Jeffrey Spaeder, MD; Samer S. Najjar, MD; Gary Gerstenblith, MD; Gail Hefter, RN, MS, CRNP; Linda Kern, PA-C; James G. Palmer, MS; Sheldon H. Gottlieb, MD; Edward K. Kasper, MD 

Am Heart J.  2006;151(4):844.e1-844.e10.  ?2006 Mosby, Inc.

Abstract and Introduction

Abstract

Background: Because of potential side effects and logistical difficulty of titrating medications, outpatients with congestive heart failure rarely receive appropriate doses of carvedilol or other β-blockers. To address these obstacles, we studied if an automated telemedicine system named TeleWatch (TW) could facilitate carvedilol titration in outpatients with left ventricular systolic dysfunction.
Methods: Forty-nine patients with New York Heart Association class II and III left ventricular systolic dysfunction, who were tolerating appropriate afterload-reducing therapy and not receiving β-blockers, were enrolled into a 3-month study. Patients were randomized to have clinic-only (CO) (n = 24) carvedilol titration or titrations which combined clinic visits with TW monitoring (n = 25). All patients were seen in clinic biweekly, and those in the TW group were remotely monitored daily. Using a predefined algorithm, patients in the CO and TW groups were eligible for carvedilol titration on a biweekly or weekly basis, respectively, by physicians blinded to group assignment.
Results: There was no statistical difference in the mean final daily dose of carvedilol between the CO and TW groups (39.4 vs 36.2 mg/d, P = .52). Because remote telemedicine titrations were as successful as titrations in the clinic, the time to reach the final dose of carvedilol was significantly shorter in the TW group (33.6 vs 63.7 days, P < .0001). There were 5 serious adverse events in the study, 4 of which were in the TW group (P = .29); however, TW prospectively detected 2 adverse events.
Conclusions: Remote monitoring with an automated telemedicine system can successfully facilitate titration of carvedilol in outpatients with New York Heart Association class II and III congestive heart failure.

Introduction

β-Adrenergic blocking agents reduce morbidity and mortality in patients with congestive heart failure (CHF) due to impaired left ventricular (LV) systolic function.^[1-7] Carvedilol is a third-generation β-blocker that improves LV systolic function, decreases hospitalizations by 15% to 27%, and reduces mortality by 25% to 73%.^[1,6,7] Because its improvement upon LV function are dose-dependent, CHF guidelines suggest a goal carvedilol dose of at least 25 mg bid.^[7,8] However, in routine outpatient practice, only 35% to 48% of patients with CHF are prescribed carvedilol or other β-blockers.^[9-11] In addition, although goal dose of carvedilol is achieved in 66% to 80% of patients with CHF in clinical trials,^[1,6] studies report that within 60 days after initiation of carvedilol, only 3% of patients receive this goal dose. Although the Coreg Heart Failure Registry reported that 49% of patients treated by cardiologists achieved the goal carvedilol dose of 25 to 50 mg bid, only 27% of noncardiologists in the same study achieved the goal dose, whereas in other epidemiologic studies, only 18% of all outpatients with CHF achieved a dose of at least 25 mg bid.^[12-15] Because the improvements in LV function are dose-dependent,^[7] the observed slow rate of titration and suboptimal dosing of carvedilol deprive patients from the maximal benefits of the medication. Therefore, new strategies are required to increase the number of patients with CHF who receive carvedilol or other β-blockers, as well as the rate with which they are titrated.

Concerns about precipitating clinical deterioration in the less intensively monitored outpatient setting and a preoccupation of addressing patient complaints and other acute clinical issues are potential reasons why healthcare providers are not more aggressive in up-titrating β-blockers in patients with CHF.^[16] Participation in disease management programs increases the proportion of patients with CHF who receive β-blocker therapy;^[17] however, personalized nurse-directed disease-management programs are resource-intensive and expensive. One method of decreasing the resources required for intensive outpatient monitoring is the use of automated telemedicine systems, which can provide moderate intensity outpatient cardiovascular monitoring.^[18,19] However, to date, there have been no published studies regarding the use of an automated telemedicine system to remotely titrate vasoactive medications in an outpatient CHF population. For this reason, the C-CUTS was performed, in which the efficacy of making titration decisions based upon data obtained from an automated telemedicine system was compared with the data obtained during a clinic visit.

