Glomerulonephritis and malignancy: a population-based analysis

An association between glomerulonephritis and malignant tumors has previously both been found and discarded in clinical series, but to our knowledge never has been tested in a population-based setting.

The Danish Kidney Biopsy Registry includes all kidney biopsies performed from 1985. Using a unique personal identification number, each person in the registry to the National Population Registry and the Danish Cancer Registry were linked. Cancer occurrence after the biopsy was compared in patients with morphological, glomerular diseases with that of the general Danish population, taking into account sex, age, calendar period and time since biopsy, and the 95% confidence interval (95% CI) for the observed-to-expected rates was calculated, assuming a Poisson distribution. Cancer occurrence was stratified to <1 year, 1 to 4, and >or=5 years after a biopsy.

A total of 102 de novo cancers were found in 1958 patients. These cancers represent a two- to threefold excess of the expected number at <1 and 1 to 4, but not >or=5 years after a biopsy. Non-Hodgkin's lymphomas were observed six to eight times more than expected. Cancer excess was seen in glomerulonephritides with a known or suspected virus etiology.

The excess cancer rate could be the result of underlying undiagnosed tumors whose antigens have initiated glomerulonephritis, or the immunosuppressive therapy that initiated or energized tumor cells. Based on the findings in our study, there is some support for an association to persistent viruses causing first the glomerulonephritides and then the malignancies, perhaps through a common pathogenesis. This calls for other studies to be done that are specifically designed to investigate this issue, with more data on patient characteristics and confounders.