European Heart Journal (1991) 12, 1266-1272
Attenuation of nitrate effect during an intermittent treatment regimen and the time course of nitrate tolerance M. A. JAMES, M. PAPOUCHADO AND J. V. JONES
Cardiology department, Bristol Royal Infirmary, Bristol BS2 8HW, U.K. KEY WORDS: Transdermal nitrate, tolerance, nitrate-free periods. The long-term efficacy of transdermal nitrate therapy, in particular the ability of a single patch to provide 24 h prophylaxis against angina, has been questioned. Two mechanisms have been suggested for this loss of effect: the development of pharmacological tolerance, and premature patch exhaustion. This study was designed to investigate this problem, and in particular to investigate the time course of treatment failure. It comprised a randomized, double-blind, cross-over comparison oftransdermal glyceryI trinitrate and matching placebo transdermal patches. Significant treatment effects were demonstrated by several criteriafor 8 h ofcontinuous therapy, with some limitedeffect persisting for 15 h. Loss of effect began to develop very soon after treatment was initiated and progressed in a steady, linear fashion so that there was virtually no treatment effect after 24 h.In contrast, during intermittent therapy, treatment effects were maintained on the second dayfollowing a nitrate-free interval. Significant benefit was demonstrated for up to 32 h (i.e. 8 h of treatment on day 2). Both nitrate-free intervals (12 and 16 h) seemed to be equally effective in maintaining efficacy after 3 h of treatment on the second day, although this was still somewhat attenuated compared with day 1. These results confirm that loss of therapeutic efficacy of transdermal nitrate is due to the development of tolerance and not premature patch exhaustion. In contrast to previous studies, however, they suggest that tolerance can only partly be reversed by intermittent therapy andalso that the onset of tolerance is so rapid that it is well established in less than a day's treatment. Introduction
There has been much debate concerning the efficacy of transdermal nitrate therapy, particularly its ability to provide 24-h prophylaxis against angina. Thus whilst some authors have been able to demonstrate efficacy for over 24 h and during continuous therapy11'2', we found no effect at 24 h'3' and neither have the majority of studies'4"6'. Pharmacological tolerance to constant blood levels has been suggested as the mechanism underlying the observed loss of therapeutic effect and has been demonstrated in all forms of nitrate therapy'7-8'. The degree of stability of the blood levels has been suggested as the prime determinant of the rapidity with which tolerance develops18-9'. Tolerance, which is now a generally accepted phenomenon'9-''', can thus be anticipated during transdermal therapy which is designed to maintain constant blood levels. If tolerance is related to the stability of nitrate blood levels it should be possible to prevent it by varying these and attention has turned to the question of whether nitrate-free periods can prevent the development of tolerance'12"14'. In practice the transdermal preparations would remain of considerable value if they were able to provide day long prophylaxis against angina with a break in treatment simply necessary overnight whilst the patient is asleep. Several studies have now suggested that this approach can be effective'12"14' in maintaining an immediate therapeutic response shortly after patch application, but the duration of effect and whether the patch can provide persistent day-long prophylaxis against angina has not been established. Submitted for publication on 3 August 1990, and in revised form 22 October 1990. Correspondence: Dr M. A. James, Cardiology Department, Bristol Royal Infirmary, Bristol BS2 8HW, U.K. 0195-668X/91 /121266 + 07 $03.00/0
It is, of course, by no means certain that tolerance is the sole cause of the loss of therapeutic effect. Although blood levels have been shown to remain stable for over 24 h'15'16', it is possible that this is true only during ideal test conditions. It has been suggested that under certain circumstances, when blood flow through the skin may be varying, the rate of absorbtion of nitrate may become unpredictable leading to 'pharmacological dumping' and premature patch exhaustion. This has been demonstrated in one study where fluctuating blood levels have been reported during conditions of extreme heat and physical exertion'17'. We have noticed that many patients have complained of marked headache shortly after exercise testing, which could represent a sudden peak in nitrate blood levels as a result of this 'dumping' phenomenon (unpublished observations). This study was therefore designed to establish firstly, how quickly the loss of therapeutic effect develops and how long therapeutic effects can be demonstrated following the application of a single patch; secondly, whether a nitrate-free interval is capable of maintaining day-long therapeutic efficacy on successive days; and finally whether tolerance is the sole cause of this loss of therapeutic effect or partly a reflection of premature patch exhaustion. Methods
Twelve male patients with chronic stable angina were studied (mean length of symptoms 62-7 months; range 12-180). Their mean age was 57-5 years (range 46-70). No patient was allowed concomitant vasodilator therapy but 10 were receiving additional beta-blocking therapy. © 1991 The European Society of Cardiology
A ttenuation of nitrate effect 1267
Doyl Patch 2
Patch 1
Patch 3
'k :B:
i
YA
t
t :
*
l2OOh Time of exercise test (3h)
l7OOh 2100 h (8h) (I2h)
l200h (3= 27h)
Sequence
A + B C+D
All
All
I
i
l700h 2IOOh (8= (12= 32W 36h) A+B
C+D
Figure I Diagramatic plan of study protocol. Treatment was administered in four treatment sequences (A,B,C,D) each lasting at least 32 h (dependent on the length of nitrate-free interval) and allowing exercise testing to be performed on two consecutive days. During the 5 days between each treatment sequence no vasodilator therapy was given. This allowed ample time for wash out of effects between successive treatment sequences. In each treatment sequence three active or dummy nitrate patches were applied continuously according to the plan above. Patch numbers 1 and 3 were always applied at 0900 h but patch number 2 was applied at 1700 or 2100 h depending on the sequence. Treatment sequences were randomized and administered double blind. £2 = sequence; §3 = active; Q = placebo.
