Cytometry Part B (Clinical Cytometry) 52B:32–36 (2003)

Normal Values of CD4 and CD8 Lymphocyte Subsets in

Healthy Indian Adults and the Effects of Sex, Age,
Ethnicity, and Smoking
S. S. Uppal,1* Shashi Verma,2 and P. S. Dhot3
1
Clinical Immunology Center and Laboratory, Command Hospital, Pune, India; and Department of Medicine,
Faculty of Medicine, Kuwait University, Kuwait
2
Department of Biostatistics, National Institute of Virology, Pune, India
3
Blood Transfusion Department, Armed Forces Medical College, Pune, India

Background: Information on lymphocyte populations (T, B, and natural killer cells) and subpopulations
(CD4 and CD8) in India is generally lacking. Measurement of T-cell subsets is important in India for
evaluating disease stage and progression in individuals with the human immunodeficiency virus (HIV).
Hence, this study was conducted to provide normal ranges of absolute and percentage values of CD4 and
CD8 T-lymphocyte subsets and the ratio of CD4 to CD8 in normal Indian adults.
Methods: Flow cytometric analysis (EPICS-XL) was used to determine the range of T-lymphocyte sub-
populations in normal Indian blood donors at Command Hospital and the Armed Forces Medical College,
Pune, India. The reference population consisted of 94 healthy HIV-seronegative blood donors. T-lymphocyte
subsets were analyzed with two-color immunophenotyping of peripheral blood lymphocytes with the use of
a lysed whole-blood technique and enumerated.
Results: For normal values of various blood components, we found mean values of 2114 cells/␮l for total
lymphocytes, 865 cells/␮l (40.2%) for CD4ⴙ lymphocytes, 552 cells/␮l (31.3%) for CD8ⴙ lymphocytes, and
1.7 for the CD4:CD8 ratio. The 95% confidence intervals for the same parameters were 1115– 4009 cells/␮l,
430 –1740 cells/␮l (30.75– 49.60%), 218 –1396 cells/␮l (20.06 – 42.52%), and 0.39 –3.02 respectively.
Females had significantly higher CD4 counts (P < 0.05), percentage of CD4 lymphocytes (P < 0.01), and
CD4:CD8 ratio (P < 0.01). Males had a significantly higher percentage of CD8 lymphocytes (P < 0.01). They
also had higher CD8 counts that did not reach significance. Age, ethnicity (Dravidian versus Aryan), smoking,
alcohol consumption, and the interval between drawing the blood sample and its analysis were factors that
did not produce statistically significant differences in the T-cell subsets studied.
Conclusions: When compared with other published series, the CD4 and CD8 values in healthy Indians were
no different from those reported in the West. These observations have important clinical implications for the
use of T-lymphocyte subset measurements in India, especially in the management of HIV infection. The
normal ranges established by this study can be used as a reference for decisions made in clinical practice.
Cytometry Part B (Clin. Cytometry) 52B:32–36, 2003. © 2003 Wiley-Liss, Inc.

Key terms: adult; flow cytometry; CD4 lymphocyte count; CD8 lymphocyte count; India; normal ranges

Information on lymphocyte populations (T, B, and nat-
ural killer cells) and subpopulations (CD4 and CD8) in
India is generally lacking. The enumeration of circulating
CD4 and CD8 lymphocyte subsets is important in India for
evaluating disease stage and progression in individuals
with the human immunodeficiency virus (HIV). To evalu-
ate for HIV, it is necessary to determine normal values and
variations thereof in the reference Indian population after
careful screening for the absence of diseases that could
alter these values. It is also of great interest to know how
our target population differs from other populations stud-
ied across the world, and whether such differences need
to be taken into account while interpreting data with
regard to the immune status of such individuals in the
Indian setting.
*Correspondence to: S. S. Uppal, M.D., F.I.C.P., Visiting Professor,
Department of Medicine, Faculty of Medicine, Kuwait University, P.O.
Box 24923, Safat 13110, Kuwait.
E-mail: uppalss@hsc.kuniv.edu.kw
Received 24 July 2002; Accepted 8 November 2002
Published online in Wiley InterScience (www.interscience.wiley.com).
DOI: 10.1002/cyto.b.10011

© 2003 Wiley-Liss, Inc.

 CD4 AND CD8 NORMAL RANGES IN INDIAN ADULTS
Table 1
Normal Values of Various Blood Components in
Healthy Individuals (N ⫽ 94)
Blood component Mean 95% Confidence interval
Hemoglobin (g/dl) 13.5 10.14–16.76
White blood cell
count 6564 4038–10.670
Absolute lymphocyte
count 2114 1115–4009
CD4 865 430–1740
% CD4 40.2 30.75–49.60
CD8 552 218–1396
% CD8 31.3 20.06–42.52
CD4:CD8 ratio 1.7 0.39–3.02
Several studies have evaluated CD4 and CD8 lympho-
cyte subsets in children (1– 6) and adults in Western na-
tions (7–15) and other countries such as Singapore (1),
Thailand (16), Malaysia (17,18), Saudi Arabia (19), Poland
(20), and Papua New Guinea (21). Except for a small study
from Manipur in northeast India (22) and another from
south India using kits based on enzyme-linked immunosor-
bent assays (23), data on normal ranges of CD4 and CD8
T-cell subsets in India are generally lacking. Hence, this
study was conducted to establish the normal absolute and
percentage values of CD4 and CD8 T-lymphocyte subsets
and the ratio of CD4 to CD8 cells in normal Indian adults
by studying healthy, volunteer, noncommercial blood do-
nors.
MATERIAL AND METHODS
Ninety-four healthy, adult, blood donors coming to the
Blood Transfusion Department, Armed Forces Medical
College, Pune, participated in the study after giving in-
formed consent. All donors were carefully screened for
the absence of any disease or illness. All were negative for
HIV, hepatitis B surface antigen, and venereal disease.
Data on smoking habits, ethanol use, and time from last
meal were recorded on a predesigned form before blood
collection. Flow cytometric enumeration of T-cell subsets
(CD4 and CD8) was done in all subjects with an EPICS-XL
Coulter flow cytometer (Coulter Cytometry, Hialeah, FL).
Blood specimens were collected by venipuncture and
anticoagulated with ethylenediamine tetraacetic acid (2
mg/ml). Anticoagulated blood samples were tested within
1 h of collection. For each sample, two 12- ⫻ 75-mm test
tubes were labeled for monoclonal antibody tests and
appropriate isotypic controls, respectively. One hundred
microliters of anticoagulated blood was pipetted into the
bottom of each properly labeled test tube, thus ensuring
that the inside surface and top of the tube were free of
blood. Next, 10 ␮l of Cyto-Stat/CoulterClone T4-RDI/T8-
fluorescein isothiocyanate and 10 ␮l of idiotypic control
(Cyto-Stat/CoulterClone) were added to the test and con-
trol tubes, respectively, vortexed, and incubated for 10
min at room temperature. Next, the tubes were placed in
the Coulter Q Prep Workstation and the 35-s cycle was
initiated to lyse the red blood cells with ImmunoPrep A
(formic acid, 1.2 ml/l), stabilize the leucocytes with Im-
33
munoPrep B (sodium carbonate, 6 g/l), and fix the cell
membranes with ImmunoPrep C (paraformaldehyde, 10
g/l). Then the prepared cells were analyzed on the Coulter
EPICS-XL flow cytometer by using dual fluorescence anal-
ysis, properly standardized (24) and gated on lympho-
cytes. Briefly, this consisted of collecting a 90-degree side
scatter (SS) versus forward scatter (FS) histogram, gating
on the lymphocyte population, collecting log-integrated
green and red fluorescences gated on lymphocytes of
90-degree SS versus FS, determining the percentage of
positively stained cells by integrating a clear fluorescence
peak (single color), and obtaining quadrant statistics (dual
color). The Cyto-Stat/CoulterClone T4-RDI/T8-fluorescein
isothiocyanate/T3-phycoerythrin reagent was used for
one sample per run for performing a three-color analysis
to cross check the accuracy of the two-color procedure. A
positive control in the form of Coulter Cyto-Trol control
cells (lyophilized lymphocytes with a known quantity of
CD4 and CD8 surface antigens) also was used to cross
check the results. Total white blood cell counts (WBCs),
absolute lymphocyte counts (ALCs), and hemoglobin val-
ues were determined on a Coulter machine (AcTDiff), and
then absolute values of CD4 and CD8 cells were calcu-
lated by multiplying the subject’s ALC by the percentage
of the particular T-cell subset obtained by flow cytometry.
Statistical analysis was done with regression coefficients
and paired t-tests. For two groups, t-test was applied to
compare means; for more than two groups, single-factor
analysis of variance was applied. Some data transforma-
tions were necessary before analysis. Percentages of CD4
and CD8 cells were transformed to arcsin (angle) values,
and WBC, ALC, CD4, and CD8 were transformed to log10.
Thus, the mean ⫾ two standard deviations for the latter
parameters are expressed as 95% confidence intervals to
match the actual values by taking antilogs of the estimates
calculated under transformation. Correlation coefficients
were computed to check for linear trend while calculating
the effect of the time since the last meal on various values.
Microsoft Office Excel was used to compile and analyze
the data.
RESULTS
Ninety-four adults, 39 of whom were female, were the
subjects of this study. The age ranges were 18 –74 years
(mean, 40.5 years) for the females and 20 – 68 years
(mean, 43.1 years) for the males.
Table 2
Normal CD4 and CD8 Values in the Present Study Versus the
Coulter Reference Range
% Positivity Cells/␮l
Coulter Present Coulter Present
Cell type range study range study
CD4 20–65 31–50 500–2000 430–1740
CD8 5.0–55.0 20–43 350–1750 218–1396
CD4:CD8 ratio 1.0–2.0 0.4–3.0
34 UPPAL ET AL.

Table 3
Means and 95% Confidence Intervals of Various Parameters in Females and Males

Blood Females (n ⫽ 39) Males (n ⫽ 55)

component Mean 95% CI Mean 95% CI P
% Hb 12.22 10.92–13.52 14.33 a
13.04–15.62 ⬍0.01
WBC 7015a 4766–10,325 6262 3691–10.625 ⬍0.05
ALC 2201 1207–4013 2055 1056–3999 NS
CD4 963a 511–1817 802 395–1627 ⬍0.05
% CD4 41.71a 37.61–45.81 39.09 34.24–43.94 ⬍0.01
CD8 530 264–1065 568 196–1646 NS
% CD8 29.71 25.45–33.97 32.41a 26.21–38.61 ⬍0.05
CD4:CD8 ratio 1.92a 1.32–2.52 1.55 0.89–2.21 ⬍0.01
*ALC, absolute lymphocyte count; CI, confidence interval; Hb, hemoglobin; NS, not significant; WBC,
white blood cell count.
a
Statistically higher mean.

Normal Values of Various Blood Components Obtained Influence of Ethnicity

The mean ⫾ two standard deviations (being the same as
95% confidence intervals) of various hematologic and im-
munologic parameters (irrespective of age, sex, or influ-
ence of other factors) are shown in Table 1, with the
important mean values being 2114 cells/␮l for ALC, 865
cells/␮l (40.2%) for CD4, 552 cells/␮l (31.3%) for CD8,
and 1.7 for CD4:CD8. Before the establishment of normal
values for our population, our laboratory had been using
the range of values recommended by the Coulter Corpo-
ration, the manufacturers of the EPICS-XL flow cytometer.
Comparison of these recommended values with the refer-
ence range obtained in our study is shown in Table 2.
Influence of Sex
Results were analyzed separately for males and females
(Table 3), and females had significantly higher absolute
WBCs (P ⬍ 0.05), CD4 counts (P ⬍ 0.05), percentages of
CD4 cells (P ⬍ 0.01), and CD4:CD8 (P ⬍ 0.01). Males had
significantly higher percentages of CD8 cells (P ⬍ 0.01)
and higher CD8 counts that did not reach significance. As
expected, they also had significantly higher hemoglobin
levels (P ⬍ 0.01) than did females.
Influence of Age
The subjects were grouped by age: 18 –27 years, 28 –37
years, 38 – 47 years, 48 –57 years, and ⱖ58 years. How-
ever, none of the parameters differed significantly in any
age group, implying that in adulthood age had no signifi-
cant influence on the various parameters in our study (P ⬎
0.05).
The Indian population can be divided broadly into Dra-
vidians (Dravidian-speaking South Indians) and the Aryans
(Indo-Aryan–speaking Hindus) according to their ethnicity
(25). When the various hematologic and immunologic
parameters were analyzed separately for these groups
(Table 4), we found no significant difference in the mean
values of any of the parameters (P ⬍ 0.05 for all parame-
ters).
Effects of Smoking, Alcohol, and Time Since Last Meal
When the examined parameters were compared be-
tween smokers and nonsmokers and between alcohol
drinkers and nondrinkers, together and separately for
males and females (because there was only one female in
each category of smoker and ethanol consumer), we
found no statistically significant difference for any param-
eter except, not surprisingly, hemoglobin, which was
higher among smokers (P ⬍ 0.01). Similarly, the interval
between the last meal and the sample collection did not
produce statistically significant differences.
DISCUSSION
Influence of Race and Ethnicity (Table 5)
Lymphocyte subset percentage ranges in adult Cauca-
sians in the United States have been reported as 28 –58%
for CD4 and 19 – 48% for CD8 cells (8). Almost similar
values have been reported from the United Kingdom (7).
Corresponding values in our study of Indians were 31–
50% for CD4 and 20 – 43% for CD8 cells. The absolute
numbers of lymphocytes and CD4 and CD8 cells were

Table 4
Means (95% Confidence Intervals) of Various Parameters in Aryan and Dravidian Subjects According to Ethnicity*
CD4:CD8
n Hb WBC ALC CD4 %CD4 CD8 %CD8 ratio
Aryan 69 13.55 6352 2052 827 39.82 540 31.46 1.66
(10.25–16.84) (3899–10,346) (1084–3886) (416–1644) (30.24–49.41) (210–1389) (20.07–42.85) (0.31–3.02)
Dravidian 9 13.53 6950 2049 873 40.91 537 31.08 1.70
(10.42–16.65) (3861–12.513) (916–4586) (363–2102) (34.18–47.64) (238–1214) (22.91–39.25) (0.65–2.75)

*P ⬎ 0.05 for all parameters. Information on ethnicity not available for 16 subjects. ALC, absolute lymphocyte count; Hb,
hemoglobin; WBC, white blood cell count.
 CD4 AND CD8 NORMAL RANGES IN INDIAN ADULTS
Table 5
Comparison of Values From the Present Study With Published Reference Values*
Present study India:
Blood Mean Singh et al. (22)
component (N ⫽ 94) 95% CI (N ⫽ 14)
ALC, cells/␮l 2114 1115–4009 NM
CD4, cells/␮l 865 430–1740 848
% CD4 40.2 30.75–49.60 36
CD8, cells/␮l 552 218–1396 427
% CD8 31.3 20.06–42.52 21
CD4:CD8 ratio 1.7 0.39–3.02 1.99
*ALC, absolute lymphocyte count; CI, confidence interval; NM, not mentioned.
1900, 830, and 560 cells/␮l, respectively, in the United
Kingdom (7). These also compared well to our values, i.e.,
2114, 865, and 552 cells/␮l, respectively. However, in one
US study, the Asian population appeared to have a lower
mean percentage of CD3 and CD4 cells, a lower CD4:CD8
ratio, and lower absolute CD4 lymphocytes as compared
with Caucasians (11). We found no such differences.
In the study published from Manipur in India (22), the
investigators measured absolute mean values of 848
cells/␮l (36%) for CD4⫹ cells and 427 cells/␮l (21%) for
CD8⫹ cells in normal healthy individuals. Their CD4 val-
ues compared well to our values, i.e., 865 cells/␮l (40.2%),
but we obtained somewhat higher CD8 values, i.e., 552
cells/␮l (31.3%).
In a Malaysian study, differences in race were observed,
with some values being significantly different in the Indi-
ans and Malays when compared with the Chinese (more
CD4 cells and higher CD4:CD8 ratio) (17). A sex differ-
ence was observed only in the Chinese population: the
CD4:CD8 ratio was significantly higher in females than in
males (18). Thus, it seems that Indian values are similar to
Caucasian values but higher than those seen in the Chi-
nese.
Overall, CD4 and CD8 values in healthy Indians seem no
different from those reported from the West.
Influence of Sex
Two studies on Asian adults (1,17), one study from
Britain (12), and one from the United States (26) have
reported that males have higher CD8 percentages than
females and that, conversely, females have higher CD4
percentages and counts than males. We also found that
females have significantly higher CD4 counts (P ⬍ 0.05),
percentage of CD4 cells (P ⬍ 0.01), and CD4:CD8 ratio
(P ⬍ 0.01), whereas males had significantly higher per-
centages of CD8 cells (P ⬍ 0.01).
Influence of Age
In a recent study of age-dependent changes with re-
spect to the expression of the major lymphocyte surface
receptors in healthy elderly subjects, T lymphocytes from
elderly individuals were found to express lower levels of
CD3 when compared with levels in young subjects (27).
Reduced T cells and lower CD4:CD8 ratios have been
described in the frail elderly (28). We discerned no age-
35
USA: UK: Coulter
Reichert et al. (8) Bofill et al. (7) reference
(N ⫽ 271) (N ⫽ 676) range
NM 1900 NM
NM 830 500–2000
43 43.6 20–65
NM 560 350–1750
33 29.5 5.0–55
1.4 1.51 1.0–2.0
related differences in our population, mainly because we
did not study the elderly as they are not relevant with
respect to HIV.
Effects of Smoking, Alcohol, and Time Since Last Meal
In a U.S. study of air force personnel, alcohol consump-
tion showed a significant correlation with suppressed
counts and functions for nearly all variables, whereas
tobacco use was associated with stimulation of T-cell
number and function (9). The absolute CD4 counts were
reported to be significantly higher in smokers than in
nonsmokers in two other studies (12,29). Smoking in-
creased the number of T cells and mainly CD4⫹ peripheral
blood lymphocytes (15). However, we found no differ-
ences related to smoking or alcohol but noted that all but
one of the smokers and alcohol consumers in our study
were male. The reasons for this difference between our
results and those of the other series are unclear. Also,
there were no differences in our study with regard to the
time interval between taking the last meal and drawing of
the blood sample.
LITERATURE CITED
1. Lee BW, Yap HK, Chew FT, Quah TC, Prabhakaran K, Chan GS, et al.
Age- and sex-related changes in lymphocyte subpopulations of
healthy Asian subjects: from birth to adulthood. Cytometry 1996;26:
8 –15.
2. Panaro A, Amati A, di Loreto M, Felle R, Ferrante M, Papadia AM, et al.
Lymphocyte subpopulations in pediatric age. Definition of reference
values by flow cytometry. Allergol Immunopathol (Madr) 1991;19:
109 –112.
3. Heldrup J, Kalm O, Prellner K. Blood T and B lymphocyte subpopu-
lations in healthy infants and children. Acta Paediatr 1992;81:125–
132.
4. Stern DA, Hicks MJ, Martinez FD, Holberg CJ, Wright AL, Pinnas J, et
al. Lymphocyte subpopulation number and function in infancy. Dev
Immunol 1992;2:175–179.
5. Robinson M, O’Donohoe J, Dadian G, Wankowicz A, Barltrop D,
Hobbs JR. An analysis of the normal ranges of lymphocyte subpopu-
lations in children aged 5–13 years. Eur J Pediatr 1996;155:535–539.
6. Yanase Y, Tango T, Okumura K, Tada T, Kawasaki T. Lymphocyte
subsets identified by monoclonal antibodies in healthy children. Pe-
diatr Res 1986;20:1147–1151.
7. Bofill M, Janossy G, Lee CA, MacDonald-Burns D, Phillips AN, Sabin C,
et al. Laboratory control values for CD4 and CD8 T lymphocytes.
Implications for HIV-1 diagnosis. Clin Exp Immunol 1992;88:243–
252.
8. Reichert T, DeBruyere M, Deneys V, Totterman T, Lydyard P, Yuksel
F, et al. Lymphocyte subset reference ranges in adult Caucasians. Clin
Immunol Immunopathol 1991;60:190 –208.
9. Wolfe WH, Miner JC, Michalek JE. Immunological parameters in
36 UPPAL ET AL.
current and former US Air Force personnel. Vaccine 1993;11:545–
547.
10. Goff LK, Habeshaw JA, Rose ML, Gracie JA, Gregory W. Normal values
for the different classes of venous blood mononuclear cells defined
by monoclonal antibodies. J Clin Pathol 1985;38:54 –59.
11. Howard RR, Fasano CS, Frey L, Miller CH. Reference intervals of CD3,
CD4, CD8, CD4/CD8, and absolute CD4 values in Asian and Non-
Asian populations. Cytometry 1996;26:231–232.
12. Maini MK, Gilson RJ, Chavda N, Gill S, Fakoya A, Ross EJ, et al.
Reference ranges and sources of variability of CD4 counts in HIV-
seronegative women and men. Genitourin Med 1996;72:27–31.
13. McNerlan SE, Alexander HD, Rea IM. Age-related reference intervals
for lymphocyte subsets in whole blood of healthy individuals. Scand
J Clin Lab Invest 1999;59:89 –92.
14. Ricalton NS, Roberton C, Norris JM, Rewers M, Hamman RF, Kotzin
BL. Prevalence of CD8⫹ T-cell expansions in relation to age in
healthy individuals. J Gerontol A Biol Sci Med Sci 1998;53:B196 –
B203.
15. Schaberg T, Theilacker C, Nitschke OT, Lode H. Lymphocyte subsets
in peripheral blood and smoking habits. Lung 1997;175:387–394.
16. Vithayasai V, Sirisanthana T, Sakonwasun C, Suvanpiyasiri C. Flow
cytometric analysis of T-lymphocytes subsets in adult Thais. Asian Pac
J Allergy Immunol 1997;15:141–146.
17. Choong ML, Ton SH, Cheong SK. Influence of race, age and sex on
the lymphocyte subsets in peripheral blood of healthy Malaysian
adults. Ann Clin Biochem 1995;32(pt 6):532–539.
18. Dhaliwal JS, Balasubramaniam T, Quek CK, Gill HK, Nasuruddin BA.
Reference ranges for lymphocyte subsets in a defined Malaysian
population. Singapore Med J 1995;36:288 –291.
19. Shahabuddin S, al Ayed IH, el Rad MO, Qureshi MI. Lymphocyte
subset reference ranges in healthy Saudi Arabian children. Pediatr
Allergy Immunol 1998;9:44 – 48.
20. Kawiak J, Rokicka-Milewska R, Zeman K, Hoser G, Derulska G, For-
nalczyk-Wachowska E, et al. Peripheral blood leukocytes and lym-
phocyte subpopulations as determined by flow cytometric measure-
ments in healthy children. Folia Histochem Cytobiol 1995;33:33–38.
21. Witt CS, Alpers MP. Lymphocyte subsets in Eastern Highlanders of
Papua New Guinea. Papua N Guin Med J 1991;34:98 –103.
22. Singh YG, Dar L, Singh NG. Levels of CD4 and CD8 among the
inhabitants of Manipur, India. J Commun Dis 2000;32:201–206.
23. Kannangai R, Prakash KJ, Ramalingam S, Abraham OC, Mathews KP,
Jesudason MV, et al. Peripheral CD4⫹/CD8⫹ T-lymphocyte counts
estimated by an immunocapture method in the normal healthy south
Indian adults and HIV seropositive individuals. J Clin Virol 2000;17:
101–108.
24. 1997 revised guidelines for performing CD4⫹ T-cell determinations
in persons infected with human immunodeficiency virus (HIV). Cen-
ters for Disease Control and Prevention. MMWR Recomm Rep 1997;
46:1–29.
25. Mehra NK, Taneja V, Kailash S, Raizada N, Vaidya MC. Distribution of
HLA antigens in a sample of the North Indian Hindu population.
Tissue Antigens 1986;27:64 –74.
26. Tollerud DJ, Clark JW, Brown LM, Neuland CY, Pankiw-Trost LK,
Blattner WA, et al. The influence of age, race, and gender on periph-
eral blood mononuclear-cell subsets in healthy nonsmokers. J Clin
Immunol 1989;9:214 –222.
27. Ginaldi L, De Martinis M, D’Ostilio A, Marini L, Loreto F, Modesti M,
et al. Changes in the expression of surface receptors on lymphocyte
subsets in the elderly: quantitative flow cytometric analysis. Am J
Hematol 2001;67:63–72.
28. Fahey JL, Schnelle JF, Boscardin J, Thomas JK, Gorre ME, Aziz N, et al.
Distinct categories of immunologic changes in frail elderly. Mech
Ageing Dev 2000;115:1–20.
29. Tollerud DJ, Clark JW, Brown LM, Neuland CY, Mann DL, Pankiw-
Trost LK, et al. The effects of cigarette smoking on T cell subsets. A
population-based survey of healthy Caucasians. Am Rev Respir Dis
1989;139:1446 –1451.