Optical coherence tomography and histological assessment of cutaneous vasculature and neural changes in long-term smokers: an exploratory study

Abstract

Smoking affects microcirculation by damaging endothelial cells and reducing blood flow. Vascular disease is also thought to contribute to the development of diabetic neuropathy, resulting in the decrease of nerve fiber density. The impact of nicotine on vasculature using optical coherence tomography angiography (OCT-A) has been extensively studied. We utilized an alternative method, reactive hyperemia OCT-A (RH-OCT-A), to provide a more complete visualization of microvasculature changes in the forearm of 5 old female smokers (mean age = 62.6yrs) and 7 young female smokers (mean age = 26.6yrs) compared against old (n = 17, mean age = 67.8yrs) and young (n = 22, mean age = 24.0yrs) nonsmoking populations. Intraepidermal nerve fiber densities (IENF) were collected from 6 old female smokers (mean age = 63.0yrs) and 12 young female smokers (mean age = 26.3yrs) and compared against 11 old female nonsmokers (mean age = 66.2yrs) and 13 young female nonsmokers (mean age = 24.2yrs). A trend was observed from vessel density measurements showing old smokers with the lowest value (15.7%), followed by old nonsmokers (16.4%), young smokers (18.8%), and finally young nonsmokers (18.9%). Additionally, the correlation within the smoking population between pack-year history and vessel density was stronger than the correlation between subject age and vessel density (R= -0.53 vs. R= -0.42). Using RH-OCT-A, we show that smoking history is associated with lower vessel density and therefore suggestive of accelerated aging.

Introduction

Aging results in the progressive reduction of vasculature and neurons in the skin^1–3. Intrinsic skin aging is determined primarily by inherent genetic factors, hormonal changes, and gradual cellular damage from crucial metabolic functions^1,3. Moreover, external and environmental factors such as ultraviolet radiation, tobacco smoking, and systemic medical disorders can affect the appearance and function of the dermal microvasculature^4–7. These factors contribute to oxidative damage and the progressive decline of collagen and elastin production which leads to a loss of skin firmness, thinning of the skin, and a reduction in the number and density of blood vessels^3,8. External influences can also lead to the formation of advanced glycation end products (AGE), which form cross-links of collagen within vascular walls, a common pathophysiology of peripheral artery diseases^9,10.

Tobacco smokers show a strong association between tobacco smoking and the production of AGEs in the skin compared to non-smokers⁹. Smoking accelerates telomere shortening with age¹¹, potentially due to an increase in oxidative stress¹². Clinical assessment of skin wrinkling in an aged cohort revealed that smoking one pack/day for 5 years is equivalent to a decade of chronological aging⁶. Smoking has a profound effect on the microcirculation by causing endothelial cell dysfunction, decreasing endothelial-dependent vasodilation and reducing blood flow¹³. Smoking cessation rapidly restores (within four weeks) some of the effects on the microcirculation as well as reduces the incidence of surgical site infection¹⁴.

Anatomically, the microcirculation composed of arterioles, capillaries, and venules provide blood flow to the skin, which can be challenging to image and quantify. In adults aged 65 and above, notable vessel reduction has been observed, particularly among the capillary loops within the dermal papillae^15–18. Since cutaneous microcirculation is regarded as being representative of microcirculation in other organ systems, assessing the health of cutaneous microvasculature may provide important insight into the vascular mechanisms in a variety of disease states¹⁹. Our group has previously demonstrated that reactive hyperemia-optical coherence tomography angiography (RH-OCT-A), a label-free imaging technique, can generate images that serve as a non-invasive alternative to histopathology for vascular analysis²⁰. RH-OCT-A measures the increase in blood flow following a brief period of vascular occlusion, allowing age-related reduction in reserve to become evident due to stress^20,21.