Patients were not studied for the following reasons: a history of recent myocardial infarction (3 months), unstable angina, heart failure and inability to perform satisfactory exercise tests (for whatever reason). Before entry to the study patients were required to perform two exercise tests, one before and one after sublingual glyceryl trinitrate (GTN), to establish that they were responsive to nitrate therapy. Four patients (out of 16 initially screened) who failed to achieve an improvement of at least 1 min in their total exercise time were not studied. The study comprised a randomised, double-blind, within-patient comparison of active Transiderm-Nitro 10 and matching placebo patches. Treatment was administered in sequences lasting for 32-36 h depending on whether a 12 or a 16 h nitrate-free interval was employed (Fig. 1). Four treatment sequences were tested in four successive weeks with 5 days between each sequence. During each sequence three successive active or dummy patches were consecutively applied in various orders. Exercise testing was performed at 3 h (1200h) and at either 8 h (1700 h) or 12 h (2100 h) after patch application depending on the treatment sequence. The protocol allowed for assessment of continuous active therapy at 3, 8, 12, 15 and 24 h. Patients were given their supply of three patches at the beginning of each sequence; they applied the first patch at 0900 h on study day 1 and attended for an exercise test at noon, returning for a second test later the same day. The second patch was applied after the second exercise test on day 1 (1700 h or 2100 h, depending on treatment sequence) and remained in place until the following morning when the third patch was applied at 0900 h on study day 2. Throughout the study, exercise testing at 3,8 and 12 h was performed at the same time of
day and by the same observer for each patient. A plan of the four study sequences is shown in Fig. 1. Sequence A tested the effect of a 16 h nitrate-free interval and sequence C the effect of a 12 h nitrate-free interval. Sequence B provided most of the placebo data, but sequence D, day 1, provided the 12 h placebo data. Sequence D also provided the data on 15 and 24 h of continuous therapy using two successive active patches, to investigate the possibility of premature patch exhaustion. Exercise testing was performed on a treadmill using the standard Bruce protocol, preceded by the addition of a 3 min 'warm up' stage (2 kph at 5% incline). A modified 12 lead electrocardiogram was recorded throughout exercise. An attempt was made to standardize successive exercise tests by asking patients to assess the severity of their angina on a simple linear scale (1-20) at the end of the first test, and they were asked to attempt to achieve the same 'angina severity score' during each subsequent test. ST segments were measured 60 ms after the J point, in the lead initially demonstrating the greatest amount of ST depression; the same lead was used for all subsequent tests. The following five exercise test parameters were measured: time to 1 mm ST depression, time to angina, total exercise time, maximum ST depression and the amount of ST depression at the maximum common exercise time. Many of these parameters, although commonly employed, are hampered by the 'soft' end points by which they are measured. Thus, total exercise time, time to angina and maximum ST segment depression are entirely dependent upon the patient's exercise performance, and this is influenced by many factors other than angina. ST depression at maximum common exercise time and time to 1 mm ST depression are parameters which are
1268 M. A. James etal.
measured at a fixed time or electrocardiographic end Table I Combined results for the 3 h (1200 h) tests during intermittent therapy (A + C vs B). The results are expressed as the point and are thus uninfluenced by the patient's overall difference between the mean values for the active and placebo tests exercise performance. (The maximum common exercise time is determined by the shortest test performed by each Day 1 Day 2 patient, all other tests exceed this time point and ST depression can thus be measured at this same time point Time to 1 mm ST depression; s + 67 5 + 50 (and same workload) in all tests.) We believe that these P = 0-0004 /> = OO35 two parameters, which we have used previously'31, provide ST depression at MCET, mm -0-45 -0-40 /> = 00001 /> = 0-009 the most reliable measurement end points because of their objective nature. Other authors have subsequently used Maximum ST depression; mm -0-32 -0-22 these variables: time to 1 mm ST depression"418"1 and ST /> = 0-006 /> = 0-043 2 121 depression at a standardized exercise time"- - . Time to onset of angina; s + 13 + 26-5 Statistical analysis was by analysis of variance, a P /> = 038 />=0081 + 24-5 + 24 value of < 0 0 5 was taken as significant. However, our Total exercise time; s P = 0-092 /> = 0-078 statistician advised that, in view of the multiple comparisons made, a lower P value should be taken as significant for individual comparisons in order to maintain an overall MCET = Maximum common exercise time. 5% significance level. Thus a P value of 0-03 was taken as significant for individual comparisons and a value of of marginal significance (Table 2). Thus, during continuous therapy significant treatment effects were seen for 8 h 003-005 was considered of borderline significance. in several parameters; some treatment effect was seen at 12 and 15 h in fewer parameters and by 24 h there was no Results significant treatment effect at all (Table 3). In contrast, Comparisons are made both within the same treat- during intermittent therapy, significant treatment effects ment sequence (day 1 vs day 2) and between treatment were seen for 32 h, that is after 8 h of treatment on day 2 sequences (i.e. sequence A: day 1 vs sequence B: day 1). (Table 2). There was no significant treatment effect after The effects of active treatment, sequence A or C, were 12 h on the second day of intermittent therapy. However, compared with placebo, sequence B (placebo data for the since treatment effects after 12 h on the first day were 12 h exercise tests were obtained from sequence D, day 1; of marginal significance this probably did not reflect an see Fig. 1). Thecombined results of active treatment at 3 h important change. Nonetheless, treatment effects on the (A + C) can be compared with placebo, which by doubling second day, even during intermittent therapy, were not as the number of data points increases the statistical power good as those on the first day. Although the difference of this comparison. The efficacy of treatment on the between day 1 and day 2 of intermittent therapy was not second day in each treatment sequence can be tested by significant, this trend towards poorer results was reflected demonstrating that treatment remains significantly better in much lower P values on the second day (Tables 1 and 2). than placebo (i.e. A vs B). The second day of therapy in Figures 2 and 3 demonstrate the results graphically. sequence D tested the effect of continuous therapy since The graphs show the treatment effect expressed as the sequence D employed the use of two consecutive active mean percentage improvement in exercise test parameters patches, so that the 3 h exercise test on day 2 actually compared with the equivalent placebo test. Figure 2 represented 15 h of continuous treatment. The effect of shows the results for continuous therapy and demon15 h of continuous treatment, compared with placebo, strates that the improvement in exercise performance could then be established by comparison with sequence B: achieved by treatment falls in a steady and consistent way day 2. The effect of 24 h continuous therapy could be throughout the 24 h of continuous therapy. In contrast, compared with sequence D: day 1. The validity of com- during intermittent therapy (Fig. 3) treatment effects are parisons between day 1 and day 2 of any given treatment maintained at similar levels on both days 1 and 2. Howsequence was checked by looking for a difference between ever, the graph also demonstrates that there is a tendency day 1 and day 2 in sequence B (which was a complete for treatment effects on the second day to be slightly placebo sequence). diminished compared with the first day during interThe results showed a highly significant treatment effect mittent therapy and that considerable tolerance develops at 3 h when analysing the combined results (A+C vs B; even during the course of a single day's treatment. Table 1) and also when analysing both comparisons There was a marked diurnal variation in exercise perindependently (A vs B and C vs B/D; Table 2). During formance during this study (Fig. 4(a) and (b)). During intermittent therapy, significant treatment effects were placebo therapy exercise test parameters were signifimaintained after 3 h on the second day of treatment cantly better at 1200 h than at 1700 or 2100 h, whilst those (Table 1), regardless of the length of nitrate free period at 1700 h were better than at 2100 h. (Table 2). However, during continuous therapy, effects at In summary, the results have shown that there is a sig3 h on day 2 (15 h of continuous treatment) were only of nificant effect of continuous therapy with Transiderm marginal significance (Table 3). nitro for over 8 h, however, after 12 h while significant On the first day of treatment significant effects were effects can still be demonstrated, these are in fewer parseen at 8 h; however, treatment effects after 12 h were only ameters and at rather lower significance levels. After 15 h
Attenuation of nitrate effect 1269
Table 2 Results summary during intermittent therapy. The results are expressed as the difference between the mean valuesfor active and placebo tests, AvsB (8 h treatment/16 hfree) and C vs D/B (12 h treatment/12 hfree). Improvement is reflected by an increase in parameters measured in s and by a reduction in parameters measured in mm Day 1 1200h
ST depression at MCET; mm Maximum ST depression; mm Time to onset of angina; s Total exercise time; s
AvsB
CvsD
+ 50 P = 00\
+ 52 P = 0024 -0-27 P = 0\2
+ 37 /> = 0-2 -019 ^ = 0-24
-0-34 /> = 0-008
-Oil ^ = 0-44
+ 004 P = 0-85 + 32 P = 0-22 -2 /> = 0-96
CvsD
AvsB
+ 48 />=0-04 -0-41 /> = 0 007
+ 52 P=0027 -019 />=0-20
+ 50 P = 006 -0-35 P = 004
P = 006 -0-45
0 />=10
/> = 0-06
-004 /> = 0-80 + 60 P=0005
+ 27 /> = 011
P = 004
-010 P = 0-4\ + 42 P = 002 + 38 /> = 0-018
-0-24 P = 006
+ 40
+ 30 P=O19 + 27 P=Q 10
CvsB
AvsB
+ 58 + 92 />=0-006 P< 00001 -0-40 -0-46 /> = 0-002 />=00004 - 0 19
CD2vsDDl 2100h
2100 h
CvsD
P=0l3
1700h
1700h
AvsB Time to 1 mm ST depression; s
Day 2 1200h
+ 40 /> = 0047 + 26 /> = 0-25
+ 50
+ 11
+1
/>=0-46
P = 0-94 + 26 /> = 0-ll
+ 10
/>=0-51
End column: C D2 = sequence C, day 2 test 2 (2100 h); D D1 = sequence D, day 1 test 2 (2100 h). MCET = maximum common exercise time.
