Slow vs standard up-titration of paroxetine in the treatment of panic disorder: A prospective randomized trial

Authors

  • Massimiliano Buoli MD,

    Corresponding author
    1. Department of Psychiatry, University of Milan, Fondazione IRCCS Ca'Granda Ospedale Maggiore Policlinico
      Massimiliano Buoli, MD, Department of Psychiatry, University of Milan, Fondazione IRCCS Ospedale Maggiore Policlinico, Via F. Sforza 35, 20122 Milan, Italy. Email: massimiliano.buoli@hotmail.it
    Search for more papers by this author
  • Bernardo Dell'Osso MD,

    1. Department of Psychiatry, University of Milan, Fondazione IRCCS Ca'Granda Ospedale Maggiore Policlinico
    Search for more papers by this author
  • Monica Francesca Bosi MD,

    1. Department of Clinical Sciences, Luigi Sacco Hospital, Milan, Italy
    Search for more papers by this author
  • Carlo Altamura MD

    1. Department of Psychiatry, University of Milan, Fondazione IRCCS Ca'Granda Ospedale Maggiore Policlinico
    Search for more papers by this author

Massimiliano Buoli, MD, Department of Psychiatry, University of Milan, Fondazione IRCCS Ospedale Maggiore Policlinico, Via F. Sforza 35, 20122 Milan, Italy. Email: massimiliano.buoli@hotmail.it

Abstract

Aim:  Patients with panic disorder (PD) might be sensitive to the stimulating effects of selective serotonin reuptake inhibitors (SSRI), thus requiring low dosages at treatment initiation. The aim of the present study was to assess eventual differences in terms of effectiveness and tolerability between a slow up-titration with paroxetine and a standard one.

Methods:  In an open randomized, multicenter, primary-care study, 60 patients (44 women and 16 men) with PD with or without agoraphobia were enrolled and randomized to receive a slow up-titration with paroxetine (increments of 2.5 mg/day every 2 days) or a standard one (increments of 10 mg/day every week) up to a maximum daily dose of 20 mg. Repeated-measures anova on sub-items scores of the Panic Attack Anticipatory Anxiety Scale (PAAS) and Dosage Record and Treatment Emergent Symptom Scale (DOTES), respectively, used as outcome measures of effectiveness and tolerability, were performed. Significance level was set at 0.05 and it was not corrected.

Results: anova showed no differences between the two treatments in terms of effectiveness and tolerability. Post hoc analysis found only one significant difference in the intensity of spontaneous panic attacks (Panic and Anticipatory Anxiety Scale) in the first 9 days of treatment between the two treatment groups, which was that this item was less intense in the slow-titration group (treatment effect: F = 4.89, P = 0.03, effect size = 0.1).

Conclusion:  Present findings suggest only a small superiority for a slow up-titration regimen of paroxetine compared to a standard one in the first 9 days of treatment but no differences at end-point.

PANIC DISORDER (PD) is a disabling condition exerting a negative impact on social, family1 and professional life2 of affected patients. PD is common in the general population with a 1-year prevalence of 2.7%3 and a lifetime prevalence of 4.7%.4 Subjects with PD, particularly those with comorbid depression,5 may show an increased risk for suicide attempts6 and substance abuse.7

Main treatments of PD include cognitive behavioral therapy (CBT) and pharmacotherapy8 with antidepressants and benzodiazepines.9,10 After being considered for many years, the first-line therapeutic option for anxiety disorders,11 benzodiazepines, are no longer regarded as the gold standard for the treatment of these conditions, given that long-term treatment with these compounds may cause physical addiction and withdrawal symptoms when abruptly discontinued.12 On the other hand, selective serotonin reuptake inhibitors (SSRI) have shown a similar efficacy in the treatment of anxiety disorders and a better tolerability compared to tricyclic antidepressants and monoamine oxidase inhibitors.13 Therefore, current guidelines consider SSRI as first-line treatment for PD due to their efficacy over anxiety symptoms and their favorable safety profile.14 According to major international treatment guidelines, SSRI treatment in PD should last for 18–24 months.9,15 Paroxetine, moreover, represents the first SSRI approved for the treatment of PD16 and its efficacy has been shown both in short- and long-term clinical trials.17,18

