After ensuring that there was no similar systematic review on PROSPERO or the Cochrane Library, a systematic search by title, abstract and keywords was conducted in the Cochrane Library, EBSCO, PubMed, Scopus and Web of Science in September 2018. Additional searches were performed in weekly automatic updates retrieved from the databases until December 2019. References of each included study were hand searched for potentially eligible studies. Appendix A (supplementary material) reports the full search strategy.

After completion of databases search, all duplicates were removed. Then, one author screened each article for the scope of the review by their title, abstract and keywords. In parallel, an independent researcher screened 10% of all abstracts for eligibility to be included in the study. Full text of the articles was assessed, and papers excluded according to the eligibility criteria. Studies were included if: (1) included healthy adults (≥18 years), (2) developed at least one predictive equation for either Pimax or Pemax (mouth pressures), and (3) were written in English, French, Portuguese or Spanish. Studies were excluded if they were qualitative studies, research protocols, thesis/dissertations, abstracts, letters to the editor, news, case studies, book chapters, guidelines, position papers and unpublished work.

In cases of uncertainty, the decision to include/exclude the article was debated between the two reviewers and a third member was consulted to reach consensus.

Data were extracted to a predesigned structured table with author's name, year and country, participant characteristics (sample size, number of women and men, age and body mass index (BMI)), smoking status, equipment and protocol used to assess Pimax and Pemax, established predictive equation(s), lower limits of normality (LLN), standard error of estimate (SEE) and coefficient of determination (R2). The table was built in an excel file format to facilitate calculations when needed and then transformed and simplified to a word format.

A second table was built to aid visualization of the equations found, their explanation coefficients and which ones complied with the American Thoracic Society/European Respiratory Society standards (Pimax measured at residual volume, Pemax measured near total lung capacity, use of a flanged mouthpiece, use of noseclip not mandatory, holding pressure for ≥1.5s but not much longer, avoid use of aneroid manometers, patients supports cheeks during maneuver, 3 maneuvers with less than 20% variability).20

The accuracy of the extracted data was verified by two reviewers and confirmed by all authors. All corresponding authors of the included studies were contacted by e-mail in case of missing data.

Two reviewers independently assessed the quality of the included studies with the Quality Assessment of Diagnostic Accuracy Studies (Quadas-2).27 This scale has two dimensions (risk of bias and applicability concerns) and four domains (patient selection, index test, reference standard and flow and timing) that are scored with unclear risk, low risk, or high risk.27 Consistency of the quality assessment performed by the two reviewers was explored with the inter-rater agreement analysis using Cohen's kappa through SPSS statistics (IBM, version 25.0) and interpreted as <0: poor agreement; 0.00–0.20: slight agreement; 0.21–0.40: fair agreement; 0.41–0.60: moderate agreement; 0.61–0.80: substantial agreement; 0.81–1.00: almost perfect agreement.28

The databases search identified 203 records. Eleven additional records were identified through hand searches of references of published articles and systematic reviews on the topic. After removing duplicates, 197 studies were screened. Of these, 164 studies were excluded by title, abstract or keywords, as they did not comply with the inclusion/exclusion criteria. Of the remaining 33 studies, 13 were excluded due to: unavailability of full-text (n=3),1–3 no equations for Pimax or Pemax being developed (n=9),4–12 and being a review of the literature (n=1).13 Twenty studies were included.24,25,29–46 A PRISMA flow diagram can be found in Fig. 1.

Figure 1.

PRISMA flow diagram of systematic search.

*Authors were contacted and no response was obtained.

(0.37MB).

A detailed view of the quality assessment is presented in Table B.1 (Appendix B). Overall, the quality of studies was good. Risk of bias was high in eight studies,24,30,34,35,39,41,43,45 mainly due to inclusion of participants by convenience, and in three studies the eligibility criteria was not clear.29,31,37 Risk of bias in the index test was high in one study.41 Concerns with applicability were unclear for patient selection and index test in eleven studies.24,29–31,34,36–38,41,45,46 Only three studies showed unclear risk for applicability concerns regarding the reference standard.31,32,43 Quality assessment of the studies revealed substantial agreement between the two reviewers (kappa=0.74, 95% CI 0.59–0.86).

