The multi‐ethnic global lung initiative 2012 (GLI-2012) norms reflect contemporary adult's Algerian spirometry

Background The validation of the multi-ethnic GLI-2012 spirometric norms has been debated in several countries. However, its applicability in Algeria has not been verified. Aim To ascertain how well the GLI-2012 norms fit contemporary adult Algerian spirometric data. Methods This was a cross-sectional study of a convenience sample of 300 healthy non-smoker adults (50% men, age range: 18–85 years) recruited from the Algiers region general population. All participants underwent a clinical examination and a plethysmography measurement. Z-scores for some spirometric data [FEV1, FVC, FEV1/FVC and forced expiratory flow at 25–75% of FVC (FEF25-75%)] were calculated. If the average Z-score deviated by “< ± 0.5” from the overall mean, the GLI-2012 norms would be considered as reflective of contemporary Algerian spirometry. Results The means±SDs of age, height, weight, FVC, FEV1, FEV1/FVC and FEF25-75% of the participants were, respectively, 48±17 years, 1.65±0.10 m, 73±14 kg, 4.04±1.04 L, 3.18±0.82 L, 0.79±0.05 and 4.09±1.09 L/s. Almost the quarter of participants were obese. The total sample means±SDs Z-scores were 0.22±0.87 for FVC, 0.04±0.88 for FEV1, -0.34±0.67 for FEV1/FVC and 0.93±0.79 for FEF25-75%. For men and women, only the means±SDs of the FEF25-75% Z-scores exceeded the threshold of “± 0.5”, respectively, 1.13±0.77 and 0.73±0.76. Conclusion Results of the present study, performed in an Algerian population of healthy non-smoking adults, supported the applicability of the GLI-2012 norms to interpret FEV1, FVC and FEV1/FVC but not the FEF25-75%.


Results
The means±SDs of age, height, weight, FVC, FEV 1 , FEV 1 /FVC and FEF 25-75% of the participants were, respectively, 48±17 years, 1.65±0.10 m, 73±14 kg, 4.04±1.04 L, 3.18±0.82 L, 0.79±0.05 and 4.09±1.09 L/s. Almost the quarter of participants were obese. The total sample means±SDs Z-scores were 0.22±0.87 for FVC, 0.04±0.88 for FEV 1 , -0.34±0.67 for FEV 1 /FVC and 0.93±0.79 for FEF 25-75% . For men and women, only the means±SDs of the FEF 25-75% Z-scores exceeded the threshold of "± 0.5", respectively, 1.13±0.77 and 0.73± 0.76. PLOS  Introduction the authors' knowledge, only one study ascertained how well the GLI-2012 norms fit contemporary adult Tunisian spirometric data [15]. However, while only Tunisian "healthy" adults were included, the authors generalized their results and concluded that GLI-2012 norms don't reflect contemporary adult's North-African spirometry [15]. This "generalization" is questionable for at least two methodological reasons. The 1 st one was the low percentage of females (19.6%) included in their sample (n = 489), which could "biased" conclusions. The 2 nd reason concerns the "unusual" recruitment method of the "healthy" adults, consisting of a population that undergoes spirometry at an occupational medicine group. Therefore, before accepting the "generalization" of the Tunisian study conclusion [15], it seems that verifying the applicability of the GLI-2012 [10] norms for the Algerian population is crucial for care activities and research, and is urgently needed. It is of unlimited meaning that the population from which the norms are derived is representative of the population under study [14]. The age scattering and other anthropometric, ethnic, socioeconomic and environmental factors should be equivalent since such factors can mark lung function [14]. Furthermore, the methodology for performing spirometric tests (eg; protocol and equipment) must be stringent [4,25]. The external validation of the GLI-2012 norms is recommended [9, 10] and further evaluations of applicability from other parts of the world (particularly the Arab one) are required in order to verify the appropriateness in these areas. Hitherto, there is no publication evaluating the applicability of the GLI-2012 norms for Algerian adults. Since the GLI-2012 norms [10] may be unsuitable for use in Algerian adults' population, it is essential that physicians are made aware of the potential consequences of adopting these norms for clinical decision-making [10].
The aim of this study was to evaluate if the GLI-2012 norms [10], although endorsed by several respiratory societies, are applicable for an adult Arab population resident in Algeria.