In spite of the efficacy of SSRI, the rate of recurrence for anxiety disorders is 27–39% of cases19,20 and early interruption of therapy is one of the major contributors to this high recurrence rate.21 As a matter of fact, poor adherence to the treatment regimen is quite frequent in psychiatric disorders, occurring in up to 60% of patients by the sixth month of therapy22 and seems to depend on the effectiveness and tolerability of the administered drug. Lin and colleagues showed that up to 60% of patients with mood and anxiety disorders discontinue antidepressant treatment due to adverse events.23 Specifically, adverse events occurring with a frequency higher than 5% during paroxetine treatment include nausea, sexual dysfunctions, sleepiness, asthenia, headache, constipation, dizziness, sweating, tremor and loss of appetite.23 Therefore, it may be assumed that efficacy of different SSRI, including paroxetine, is closely related to patients' compliance. In fact, early discontinuation of therapy and poor adherence to treatment contribute to treatment failure.21 Specifically, in patients with PD there are several possible causes of early discontinuation of antidepressant therapy, including lack of efficacy, adverse events and the so-called ‘jitteriness syndrome’, characterized by exacerbation of anxiety symptoms during the initial phase of treatment.24 In order to minimize this phenomenon, current guidelines for PD15 recommend that the starting dose for paroxetine should be 10 mg/day, which should be maintained for approximately 3–7 days, then gradually increased (e.g. in weekly increments) to a higher daily dose, adjusting the titration schedule according to the patient's ability to tolerate it.25 Recently, an oral liquid formulation of paroxetine was approved in many European countries by the European Medicines Agency (EMEA). This formulation allows an up-titration with a starting dose and subsequent dose increases lower than traditionally recommended. A slow titration could be advantageous to avoid the initial side-effects due to the rapid occupation of post-synaptic serotonin receptors with a molecule like paroxetine that is 2–23 times more potent on serotonin reuptake inhibition than other SSRI.17

The aim of the present study was to evaluate the effectiveness and tolerability of paroxetine during the phase of up-titration in patients with PD, by comparing two different dosage regimens as well as to investigate whether a slower titration might show a more favorable impact in the first weeks of treatment.

METHODS

An open randomized multicenter comparative study of two start-up therapeutic regimens of paroxetine (slow vs standard titration) was conducted in a primary care setting with a sample of 60 outpatients with a DSM-IV diagnosis of PD (44 women and 16 men) who were selected from those attending 14 community primary care services in Milan, Italy.

Subjects: had to be aged at least 18 years; had to present a diagnosis of PD with or without agoraphobia in accordance with the DSM-IV criteria;26 and had to have had at least two panic attacks in the last 2 weeks before the screening visit. Study exclusion criteria are summarized in Table 1. The protocol was approved by the ethics committee of Luigi Sacco Hospital, Milan. All patients provided written informed consent for participation in the study. Diagnoses were assessed through the administration of a semi-structured interview based on DSM-IV criteria (SCID-I) administered by the community care doctors with the assistance of trained psychiatrists.27 Patients meeting criteria for being included in the study were recruited by a general practitioner and were then randomized by psychiatrists through a computerized 1:1 random-number generator to start a slow or a standard titration treatment. Paroxetine was administered to patients randomized to the slow-titration group through an oral drop solution at a dose of 2.5 mg daily for 2 days (5 drops; 2 drops = 1 mg); then the dose was progressively increased by 2.5 mg daily, every other day, up to 20 mg daily (achieved on the 15th day of treatment). Patients in the standard-titration group received paroxetine drops at a starting dose of 10 mg daily for the first week, and then 20 mg daily from the 8th day of treatment until the end of the up-titration phase, which lasted 18 days and included three visits on days 0, 9 and 18.

Table 1.  Exclusion criteria for participating in the study
 (1) A current comorbid psychiatric condition
 (2) A previous psychotic episode
 (3) A history of epilepsy
 (4) Substance abuse in the last 6 months
 (5) Fluoxetine use in the last 5 weeks and other antidepressants and benzodiazepines administration in the last 2 weeks preceding the randomization visit
 (6) Concomitant therapy with drugs that inhibit the hepatic enzyme CYP450 2D6
 (7) Concomitant therapy with full dosage oral anticoagulants (to reach international normalized ratio >2.5)
 (8) Treatment with other investigational drugs or patient inclusion in other clinical trials
 (9) Hypersensitivity to paroxetine or any of the excipients
(10) Bleeding diathesis
(11) Renal failure (creatine >1.5 mg/dL)
(12) Known active gastric or duodenal ulcer
(13) Severe liver insufficiency (clinical history and lab)
(14) Investigator's lack of confidence on the patient compliance to follow the study requirements in terms of visits/compliance to treatment.
(15) Pregnancy and breastfeeding.