Table 3 Results during continuous therapy expressed as the difference between the mean values for active and placebo tests at the same times of day. Resultsfor 3,8 and 12 h tests are obtained from day 1 of treatment during sequences A and C. Results for 15 and 24 h of treatment are obtained from the 3 h (1700 h) and 12 h (2100 It) exercise tests on day 2 of sequence D. As shown before, improvement is reflected by an increase in parameters measured in s and by a reduction in parameters measured in mm Hours
Time to 1 mm ST depression; s ST depression at MCET; mm Maximum ST depression; mm Time to onset of angina; s Total exercise time; s
3
8
12
15
24
+ 58 P = 0-006 -0-40 /> = 0-002
+ 48 f = 004 -0-41 />=0-007
+ 52 /> = 0-027 -019 ^ = 0-20
+ 25 /> = 0-35 -0-22 ^ = 0-22
-3 P = 0-9\ -006 /> = 0-73
-019 /> = 0-13 + 30 P = 0 19 + 27 P = 0\0
-004 p = 0% + 60 /> = 0-005
0 P=\0 + 40 /> = 0-047
-003 /»=0-78 + 40 ^ = 003 + 38 /> = 0-016
+ 006 P = 0-68 + 27
+ 50 ^ = 0-04
+ 26
P= 25
°
P = 0\6
+ 10 /> = 0-50
MCET = maximum common exercise time.
of treatment a significant effect was noted in one parameter, but most parameters showed little difference from pretreatment levels, and after 24 h no discernible effect of treatment could be demonstrated (Table 3). By contrast, during intermittent therapy, significant effects of treatment could be demonstrated after 27 h of treatment, regardless of whether a 12 or 16 h break in therapy had been employed (Table 2) and treatment effects could be demonstrated for up to 32 h (8 h of day 2). However, treatment effects on day 2 did seem to be somewhat attenuated compared with day 1 even during intermittent therapy (Tables 1 and 2).
Discussion Several studies have suggested that the therapeutic effect of transdermal nitrate treatment can be maintained during long-term usage by intermittent or interval therapy. However, the major weakness in most of these studies is that treatment effects have only been assessed for a relatively short time after patch application. Thus it remains unknown whether nitrate patches are capable of providing day-long prophylaxis against angina, or how long a single patch may be expected to be effective. This study, whilst confirming these earlier reports, that significant treatment
1270 M. A. James etal.
30
( a ) 600
500 -
20
E
S
H
400 -
10
300 10
20
30
Time of test (h)
Figure 2 The mean percentage improvement in exercise test parameters during continuous therapy The graph shows the mean percentage improvement in exercise parameters compared with the equivalent placebo tests during continuous treatment. The results are presented for the meaned percentage improvements of all five measured parameters ( ) and also for the two we considered most important ( ).
30
o *(
Day2
Day 1
\ 25 -
F
\
1-4
\
/
20 -
/
\
s.
\
•
V\ \
15 -
\ V
\ \
Figure 4(a) and (b) The diurnal variation in exercise test parameters. The graphs demonstrate the difference in the results of all five exercise test parameters at the three times of day they were tested, (a) A = total exercise time; • = time to angina; • = time to 1 mm ST depression; (b) • = maximum ST depression; D = S T depression at maximum common exercise time.
\
/
V 10
3
8
112
1 20
?7 ±32 30
b 40
Time of test (h)
Figure 3 The mean percentage improvement in exercise test parameters during intermittent therapy The graph demonstrates the mean percentage improvement in the time to 1 mm ST segment depression and the STdepression at maximum common exercise time. Improvement in performance is maintained on day 2 during both intermittent treatment regimens (cf. Fig I). However, the percentage improvement is slightly smaller on day 2 compared with day 1 during both treatment regimens. Note also the deterioration in the improvement measured during the course of a single day • . 8— 16 h schedule; O , 12-12 h schedule.
