¹Drug Safety Research Unit, Bursledon Hall, Southampton SO31 1AA, UK

²London School of Hygiene and Tropical Medicine, London, UK

Correspondence to: Dr PN Biswas, Drug Safety Research Unit, Bursledon Hall, Blundell Lane, Southampton S031 1AA, UK. E-mail: pipasha.biswas@dsru.org

Valsartan is a second class of angiotensin II receptor antagonist, indicated for the treatment of hypertension. The objective of the study was to monitor the safety of valsartan using the technique of prescription event monitoring (PEM), in patients who were prescribed this drug by general practitioners (GPs) in England. PEM is a noninterventional observation cohort technique. Exposure data were obtained from dispensed prescriptions issued between December 1996 and November 1998. Outcome data were obtained by sending questionnaires to prescribing GPs. The cohort comprised 12 881 patients. Events most frequently reported as suspected adverse drug reactions were malaise/lassitude (37; 0.3% of total cohort), dizziness (19; 0.1%), and unspecified side effects (57; 0.4%). Events with the highest incidence density (ID₁ per 1000 patient-months of treatment) in the first month of treatment were malaise/lassitude (15.6), dizziness (11.8), and headache/migraine (10.9). Most frequent reasons for stopping valsartan were not effective (847; 6.6% of total cohort), malaise/lassitude (265; 21%), and dizziness (146; 1.1%). No unexpected serious adverse events were identified. Other events assessed as possibly related to valsartan use were impotence (37), dizziness (19), cough (9), facial oedema (5), hyperkalaemia (3), and angioneurotic oedema (1). There were four reports of exposure during pregnancy and 203 deaths (1.5%) in this cohort. In conclusion, this study monitored the safety profile of valsartan in a large cohort of patients in general practice in England. No untoward features other than dizziness were identified that were not mentioned in the prescribing guidance.

Journal of Human Hypertension (2002) 16, 795-803. doi:10.1038/sj.jhh.1001490

valsartan; prescription event monitoring; safety; adverse drug reactions

Introduction

The safety of new antihypertensive agents is of special importance as they are likely to be used long term in considerable number of patients in general medical practice. The number of subjects forming the safety database for successful product licences for drugs containing new active substances is only limited¹ and small compared with the number of patients likely to receive a new drug for a common condition such as hypertension; also the duration in clinical trials is short.

Valsartan is the second of the class of angiotensin II receptor antagonists to be licensed in the UK for the treatment of hypertension. Valsartan acts selectively at the AT₁ receptor subtype, inhibiting the vasoconstrictor and pressor response exerted by angiotensin II. It also blocks angiotensin-II-induced promotion of sodium retention, and stimulation of aldosterone secretion. It has been suggested that the site specificity of valsartan could reduce the incidence of cough and angioneurotic oedema associated with ACE inhibitors.² Hence, pharmaco-logically, it offers advantages of increased selectivity, specificity, and maintained blockade of the circulating and tissue renin-angiotensin system at the AT₁ receptor without the adverse reactions normally associated with angiotensin-converting enzyme (ACE) inhibitors.³

Also, valsartan does not undergo significant hepatic metabolism.⁴ The drug is contraindicated in pregnancy and in patients with severe hepatic impairment, cirrhosis, and biliary obstruction.

This paper reports the results of a non-interventional observational cohort study undertaken to monitor the safety of valsartan in a large population of patients treated in general practice in England. The study began as soon as the drug was marketed in the UK, and so the patients who formed the cohort are representatives of the first patients who were prescribed the drug in England. Also, there were no exclusion criteria for a patient's participation in the study. The patients were identified from dispensed prescriptions, so there was no influence on general practitioners (GPs) at the time of prescribing.

Method

An observational cohort study was undertaken using the technique of prescription event monitoring (PEM).⁵ In PEM, the patients are identified from dispensed National Health Service prescriptions written by GPs in England. Information from the prescriptions for valsartan issued between December 1996 to November 1998 was supplied in confidence by the Prescription Pricing Authority (PPA). This comprised the exposure data.

