TheophyllineConcentrations in PatientsWith AcuteExacerbation of COPD CHARLES L. EMERMAN, MD,* ALFRED F. CONNORS, MD,t THOMAS W. LUKENS, MD,* MICHAEL E. MAY, MD,* DAVID EFFRON, MD* Patients with acute exacerbation of chronic obstructive pulmonary disease (COPD)fmquentiy reporf a history of theophyfiine use prior to arrival in the emergency department. The miiabllity of wing tftis history to guide aminophyliine therapy is, however, unclear. Tbe authorsstudied patients with an acute exacerbation of COP0 to determine the relation frotweenthe inltfai tboofrbyiiine level and Dre medkatioo history. 6eventyuiee patients with a moan age of 64.3 years wore entered ink tbe stedy. Tire average Woopbyiiine level was 10.8 f@mL. Forty-seven percent of the patients had &therapeutic levels, 46% bad tbempeutic levels, and 7% had toxic levels. Them wem weak aitbough statistically significant cormlotions between the theophyiiine level aed the iotetvai since tire last Dtoophyflinedose (r = -.40, P < .Wl), as well as with tffe lost mcorded outpatient theophygine level (r = 46, P < 6fJl). However, tire theof& yiline level could not be predicted accomteiy In a number of patients. The authors conclude, because the pmdktion rule was often inaccomte, aminopfryiiinetherapy in patients with acute exacerbation of COPDshoeid be based on dimcf measumment of the semm theophyiline level. (Am J Emerg Med 1990;8:066_M)O. 0 1990 by W.8. Saunders Company.)
obstructive lung disease. Factors which affect theophylline clearance such as advanced age, and cigarette use may be more common in COPD patients as compared with asthma patients.4 Studies in patients with hypertension and diabetes indicate that older patients tend to be more compliant with medication use, although cigarette users and men tend to be less compliant.5,6 Older asthmatics have been found to be more compliant in the use of metered dose inhalers.’ The overall effect of these factors on serum theophylline levels in patients with COPD has not been extensively studied. The high prevalence of theophylline use in patients with COPD and the risk of toxicity prompted us to investigate the role of empiric formulas in these patients. The purpose of this study, therefore, was to determine whether the history of theophylline use could be used to predict theophylline levels in patients with an acute exacerbation of COPD. METHODS
Patients with an acute exacerbation of chronic obstructive pulmonary disease (COPD) frequently present with a history of previous outpatient theophylline use. However, there are few studies addressing the compliance of these patients with their drug regimen or of the efficacy of usual oral doses of theophylline in achieving therapeutic levels. Consequently it is not clear how to use the history of theophylline use in adjusting the initial theophylline dose in patients with an acute exacerbation of COPD. Evidence in asthmatic patients indicates that, although most patients will present with subtherapeutic levels, the history of theophylline use is a poor predictor of serum theophylhne levels.’ In some studies empiric formulas have been found unreliable for calculating the aminophylline dosage for asthmatic patients.’ However, another study has recently found that asthmatic patients reporting an interval of 12 or 24 hours since the last dose of short or long acting theophylline, respectively, may be administered a full loading dose safely.3 There are several reasons to question whether the same findings are true for patients with chronic From the Departments of Emergency Medicine” and Medicine,T MetroHealth Medical Center, (formerly Cleveland Metropolitan General Hospital), and the Departments of Surgery’ and Medicinet Case Western Reserve University, Cleveland, OH. Manuscript received June 29, 1989; revision accepted August 29, 1989. Address reprint requests to Dr Emerman, Department of Emergency Medicine, MetroHealth Medical Center, 3395 Scranton Rd, Cleveland, OH 44109. Key Words: Theophylline, COPD, respiratory distress, obstructive lung disease. 8 1990 by W.B. Saunders Company. 0735-6757/90/0804-ooo4$5.og/o
Patients presenting to the emergency department of MetroHealth Medical Center with a clinical history of COPD’ and a history of outpatient theophylline use were considered for entry into the study. All patients were at least 50 years of age and presented with a chief complaint of respiratory distress. Patients were excluded from the study if they had a history of asthma’ or a history of onset of episodes of respiratory distress before age 35. Patients were also excluded from the study if they were found to have pneumonia, acute congestive heart failure, or acute conditions other than exacerbation of COPD. All patients gave written, informed consent to a protocol approved by the hospital’s Human Investigation Committee. Before beginning any therapy, all patients had blood drawn for a complete blood count, and to measure arterial blood gases, and serum theophylline level. The serum theophylline levels were analyzed in our clinical chemistry laboratory using an enzyme-mediated immuno assay technique. A detailed history of theophylline intake was obtained and prior serum theophylline levels were obtained from the hospital medical record since other studies have indicated that this may be useful in predicting levels.’ Spirometry was performed on all patients using a Fleisch-pneumotach type computerized spirometer (Spiroscan 1000, Brentwood Instruments, Portland, OR). Spirometry was performed with the patient seated and wearing nose clips. At least two, and usually three, curves were obtained. Those curves with the highest forced vital capacity (FVC) and forced expiratory volume (FEV,,,) were used for analysis. Continuous data was analyzed using student’s f-test, and categorical data was analyzed using the x2 test. Correlations were made using Pearson’s product moment correlation co289
AMERICAN JOURNAL OF EMERGENCY MEDICINE H Volume 8, Number 4 n July 1990
290
efficient. Stepwise linear regression was used to identify variables most predictive of the measured theophylline level.” A P < .05 was taken to indicate statistical significance. AU data is reported as mean 2 standard deviation.