effects can be maintained for at least 3 h during intermittent therapy, also demonstrates two further important findings. Firstly, tolerance develops so rapidly that although significant treatment effects are maintained for several hours, they become markedly attenuated in less than a day. Secondly, although significant treatment
effects are maintained into the second day of treatment during intermittent therapy, the results are consistently attenuated compared with the initial (first day) response (also reflected in lower P values on the second day). In addition, the results confirm that loss of therapeutic efficacy during treatment is due to the development of tolerance rather than patch exhaustion. Continuous therapy with consecutive fresh patches does not lead to a significant difference from placebo on the second day of treatment even at 3 h, whilst during intermittent therapy significant treatment effects are maintained on the second day. Tolerance to nitrates began to develop very soon after treatment had started and became well developed between 8 and 12 h. Both 12 and 16 h nitrate-free intervals were sufficiently long to prevent the occurrence of tolerance at 3 h after restarting treatment. However, as tolerance was already well established by 12 h of treatment there would
Attenuation of nitrate effect 1271
seem to be no advantage in a treatment regimen employing a shorter nitrate-free interval. Furthermore, the development of tolerance did not appear to be a threshold phenomenon, but a continuous process, beginning probably from the moment the patch was applied (although we of course only began testing from 3 h). Figure 2 shows a graph of the mean percentage change in exercise test parameters during continuous therapy. The results are presented for the mean percentage improvement in all five exercise parameters and also for the two exercise parameters which we consider to be the most reliable (ST depression at maximum common exercise time and time to 1 mm ST segment depression), as previously explained. The graph shows the apparent steady and continuous decline in therapeutic effect throughout the study period. This steady decline in effect is observed regardless of how many parameters are analysed. However, treatment effects are apparently greater when restricting analysis to the two 'important' exercise parameters and this is due to the greater consistency of these measurements. This finding is important because previous studies have not generally evaluated the time course of development of tolerance. Thus, although most studies have demonstrated a significant treatment effect 0-4 h after patch application during intermittent therapy, the conclusions drawn, that this therefore implied day-long efficacy are not justified. In this study, therapeutic efficacy was also demonstrated at 3 h during intermittent therapy. However, whilst significant treatment effects were maintained for 12-15 h these effects were markedly attenuated compared with the 'full' nitrate effect earlier in the day. Only three other studies1'418191 have reported the effect of nitrate patches later in the day during intermittent therapy. Schaer et a/.'141 performed exercise tests at 4 and 8 h and also demonstrated attenuation of effect after 8 h, although the authors made no comment on this finding. Waters et a/.1'81 performed exercise tests after 12 or 16 h of treatment and found almost no significant effects at all, although their data also showed a trend toward better results at 12 h than at 16 h. The most recent study (de Mots et a/.)'"1 demonstrated efficacy up to 12 h on thefirstday of treatment, but only at 4 and 8 h during long-term therapy. Thus, although two of these studies purported to show that intermittent therapy was efficacious, all three studies in fact provide support for the findings of this study, namely, that therapeutic effects are lost rapidly, within the course of a single day, and that intermittent therapy can only maintain therapeutic responsiveness for part of the day. Although results on the second day were not significantly different from results on the first day, there was a suggestion that they were persistently slightly worse on day two. This would suggest that even during intermittent therapy, tolerance may only be partly averted, with persistent attenuation of effect despite the nitrate-free interval. It is possible that this attenuation could be a cumulative phenomenon, but this would require further prospective evaluation. In the longer term, cumulative attenuation would result in gradual loss of all nitrate effect and might lead to a rebound increase in ischaemia and angina. This
could only be evaluated by a longer follow-up period. However, previous long-term studies have not commented on a rebound increase in angina, and have generally shown that the immediate (0-4 h) response is maintained over periods of 1-4 weeks,'1214"1 suggesting that cumulative attenuation may not be a problem. All statistical comparisons in this study were made with placebo tests performed at the same time of day. The graphs in Figs 2 and 3 were made up using values for percentage improvement compared with a placebo test at the same time of day. This method of analysis avoids the problems of the variability in absolute values of the exercise data derived from different variables and at different times of day. Statistical comparisons have not been made on this derived data; however, it is interesting to note that the results of the derived data reflect exactly the same pattern as the results of the statistical comparisons on the actual data given in Tables 1-3. Finally, the marked diurnal variation in exercise test performance observed in this study is important to all studies employing comparative exercise testing. This diurnal variability has been reported previously120'211 and the current findings serve to underline the importance of uniform exercise test times during comparative studies employing exercise testing. We thank Ciba-Geigy Pharmaceuticals who provided the active and placebo nitrate patches and whose department of statistics performed the statistical analyses.
References [1] Rezakovic D, Stalec J. Acute and chronic efficacy of 10 mg nitroglycerin patch in stable angina pectoris. Am J Cardiol 1988;61:52E-58E. [2] Greco R, D'Alterio D, Schiattarella M, Boccia A, Greco L, Marsico F. Efficacy of a new transdermal nitroglycerin patch (Deponit 10) for stable angina pectoris. Am J Cardiol 1988; 61: 44E-51E. [3] James MA, Walker PR, Papouchado M, Wilkinson PR. Efficacy of transdermal glyceryl trinitrate in the treatment of chronic stable angina pectoris. Br Hear J 1985; 53: 631-5. [4] Parker JO, VanKoughnett KA, Fung H-L. Transdermal isosorbide dinitrate in angina pectoris: effect of acute and sustained therapy. Am J Cardiol 1984; 54: 8-13. [5] Abrams J. The brief saga of transdermal nitroglycerin discs: paradise lost. Am J Cardiol 1984; 54: 220-4. [6] Thadani U. The effectiveness of transcutaneous nitrate preparations for angina pectoris. Int J Cardiol 1987; 14:9-14. [7] Packer M, Lee WH, Kessler PD, Gottlieb SS, Medina N, Yushak M. Prevention and reversal of nitrate tolerance in patients with congestive heart failure. New Eng J Med 1987; 317:799-804. [8] Parker JO, Fung H-L, Ruggirello D, Stone JA. Tolerance to isosorbide dinitrate, rate of development and reversal. Circulation 1983; 68: 1074-80. [9] Parker JO. Nitrate tolerance. Am J Cardiol 1985; 56: 28L-31L. [10] Desilvio A, Barlattani MP. How best to use nitrates. Br Med J 1987; 295: 1163-4. [11] Anon. Nitrates: the problem of tolerance. Drug Ther Bulletin 1988; 26: 57-9. [12] Cowan JC, Bourke JP, Reid DS, Julian DS. Prevention of tolerance to nitroglycerin patches by overnight removal. Am J Cardiol 1987; 60: 271-5. [13] Luke R, Sharpe N, Coxon R. Transdermal nitroglycerin in angina pectoris: efficacy of intermittent application. J Am Coll Cardiol 1987; 10:642-6.
1272 M. A. James etal.
[14] Schaer DH, Buff LA, Katz RJ. Sustained anti-anginal efficacy of transdermal nitroglycerin patches using an overnight 10 hour nitrate free interval. Am J Cardiol 1988; 61:46-50. [15] Wolff M, Cordes G, Luckow V. In vitro and In vivo release of nitroglycerin from a new transdermal therapeutic system. Pharm Res 1985; 1: 1-54. [16] Shaw JE. Transdermal therapeutic systems. In: Elsdon Drew RW, Birdwood GFB, eds. Transdermal nitrates in ischaemic heart disease International congress and symposium series, London Royal Society of Medicine. 1983; 59: 33-5. [17] Barkve TF, Langseth-Manrique K, Bredesen JE, Gjesdal K. Increased uptake of transdermal glyceryl trinitrate during physical exercise and during high ambient temperature. Am Heart J 1986; 112:537-41.
[18] Waters DD, Juneau M, Gossard D, Choquette G, Brien M. Limited usefulness of intermittent nitroglycerin patches in stable angina. J Am Coll Cardiol 1989; 13:421-5. [19] DeMots H, Glasser SP and the Transiderm-Nitro study group. Intermittant transdermal nitroglycerine therapy in the treatment of chronic stable angina. J Am Coll Cardiol 1989; 13: 786-93. [20] Jones EF, Joy MD. Parasympathetic blockade with benzhexol hydrochloride in the management of stable angina pectoris. Clin Sci 1987; 72(Suppl 16): 57. [21] Joy M, Pollard CM, Nunan TO. Diurnal variation in exercise responses in angina pectoris Br Heart J 1982; 48: 156-60.