The outcome were obtained by sending a simple questionnaire ('green form') to the prescribing GP at least 6 months after the date of the first prescription for each individual patient. The green form requested information on the age and gender of the patient, the reason for stopping the drug if it had been stopped, the GP's opinion of the effectiveness of the drug, and all medical events that had occurred after valsartan was prescribed.

Only one green form was sent to each patient. The study cohort was formed of those patients for whom the GP returned a green form containing clinical data. Green form questionnaires that were blank and those which informed that the patient was no longer registered at the practice or that the drug had not been taken were classified as void and not considered further.

The term 'event' was defined as including any new diagnosis, any reason for referral to a consultant or admission to hospital any unexpected deterioration (or improvement) in a concurrent illness, any suspected drug reaction, or any complaint considered to be of sufficient importance to enter into the patient's notes.

All events reported on green forms were coded onto the computer using the in-house hierarchical dictionary arranged in a system-organ classification. The event data are presented as the number of first reports for each event in each of the first 6 months of the observation period and the total number of events reported. Reasons for stopping treatment and events reported as suspected adverse drug reactions (ADRs) were reviewed in detail. Additional information was obtained for selected adverse events by sending a supplementary questionnaire to the GP who first prescribed valsartan. For these events, medically qualified staff assessed the causal relationship between valsartan use and the event. The causality assessment considered criteria such as temporal relationship, the effects of concomitant medication, the underlying illness or concurrent illness, and the quality of supportive information such as results of laboratory tests and other investigations. All green forms were reviewed by a medically qualified clinical research fellow who followed up medically important events together with reactions associated with products in the same therapeutic class, reactions highlighted during premarketing development and in postmarketing phase in other countries. Further details of medical history for those patients experiencing these selected events were obtained in confidence by a detailed anonymised follow-up questionnaire to the reporting GP. Individual case reports were assessed for causality using four basic considerations (temporal relationship, pharmacological plausibility, clinical and pathological characteristics of the event, and the likelihood of exclusion of other possible causes). This assessment took into account the time to onset of the event in relation to the start of treatment, past medical history, concurrent illnesses, concomitant medication, whether the event was the reason for stopping treatment, if yes the resolution of the event (with or without specific treatment), whether the symptoms returned on re-challenge if the drug was taken again, whether it was pharmacologically plaus-ible or whether an alternative cause was specified, and the GP's opinion. The relationship between the event and valsartan exposure was categorised as probable, possible, unlikely or not assessible.

Specific supplementary questionnaires were sent to determine the outcomes of pregnancies. Further information was obtained regarding the cause of death where the cause was not specified.

This study was conducted in accordance with the International Ethical Guidelines for Biomedical Research prepared by the Council for International Organizations of Medical Sciences in collaboration with the World Health Organization.⁶ The method also complies with the guidelines on the Practice of Ethics Committees in Medical Research involving Human subjects, as issued by the Royal College of Physicians of London,⁷ and the recently published Multicentre research ethics guidance notes.⁸

Analysis

Incidence densities (IDs) were calculated for a given time period (t) for all events reported during the treatment for patients for whom either the date of stopping the drug is known or who continue to take the drug.⁹ The figures are expressed as ID per 1000 patient-months of treatment. IDs were calculated for all events reported during the treatment in the first month after the start of therapy (ID₁), for the second to sixth months (ID₂), and for the overall treatment period (ID_A).

The difference between the ID₁ and ID₂ and the 99% confidence intervals around the point estimate were calculated to examine whether the event rate between these two periods was increasing or decreasing. If the 99% confidence limits around the point estimate of a positive difference between the rate of events in month 1 (ID₁) and months 2-6 (ID₂) did not include the null value, this indicated that the rate of events in the first month was significantly greater than that in the second to sixth months. This can be considered to be a signal of an adverse event associated with starting valsartan.

Results

A total of 14 127 (55%) of the 25 838 green forms posted were returned. In all, 1246 (9%) were classified as void (patient no longer registered with doctor, 762; blank forms, 246; no record of treatment in notes, 166; valsartan prescribed but not taken, 22; duplicate green form for patient, 33; patient's doctor died, moved, or retired, 17), and so useful clinical information was available for 12 881 patients.

