Techniques in pharmacovigilance
The Pharmaceutical Journal Vol 264 No 7101p922-923
June 17, 2000 Articles
Techniques in pharmacovigilance
By Ian C. K. Wong, PhD, MRPharmS, and Dimah Sweis, BSc
This article introduces study designs and methods used in pharmacovigilance studies
In a previous article,1 we introduced readers to pharmacovigilance resources in the United Kingdom. The present article introduces the study designs and methods utilised in undertaking pharmacovigilance. Only epidemiological techniques, such as cohort and case-control studies, are able to quantify the risks of adverse drug reactions (ADRs). Therefore, it is not too surprising that most of the techniques and study designs employed in pharmacovigilance are those used in epidemiology.
Study designs used in pharmacovigilance
Case reports and case series Case reports are concerned with single patients who have been exposed to a drug and experienced an adverse reaction. Case series can refer to a collection of patients who have been exposed to the same drug and whose clinical outcomes are then evaluated and described. Alternatively, a case series can be a collection of patients with a single outcome, whose history is then checked to ascertain previous drug exposure.
Case reports and case series are useful in the generation of drug safety signals. However, they are rarely able to determine whether or not the drug was the cause of the adverse outcome, nor can they quantify the incidence of adverse outcomes. Hence, more rigorous studies are normally undertaken to confirm or refute the findings of the case reports or case series. For example, it has been reported that some patients with epilepsy had developed irreversible visual field constriction after long-term treatment with vigabatrin.2,3 These case reports were useful in producing a drug safety signal. As a result, more rigorous studies were conducted to investigate the link between the drug and the adverse event and to determine the incidence of this event.4
Analysis of secular trends The type of study known as analysis of secular trends examines trends in an exposure that is a presumed cause of an ADR and trends in a disease that is a presumed effect of an ADR, testing whether these trends coincide. Vital statistics, such as mortality rates, are often used for these studies. A good example is the correlation between the introduction of isoprenaline forte and fenoterol inhalers and the incidence of death from asthma in New Zealand.5 Analysis of secular trends is useful for rapidly providing evidence for or against a hypothesis. However, these studies lack data on individuals; therefore, it is not possible to control for confounding factors, making the results open to misinterpretation.
Case-control studies Case-control studies are retrospective studies where a group of patients with a particular disease (or ADR) are compared with a group of patients who do not have the disease, and their histories of previous exposure to a "risk factor" are compared. For example, the World Health Organisation identified 1,143 female patients who had developed idiopathic venous thromboembolic events (VTE) in 21 countries. The WHO then selected 2,998 controls (women who did not experience VTE). Previous exposure to oral contraceptives was then compared between the two groups. Patients who developed VTE were more likely to have had exposure to oral contraceptives; therefore, the study suggested that oral contraceptives were associated with VTE.6 Strengths associated with case control studies include that:
- They can be used to study multiple drug exposures
- The can be used to study uncommon diseases
- They are relatively easy, fast and cheap in terms of data collection
However, case control studies also have their weaknesses:
- Finding appropriate matched controls may be difficult
- They are prone to "recall bias", ie, people with a disease remember things differently from those not affected
- As data are collected retrospectively, some information may not available
Cohort studies Cohort studies compare a group of individuals with a particular drug exposure to a group without the same exposure in terms of adverse outcomes. The study can be either prospective or retrospective. For example, a cohort study compared 90,000 patients who had been exposed to co-amoxiclav with 360,000 patients who had been exposed to amoxycillin. The study found that the co-amoxiclav users were six times more likely to experience acute liver toxicity.7 The strengths of cohort studies are that:
- They allow calculation of incidence rates and precise risks
- They can be used to study multiple outcomes
- They provide unbiased drug exposure data
- Selection bias is less likely than in case-control studies
Their weaknesses are that:
- They are relatively more expensive, especially the ad hoc studies
- It is difficult to gather a large number of subjects - if cohort studies are done prospectively, thay can take years to complete
- Outcome data may be biased, especially in retrospective studies
Meta-analyses In this era of evidence-based medicine, meta-analyses allow investigators to review clinical trial data systematically. Although most investigators are interested in the efficacy of treatments, it is also possible to investigate adverse outcomes systematically. A systematic review of randomised controlled trials involving administration of human albumin in critically ill patients has been undertaken.8 In this review, 30 randomised controlled trials were analysed, including 1,419 randomised patients. It was found that for each patient category, the risk of death in the albumin treated group was higher than in the comparison group. The data gathered suggested that for every 17 critically ill patients treated with albumin there is one additional death. It therefore suggested that the use of human albumin in critically ill patients should be urgently reviewed. This particular meta-analysis generated a lot of interest and debate regarding the methodology and conclusions.