Serial efficacy, safety and tolerability assessments were performed at baseline (visit 1), after 9 (visit 2) and 18 days (visit 3) of treatment. Outcome measures were represented by sub-item scores of the Panic Attack Anticipatory Anxiety Scale (PAAS)28 and scores of global judgment of the Dosage Record and Treatment Emergent Symptom Scale (DOTES).29 These scales are interviewer-administered scales and the first evaluates intensity, number and duration in minutes of the panic attacks, while the second was used to assess tolerability.

Statistical analysis

The main demographic and clinical variables (age, sex and duration of illness) of the sample were analyzed and compared between the two sub-groups of patients using χ2-tests for dichotomous variables and Student's t-tests for independent samples for continuous ones. Repeated-measures anova on sub-items total scores of PAAS and DOTES (global judgment) were performed to evaluate eventual differences in terms of efficacy and tolerability between the treatment groups. Post hoc analysis between baseline–time 1 (9 days) and time 1–end-point (18 days) were also carried out. Significance level was set at 0.05 and it was not corrected.

RESULTS

Thirty-one patients were randomized to slow titration and 29 to standard titration. In the slow-titration group five patients did not complete the trial: three for protocol violation, one for lack of compliance and one due to side-effects (constipation) (rate of discontinuation 16.1%). In the standard-titration group three patients did not complete the trial: two for protocol violation and one due to side-effects (anxiety) (rate of discontinuation 10.3%). The rate of discontinuation of the total sample was 13.3%. Among the patients that concluded the protocol, two patients developed side-effects in the slow-titration group (one dizziness and one xerostomia) and two patients developed side-effects in the standard-titration group (one constipation and one tachycardia).

The two groups were not different at baseline in relation to gender distribution (χ2 = 0.02, d.f. = 1, P = 1.00), age (t = −0.46, P = 0.65), duration of illness (t = −0.44, P = 0.66) or severity of symptoms, which included: number of situational panic attacks per week (t = −0.30, P = 0.77) and per month (t = −0.62, P = 0.54), duration of situational panic attacks in minutes (t = 0.88, P = 0.39), intensity of situational panic attacks (t = −0.85, P = 0.40), number of spontaneous panic attacks per week (t = −0.63, P = 0.53) and per month (t = −0.42, P = 0.68), duration of spontaneous panic attacks in minutes (t = 0.91, P = 0.37) and intensity of spontaneous panic attacks (t = −1.28, P = 0.21) (Table 2).

Table 2.  Clinical and demographic characteristics of the sample subgroups at baseline
VariableSlow titrationStandard titrationP-value
n = 31n = 29
  1. P-values are referred to Student's t-tests for independent samples for continuous variables or χ2-test for dichotomous ones.

  2. Standard deviations of continuous variables are presented in parentheses.

Age50.65 (±18.58)52.69 (±15.30)0.65
Duration of illness (months)48.42 (±19.34)50.45 (±16.37)0.66
GenderMale8 (25.8%)8 (27.6%)1.00
Female23 (74.2%)21 (72.4%)
Spontaneous panic attacksDuration (min)17.86 (±33.88)11.50 (±15.91)0.37
Intensity4.83 (±3.13)5.93 (±3.39)0.21
Frequency (week)1.61 (±2.08)2.1 (±3.76)0.53
Frequency (month)4.71 (±6.54)5.41 (±6.54)0.68
Situational panic attacksDuration (min)17.45 (±38.36)10.67 (±12.48)0.39
Intensity5.21 (±3.05)5.89 (±2.81)0.40
Frequency (week)2.68 (±2.74)2.45 (±3.21)0.77
Frequency (month)7.94 (±9.21)9.76 (±13.19)0.54