The age distribution and sex of the patients in the cohort are shown in Figure 1. The mean age (±s.d.) for males was 61.1±12.1 years, and the mean age for females was 65.4±12.5 years. The age was not recorded for 1501 (11.7%) of the patients. Of the 12 881 patients, 5215 (40.5%) were males and 7600 (59.0%) were females. Sex was not specified for 66 (0.5%) of the patients.

The major indication for prescribing valsartan was as expected, the licensed indication, hypertension (64.3%). Cough was the indication in 249 (1.9%) & for hypertension 8280 (64.3%) of patients. The indication was not specified for 3765 (29.2%) of the patients (See Table 1).

Valsartan appeared to be well tolerated as judged by the length of time the patients continued with treatment. After 6 months, 10 056 (81.1%) of the 12 407 patients for whom it was recorded that treatment was continuing or in whom the date of stopping medication was given were still being prescribed valsartan.

Where GPs gave an opinion on the green forms regarding effectiveness (86.9% of the total cohort), valsartan was reported to have been effective in 9489 (84.8%) of the patients.

An event was coded as an ADR if the GP specified on the green form that the event was attributed to a drug. In total, 295 events in 209 (1.6% of the cohort) patients were reported by the GPs to have been adverse reactions to valsartan. Of the 295 (4.4%) events, 13 were documented on the green form as having been reported to the Committee on Safety of Medicines (CSM). The most frequently reported adverse reaction to valsartan was unspecified side effects in 57 (0.4% of total cohort) patients, followed by malaise in 20 (malaise/lassitude 37 patients, 0.3%), and dizziness in 19 (0.1%). The most frequently reported adverse reactions are shown in ranked order in Table 2. There were also two reports of drug interaction, one with ibuprofen causing indigestion and heartburn, and the other with warfarin caused 'deranged INR' reported as suspected adverse drug reactions.

The most frequent reasons for stopping treatment with valsartan were 'not effective'¾847 reports (6.5% of the total cohort), followed by malaise/lassitude¾265 cases (2.0% of the total cohort), and dizziness¾146 cases (1.1% of the total cohort) as shown in ranked order in Table 3. A total of 3196 reasons for discontinuing valsartan were reported for 2562 (19.9%) patients.

Other events for which a high proportion of patients discontinued valsartan were hyperkalaemia (10/18; 55.5% of the total number of events reported), impotence (27/55; 49.0%), malaise/lassitude (265/539; 49.0%), dizziness 146/350; 41.7%), headache/migraine (134/358; 37.4%), palpitation (35/111; 31.5%), epistaxis (20/64; 31.2%), and cough (114/359; 30.0%).

Excluding indication-related events (hypertension and dose increased), malaise/lassitude was the adverse event with the highest ID in the first month of treatment (15.6 per 1000 patient-months of exposure), followed by dizziness (11.8 per 1000 patient-months of exposure). ID₁ was significantly greater than ID₂ for the adverse events; malaise/lassitude, dizziness, cough, nausea/vomiting, intolerance, diarrohea, dyspnoea, impotence/ejaculation failure, dyspepsia, oedema, and rash (Table 4), although some of these events are likely to be confounded by indication. All of these events except dizziness and rash are listed in the Summary of Product Characteristics (SmPC) as adverse events observed in clinical trials.

The events with the highest proportion reported during the first month of treatment were angioneurotic oedema (60.0%), palpitation (45.9%), dizziness (40.5%), headache/migraine (36.8%), nausea/vomiting (35.7%), impotence (35.1%), malaise/lassitude (34.8%), myalgia (32.8%), cough (32.5%), insomnia (27.7%), abnormal taste (25.0%), rash (23.8%), oedema (20.9%) and depression (17.7%). The events that were listed in the SmPC were malaise/lassitude, epistaxis, abnormal liver function tests (LFTs), hyperkalaemia, hyponatraemia, and angioneurotic oedema (Table 5).

