient was contacted by telephone to obtain relevant information. Patients whose records were not available and who were not able to recall events from their hospital admission for episodes of pancreatitis were excluded from the study.
Using the cohort database, clinical and laboratory measures of SLE patients who developed pancreatitis were compared to SLE patients who did not develop pancreatitis.
Statistical analysis
Statistical analysis was done using the JMP software system (SAS, Cary, NC, USA). The chi-square test or Fisher’s exact test was used for dichotomous variables and the Student t test for continuous variables. For demographic, physical, and laboratory variables that were different between patients with and without pancreatitis, multiple logistic regression analysis was used to identify the variables that independently predicted pancreatitis. A 2-tailed p value < 0.05 was considered statistically significant in all comparisons.
RESULTS
Out of a total of 1811 patients with SLE, 76 (4.2%) developed one or more episodes of pancreatitis. One patient was excluded because of the presence of isolated hyperamylasemia without evidence of pancreatitis on CT of the abdomen. An additional 4 patients were excluded because of nonavailability of records (Figure 1).
Selection of participants.
The 71 remaining SLE patients developed 152 episodes of pancreatitis, with a mean of 2.1 episodes/person and an average frequency of 0.10 episodes/person/year. The etiology of pancreatitis was found to be idiopathic, and hence related to SLE, in 63 patients (3.5% of the cohort), while it was attributed to other causes in 7 patients (2 alcohol abuse, 1 post-traumatic, 1 sepsis, 1 cholelithiasis, 1 related to Depakote toxicity, and 1 due to abnormal anatomy consistent with annular pancreas and abnormal pancreaticobiliary junction). The complete evaluation to exclude other etiologies was not available in one patient.
Of all 63 patients with pancreatitis due to SLE, only 1 (1.6%) patient had hypercalcemia (the level was only mildly elevated and was not presumed to be the cause for pancreatitis). Elevated triglyceride levels were noted in 12 (19%) patients. None had triglyceride levels > 200 mg/dl, nor were they on therapy for hypertriglyceridemia. Eight patients had a history of alcohol abuse, but only 2 were consuming any alcohol at the time of their pancreatitis. Eight patients had a history of cholecystectomy, 6 for cholelithiasis and 2 for abdominal pain of unknown etiology. Twenty-nine (46%) patients were positive for antiphospholipid antibodies, of which 21 (72%) had aCL, 12 (41%) had lupus anticoagulant, and 3 (10%) had anti-ß2-glycoprotein I antibody. Eleven patients (17%) were taking an immunosuppressive drug during their episode of pancreatitis, including 5 on azathioprine, 2 cyclophosphamide, and 1 each taking methotrexate, mycophenolate mofetil, chlorambucil and cyclosporine.
Forty-six (73%) patients showed complete resolution of pancreatitis whereas the others developed complications, recurrences, or chronicity. Nine patients developed various complications of pancreatitis including 6 with pancreatic pseudocysts, 4 with secondary diabetes mellitus, and 2 with exocrine pancreatic insufficiency. Twenty-seven (43%) patients developed recurrent pancreatitis (more than one episode) and 9 (14%) developed chronic pancreatitis. The pancreatitis was severe enough to lead to mortality in 2 patients (3%).
Table 1 shows the demographic comparison of the 63 SLE patients with pancreatitis due to SLE versus 1740 SLE patients without pancreatitis. Pancreatitis attributed to SLE was associated with lower income, less private insurance, more disability, and smoking.
Demographic characteristics of SLE patients with and without SLE-attributed pancreatitis.
Table 2 summarizes the clinical manifestations and conditions associated with pancreatitis attributed to SLE. These included fever, vasculitis, pleurisy, psychosis, organic brain syndrome, cognitive impairment, anemia, Sjögren’s syndrome, anti-La, hypertension, and diabetes mellitus. Table 3 summarizes the comparison of laboratory measures. Hypertriglyceridemia was strongly associated with pancreatitis, but antiphospholipid antibodies were not.
Comparison of clinical manifestations in SLE-attributable pancreatitis.
Comparison of laboratory measures in SLE patients with and without SLE-attributable pancreatitis.
Table 4 shows that SLE patients with pancreatitis attributed to SLE have accrued more permanent organ damage, including pancreatic insufficiency.
Prevalence of organ damage in SLE patients with and without SLE-attributable pancreatitis.
Table 5 shows the multivariate model of the most significant variables. This model included hypertriglyceridemia, psychosis, pleurisy, gastritis, and anemia.
Multivariate regression model of variables associated with SLE-attributable pancreatitis.
DISCUSSION
In this large series of patients with SLE we found that SLE-attributable pancreatitis was rare: 3.5% (63/1810). Chronic pancreatitis occurred in 14% and death in 3% of these patients. The mortality rate was lower in comparison to other studies21, but this may be as a result of close monitoring and followup of these SLE patients enrolled in a prospective cohort study, increasing the likelihood of earlier diagnosis and treatment. Appropriate therapeutic strategy, particularly treatment with corticosteroids, which was done in a relatively large proportion of our patients with pancreatitis, may have also added a survival benefit.
Patients with SLE-attributable pancreatitis had a significantly higher Damage Index score (3.54 vs 1.98; p ≤ 0.001), which represents an overall higher disease burden and disease severity in this subset of SLE patients.
This series sheds some light on potential mechanisms. First, there was no association with any antiphospholipid antibody. However, events associated with hypercoagulability (deep venous thrombosis and stroke) were significantly more common with SLE-attributable pancreatitis. In univariate analyses, cutaneous vasculitis had occurred significantly more often in SLE-attributable pancreatitis (27% vs 14%; p = 0.015). Vasculitis has been a proposed mechanism for SLE pancreatitis4,12,13. However, cutaneous vasculitis did not remain independently associated with SLE-attributable pancreatitis in our multiple regression model.
Our series found that secondary Sjögren’s syndrome (p = 0.0082) and anti-La (p = 0.0175) were associated with pancreatitis. Pancreatitis is a known complication of primary Sjögren’s syndrome22. These variables, however, did not retain their significance in the multiple regression model.
The variable with the strongest association was triglyceridemia, which was very significant in both univariate (p < 0.0001) and multiple regression models (p < 0.0001). The triglyceridemia was modest in comparison to the hypertriglyceridemia known to cause pancreatitis23.
Some other SLE manifestations were independent associates of SLE-attributable pancreatitis, namely psychosis (p = 0.0032) and pleurisy (p = 0.0004). Two nonspecific conditions, gastritis and anemia, also remained in the multiple regression model. Physicians treating patients with SLE with a history of these conditions who develop abdominal pain should have a high index of suspicion of pancreatitis.
Our series showed that SLE-attributed pancreatitis, although rare, identified a subset of SLE patients with much higher degree of organ damage and mortality. Antiphospholipid antibodies were not found to be associated, but a history of thrombosis, cutaneous vasculitis, secondary Sjögren’s syndrome, and triglyceridemia were all suggested as potential mechanisms. Only triglyceridemia remained significant in the multiple regression model for a strong association with pancreatitis attributable to SLE.
Footnotes
-
Supported by grant UL1 RR025005 from the National Center for Research Resources. The Hopkins Lupus Cohort is supported by a grant from the National Institutes of Health (NIH AR43727).
- Accepted for publication September 18, 2009.
