Association of sC5b-9 plasma levels with laboratory parameters and with clinical conditions in 58 patients with TTP during acute phase. Linear regression analyses (left) showing association with creatinine (A), platelets (B), LDH (C), and Hb (D) are shown. For ln-transformed continuous end points (creatinine, platelets, and LDH), the estimate should be interpreted as percent change per 1 ln(sC5b-9) increase; for Hb the change is in g/dL. Continuous lines in the graphs represent predicted lines from crude linear regression models, shaded areas represent 95% confidence bands. In the table (bottom), each end point reports the number of observation (N), the percent changes per 1 ln(sC5b-9) (crude estimate from univariate analysis and adjusted estimate from multivariable analysis), the 95% CIs, and the P value (P). Multivariable models were adjusted for sex, age (continuous), and type of episode (first TTP event vs relapse). Logistic regression analyses (right) showing ORs and 95% CIs per 1 ln(sC5b-9) increase, for systemic symptoms (fatigue, fever, abdominal pain, headache, jaundice, and vomiting), neurological alterations (stroke, seizures, coma, personality change, focal neurological signs, and transitory ischemic attack), cardiac alterations (acute coronary syndrome and electrocardiographic ischemic abnormalities), renal alterations (an increase in serum creatinine ≥0.3 mg/dL within 48 hours, ≥50% within 7 days, or a urine output of <0.5 mL/kg per hour for >6 hours, per the Kidney Disease Improving Global Outcomes guidelines), and bleeding (hematuria, meno-metrorrhagia, mucosal bleeding, gastrointestinal tract bleeding, ecchymosis, and purpura) are shown. In the table (bottom), each end point reports the number of patients (N), the percent changes per 1 ln(sC5b-9) (crude estimate from univariate analysis and adjusted estimate from multivariable analysis), the 95% CI, and the P value. Multivariable models were adjusted for sex, age (continuous), and type of episode (first TTP event vs relapse).