Role of Smoking in Androgenetic Alopecia: A Systematic Review

Abstract

Smoking and its role in Androgenetic Alopecia has long been debated. Smoking may lead to hair loss by vasoconstriction, by forming DNA adducts, free radical damage to hair follicle, by enhancing senescence and hormonal effects. We have reviewed the available literature on AGA and smoking. Data available show that there is a significant association between smoking and AGA. However, studies demonstrating the benefit of avoidance of smoking in improving hair loss are lacking. Furthermore, large controlled studies with histological documentation are still unavailable to affirm the findings.

INTRODUCTION

Smoking has generally been considered inimical to hair growth and its role in hair loss has long been debated. This article reviews the literature to analyze the role of smoking in hair loss, possible mechanism of how smoking affects hair, and the available evidence for such an association.

METHODS OF STUDY

A thorough research was made in February 2021 using databases of PUBMED/MEDLINE and CINAHL. The search item used was “smoking” or “nicotine” and “hair loss” or “androgenic alopecia (AGA).” Exclusion criteria include studies correlating smoking and premature graying of hairs, smoking, and association with frontal fibrosing alopecia, animal-based studies and studies whose literature was not in English. Total 15 trials and one review literature were found suitable and have been reviewed.

Tobacco smoke composition

Tobacco smoke has both a solid particulate phase and volatile gas phase. The main contents of solid phase include nicotine, phenol, catecho, quinolone, aniline, toluidine, nickel, N-nitrosodimethylamine, benzanthracene, benzopyrenes, and 2-naphthylamine. The main toxic contents of gas phase include carbon monoxide, carbon dioxide, nitrogen oxides, acetone, hydrogen cyanide, acrolein, ammonium, pyridine, 3 vinyl pyridine, formaldehyde, N-nitrosodimethyamine, and N-nitrosopyrrolidine.[1] Many of these constituents may have a role on hair loss as discussed below.

Mechanism of absorption of nicotine and its metabolites

Nicotine and other metabolites enter the body through ingestion, inhalation, transdermal patches, topical cream, etc., Nicotine can enter hair by either absorption through blood[2,3,4,5,6] or from exposure to environmental smoke.[7,8,9] Absorption directly from blood occurs through passive diffusion from vessels present at base of the hair follicle into the growing hair cells with average blood flow determining the concentration of nicotine in hairs.[6] Nilsen et al.[9] found concentration of nicotine more in distal end of hair as compared to proximal end which was attributed to longer exposure of distal hair to environmental smoke. Based on this, studies have suggested that the level of nicotine in hair can be a possible marker of long-term smoke exposure.[2,3,4,5]

Mechanisms of smoking leading to hair loss

There are multiple mechanisms by which smoking can accelerate hair loss [Table 1].

Table 1.

Mechanism of action by which smoking causes hair loss

Mechanism of action

Vasoconstrictor action by smoke metabolites

Retention of DNA adducts causing DNA damage

Oxidative stress and disequilibrium of antioxidant system

Sustained microinflammation causing perifollicular fibrosis

Desensitization of nicotinic acetylcholine receptors

Androgen-dependent hair thinning

Vasoconstrictor effect

Vasoconstriction of cutaneous microvasculature by nicotine metabolites is an important effect of acute and chronic smoking. Nicotine acts by neither amplifying the vasoconstrictor effect of nor epinephrine and also attenuating acetylcholine-induced endothelium-dependent skin vasodilation.[10] A study on preauricular flaps for hair restoration documented higher complication rate in smokers as compared nonsmokers. Nicotine-induced vasoconstriction is therefore regarded as a risk factor postoperative necrosis after flap surgery.[11] It is now a routine advice to patients to stop smoking before hair transplantation, even though studies specifically in hair transplantation are lacking.

Effect on DNA

Cigarette smoking results in retention of DNA adducts which can cause DNA damage of both nuclear and mitochondrial DNA.[12] Cigarette smoke induces imbalance is protease/anti-protease systems which play an important role in extracellular matrix remodeling during hair follicle regression (catagen phase), thus affecting hair follicle growth cycle.[13]

Effect on free radicals and antioxidants

Many studies are available in literature to substantiate the effect of smoking on free radicals and the redox system which is important in hair cycle. Smoke metabolites can induce oxidative stress by generation of free radicals and also cause imbalance of the antioxidant systems. This may cause follicular keratinocytes to release immunomodulatory cytokines such as interleukin (IL)-1 alpha, IL-1beta, and TNF alpha which are potent inhibitors of hair follicle growth.[14] Free radicals, in particular reactive oxygen species (ROS), interact with nucleic acids (both in the mitochondria and in the nucleus) resulting in mutation predisposing DNA strands to break.[15] Oxidation of lipids by ROS can also cause premature cell death by damaging phospholipid cellular membranes,[15] thus affecting hair follicle molecular balance.

Smoking-induced free radical generation may cause early apoptosis of hair follicle resulting in early catagen phase. Naito et al.[16] using TUNEL staining found that lipid peroxides can induce apoptosis of hair follicle cells and of human epidermal keratinocytes by up-regulating apoptosis-related genes. They also observed that topical application of linolein hydroperoxides one of the lipid peroxides causes early catagen phase in murine hair cycle.

Smoking may enhance senescence

Bahta et al.[17] cultured dermal hair papilla cells (DPCs) from balding and nonbalding scalp and demonstrated that balding DPCs grew slower in vitro than nonbalding DPCs. The finding of premature senescence of balding DPCs along with the expression of oxidative stress markers and DNA damage in vitro suggests that balding DPCs are sensitive to environmental stress. Nicotine can cause sustained microinflammation of the hair follicle, inflammatory cell infiltrate, connective tissue remodeling, and interplay of collagenases culminating into perifollicular fibrosis.[18] Nicotine can cause overstimulation of cellular nicotinic acetylcholine receptors leading to their desensitization of receptors. This leads to hair follicle destruction by activation of programmed cell death pathways present in keratinocytes.[19]

Hormonal effects

Smoking may lead to an increase in androgen-dependent hair thinning. The hyp-estrogenic state may be due to increased hydroxylation of estradiol and inhibition of the enzyme aromatase.[20] Smoking may cause increased levels of androgens, so possibility of AGA can be correlated to increased testosterone levels.[21]

Effect of smoking on hair loss

A number of studies have been published on this subject both in animals and humans. While many support the view that smoking induces hair loss, others have not supported this view. These are reviewed below.

Studies supporting the role of smoking in inducing hair loss

A number of studies have been published which demonstrate the causative role of smoking in hair loss, which are reviewed below [Table 2].

Table 2.

Studies showing positive correlation between smoking and hair loss

Author	Place of study	Year of publication	Study design	Sample size	Duration of study	Study subject details	Smoking status	Observation and results
Mosley and Gibbs[22]	United Kingdom	1996	Controlled retrospective trial	n=606 (male=268, female=338)	3 months	Age group>30 years 152 males and 152 females were smokers. Hamilton badness scale. Category III and above regarded as bald. Bald smokers male 103; female 2 Bald nonsmokers male 63, female 2	Not included	Odds ratio for association of smoking and baldness in men was 1.93 (95% confidence interval 1.13-3.28). For females, corresponding calculation was not done
Su and Chen[23]	Taiwan	2007	Uncontrolled retrospective trial	n=740	2 months (April 2005-June 2005)	Mean age 65 years Norwood and Ludwig classification system. In age group 40-49, 50-59, 60-69, and >70, the age specific prevalence of Norwood type A variant were 5.1, 1.1, 8.1, and 7.4%, respectively, and those of female pattern AGA were 0.0, 2.1, 1.6, and 2.4%, respectively	Smoking status categorized as never, quit, current smoker of <20 cigarettes and smokers of >20 cigarette a day	Smokers had an increased risk of moderate or severe AGA (Norwood type 4) and smoking status (OR, 1.77; 95% CI, 1.14-2.76), current cigarette smoking of 20 cigarettes or more per day (OR, 2.34; 95% CI, 1.19-4.59), and smoking intensity (OR, 1.78; 95% CI, 1.03-3.07)
Fortes et al.[24]	Italy	2017	Uncontrolled, retrospective trial	n=351 (237 males, 114 females)	1 year (2010-2011)	Mean age: 35 years Male: 133 subjects with mild (Hamilton-Norwood classifications I-III) and 104 subjects with moderate and severe (Hamilton Norwood classification IV-VII). Female: 104 subjects with mild (Ludwig type 1) and 10 subjects with moderate/severe (Ludwig types II and III)	Subjects categorized as never, quit, current smokers of <10 cigarettes and smokers of >10 cigarette per day	Subjects with severe AGA smoked more as compared to less severe AGA (17.5 versus 7.6%, P=0.05). Furthermore, AGA subjects who were heavy smokers had almost 3 times an increased risk of having a moderate/severe alopecia in comparison to who never smoked
Salem et al.[25]	Egypt	2021	Controlled prospective trial	n=1000 (smokers 500; nonsmokers 500)	1 year 6 months (January 2018-September 2019)	Mean age 25 years of smokers group 235 (47%) had grade 3 AGA and 120 subjects (24%) had grade4 AGA. In the nonsmoker group, 100 subjects (20%) had grade 2 AGA and 50 subjects (10%) had either grade 3 or 4 AGA	Smokers were who smoked >10 cigarette over >1 year	Majority of smokers (425) had a form of AGA while only (200) nonsmokers had a degree of AGA (P<0.01)
Park et al.[26]	Korea	2016	Controlled retrospective trial	n=1884 (male=915, female=969)	2 years (October 2012-December 2014)	Mean age of 56.6 years 991 was AGA patients (613 males, 378 females) of the 991 AGA patients, 443 subjects were classified as having mild AGA, 381 had moderate AGA, and 167 had severe AGA. BASP classification was used to evaluate hair loss	Divided as presence or absence of smoking	Smokers in AGA group 334/991 (33.7%) were higher than the non-AGA group 214/892 (24%) (P=0.0000). Subjects with more smoking had more severe form of AGA [mild: 125/443 (28.2%), moderate: 143/381 (37.5%), severe 66/167 (39.5%) (P=0.004)
Gatherwright et al.[27]	Ohio	2012	Controlled retrospective trial	n=98 female	2 years (2009-2011)	98 female identical twins (49 sets) mean age 53 years	Not included	Frontal hair loss noted less in twins who had never smoked (P=0.021; n=8). Increased frontal hair thinning was also significant in smokers. Temporal hair loss more in smokers (P=0.096, n=10)
Vora et al.[28]	Gujarat, India	2019	Controlled prospective trial	n=100 male	2 years (June 2014-May 2016)	50 subjects with AGA and 50 age-matched control subjects were included with a mean age of 29.32 and 31.5 years, respectively. The degree of androgenic alopecia was based on the Norwood scale (3-7). AGA developing before 36 years of age and reaching at least stage 3 of Hamilton-Norwood classification is termed as early onset AGA	Not included	The prevalence of smoking was higher in male patients with early onset AGA. 26 (52%) as compared to controls 11 (22%) (P=0.002)
Gatherwright et al.[29]	Ohio	2013	Controlled retrospective trial	n=92 male	2 years (2009-2011)	92 male identical twins (46 sets). Mean age 52 years	Not included	Increased smoking duration and frontal hair loss showed positive co relation (P<0.001), but quantity of smoking was not statistically significant. Twins who smoked had more vertex hair loss than their nonsmoking siblings (P=0.047, n=20)

AGA – Androgenetic alopecia; OR – Odds ratio; CI – Confidence interval; BASP – Basic and specific

In an animal study, D’Agostini et al.[30] showed that C57BL/6 mice exposed to cigarette smoke-developed hair loss, while mice that were exposed to smoke and given N-acytlcysteine and sham-exposed mice developed no alopecia. Smoke-exposed mice had extensive atrophy of the epidermis, reduced thickness of the subcutaneous tissue. Furthermore, there was a decrease in density of hair follicle mostly present in dystrophic anagen phase.

There are several studies which have looked into different aspects of smoking and AGA: smoking, number of cigarettes smoked, AGA, severity of AGA, onset of AGA, associated factors such as obesity and gender.

The first to report significant relationship between smoking and hair loss in humans was a study by Mosley and Gibbs[22] in the United Kingdom. Their study was designed to test the hypothesis that premature hair changes and hair loss may casually be related to smoking. They studied 606 patients aged more than 30 years of age who visited surgical outpatient department (OPD) over 3 months. The odds ratio for the association of smoking and baldness was 1.93 (95% confidence interval [CI]: 1.13–3.28). Another observation was made in a 52-year-old identical male twin pair with one brother being heavy smoker having balding while nonsmoker twin doing pretty well.[13]

Studies have also looked at amount of smoking and AGA. In 2005, a population-based cross-sectional study of Asian men 40 years or older was done. They observed that for the development of moderate to severe AGA, smoking status, current amount of cigarette smoking, and smoking intensity were important. There were statistically significant positive associations between moderate or severe AGA and smoking status (odds ratio [OR], 1.77; 95% CI, 1.14–2.76), current cigarette smoking of 20 cigarettes or more per day (OR, 2.34; 95% CI, 1.19–4.59), smoking intensity (OR, 1.78; 95% CI, 1.03–3.07) after adjusting for age and family history of AGA. A statistically significant association between smoking status and moderate or severe AGA (OR, 1.63; 95% CI, 1.00–2.65) was present even when age, family history of AGA, dyslipidemia, and betel nut chewing variables were adjusted.[23]

It is well known that AGA is associated with metabolic syndrome. A study from Italy done by Fortes et al.[24] concluded that overweight and smoking are associated with increased severity of androgenetic alopecia. Heavy smokers (≥10 cigarettes per day) had almost three times an increased risk of having a moderate/severe alopecia (OR: 2.56; 95% CI: 1.27–5.16) in comparison with the ones who never smoked.[24]

Another study from Egypt with sample size of 1000 patients attending dermatology OPD found a statistically high significant difference (P < 0.001) for androgenic alopecia among smokers and nonsmokers. Importantly, the study found an association between smoking and more severe grade of AGA. More patients in the smoking group (425) had AGA, while among nonsmokers, only 200 patients had AGA. In smokers group, as many as 71% had grade III or more severe grade of AGA (235 [47%] had grade III AGA and 120 subjects [24%] had grade IV AGA). In the nonsmokers group, only (10%) had either grade III or IV AGA. However, they did not find significant association between cigarette smoked per day (in packs) and severity of androgenic alopecia.[25]

However, a controlled retrospective from Korea demonstrated smokers with more severe smoking had more severe grades of AGA. The study done from 2012 to 2014 investigated 1884 subjects visiting hospital on the basis of a questionnaire, anthropometric measures, and blood test.[26]

A controlled retrospective trial compared the relationship of AGA, smoking gender in 154 subjects. The study found tobacco use was higher in males with androgenic alopecia (P = 0.48) than control group. Interestingly, hair loss was more common in female nonsmokers (P = 0.09).[31] A 98 identical female twin study has been reported from Ohio which was based on extensive questionnaire and sputum testosterone levels. Frontal hair loss and temporal hair loss were statically significant in twins who smoked.[27]

A controlled prospective trial studied the onset of AGA and hair loss in 100 males. Smoking was higher in men with early-onset AGA as compared to the control group (P = 0.002).[28]

A study in 92 male twins in Ohio analyzed frontal, temporal, and vertex thinning separately and found that increased smoking duration was significantly associated with increased frontal hair loss (P < 0.001) and increased vertex hair thinning as well (P = 0.047).[29]

A review of literature was published in 2020 regarding relationship between smoking, premature graying of hairs, and alopecia both androgenic alopecia and frontal fibrosing alopecia. They concluded a positive correlation between length of smoking and alopecia severity in patient of AGA.[19]

Key points

Thus, there are a number of studies including controlled trials supporting the causative role for smoking in hair loss. There are studies which have found association with quantity of smoking, onset of AGA, gender, pattern of AGA, etc., However, it is important to note that the authors could not find any study which demonstrated improvement in hair growth after cessation of smoking in these patients. Furthermore, while hair transplant patients are routinely advised to stop smoking before hair loss, there is no publication which studied the effect of smoking on hair growth after hair transplantation.

Studies which failed to demonstrate association between smoking and hair loss

The authors also reviewed a few studies which failed to demonstrate an association between smoking and AGA [Table 3].

Table 3.

Studies showing negative co relation between smoking and hair loss

Author	Place of study	Year of publication	Study design	Patient	Duration of study	Study subject details	Smoking status	Observation and results
Matilainen et al.[32]	Finland	2003	Controlled retrospective trial	n=324 female	Part of cohort study which was started in 1990	Ludwig’s scale. Modification was used to asses hair status. The prevalence of extensive hair loss at least grade II or III on Ludwig’s scale was high (31.2%)	Not included	Both groups of normal hair and extensive hair loss, prevalence of smoking was low and equal (P=0.918)
Severi et al.[33]	Australia	2003	Controlled retrospective trial	n=1390 men	Age stratified, population-based case-control study. Cases who were diagnosed Adenocarcinoma of prostrate before 70 years during 1994-1997 were included	Interviewers scored AGA according to the Hamilton-Norwood scale as described below. No balding n=350 (Hamilton-Norwood stage I), front balding n=447 (Hamilton-Norwood stage II, III, IIIa, and Iva) vertex balding n=238 (Hamilton Norwood stage III-vertex and 5) and frontal AGA with vertex AGA n=335 (Hamilton Norwood stage 4, 5, 5a, 6, and 7)	Defined as either nonsmokers, current smokers, or ex-smokers	Prevalence of frontal only AGA was higher in nonsmokers (199/447) than in current (73/447) and ex-smoker groups (175/447). Vertex only AGA was higher in nonsmokers (106/238) than in current (41/238) and ex-smokers (91/238). Frontal AGA along with vertex AGA was higher in nonsmokers (173/335) than current (47/335) and ex-smokers (135/335) (P>0.05)
Nargis et al.[21]	Manglore	2017	Uncontrolled prospective study	n=103	1.6 years	The questionnaire was based on classification of AGA, age at onset of AGA, duration of hair loss, smoking, and other risk factors	Smoking status classified as never or current and further categorized as <20 cigarette and >20 cigarette	Among 41 smokers, 22 smoked<20 cigarette/day and had AGA before 35 years (P=0.014) Moreover, in group of >20 cigarettes per day, 13 had onset before 35 years and 6 patients had onset after 35 years, statically insignificant
Danesh- shakiba et al.[34]	Iran	2020	Controlled prospective trial	n=512 male	8 months (January 2018-September 2018)	256 patients with AGA and 256 age-matched control subjects were included with mean age of 38.3 and 38.4 years, respectively. The Norwood-Hamilton scale was used to determine the presence and severity of AGA by single physician	Current smokers who smoked cigarette in past 30 days	Smoking was not statistically significant among AGA subjects 72 (28.1%) compared to control subjects 66 (25.8%) (P=0.550). Smoking was not significantly more common in patients with mild-to-moderate hairless
Salman et al.[35]	Turkey	2017	Controlled retrospective trial	n=954 (419 males, 535 females)	8 months (October 2013-May 2014)	Average age of 37.7 years. AGA prevalence found was 67.1% (n=281) in men and 23.9% (n=128) in women. Severity of AGA was evaluated with Norwood-Hamilton scale in men and Ludwig classification in women	Not included	No significant difference was seen between smokers and nonsmokers
Lai et al.[36]	Taiwan	2013	Controlled prospective trial	n=354 male	5 years	Age range was 35-65 years. The Norwood-Hamilton scale was employed to asses severity	Not included	In subjects with Hamilton-Norwood scale score of I-III, 175 (59.5%) subjects were smokers and 119 (40.5%) subjects were nonsmokers. Subjects with a Hamilton- Norwood scale of IV-VII, 35 (58.3%) subjects were smokers and 25 (41.7%) were nonsmokers

AGA – Androgenetic alopecia

A study from Finland on female pattern hair loss and its association with insulin resistance linked different parameters and with paternal and maternal family history of alopecia. The study failed to demonstrate statistically significant relation between smoking and hair loss in women.[32]

A population-based case-controlled study for androgenic alopecia and associated risk factors among middle-aged men failed to establish relation between smoking and hair loss.[33] On the contrary, the study reported a lower risk of AGA among current smokers (OR, 0.86; 95% CI, 0.54–1.38) and ex-smokers (OR, 0.91; 95% CI, 0.65–1.29), although the results were not statistically significant.

A 2015 study from India on early-onset androgenic alopecia and associated risk factors in 103 patients failed to demonstrate significant relation between onset of AGA and smoking status. Quantity of cigarettes smoked was studied, but no correlation was found.[21]

A case–control study in 256 men with AGA and 256 age-matched healthy controls from Iran in 2018 also failed to establish any association between smoking and AGA.[34] A study from turkey in 954 patients[35] and a controlled prospective trial in 354 Taiwanese male subjects also found no significant association.[36]

Summary

Androgenetic alopecia is thought to be caused by twin factors of heredity and androgens. Other factors such as environmental factors are also regarded as important, and smoking is believed to be one of them. This review shows that smoking may have a significant role to play in androgenetic alopecia. There is also some evidence to support a relation between amount of smoking and severity of baldness and also age of onset of AGA. There are a number of plausible mechanisms through which smoking could affect hair loss. The dermatologist should therefore advise patients of AGA to stop smoking.

However, there are also no large controlled studies. Proper histological studies are also lacking. There are few other studies which do not demonstrate such a role. Importantly, there are no studies to suggest that hair loss improves after stopping smoking. The authors, therefore, feel that larger controlled studies, which include histopathological studies and longer follow-up and which demonstrate the improvement in hair loss after stopping smoking are needed to conclusively prove this causation.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.