Repeated-measures anova showed no differences between treatment groups neither on scores of sub-items of PAAS nor on scores of item global judgment of DOTES (Tables 3,4). However, post hoc analysis showed a significant difference in the intensity of spontaneous panic attacks (PAAS), which were less intense in the slow-titration group, in particular in the first 9 days of treatment (time effect: F = 27.72, P < 0.001, effect size = 0.40; time × treatment effect: F = 0.15, P = 0.86, effect size = 0.004; treatment effect: F = 3.81, P = 0.06, effect size = 0.08) (post hoc: Visit 1 vs Visit 2: time effect: F = 28.24, P < 0.001, effect size = 0.39; time × treatment effect: F = 0.02, P = 0.9, effect size = 0.001; treatment effect: F = 4.89, P = 0.03, effect size = 0.1; Visit 2 vs Visit 3: time effect: F = 4.81, P = 0.03, effect size = 0.1; time × treatment effect: F = 0.32, P = 0.58, effect size = 0.01; treatment effect: F = 2.47, P = 0.12, effect size = 0.05; Visit 1 vs Visit 3: time effect: F = 53.70, P < 0.001, effect size = 0.56; time × treatment effect: F = 0.52, P = 0.48, effect size: 0.01; treatment effect: F = 3.18, P = 0.08, effect size 0.07) (Fig. 1). No further significant differences between groups in the scores of sub-items of PAAS and DOTES were found in the post hoc analysis.

Table 3.  Mean scores of sub-items of PAAS and DOTES in the two treatment groups at Visit 1, Visit 2 and Visit 3
PAAS Visit 1 (n = 60)Visit 2 (n = 54)Visit 3 (n = 52)
  1. Standard deviations are presented in parentheses.

  2. DOTES, Dosage Record and Treatment Emergent Symptom Scale; PAAS, Panic Attack Anticipatory Anxiety Scale; SL, slow-titration group; ST, standard-titration group.

Spontaneous panic attacksDuration (min)SL17.86 (±33.88)3.65 (±7.03)2.36 (±6.38)
ST11.50 (±15.91)6.00 (±13.11)4.85 (±13.75)
IntensitySL4.83 (±3.13)2.38 (±2.73)1.56 (±2.90)
ST5.93 (±3.39)4.14 (±3.26)2.38 (±3.28)
Frequency (week)SL1.61 (±2.08)0.52 (±1.12)0.19 (±0.57)
ST2.10 (±3.76)0.74 (±1.95)0.23 (±0.65)
Frequency (month)SL4.71 (±6.54)1.89 (±2.91)1.00 (±2.47)
ST5.41 (±6.54)1.96 (±2.67)0.58 (±1.14)
Situational panic attacksDuration (min)SL17.45 (±38.36)4.42 (±4.73)1.48 (±3.15)
ST10.67 (±12.48)6.38 (±13.07)5.67 (±13.57)
IntensitySL5.21 (±3.05)2.42 (±2.30)1.33 (±2.48)
ST5.89 (±2.81)3.08 (±2.91)2.00 (±3.07)
Frequency (week)SL2.68(±2.74)0.85 (±1.56)0.62 (±1.63)
ST2.45 (±3.21)0.44 (±0.89)0.31 (±0.74)
Frequency (month)SL7.94 (±9.21)3.11 (±4.89)1.96 (±4.75)
ST9.76 (±13.19)2.59 (±3.94)0.88 (±2.05)
Sub-threshold spontaneous panic attacksDuration (min)SL8.73 (±13.92)3.93 (±7.27)1.92 (±4.41)
ST7.27 (±12.55)5.65 (±13.45)1.33 (±2.54)
IntensitySL4.80 (±3.36)2.11 (±2.61)0.14 (±0.13)
ST4.65 (±3.16)2.91 (±3.10)0.13 (±0.14)
Frequency (week)SL1.87 (±2.90)0.56 (±0.75)0.31 (±0.62)
ST2.00 (±2.65)0.96 (±2.77)0.31 (±0.84)
Frequency (month)SL6.40 (±10.71)1.74 (±2.73)1.27 (±2.18)
ST5.79 (±8.33)3.37 (±6.93)1.88 (±6.13)
Sub-threshold situational panic attacksDuration (min)SL9.21 (±12.34)4.37 (±6.42)2.15 (±6.05)
ST7.32 (±7.42)7.56 (±13.26)2.70 (±6.88)
IntensitySL5.72 (±2.60)3.00 (±2.60)1.32 (±2.30)
ST5.80 (±2.75)3.79 (±3.18)2.10 (±2.88)
Frequency (week)SL2.97 (±3.48)1.70 (±4.28)0.46 (±1.10)
ST2.90 (±3.80)1.52 (±3.01)0.58 (±1.53)
Frequency (month)SL9.50 (±11.48)3.07 (±4.29)1.92 (±3.63)
ST8.93 (±3.53)4.89 (±7.47)2.35 (±6.49)
Anticipatory anxiety% TimeSL22.42 (±21.86)9.44 (±12.04)5.58 (±8.04)
ST25.34 (±25.55)10.56 (±14.10)3.73 (±6.69)
IntensitySL4.97 (±2.90)2.77 (±2.44)1.76 (±2.39)
ST4.75 (±2.69)2.63 (±2.63)1.65 (±2.78)
DOTES     
 Global judgmentSeveritySL 0.64 (±0.86)0.67 (±0.82)
ST 0.76 (±1.09)0.71 (±1.04)
DiscomfortSL 0.76 (±0.97)0.92 (±0.97)
ST 0.83 (±1.13)0.71 (±1.08)
Table 4.  Results from repeated-measures anova on mean scores of sub-items of PAAS and DOTES between the treatment groups
PAASTreatment effect (F)P-value
  1. DOTES, Dosage Record and Treatment Emergent Symptom Scale; PAAS, Panic Attack Anticipatory Anxiety Scale.

