To evaluate the association between inflammatory status, as measured by C-reactive protein (CRP), during inpatient admission and subsequent inpatient outcome and associated resource use. Probabilistic record linkage was used to match hospital episode data, laboratory reports and mortality statistics in a large urban population of 424,000 people in South Wales, UK. Inpatient mortality, length of stay, emergency readmissions and subsequent 1-year hospital bed day occupancy were assessed as a function of CRP status. Between 2001 and 2005, in total there were 432,272 CRP observations from 98,505 people; 69,593 admissions had at least one CRP measurement, affecting 47,100 individual patients. Across all ICD-10 primary diagnoses, CRP was acutely high (> 10 mg/L) in three-quarters of admissions. Acutely high CRP was associated with an eight-fold increase in risk of hospital mortality (p < 0.001) and a doubling of length of stay (p < 0.001) compared to normal CRP levels, after standardising for age and gender. Across the range of observed maximum CRP values measured during admissions (1 mg/L to > 400 mg/L) the likelihood of emergency readmission within 28 days of discharge increased by 50% (p < 0.001), and the predicted number of subsequent bed days occupied in the year following discharge increased by 30-58% across the range of CRP measurement (p = 0.004). CRP has been found to be clearly associated with hospital resource use. Furthermore, CRP also predicted in-hospital mortality. This may imply that better management of systemic inflammation would result in resource savings in inflammatory diseases such as rheumatoid arthritis.