Attenuation of nitrate effect during an intermittent treatment regimen and the time course of nitrate tolerance M. A. JAMES, M. PAPOUCHADO AND J. V. JONES
Cardiology department, Bristol Royal Infirmary, Bristol BS2 8HW, U.K. KEY WORDS: Transdermal nitrate, tolerance, nitrate-free periods. The long-term efficacy of transdermal nitrate therapy, in particular the ability of a single patch to provide 24 h prophylaxis against angina, has been questioned. Two mechanisms have been suggested for this loss of effect: the development of pharmacological tolerance, and premature patch exhaustion. This study was designed to investigate this problem, and in particular to investigate the time course of treatment failure. It comprised a randomized, double-blind, cross-over comparison oftransdermal glyceryI trinitrate and matching placebo transdermal patches. Significant treatment effects were demonstrated by several criteriafor 8 h ofcontinuous therapy, with some limitedeffect persisting for 15 h. Loss of effect began to develop very soon after treatment was initiated and progressed in a steady, linear fashion so that there was virtually no treatment effect after 24 h.In contrast, during intermittent therapy, treatment effects were maintained on the second dayfollowing a nitrate-free interval. Significant benefit was demonstrated for up to 32 h (i.e. 8 h of treatment on day 2). Both nitrate-free intervals (12 and 16 h) seemed to be equally effective in maintaining efficacy after 3 h of treatment on the second day, although this was still somewhat attenuated compared with day 1. These results confirm that loss of therapeutic efficacy of transdermal nitrate is due to the development of tolerance and not premature patch exhaustion. In contrast to previous studies, however, they suggest that tolerance can only partly be reversed by intermittent therapy andalso that the onset of tolerance is so rapid that it is well established in less than a day's treatment. Introduction
There has been much debate concerning the efficacy of transdermal nitrate therapy, particularly its ability to provide 24-h prophylaxis against angina. Thus whilst some authors have been able to demonstrate efficacy for over 24 h and during continuous therapy11'2', we found no effect at 24 h'3' and neither have the majority of studies'4"6'. Pharmacological tolerance to constant blood levels has been suggested as the mechanism underlying the observed loss of therapeutic effect and has been demonstrated in all forms of nitrate therapy'7-8'. The degree of stability of the blood levels has been suggested as the prime determinant of the rapidity with which tolerance develops18-9'. Tolerance, which is now a generally accepted phenomenon'9-''', can thus be anticipated during transdermal therapy which is designed to maintain constant blood levels. If tolerance is related to the stability of nitrate blood levels it should be possible to prevent it by varying these and attention has turned to the question of whether nitrate-free periods can prevent the development of tolerance'12"14'. In practice the transdermal preparations would remain of considerable value if they were able to provide day long prophylaxis against angina with a break in treatment simply necessary overnight whilst the patient is asleep. Several studies have now suggested that this approach can be effective'12"14' in maintaining an immediate therapeutic response shortly after patch application, but the duration of effect and whether the patch can provide persistent day-long prophylaxis against angina has not been established. Submitted for publication on 3 August 1990, and in revised form 22 October 1990. Correspondence: Dr M. A. James, Cardiology Department, Bristol Royal Infirmary, Bristol BS2 8HW, U.K. 0195-668X/91 /121266 + 07 $03.00/0
It is, of course, by no means certain that tolerance is the sole cause of the loss of therapeutic effect. Although blood levels have been shown to remain stable for over 24 h'15'16', it is possible that this is true only during ideal test conditions. It has been suggested that under certain circumstances, when blood flow through the skin may be varying, the rate of absorbtion of nitrate may become unpredictable leading to 'pharmacological dumping' and premature patch exhaustion. This has been demonstrated in one study where fluctuating blood levels have been reported during conditions of extreme heat and physical exertion'17'. We have noticed that many patients have complained of marked headache shortly after exercise testing, which could represent a sudden peak in nitrate blood levels as a result of this 'dumping' phenomenon (unpublished observations). This study was therefore designed to establish firstly, how quickly the loss of therapeutic effect develops and how long therapeutic effects can be demonstrated following the application of a single patch; secondly, whether a nitrate-free interval is capable of maintaining day-long therapeutic efficacy on successive days; and finally whether tolerance is the sole cause of this loss of therapeutic effect or partly a reflection of premature patch exhaustion. Methods
Twelve male patients with chronic stable angina were studied (mean length of symptoms 62-7 months; range 12-180). Their mean age was 57-5 years (range 46-70). No patient was allowed concomitant vasodilator therapy but 10 were receiving additional beta-blocking therapy. © 1991 The European Society of Cardiology
A ttenuation of nitrate effect 1267
Doyl Patch 2
Patch 1
Patch 3
'k :B:
i
YA
t
t :
*
l2OOh Time of exercise test (3h)
l7OOh 2100 h (8h) (I2h)
l200h (3= 27h)
Sequence
A + B C+D
All
All
I
i
l700h 2IOOh (8= (12= 32W 36h) A+B
C+D
Figure I Diagramatic plan of study protocol. Treatment was administered in four treatment sequences (A,B,C,D) each lasting at least 32 h (dependent on the length of nitrate-free interval) and allowing exercise testing to be performed on two consecutive days. During the 5 days between each treatment sequence no vasodilator therapy was given. This allowed ample time for wash out of effects between successive treatment sequences. In each treatment sequence three active or dummy nitrate patches were applied continuously according to the plan above. Patch numbers 1 and 3 were always applied at 0900 h but patch number 2 was applied at 1700 or 2100 h depending on the sequence. Treatment sequences were randomized and administered double blind. £2 = sequence; §3 = active; Q = placebo.
