Folate deficiency has long been postulated to play a role in the etiology of cervical cancer, the third most frequent cancer among women worldwide. In a large, multiethnic community-based case-control study of invasive cervical cancer in five U.S. areas, we assessed accepted and postulated risk factors with an in-home interview and successfully obtained blood samples, at least 6 mo after completion of cancer treatment, from 51 and 68%, respectively, of interviewed cases and controls. Cases with advanced disease (6%) and/or receiving chemotherapy (4%) were excluded, leaving 183 cases and 540 controls. Serum and red blood cell folate were measured with both microbiologic and radiobinding assays. For all four folate measures, risk was moderately, but nonsignificantly, elevated for women in the lowest quartile, compared to the highest [fully adjusted relative risks (RR), including serologic human papillomavirus (HPV)-16 status = 1.2-1.6]. However, for women in the upper three homocysteine quartiles (>6.31 micro mol/L), risk of invasive cervical cancer was substantially and significantly elevated (fully adjusted RR, including serologic HPV-16 status = 2.4-3.2; P for trend = 0.01). This strong relationship suggests that circulating homocysteine may be 1) an especially accurate indicator of inadequate folate, 2) an integratory measure of insufficient folate in tissues or 3) a biomarker of disruption of one-carbon metabolism. The contribution of common polymorphisms in one-carbon pathway genes, as well as inadequate vitamin B-6, vitamin B-12 and/or riboflavin, to elevated homocysteine, inefficient one-carbon metabolism and increased cervical cancer risk merits further exploration.