Methods

C-CUTS was a randomized, physician-blinded, controlled study of the initiation and dose escalation of carvedilol in outpatients with impaired LV systolic function using the TeleWatch (TW) telemedicine system combined with a titration algorithm. The study had 2 phases: a titration phase, during which carvedilol dose was titrated and physicians were blinded to patient group assignment and a follow-up phase, which started after the completion of carvedilol titration and continued for 3 months after enrollment.

Eligibility criteria for this study included male or female patients who understood spoken English, were at least 21 years of age with New York Heart Association (NYHA) class II or III heart failure and an ejection fraction of <40%, and who were clinically euvolumic and receiving appropriate afterload reducing agents (angiotensin-converting enzyme inhibitors, angiotensin receptor blockers or hydralazine/nitrates) of at least intermediate dose (equivalent to a dose of enalapril of at least 5 mg bid). Exclusion criteria included treatment with any β-blocking agent within the 6 months before enrollment, unstable angina or myocardial infarction within 6 weeks, contraindications to β-blocker therapy, a pulse <60 beat/min, and a systolic blood pressure (SBP) <90 mm Hg. Patients who were pregnant, as well as those with active substance abuse, uncontrolled psychiatric illness, or a life expectancy of <12 months were also excluded from the study. Potential study participants were identified from the inpatient medicine services at Johns Hopkins Hospital and Johns Hopkins Bayview Medical Center, who were admitted with the diagnosis of decompensated heart failure as well as patients referred to the outpatient heart failure clinics of these 2 institutions. Participants who were identified as inpatients were enrolled into the study only after discharge from the hospital and when clinically stable. The institutional review board of the study sites reviewed and approved the study protocol. An independent data safety monitoring board reviewed all serious adverse events at prespecified time points and approved the continuation of the study after each review.

After qualified patients signed informed written consent, they were randomized to carvedilol titration performed exclusively at biweekly visits in a CHF clinic (hereafter referred as Clinic-only [CO] group) or titration performed during biweekly visits to a CHF clinic and during the intervening week at home through use of the TW telemedicine system (hereafter referred as TW group). The primary end point of the study was the titration time, which was the time from initiation of carvedilol to achieving the final dose of carvedilol. Although the frequency of titration was greater in the TW group, if the final dose of carvedilol was the same in both groups, then a shorter titration time would occur in the TW group only if remote titration of carvedilol was as effective as titrations which occurred during clinic visits.

The TW System is a telephone-based, automated, voice-interactive, 2-way store and forward telemedicine system, which was exempted from Food and Drug Administration regulation for clinical use in 2001 and has been described previously.^[20] TeleWatch version 1.0 was used in this study, which was written in Microsoft Visual Basic 6.0 (Redmond, Wash) with data stored in a Microsoft Access 2000 database. The software application resided on a Dell Optiplex GX 110 Mini-Tower PC (Round Rock, TX) which operated at 733 MHz and was accessed by patients through analog telephone lines connected to an Intel/Dialogic D/4PCI 4-line telephone interface board (Santa Clara, CA) (Figure 1). The database was accessible through a locally secured intranet with preconfigured client-server connectivity. Patients self-collected appropriate physiologic data (weight, pulse, blood pressure) using digital scales and Omron HEM 601 automated wrist sphygmomanometers (Bannockburn, IL). Patients then accessed the TW System by calling a telephone number which connected the patient with the computer application. Once the patient's identity was verified with a unique user ID and passcode, TW sequentially played 12 to 15 prerecorded questions which had numeric, yes/no, or multiple-choice answers to which the patient responded using the telephone keypad. These questions related to 4 main topics: physiologic parameters (weight, pulse, blood pressure); evidence of CHF exacerbation (dyspnea, lower extremity edema, paroxysmal nocturnal dyspnea); medication adherence; and medication side effects. The system analyzed the responses using a data validation algorithm to insure proper data entry, and abnormal data prompted the system to ask additional prerecorded questions in a rule-based fashion (eg, evidence of increasing fluid accumulation prompted questions about dietary intake, adherence with fluid restriction, and medication adherence). The system generated alerts to the study nurse when prespecified symptoms or physiologic changes were detected. Patients in the TW group were requested to call the system on a daily basis; however, patients in both groups were able to directly contact the study nurse.

