USe of a symptom questionnaire to identify malabsorption in human immunodeficiency virus (HIV) infected patients with diarrhea

BACKGROUND: Diarrhea is a common complication of infection with HIV and may be secondary to either small bowel disease with malabsorption or colitis. We evaluated the ability of a questionnaire to identify HIV patients with malabsorption as well as its ability to distinguish specific pathogens. METHODS; 178 AIDS patients sequentially referred for evaluation of gastrointestinal symptoms were administered a questionnaire to determine demographics and the presence of diarrhea! symptoms. Subsequently, patients underwent gastrointestinal evaluation including multiple stool examinations for enteric pathogens, D-xylose test, and small and large intestinal endoscopie biopsy. Univariate analysis was performed using the chi square test and t-test. RESULTS: 143/178 (80%) of the patients had diarrhia, 97/170 (53.9%) had at least one pathogen isolated, and 83/142 (58.4%) had an abnormal D-xylose test. Symptoms predictive of an abnormal Dxylose test included patient self-report of diarrhea (p=.001, sensitivity (sn)=91%, specificity (sp)=31% ), nocturnal diarrhea (p<.001, sn=90%, sp=31%), large volume diarrhea (p<.05, sn=75%, sp=44%), recent decrease in food intake (p<05, sn=50%, sp=70%), and frequent diarrhea (p< .05, sn=78%, sp=38%). Combination of the presence of diarrhea and a recent decrease in food intake resulted in a sn=90% and sp=59%. More severe immunodeficiency (mean CD4 34.8 + 40 vs 163 + 216 Ip<001]) also predicted an abnormal D-xylose test. Patieni age, duration of AIDS, and weight loss were not associated with an abnormal D-xylose test. Presence or absence of a pathogen was predicted only by nocturnal diarrhea (p=.001) and a low CD4 count (p<.001). None of the patient reported symptoms could reliably identify patients with cryptosporidiosis, microsporidiosis or cytomegalovirus. CONCLUSIONS: A symptom questionnaire can identify HIV patients who are malabsorbers with adequate sensitivity (90%), but the specificity of the current questions is too low (59%) to be clinically useful. In addition, symptoms do not predict which etiologic agent will be isolated from patients. Modification of the questionnaire may improve the ability of the questions to identify malabsorbers and may prove useful in directing the evaluation of diarrhea in HIV patients.