Patients were not studied for the following reasons: a history of recent myocardial infarction (3 months), unstable angina, heart failure and inability to perform satisfactory exercise tests (for whatever reason). Before entry to the study patients were required to perform two exercise tests, one before and one after sublingual glyceryl trinitrate (GTN), to establish that they were responsive to nitrate therapy. Four patients (out of 16 initially screened) who failed to achieve an improvement of at least 1 min in their total exercise time were not studied. The study comprised a randomised, double-blind, within-patient comparison of active Transiderm-Nitro 10 and matching placebo patches. Treatment was administered in sequences lasting for 32-36 h depending on whether a 12 or a 16 h nitrate-free interval was employed (Fig. 1). Four treatment sequences were tested in four successive weeks with 5 days between each sequence. During each sequence three successive active or dummy patches were consecutively applied in various orders. Exercise testing was performed at 3 h (1200h) and at either 8 h (1700 h) or 12 h (2100 h) after patch application depending on the treatment sequence. The protocol allowed for assessment of continuous active therapy at 3, 8, 12, 15 and 24 h. Patients were given their supply of three patches at the beginning of each sequence; they applied the first patch at 0900 h on study day 1 and attended for an exercise test at noon, returning for a second test later the same day. The second patch was applied after the second exercise test on day 1 (1700 h or 2100 h, depending on treatment sequence) and remained in place until the following morning when the third patch was applied at 0900 h on study day 2. Throughout the study, exercise testing at 3,8 and 12 h was performed at the same time of
day and by the same observer for each patient. A plan of the four study sequences is shown in Fig. 1. Sequence A tested the effect of a 16 h nitrate-free interval and sequence C the effect of a 12 h nitrate-free interval. Sequence B provided most of the placebo data, but sequence D, day 1, provided the 12 h placebo data. Sequence D also provided the data on 15 and 24 h of continuous therapy using two successive active patches, to investigate the possibility of premature patch exhaustion. Exercise testing was performed on a treadmill using the standard Bruce protocol, preceded by the addition of a 3 min 'warm up' stage (2 kph at 5% incline). A modified 12 lead electrocardiogram was recorded throughout exercise. An attempt was made to standardize successive exercise tests by asking patients to assess the severity of their angina on a simple linear scale (1-20) at the end of the first test, and they were asked to attempt to achieve the same 'angina severity score' during each subsequent test. ST segments were measured 60 ms after the J point, in the lead initially demonstrating the greatest amount of ST depression; the same lead was used for all subsequent tests. The following five exercise test parameters were measured: time to 1 mm ST depression, time to angina, total exercise time, maximum ST depression and the amount of ST depression at the maximum common exercise time. Many of these parameters, although commonly employed, are hampered by the 'soft' end points by which they are measured. Thus, total exercise time, time to angina and maximum ST segment depression are entirely dependent upon the patient's exercise performance, and this is influenced by many factors other than angina. ST depression at maximum common exercise time and time to 1 mm ST depression are parameters which are
1268 M. A. James etal.
measured at a fixed time or electrocardiographic end Table I Combined results for the 3 h (1200 h) tests during intermittent therapy (A + C vs B). The results are expressed as the point and are thus uninfluenced by the patient's overall difference between the mean values for the active and placebo tests exercise performance. (The maximum common exercise time is determined by the shortest test performed by each Day 1 Day 2 patient, all other tests exceed this time point and ST depression can thus be measured at this same time point Time to 1 mm ST depression; s + 67 5 + 50 (and same workload) in all tests.) We believe that these P = 0-0004 /> = OO35 two parameters, which we have used previously'31, provide ST depression at MCET, mm -0-45 -0-40 /> = 00001 /> = 0-009 the most reliable measurement end points because of their objective nature. Other authors have subsequently used Maximum ST depression; mm -0-32 -0-22 these variables: time to 1 mm ST depression"418"1 and ST /> = 0-006 /> = 0-043 2 121 depression at a standardized exercise time"- - . Time to onset of angina; s + 13 + 26-5 Statistical analysis was by analysis of variance, a P /> = 038 />=0081 + 24-5 + 24 value of < 0 0 5 was taken as significant. However, our Total exercise time; s P = 0-092 /> = 0-078 statistician advised that, in view of the multiple comparisons made, a lower P value should be taken as significant for individual comparisons in order to maintain an overall MCET = Maximum common exercise time. 5% significance level. Thus a P value of 0-03 was taken as significant for individual comparisons and a value of of marginal significance (Table 2). Thus, during continuous therapy significant treatment effects were seen for 8 h 003-005 was considered of borderline significance. in several parameters; some treatment effect was seen at 12 and 15 h in fewer parameters and by 24 h there was no Results significant treatment effect at all (Table 3). In contrast, Comparisons are made both within the same treat- during intermittent therapy, significant treatment effects ment sequence (day 1 vs day 2) and between treatment were seen for 32 h, that is after 8 h of treatment on day 2 sequences (i.e. sequence A: day 1 vs sequence B: day 1). (Table 2). There was no significant treatment effect after The effects of active treatment, sequence A or C, were 12 h on the second day of intermittent therapy. However, compared with placebo, sequence B (placebo data for the since treatment effects after 12 h on the first day were 12 h exercise tests were obtained from sequence D, day 1; of marginal significance this probably did not reflect an see Fig. 1). Thecombined results of active treatment at 3 h important change. Nonetheless, treatment effects on the (A + C) can be compared with placebo, which by doubling second day, even during intermittent therapy, were not as the number of data points increases the statistical power good as those on the first day. Although the difference of this comparison. The efficacy of treatment on the between day 1 and day 2 of intermittent therapy was not second day in each treatment sequence can be tested by significant, this trend towards poorer results was reflected demonstrating that treatment remains significantly better in much lower P values on the second day (Tables 1 and 2). than placebo (i.e. A vs B). The second day of therapy in Figures 2 and 3 demonstrate the results graphically. sequence D tested the effect of continuous therapy since The graphs show the treatment effect expressed as the sequence D employed the use of two consecutive active mean percentage improvement in exercise test parameters patches, so that the 3 h exercise test on day 2 actually compared with the equivalent placebo test. Figure 2 represented 15 h of continuous treatment. The effect of shows the results for continuous therapy and demon15 h of continuous treatment, compared with placebo, strates that the improvement in exercise performance could then be established by comparison with sequence B: achieved by treatment falls in a steady and consistent way day 2. The effect of 24 h continuous therapy could be throughout the 24 h of continuous therapy. In contrast, compared with sequence D: day 1. The validity of com- during intermittent therapy (Fig. 3) treatment effects are parisons between day 1 and day 2 of any given treatment maintained at similar levels on both days 1 and 2. Howsequence was checked by looking for a difference between ever, the graph also demonstrates that there is a tendency day 1 and day 2 in sequence B (which was a complete for treatment effects on the second day to be slightly placebo sequence). diminished compared with the first day during interThe results showed a highly significant treatment effect mittent therapy and that considerable tolerance develops at 3 h when analysing the combined results (A+C vs B; even during the course of a single day's treatment. Table 1) and also when analysing both comparisons There was a marked diurnal variation in exercise perindependently (A vs B and C vs B/D; Table 2). During formance during this study (Fig. 4(a) and (b)). During intermittent therapy, significant treatment effects were placebo therapy exercise test parameters were signifimaintained after 3 h on the second day of treatment cantly better at 1200 h than at 1700 or 2100 h, whilst those (Table 1), regardless of the length of nitrate free period at 1700 h were better than at 2100 h. (Table 2). However, during continuous therapy, effects at In summary, the results have shown that there is a sig3 h on day 2 (15 h of continuous treatment) were only of nificant effect of continuous therapy with Transiderm marginal significance (Table 3). nitro for over 8 h, however, after 12 h while significant On the first day of treatment significant effects were effects can still be demonstrated, these are in fewer parseen at 8 h; however, treatment effects after 12 h were only ameters and at rather lower significance levels. After 15 h
Attenuation of nitrate effect 1269
Table 2 Results summary during intermittent therapy. The results are expressed as the difference between the mean valuesfor active and placebo tests, AvsB (8 h treatment/16 hfree) and C vs D/B (12 h treatment/12 hfree). Improvement is reflected by an increase in parameters measured in s and by a reduction in parameters measured in mm Day 1 1200h
ST depression at MCET; mm Maximum ST depression; mm Time to onset of angina; s Total exercise time; s
AvsB
CvsD
+ 50 P = 00\
+ 52 P = 0024 -0-27 P = 0\2
+ 37 /> = 0-2 -019 ^ = 0-24
-0-34 /> = 0-008
-Oil ^ = 0-44
+ 004 P = 0-85 + 32 P = 0-22 -2 /> = 0-96
CvsD
AvsB
+ 48 />=0-04 -0-41 /> = 0 007
+ 52 P=0027 -019 />=0-20
+ 50 P = 006 -0-35 P = 004
P = 006 -0-45
0 />=10
/> = 0-06
-004 /> = 0-80 + 60 P=0005
+ 27 /> = 011
P = 004
-010 P = 0-4\ + 42 P = 002 + 38 /> = 0-018
-0-24 P = 006
+ 40
+ 30 P=O19 + 27 P=Q 10
CvsB
AvsB
+ 58 + 92 />=0-006 P< 00001 -0-40 -0-46 /> = 0-002 />=00004 - 0 19
CD2vsDDl 2100h
2100 h
CvsD
P=0l3
1700h
1700h
AvsB Time to 1 mm ST depression; s
Day 2 1200h
+ 40 /> = 0047 + 26 /> = 0-25
+ 50
+ 11
+1
/>=0-46
P = 0-94 + 26 /> = 0-ll
+ 10
/>=0-51
End column: C D2 = sequence C, day 2 test 2 (2100 h); D D1 = sequence D, day 1 test 2 (2100 h). MCET = maximum common exercise time.