Figure 1. 

TeleWatch system.

     

After randomization, treatment with carvedilol 3.125 mg bid was initiated during the enrollment clinic visit and was subsequently titrated in clinic visits or remotely with TW to doses of 6.25 mg bid, 12.5 mg bid, and then to a goal dose of 25 mg bid. Study participants from both groups were seen in the study clinic every 2 weeks during carvedilol titration (titration phase) and then monthly thereafter (follow-up phase). During the titration phase, the CO patients were eligible for carvedilol titration only during the biweekly clinic visits, and the TW patients were eligible for carvedilol titration on a weekly basis (during biweekly clinic visits and once a week on a specific day during the intervening week). On days when patients were eligible for carvedilol titration, clinical information was obtained from either the clinic visit or the TW system by the study nurse who summarized and presented data concerning heart rate and blood pressure as well as any symptoms related to carvedilol side effects in a standard format to one of the investigators without revealing the patient's name or group assignment. Using a predefined titration algorithm and while blinded to the patient's group assignment, the investigator used these data to make a titration decision which the study nurse then implemented. This titration algorithm dictated that carvedilol be up-titrated when SBP >100 mm Hg and pulse >65 beat/min in the absence of symptoms of hypotension (presyncope, increasing dyspnea, or worsening fatigue). If SBP was between 85 to 99 mm Hg, or the pulse was between 55 and 64 beat/min and the patient did not have any symptoms of hypotension, additional information regarding the patient's change in vital signs after the previous carvedilol titration was relayed by the nurse, and the titration decision was deferred to the cardiologist; however, at no time did the cardiologist have access to any data that would unmask the patient's group assignment. Aside from the frequency of possible titrations, the titration algorithm was the same for both groups.

If the patient's condition did not meet predefined titration parameters on the day when he/she was eligible for titration, no titration was attempted, and the patient was deemed to have failed a titration attempt. Similarly, if the carvedilol dose was increased and the patient was unable to tolerate the higher dose, the attempt was also considered a failure and the dose of carvedilol was decreased to the previously tolerated level. If on the day of potential carvedilol titration the patient was experiencing a transient, reversible medical condition (eg, an infectious process or poor oral intake), no titration was attempted, but it was not considered a titration failure and the patient was eligible for titration at the next scheduled time. After a patient failed 2 sequential titrations of the same dose of carvedilol, no further titrations were attempted, and the patient was monitored on a monthly basis in the follow-up phase of the study while taking the maximally tolerated dose (MTD) of carvedilol.

Exercise capacity at entry was measured by the 6-minute walk, as described by Guyatt et al.^[21] Quality of life and depression were, respectively, assessed by the Kansas City Cardiomyopathy Questionnaire^[22] and the Zung Depression Scale.^[23]

Statistical Analysis

To show a 50% reduction in the titration time compared with the CO group, a Wilcoxon rank sum test was used for sample-size calculations, which projected that 42 patients (21 in each group) would provide an 80% power at the a = .05 level of significance. To account for drop out, target enrollment was 50 patients (25 in each group).

Statistical analyses were performed using STATA (Version 8, STATA Corp, College Station, TX). Baseline and outcome variables were reported as mean (+SD). Differences in mean values were compared with 2-tail Student t test and reported as the mean difference with 95% CIs. Variables which did not have a normal distribution (defined as P < .05 by Shapiro-Wilk test) were reported as means (interquartile ranges) and compared with the Wilcoxon rank sum test. Binary data were reported as total number and percentage and were compared with the Pearson χ² test or the Fisher exact test when there were <5 events. Univariate regression analysis was performed to assess which demographic and clinical variables influenced the time to attain the goal dose or MTD. The final regression model was derived from backward stepwise regression, which initially included all the variables from the univariate analysis with the least statistically significant variable sequentially removed if it was not statistically significant (P > .051). Statistical significance was defined at the P < .05 value.

