To the Editor:
Spirometry is an effective and widely available technique to measure lung function. Correct interpretation of spirometry is imperative when used to diagnose and manage lung pathology. The Global Lung Initiative 2012 (GLI2012) provides robust and representative reference equations for lung function in four ethnic groups; however, the GLI2012 is limited in data from African populations, and “Black” equations in GLI2012 were solely derived using data from African Americans. For populations lacking reference range equations and for individuals of mixed ethnic origin, the GLI2012 taskforce provided a composite “Other” equation (1).
The Pan African Thoracic Society is reluctant to endorse the use of the GLI2012 “Other” or “Black” equations in Africa without evidence of their applicability in African populations (2). In this study, we aimed to collect spirometry data in healthy South Africans to determine if the “Black” or “Other” GLI2012 reference equations were a good fit, or whether new reference equations are required. We hypothesized that the GLI2012 “Black” reference equations will not fit black South African adults and children. Some of the results of this study have been previously reported in the form of abstracts (3–5).
In this cross-sectional population-based study, healthy children and adults between the age of 5 and 95 years were recruited from two provinces in South Africa: KwaZulu-Natal and the Western Cape. South Africa has a population of over 57 million people who belong to four major ethnic groups: Black African (80.9%), Mixed Ethnicity (8.8%), Caucasian (7.8%), and Indian/Asian (2.5%) (6). In line with GLI2012 recommendations, we recruited a representative sample of at least 300 participants for each ethnic population (7). Participants were recruited between August 1, 2017, and July 31, 2018.
Anthropometric measurements were obtained and spirometry was performed as per international recommendations. Spirometry data were converted to z-scores using the GLI2012 Desktop Software for Large Datasets (version 1.3.4 Build 3, April 7, 2013) and summarized by ethnic group. A good fit was determined if the average z-score was not statistically or physiologically different from an average z-score of zero (SD of 1). A difference of more than 0.5 z-scores from zero was considered to be clinically significant, as it represents a difference greater than sampling variability (7).
A total of 4,223 participants were recruited; of these, 546 (13%) were excluded. Exclusions included those who were acutely unwell or had a previous diagnosis of respiratory, cardiac, or neuromuscular disease. Past and current smokers were also excluded as per GLI methodology and the American Thoracic Society recommendations (8). Tests with missing data, failing quality control, or with z-scores greater than ±5 were excluded. Demographic characteristics of the final cohort (3,676 participants) are included in Table 1. Observed z-scores from Black African participants (n = 2,116) showed that the GLI2012 “Other” had the best fit for this group (Figure 1; mean z-score ± SD of 0.13 ± 1.28 for FEV1, 0.13 ± 1.32 for FVC, and −0.01 ± 0.87 for FEV1/FVC).
Table 1.