A detailed description of the studies can be found in Table 1. The 20 studies included a total of 9643 healthy individuals (5146 women, 4497 men). Most studies included young and older adults with ages between 10 and 90 years old (n=16),24,25,29–31,34–40,43–47 one study had only young adults (18–34 years old)41 and three studies had only middle-aged and older adults (45–85 years old).32,33,42 Seventeen studies reported the BMI, which ranged from 18 to 41.7kg/m2.24,25,30,32–36,38–46 Twelve studies excluded smokers.24,25,32–35,37,42–46 Seven included smokers and reported no differences in maximum respiratory pressures between smokers and non-smokers,29–31,36,38–40 and one study did not report this characteristic.41

Of the 20 studies, ten used digital manometers,24,36,37,39,40,42–46 eight used mechanical gauges,25,29–35 one used a gauge associated with a pneumotachograph,38 whilst in another the gauge was associated with a spirometry system.41 Assessment protocol varied across studies. The vast majority of studies performed measurements with the participants in a sitting position (n=18)24,25,29–35,37–40,42–46 and using a nose clip.24,25,29,31–35,37–40,42–46 Two studies did not report either the position or use of a nose clip.36,41 Only one study performed the measurement at different volumes for both Pimax and Pemax (percentage of vital capacity),41 whilst 19 studies measured Pimax at residual volume24,25,29–40,42–46 and 16 measured Pemax at total lung capacity.24,25,29,31–37,39,41,43–46,48 Half of the studies reported holding pressures for ≥1s,25,29,30,34,37,39,40,42–44 eight for 2s,32,33,35,36,38,41,45,46 one for ≥1.5s,24 and one study did not report the holding time31. Number of repetitions of the procedure varied between 2 and 10, with most studies (n=18) reporting at least 3 repetitions.24,30–46

A total of 76 predictive equations were found, 42 for Pimax and 34 for Pemax. Variables most frequently used in the equations were: age (Pimax n=22; Pemax n=17),24,25,29–40,42–46 weight (n=11; n=9),24,25,32–36,39,40,42,44,46 height (n=5; n=3)30,34,36,42,46 and sex (n=4; n=4).24,31,39,46 Most studies produced distinct equations for each sex (n=15),25,29,30,32–38,40,42–45 with only five studies reporting a single equation for both.24,31,39,41,46 BMI was used in five Pimax equations25,38,46 and in two Pemax equations,25 and waist circumference in one Pemax equation.24 Body surface area was also used in 2 Pimax and 2 Pemax equations.31 Only 9 studies reported LLN values,24,33,36,38,40,42–44 12 reported the SEE24,25,31,32,35–37,39,40,43,44,46 and almost all (n=19) the R2.24,25,30–46

Overall the proposed equations explained between 3 and 96% of the variance of Pimax/Pemax, with most studies (n=14) explaining less than 50%. Equations were developed with individuals from 11 countries, with the most representative continents being North America (n=6), Europe (n=6), South America (n=5) and Asia (n=4). No developed equations were found in Northern Asia, Africa or Oceania. A world map of the distribution of the existent predicted equations and respective range of R2 can be found in Fig. 2. Of the 42 Pimax and 34 Pemax equations, 12 and 8 respectively, were generated using procedures that globally complied with the international standards. Of the remaining 56 equations using procedures that have not followed the ATS/ERS guidelines, only 36 were developed before publication of the guidelines. Table 2 summarizes the predictive equations found for respiratory muscle strength.

Figure 2.

World map of studies with predictive equations for respiratory muscle strength and range of coefficients of determination. Numbers above range of coefficients represent the number of studies.

(0.33MB).