Spontaneous panic attacksDuration (min)0.370.55
Intensity3.810.06
Frequency (week)0.400.53
Frequency (month)0.450.83
Situational panic attacksDuration (min)<0.010.97
Intensity0.650.43
Frequency (week)0.730.40
Frequency (month)<0.010.99
Sub-threshold spontaneous panic attacksDuration (min)0.050.82
Intensity<0.010.96
Frequency (week)0.420.52
Frequency (month)0.350.56
Sub-threshold situational panic attacksDuration (min)0.020.88
Intensity0.030.88
Frequency (week)0.060.81
Frequency (month)0.020.89
Anticipatory anxiety% Time0.010.93
Intensity1.130.30
DOTES   
 Global judgmentSeverity<0.010.96
Discomfort0.080.78
Figure 1.

The intensity of spontaneous panic attacks (PAAS) during the up-titration phase. Statistics (three-time anova): time effect: F = 27.72, P < 0.001, effect size = 0.40; time × treatment effect: F = 0.15, P = 0.86, effect size = 0.004; treatment effect: F = 3.81, P = 0.06, effect size = 0.08. (inline image) SL, slow-titration group; (inline image) ST, standard-titration group.

DISCUSSION

To the best of our knowledge, the current study is the first to compare two different modalities of up-titration using paroxetine drops. Study results showed no differences at end-point in the two treatment groups in terms of effectiveness and tolerability. The only item in which the slow titration was superior to the standard titration was that of spontaneous panic attacks in the first 9 days of treatment. This result might be a false positive due to multiple comparisons without adjustment and to the use of a completers analysis. However, this specific type of panic attack is very disabling for patients as they cannot be prevented through avoidant behaviour30 and are often accompanied by significant respiratory and autonomic abnormalities.31 This finding may have important clinical and practical implications as the current literature points out that a reduced severity of spontaneous panic attacks may increase treatment adherence.32 In light of these considerations the slow titration would be perhaps preferred in terms of major efficacy and compliance.

The rate of treatment discontinuation of the total sample was 13.3%, but few patients in our sample dropped out during the trial due to side-effects (n = 2) confirming the overall good tolerability of paroxetine treatment, especially in the initiation phase.33 No differences were found between treatments on DOTES scores so that we can conclude that in the present study standard and slow titration were equivalent in terms of tolerability.

From a neurobiological point of view, a greater intensity of spontaneous panic attacks in patients receiving a standard schedule might be explained by an increased serotonin neurotransmission at the postsynaptic receptors resulting from a rapid and massive serotonin reuptake inhibition.34 Moreover, results from different studies indicate that the pathogenesis of PD includes at least two identifiable components: a biological component represented by spontaneous (unexpected) panic attacks, which is supposed to be more improved by antidepressants, and a cognitive component represented by situational attacks.35,36

This study may have some limitations: in particular the possible presence of a type I error as well as the open design, the relatively small sample size and the lack of a placebo arm. The choice to conduct the study in a primary care setting was due to the high prevalence of PD among patients of general practitioners as reported in a recent survey.37 In fact, patients with panic attacks or PD frequently present to outpatient primary care settings reporting the somatic manifestations of their panic attacks.38 Further double-blind studies with larger samples are warranted to confirm present data and to optimize and develop novel approaches that could expand the array of treatment options.

ACKNOWLEDGMENTS

This trial was partially supported by Italfarmaco S.p.A., Via dei Lavoratori, 54, 20092, Milan, Italy.

Ancillary