The events that were not listed in the SmPC and were generated as potential signals in this study were cough, headache/migraine, dizziness, nausea/vomiting, oedema, rash, depression, palpitations, myalgia, impotence, insomnia, and abnormal taste. In our study, we found that high proportion of these events were reported during the first month of treatment and were also the reason for discontinuing valsartan (Table 6).

Selected events of interest that were reviewed and followed up, and after assessment were thought to be causally associated with valsartan use were rash, insomnia, abnormal taste, abnormal LFTs, impotence, facial oedema, allergy, myalgia, and neutropaenia.

In our study, there were 22 reports of abnormal LFTs reported during treatment with valsartan, of which only two were reported in the first 30 days of treatment, although there were four reports of stopping valsartan. One case of preexisting mildly abnormal LFTs, which became worse on treatment and returned to pretreatment levels off-treatment, was assessed as being a possible side effect of valsartan. In the further 12 cases, it was thought that valsartan may have been possibly associated with the event, while in the remaining nine reports the cause was unlikely to be associated with valsartan use.

There were two reports of hepatitis, four reports of jaundice and one of obstructive jaundice during treatment with valsartan. Of the four cases of jaundice, in one case in which it was the reason for stopping valsartan, the jaundice resolved on stopping the drug. The remaining two cases of jaundice were assessed as unlikely to be due to valsartan exposure, as alternative causes were given.

There were two reports of hepatitis of which one was reported by the GP as an ADR to valsartan (this case was also reported to CSM), while in the other case the patient was diagnosed to have infectious hepatitis and thus unlikely to be related to valsartan exposure. There was one report of obstructive jaundice that was due to the pancreatic cancer.

We also examined the report indicative of spontaneous bleeding disorders, to see whether these events were associated with the use of valsartan. The events reviewed under the spontaneous bleeding disorder category were haematuria (29 reports), subconjunctival haemorrhage (14 reports), and haemoptysis (seven reports), bleeding gums (four reports), spontaneous haemorrhage (four reports), and retinal haemorrhage (one report). At the time of launch of valsartan, spontaneous bleeding was not listed in the SmPC as a possible adverse event.

Pregnancies

There were four pregnancies reported during the study. All four mothers took valsartan during the first trimester.

In the first patient, the pregnancy was therapeutically terminated because of hypertension. There were no fetal abnormalities and the mother continued to take valsartan.

In the second case, the patient became pregnant and was also taking bendrofluazide. The pregnancy ended with a spontaneous first trimester miscarriage. In the third case also, the patient ended with a late first trimester incomplete miscarriage.

In the fourth case, the patient had a live male infant born by normal vaginal delivery with no abnormalities.

Deaths

There were a total of 203 deaths reported during the total period of observation for valsartan. The cause of death was established for 198 of these cases (97.5%); cardiovascular causes accounted for 121 (59.6%) of the deaths, and cancer for 39 (19.2%). The cardiovascular causes included myocardial infarction, 39; ischaemic heart disease, 26; cardiac failure, 18; and hypertension, 2.

Discussion

The present study was a prescription-based observational cohort study in pharmacovigilance. There was no interference with the decision of the doctors about the drug to be prescribed for their individual patients, and so avoided this potential selection bias.

The study included 12 881 patients (10 1429 patient-months of exposure to valsartan). The study was of national proportions and systematic in that the entire cohort for whom prescriptions were available represented the first patients given valsartan after its introduction into clinical practice in England. The analysis was based upon event data and so is theoretically capable of identifying signals that doctors do not suspect to be caused by the drug. The methodology allows follow-up of events of interest, deaths, and pregnancies during or following treatment with valsartan.

A possible weakness of the study was the response rate. Only 14 127 (55.0%) of the 25 838 green forms that were posted were returned, which could conceal biases.¹⁰ If the patients whose GPs failed to return the green forms were different from the patients whose GPs supported the study, then this difference could have remained undetected. Nevertheless, the response rate is substantial compared to the rate of reporting in spontaneous adverse-reaction reporting schemes. It is also satisfactory compared to the response rate in GP postal surveys in general.¹¹ The response rate for the losartan PEM study was similar at 54.6%.¹² The green forms were sent to the prescribers 6 months after the first prescription of valsartan; thus, a limitation of the study is the lack of capture of any events which occurred more than 6 months after the first prescription.

This study is the largest safety study of valsartan so far reported (MEDLINE and EMBASE search criteria). In this study, valsartan appeared to be well tolerated with majority of the patients as 81.1% of the 10 056 patients for whom date of the stopping was known were still being prescribed the drug after the end of 6 months. This is similar to the results with losartan (82.5%).¹² Also, it was said to have been effective in 84.8% of the 11 191 patients for whom the GPs gave an opinion regarding effectiveness.

The events were reported most frequently throughout the treatment period were hypertension, dose increase, malaise/lassitude, dizziness, headache/migraine, and nausea/vomiting. These were also the events (excluding indication-related events) with the highest incidence rate in the first month of treatment, with the highest number of reports in the first month of treatment and as the reasons for stopping treatment. These were also the events reported frequently as adverse events in clinical trials.

Dizziness was the most frequently reported event. Data from published studies of all angiotensin II receptor antagonists report dizziness as the most frequent clinical adverse event.¹³ Dizziness did not occur more commonly than placebo in clinical trials,¹⁴ nor did it appear as a side effect on the SmPC for valsartan.⁴ In this study, dizziness is reported as an event in the first month of treatment (ID₁-ID₂ 9.1 per 1000 patient-months treatment and 99% CI 6.5-11.7) and also as a reason for stopping valsartan in 146 patients, which is similar to losartan (which included dizziness as an ADR in its SmPC).

Cough was one of the frequently reported events during this study, but may represent a 'carry over' symptom from previous ACEI therapy complicated by cough.¹⁵ Cough is a class effect of ACE inhibitors to the nonspecificity of the ACE. One clinical trial has shown that the incidence of cough under valsartan was similar to that observed under diuretic therapy, but significantly lower than that observed under lisinopril.¹⁶

Angioedema is a well-recognized adverse effect of ACE inhibitors thought to be due to the accumulation of bradykinin. There are several reports in the literature of angioedema¹⁷ during exposure to losartan, including four possible cases in the PEM study of losartan¹². Angioedema occurred once in 3204 patients exposed to valsartan in the randomised controlled trails quoted in the valsartan SmPC.⁴ Angioedema can occur with may substances including drugs and some foods; thus, it is difficult to determine whether these episodes truly represent a valsartan-induced ADR. In this study, there were five reports of angioedema, of which three occurred in the first 30 days of starting valsartan. In one patient, there was re-challenge with valsartan with a re-occurrence of angioedema, which makes it probable that this was an ADR to the drug. In the other two cases, it was assessed as possibly associated with valsartan use.

Epistaxis was reported by 0.5% of the cohort. It is a labelled side effect of valsartan but not of the other angiotensin II antagonists. It is also mentioned in the patient information leaflet for valsartan. Spontaneous bleeding disorders as an adverse event was not listed in the SmPC for valsartan, and there were also no reports of its association in the literature search. It is sometimes difficult to assess the causality of spontaneous bleeding with the use of valsartan, because of confounding and limited information from spontaneous reports. A further investigation of spontaneous bleeding disorders is planned.

Fatigue was a common drug-related adverse effect during clinical trials of angiotensin II and is a labelled event for valsartan occurring in 1-10% of patients. Fatigue (coded as lassitude) in this study was reported by 2.5% of the patients and so the frequency is within the range found in clinical trials.

In this study, rash was reported in 147 patients (1.1% of the total cohort) during treatment with valsartan, of which 35 were reported in the first months of starting valsartan. Over one-third (53 reports; 36.0%) of the patients discontinued treatment with valsartan. Although rash is not listed in the SmPC for valsartan, 19 reports of rash have been sent to the CSM by 6/12/1999 (Drug Analysis Print from the MCA).¹⁸

Insomnia was reported in 0.4% of the cohort with 54 reports during the treatment period with valsartan, of which 14 (27.7%) were reported in the first month of treatment. Again, there is no mention of insomnia as an adverse event in the SmPC for valsartan.

Taste abnormality has not been listed as an adverse effect in the SmPc, but three reports of taste alteration have been sent to the CSM by 6/12/1999. In this study, there were 12 reports of abnormal taste, of which there were reported in the first month of treatment. It was also the reason for stopping treatment in five cases. Ten cases of abnormal taste were assessed as being possibly caused by exposure to valsartan, in this study.

In controlled clinical trials in 3204 patients with hypertension, there were rare cases of valsartan being associated with decrease in haemoglobin and occasional neutropaenia.⁴

In this study, four cases of neutropaenia were reported and after assessing all the four cases, only two cases of mild neutropaenia were assessed as possibly due to exposure to valsartan. Both these cases resolved on stopping valsartan.

Valsartan is contraindicated in patients with severe hepatic impairment, cirrhosis, and biliary obstruction. Occasional elevations of liver function tests, creatinine, and bilirubin were seen with the use of valsartan and abnormal LFTs are listed in the SmPC for valsartan and it mentions that clinical trials showed occasional elevations of LFTs but no more than in patients taking placebo.⁴ A total of 13 cases of abnormal LFTs were assessed as possible related to valsartan use in our study, while in the remaining cases it was difficult to evaluate the role of exposure to valsartan as there was confounding by indication (hepatic engorgement from heart failure), by other illness (eg diabetes, obesity, gall stones) or by other exposures such as alcohol.

There is no mention of the musculoskeletal events during the clinical trials of valsartan and they are not labelled possible side effects for valsartan.⁴ Myalgia is a labelled side effect for losartan and back pain is for candesartan. Until 6/12/99, the following number of suspected ADRs of musculo-skeletal events had been reported to the CSM for valsartan: joint stiffness, 2; muscle cramps, 3; and myalgia/myalgia aggravated, 5. In our study, there were four cases of myalgia and five of joint pain assessed as possibly associated with valsartan use.

The SmPC,⁴ suggests that serum potassium should be monitored in patients with renal impairment, the elderly, or those on potassium supplements. Clinical trials showed occasional elevation of serum potassium, but no patient discontinued treatment because of this. Occasional minor elevations of creatinine were also seen in controlled trials. In this study, three patients stopped valsartan because of hyperkalaemia and two because of raised creatinine, which then reverted to previous levels.

This is likely to have been confounding by indication for cardiovascular-related events such as palpitations, chest pain, and dyspnoea. Also, worsening of cardiac failure probably represents a therapeutic failure rather than a possible ADR. Reports of these events were also likely to be confounded by indication. This may also apply to impotence. Follow-up of cases of impotence suggests that in the majority of the cases this was not a drug-related event.

The use of valsartan is contraindicated during pregnancy, but there were four pregnancies reported during this study. All four mothers were exposed to valsartan during the first trimester of pregnancy, and the outcome was one live birth with no abnormalities, one therapeutic termination of pregnancy, and two spontaneous first trimester miscarriages.

Conclusion

In this PEM study, valsartan appears to be well tolerated. The majority of events reported were assessed as not adverse drug reactions, and the most frequently reported events were as a result of confounding due to reporting of symptoms associated with the disease being treated. Dizziness is a labelled adverse effect for losartan but not for valsartan. There is likely to have been confounding by indication for cardiovascular-related events such as palpitations. Confounding by indications may apply to impotence. Myalgia may be a rare adverse drug effect to valsartan. Cough was a commonly reported event, but this may be due to carry-over effect from former use of ACE inhibitors. A further investigation of cough and spontaneous bleeding is planned. This study defines the safety profile of valsartan as was reported by general practitioners in England.