30 f
.
22 =
RESULTS
.
E
-
.
.
20-
.
*. ..
0’ . : .. 0 . 3 . . .O* . . .O FE ,~_~~~~~~~~~~~~~~~~~ r 0 o. :. . 8 . .P . 2 . . I*. 7
Seventy-nine patients were entered into the study with an average age of 64.3 & 8.1 years. There were 42 males and 37 females with an average of 57.3 2 31.8 pack years of cigarette use. Ninety-six percent of the patients gave a history of chronic bronchitis,8 63% of the patients had a history of emphysema, 82% of the patients were currently using a P-agonist inhaler, and 27% of the patients had a history of steroid use. All of the patients on outpatient theophylline products were using long-acting preparations. The average serum theophylline level on arrival in the emergency department (ED) was 10.8 & 6.2 pg/mL. Fortyseven percent of the patients had subtherapeutic initial levels (20 pg/mL). There was no significant difference in age, history of steroid or P-agonist use, or pack years of cigarette use between those with subtherapeutic levels and those with levels of 10 pg/mL or greater (Table 1). There was no correlation between the initial FEV, (r = 0.14, NS), or the percent predicted FEV, (r = -0.08, NS) (Fig 1) and the theophylline level. There was no correlation between the initial theophylline level and the subsequent change in FEV,,, with treatment. There was a weak, although statistically significant correlation between the theophylline level and the time since the last outpatient dose (r = - .40, P < .OOl) (Fig 2). Six patients stated that it had been more than 24 hours since their last theophylline dose. One (17%) of these patients had a therapeutic level (16 ug/mL). Sixty patients had theophylline levels recorded on outpatient visits before the study. Thirty of these prior theophylline levels were subtherapeutic, 26 were in the therapeutic range, and two were toxic. Sixty-three percent of the patients whose earlier theophylline levels were in the therapeutic range had therapeutic levels on entry into the study as
.
0 I
1
0
10
. .. . I
.
.
*
.
: . ..
.
. I
L
I
I
I
70
50
30 Percent Predicted FEV1.g
FIGURE 1. Relation between the initial theophylline level and the percent predicted FEV,., (r = -0.08, NS). Open circles indicate multiple observations at this point.
compared to 26% of the patients whose previous levels were subtherapeutic (P < .Ol). Of the patients whose ED levels were in the toxic range, one had subtherapeutic levels on outpatient testing, three had therapeutic levels, and one had toxic levels. There was a significant correlation between the theophyliine level and the last outpatient level (r = 0.45, P < .OOl) (Fig 3). We analyzed the usefulness of a multiple regression model to predict the serum theophylline level. The model was developed using univariate analysis to identify those variables significantly associated with the theophylline levels. Only the time since last theophylline dose, prior outpatient theophylline level, and history of cigarette use remained in the model after stepwise linear regression. The equation for this model was: theophylline level = 12.3 - (0.1 x hours since last dose) + (0.5 x last outpatient level) + (5.3 if the patient is a cigarette user). This model had an R* of 0.40 (P < .OOOl) (Fig 4). Thirty-three percent of the patients predicted to have subtherapeutic levels by this model actually had therapeutic levels, while 2% of the patients predicted to have levels greater than 10 pg/mL actually had subtherapeutic levels.
TABLE1. Comparison of Patients With Theophylline Levels Lesser or Greater Than 10 FglmL on Presentation to the ED Theophylline Level Lessthan 10 p@mL
~~0, POP Percent predicted FEV,,, Hours since last outpatient theophylline dose Last outpatient theophylline level (pg/mL) ‘*SD.