Study design
A cross sectional study was performed in the Department of Pneumology, Phthisiology and Allergology at the Rouiba Hospital, Algiers (186 m above sea level), Algeria. The study was conducted in compliance with the 'Ethical principles for medical research involving Human subjects' of the Helsinki Declaration (available from: http://www.wma.net/en/30publications/ 30ethicsmanual/pdf/ethics_manual_arabic.pdf; last visit: 7 th August 2018). The study was approved (approval number: 0601/2014) by the Rouiba Hospital (Algiers) Medical Advice and Ethics Commission [president: Pr. Ferhat Zebboudj (zebboudjferhat@gmail.com)]. Written informed consent was obtained from all participants.
The present project comprised two parts. The first one (the aim of this study) was to verify the applicability of the GLI-2012 norms [10] for some spirometric data in a sample of Algerian adults' healthy participants (GLI-2012 validation group). The second part aimed to generate plethysmographic norms for Algerian adults' healthy population (plethysmographic norms group).

Study population
The target population consisted of a group of healthy participants aged 18 years and more. They were selected by convenience sampling from the acquaintances of patients hospitalized at the Department of Pulmonology, Phthisiology and Allergology, during the visit-period for example.
Only healthy participants with technically acceptable and reproducible spirometry maneuvers were included. The presence of (i) acute or past chronic diseases of the respiratory system (eg, presence of physician-diagnosed respiratory disease (such as asthma, chronic bronchitis, chronic obstructive pulmonary disease, emphysema, or tuberculosis); hospitalization for lung or chest conditions), (ii) heart diseases which may influence the respiratory system (eg, heart failure, arrhythmia, unstable angina or myocardial infarction, uncontrolled blood hypertension), (iii) a cigarette smoking history of more than one pack-years, (iv) obesity levels 2 or 3, and (v) a higher level of sports practice (> 5 hours per week) were applied as non-inclusion criteria [15,26,27].

Data collection procedures
Medical data were collected using a simplified and modified medical questionnaire derived from the ATS division lung diseases questionnaire [28].
Plethysmography was carried out in the sitting position, and a nose clip was applied. All tests were performed between 9.00 am and 3.00 pm by only one qualified person (AK in the authors' list). Plethysmographic measurements were performed with a body plethysmograph (Body-box 5500, MediSoft, Belgium), carefully following the ATS/ERS recommendations [31,32]. The spirometer was calibrated daily with a 3-L syringe. The plethysmographic technique and especially the FVC maneuver, were previously described [15,[31][32][33][34][35][36]. Briefly, at least three reproducible FVC measurements were obtained [31]. FVC and FEV 1 , the best two out of the three selected tests, did not differ by more than 0.150 L (if FVC ! 1 L) or 0.100 L (if FVC < 1 L). The highest FVC and FEV 1 were computed, even though the two data did not come from the same flow-volume curve [31]. The following flow-volume curve data were measured and/ or calculated: FEV 1 (L), FVC (L), FEV 1 /FVC ratio (absolute value), forced expiratory flow at 25-75% of FVC (FEF 25-75% , L/s) and Z-scores (without unit). Algorithms and stand-alone software for the GLI-2012 norms [10] are freely available from www.lungfunction.org (last visit: 7 th August 2018). For the GLI-2012 [10], software calculated Z-scores for FEV 1 , FVC, FEV 1 /FVC and FEF 25-75% , and exported the results to a ".csv file" for manipulation in a spreadsheet.

Statistical analysis
The distribution of quantitative variables was normal and results were expressed by their means±SDs and 95% confidence interval (95%CI). The obesity status results were expressed as numbers (relative frequencies).
The chi-square test was used to compare percentages. The Student t-test was used to compare anthropometric and spirometric data of men and women.
Height-, age-and sex-specific Z-scores for spirometric data were calculated using the GLI-2012 norms [10]. If there was an offset between the GLI-2012 norms [10] and test population (measured spirometric data), the expected Z-scores of the test population would have a mean of "> zero" and a SD of "> one" and would therefore be considered as statistically significant [10]. As done in some studies [10, 15, 16, 37] and according to a consensus established by the GLI scientific advisory panel (http://www.lungfunction.org; last visit: 7 th August 2018), a Zscore of "> ± 0.5" was arbitrarily considered to be clinically significant.
All mathematical computations and statistical procedures were performed using a statistical software (Statistica Kernel version 6; Stat Software. France). Significance was set at the 0.05 level.