Also present in the superficial dermis are nerve fibers establishing close connections with the microvasculature. Both intrinsic aging and photoaging are implicated in neuronal changes within the skin of select anatomical sites^22,23. Age and gender associated decreases in epidermal innervation^24,25 can be observed in certain anatomical sites, such as the distal leg and face^26,27. The density of intraepidermal nerve fibers (IENF) is an important marker for neuronal health and is particularly reduced in disorders related to small-fiber neuropathy²⁸. Notably, subsequent investigations in humans demonstrated that vascular factors play a crucial role in the development of diabetic neuropathy, showing a strong correlation between the degree of microangiopathic changes with subsequent myelinated nerve fiber loss in diabetic nerves^29,30. Furthermore, as diabetic neuropathy becomes more severe, there is a deterioration in endothelial function, decreased expression of vascular endothelial growth factor, and progressive loss of IENF in the foot skin of affected patients³¹. While this study does not include diabetic patients, these findings suggest there is a relationship between vasculature and neuronal health and function.

Thus, we sought to investigate the effects of age, smoking habits, and UV-exposure on the skin by comparing quantitative changes in the skin’s microcirculation and intraepidermal nerve fiber density. To our knowledge, this is one of the first studies utilizing RH-OCT-A to assess smoking-related skin aging.

Methods

Study design and subjects

Female Adults (≥ 18years) of any Fitzpatrick skin type were recruited to examine the potential effects of aging (Tables 1 and 2). A group of old (≥ 50 years old) and young (18–49 years old) subjects were enrolled in the IRB-approved clinical trial. Active daily smokers within each age group were enrolled with median pack years of 50 and 6 respectively. Calculating pack years for smokers is done by multiplying the number of packs (20 cigarettes) smoked per day by the number of years someone has smoked and allows for the standard clinical quantification of smoking history. The calculation is adjusted slightly to account for vapers: the number of puffs per day is divided by 10 before multiplying by the number of years vaped. Smoking activities could include e-cigarettes or vapes with nicotine use. The old and young nonsmoker groups had no history of cigarette or e-cigarette smoking; tobacco use nor use of other types of nicotine replacement therapy in the past 15 years.

Table 1.

Impact of age and smoking on vasculature assessed with optical coherence tomography (OCT-A) and reactive hyperemia optical coherence tomography angiography (RH-OCT-A).

Cohort		OCT-A			RH-OCT-A
	Sample Size	Vessel Density (%)	Length of Vascular Network (mm)	Number of Branch Points	Vessel Density (%)	Length of Vascular Network (mm)	Number of Branch Points
Young	29	10.2 ± 3.4	5.0 ± 1.4	1619 ± 675	18.6 ± 3.5	8.2 ± 1.6	3364 ± 918
Old	22	8.6 ± 2.9	4.3 ± 1.3	1257 ± 463	16.2 ± 3.6	7.6 ± 1.9	2812 ± 970
Smoker	12	9.3 ± 3.3	4.6 ± 1.6	1413 ± 532	17.5 ± 2.9	7.6 ± 2.1	3009 ± 1053
Nonsmoker	39	9.6 ± 3.2	4.7 ± 1.4	1475 ± 641	17.8 ± 3.6	8.1 ± 1.7	3158 ± 956
Old Smoker	5	9.6 ± 2.0	4.8 ± 1.1	1407 ± 273	15.7 ± 2.2	6.8 ± 1.4	2451 ± 727
Old Nonsmoker	17	8.4 ± 3.1	4.2 ± 1.4	1215 ± 499	16.4 ± 3.8	7.7 ± 1.9	2865 ± 998
Young Smoker	7	9.0 ± 4.1	4.5 ± 1.8	1417 + 670	18.8 ± 2.5	8.3 ± 1.0	3397.4 ± 712.6
Young Nonsmoker	22	10.6 ± 3.1	5.2 ± 1.3	1682 ± 671	18.9 ± 3.0	8.4 ± 1.4	3391 ± 862

Table 2.

Impact of age and smoking on advanced glycation end products (AGEs) and intraepidermal nerve fiber density (IENF) assessed with histology.

