Author + information
- Angelo Auricchio, MD, PhD and
- Catherine Klersy, MD, MSc* ()
- ↵*IRCCS Fondazione Policlinico San Matteo, Biometry & Clinical Epidemiology, viale Golgi 19, Pavia, PV 27100, Italy
We thank Drs. Winkler and Koehler for their interest in our report (1) regarding the assessment of effect of remote patient monitoring on the outcome of chronic heart failure patients. We appreciated their provocative thoughts about unmet needs in structured disease management program (e.g., identification of the patients who most likely benefit from the technology, determinants of outpatient responsiveness, what makes interventions effective), which we all share; however, we are afraid that none of these questions might have found an answer, given the lack of published data. As far as the differences in outcome of 2 of the most recently published studies and the results of our meta-analysis are concerned, we believe that they are much less than what Drs. Winkler and Koehler perceived.
The study by Mortara et al. (2) showed a similar outcome between usual care and remote monitoring (indicated in the study as home telemonitoring). However, patients in the Mortara et al. study were at least 5 to 10 years younger than those included in our meta-analysis, and they were in a much lower New York Heart Association (NYHA) functional class (ranging from 34% to 49% in NYHA functional class >3 compared with 54% [randomized controlled trials] and 83% [observation cohorts] in our meta-analysis). Moreover, there was an unexplained imbalance, as already emphasized by Mortara et al., in baseline characteristic in the large Polish cohort as indicated by a more advanced NYHA functional class, significantly lower left ventricular ejection fraction, higher dyspnea score, and much lower sodium plasma level for those patients assigned to home telemonitoring. A post hoc analysis revealed a highly significant interaction between home telemonitoring and country in the association with the number of hospital stays (p = 0.004) and in the combined end point of cardiac death and heart failure hospital stay (p = 0.004). If one would put in perspective the outcome of the Italian cohort of the study by Mortara et al. with the results of our meta-analysis, an impressive similar benefit of remote monitoring compared with usual care would be found.
The study by Dar et al. (3) was a small, prospective, randomized controlled study including 182 patients randomized to usual care versus home monitoring. Although the baseline demographic characteristics of these patients were similar to those reported in the studies included in our meta-analysis, only 74 patients in the home monitoring arm and 79 patients in the usual care completed 180-day follow-up. Thus, the relative weight of the study by Dar et al. (3) in our meta-analysis would be relatively low, and importantly, approximately 50% of the studies we meta-analyzed had a similar duration of follow-up. Of note, there were 14 deaths in the home monitoring arm and only 4 deaths in the usual care group. Overall, this was an extremely high death rate for a very short follow-up but also impressively different between 2 treatments. We were not able to find any comparative study in our meta-analysis reporting similar death rates, which let us question about the reasons (not addressed in the study). Indeed, the death rate at 12 months in all randomized controlled studies we meta-analyzed ranged from 14.1% (95% confidence interval [CI]: 12.8 to 15.4) to 11.7% (95% CI: 10.7 to 12.9) in the usual care and home monitoring arm, respectively, and in observational studies it ranged from 13.0% (95% CI: 10.9 to 15.3) to 6.8% (95% CI: 5.3 to 8.6).
In conclusion, there is no doubt that the studies by Mortara et al. (2) and Dar et al. (3) both represent important contributions to implementation of remote monitoring. Their relative weight needs to be defined in future meta-analysis, keeping in mind some important methodological issues of each study.
- American College of Cardiology Foundation