Biomarkers and Thresholds to Identify Bone Turnover: A Retrospective Cohort Study of Bone Biopsies

Fig 2 from Sprague et al AJKD, © National Kidney Foundation.

Fig 2 from Sprague et al AJKD, © National Kidney Foundation.

Bone mineral disease (MBD) is common in patients with chronic kidney disease (CKD), yet remains a diagnostic and therapeutic enigma for clinicians and scientists. In contrast to complications such as anemia and acidosis, a definitive diagnosis of CKD-MBD cannot be made by blood tests alone, and require a bone biopsy. These biopsies are invasive, rarely done, and as such, treatment is often based on a presumed diagnosis made on the level of markers of bone turnover. The diagnostic accuracy of these markers, and the appropriateness of therapy to limit bone turnover, is largely unknown.

In a recent AJKD article, Sprague et al utilized bone biopsy data from 492 dialysis patients from four countries, and combined it with stored serum samples to determine the diagnostic accuracy of alkaline phosphatase (ALP) and intact parathyroid hormone (iPTH) in distinguishing low turnover from non-low, and high turnover from non-high bone disease on bone biopsy. They found that both markers showed modest discrimination (AUROC > 0.70, but < 0.80) for low or high turnover disease, and combinations of the markers were not more helpful than individual markers alone.

The authors also compared the accuracy of the KDOQI and KDIGO guidelines for intact and whole PTH for both low and high turnover disease, and found similarly modest results. Guideline thresholds for low turnover disease (150 pg/mL or < 2 times cutoff) were modestly sensitive (65-73%), but not very specific (57-67%). In contrast, thresholds for high turnover disease (300 pg/mL or 9 times cutoff) were poorly sensitive (30 to 58%), but more specific (78-88%). These findings suggest that patients with PTH values higher than 2 times normal are unlikely to have low turnover disease, and those with levels more than 9 times are very likely to have high turnover disease. For patients with values in between these cutoffs, the diagnostic tests appear to be not helpful.

Importantly, the authors also examined the diagnostic accuracy of P1NP, and found no improvement over traditional markers, but were unable to evaluate other novel biomarkers such as FGF-23 or klotho due to a lack of available serum. Future research using bone biopsy samples should examine these markers, and compare their accuracy to ALP and PTH. Other limitations of the study include indication bias, which is typical of all biopsy studies, as these patients undergoing a bone biopsy for clinical reasons may not be representative of all patients on dialysis. Nonetheless, the study by Sprague et al provides clinicians useful data to guide clinical decisions about PTH-lowering therapy in patients on dialysis, and provides cutoffs which may help guide treatment.

Navdeep Tangri, MD
AJKD Blog Contributor

To view the article abstract or full-text (subscription required), please visit

Leave a Reply

%d bloggers like this: