Co-prescribing of SSRIs and TCAs in Australia: how often does it occur and who is doing it?

Abstract

Aims

To determine the frequency with which the selective serotonin re-uptake inhibitor (SSRI) antidepressants are used as add-on therapy to the tricyclic antidepressants (TCA) rather than as replacement therapy.

Methods

The data analysed were profiles of prescription records by date of supply to the patient. From within the national administrative dispensing claims database, the subset eligible for social security entitlements was identified as individuals by means of their coded permanent identification numbers (PINs). Following the initial supply of an SSRI in January 1996, the subsequent 6 months dispensing of SSRI and TCA antidepressants to these individuals was examined. The main outcome measure was the proportion of individuals to whom SSRIs and TCAs were dispensed concurrently, as an indirect measure of coprescription. In instances where a patient was receiving prescriptions for SSRIs and TCAs that had been written by the one doctor only, the major specialty of the doctor was investigated.

Results

55 271 PINs were dispensed 63 865 SSRI prescriptions in January 1996 which represented over half (52%) of the total community SSRI prescriptions dispensed in that month. The number of these patients meeting the criteria for coprescription of SSRIs and TCAs over the next 6 months was 2773 (5%). The coprescribing instances were highest in Queensland and the prescribers most frequently involved had psychiatry major specialty codes.

Conclusions

Among SSRI users there is a cohort of patients who, within the same time frame, are receiving supplies of a TCA, the nonselective drug that the SSRIs were designed to replace. This is indirect evidence of probable coprescription. Such combination use is of uncertain clinical and cost effectiveness, and carries additional risks. The SSRIs were included on the subsidy list in Australia on the basis of reasonable cost effectiveness as monotherapy compared with the TCAs. Our data imply that for some patients, antidepressant prescribing is inconsistent with the basis on which government subsidy was approved.

Introduction

Fluoxetine, the first of a new class of antidepressant drugs – the selective serotonin re-uptake inhibitors (SSRIs), was approved for marketing in Australia in 1990. The rationale underlying the development of the SSRIs was that the receptor selectivity resulted in relative safety in overdose compared with the tricyclic antidepressants (TCAs), and to some extent a more favourable adverse effect profile. It was generally accepted that the SSRIs were no more effective than the tricyclic antidepressants in the treatment of moderate to severe depressive illness [1]. The SSRIs are substantially more expensive than the older agents, but this cost differential may be justified if it provides safer and better tolerated therapy, and if there are cost offsets elsewhere in the health system.

The market uptake of the SSRIs has been rapid and, remarkably, accompanied by only a relatively small decrease in the use of the TCAs (Figure 1). As a result, the overall antidepressant market has expanded dramatically, with utilization being nearly three times greater in 1998 than in 1990. Over this period, five SSRIs have entered the Australian market each with its accompanying promotional pressures. There has even been some suggestion of the possibility of legal liability in prescribing the older drugs if they have a greater likelihood of adverse outcome [2].

Figure 1.

Trends in antidepressants dispensed through community pharmacies in defined daily doses* (DDDs) per 1000 population per day, 1990–1998. TCAs, □; SSRIs, ; Moclobemide, other, ▪. Reference: Commonwealth Department of Health and Aged Care. Australian Statistics on Medicines 1998. AusInfo, Canberra 1999. *The defined daily dose (DDD) is based on the assumed average daily dose of the drug when used for its main indication by adults. The number of DDDs per 1000 of the population per day is the World Health Organization approved measurement unit for drug use studies.

In 1996 Parkes, using the Department of Veterans' Affairs dispensing claims database, noted that some new SSRI users who were previously taking TCAs appeared to be continuing on both drugs rather than switching from a TCA to an SSRI [3].

The combined use of SSRIs and TCAs is against the logic of subsidy approval and the recommendation of manufacturers [4]. The claimed advantages of the SSRIs relate to their receptor selectivity, and this is negated if they are used in combination with the less selective agents they were designed to supersede. Also, a clinically significant adverse drug interaction occurs between the two antidepressant classes, further reducing any safety or tolerance advantage from using the newer agents [5, 6].

Here we describe an individual patient tracking study, using a national dispensing claims database, to determine the frequency with which the SSRI antidepressants were used as add-on therapy with the tricyclic antidepressants rather than as replacement therapy. The major specialty of the doctors most frequently involved in such combination therapy was also investigated.

Methods

Within the subsidized prescription database maintained by the Health Insurance Commission (HIC) there is a group of patients who are assigned a unique identification number. These are the people who are eligible for social security entitlements and as such receive prescriptions through the Pharmaceutical Benefits Scheme (PBS) at a reduced patient contribution. As each prescription supplied to these patients includes their concession card number, a permanent identification number (PIN) can be assigned that allows a patient prescription profile to be assembled. The PIN is an internal processing number which is used to represent an individual but cannot be de-coded by the Department of Health & Aged Care to identify a particular named patient.

Although patients in the study were over-represented by the elderly and by females compared with the general population, a survey of general practice activity in 1998/99 showed that the age/sex distribution was close to that of patients who were prescribed antidepressants in Australia (Table 1).

Table 1.

Age/sex distribution of the study population relative to the general population (1996) and for an estimate of patients prescribed antidepressants in primary care in Australia (1989/1999)

	Study population 1996	Australian population 1996 a	Patients prescribed antidepressants in Australia 1998/99b
≤ 24 years	1.88%	35.97%	7.81%
25–44 years	30.21%	30.99%	33.78%
45–64 years	32.71%	21.01%	33.86%
65–74 years	13.5%	7.07%	12.34%
≥ 75 years	21.03%	4.96%	12.21%
Age missing	0.67%
Male	33.77%	49.74%	32.98%
Female	65.64%	50.26%	67.02%
Missing	0.59%
Total	55 271	18 310 714	5112

^aAustralian Bureau of Statistics (ABS). Estimated resident population by age groups, 30 June 1996. Australian Demographic Statistics, ABS catalogue 3101.0 September quarter 1997. Canberra, Australia.

^bBritt H, Sayer GP, Miller GC et al. General Practice activity in Australia 1998–99. AIHW Catalogue number GEP 2. Canberra: Australian Institute of Health and Welfare (General Practice Series no. 2).

Persons covered by concession cards are high volume users of prescription medicines in Australia. In 1996 they represented 39% of the population but received 64% of the prescriptions dispensed through community pharmacies. In this group, the low patient copayment allows the capture of all their antidepressant prescription use (other than private prescriptions).

Most of the concession card categories involve long-term recipients of pharmaceutical benefits including aged pensioners, returned war veterans, persons with disabilities and sole parents. For the time frame of this study (6 months), other groups such as the unemployed and those with low incomes (predominantly students) were treated as continuously entitled persons.

Details of all the individuals (PINs) who were supplied a selective serotonin re-uptake inhibitor (SSRI) in January 1996 were extracted from the HIC claims database. Dispensing of SSRI and TCA prescriptions to these patients was then examined over the 6 months period from January to June 1996. The main outcome measure was the proportion of individuals to whom SSRIs and TCAs were dispensed concurrently, as an indirect measure of coprescription. Within the database used for this study, all SSRI and TCA prescriptions dispensed to an individual were sorted in chronological order by date of supply over the 6 month period. Each occasion the patient had a prescription dispensed that was from an antidepressant class different to that supplied immediately prior to this script (SSRI dispensed when the previous prescription was a TCA and vice versa) this was viewed as a change or switch in the class of antidepressant being supplied. Concurrent dispensing was defined by four or more of these changes (from SSRI to TCA and vice versa) and by more than 30 days between the first and last change within this 6 month period.

These criteria were devised following careful consideration of the type of data being examined (prescription data by date of supply) and incorporated reasonable safeguards to exclude patients who were in genuine transition from one antidepressant drug class to another. National guidelines in Australia recommend no period of overlap between these drug classes so the approach taken, in requiring at least four switches between the classes over a period of at least a month, was conservative [7].

Where a patient was receiving prescriptions for SSRIs and TCAs that had been written by the one doctor only, the major specialty of the doctor was determined from an algorithm based on their medical service claims over the study period.

Results

Patients

There were 55 271 individual PINs to whom 63 865 SSRI prescriptions were dispensed in January 1996. These patients would be a mixed group of new starters and patients on established therapy and their prescription use represented just over half (52%) of the total community SSRI prescriptions dispensed in that month (123 334). Those PINs who received an SSRI in January 1996 and also recorded the dispensing of a tricyclic antidepressant within the 6 months January to June 1996 inclusive were then identified (7252 individuals, 61 171 prescription records). This subgroup was then examined against the coprescribing criteria.

The number of PINs meeting the criteria for coprescription of SSRIs and TCAs (four or more switches between these antidepressants and greater than 30 days between the first and last switch) was 2773 (5.02% of the total SSRI group).

The percentage of patients being coprescribed SSRIs and TCAs (Table 2) was statistically different between the States (P < 0.0001) with coprescribing being 55% more common in Queensland than in Victoria. The difference is unexplained by the doctor mix as both psychiatrist and general practitioner services per head of population in Queensland were slightly less than that of the national average in 1996 (GP services per head of population: Qld 4.79, Aust 4.83; psychiatrist services per head of population: Qld 0.10, Aust 0.12) [8].

Table 2.

The number of PINS supplied an SSRI in January 1996 and of these the number satisfying the TCA/SSRI coprescribing criteria in the period January to June 1996, by State

State	PINs supplied an SSRI in January 1996	PINs coprescribed SSRIs and TCAs
New South Wales/Australian Capital Territory	18291	852 (4.66%)
Victoria	14483	624 (4.31%)
Queensland	10091	677 (6.71%)
South Australia/Northern Territory	4800	220 (4.58%)
Western Australia/Tasmania	7606	400 (5.25%)
Total	55271	2773 (5.02%)

The percentage of the Queensland population covered by entitlement cards (35%) was similar to that seen in Australia overall (34%).

Drugs

Fluoxetine, sertraline and paroxetine were the SSRIs most commonly involved in coprescribing, while dothiepin, doxepin and amitripyline were the most common tricyclic antidepressants. The majority of coprescribing instances (84%) involved only one SSRI and one TCA, although some involved differing strengths of these drugs.

The median number of TCA and SSRI prescriptions dispensed to the coprescribed patients within the 6 month study period was 11 (first quartile (Q1) 9 prescriptions; third quartile (Q3) 14 prescriptions) and the median number of switches between SSRIs and TCAs or vice versa was seven (Q1 5 switches; Q3 9 switches). The median number of days between first and last switch was 128 days (Q1 94 days; Q3 147 days).

Prescribers

For 1133 (40%) of the patients coprescribed TCAs and SSRIs, only one doctor was involved in the prescribing, while 284 (10%) had the SSRI and the TCA prescriptions prescribed separately by different doctors, i.e. there was no coprescribing doctor. The remaining 1356 patients (50%) had multiple doctors prescribing the SSRI or TCA or both in the 6 months study period. Forty-seven (1.7%) of the patients who were coprescribed SSRIs and TCAs presented prescriptions from more than five different doctors. The number of prescription records that had no prescriber identification was 871 (1.4%).

In coprescribing situations where only the one doctor was responsible for prescribing all the SSRI and TCA prescriptions for an individual patient, the specialty of the doctor (general practitioner, psychiatrist and other) was determined. Where multiple doctors were involved, the individual practitioners may not have been coprescribing, as they may have been changing merely from one drug group to the other.

Of the single doctor instances, 56 doctors (13 general practitioners, 1 other nonspecialist and 42 psychiatrists) were involved in coprescribing SSRIs and TCAs to three or more patients.

The top 10 coprescribing doctors had psychiatry major specialty codes. However, when the frequency of these coprescribing instances was expressed as a proportion of the concession card patients to whom the doctor had prescribed SSRIs in the month of January 1996, the top 10 coprescribers were 6 vocationally registered GPs, 1 other nonspecialist and 3 psychiatrists.

Pharmacies

Most of the PINs (1607–58%) used a pharmacy in the same postcode area for all their SSRI and TCA prescriptions, while 78 (2.8%) had their prescriptions dispensed at pharmacies in five or more different postcode areas.

The State analyses were based on the leading digit of the PIN that is a State identifier, the majority of which were assigned in 1984. To describe the extent of relocation to other states that may have occurred between then and 1996, the postcode of the primary dispensing pharmacy used by the patient in the period January to June 1996 was examined. There were 272 PINs (9.8%) using a primary pharmacy postcode from a State different from that indicated on their identification number. Around half (51%) of these changes indicated a transfer to Queensland postcodes from the original state, followed by NSW (20%) and WA/Tas (13%).

Although the PIN separately identifies the primary card holder from any dependents also covered, the PBS safety net that reduces the patient copayment after a specified number of prescriptions have been dispensed is a family based system. A sensitivity analysis was conducted to exclude the possibility of contamination of the primary card holder's record with the prescription use of dependents by removing those PINs who had reached the safety net threshold within the 6 months study period. The number of PINs meeting the criteria for being coprescribed SSRIs & TCAs fell only marginally from 2773 (5.02%) to 2421 (4.54%) and did not alter the level of statistical significance of the difference in coprescribing between the States.

Discussion

There is little evidence that any antidepressant drug group, including the SSRIs, has greater therapeutic efficacy than the others [1, 9]. The clinical advantages used to justify the higher cost of the SSRIs compared with the TCAs are the lower incidence of adverse effects (particularly sedation and anticholinergic effects) and the greater safety in overdose, both of which are based on the greater receptor selectivity of the SSRIs [10]. Despite this, the present study has revealed that about 5% of SSRI users in Australia were also being prescribed a TCA, the nonselective drugs that the SSRIs were designed to replace. In such a situation the claimed advantages of the new antidepressants are lost, and the different adverse effects of the SSRIs are added to those of the TCAs. In addition there is a well-documented adverse drug interaction between the two groups of drugs. Some, but not all, of the SSRIs are potent inhibitors of cytochrome P4502D6, the major enzyme involved in metabolism and clearance of the TCAs [6]. This inhibitory metabolic interaction results in elevated plasma concentrations of the TCAs sometimes causing serious toxicity.

In some cases the coprescribing of the SSRIs and TCAs may not be intentional. Ten percent of patients (284) coprescribed antidepressants always received the SSRI from one doctor and the TCA from another. It is likely that these coprescribing episodes were inadvertent and that this number was an underestimate as many more patients had multiple doctors.

Unlike the capitation system in the UK there is no limit to the number of doctors that patients in Australia may attend, a situation that increases the likelihood of two (or more) different doctors unwittingly prescribing these different drugs to the same patient. This raises a major problem of doctor-doctor and doctor-patient communication and reinforces the need to take a thorough medication history before adding another drug.

However, in 40% of patients (1133) only one doctor was responsible for prescribing both the SSRIs and the TCAs, suggesting that coprescribing was an intentional therapeutic strategy in these cases. This is supported by the finding that these doctors were more likely to be psychiatrists.

Several approaches involving combination therapy have been proposed for the treatment of resistant depression. Although there are reports of the value of the combination of SSRIs and TCAs in patients with treatment resistant depression [11, 12], the best studied augmentation strategy involves the addition of lithium to existing antidepressants including SSRIs [13, 14]. More recently pindolol has been reported as enhancing the effect of SSRIs in those with treatment resistance [15].

There were anecdotal data from the State HIC centres that some coprescribing GPs reported using the combination to overcome the sleep disturbance caused by the SSRIs. Others thought that the combination was acceptable as long as the patient was monitored and the doses were in the low to middle range for each drug (personal communication: Leone Coper, formerly Health Insurance Commission).

There are reports of SSRIs being coprescribed with other antidepressants such as the use of low dose trazodone in the US to treat insomnia [16]. Others, however, harbour concerns about the potential for serotonin syndrome with this combination [17].

The problems involved with using a SSRI/TCA combination would need to be balanced by good evidence of enhanced efficacy to justify combined use of the two drug groups. Similarly, good evidence of improved clinical outcomes would be needed to justify the substantially greater cost incurred by adding an SSRI to a TCA.

There are no randomised controlled trial data evaluating the effect of combination SSRI/TCA treatment for patients who are unresponsive to first-line antidepressant medication. The discussion in the literature has been based on a few case studies and case series [11, 12]. Therefore, the evidence for benefit from such a combination is weak while there are well-documented adverse outcomes in some patients [18]. Our current study indicates that, despite the weak evidence base, tricyclic augmentation of SSRIs has become a surprisingly frequent approach among at least some psychiatrists and some general practitioners.

Recent evidence-based guidelines on the treatment of depression, such as those of the American Psychiatric Association [19], highlight the potential complications of the SSRI/TCA combination, and suggest options for those with resistant depression such as changing to a different antidepressant, or augmenting with lithium or thyroid hormone. For those instances where the coprescription may have been inadvertent, Australia is in the process of establishing an information management system that will offer health professionals, with the appropriate consent, an accurate and up-to-date medication history for their patients [20].

Finally, nearly 99% of SSRI prescriptions dispensed through community pharmacies are subsidized via the Pharmaceutical Benefits Scheme or the Repatriation PBS. The SSRIs were included on the subsidy list in Australia on the basis of reasonable cost effectiveness as monotherapy compared with the TCAs. Use in combination therapy is of uncertain clinical and cost effectiveness. Similar issues relating to leakage outside indications where cost effectiveness has been established occur with other drug groups [21].