10
than
&mL
63.836 7.4 59.4 +- 36.6
42 64.3 + 6.4 54.1 + 25.6
NS NS
61% 22% 7.42 + 64 41.3 + 6.6 66.5 + 16.0 25.2 2 11.4
65% 31% 7.45 f .06 39.4 f 9.6 62.5 f 13.3 26.8 f 12.9
NS NS NS NS NS NS
18.4 f 32.0
6.0 + 6.2
P < .05
6.1 + 4.6
10.9 + 6.6
N Age (years) Cigarette use (pack years) History of B-agonist inhaler use History of steroid use
Greater
P i
.Ol
3 2
/ ,o
i *. . :: . . . . .
.
.
1_‘__L_-L__-____
O’C.
28
z
.
. :
0
I
0
0 .
.
.
0
.
.
I
I
8
I
16
I
I
24
Hours since last Theophylline Dose
FIGURE 2. Relation between the theophylline level and the time since the last medication dose (r = - 0.40, P
obstructive lung disease. Factors which affect theophylline clearance such as advanced age, and cigarette use may be more common in COPD patients as compared with asthma patients.4 Studies in patients with hypertension and diabetes indicate that older patients tend to be more compliant with medication use, although cigarette users and men tend to be less compliant.5,6 Older asthmatics have been found to be more compliant in the use of metered dose inhalers.’ The overall effect of these factors on serum theophylline levels in patients with COPD has not been extensively studied. The high prevalence of theophylline use in patients with COPD and the risk of toxicity prompted us to investigate the role of empiric formulas in these patients. The purpose of this study, therefore, was to determine whether the history of theophylline use could be used to predict theophylline levels in patients with an acute exacerbation of COPD. METHODS
Patients with an acute exacerbation of chronic obstructive pulmonary disease (COPD) frequently present with a history of previous outpatient theophylline use. However, there are few studies addressing the compliance of these patients with their drug regimen or of the efficacy of usual oral doses of theophylline in achieving therapeutic levels. Consequently it is not clear how to use the history of theophylline use in adjusting the initial theophylline dose in patients with an acute exacerbation of COPD. Evidence in asthmatic patients indicates that, although most patients will present with subtherapeutic levels, the history of theophylline use is a poor predictor of serum theophylhne levels.’ In some studies empiric formulas have been found unreliable for calculating the aminophylline dosage for asthmatic patients.’ However, another study has recently found that asthmatic patients reporting an interval of 12 or 24 hours since the last dose of short or long acting theophylline, respectively, may be administered a full loading dose safely.3 There are several reasons to question whether the same findings are true for patients with chronic From the Departments of Emergency Medicine” and Medicine,T MetroHealth Medical Center, (formerly Cleveland Metropolitan General Hospital), and the Departments of Surgery’ and Medicinet Case Western Reserve University, Cleveland, OH. Manuscript received June 29, 1989; revision accepted August 29, 1989. Address reprint requests to Dr Emerman, Department of Emergency Medicine, MetroHealth Medical Center, 3395 Scranton Rd, Cleveland, OH 44109. Key Words: Theophylline, COPD, respiratory distress, obstructive lung disease. 8 1990 by W.B. Saunders Company. 0735-6757/90/0804-ooo4$5.og/o
Patients presenting to the emergency department of MetroHealth Medical Center with a clinical history of COPD’ and a history of outpatient theophylline use were considered for entry into the study. All patients were at least 50 years of age and presented with a chief complaint of respiratory distress. Patients were excluded from the study if they had a history of asthma’ or a history of onset of episodes of respiratory distress before age 35. Patients were also excluded from the study if they were found to have pneumonia, acute congestive heart failure, or acute conditions other than exacerbation of COPD. All patients gave written, informed consent to a protocol approved by the hospital’s Human Investigation Committee. Before beginning any therapy, all patients had blood drawn for a complete blood count, and to measure arterial blood gases, and serum theophylline level. The serum theophylline levels were analyzed in our clinical chemistry laboratory using an enzyme-mediated immuno assay technique. A detailed history of theophylline intake was obtained and prior serum theophylline levels were obtained from the hospital medical record since other studies have indicated that this may be useful in predicting levels.’ Spirometry was performed on all patients using a Fleisch-pneumotach type computerized spirometer (Spiroscan 1000, Brentwood Instruments, Portland, OR). Spirometry was performed with the patient seated and wearing nose clips. At least two, and usually three, curves were obtained. Those curves with the highest forced vital capacity (FVC) and forced expiratory volume (FEV,,,) were used for analysis. Continuous data was analyzed using student’s f-test, and categorical data was analyzed using the x2 test. Correlations were made using Pearson’s product moment correlation co289
AMERICAN JOURNAL OF EMERGENCY MEDICINE H Volume 8, Number 4 n July 1990
290
efficient. Stepwise linear regression was used to identify variables most predictive of the measured theophylline level.” A P < .05 was taken to indicate statistical significance. AU data is reported as mean 2 standard deviation.