Characteristics of the Study Population (KwaZulu-Natal and Western Cape Province, South Africa)
| Black African (n = 2,116) | Caucasian (n = 343) | Mixed Ethnicity (n = 693) | Indian (n = 524) | Total (N = 3,676) | |
|---|---|---|---|---|---|
| Sex, F | 1,200 (56.6%) | 153 (44.6%) | 404 (58.3%) | 326 (62.2%) | 2,083 (56.7%) |
| Age | |||||
| <25 yr | 1,128 (53.3%) | 212 (61.8%) | 440 (63.5%) | 243 (46.4%) | 2,023 (55.0%) |
| >25 yr | 988 (46.7%) | 131 (38.2%) | 253 (36.5%) | 281 (53.6%) | 1,653 (45.0%) |
| Weight-for-age z-score | 0.04 ± 0.41 | 0.10 ± 0.66 | 0.09 ± 0.47 | 0.12 ± 0.56 | 0.07 ± 0.48 |
| Height-for-age z-score | −0.27 ± 0.84 | 0.08 ± 0.87 | −0.40 ± 1.12 | −0.25 ± 0.95 | −0.26 ± 0.92 |
| BMI for age z-score | 0.99 ± 1.28 | 0.71 ± 1.15 | 0.90 ± 1.034 | 1.01 ± 1.28 | 0.95 ± 1.29 |
| Cormic index | 0.51 ± 0.03 | 0.52 ± 0.04 | 0.50 ± 0.03 | 0.52 ± 0.03 | 0.51 ± 0.03 |
| Stunting* | 110 (5.2%) | 5 (1.5%) | 64 (9.2%) | 37 (7.1%) | 216 (5.9%) |
| Province | |||||
| KwaZulu-Natal | 1,260 (59.6%) | 236 (68.8%) | 306 (44.2%) | 517 (98.7%) | 2,319 (63.1%) |
| Western Cape | 856 (40.4%) | 107 (31.2%) | 387 (55.8%) | 7 (1.3%) | 1,357 (36.9%) |
| Living region | |||||
| Rural | 1,020 (48.2%) | 11 (3.2%) | 192 (27.7%) | 0 (0.0%) | 1,223 (33.3%) |
| Urban | 1,096 (51.8%) | 332 (96.8%) | 501 (72.3%) | 524 (100.0%) | 2,453 (66.7%) |
| Housing type† | |||||
| Temporary | 136 (6.4%) | 1 (0.3%) | 10 (1.5%) | 1 (0.2%) | 148 (4.0%) |
| Wooden | 17 (0.8%) | 0 (0.0%) | 16 (2.3%) | 5 (1.0%) | 38 (1.0%) |
| Brick | 1,891 (89.5%) | 338 (98.5%) | 658 (95.1%) | 510 (97.8%) | 3,397 (92.6%) |
| Other | 68 (3.2%) | 4 (1.2%) | 8 (1.2%) | 5 (1.0%) | 85 (2.3%) |
| Household size† | |||||
| 1–3 people | 248 (11.8%) | 94 (27.4%) | 102 (14.7%) | 90 (17.2%) | 534 (14.6%) |
| 4–5 people | 881 (42.0%) | 210 (61.2%) | 368 (53.1%) | 284 (54.3%) | 1,743 (47.6%) |
| 6 or more people | 970 (46.2%) | 39 (11.4%) | 223 (32.2%) | 149 (28.5%) | 1,381 (37.8%) |
| Access to electricity | 2,058 (97.4%) | 341 (99.4%) | 687 (99.1%) | 520 (99.6%) | 3,606 (98.2%) |
| Heating/lighting fuel | |||||
| Wood | 139 (6.6%) | 1 (0.3%) | 4 (0.6%) | 1 (0.2%) | 145 (4.0%) |
| Paraffin | 11 (0.5) | 0 (0.0%) | 2 (0.3%) | 0 (0.0%) | 13 (0.4%) |
| Gas | 75 (3.6%) | 37 (10.8%) | 16 (2.3%) | 14 (2.7%) | 142 (3.9%) |
| Electricity | 1,967 (93.1%) | 325 (94.8%) | 680 (98.1%) | 514 (98.1%) | 3,486 (94.9%) |
| Cooking fuel | |||||
| Electricity | 1,988 (94.0%) | 319 (93.0%) | 682 (98.4%) | 516 (98.5%) | 3,505 (95.3%) |
| Coal | 6 (0.3%) | 2 (0.6%) | 1 (0.1%) | 0 (0.0%) | 9 (0.3%) |
| Wood | 169 (8.0%) | 1 (0.3%) | 0 (0.0%) | 0 (0.0%) | 170 (4.6%) |
| Paraffin/gas | 26 (1.2%) | 30 (8.8%) | 7 (1.0%) | 12 (2.3%) | 75 (2.1%) |
Definition of abbreviation: BMI = body mass index.
Children and young adults (<25 yr) constituted 55% of the population. Data are shown as mean ± SD or n (%).
Stunting defined as height-for-age z-score <−2.
Missing data.
Figure 1.
Distribution of z-scores of FEV1, FVC, and FEV1/FVC for Black African, Caucasian, Indian, and Mixed Ethnicity individuals using the GLI2012 reference equations. The equations that resulted in the closest fit to a mean z-score of zero and an SD of one were selected as best fit. GLI2012 = The Global Lung Initiative 2012; NE = Northeast; SE = Southeast.
