Intended for healthcare professionals

CCBYNC Open access

Use of N-nitrosodimethylamine (NDMA) contaminated valsartan products and risk of cancer: Danish nationwide cohort study

BMJ 2018; 362 doi: (Published 12 September 2018) Cite this as: BMJ 2018;362:k3851

Linked editorial

Regulatory response to contaminated valsartan

Linked opinion

Fast-tracked pharmacoepidemiology in drug safety issues

Epidemiologic Study of NDMA Contaminated Valsartan Use And Risk of Cancer. Reassuring or Inconclusive?


We enjoyed reading the study by Pottegard et al.1 that examined the risk of NDMA contaminated valsartan and the risk of cancer. We commend the authors for addressing an important and timely question. A number of methodological issues must be addressed to better put the results into context.

First, an intention-to-treat approach used in this study is the not the optimal analytical approach to examine diseases with long latency like cancer. This is especially important with this particular study question as it is possible that patients taking NDMA valsartan may switch to other generic formulations during follow up increasing the possibility of exposure misclassification. The strongest analysis in the study is the dose response analysis which demonstrates adequate exposure over time. However, from the person time information provided the cumulative exposure follow up time spans from less than one year to 1.5 years for the lowest and highest dose categories respectively. These follow up times are not particularly sufficient for the initiation and detection of a malignant process that might result from NDMA contaminated valsartan use. Furthermore, since a competing risk bias analysis was not presented it is not known how many patients died of other causes and were not at risk of developing cancer.

All these reasons might have contributed to a small number of cancer events and the possibility of sparse data bias2. In their conclusion the authors state that there was a non-statistically significant risk with colorectal and uterine cancers. However, according to Figure 3 the upper bounds of the confidence intervals with all solid tumor risks presented are high and cannot exclude the possibility of an increased risk. The increase in the risk of cancer with incident users is concerning and studies with a larger sample and longer follow up that can also address competing risk bias are needed to better answer this question.


1. Pottegård A, Kristensen KB, Ernst MT, Johansen NB, Quartarolo P, Hallas J. Use of N-nitrosodimethylamine (NDMA) contaminated valsartan products and risk of cancer: Danish nationwide cohort study. BMJ 2018; 362 doi:

2. Greenland S, Mansournia MA, Altman DG. Sparse data bias: a problem hiding in plain sight.BMJ 2016;352:i1981.

Competing interests: No competing interests

19 September 2018
Mahyar Etminan
Associate Professor
Mohammad Ali Mansournia
Ophthalmology and Visual Sciences and Pharmacology, University of British Columbia
Departement of Ophthalmology and Visual Sciences, University of British Columbia, 323-2550 Willow Street, Vancouver, Canada