30 f
.
22 =
RESULTS
.
E
-
.
.
20-
.
*. ..
0’ . : .. 0 . 3 . . .O* . . .O FE ,~_~~~~~~~~~~~~~~~~~ r 0 o. :. . 8 . .P . 2 . . I*. 7
Seventy-nine patients were entered into the study with an average age of 64.3 & 8.1 years. There were 42 males and 37 females with an average of 57.3 2 31.8 pack years of cigarette use. Ninety-six percent of the patients gave a history of chronic bronchitis,8 63% of the patients had a history of emphysema, 82% of the patients were currently using a P-agonist inhaler, and 27% of the patients had a history of steroid use. All of the patients on outpatient theophylline products were using long-acting preparations. The average serum theophylline level on arrival in the emergency department (ED) was 10.8 & 6.2 pg/mL. Fortyseven percent of the patients had subtherapeutic initial levels (20 pg/mL). There was no significant difference in age, history of steroid or P-agonist use, or pack years of cigarette use between those with subtherapeutic levels and those with levels of 10 pg/mL or greater (Table 1). There was no correlation between the initial FEV, (r = 0.14, NS), or the percent predicted FEV, (r = -0.08, NS) (Fig 1) and the theophylline level. There was no correlation between the initial theophylline level and the subsequent change in FEV,,, with treatment. There was a weak, although statistically significant correlation between the theophylline level and the time since the last outpatient dose (r = - .40, P < .OOl) (Fig 2). Six patients stated that it had been more than 24 hours since their last theophylline dose. One (17%) of these patients had a therapeutic level (16 ug/mL). Sixty patients had theophylline levels recorded on outpatient visits before the study. Thirty of these prior theophylline levels were subtherapeutic, 26 were in the therapeutic range, and two were toxic. Sixty-three percent of the patients whose earlier theophylline levels were in the therapeutic range had therapeutic levels on entry into the study as
.
0 I
1
0
10
. .. . I
.
.
*
.
: . ..
.
. I
L
I
I
I
70
50
30 Percent Predicted FEV1.g
FIGURE 1. Relation between the initial theophylline level and the percent predicted FEV,., (r = -0.08, NS). Open circles indicate multiple observations at this point.
compared to 26% of the patients whose previous levels were subtherapeutic (P < .Ol). Of the patients whose ED levels were in the toxic range, one had subtherapeutic levels on outpatient testing, three had therapeutic levels, and one had toxic levels. There was a significant correlation between the theophyliine level and the last outpatient level (r = 0.45, P < .OOl) (Fig 3). We analyzed the usefulness of a multiple regression model to predict the serum theophylline level. The model was developed using univariate analysis to identify those variables significantly associated with the theophylline levels. Only the time since last theophylline dose, prior outpatient theophylline level, and history of cigarette use remained in the model after stepwise linear regression. The equation for this model was: theophylline level = 12.3 - (0.1 x hours since last dose) + (0.5 x last outpatient level) + (5.3 if the patient is a cigarette user). This model had an R* of 0.40 (P < .OOOl) (Fig 4). Thirty-three percent of the patients predicted to have subtherapeutic levels by this model actually had therapeutic levels, while 2% of the patients predicted to have levels greater than 10 pg/mL actually had subtherapeutic levels.
TABLE1. Comparison of Patients With Theophylline Levels Lesser or Greater Than 10 FglmL on Presentation to the ED Theophylline Level Lessthan 10 p@mL
~~0, POP Percent predicted FEV,,, Hours since last outpatient theophylline dose Last outpatient theophylline level (pg/mL) ‘*SD.
10
than
&mL
63.836 7.4 59.4 +- 36.6
42 64.3 + 6.4 54.1 + 25.6
NS NS
61% 22% 7.42 + 64 41.3 + 6.6 66.5 + 16.0 25.2 2 11.4
65% 31% 7.45 f .06 39.4 f 9.6 62.5 f 13.3 26.8 f 12.9
NS NS NS NS NS NS
18.4 f 32.0
6.0 + 6.2
P < .05
6.1 + 4.6
10.9 + 6.6
N Age (years) Cigarette use (pack years) History of B-agonist inhaler use History of steroid use
Greater
P i
.Ol
3 2
/ ,o
i *. . :: . . . . .
.
.
1_‘__L_-L__-____
O’C.
28
z
.
. :
0
I
0
0 .
.
.
0
.
.
I
I
8
I
16
I
I
24
Hours since last Theophylline Dose
FIGURE 2. Relation between the theophylline level and the time since the last medication dose (r = - 0.40, P