The “Other” equations were also the best fit for the Mixed Ethnicity group (n = 693) (Figure 1; mean z-scores were 0.22 ± 1.44 for FEV1, 0.24 ± 1.56 for FVC, and −0.02 ± 0.85 for FEV1/FVC). The “Northeast Asian” equations had a similar average z-score but had much wider variability. The Caucasian participants (n = 343) demonstrated a good fit with the GLI2012 “Caucasian” equation (Figure 1; mean z-scores were 0.21 ± 1.22 for FEV1, 0.19 ± 1.24 for FVC, and 0.02 ± 0.91 for FEV1/FVC). Participants of Asian ancestry (n = 524) demonstrated a good fit to the “Southeast Asian” and “Black” equation. (Figure 1; Southeast Asian mean z-scores were −0.18 ± 1.03 for FEV1, −0.13 ± 1.09 for FVC, and −0.1 ± 0.93 for FEV1/FVC; Black equation mean z-scores were 0.15 ± 1.03 for FEV1, 0.04 ± 1.07 for FVC, and 0.23 ± 0.87 for FEV1/FVC). Across all ethnic groups, the FEV1/FVC ratio z-scores were close to zero (Figure 1).
In this large, representative sample of the South African population, we found that the GLI2012 “Caucasian” fit the Caucasian population well. For the Indian population, both the Black and the Southeast Asian equations demonstrated a good fit. As the Southeast Asian data reflects the ethnic background of the Indian population best, we determined that Southeast Asian showed the best fit but that a larger data set would be useful to confirm this. The GLI2012 “Other” equations fit the Black African and Mixed Ethnicity populations well. In South Africa, the black population largely represents a mixture of Bantu and Khoi-San ancestry. As these genetic groups predominate in wider Southern Africa, it may be appropriate to extrapolate our conclusions to the Southern African region.
However, previous studies investigating the use of GLI2012 equations in Africa are relatively scarce and have provided conflicting results from cohorts in different regions of Africa (9–12). We previously found that there is wide regional variability in lung function in Africa and that no single GLI2012 equation can be used for the African continent (13).
Ethnicity may be an important determinant of optimal lung function; it can also be confounded by different environmental and socioeconomic exposures that affect lung development and health. We note that in our study population, the majority lived in brick housing and had access to electricity, which may not be representative of other parts of Southern Africa. Some caution may therefore need to be applied in extrapolating our study results to the populations of different environmental and socioeconomic backgrounds. Further studies across Africa would help to determine best practices for each region or population.
The GLI2012 “Black” equation was derived entirely from African American subjects from the United States’ data sets, and it may be that this population is genetically closer to those in Central and West Africa. It is likely that individuals living in the United States experience a different nutritional and socioeconomic environment than those living in South Africa, and this may lead to further disparity in their lung function. We recommend that GLI should relabel the GLI2012 “Black” equations to reflect the population from which the data was derived (i.e., African American).
Despite including a wide range of participants (5–95 yr), our study was limited by few participants over the age of 50 years. As the health outcomes in South Africa continue to improve, further data in older individuals will be necessary to confirm validity. Furthermore, smoking habits were self-reported, and we did not screen for HIV infection. The strengths of this study are the large sample size, standardized equipment and measurement protocols, and data quality control.
Conclusions
The findings of this study will help to inform clinical practice in South Africa.
Supplementary Material
Acknowledgments
Acknowledgment
The authors thank Lindsay Zurba and her team of technicians for conducting the field work for the study.
Footnotes
Supported by a grant from the Medical Research Council of South Africa and Astrazeneca Investigator Self-Initiated Research Grant. S.-J.S. was supported by a Wellcome Trust Institutional Strategic Support Fund from Imperial College, London, UK.
Author Contributions: S.-J.S.: data handling, analysis, and interpretation and drafting and revising the manuscript. D.M.G.: study conception and design, data analysis and interpretation, and drafting and revising the manuscript. R.P.M.: data analysis and interpretation and revising the manuscript. G.L.H.: study conception and design, data analysis and interpretation, and drafting and revising the manuscript. S.S.: study conception and design, data analysis and interpretation, and drafting and revising the manuscript. R. Mphahlele: data interpretation and drafting and revising the manuscript. R. Masekela: study conception and design, data analysis and interpretation, and drafting and revising the manuscript.
Originally Published in Press as DOI: 10.1164/rccm.202005-2085LE on August 6, 2020
Author disclosures are available with the text of this letter at www.atsjournals.org.
References
- 1.Quanjer PH, Stanojevic S, Cole TJ, Baur X, Hall GL, Culver BH, et al. ERS Global Lung Function Initiative. Multi-ethnic reference values for spirometry for the 3-95-yr age range: the global lung function 2012 equations. Eur Respir J. 2012;40:1324–1343. doi: 10.1183/09031936.00080312. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Masekela R, Zurba L, Gray D. Dealing with access to spirometry in Africa: a commentary on challenges and solutions. Int J Environ Res Public Health. 2018;16:62. doi: 10.3390/ijerph16010062. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Smith S, Zurba L, Hall G, Stanojevic S, Gray D, Masekela R.How well do the 2012 GLI multi-ethnic spirometry reference equations fit healthy South African school children? Presented at the Abstracts of the combined SATS & PATS CongressApril 12–15.2018Durban, South Africa. p. 45. [Google Scholar]
- 4.Smith SJ, Gray D, Hall G, Stanojeic S, Mphahlele R, Zurba L, et al. Late breaking abstract-do the 2012 GLI multi-ethnic reference equations show a good fit for healthy Black South Africans? Presented at the European Respiratory Society CongressSeptember 15–19.2018Paris, France [Google Scholar]
- 5.Masekela R, Smith S, Stanojevic S, MacGinty R, Mphahlele R, Gray D.The validation of the GLI2012 ‘other’ reference equation in Black and mixed-race South Africans: the Paediatric and Adult African Spirometry (PAAS)II studyPresented at the American Thoracic Society International Conference Abstracts. May 15–20, 2020, Philadelphia, PA [Google Scholar]
- 6.Statistics South Africa Statistical release P0302: mid-year population estimates, South Africa 2005 [Internet] Pretoria, South Africa: Statistics South Africa; 2005[accessed 2020 May 28]. Available from: http://www.hst.org.za/indicators/StatsSA/midyearpop2005.pdf [Google Scholar]
- 7.Quanjer PH, Stocks J, Cole TJ, Hall GL, Stanojevic S Global Lungs Initiative. Influence of secular trends and sample size on reference equations for lung function tests. Eur Respir J. 2011;37:658–664. doi: 10.1183/09031936.00110010. [DOI] [PubMed] [Google Scholar]
- 8.Miller A. Lung function testing: selection of reference values and interpretative strategies. Am Rev Respir Dis. 1992;146:1368–1369. doi: 10.1164/ajrccm/146.5_Pt_1.1368b. [DOI] [PubMed] [Google Scholar]
- 9.Arigliani M, Canciani MC, Mottini G, Altomare M, Magnolato A, Loa Clemente SV, et al. Evaluation of the global lung initiative 2012 reference values for spirometry in African children. Am J Respir Crit Care Med. 2017;195:229–236. doi: 10.1164/rccm.201604-0693OC. [DOI] [PubMed] [Google Scholar]
- 10.Ratomaharo J, Linares Perdomo O, Collingridge DS, Andriamihaja R, Hegewald M, Jensen RL, et al. Spirometric reference values for Malagasy adults aged 18-73 years. Eur Respir J. 2015;45:1046–1054. doi: 10.1183/09031936.00114914. [DOI] [PubMed] [Google Scholar]
- 11.Madanhire T, Ferrand RA, Attia EF, Sibanda EN, Rusakaniko S, Rehman AM. Validation of the global lung initiative 2012 multi-ethnic spirometric reference equations in healthy urban Zimbabwean 7-13 year-old school children: a cross-sectional observational study. BMC Pulm Med. 2020;20:56. doi: 10.1186/s12890-020-1091-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Fawibe AE, Odeigah LO, Saka MJ. Reference equations for spirometric indices from a sample of the general adult population in Nigeria. BMC Pulm Med. 2017;17:48. doi: 10.1186/s12890-017-0390-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Masekela R, Hall GL, Stanojevic S, Sartorius B, MacGinty R, Saad HB, et al. An urgent need for African spirometry reference equations: the Paediatric and Adult African Spirometry study. Int J Tuberc Lung Dis. 2019;23:952–958. doi: 10.5588/ijtld.18.0442. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.