Results

Randomization began on November 1, 2001 and the study ended on October 30, 2003. Forty-nine patients were enrolled in the study, with 24 patients (49%) randomized to the CO group and 25 patients (51%) to the TW group. Table I presents the baseline characteristics of these patients. The average age was 54.6 (?15.8) years and the mean ejection fraction was 22.2% (?7.0). The groups were equivalent with respect to all enrollment variables. Two patients withdrew from the study, one each from the CO and TW groups, and 1 patient from the TW group was withdrawn by investigators because of a violation of eligibility criteria. None of the 3 patients who withdrew from the study suffered any adverse events.

Of the 46 patients who underwent carvedilol titration (Figure 2), 2 patients (4.3%), 1 from each group, were unable to tolerate the lowest dose of carvedilol. Fourteen patients (30.4%) were unable to achieve the dose of carvedilol 25 mg bid, 6 in the CO group, and 8 in the TW group (P = .54). Thirty patients (65.2%) were able to achieve the 25 mg bid carvedilol dose, 16 in the CO group and 14 in the TW group (P = .54). At the end of titration, there was no difference between the mean daily dose of carvedilol in the CO group and in the TW group ( Table II ).

Figure 2. 

Randomization schema.

     

TeleWatch Use

During the study, patients in the CO group who tolerated the initial dose of carvedilol were eligible for titration, an average of 5.1 (?1.4) times, before achieving the final dose of carvedilol, whereas the patients in the TW group who tolerated the initial dose of carvedilol were eligible for an average of 5.2 ? (1.8) possible titrations (P = .85). Of the possible titrations in the TW group, 1 occurred at enrollment, a mean of 2.3 (?1.2) occurred during subsequent clinic visits (TW clinic visits), and 2.0 (?0.9) occurred telephonically (TW telemonitoring) ( Table III ).

Based upon the titration algorithm, during the titration phase, carvedilol titration was attempted, respectively, during 84 (74.3%), 53 (73.6%), and 28 (62.2%) of the CO visits, TW clinic visits, and TW telemonitoring interactions (P = .28) ( Table III ). The success rate of these titrations in the CO visits, TW clinic visits, and TeleWatch telemonitoring interactions was between 89.3% to 92.8%, which was not statistically different (P = .83).

During the titration phase, patients in the TW group called into the telemonitoring system an average of 0.89 (+ 0.22) times a day. Overall, during the entire 3-month study, patients in the TW group made an average of 75.9 (+ 20.6) calls during 91.1 (?1.1) days of study enrollment for an average of 0.83 (?0.23) calls per patient per day of study enrollment.

Effect of Outpatient Monitoring on Duration of Carvedilol Titration

The time to achieve the goal or MTD of carvedilol in the CO group was 63.7 days (?20.2), whereas in the TW group, it was 33.6 days (?16.6), with the difference of 30.1 days being statistically significant (P < .0001). In the 30 patients who achieved the carvedilol dose of 25 mg bid, the mean titration time was 62.2 (?18.6) days in the CO group and 34.1 (?13.6) days in the TW group with the TW patients achieving goal dose a mean of 28.0 days sooner than the CO patients (95% CI 15.7-40.4, P < .0001). In patients unable to achieve the dose of carvedilol 25 mg bid, the mean titration time in the CO group was 67.8 days (?25.7), whereas it was 32.5 days (?22.0) in the TW group, with a statistically significant difference of 35.3 days (95% CI 7.5-63.1, P = .02).

As shown in Table IV , simple correlations showed that only group assignment was correlated to the titration time with univariate analysis. The number of titration steps required to achieve the final dose of carvedilol (ie, final dose of 3.125 mg bid would be 1 titration step; 6.25 mg bid would be 2 steps, etc, to a maximum of 4 steps for 25 mg bid regardless of the number of attempts required to achieve this dose of carvedilol, etc) was not correlated to the titration time on either univariate or multivariate analysis. The final regression model showed that group assignment, presence of a pacemaker, and NYHA class were independently associated with titration time. In this model, patients randomized to the TW group achieved the final dose of carvedilol 32 days sooner than did CO patients after adjusting for presence of a pacemaker and NYHA class.

