Normal Respiratory Rate and Peripheral Blood Oxygen Saturation in the Elderly Population

Age-specific normal limits for a number of vital signs and physiological parameters have not been established in the elderly population. The limits for younger adults are not always applicable because of age-associated physiological changes and the increase of interindividual differences with age.1 Regarding the respiratory system, there are few data on normal respiratory rate at rest (RR) and peripheral pulse oximetry values (SpO2), which are major parameters in clinical practice and easy to measure, and become altered quickly in respiratory and cardiac diseases. (Increased respiratory rate is often the only visible sign of a respiratory infection.)2, 3 This was a cross-sectional study of 791 noninstitutionalized individuals aged 65 and older living in Spain to establish the limits of normal RR and SpO2 in the elderly population. The sample was collected using multistaged probabilistic sampling and stratified according to sex, size of place of residence (rural, urban, or big city), and geographic location with a nonproportional age stratum (523 subjects aged ≥80). A sample of 576 participants was considered necessary to estimate RR and SpO2 with 5% error and a design effect of 1.5. Survey data were collected between 2007 and 2009. The survey was carefully designed to reduce nonsampling errors, the survey takers received specific training, and the field work was thoroughly supervised. RR and the SpO2 were measured with the participant in a seated position after a rest of at least 10 minutes. SpO2 was measured using a pulse oximeter (9500; Nonin Medical, Plymouth, MN), and RR was measured by directly observing thoracic movements for a 30-second period. As a distraction maneuver, the survey takers pretended to measure the radial pulse, so that participants would not be aware that their respiratory rate was being measured.3 All information about participants' medical background was collected as control variables. Two consecutive analyses were conducted. First, all participants with pathologies that proved to affect RR or SpO2 independently in multivariable models were excluded. A subsequent more-restricted analysis was performed by excluding all individuals who had any clinical factor showing significant influence in bivariate analyses. Participants with dyspnea during the examination were excluded from all calculations. Normal RR limits were represented according to percentiles that delimit 95% of the sample (2.5–97.5) and percentiles that delimit 99% of the sample (0.5–99.5). Limits of SpO2 were represented according to the first and fifth percentiles. Calculations were weighted according to age, sex, and size of place of residence. History of chronic obstructive pulmonary disease (COPD) was the only variable that independently influenced RR and SpO2 in the multivariate models. Once individuals with COPD were excluded, the RR distribution appeared bell-shaped, with 0.67 kurtosis and 0.43 asymmetry, and was significantly different from the theoretical normal distribution according to the Kolmogorov contrast test. Percentiles 2.5 and 97.5 were 12 respirations per minute (rpm) (95% CI = 10–12 rpm) and 28 rpm (95% CI = 28–32 rpm), respectively. Percentiles 0.5 and 99.5 were 8.5 rpm (95% CI = 8–10 rpm) and 32 rpm (95% CI=32–32 rpm), respectively. These values did not change significantly in the second more-restrictive analysis (Table 1). Detailed data are shown in Table 1 according to sex and age. The oxygen saturation distribution was not normal; the fifth percentile was 91% (95% CI = 89.5–92.0%), and the first percentile was 85% (95% CI = 80.0–87.8%). The results of the restrictive analysis and detailed results according to age and sex are shown in Table 2. The reported RR values were higher than those established for younger populations. Earlier articles reported RRs of up to 25 in institutionalized populations, with 28 to 30 rpm as a criterion for considering severe pneumonia.3-5 The current results, together with the values previously reported, suggest that 28 rpm could be considered the limit between normal respiratory rate and tachypnea. Regarding SpO2, a previous study6 reported a 92% fifth percentile, which is similar to the results of the current study, although that sample was significantly younger than that of the current study (included participants aged 40–79). Although the difference in RR between the two methods of analysis were not clinically relevant, the SpO2 results improved some in the more-restrictive analysis, making it advisable that the latter values be considered the normal limit for SpO2. In summary, elderly adults have a higher RR than younger ones, with the limit for tachypnea established as 28 rpm. The limit of normal for SpO2 was 91%, similar to that of the younger population. We would like to thank Dr. Antonio Yuste for the hard work done in research management and administration. We also thank Natalia Gonzalo and Esther Valldosera for their invaluable work in data collection and quality control during field work. Conflict of Interest: The authors have no commercial associations or sources of support that might pose a conflict of interest. This work was supported by the European Commission within the 6th Framework Programme (project CAALYX (IST-2005–045215)) and the Spanish Geriatric and Gerontology Society. Author Contributions: Rodríguez-Molinero conceived and designed the study, contributed to field work supervision, contributed to data analysis, and drafted and reviewed the submitted material for publication. Narvaiza contributed to design and data analysis and reviewed the submitted material for publication. Ruíz performed the statistical analyses and reviewed the submitted material for publication. Gálvez-Barrón contributed to quality control, database preparation, and statistical analysis and reviewed the submitted material for publication. Sponsor's Role: The sponsors had no role in the design of the study, methods, recruitment of subjects, data collection, statistical analysis, or preparation of paper.