Cohort	Nerve Fiber Density		AGE Reader
	Sample Size	IENF (mm2)	Sample Size	Age
Young	25	267.2 ± 82.4	32	2.0 ± 0.5
Old	17	183.2 ± 106.4	20	2.8 ± 0.6
Smoker	18	260.9 ± 114.3	8	2.2 ± 0.7
Nonsmoker	24	212.5 ± 85.3	44	2.3 ± 0.7
Old Smoker	6	228.1 ± 125.8	2	2.7 ± 1.0
Old Nonsmoker	11	158.8 ± 84.7	18	2.8 ± 0.6
Young Smoker	12	277.3 ± 104.2	6	2.1 ± 0.6
Young Nonsmoker	13	257.9 ± 53.5	26	2.0 ± 0.5

Three optical coherence tomography (OCT) images and one 3 mm skin punch biopsy were obtained from the inner (UV-protected) forearm and the outer (UV-exposed) forearm for a total of six OCT images and two skin biopsies. AGE accumulation was measured noninvasively on the inner forearm using an AGE Reader. Sample sizes for these different analyses (OCT, IENF, AGE) vary: Old Smoker group (n = 5, n = 6, n = 2); Young Smoker group (n = 7, n = 12, n = 6); Old Nonsmoker group (n = 17, n = 11, n = 18); Young Nonsmoker group (n = 22, n = 13, n = 26). Blood pressure measurements were taken prior to biopsy acquisition.

It is important to note that the small sample size of old smokers (n = 6) and young smokers (n = 12) is a limitation. Thus, the results of this study are considered exploratory.

OCT acquisition on the inner and outer forearms

Optical coherence tomography angiography (OCT-A) was performed to image cutaneous microvasculature. OCT-A works by delivering low-coherence light to the skin, causing scattering from moving blood cells and allowing the scanner to distinguish areas of blood flow in a 2D image^32,33. A reactive hyperemia method was used during OCT acquisition as this has been shown to provide a more comprehensive visualization of the skin microvasculature²⁰.

To acquire angiographic volumes, an OCT scanner (TEL220C1, Thorlabs) with a full contact glass OCT probe (OCT-IMM3, Thorlabs Inc., Newton, NJ) was used. OCT volumes of 6 × 6 × 1.5 mm³ were acquired at an A-scan rate of 76 kHz. A small quantity of glycerol was applied to the probe to ensure optical coupling between the probe and the skin surface. Reactive hyperemia during OCT imaging acquisition has been shown to improve visualization of the microvascular network^20,33–35 and so a series of three images were collected to increase blood perfusion. After obtaining a baseline perfusion image, mechanical pressure was applied to the skin for two minutes to occlude the blood flow. A second image was acquired while pressure was applied. Finally, a third image was collected immediately upon the release of the pressure to visualize the reactive hyperemia response. Images were taken on the inner and outer forearms for a total of six images per subject.

Vessel segmentation and image analysis³⁶ were performed on the baseline perfusion images (OCT-A) and the reactive hyperemia images (RH-OCT-A) to acquire vessel density (VD), length of vascular network (LVN), and number of branch points (NBP). Vessel density is represented as a percentage of the area of segmented vessels (red pixels) within the entire image.

AGE reader

An AGE Reader device (Diagnoptics Technologies, The Netherlands) noninvasively measured skin accumulation of advanced glycation end products (AGEs) using skin autofluorescence. Subjects’ inner forearms were placed over the measurement window of the device with the elbow aligned to the edge of the armrest. The skin over the measurement window was exposed to a light source with a wavelength of approximately 370 nm, resulting in the emission of AGEs at 420-600nm³⁷. The forearm was held steady for 12–15 s per measurement for a total of three measurements. The AGE reader expresses skin autofluorescence in arbitrary units (AU).

Histology

Skin biopsies from the inner and outer forearm were fixed in Zamboni’s solution (Newcomer Supply, Wisconsin) for up to two days and then embedded in optimum cutting temperature solution (O.C.T. compound, Tissue-Tek). Samples were frozen in a −80 °C freezer. Vertical sections of 50 μm thickness were cut and stained with PGP 9.5 antibody (ab108986, Abcam) and visualized with immunoperoxidase and SG chromogen. Counts were performed blinded to the study cohort with at least 3 sections per biopsy averaged. Intraepidermal nerve fibers were counted according to standard practice with a Zeiss Axio Microscope at 40x magnification. The number of nerve fibers crossing the dermo-epidermal junction was standardized to the length of the epidermis in compliance with published EFNS guidelines³⁸.

Statistical analysis

Welch’s T-tests and ANOVA multiple comparison tests were used to compare vessel density, AGE metrics, and nerve fiber density measurements between age groups (old and young) and disease groups (smoker and nonsmoker). Tukey’s correction was applied to the ANOVA multiple comparison tests. To evaluate the strength of associations of subject groups with vessel density, IENF density, and AGE metrics, the correlation coefficient (R) was calculated.

IRB statement

All procedures were approved by the Institutional Review Board of the Massachusetts General Hospital (Protocol No.:2018P001115) and informed consent was obtained from all subjects. All research methods were performed in accordance with relevant guidelines and regulations.

Results

Smoking sample sizes

Vessel density analysis was performed on only 5 of the old smokers and 7 of the young smokers due to abnormalities during the OCT acquisition (too much subject movement, uneven pressure applied across the imaging area) (Table 1). Histological IENF analysis was performed on all 6 old smokers and all 12 young smokers. AGE measurements were acquired from only 2 old smokers and 6 young smokers as this method was added to the study procedure after data collection for some subjects had already occurred (Table 2).

Effects of sun-exposure on vasculature and innervation

Vessel density and IENF density from the inner and outer forearm were assessed with OCT image analysis and histological evaluation respectively for all four study groups: old smokers, young smokers, old nonsmokers, and young nonsmokers. We found no significant difference in vessel density or IENF density between inner and outer forearms within individual study groups nor when subjects from all study groups were combined. Therefore, inner and outer measurements were averaged and analyzed as one dataset.

Effects of subject age and smoking on vasculature, innervation, and ages

Vessel density measurements collected from the RH-OCT-A images were used in the following analysis. Subject age correlated with vessel density (R= −0.31), IENF density (R= −0.50), and AGE measurements (R = 0.65) (Fig. 1). Furthermore, when comparing the metrics of old subjects and young subjects, we observed a significant difference in vessel density (Mean_old =16.2%, Mean_young=18.8%, p = 0.009), nerve fiber density (Mean_old =183.2mm², Mean_young=267.2mm², p = 0.01), and AGE measurements (Mean_old =2.8, Mean_young=2.0, p < 0.0001).

Fig. 1.

Impact of age on dermal aging metrics. a Vessel density shows an age-related decline (p = 0.009), with a d corresponding negative correlation (R = −0.31). b Intraepidermal nerve fiber (IENF) density also decreases significantly with age (p = 0.01), exhibiting a e negative correlation (R = −0.50). c Advanced glycation end-product (AGE) accumulation increases significantly with age (p < 0.0001) and demonstrates a f robust positive correlation (R = 0.65).

OCT images of representatives from each study cohort are shown in Fig. 2. We found no significance among smoker and nonsmoker populations between vessel densities. However, within age groups (old vs. young) the RH-OCT-A values trend higher in the nonsmoker population than in the smoker group (Fig. 2). The specific trend observed in the vessel density measurements follows that the old smoker group results in the lowest value (15.7%), followed by the old nonsmoker group (16.4%), the young smoker group (18.8%), and finally the young nonsmoker group (18.9%). Additionally, a correlation was found in the smoking population (5 old smokers and 7 young smokers) between the number of pack years and the vessel density (R=−0.53). This was stronger than the correlation in the smoking population between the subject age and vessel density (R=−0.42).

Fig. 2.

Impact of smoking and age on microvasculature. a Blood vessels are visualized with baseline perfusion images (OCT-A) and reactive hyperemia images (RH-OCT-A) from representatives of the four study cohorts: Old Smoker (age: 55years; pack-years: 43years), Old Nonsmoker (age: 50years), Young Smoker (age: 26years; pack-years: 7years), and Young Nonsmoker (age: 25 years). b A trend is observed wherein vessel density decreases with age and smoking history. c Within the smoking population (n = 12), there is a stronger correlation between RH-OCT-A vessel density measurements and pack-year history (R= −0.53) than there is between d RH-OCT-A vessel density measurements and subject age (R= −0.42).

The difference in IENF counts between smoker and nonsmoker populations was not statistically significant. The only significance observed within IENF counts was between the old nonsmokers and the young nonsmokers (p = 0.004) (Fig. 3). Notably, the nonsmoker populations within each age group (old and young) had lower IENF densities than the smoking populations (Mean_{old smoker} =228.1mm², Mean_{old nonsmoker}=158.8mm², Mean_{young smoker}=277.3mm², Mean_{young nonsmoker}=257.9mm²). Correlations between the number of pack years and the IENF as well as between the subject age and IENF were equivalent within the smoker population (R=−0.44).

Fig. 3.

Impact of smoking and age on IENF. a There is a statistically significant difference (p = 0.004) between the old nonsmoker IENF and the young nonsmoker IENF. b Within the smoking population (n = 18) the correlation observed between IENF and pack-year history (R=−0.44) is equivalent to c the correlation between IENF and subject age (R=−0.44).

We also looked for relationships between vessel density, IENF, and AGE readings. The strongest correlations were observed between vessel density and AGE accumulation (R= −0.37) and between IENF and AGE accumulation (R= −0.25) (Fig. 4). As AGE accumulation increased, IENF and vessel density decreased (Fig. 4). No correlation was found between vessel density and IENF.

Fig. 4.

Correlations between AGE (AU), Vessel density (%), and IENF (mm²). Observed correlations between measured metrics occurred between a AGE and IENF (R=−0.25) and between b AGE and vessel density (R=−0.37).

Discussion

This study reports on the use of RH-OCT-A to quantify changes in cutaneous vessel densities among smoking and nonsmoking populations. Previous research is divided on the reduction in vessel density from UV exposure with some studies showing the presence of this effect and others not^15,20,39. We found no significance between metrics from the inner (UV-protected) and the outer (UV-exposed) forearms. Most likely this is because our chosen treatment areas were not sufficiently contrasting enough. Future studies should consider alternative treatment locations for assessment.

In addition to vessel density, we also measured innervation and AGE accumulation between the four study cohorts (old smokers, young smokers, old nonsmokers, and young nonsmokers). We found statistical significance between the old and young datasets for vessel density, nerve fiber density, and AGE scores. This aligns with numerous other studies that have shown vessel rarefaction²⁰, innervation reduction^24,26, and AGE accumulation^9,40 in older populations compared to young populations. Additionally, we concluded that as subjects age, nerve fiber and vessel densities decrease while AGE accumulation increases. This aligns with research which shows the decline of nerve and vessel densities^3,8,24 and the accumulation of AGEs with increasing age⁴¹.

When comparing the smoker and nonsmoker groups, we expected to see a reduction in vessel density in the smoker group given that various publications have shown smoking habits lead to a reduction of blood flow and oxygen delivery, specifically during wound healing^42,43. Though we observed only a small reduction in vessel density in the smoker group compared to the nonsmoker group, this could be influenced by the different ages of the cohorts as we did not age-match subjects. The similar vessel densities among smoking and nonsmoking groups reflects the findings of Isik et al.⁴⁴ who showed that OCT-acquired vessel densities in the retinal layers of 45 smokers at least one hour after smoking were similar to vessel densities of 45 nonsmokers. However, other studies utilizing OCT to assess ophthalmic vasculature changes differed in their conclusions: Cinar et al.⁴⁵ reported a decrease in macular vessel density after one hour of nicotine gum chewing compared to a control group, and Kaymaz et al.⁴⁶ showed significant changes in superficial and deep foveal vessel densities after one hour of smoking. Both of these studies were performed on non-smoking healthy individuals while Isik et al. tested smokers with an average history of 2.2 years. While these studies might align with our results as they find vessel densities stay relatively the same or decrease from nicotine use, our study draws from a pool of daily smokers (Young smokers with an average of 6.2 pack years and old smokers with an average of 50.3 pack years) to assess long-term skin changes rather than those changes observed immediately post-smoking in the eye anatomy. An additional study, Mirbod et al.⁴⁷ found no significant difference between the vascular density of smokers and nonsmokers in vestibular gingival biopsy samples which underwent CD34 immunohistochemical staining techniques. These varied results suggest that vessel density reduction from smoking may be most notable immediately after nicotine exposure and that the effects may be location dependent. Also, all of these studies observe nicotine effects from baseline vasculature (OCT-A images) and do not consider effects using the reactive hyperemia method (RH-OCT-A). The RH-OCT-A method on the eye is not feasible but given that baseline vasculature flow is influenced by various extrinsic factors, it is not surprising that these studies have reported fluctuating results.

Vessel densities in the smoking groups are lower than the nonsmoking groups within age cohorts. This difference is only slight within the young age group and more pronounced within the old age group. Our findings support those of Rossi et al.⁴⁸ who compared post-occlusive reactive hyperemia blood flow between smokers and nonsmokers. Rossi also used Doppler fluximetry to determine that blood vessel function was negatively correlated (R= −0.33) to pack years within their smoker population. We found a correlation between pack years and vessel density (R= −0.53) within the smoker population as well, which was notably stronger than the correlation between subject age and vessel density (R= −0.42). Though we cannot definitively conclude that smoking history has an impact on skin vasculature, our data indicates that it is associated with lower vessel density and thus suggestive of accelerated aging. However, given the cross-sectional design, causality cannot be inferred.

Unexpectedly, we did not find a correlation between vessel density and IENF, contradicting studies which have suggested associations between the two metrics. Specifically, connections have been shown between microangiopathy and neuropathy^30,31. This may be because we found lower IENF measurements in nonsmokers than smokers within each age group. However, these results include only 6 old smokers with one measurement noticeably greater than the others but not statistically considered an outlier given the small data set. These small sample sizes are therefore a limitation of this study, and subsequent longitudinal studies should be undertaken. Furthermore, we did not adjust for hypertension or other confounding factors such as exercise, diet, or pre-diabetic conditions, which could all account for our smokers having higher IENF counts. Finally, seven out of twelve young smokers had histories of vaping rather than tobacco smoking while all old smokers had histories of tobacco smoking. The prevalence of vaping in our young smoker study population may have influenced our results and a future study might test the effects on aging factors of only young vapers compared to young non-vapers especially given the increased usage of vapes among young adults⁴⁹.

This study reports on one of the first applications of RH-OCT-A to assess smoking-related skin aging and presents this method as a noninvasive and exploratory approach. Previous research assessing the impact of smoking on vessel density has largely been limited to ophthalmic studies, which negate the influence of reactive hyperemia on the results, or to investigations using alternative methods such as Doppler fluximetry to measure skin aging as a function of vessel density. Thus, we propose an alternative methodology for noninvasively quantifying smoking-related skin aging factors.

Author contributions

DM, SBE, and MWE conceptualized the study. Data was collected by AD, MWE, PP, AC, LAG, AVC, and MBA. Tissue samples were processed by HD, while AD and MWE analyzed and interpreted the data. DM and MWE supervised the work. AD, PP, and MWE prepared the manuscript. All authors have read and approved the final version of the manuscript.

Funding

No outside funding was received in support of this research.

Data availability

Raw data were generated at Massachusetts General Hospital (MGH). Data is available upon request from the corresponding author (A.D. and M.W-E.)

Declarations

Competing interests

The authors declare no competing interests.