Modeling changes in resting heart rate and its relationship to all-cause mortality in humans
| dc.contributor.advisor | Derman, Wayne | en_ZA |
| dc.contributor.advisor | Heine, Martin | en_ZA |
| dc.contributor.advisor | Nyabadza, Farai | en_ZA |
| dc.contributor.author | Ristow, Brandon Mark | en_ZA |
| dc.contributor.other | Stellenbosch University. Faculty of Medicine and Health Sciences. Dept. of Biomedical Sciences: Medical Physiology. | en_ZA |
| dc.date.accessioned | 2021-06-07T10:49:36Z | |
| dc.date.available | 2021-06-07T10:49:36Z | |
| dc.date.issued | 2021-03 | |
| dc.description | Thesis (MMed)--Stellenbosch University, 2021. | en_ZA |
| dc.description.abstract | Background Resting heart rate (RHR) is an easily measurable proxy of cardiovascular and general health. Extensive cohort studies have indicated that a higher RHR among individuals, results in an increased risk of all-cause mortality. RHR has been found to change over time based on health and disease states, however it is still unknown whether a change in RHR will impact the relative risk of mortality and what lifestyle choices contribute the most to long term changes in RHR. Therefore, in this thesis we set out to I) understand the relative all-cause mortality risk associated with an increasing RHR over time; II) understand if a mechanism exists to standardise the way in which RHR is measured among studies; and III) model the daily changes in RHR using lifestyle proxies. Methods In chapter 2, a meta-analysis was performed to identify studies reporting the relative risk of all-cause mortality associated with long-term changes in RHR. These studies were required to have cohorts of adult humans with no known disease, RHR measurements taken at least one year apart, and follow-up for all-cause mortality over a minimum of one year. Relative risk of participants with increasing and decreasing RHR were compared to those with an unchanged RHR. Chapter 3 analyses a set of wearable device data from a representative cohort and determines the daily RHR as well as lifestyle proxies related to sleep and physical activity. Using these measures, daily changes in RHR were forecast under three different methods (population-level, grouped level and individual level) using generalized additive models created with the lifestyle proxies. Results In chapter 2, nine studies fulfilled the criteria for the meta-analysis, with a total of 108 625 participants. It was found that an increased RHR over time had a hazard ratio of 1.21 (95% CI 1.09-1.35, p<.01) compared to an unchanging RHR. When an unchanged RHR was compared to a decreased RHR over time, no significant results were found. The modeling of RHR at a population-level resulted in a best model accuracy of R2 = 0.039. When attempting to first cluster individuals into groups before creating multiple models, the best model accuracy was R2 = 0.114. When creating unique model’s for each individual, the best model’s had a mean accuracy of R2 = 0.426. It was found that a daily point measurement of RHR using heart rate values at the time of waking resulted in the best model accuracy. Conclusion A RHR that is increasing over time will lead to an increased risk of all-cause mortality, however it is uncertain what effect a decreasing RHR has on mortality risk. When attempting to understand the lifestyle mechanisms that influence changes in RHR, it is best to approach things from an individual level, as there exists a great deal of variability between people. While there is no known method for modeling next day RHR for all individuals, it is feasible for some. | en_ZA |
| dc.description.abstract | Agtergrond Rustende harttempo (RHT) is ‘n eenvoudig meetbare waardenorm wat as goeie aanduiding vir kardiovaskulêre en algemene gesondheid dien. Omvattende kohortstudies dui daarop dat individue wat verhoogde RHT toon, ook ‘n verhoogde risiko vir oorkoepelende mortaliteit loop. Gesondheids- en siektetoestande lei tot ‘n verandering in RHT met die verloop van tyd, maar dit is egter steeds onbekend of ‘n verandering in RHT a) die relatiewe risiko vir mortaliteit sal beïnvloed en, b) watter leefstylkeuses mees beduidend tot langtermyn veranderinge in RHT bydra. Metodes Hoofstuk 2 is gewy aan ‘n meta-analise, met die doel om studies wat die relatiewe risiko vir oorkoepelende mortaliteit wat met langtermyn veranderinge in RHT verband hou, te identifiseer. Die betrokke studies moes voldoen aan die volgende vereistes: a) kohorte bestaande uit volwasse mense wat aan geen bekende siektes ly nie en, b) RHT metings moes ten minste een jaar uitmekaar geneem word, met die opvolgmeting vir oorkoepelende mortaliteit oor die verloop van minstens een jaar. Die relatiewe risiko van deelnemers met toenemende en afnemende RHT is met dié van deelnemers met onveranderde RHT vergelyk. Hoofstuk 3 is gewy aan die ondersoek van data, insluitend geteoritiseerde daaglikse metings vir RHT en leefstyl waardenorme, wat van ‘n kohort bestaande uit 933 individue, deur middel van draagbare toestelle, verkry is. Hierdie metings is gebruik om daaglikse veranderinge in RHT te voorspel deur die gebruik van veralgemeende toevoegings modelle wat deur middel van die genoemde leefstyl waardenorme geskep is. Resultate Soos gedokumenteer in hoofstuk 2 het nege studies, met ‘n totaal van 108 625 deelnemers, aan die kriteria van die meta analise voldoen. Dit is bevind dat ‘n toememende RHT met die verloop van tyd ‘n gevaarverhouding van 1.21 (95% Cl 1.09- 1.35, p<.01) getoon het, in vergelyking met ‘n onveranderde RHT. Wanneer ‘n onveranderde RHT vergelyk is met ‘n RHT wat afneem met die verloop van tyd, is geen meningsvolle verskille gevind nie. Die modellering van RHT op populasievlak het ‘n beste model akkuraatheid van R2 = 0.039 opgelewer. Wanneer individue in groepe saamgebondel is voordat veelvuldige modelle geskep is, was die beste model akkuraatheid R2 = 0.114. Wanneer ‘n unieke model vir elke individu geskep is, het die beste modelle ‘n gemiddelde akkuraatheid van R2 = 0.426 getoon. ‘n Daaglikse puntmenting van RHT by ontwaking het die beste model akkuraatheid opgelewer. Gevolgtrekking ‘n RHT wat toeneem met die verloop van tyd, lei tot ‘n verhoogde risiko van oorkoepelende mortaliteit. Dit is egter onduidelik wat die presiese effek van ‘n afnemende RHT op mortaliteitsrisiko is. In ‘n poging om die leefstylmeganismes wat die veranderinge in RHT beïnvloed te verstaan, is dit raadsaam om elke persoon se datastel individueel te benader, vanweë die geweldige variansie tussen verskillende mense. Alhoewel daar geen bekende metode vir die modellering van opeenvolgende-dag RHT vir alle individue bestaan nie, is dit wel in sommige gevalle haalbaar. | af_ZA |
| dc.description.version | Masters | en_ZA |
| dc.format.extent | 65 pages | en_ZA |
| dc.identifier.uri | http://hdl.handle.net/10019.1/110555 | |
| dc.language.iso | en_ZA | en_ZA |
| dc.publisher | Stellenbosch : Stellenbosch University | en_ZA |
| dc.rights.holder | Stellenbosch University | en_ZA |
| dc.subject | Humans -- Mortality | en_ZA |
| dc.subject | UCTD | en_ZA |
| dc.subject | Heart rate monitoring | en_ZA |
| dc.subject | Cardiovascular system | en_ZA |
| dc.subject | Resting heart rate | en_ZA |
| dc.title | Modeling changes in resting heart rate and its relationship to all-cause mortality in humans | en_ZA |
| dc.type | Thesis | en_ZA |