Acknowledgements

We thank the very many GPs who have, without payment, participated in this study. We also thank the prescription pricing Authority, The office of the National Statistics, and the Regional Authorities of the NHS, without whose support PEM could not have been undertaken. Special thanks goes to Mrs. Gillian Pearce, Data Processing Manager, Mr. Shayne Freemantle, IT Manager for providing the data required for this study, and Mrs. Lesley Flowers for typing the manuscript. The DSRU is a medical charity, which has received unconditional grants from many pharmaceutical companies, including the manufacturer of valsartan. The organisations have no control over the decision to undertake a study, or the conduct and reporting of the studies.

1 Rawlins MD, Jefferys DB. United Kingdom product license applications involving new active substances, 1987-1989: their fate after appeals. Br J Clin Pharmacol 1993; 35: 599-602. MEDLINE

2 Ramsay LE, Yeo WW. ACE inhibitors, angiotensin II antagonists and cough. J Hum Hypertens 1995; 9: ((Suppl 5)) 551-554.

3 Lacourciere Y, Lefebvre J. Modulation of the renin-angiotensin-aldosterone system and cough. Can J Cardiol 1995; 11: ((Suppl F)) 33F-39F. MEDLINE

4 Diovan Summary of Product Characteristics. Novartis, 1996.

5 Mann RD. Prescription Event Monitoring¾recent progress and future horizons. Br J Clin Pharmcol 1998; 46: 195-201.

6 WHO. International Ethical Guidelines for Biomedical Research Involving Human Subjects Geneva: CIOMS/WHO, 1993.

7 Guidelines on the practice of Ethical Committees in Medical Research involving Human Subjects. Royal College of Physicians of London, August 1996.

8 Multicentre Research Ethics Committee. Examples of enquiries and surveys in the public interest where no reference to a Research Ethics Committee is necessary. Multi-Centre Research Ethics Committees Guidance Notes 2000; Appendix C: 19.

9 Freemantle SN et al. The incidence of the most commonly reported events with 40 newly marketed drugs¾study by Prescription-Event Monitoring. Pharmacoepidemiol Drug Safety 1997; 6: ((Suppl 1)) 1-8.

10 Templeton L et al. Surveying general practitioners: does a low response rate matter? Br J Gen Pract 1997; 47: 91-94. MEDLINE

11 McAvoy BR, Kramer EFS. General practice postal surveys: a questionnaire too far? BMJ 1997; 313: 732-733.

12 Mann RD et al. Losartan: a study of pharmacovigilance data on 14 522 patients. J Hum Hypertens 1999; 13: 551-557. Article MEDLINE

13 Mazzolai L, Burnier M. Comparative safety and tolerability of Angiotensin II receptor antagonists. Drug Safety 1999; 21: 23-33. MEDLINE

14 Markham A, Goa KL. Valsartan: a review of its use in essential hypertension. Drugs 1997; 54: 299-311. MEDLINE

15 Mackay FJ, Pearce GL, Mann RD. Cough and angiotensin II inhibitors: cause or confounding? Br J Clin Pharmacol 1999; 47: 111-114. Article MEDLINE

16 Benz J et al. Valsartan, a new angiotensin II receptor antagonist: a double blind study comparing the incidence of cough with lisinopril and hydrochlorothiazide. J Clin Pharmacol 1997; 37: 101-107. MEDLINE

17 Israili ZH, Hall WD. Cough and angioneurotic oedema associated with angiotensin-converting enzyme inhibitor therapy: a review of the literature and pathophysiology. Ann Intern Med 1992; 117: 234-242. MEDLINE

18 Heeley E et al. Prescription event monitoring and reporting of adverse drug reactions. Lancet 2001; 358: 1872-1873. Article MEDLINE

Figure 1 Age distribution and sex of patients for valsartan.

Table 1 Indications analysed by sex

Table 2 Adverse reactions (ADRs) to valsartan

Table 3 Reasons for stopping valsartan

Table 4 Incidence densities (IDs) ranked for valsartan in order of ID₁ per 1000 patient-months treatment

Table 5 Proportion of reported events as reason for stopping, suspected ADRs and events included in the SmPC

Table 6 Proportion of reported events as reason for stopping, suspected ADRs. Adverse events not listed in the SmPC