AJTCCM VOL. 30 NO. 3 2024 99
ORIGINAL RESEARCH: ARTICLES
Background. ere has been a growing interest in nutritional/lifestyle factors, including vitaminD, that may aect chronic obstructive
pulmonary disease (COPD). Most data are from Caucasian populations and temperate climates, with minimal African data.
Objectives. e primary objective was to determine the prevalence of vitaminD deciency (25-hydroxyvitaminD (25(OH)D) ≤20 ng/mL)
and insuciency (25(OH)D 21 - 29 ng/mL) among patients with COPD. Secondary objectives were to investigate the association between
vitaminD and demographic/lifestyle factors, lung function parameters, markers of COPD severity and corticosteroid use.
Methods. A prospective, cross-sectional study of 76 patients with COPD was conducted at a tertiary hospital in Johannesburg, South Africa.
Patients were interviewed regarding demographic/lifestyle factors, COPD severity markers and corticosteroid therapy. e most recent
spirometry result was recorded. Blood samples were taken for measurement of calcium, alkaline phosphatase and vitaminD levels. Patients
were stratied according to vitaminD status (deciency and non-deciency (25(OH)D >20 ng/mL, i.e. combined insuciency and adequate
levels)), and statistical analysis was performed to assess for associations.
Results. e sample included 72% males and 63% black African patients. e prevalences of vitaminD deciency and insuciency were 48%
(95% condence interval (CI) 42 - 54) and 35% (95% CI 30 - 41), respectively. A Modied Medical Research Council (mMRC) dyspnoea score
≥2 was associated with a relative risk of 1.34 (95% CI 1.05 - 1.7) for vitaminD deciency in univariate analysis. In multivariate regression
analysis, only sunlight exposure (<1 hour/day) was an independent predictor of vitaminD deciency (odds ratio 2.4; 95% CI1.3 - 4.5).
Conclusion. ere was a high prevalence of suboptimal vitaminD levels in this COPD sample population. A higher mMRC score was associated
with an increased risk of vitaminD deciency, while low sunlight exposure was the only independent predictor of vitaminD deciency.
Keywords. Vitamin D, vitaminD deciency, vitaminD deciency prevalence, 25(OH)D, COPD, chronic obstructive pulmonary disease.
Afr J Thoracic Crit Care Med 2024;30(3):e1141. https://doi.org/10.7196/AJTCCM.2024.v30i3.1041
Globally, >170 million people are affected by chronic obstructive
pulmonary disease (COPD), and COPD accounted for ~3.2 million
deaths in 2015.[1] Low- and middle-income countries bear a signicant
burden of COPD mortality. There has been a growing interest in
nutritional and lifestyle factors that may affect COPD. Globally,
numerous studies have demonstrated a high prevalence of vitaminD
deciency in patients with COPD.[2-10] Notably, most data are from
Caucasian populations in countries with developed economies and
Vitamin D status in patients with chronic obstructive pulmonary
disease at Chris Hani Baragwanath Hospital, Johannesburg,
SouthAfrica
I Kola,1 MB BCh, Dip HIV Man (SA), FCP (SA), MMed (Int Med) ;
S A van Blydenstein,2 MB BCh, DCH, FCP (SA), MMed (Int Med), Cert Pulmonology (SA); M Kola,3 MB ChB;
S Kooverjee,1 MB BCh, FCP (SA); S Omar,4 MB ChB, FC Path (SA) Chem, DA (SA), Critical Care (SA)
1 Department of Internal Medicine, School of Clinical Medicine, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa
2 Division of Pulmonology, Department of Internal Medicine, School of Clinical Medicine, Faculty of Health Sciences, University of the Witwatersrand and
ChrisHani Baragwanath Academic Hospital, Johannesburg, South Africa
3 Private general practitioner, Newclare, Johannesburg, South Africa
4 Division of Critical Care, School of Clinical Medicine, Faculty of Health Sciences, University of the Witwatersrand and Chris Hani Baragwanath Academic
Hospital, Johannesburg, South Africa
Corresponding author: I Kola (imraankola@gmail.com)
Study synopsis
What the study adds. is is the rst study to provide prevalence data regarding vitaminD status in COPD patients in sub-Saharan Africa.
e study highlights a relationship between vitaminD status and both symptom severity and sunlight exposure.
Implications of the ndings. Owing to the high prevalence of suboptimal vitaminD status among COPD patients, it may be useful to
screen patients for vitaminD deciency, especially those with a more severe phenotype. ere may be scope for further studies to evaluate
whether vitaminD supplementation corrects the deciency and provides any clinical outcome benet.
100 AJTCCM VOL. 30 NO. 3 2024
ORIGINAL RESEARCH: ARTICLES
temperate climates. ere is a paucity of data on the prevalence of
COPD and vitaminD deciency in Africa and among non-Caucasians.
Vitamin D3 is predominantly synthesised in the skin. Ultraviolet
(UVB) radiation converts pre-vitaminD3 to vitaminD3 (cholecalciferol).
is step is inuenced by the melanin content in the skin and sunlight
exposure. e other less signicant source of vitaminD is dietary
intake. Vitamin D3 is hydroxylated in the liver to 25-hydroxyvitaminD
(25(OH)D). 25(OH)D is the major storage form and is used to ascertain
vitaminD status in populations.[11,12] Vitamin D is predominantly
hydroxylated in the kidney to 1,25-dihydroxyvitaminD. is represents
the active form of vitaminD, which enhances gastrointestinal absorption
of calcium and phosphate and has a positive eect on bone turnover and
bone mineral density.[13]
ere has been a growing interest in the non-calcaemic eects of
vitaminD, which include immunomodulation.[14] Vitamin D has
been shown to combat mycobacterial and other respiratory infections
through the production of cathelicidin (antimicrobial peptide).[14] is
interaction between vitaminD and cathlicidin may be relevant, because
infectious exacerbations are linked to COPD disease progression. In
addition, vitaminD deciency has been associated with other respiratory
conditions including tuberculosis,[15] sarcoidosis,[16] childhood asthma,
cystic brosis[13] and recently COVID-19 .[17]
Spirometric correlates concerning the association between the forced
expiratory volume in the 1st second (FEV1) and vitaminD have been
conflicting in COPD patients.[2,3,5,7,8] Most studies demonstrated a
correlation between vitaminD levels and symptom scores,[6,8,9,18] as well
as ethnicity and sunlight exposure,[5-7,9,18] among COPD patients.
Objectives
We therefore studied the prevalence of vitaminD deficiency and
insuciency (25(OH)D ≤20 ng/mL and 21 - 29 ng/mL, respectively) in
COPD patients in South Africa (SA). Secondary objectives were to look
for an association between vitaminD and demographic/lifestyle factors,
lung function parameters, markers of COPD severity, and corticosteroid
type and dosage.
Methods
Study design
is was a prospective, cross-sectional study of COPD patients.
Study population
Patients with spirometry-conrmed COPD were included, as per the
Global Initiative for Obstructive Lung Disease (GOLD) 2019 guideline.
[19] For inclusion, participants had to have a post-bronchodilator FEV1/
forced vital capacity (FVC) ratio <70% together with at least one of
the classic symptoms of COPD – chronic cough, dyspnoea or chronic
sputum production. e spirometry result was the most recent in the
patient’s le, or spirometry was repeated on the date of the interview if
no result could be traced in the records. Patients had to be >18 years
of age. Patients were excluded if they had reversible airow limitation,
current active pulmonary tuberculosis, a concomitant diagnosis of
asthma, active malignancy, malabsorption or a history of pancreatic
insuciency, or if they were already on vitaminD supplementation.
Study setting
e study was conducted at a tertiary academic hospital in Johannesburg,
SA. Patients were recruited from the outpatient clinic and inpatient
wards. Johannesburg is at a latitude of 26.2˚ south of the Equator.
Sample size
Based on previous studies, ≥7% of COPD patients have normal
vitaminD levels. Using an estimate of 7%, a precision of 5% and a
condence level of 95%, a sample size of 100 was initially aimed for.
Unfortunately, owing to COVID-19 spirometry limitations and
COVID-19 elective patient number curtailments, a sample size of only
76 patients was reached. Consecutive patients were recruited between
November 2020 and July 2022.
Data collection
Data were collected by the investigators using a data sheet. Spirometry
was performed if no result was available in the records. Pulmonary
function tests were performed using a JAEGER Vyntus SPIRO PC
spirometer (Vyaire Medical, USA) with a calculation of the percentage
of the predicted FVC and FEV1 values according to American oracic
Society/European Respiratory Society recommendations. A400µg dose
of salbutamol was administered for post-bronchodilator spirometry.
Patient demographic/lifestyle factors, clinical and COPD severity
markers, spirometry results, laboratory data and treatment information
were collected.
A 3 - 5 mL cued venous blood sample was taken in an acid citrate
dextrose tube from participants. Samples were transported to the
laboratory under cold-chain conditions.
Denitions
Ethnicity. Ethnicity was self-reported by the participants.
Season of sample collection. Dates of blood sample collections/
interviews were categorised into seasons. Seasons were defined as
follows: summer (December - February), autumn (March - May),
winter (June - August), and spring (September - November).
Sunlight exposure. Sunlight exposure was self-reported by the
participant and based on recall of the week preceding the interview.
Categories were 1 - 4 hours/week, 5 - 6 hours/week, 1 - 2 hours/day, 3 - 5
hours/day and ≥6 hours/day.
Exacerbation. An exacerbation was based on patient reports/le notes
and included episodes requiring oral/intravenous (IV) corticosteroids,
antibiotics, a casualty visit, medical practitioner consultation or hospital
admission, or worsening of cough, dyspnoea or sputum production (>2
days).
Smoking status. A patient was regarded as an ex-smoker if they had not
smoked for >3 months. Never-smoker was dened as <100cigarettes
consumed during the course of the patient’s life.
Systemic corticosteroids. This was defined as use of IV/oral
corticosteroids during the year preceding the interview.
Low-dose inhaled corticosteroids. This was defined as low- and
medium-dose inhaled corticosteroids (fluticasone ≤250 µg/day,
budesonide 160 µg/day, and beclomethasone 200 - 400 µg/day).
High-dose inhaled corticosteroids. is was dened as uticasone
≥500 µg/day, budesonide 320 µg/day, and beclomethasone ≥400µgday.
Vitamin D testing and denition
25(OH)D was measured with a double-sandwich immunoassay using
a chemiluminescent label at a South African National Accreditation
AJTCCM VOL. 30 NO. 3 2024 101
ORIGINAL RESEARCH: ARTICLES
System (ISO15189)-approved laboratory.
e instrument used was the ARCHITECT
i2000 (Abbott, USA). This method is
traceable to the reference method, namely
liquid chromatography-mass spectrometry,
and meets the required standards for
clinical testing.[20] ere is some variation in
levels quoted for vitaminD deficiency and
insufficiency in the literature. Definitions
used for 25(OH)D in the present study were
deciency ≤20 ng/mL, insuciency 21 - 29 ng/
mL, and adequate ≥30 ng/mL.[21]
Statistical analysis
All data obtained were entered onto an Excel
spreadsheet, Microsoft 365, version 2405
(Microsoft, USA), by the first author (IK),
and then into Statistica v13.3 (StatSo, USA,
currently maintained by TIBCO Soware Inc.,
USA). e prevalence data were provided as
percentages with 95% confidence intervals
(CIs). e distribution of data was determined
from histograms, using the Shapiro-Wilk and
Lilliefors tests. Categorical variables were
presented as counts (n) and percentages, and
comparisons were made in vitaminD-decient
and non-decient (25(OH)D >20 ng/mL, i.e.
combined insuciency and adequate levels)
groups using the χ
2
test. Continuous variables
were summarised as means with standard
deviations (SDs) for normally distributed data
and medians with interquartile ranges (IQRs)
for non-normally distributed data. Independent
variables were compared in decient v. non-
deficient groups using the Mann Whitney
U-test for independent medians and Student’s
t-test for independent means.
For multivariate analysis, eight variables
were used to predict the presence of vitaminD
deciency or insuciency. e choice of these
variables was based on pathophysiological
plausibility and prominence in the literature
review regarding their effect on vitaminD
levels. There were six continuous variables:
age in years, body mass index (BMI), waist
circumference, smoking pack history in years,
FEV1 and the Modified Medical Research
Council (mMRC) dyspnoea score, and two
categorical variables: sunlight exposure (<1
hour/day v. ≥1 hour/day) and ethnicity (black
v. non-black). Six variables with a p-value <0.2
on the univariate model were selected for the
nal multivariate model.
Ethical considerations
Ethics approval was obtained from the Human
Research Ethics Committee (Medical) at
the University of the Witwatersrand (ref. no.
M200112) before commencement of the study.
Written informed consent was obtained from
each participant before recruitment.
Results
We included all 76 patients in our analysis. e
study patient characteristics are summarised
in Table1. ere were 55 males (72%). e
ethnicity prole was as follows: black African
48 patients (63%), coloured (mixed race) 17
(22%), Indian/Asian 8 (11%), and white 3
(4%). e mean (SD) corrected calcium level
was 2.25 (0.15) mmol/L and the median (IQR)
alkaline phosphatase level 92 (76 - 121) U/L.
Fig.1 shows the sunlight exposure of the study
participants.
Primary objective
e prevalence of vitaminD deciency and
insuciency (25(OH)D <30ng/mL) was 84%
(95% CI 80 - 88). Table2 summarises the
prevalence and median vitaminD level in each
status category. e median (IQR) 25(OH)D
level in our sample was 21 (14.5 - 26.5) ng/mL.
Secondary objectives
Univariate analysis
The differences between the vitamin
D-deficient and non-deficient groups are
provided in Table 3. ere was a relative risk
(RR) of vitamin D deficiency for patients
with daily sunlight exposure of <1hour/day
compared with ≥1 hour/day of 1.62 (95% CI
1.02 - 2.57) (Fig.1). ere was no dierence
in vitaminD deciency between black African
and non-black African ethnicity (p=0.11).
Smoking status (never-, ex- or current
smoker) was not compared in the deciency
v. non-deficiency analysis, as two groups
with a similar number of participants to
facilitate meaningful comparison could not be
constituted. e majority of participants were
ex- or current smokers (n=68; 89%), and only
8 (11%) had never smoked.
There were no significant differences in
spirometry parameters (FEV1 and FEV1
percentage predicted, FVC and FVC
percentage predicted, FEV1/FVC ratio)
between the vitaminD-deficient and non-
decient groups.
Table3 summarises the severity features
and therapeutic differences between the
vitamin D-deficient and non-deficient
groups. An mMRC dyspnoea score of ≥2 was
associated with an RR of 1.34 (95% CI 1.05 -
1.7) for vitaminD deciency compared with
a score of <2. Nosignicant dierence in the
deciency v. non-deciency groups was noted
in number of exacerbations in the preceding
year, GOLD grade, GOLD group, use of
Vitamin D decient (25(OH)D ≤20 ng/mL)
Vitamin D non-decient (25(OH)D >20 ng/mL)
1 - 4
hours/week
Sunlight exposure categories during week preceding interview
(participant reported), in ascending order
Participants in each vitamin D category, %
100
90
80
70
60
50
40
30
20
10
0
n=15
n=7
5 - 6
hours/week
n=6
n=6
1 - 2
hours/day
n=10
n=12
3 - 5
hours/day
n=5
n=11
>6
hours/day
n=1
n=3
Fig.1. Vitamin D status in relation to increasing sunlight exposure. (25(OH)D = 25-hydroxyvitaminD.)
102 AJTCCM VOL. 30 NO. 3 2024
ORIGINAL RESEARCH: ARTICLES
inhaled or systemic corticosteroids during
the preceding year, or inhaled corticosteroid
dosage.
Multivariate analysis
Six variables with a p-value <0.2 on the
univariate model were selected for the nal
multivariate model (Table4). Only sunlight
exposure (<1hour/day) was an independent
predictor of vitaminD deciency (odds ratio
2.4; 95% CI 1.3 - 4.5).
Discussion
e main nding from this study was the high
prevalence of vitaminD deciency (48%) and
insuciency (35%) in the study population.
In an analysis of the Subpopulations and
Intermediate Outcome Measures in COPD
Study (SPIROMICS) cohort,[5] with a large
sample size of 1 609, 20.6% and 33.2% of
participants were vitaminD deficient and
insucient, respectively. e same denitions
for vitaminD levels were used in the present
study. The proportion with deficiency was
signicantly lower in the SPIROMICS cohort
than in the present study. Some reasons that
may account for this discrepancy in ndings
are a higher proportion of patients with a
milder COPD phenotype (majority GOLD
grade 2) and a majority of Caucasians (lower
skin melanin content) in the US study.
In another large US study, Kunisaki etal.[6]
found 40.4% of participants to be vitaminD
deficient and 33.1% to be insufficient. This
is not dissimilar to our study findings, as
the population matched our cohort’s COPD
disease severity (majority GOLD grade 3 and
with a high exacerbation risk). e marginally
lower prevalence of deficiency may be
accounted for by ethnicity (majority Caucasian
participants).
Gawron et al.[22] in a small Polish case-
control study found the highest reviewed
rates of vitamin D deficiency (90.2%) in
COPD patients. Controls had similarly high
vitaminD deciency rates, and these ndings
may be accounted for by winter-only, nadir
vitaminD sampling in a temperate location.
Holick[14] in a review article quoted a
vitamin D deficiency prevalence of 40 -
100% in elderly non-institutionalised healthy
people in the USA, and stated that >50% of
postmenopausal women with osteoporosis
Table1. Study patient characteristics (N=76)
Variable n (%)*
25(OH)D (ng/mL), median (IQR) 21 (14.5 - 26.5)
Age (years), mean (SD) 62 (10)
BMI (kg/m2), median (IQR) 21 (18 - 25)
Waist circumference (cm), median (IQR) 83 (73 - 92)
Smoking pack-years, median (IQR) 19 (5 - 37)
Season of blood collection
Summer (December - February) 24 (32)
Autumn (March - May) 30 (39)
Winter (June - August) 12 (16)
Spring (September - November) 10 (13)
Lung function parameters, median (IQR)
FEV (L) 1.1 (0.8 - 1.5)
FEV1 (% predicted) 41.9 (31.1 - 65.2)
FVC (L) 2.5 (1.9 - 3.1)
FVC (% predicted) 86.1 (59.9 - 99.9)
FEV1/FVC (%) 44.6 (35.6 - 55.5)
Number of exacerbations in past year, median (IQR) 2 (2 - 3)
mMRC dyspnoea score
0 1 (1)
1 16 (21)
2 15 (20)
3 39 (51)
4 5 (7)
GOLD† grade
1 7 (9)
2 23 (30)
3 30 (39)
4 16 (21)
GOLD† group
A 7 (9)
B 33 (43)
C 9 (12)
D 27 (36)
25(OH)D = 25-hydroxyvitaminD; IQR = interquartile range; SD = standard deviation; BMI = body mass index; FEV1 = forced
expiratory volume in the 1st second;
FVC = forced vital capacity; mMRC = Modied Medical Research Council; GOLD = Global Initiative for Obstructive Lung Disease
2019.
*Except where otherwise indicated. Mean (SD) for normally distributed data, median (IQR) for non-normally distributed data.
†Singh etal.[19]
Table2. Prevalence of vitaminD deciency and insuciency in the study population (N=76)
Vitamin D status Prevalence, % (95% CI) n
25(OH)D (ng/mL),
median (IQR)
Deciency and insuciency (<30 ng/mL) 84 (80 - 88) 64 18 (12.5 - 23.5)
Deciency (≤20 ng/mL) 48 (42 - 54) 37 14 (11 - 17)
Insuciency (21 - 29 ng/mL) 35 (30 - 41) 27 25 (22 - 27)
Adequate levels (≥30 ng/mL) 16 (12 - 20) 12 37.5 (31.5 - 38.5)
CI = condence interval; 25(OH)D = 25-hydroxyvitaminD; IQR = interquartile range.
AJTCCM VOL. 30 NO. 3 2024 103
ORIGINAL RESEARCH: ARTICLES
Table3. Dierences between vitaminD-decient and non-decient (insucient and adequate) groups
Variable
Vitamin D
deciency (25(OH)D
≤20 ng/mL) (n=37),
n (%)*
Vitamin D
non-deciency
(25(OH)D >20 ng/ml)
(n=39), n (%)* p-value/RR (95% CI)†
Demographic and lifestyle factors
Age (years), mean (SD) 64 (11) 61 (8) 0.10
Gender male 27/37 (73) 28/39 (72) 0.91
Weight (kg), median (IQR) 58(53 - 72) 58(47 - 68) 0.44
Height (m), mean (SD) 1.66 (0.08) 1.66 (0.09) 0.90
BMI (kg/m2), median (IQR) 20 (19 - 26) 21 (17 - 23) 0.56
Waist circumference (cm), median (IQR) 83 (76 - 94) 83 (72 - 90) 0.44
Smoking pack-years, median (IQR) 24 (7.5 - 40) 17 (5 - 27) 0.13
Lung function parameters, median (IQR)
FEV1 (L) 0.94 (0.69 - 1.49) 1.11 (0.77 - 1.75) 0.18
FEV1 (% predicted) 41.20 (28.50 - 53) 42.90 (32.20 - 70) 0.22
FVC (L) 2.39 (1.71 - 3.01) 2.56 (1.95 - 3.32) 0.22
FVC (% predicted) 82 (57.10 - 97.30) 89 (66.14 - 100.40) 0.22
FEV1/FVC (%) 42 (36.27 - 55.50) 47.98 (34.14 - 55.56) 0.72
COPD severity markers
mMRC dyspnoea score
<2 4/37 (11) 13/39 (33)
≥2 33/37 (89) 26/39 (66) RR of deciency: 1.34 (1.05 - 1.7)
Number of exacerbations in past year 0.26
≤1 21/37 (57) 27/39 (69)
>1 16/37 (43) 12/39 (31)
GOLD‡ grade GOLD 1 and 2 v. 3 and 4: 0.45
1 3/37 (8) 4/39 (10)
2 10/37 (27) 13/39 (33)
3 14/37 (38) 16/39 (41)
4 10/37 (27) 6/39 (15)
GOLD‡ group GOLD A and B v. C and D: 0.80
A 2/37 (5) 5/39 (13)
B 18/37 (49) 15/39 (38)
C 1/37 (3) 8/39 (21)
D 16/37 (43) 11/39 (28)
erapy
Use of inhaled corticosteroids in past year 0.38
Yes 33/37 (89) 32/39 (82)
No 4/37 (11) 7/39 (18)
Use of systemic corticosteroids in past year 0.80
Yes 21/37 (57) 21/39 (54)
No 16/37 (43) 18/39 (46)
Inhaled corticosteroid dose 0.39
Low§8/37 (22) 5/39 (13)
High25/37 (68) 27/39 (69)
25(OH)D = 25-hydroxyvitaminD; n = number in category; RR = relative risk; CI = condence interval; SD = standard deviation; IQR = interquartile range; BMI = body mass index,
FEV1 = forced expiratory volume in the 1st second; FVC = forced vital capacity; mMRC = Modied Medical Research Council; GOLD = Global Initiative for Obstructive Lung Disease 2019.
*Except where otherwise indicated. Mean (SD) for normally distributed data, median (IQR) for non-normally distributed data.
†All values are p-values except mMRC score, which is RR (95% CI).
‡Singh etal.[19]
§Includes low- and medium-dose inhaled corticosteroids (uticasone ≤250 µg/day, budesonide 160 µg/day, beclomethasone 200 - 400 µg/day).
Fluticasone ≥500 µg/day, budesonide 320 µg/day, beclomethasone ≥400 µg/day.
104 AJTCCM VOL. 30 NO. 3 2024
ORIGINAL RESEARCH: ARTICLES
(without COPD) were reported as having suboptimal vitaminD
(25(OH)D <30 ng/mL). A large African meta-analysis[23] found a
population prevalence of 59% for combined vitaminD deciency and
insuciency, compared with 84% in our COPD study. Similar ndings
were noted in case-control studies.[4,9,18] These general population
vitaminD deciency/insuciency prevalence estimates appear lower
than in the COPD population.
Some of the postulated mechanisms for a higher prevalence of
vitaminD deciency in COPD patients include a sicker phenotype,
resulting in less sunlight exposure; a poorer diet; smoking resulting
in pigmentary skin changes and decreased cutaneous pre-vitaminD3
activation; possible increased vitamin D catabolism due to
corticosteroid use; and lower BMI and hence lower fat/muscle stores
of the vitamin.[13]
Sunlight exposure was significantly associated with vitaminD
deciency in univariate analysis in the present study, and remained
the only independent predictor in the multivariate model. This
association is evident in many European studies. Jollie etal.,[7] in a
multicentre cross-sectional study in London with 278 participants,
showed that the absence of a recent sunny holiday correlated with
vitaminD deciency in a COPD cohort. Kentson etal.[9] in a Swedish
case-control study with 38 COPD patients also found vitaminD
deficiency to be associated with a lower ultraviolet score (UVS).
The UVS was a composite measure of seasonality and sunlight
(UV) exposure. e association between low sunlight exposure and
vitaminD is attributable to the vitaminD pathway and UV-dependent
skin activation, as mentioned previously.
e present study found a trend towards an association between
vitaminD deciency and a larger number of smoking pack-years, lower
FEV1 and older age. In an Italian cohort, Malinovschi etal.[3] found
no association between vitaminD level and age or smoking history.
A Belgian study[4] also found no age or current smoking association
with vitaminD. Burkes etal.[5] found an inverse trend to our study,
with younger age associated with vitaminD deciency. is trend in
our study may be explained physiologically, as age and disease severity
may limit mobility, resulting in decreased outdoor sunlight exposure.
Also,skin pigmentation can become darker with ageing, decreasing
cutaneous vitaminD UV activation. Burkes etal.,[5] and Persson etal.
[18] in a multivariate analysis, also showed an association between
vitaminD levels and smoking status.
e majority of reviewed studies showed an association between a
lower FEV1 and vitaminD deciency.[2,5,7-9,18] ree studies showed no
vitaminD-FEV1 association.[3,22,24] e relationship between FEV1 and
vitaminD deciency in our study may not have reached signicance
owing to disparate sampling times and a limited sample size.
Our data showed that an mMRC dyspnoea score ≥2 was associated
with an increased risk of vitaminD deciency. Kentson etal.[9] also
demonstrated an association between higher symptom scores (which
include dyspnoea as a component) and vitaminD deciency (COPD
Assessment Test and mMRC dyspnoea score if not on vitaminD
supplementation). Kunisaki etal.[6] made a similar association, but
with the St George’s Respiratory Questionnaire (SGRQ). In their
large Norwegian case-control study, Persson etal.[18] found the same
association between mMRC dyspnoea score and vitaminD status in
univariate analysis, but not in multivariate analysis, similar to our data.
Hyun etal.,[8] in a South Korean study, also found high brinogen and
low vitaminD to be associated with higher mMRC dyspnoea scores.
e outlier in the literature[7] found no correlation between vitaminD
and the SGRQ (with dyspnoea as a component).
Strengths of the present study include an SA context (where data
are sparse) and a wide variety of factors investigated. Additionally, a
diverse sample of COPD severity was included. Vitamin D levels were
measured over all seasons in the study population as a whole (on the
date of interview for each patient), which may be more representative
of the prevalence of vitaminD deficiency than vitaminD nadir
(winter/spring)-only sampling. The majority of patients were
recruited from an outpatient department (OPD) setting, reducing
the confounding of acute illness. e same investigator conducted
surveys and measured parameters, hence negating the eects of inter-
investigator inconsistency/variability.
Limitations of this study include a smaller sample size due to
COVID-19 clinic number curtailments and spirometry restrictions. It
was a cross-sectional study, and vitaminD levels were only measured
at a single time/season in each patient. There may be seasonal
variations in vitaminD levels in individual patients. No inferences
about causality can be made, as vitaminD-deficient participants
were not followed up prospectively. No analysis of the comorbidities
of participants was made, and these could have aected the results.
Asingle-centre tertiary hospital study may limit the transferability of
ndings to other COPD populations.
Conclusion
ere was a high prevalence of vitaminD deciency and insuciency
in this COPD sample population. A higher mMRC score was associated
Table4. Binomial logistic regression for vitaminD deciency (six variables)
Variable Beta value SE p-value
Intercept –1.97
Age 0.03 0.03 0.29
Smoking pack-years 0.02 0.01 0.25
FEV1 –0.64 0.56 0.26
mMRC dyspnoea score 0.25 0.36 0.48
Ethnicity (black v. non-black) –0.055 0.35 0.11
Sunlight exposure* 0.89 0.31 0.005
SE = standard error; FEV1 = forced expiratory volume in 1 second; mMRC = Modied Medical Research Council.
*Sunlight exposure <1 hour/day v. ≥1 hour/day.
AJTCCM VOL. 30 NO. 3 2024 105
ORIGINAL RESEARCH: ARTICLES
with an increased risk of vitaminD deciency, while sunlight exposure
was the only independent predictor of vitaminD deciency.
Recommendations
ere is scope for case-control studies to evaluate the prevalence of
vitaminD deciency among healthy/hospitalised African participants
compared with their COPD counterparts. Given the high prevalence
of vitaminD deciency in COPD patients, routine testing in high-risk
groups may be valuable. ese ndings need to be validated in a milder
phenotype COPD population. Future studies should focus on the
clinical impact of vitaminD replacement in decient patients.
Declaration. e research for this study was done in partial fullment of
the requirements for IK’s MMed (Int Med) degree at the University of the
Witwatersrand.
Acknowledgements. Special thanks to Prof. M Wong, Head of Pulmonology
at Chris Hani Baragwanath Academic Hospital, Johannesburg, for allowing
the study to take place and for assisting at the respiratory outpatient
department. Special thanks also to Mohau La-Donna Kapa, Amore Visagie
and Annie Maphthu, respiratory technicians, who assisted with spirometry
and data collection.
Author contributions. IK wrote the proposal, obtained ethics clearance,
collected data, and wrote the final article. SAvB contributed to the
conceptualisation of the study, and supervised the protocol and article
writing. SO contributed to the conceptualisation of the study, supervised the
protocol, assisted in data interpretation, and supervised the article writing.
MK assisted with data collection and editing the protocol and article. SK
assisted with data collection.
Funding.Funding was obtained from the University of the Witwatersrand
Internal Medicine Research Incentive (RINC) fund. All printing and
transport costs were nanced by the rst author (IK).
Conicts of interest.None.
1. Soriano JB, Abajobir AA, Abate KH, etal. Global, regional, and national deaths,
prevalence, disability-adjusted life years, and years lived with disability for chronic
obstructive pulmonary disease and asthma, 1990 - 2015: A systematic analysis for the
Global Burden of Disease Study 2015. Lancet Respir Med 2017;5(9):691-706. https://doi.
org/10.1016/S2213-2600(17)30293-X
2. Jung JY, Kim YS, Kim SK, etal. Relationship of vitaminD status with lung function and
exercise capacity in COPD. Respirology 2015;20(5):782-789. https://doi.org/10.1111/
resp.12538
3. Malinovschi A, Masoero M, Bellocchia M, etal. Severe vitaminD deficiency is
associated with frequent exacerbations and hospitalisation in COPD patients. Respir
Res 2014;15(1):131. https://doi.org/10.1186/s12931-014-0131-0
4. Janssens W, Bouillon R, Claes B, etal. Vitamin D deciency is highly prevalent in COPD
and correlates with variants in the vitaminD-binding gene. orax 2010;65(3):215-220.
https://doi.org/10.1136/thx.2009.120659
5. Burkes RM, Ceppe AS, Doerschuk CM, etal. Associations among 25-hydroxyvitaminD
levels, lung function, and exacerbation outcomes in COPD. Chest 2020;157(4):856-665.
https://doi.org/10.1016/j.chest.2019.11.047
6. Kunisaki KM, Niewoehner DE, Connett JE. Vitamin D levels and risk of acute
exacerbations of chronic obstructive pulmonary disease. Am J Respir Crit Care Med
2012;185(3):286-290. https://doi.org/10.1164/rccm.201109-1644OC
7. Jollie DA, James WY, Hooper RL, etal. Prevalence, determinants and clinical correlates
of vitaminD deciency in patients with chronic obstructive pulmonary disease in
London, UK. J Steroid Biochem Mol Biol 2018;175:138-145. https://doi.org/10.1016/j.
jsbmb.2017.01.019
8. Hyun DG, Oh YM, Lee SW, Lee SD, Lee JS. Clinical phenotypes, comorbidities, and
exacerbations according to serum 25-OH vitaminD and plasma brinogen levels in
chronic obstructive pulmonary disease. J Korean Med Sci 2019;34(29):e195. https://doi.
org/10.3346/jkms.2019.34.e195
9. Kentson M, Leanderson P, Jacobson P, Persson HL. e inuence of disease severity
and lifestyle factors on the peak annual 25(OH)D value of COPD patients. Int J Chron
Obstruct Pulmon Dis 2018;13:1389-1398. https://doi.org/10.2147/COPD.S156121
10. Puhan MA, Siebeling L, Frei A, Zoller M, Bischo-Ferrari H, ter Riet G. No association
of 25-hydroxyvitaminD with exacerbations in primary care patients with COPD. Chest
2014;145(1):37-43. https://doi.org/10.1378/chest.13-1296
11. Ferrari D, Lombardi G, Ban G. Concerning the vitaminD reference range: Pre-
analytical and analytical variability of vitaminD measurement. Biochem Med (Zagreb)
2017;27(3):030501. https://doi.org/10.11613/BM.2017.030501
12. Jones G. Interpreting vitaminD assay results: Proceed with caution. Clin J Am Soc
Nephrol 2015;10(2):331-334. https://doi.org/10.2215/CJN.05490614
13. Kokturk N, Baha A, Oh YM, Young Ju J, Jones PW. Vitamin D deciency: What does
it mean for chronic obstructive pulmonary disease (COPD)? A comprehensive review
for pulmonologists. Clin Respir J 2018;12(2):382-397. https://doi.org/10.1111/crj.12588
14. Holick MF. Vitamin D deciency. N Engl J Med 2007;357(3):266-281. https://doi.
org/10.1056/NEJMra070553
15. Tenforde MW, Yadav AM, Dowdy DW, etal. Vitamin A and D deciencies associated
with incident tuberculosis in HIV-infected patients initiating antiretroviral therapy in
multinational case-cohort study. J Acquir Immune Dec Syndr 2017;75(3):e71-e79.
https://doi.org/10.1097/QAI.0000000000001308
16. Kamphuis LS, Bonte-Mineur F, van Laar JA, van Hagen PM, van Daele PL. Calcium and
vitaminD in sarcoidosis: Is supplementation safe? J Bone Miner Res 2014;29(11):2498-
2503. https://doi.org/10.1002/jbmr.2262
17. Kalichuran S, van Blydenstein SA, Venter M, Omar S. Vitamin D status and COVID-19
severity. South Afr J Infect Dis 2022;37(1):359. https://doi.org/10.4102/sajid.v37i1.359
18. Persson LJP, Aanerud M, Hiemstra PS, Hardie JA, Bakke PS, Eagan TML. Chronic
obstructive pulmonary disease is associated with low levels of vitaminD. PLoS ONE
2012;7(6):e38934. https://doi.org/10.1371/journal.pone.0038934
19. Singh D, Agusti A, Anzueto A, etal. Global strategy for the diagnosis, management, and
prevention of chronic obstructive lung disease: e GOLD science committee report
2019. Eur Respir J 2019;53(5):1900164. https://doi.org/10.1183/13993003.00164-2019
20. Avci E, Demir S, Aslan D, Nar R, enol H. Assessment of Abbott Architect 25-OH
vitaminD assay in dierent levels of vitaminD. J Med Biochem 2020;39(1):100-107.
https://doi.org/10.2478/jomb-2019-0039
21. Holick MF, Binkley NC, Bischoff-Ferrari HA, etal. Evaluation, treatment, and
prevention of vitaminD deciency: An Endocrine Society clinical practice guideline.
J Clin Endocrinol Metab 2011;96(7):1911-1930. https://doi.org/10.1210/jc.2011-0385
22. Gawron G, Trzaska-Sobczak M, Sozańska E, Śnieżek P, Barczyk A. Vitamin D
status of severe COPD patients with chronic respiratory failure. Adv Respir Med
2018;86(2):78-85. https://doi.org/10.5603/ARM.2018.0010
23. Mogire RM, Mutua A, Kimita W, etal. Prevalence of vitaminD deciency in Africa: A
systematic review and meta-analysis. Lancet Glob Health 2020;8(1):e134-e142. https://
doi.org/10.1016/S2214-109X(19)30457-7
24. Moberg M, Elango P, Ferrucci L, Spruit MA, Wouters EF, Rutten EPA. Vitamin D
deciency and airow limitation in the Baltimore Longitudinal Study of Ageing. Eur J
Clin Invest 2015;45(9):955-963. https://doi.org/10.1111/eci.12498
Received 4 May 2023. Accepted 20 May 2024. Published 11 October 2024.