It is important to note that, although a meta-analysis increases the statistical power to detect a difference between treatments, it is still insufficient to detect rare events. A good example is the incidence of suicide with fluoxetine.9 The study analysed pooled data from 17 double-blind clinical trials in patients with major depressive disorder, comparing fluoxetine (n=1,765) with tricyclic antidepressants (n=731) or placebo (n=569), or both. The results of the meta-analysis suggested that fluoxetine was not associated with a greater risk of suicide when compared with tricyclic antidepressants. However, the statistical power of this may have been too low to detect a difference between the two groups.10
In the past two decades, pharmacovigilance scientists have increasingly adopted the fundamentals of epidemiological research and applied them to the study of drugs. These techniques give us better understanding of the beneficial and adverse outcomes of medicines use.
No doubt they will continue to play important roles in pharmacovigilance.
Ian Wong is lecturer in pharmacy practice and Dimah Sweis is a research student at the Pharmacy Practice Research Unit, School of Pharmacy, University of Bradford, Bradford BD7 1DP. Correspondence to Dr Wong (e-mail: email@example.com)
|1.Wong ICK. Pharmacovigilance resources in the United Kingdom. Pharm J 1999;263:285-8.|
|2.Eke T, Talbot JF, Lawden MC. Severe persistent visual field constriction associated with vigabatrin. BMJ 1997;314:180-1.|
|3.Wong ICK, Mawer GE, Sander JWAS. Severe persistent and visual field constriction associated with vigabatrin - reaction might be dose dependent [letter]. BMJ 1997;314:1693-4.|
|4.Wilton LV, Stephens MD, Mann RD. Visual field defect associated with VGB: observational cohort study. BMJ 1999;319: 1165-6.|
|5.Beasley R, Smith K, Pearce N, Crane J, Burgess C, Culling C. Trends in asthma mortality in New Zealand, 1908-1986. Med J Aust 1990;152:572-3|
|6.World Health Organisation Collaborative Study of Cardiovascular Disease and Steroid Hormone Contraception. Venous thromboembolic disease and combined oral contraceptives: results of international multicentre case-control study. Lancet 1995;346:1575-82.|
|7.Garcia Rodriguez LA, Stricker A, Zimmerman HJ. Risk of acute liver injury associated with the combination of amoxicillin and clavulanic acid. Arch Intern Med 1996;156:1327-32.|
|8.Cochrane Injuries Group Albumin Reviewers and Chochrane Injuries Group Albumin Reviewers. Human albumin administration in critically ill patients; systematic review of randomised controlled trials: Why albumin may not work. BMJ 1998;317:235-40.|
|9.Beasley CM, Dornseif BE, Bosomworth JC, Sayler ME, Rampey AH, Heiligenstein JH et al. Fluoxetine and suicide: Meta-analysis of controlled trials of treatment for depression. BMJ 1991;303:685-92.|
|10.Li Wan Po A. Fluoxetine and suicide: meta-analysis and Monte-Carlo simulations. Pharmacoepidemiol Drug Safety 1993;2:79-84.|
Citation: The Pharmaceutical Journal URI: 20001845
Recommended from Pharmaceutical Press