Re: Use of proton pump inhibitors and risk of hip fracture in relation to dietary and lifestyle factors: a prospective cohort study
Dear editor
I am concerned about the potential impact of recall bias with the collection of exposure and outcome data in the study by Khalili and colleagues, which reported on the association of PPI use and hip fracture[1]. They suggest in their introduction that previous epidemiological studies had “important limitations” because they had a “retrospective design”, whereas the Nurses Health Study is a “prospective cohort study”.
I do not agree with the suggestion of soundness (“prospective”) and unsoundness (“retrospective”). Various “previous epidemiological studies” that examined hip fracture with PPI use [2-4], have collected their data prospectively. However, the statistical analysis was conducted retrospectively. This is a well accepted and correct method under the assumption that there is no a priori association between exposure (PPI use) and outcome (hip fracture). In contrast, the Nurses Health Study retrospectively collected information on exposure and outcomes, during interviews that were conducted every two years (prospectively).
Key is the retrospective data collection (i.e. during the last two years) on both exposure and outcome using questionnaires: this makes the design particularly prone to recall bias (e.g. nurses with a hip fracture may remember their drug exposure differently as compared to those without a hip fracture). The direction of this differential misclassification is hard to predict. Of course, all studies [1-4] are also prone to non-differential misclassification of exposure, but its direction is at least predictable (a bias of the point estimate towards the null value).
With kind regards
Frank de Vries, PhD PharmD
[1] Khalili H, Huang ES, Jacobson BC, et al. Use of proton pump inhibitors and risk of hip fracture in relation to dietary and lifestyle factors: a prospective cohort study. BMJ2012;344:e372
[2] Pouwels S, Lalmohamed A, Souverein P, Cooper C, Veldt BJ, Leufkens HG, et al. Use of proton pump inhibitors and risk of hip/femur fracture: a population-based case-control study. Osteoporos Int 2011;22:903-10.
[3] Kaye JA, Jick H. Proton pump inhibitor use and risk of hip fractures in patients without major risk factors. Pharmacotherapy 2008;28:951-9.
[4] Vestergaard P, Rejnmark L, Mosekilde L. Proton pump inhibitors, histamine H2 receptor antagonists, and other antacid medications and the risk of fracture. Calcif Tissue Int 2006;79:76-83.
Competing interests:
No competing interests
08 February 2012
Frank de Vries
Pharmacist
Department of Clinical Pharmacy and Toxicology, Maastricht UMC+, Maastricht, Netherlands
Rapid Response:
Re: Use of proton pump inhibitors and risk of hip fracture in relation to dietary and lifestyle factors: a prospective cohort study
Dear editor
I am concerned about the potential impact of recall bias with the collection of exposure and outcome data in the study by Khalili and colleagues, which reported on the association of PPI use and hip fracture[1]. They suggest in their introduction that previous epidemiological studies had “important limitations” because they had a “retrospective design”, whereas the Nurses Health Study is a “prospective cohort study”.
I do not agree with the suggestion of soundness (“prospective”) and unsoundness (“retrospective”). Various “previous epidemiological studies” that examined hip fracture with PPI use [2-4], have collected their data prospectively. However, the statistical analysis was conducted retrospectively. This is a well accepted and correct method under the assumption that there is no a priori association between exposure (PPI use) and outcome (hip fracture). In contrast, the Nurses Health Study retrospectively collected information on exposure and outcomes, during interviews that were conducted every two years (prospectively).
Key is the retrospective data collection (i.e. during the last two years) on both exposure and outcome using questionnaires: this makes the design particularly prone to recall bias (e.g. nurses with a hip fracture may remember their drug exposure differently as compared to those without a hip fracture). The direction of this differential misclassification is hard to predict. Of course, all studies [1-4] are also prone to non-differential misclassification of exposure, but its direction is at least predictable (a bias of the point estimate towards the null value).
With kind regards
Frank de Vries, PhD PharmD
[1] Khalili H, Huang ES, Jacobson BC, et al. Use of proton pump inhibitors and risk of hip fracture in relation to dietary and lifestyle factors: a prospective cohort study. BMJ2012;344:e372
[2] Pouwels S, Lalmohamed A, Souverein P, Cooper C, Veldt BJ, Leufkens HG, et al. Use of proton pump inhibitors and risk of hip/femur fracture: a population-based case-control study. Osteoporos Int 2011;22:903-10.
[3] Kaye JA, Jick H. Proton pump inhibitor use and risk of hip fractures in patients without major risk factors. Pharmacotherapy 2008;28:951-9.
[4] Vestergaard P, Rejnmark L, Mosekilde L. Proton pump inhibitors, histamine H2 receptor antagonists, and other antacid medications and the risk of fracture. Calcif Tissue Int 2006;79:76-83.
Competing interests: No competing interests