Table 3 Results during continuous therapy expressed as the difference between the mean values for active and placebo tests at the same times of day. Resultsfor 3,8 and 12 h tests are obtained from day 1 of treatment during sequences A and C. Results for 15 and 24 h of treatment are obtained from the 3 h (1700 h) and 12 h (2100 It) exercise tests on day 2 of sequence D. As shown before, improvement is reflected by an increase in parameters measured in s and by a reduction in parameters measured in mm Hours
Time to 1 mm ST depression; s ST depression at MCET; mm Maximum ST depression; mm Time to onset of angina; s Total exercise time; s
3
8
12
15
24
+ 58 P = 0-006 -0-40 /> = 0-002
+ 48 f = 004 -0-41 />=0-007
+ 52 /> = 0-027 -019 ^ = 0-20
+ 25 /> = 0-35 -0-22 ^ = 0-22
-3 P = 0-9\ -006 /> = 0-73
-019 /> = 0-13 + 30 P = 0 19 + 27 P = 0\0
-004 p = 0% + 60 /> = 0-005
0 P=\0 + 40 /> = 0-047
-003 /»=0-78 + 40 ^ = 003 + 38 /> = 0-016
+ 006 P = 0-68 + 27
+ 50 ^ = 0-04
+ 26
P= 25
°
P = 0\6
+ 10 /> = 0-50
MCET = maximum common exercise time.
of treatment a significant effect was noted in one parameter, but most parameters showed little difference from pretreatment levels, and after 24 h no discernible effect of treatment could be demonstrated (Table 3). By contrast, during intermittent therapy, significant effects of treatment could be demonstrated after 27 h of treatment, regardless of whether a 12 or 16 h break in therapy had been employed (Table 2) and treatment effects could be demonstrated for up to 32 h (8 h of day 2). However, treatment effects on day 2 did seem to be somewhat attenuated compared with day 1 even during intermittent therapy (Tables 1 and 2).
Discussion Several studies have suggested that the therapeutic effect of transdermal nitrate treatment can be maintained during long-term usage by intermittent or interval therapy. However, the major weakness in most of these studies is that treatment effects have only been assessed for a relatively short time after patch application. Thus it remains unknown whether nitrate patches are capable of providing day-long prophylaxis against angina, or how long a single patch may be expected to be effective. This study, whilst confirming these earlier reports, that significant treatment
1270 M. A. James etal.
30
( a ) 600
500 -
20
E
S
H
400 -
10
300 10
20
30
Time of test (h)
Figure 2 The mean percentage improvement in exercise test parameters during continuous therapy The graph shows the mean percentage improvement in exercise parameters compared with the equivalent placebo tests during continuous treatment. The results are presented for the meaned percentage improvements of all five measured parameters ( ) and also for the two we considered most important ( ).
30
o *(
Day2
Day 1
\ 25 -
F
\
1-4
\
/
20 -
/
\
s.
\
•
V\ \
15 -
\ V
\ \
Figure 4(a) and (b) The diurnal variation in exercise test parameters. The graphs demonstrate the difference in the results of all five exercise test parameters at the three times of day they were tested, (a) A = total exercise time; • = time to angina; • = time to 1 mm ST depression; (b) • = maximum ST depression; D = S T depression at maximum common exercise time.
\
/
V 10
3
8
112
1 20
?7 ±32 30
b 40
Time of test (h)
Figure 3 The mean percentage improvement in exercise test parameters during intermittent therapy The graph demonstrates the mean percentage improvement in the time to 1 mm ST segment depression and the STdepression at maximum common exercise time. Improvement in performance is maintained on day 2 during both intermittent treatment regimens (cf. Fig I). However, the percentage improvement is slightly smaller on day 2 compared with day 1 during both treatment regimens. Note also the deterioration in the improvement measured during the course of a single day • . 8— 16 h schedule; O , 12-12 h schedule.
