Study population
Our study population were obtained from National Health and Nutrition Examination Survey 1999-2002 (NHANES), which is an observational study of noninstitutionalized US civilian and a typical sample of US population. The National Center for Health Statistics of the Centers for Disease Control and Prevention has conducted the study. The National Center for Health Statistics Institutional Review Board approved all the NHANES study protocols and details of its protocols and consents are available on the NHANES website. We excluded individuals without data of modified oral frailty, comorbidities, muscle strength, muscle mass, telomere length and recreational activity. A total of 8,382 suitable participants were initially recruited in our study.
Components of modified oral frailty
A longitudinal study in Japan defined oral frailty as containing 3 or more of 6 characteristics which included the number of natural teeth, tongue pressure, articulatory skill, chewing ability and perceived eating and swallowing difficulties [21]. Base on the previous study, we modified oral frailty according to NHANES databases. In this study, the modified oral frailty is composed of 3 components, limited eating ability, dry mouth and difficult swallowing, which were determined by self-reported questionnaires. Participants were asked the questions “How often do you limit the kinds or amounts of food you eat because of problems with your teeth or dentures?“, “Do you have difficulties swallowing any foods?“ and “Does your mouth feel dry when you eat a meal?”. Participants responding ‘always, very often and often’ were categorized as limited eating ability; responding ‘yes’ as dry mouth or difficult swallowing.
Covariables
Sociodemographic covariables including age, sex, race, smoking history and comorbidities were collected from self-reported questionnaires. Smoking status was assessed by the questions “Have you smoked at least 100 cigarettes in your entire life?”. Past medical comorbidities included several cardiovascular diseases, arthritis and stroke. Recreational activity was defined as self-reported participation in moderate intensity exercise which increased pulse or breathing rate, like swimming, jogging or brisk walking for more than 10 minutes without stopping. The biochemistry profiles (creatinine, alanine aminotransferase, serum fasting glucose, total cholesterol and total calcium) were analyzed with standardized guidelines and protocols. The analysis of telomere length which compares the telomere length of participants relative to standard reference DNA (T/S ratio) was using by quantitative polymerase chain reaction (qPCR). Detailed information is available from the NHANES website.
Isokinetic strength testing of right quadriceps muscle assessing by A Kin Com MP dynamometer (Chattanooga, TN) was used to determine muscle strength. Appendicular skeletal muscle mass assessing by dual-energy X-ray absorptiometry (DEXA) QDR-4500A Hologic scanner (Bedford, MA) was used to determine muscle mass. Body composition measures including bone mineral content, fat mass and lean muscle mass was determined by DEXA.
Outcome assessment
The primary outcome of our study was all-cause mortality. The mortality status was obtained from the time of the study enrollment to death or 31 December 2006 through probabilistic matching between NHANES database and National Death Index death certificate records. The secondary endpoint was the association among muscle strength, muscle mass and telomere length with each component of modified oral frailty.
Statistical analysis
SPSS version 18 (SPSS Inc., Chicago, IL, USA) was used for executing all statistics. The one-way ANOVA and chi-square test were applied for analyzing socio-demographic characteristics, laboratory variables and medical comorbidities. The threshold of significance level was two-sided p values < 0.05. Kaplan-Meir curve stratified modified oral frailty was plotted. When the participants without any component of modified oral frailty were regarded as reference, the relationship between modified oral frailty and all-cause mortality was assessed by performing Cox proportional hazard models. Multivariate regression analysis was performed for the association among muscle strength, muscle mass and telomere length with each component of modified oral frailty. Three covariate-adjusted models were used for potential confounders: Model 1 = unadjusted; Model 2 = age, race, gender, creatinine, alanine aminotransferase, serum fasting glucose, total cholesterol and total calcium; Model 3 was Model 2 and adjustment for smoking, recreational activity, and past medical histories.