Results
Among the 608 explored adults, 491 (80.76%) were considered as healthy participants with normal spirometry (they form the "plethysmography norms group"). Among them, 300 adults (150 women) were randomly included in the GLI-2012 validation group.  Table 1 exposes the anthropometric data of the GLI-2012 validation group. Women and men were age-, BMI-and obesity status-matched. Compared to women, men were significantly taller and heavier. Table 2 exposes the absolute values and Z-scores of some flow-volume curve data. Its main conclusion was that only FEF 25-75% was out of the range considered to be significantly different (< ± 0.5). Yes to adopt GLI-2012 norms in Algeria Table 3 exposes the "r" between spirometric Z-scores and anthropometric data. "Fair" correlations were found between age and FEV 1 (total sample and women), between age and FEV 1 /FVC (total sample and men) and between height and FEF 25-75% (men and women). "Good" correlations were found only between age and FEF 25-75% (total sample, men and women). The spirometry Z-scores were not related to sex, except for FEF 25-75% and FEV 1 /FVC ( Table 2).

Discussion
The results of this study, performed in an Algerian population of 300 healthy non-smoking adults, supported the use of the GLI-2012 norms to interpret FEV 1 , FVC and FEV 1 /FVC but not the FEF 25-75% .
To the best of the authors' knowledge, only few studies [14-22, 24] aimed at evaluating the applicability of the GLI-2012 norms in healthy adults' populations. Table 4 presents the main characteristics and results of some studies reporting Z-scores data [14-20].

Methodology discussion
One of the main strong points of this study, as done in scarce relative ones [14,15], is its prospective design. Most of the remaining similar studies were retrospectives [16][17][18][19][20]. For example, all spirometric tests included in the 2012-Australian study [16] were performed in the year where a history of less than 10 pack-years of smoking was allowed. Moreover, in the French study [17], and in order to apply the GLI-2012 norms to "real-life" conditions in a general population, the authors have not taken into account smoking status. While, they noted that in a subgroup of non-smokers (n = 1081), the study sensitivity showed same results [17], their approach is questionable. In the present study, the percentage of participant with an obesity level-1 was almost 25%. On the one hand, this was similar to the percentage reported in the Tunisian study [15], where 20.3% of "healthy" participants were obese, and in line with the Finnish study [19], where some participants with levels -1 and -2 were included (percentages not reported). On the other hand, 30% of the Algerian adults showed obesity [42], and the present study group composition reflected this ''healthy" population as they exist in the real population. This increases the external validity of the present study.  Yes to adopt GLI-2012 norms in Algeria the following hypothesis: "if the GLI-2012 norms are appropriate, mean±SD Z-scores should approximate 0±1 across the entire age and height range studied". In the Swedish study [14], the agreement between the observed data in the local population and the GLI-2012 norms was verified and it was judged "perfect" if the mean Z-scores was zero and the SD was one. Moreover, relationships between Z-scores and age, height, weight and sex were examined and the lack of any such relationship was in favor of the GLI-2012 norms application [14]. In the Finnish study [19], the difference between predicted FVC and FEV 1 from their study and from GLI-2012 norms [10] was plotted as described by Bland and Altman [44]. In addition, the GLI-2012 norms were derived from cross-sectional data and application on longitudinal data, as done by Huls et al. [18] is encouraged at least for two reasons. Firstly, this could offer an original option to make longitudinal change of lung function comparable between different age groups and thereby substantially improve epidemiological analysis for respiratory risk factors [18]. Second, the use of norms makes it possible to appraise whether the change of lung function deviates from the its normal age-related decline [18]. This study presented two limitations. The 1 st one concerned the non-determination of the participants' socioeconomic levels and/or occupational status. As observed in the Tunisian study [15], this could slightly influence the results, since there are significant differences in some spirometric data (eg, FEV 1 /FVC) depending on the general socioeconomic status [45]. The 2 nd limitation concerned the non-exclusion of participants with "possible" restrictive ventilatory defect or "lung hyperinflation", as previously done in one study [15]. The main advanced reason to such choice was that the available local norms for lung volumes was published for the Eastern region of Algeria (Constantine, 649 m above sea level) and "seem" to be unsuitable for the Algiers region (186 m above sea level) [11].