Safety

There were 5 serious adverse events in the study ( Table V ), 3 of which occurred during the titration phase and 2 during the follow-up phase, there being no statistically different event rate between the CO and TW group (P = .29). Of the 2 serious adverse events in the TW group during the titration phase, 1 patient developed decompensated heart failure 7 days after carvedilol was initiated during the enrollment clinic visit and before any remote titrations were attempted. Of the 2 serious adverse events in the TW group during the follow-up phase, 1 was due to a myocardial infarction that occurred on the last day of the study in a patient who was awaiting a cardiac catheterization prompted by a positive stress test, and the other serious adverse event occurred in a patient whose last adjustment of carvedilol was 23 days earlier. None of the serious adverse events in the TW group were deemed to be the result of the speed of the carvedilol titration.

In 2 of the 4 instances where patients in the TW group required hospitalization, remote monitoring prospectively identified clinical deteriorations, and outpatient adjustment of medications was attempted. All adverse events occurred in patients receiving a carvedilol dose of 3.125 mg bid.

Discussion

This is the first study to show that an automated telemedicine system can facilitate titration of a vasoactive medication in patients with CHF as effectively as traditional clinic visits. This study also demonstrates that carvedilol can be successfully initiated, as well as the maximum tolerated dose reliably and rapidly achieved in patients with CHF using a titration algorithm, in combination with intensive outpatient monitoring provided by either biweekly clinic visits alone or in combination with home monitoring provided by an automated telemedicine system. Although the study design permitted more frequent carvedilol titration attempts in the TW group, this translated into a significantly shorter titration time only because the success of clinic-based and remote-telemedicine carvedilol titration were similar.

Achieving Goal Dose

Ninety-six percent of patients enrolled in this study tolerated carvedilol therapy, which is comparable with the 94% tolerance reported in the US Carvedilol Study.^[1] Furthermore, titration of carvedilol using daily home telemonitoring with the TW System or biweekly clinic visits resulted in 65% of patients achieving the goal carvedilol dose of 25 mg bid (69.6% of patients in the CO group and 60.8% of patients in the TW group). This finding is consistent with the 66% to 80% of patients achieving the goal carvedilol dose reported in previous CHF clinical trials^[1,6] but represents a substantial improvement compared with epidemiologic studies of routine outpatient β-blocker therapy in patients with CHF. For example, the IMPACT-HF trial investigators report that after hospital discharge of patients with stabilized CHF, only 11.5% of patients whose β-blocker was initiated in the hospital and only 3.2% of patients in the usual care group whose β-blocker was initiated and titrated at the discretion of their outpatient physicians achieved goal β-blocker dose within 60 days.^[12]

One of the methods used to increase the percentage of patients with CHF who receive goal dose of carvedilol and other β-blockers has been to increase the frequency and number of interactions between patients and healthcare providers. In one study, specially trained nurses working in a dedicated CHF clinic were able to increase the percentage of patients receiving the goal dose of β-blocker therapy from 10% to 43% by interacting with patients more frequently and implementing accepted practice guidelines.^[24] More intense outpatient monitoring of patients with CHF by nurses who interacted with patients by telephone 3 times a week resulted in 71% of patients achieving goal dose of carvedilol.^[25] This is consistent with the 69.6% of the CO patients in this study who achieved goal carvedilol dose while being monitored with biweekly clinic visits. However, outpatient monitoring strategies, which rely upon frequent patient-provider interactions, are labor intensive, expensive to operate, and difficult to expand to larger populations. For these reasons, an automated home telemedicine monitoring system such as TW, which can achieve an equivalent percentage of patients on goal dose of carvedilol as can biweekly clinic visits, combines the benefits of intensive outpatient monitoring with a highly scalable and efficient outpatient monitoring capability.

