Objective: To determine which neurologic signs or symptoms are predictive of new focal lesions on head CT in HIV-infected patients.
Methods: Prospective study with convenience sample enrollment of HIV-infected patients who presented to a large inner-city university-based ED over an 11-month period. Patients were assessed using a standardized neurologic evaluation to ascertain whether they had developed new or changed neurologic signs or symptoms. Patients with any new or changed neurologic findings had a head CT scan in the ED. The association between individual complaints or findings and new focal lesions on head CT was assessed by univariate analysis, and sensitivity, specificity, and positive predictive values were calculated. Stepwise logistic regression analysis was then carried out to estimate the relative risk for those variables independently associated with new focal lesions on CT scans. A decision guideline was developed incorporating those variables.
Results: One hundred ten patients were identified as having new or changed neurologic signs or symptoms and had a head CT done in the ED. Twenty-seven patients (24%) had focal lesions on head CT, of which 19 (18%) were identified as new focal lesions; eight of these (7%) demonstrated a mass effect. Clinical findings most strongly associated with new focal findings on head CT were: 1) new seizure, relative risk (RR) = 73.5, 95% CI = 6.2 to 873.0; 2) depressed or altered orientation, RR = 39.1, 95% CI = 4.6 to 330.0; and 3) headache, different in quality, RR = 27.0, 95% CI = 3.2 to 230.1. Use of these three findings as a screen for ordering head CT in the ED would have identified 95% (18/19) of the patients with new focal intracranial lesions, and resulted in a 53% reduction in the number of head CTs ordered in the ED. Inclusion of one additional parameter (prolonged headache, > or =3 days), would have resulted in identification of 100% of all new focal lesions, with a 37% reduction in the number of head CTs ordered. Among those patients with new focal findings, 74% required emergent management (i.e., seizure control, IV antibiotics, IV steroids or surgery). The most common intracranial lesion among patients with CD4 counts less than 200 cells/microL was toxoplasmosis, while cerebrovascular accidents (ischemic or hemorrhagic) were most common in those with CD4 counts greater than 200 cells/microL.
Conclusion: Specific clinical signs and symptoms were associated with the presence of new intracranial lesions in a group of HIV-infected patients who presented to the ED with neurologic complaints. These clinical findings can be incorporated into guidelines for determining the need for emergent head CT. Validation and widespread application of these guidelines could result in limiting the use of emergent neuroimaging to a more well-defined HIV-infected patient population.