effects can be maintained for at least 3 h during intermittent therapy, also demonstrates two further important findings. Firstly, tolerance develops so rapidly that although significant treatment effects are maintained for several hours, they become markedly attenuated in less than a day. Secondly, although significant treatment
effects are maintained into the second day of treatment during intermittent therapy, the results are consistently attenuated compared with the initial (first day) response (also reflected in lower P values on the second day). In addition, the results confirm that loss of therapeutic efficacy during treatment is due to the development of tolerance rather than patch exhaustion. Continuous therapy with consecutive fresh patches does not lead to a significant difference from placebo on the second day of treatment even at 3 h, whilst during intermittent therapy significant treatment effects are maintained on the second day. Tolerance to nitrates began to develop very soon after treatment had started and became well developed between 8 and 12 h. Both 12 and 16 h nitrate-free intervals were sufficiently long to prevent the occurrence of tolerance at 3 h after restarting treatment. However, as tolerance was already well established by 12 h of treatment there would
Attenuation of nitrate effect 1271
seem to be no advantage in a treatment regimen employing a shorter nitrate-free interval. Furthermore, the development of tolerance did not appear to be a threshold phenomenon, but a continuous process, beginning probably from the moment the patch was applied (although we of course only began testing from 3 h). Figure 2 shows a graph of the mean percentage change in exercise test parameters during continuous therapy. The results are presented for the mean percentage improvement in all five exercise parameters and also for the two exercise parameters which we consider to be the most reliable (ST depression at maximum common exercise time and time to 1 mm ST segment depression), as previously explained. The graph shows the apparent steady and continuous decline in therapeutic effect throughout the study period. This steady decline in effect is observed regardless of how many parameters are analysed. However, treatment effects are apparently greater when restricting analysis to the two 'important' exercise parameters and this is due to the greater consistency of these measurements. This finding is important because previous studies have not generally evaluated the time course of development of tolerance. Thus, although most studies have demonstrated a significant treatment effect 0-4 h after patch application during intermittent therapy, the conclusions drawn, that this therefore implied day-long efficacy are not justified. In this study, therapeutic efficacy was also demonstrated at 3 h during intermittent therapy. However, whilst significant treatment effects were maintained for 12-15 h these effects were markedly attenuated compared with the 'full' nitrate effect earlier in the day. Only three other studies1'418191 have reported the effect of nitrate patches later in the day during intermittent therapy. Schaer et a/.'141 performed exercise tests at 4 and 8 h and also demonstrated attenuation of effect after 8 h, although the authors made no comment on this finding. Waters et a/.1'81 performed exercise tests after 12 or 16 h of treatment and found almost no significant effects at all, although their data also showed a trend toward better results at 12 h than at 16 h. The most recent study (de Mots et a/.)'"1 demonstrated efficacy up to 12 h on thefirstday of treatment, but only at 4 and 8 h during long-term therapy. Thus, although two of these studies purported to show that intermittent therapy was efficacious, all three studies in fact provide support for the findings of this study, namely, that therapeutic effects are lost rapidly, within the course of a single day, and that intermittent therapy can only maintain therapeutic responsiveness for part of the day. Although results on the second day were not significantly different from results on the first day, there was a suggestion that they were persistently slightly worse on day two. This would suggest that even during intermittent therapy, tolerance may only be partly averted, with persistent attenuation of effect despite the nitrate-free interval. It is possible that this attenuation could be a cumulative phenomenon, but this would require further prospective evaluation. In the longer term, cumulative attenuation would result in gradual loss of all nitrate effect and might lead to a rebound increase in ischaemia and angina. This
could only be evaluated by a longer follow-up period. However, previous long-term studies have not commented on a rebound increase in angina, and have generally shown that the immediate (0-4 h) response is maintained over periods of 1-4 weeks,'1214"1 suggesting that cumulative attenuation may not be a problem. All statistical comparisons in this study were made with placebo tests performed at the same time of day. The graphs in Figs 2 and 3 were made up using values for percentage improvement compared with a placebo test at the same time of day. This method of analysis avoids the problems of the variability in absolute values of the exercise data derived from different variables and at different times of day. Statistical comparisons have not been made on this derived data; however, it is interesting to note that the results of the derived data reflect exactly the same pattern as the results of the statistical comparisons on the actual data given in Tables 1-3. Finally, the marked diurnal variation in exercise test performance observed in this study is important to all studies employing comparative exercise testing. This diurnal variability has been reported previously120'211 and the current findings serve to underline the importance of uniform exercise test times during comparative studies employing exercise testing. We thank Ciba-Geigy Pharmaceuticals who provided the active and placebo nitrate patches and whose department of statistics performed the statistical analyses.
References [1] Rezakovic D, Stalec J. Acute and chronic efficacy of 10 mg nitroglycerin patch in stable angina pectoris. Am J Cardiol 1988;61:52E-58E. [2] Greco R, D'Alterio D, Schiattarella M, Boccia A, Greco L, Marsico F. Efficacy of a new transdermal nitroglycerin patch (Deponit 10) for stable angina pectoris. Am J Cardiol 1988; 61: 44E-51E. [3] James MA, Walker PR, Papouchado M, Wilkinson PR. Efficacy of transdermal glyceryl trinitrate in the treatment of chronic stable angina pectoris. Br Hear J 1985; 53: 631-5. [4] Parker JO, VanKoughnett KA, Fung H-L. Transdermal isosorbide dinitrate in angina pectoris: effect of acute and sustained therapy. Am J Cardiol 1984; 54: 8-13. [5] Abrams J. The brief saga of transdermal nitroglycerin discs: paradise lost. Am J Cardiol 1984; 54: 220-4. [6] Thadani U. The effectiveness of transcutaneous nitrate preparations for angina pectoris. Int J Cardiol 1987; 14:9-14. [7] Packer M, Lee WH, Kessler PD, Gottlieb SS, Medina N, Yushak M. Prevention and reversal of nitrate tolerance in patients with congestive heart failure. New Eng J Med 1987; 317:799-804. [8] Parker JO, Fung H-L, Ruggirello D, Stone JA. Tolerance to isosorbide dinitrate, rate of development and reversal. Circulation 1983; 68: 1074-80. [9] Parker JO. Nitrate tolerance. Am J Cardiol 1985; 56: 28L-31L. [10] Desilvio A, Barlattani MP. How best to use nitrates. Br Med J 1987; 295: 1163-4. [11] Anon. Nitrates: the problem of tolerance. Drug Ther Bulletin 1988; 26: 57-9. [12] Cowan JC, Bourke JP, Reid DS, Julian DS. Prevention of tolerance to nitroglycerin patches by overnight removal. Am J Cardiol 1987; 60: 271-5. [13] Luke R, Sharpe N, Coxon R. Transdermal nitroglycerin in angina pectoris: efficacy of intermittent application. J Am Coll Cardiol 1987; 10:642-6.
1272 M. A. James etal.
[14] Schaer DH, Buff LA, Katz RJ. Sustained anti-anginal efficacy of transdermal nitroglycerin patches using an overnight 10 hour nitrate free interval. Am J Cardiol 1988; 61:46-50. [15] Wolff M, Cordes G, Luckow V. In vitro and In vivo release of nitroglycerin from a new transdermal therapeutic system. Pharm Res 1985; 1: 1-54. [16] Shaw JE. Transdermal therapeutic systems. In: Elsdon Drew RW, Birdwood GFB, eds. Transdermal nitrates in ischaemic heart disease International congress and symposium series, London Royal Society of Medicine. 1983; 59: 33-5. [17] Barkve TF, Langseth-Manrique K, Bredesen JE, Gjesdal K. Increased uptake of transdermal glyceryl trinitrate during physical exercise and during high ambient temperature. Am Heart J 1986; 112:537-41.
[18] Waters DD, Juneau M, Gossard D, Choquette G, Brien M. Limited usefulness of intermittent nitroglycerin patches in stable angina. J Am Coll Cardiol 1989; 13:421-5. [19] DeMots H, Glasser SP and the Transiderm-Nitro study group. Intermittant transdermal nitroglycerine therapy in the treatment of chronic stable angina. J Am Coll Cardiol 1989; 13: 786-93. [20] Jones EF, Joy MD. Parasympathetic blockade with benzhexol hydrochloride in the management of stable angina pectoris. Clin Sci 1987; 72(Suppl 16): 57. [21] Joy M, Pollard CM, Nunan TO. Diurnal variation in exercise responses in angina pectoris Br Heart J 1982; 48: 156-60.