Results discussion
The precision with which spirometry data are interpreted hinges on the suitability of the selected norms [4,37]. Mistakes in interpretation, with respect both to overestimation and underestimation of lung function abnormalities, can arise if inappropriate norms are applied [4,37].
How well did the GLI-2012 norms fit contemporary Algerian spirometric data?. The ERS-GLI task force noted that data from some regions (eg; the Arab World) are urgently required [10]. This study results demonstrate that the GLI-2012 norms [10] are "well" matched to some spirometry outcomes obtained in a contemporary Algerian population using modern equipment and in accordance with international spirometry guidelines [31].
As found in some studies [16][17][18]20] (Table 4), where means Z-scores for all measured spirometric data were "< ± 0.5", in this study, FEV 1 , FVC and FEV 1 /FVC means Z-scores were less than the within test variation accepted in spirometry testing. Furthermore, the observed variability (SD of the Z-score) of the above outcomes (Table 2) was close to one, indicating a good overall fit. However, the above conclusion cannot be applied for the FEF 25-75% since its mean Z-score was "> 0.5" ( Table 2). This result was in opposition with the unique study [16] reporting FEF 25-75% data where its mean Z-score was 0.07±0.95 (Table 4).
Z-scores point out how many SDs a measurement is from its normal value [10]. Compared to the percent predicted, they reduce bias due to age, height, sex and ethnic group, and are thus mainly helpful in defining the lower and upper limits of normal; they also simplify uniform interpretation of spirometry results [15]. In this study, there were good significant associations only between age and FEF 25-75% (Table 3) and only FEF 25-75% and FEV 1 /FVC Z-scores were related to sex ( Table 2). These results support the use of the GLI-2012 norms to interpret FEV 1 , FVC and FEV 1 /FVC data in the Algerian population. Other authors tested this kind of association and found controversial results [14-20] (Table 4). While some authors found some weak, but statistically significant, associations between the spirometry Z-scores and age [14][15][16], height [14-16], weight [14], or sex [14,16], others didn't find any association between the spirometry Z-scores and age [20] or height [20] or sex [15]. In the mutually adjusted multivariable models for some anthropometric data (ie; height, weight, age and sex), Thompson et al. [37] noted statistically significant but small associations for each of the spirometry Z-score results (FEV 1 Z-scores declined with height and were lower in women, FVC Z-scores declined with height, FEV 1 /FVC Z-scores increased with age and were lower in women). According to some authors [16,37], the magnitude of any differences related to such associations was small and of no physiological importance. Two possible mechanisms for the observed relationship were advanced [37]: increased variability of spirometry data with age [46] or that the all-age norms [46] didn't have sufficient data in the participants aged 60 years and more to accurately define the change in spirometry data with age.
Why did the GLI-2012 norms fit contemporary Algerian FEV 1 , FVC and FEV 1 /FVC data?. GLI-2012 datasets were obtained from 72 centers in 33 countries including Algeria [10]. Four ethnic groups were formed and the Algerian data [11] (n = 273) were included in the Caucasian group (n = 57395). Although representing almost 0.5% of the Caucasian data, the authors think that this inclusion could partially explain why the GLI-2012 norms fit contemporary Algerian spirometric data. Despite the ethnic, geographical, environmental, socioeconomic status similarities between Tunisia and Algeria and despite very close anthropometric data between this study (Table 1) and the Tunisian one [15] (Table 4), the two studies conclusions were opposite. One additional explanation, in addition to the above methodological differences, could be the existence of different subgroups in the North African population (Arab, Berber, Turkish descent) [15].
Why didn't the GLI-2012 norms fit contemporary Algerian FEF 25-75% data?. Two explanations could be advanced. The 1 st one is related to the high inter-test and intra-test variability of FEF 25-75% [31]. It is highly dependent on the validity of the FVC measurement and the level of expiratory effort [31]. For example, the FEF 25-75% between subject coefficient of variation varies between 20 and 62% [10]. This explains why it is not among the indices recommended by the ATS/ERS [4]. Moreover, the GLI group included it in their analyses only in response to requests from colleagues, especially those caring for children [10]. The 2 nd explanation is related to the effects of obesity on the FEF 25-75% [43]. On the one hand, almost 64% of the included participants were overweight or obese (Table 1). On the other hand, it was shown that BMI was negatively associated with the FEF 25-75% [47] (eg, in this study the "r" between the BMI and the FEF 25-75% was significant at -0.12).

Recommendation
In order to simplify comparative studies between countries, to avoid mistakes due to agerelated gaps in norms [48] and to simplify the conversion to norms for diverse ethnic groups, the authors acclaim implementation of the GLI-2012 spirometric norms in healthcare in Algeria. While, local spirometric data are available for children aged 6 to 16 years [49], there is a need to evaluate the applicability of the GLI-2012 in that age range.
In conclusion, the results of the current study support the use of the GLI-2012 norms to interpret clinical and research results in contemporary Algerian adults.
Supporting information S1 File. Spirometric data of the 300 Algerian adults.

Acknowledgments
Authors wish to thank Prs Béchir SAADAOUI and Farida HELLAL (Freelance Translators) for their invaluable contribution in the improvement of the quality of the writing in the present paper.
The present research data is accepted as a poster, in the upcoming Congress of the European Respiratory Society (ERS) International Congress (September 2018, Paris, France). First author: Abdelbassat Ketfi.