Rate of Titration

Because carvedilol has been shown to improve LV systolic function and reduce mortality in a dose-dependent manner,^[7] there is a possibility that achieving the MTD of carvedilol more rapidly may have beneficial effects. Whether there are demonstrable clinical benefits of achieving the MTD of carvedilol in 30 days is currently unknown, and this study was not powered to detect the clinical benefit of more rapid carvedilol titration. Nevertheless, in this study, the percentage of patients who were able to achieve the goal dose of carvedilol, both the biweekly CO group and the TW group were able to achieve the goal and MTD much more rapidly than previously reported. In the COPERNICUS trial, which enrolled patients with class IV heart failure, 66% of enrollees achieved the target dose of carvedilol within 12 weeks.^[6] Similar titration times were reported in retrospective studies of outpatient carvedilol titration, in which between 82 to 89 days were required to achieve the final dose of carvedilol.^[15,26] As was the case with attaining the goal dose of carvedilol, increased outpatient interactions with healthcare providers reduced the time required to achieve that dose. In a prospective study in which patients called an advance practice nurse 3 times a week and titrations were performed every 2 weeks, the target carvedilol dose was achieved in 71% of patients within 7.8 weeks (54.6 days).^[25] Compared with previously reported results, the titration time of 32.4 days in the TW group represents a substantial decrease in the time required to achieve final dose of carvedilol (Figure 3).

Figure 3. 

Titration time required to achieve final dose of carvedilol.

     

In this study, the percentages of patients achieving a goal dose of carvedilol in both the CO and TW groups were higher than that reported in routine practice, and the times required to achieve final dose of carvedilol were also significantly reduced, particularly in the TW group. Three components were responsible for these improvements. First, carvedilol titration was an explicit goal of the study and regular feedback was provided to physicians regarding the adequacy of each patient's β-blockade. Second, experienced clinicians who used an evidence-based titration algorithm helped overcome a reluctance to increase carvedilol in patients with borderline blood pressures and heart rates. Finally, close outpatient monitoring was provided to the patients in the form of either biweekly clinic visits or outpatient telemonitoring combined with the biweekly visits.

Although the same titration algorithm was used in both groups, the more rapid titration in the TW group was dependent upon the monitoring provided by the telemonitoring system. Although it is conceivable that weekly clinic visits would have produced a titration rate equivalent to that achieved in the telemedicine group, weekly clinic visits are not a practical option for most patients and clinic practices. For instance, in the Veterans Administration population, patients with CHF were seen in a medical clinic between 8.6 to 9.6 times a year, or approximately 1 visit every 5 to 6 weeks.^[27] Therefore, scheduling patients for weekly ambulatory care visits to titrate medications would necessitate a substantial increase in CHF clinic capacity.

Similarly, patients in the TW group were seen in clinic on a biweekly basis, which, as previously described, represented more frequent clinic visits than what is likely practical for larger populations. However, this study demonstrates that telemedicine monitoring facilitates the same number of medication titration opportunities, which have similar rates of success and comparable safety with fewer outpatient clinic visits. This suggests that carvedilol titration could have been done primarily with telemonitoring with fewer than biweekly clinic visits. These results also suggest that outpatient adjustment of medications for other chronic diseases might be successfully performed with an automated telemedicine system.

Although previous studies have shown that telephonic monitoring and remote titration of carvedilol is feasible,^[25] such an approach is very labor-intensive and limits the number of patients who can be monitored at any given time. However, automated telemedicine systems such as TW can monitor multiple patients simultaneously, and such systems can feasibly manage larger patient populations than what is possible within the logistical and financial constraints of the existing outpatient healthcare delivery system. In this study, 38.5% of the potential titrations in the TW group (2.0 [?0.9] of 5.2 [?1.6] titrations) were performed using the telemedicine system, which highlights the logistical and operational advantage of such automated systems compared with nurses manually making outgoing calls to patients.

Telemonitoring Adherence and Impact on Clinical Care

Patient adherence to calling into the TW system was very high: 0.89 calls per patient per day during the titration phase and 0.83 calls per patient per day during the remainder of the study period. The adherence with outpatient telemonitoring has not been well investigated; however, the results from this study are similar to the average daily rates of 75% to 94% reported in other telemedicine heart-failure studies.^[28,29]

The rate of success of remote telemedicine carvedilol titration was 89.3%, which did not differ statistically from the success of carvedilol titration observed in clinic visits of the CO group (92.8%) or in the clinic visits of the TW group (92.4%). This is an important finding, which was responsible for the shorter titration time in the TW group. The equivalent success of clinic-based and remote medication titration has implications for other medications which have potential hemodynamic or other physiological side effects and suggests a novel method of adjusting medical therapy using remote monitoring in the place of in-person clinic visits.

Of the 4 serious adverse events in the TW group, clinical deteriorations were prospectively detected by TW in 2 cases before the patients contacted a healthcare provider. None of these events were related to the speed with which the carvedilol was titrated, and it is of note that all the serious adverse events occurred in patients receiving the lowest dose of carvedilol. Heart failure symptoms including weight gain, lower extremity edema, and increasing dyspnea are generally present a mean of 8 to 12 days before hospitalization.^[30] Because patients contacted TW an average of 5 to 6 times a week, it is not surprising that the telemedicine system detected these symptoms before the patient's clinical condition deteriorated to the point of requiring hospitalization. The inability to prevent subsequent hospitalizations for these 2 events detected prospectively by TW does not diminish the monitoring capability of telemedicine, as early intervention, prompted by TW, likely prevented further clinical deteriorations which may have been associated with more profound outcomes. Furthermore, the fact that in 3 of the 4 serious adverse events in the TW group, patients received more intense outpatient management after TW detecting abnormal clinical data highlights the surveillance utility of automated telemonitoring to identify patients at risk for clinical deterioration from a large population and to focus healthcare resources on them. In this way, healthcare personnel and resources are used most efficiently while at the same time enhancing the quality of care delivered.

Finally, it must be recognized that this study did not use the full capability of the TW system because the requirements for physician blinding necessitated review of TW data only from the day when the patient was eligible for outpatient titration. Although the impact of integrating information from prior days in the decision-making process cannot be known, it is likely that this would have assisted the physician in making more informed titration decisions.

Limitations

Patients were excluded from the study if they received β-blocker therapy within the previous 6 months. Because β-blocker therapy is a mainstay in the management of coronary artery disease, this prohibition significantly reduced the number of study participants who had coronary artery disease (20% ischemic cardiomyopathy in study cohort compared with 48% in the US Carvedilol Study).^[1] In addition, the patients enrolled in this study were approximately 3.4 years younger (mean age 54.6 ? 15.8 years) than observed in the US Carvedilol Study (mean age 58.0 ? 12.2 years)^[1] and significantly younger than reported in the Coreg Heart Failure Registry (mean age 67 ? 13 years).^[15] This shift toward younger patients is likely due to the increased number of patients with nonischemic cardiomyopathy as well as newly diagnosed patients with coronary artery disease who had not yet been placed on a β-blocker. However, increasing age has not been associated with decreased tolerance of carvedilol,^[26] and it is therefore likely that the benefit derived from telemonitoring observed in this study would be similar in an older patient population.

Additional limitations to this study include the fact that although it enrolled a diverse patient population, the participants were drawn from 2 closely affiliated medical institutions and that all patients received more intensive care delivered by highly specialized caregivers than what is routinely available to patients with CHF. Finally, because of the small sample size, it is not possible to make a definitive statement regarding safety.

Conclusions

In summary, both TW monitoring and biweekly clinic visits substantially increased the number of patients with NYHA class II and III CHF who achieved the carvedilol dose of 25 mg bid as compared with typical patients with CHF receiving usual care reported in epidemiologic studies. However, the TW telemedicine system used in conjunction with a titration algorithm significantly decreased the time required to achieve the final carvedilol dose by 30.1 days, compared with biweekly clinic visits only. Because the beneficial effects of carvedilol on LV systolic function are dose-dependent, this more rapid titration has clinical implications, but this study was not powered to detect the magnitude of this clinical benefit. However, more importantly, the ability to successfully perform these titrations remotely as successfully as in-clinic visits has implications not only for carvedilol titration in patients with CHF, but also for a variety of medications that have benefits that are dose-dependent but also have potentially significant but measurable side effects. These results suggest that automated telemedicine systems can improve the pharmacologic management of patient populations with CHF and other chronic illnesses.