Association of Smoking With Phenotype at Diagnosis and Vascular Interventions in Patients With Renal Artery Fibromuscular DysplasiaNovelty and Significance
The pathogenesis of fibromuscular dysplasia (FMD) remains unclear, but tobacco use is thought to be involved. This retrospective cross-sectional study aimed to evaluate smoking first as a risk factor for renal artery FMD diagnosis and second as a modifier of the clinical and radiological phenotype of this disease. We retrieved 337 adult patients diagnosed with FMD in a referral center for hypertension management, who were first individually matched to controls with essential hypertension for sex, age, systolic blood pressure, number of antihypertensive drugs, and year of visit. Smoking status and other relevant data were collected at first visit. The proportion of current smokers was higher for patients with FMD than for the controls (30% and 18%, respectively, P<0.001; odds ratio, 2.5 [95% confidence interval, 1.6–3.9]). Second, characteristics of FMD were compared between current smokers and other patients. Among patients with multifocal FMD, current smokers experienced an earlier diagnosis of hypertension (36 versus 42 years, respectively; P<0.001) and FMD (43 versus 51 years; P<0.001) than other patients, and a greater likelihood of renal artery interventions (57% versus 31%; P<0.001) and of kidney asymmetry (21% versus 4%; P=0.001). In conclusion, current smoking is associated with a higher likelihood of renal artery FMD diagnosis. Rather than a higher incidence of FMD, this may reflect a more aggressive course in smokers, who have earlier hypertension leading to increased and earlier recognition of the disease. Smoking cessation should be strongly encouraged in patients with FMD.
Fibromuscular dysplasia (FMD) is a group of idiopathic, segmental, and nonatherosclerotic diseases of the arterial walls, leading to the stenosis of medium-sized arteries, including the renal and internal carotid arteries in particular.1 The pathogenesis of this rare vascular disease is unknown,2,3 but it has been suggested that a genetic predisposition4–7 or environmental factors, such as exposure to estrogens3,8 or repeated mechanical trauma, may be involved.8,9 Smoking, a recognized potent cause of vascular disease,10 was also identified as a potential factor contributing to FMD in previous small, observational studies.9,11–13
The aim of this retrospective cross-sectional study was to evaluate smoking as a risk factor for renal artery FMD diagnosis and as a disease modifier. We first compared the prevalence of current smoking between patients with FMD and matched controls with essential hypertension (EH). We then compared the clinical and radiological phenotype between current smokers and other patients with FMD.
Patients and Methods
Patients With FMD
The screening and selection of patients with FMD have been described elsewhere.14 Briefly, we reviewed the medical charts of all adult patients for whom a diagnosis of FMD was considered between January 1, 1986 and November 30, 2011 at a single institution. In accordance with the current definition of renal artery FMD, we accepted the diagnosis in patients with nonatherosclerotic stenosing lesions affecting the trunk or branches of renal arteries in the absence of aortic wall thickening or biochemical evidence of inflammation and in the absence of known syndromic arterial disease.15 Isolated renal artery aneurysm or dissection was not considered sufficient to diagnose FMD.16 The classification of FMD as unifocal (presence of a single stenosis on a given vessel, regardless of its length) or multifocal (presence of ≥2 stenoses on a given vessel segment) was based on radiological data.14
Controls With EH
Each case with renal artery FMD was paired with a control patient with EH matched for sex, age (±2 years), systolic blood pressure (BP; ±10 mm Hg), number of antihypertensive drugs used, and year of first visit (±2 years). Patients with EH were identified by screening the electronic medical record database of our institution. The medical charts of selected controls were reviewed manually to check that these individuals were not pregnant and had not subsequently been diagnosed with secondary hypertension.
Retrieval of Clinical Data
We extracted clinical and biological data collected from patients during their first visit to our unit. For patients referred after FMD had been diagnosed elsewhere, data were collected only if the first visit to our unit occurred within 1 year of the diagnosis of FMD and if no renal artery intervention had been performed during this period. For the matching procedure, multivariate imputation was used to attribute initial systolic BP values or numbers of antihypertensive drugs to patients with FMD when these data were missing. The imputed values were used for the matching procedure only and are not reported in the descriptive statistics.
Patients were asked whether they did currently smoke tobacco on a regular basis (the vast majority smoke on a daily basis). Those who answered positively were classified as current smokers. Patients who declared having quit tobacco smoking were classified as former smokers, regardless of the time delay since smoking cessation. Ever smokers included current and former smokers. The remaining patients were classified as never smokers.
Creatinine clearance was estimated with the Cockcroft–Gault formula,17 normalized for body surface area, because most creatinine measurements were not calibrated to isotope dilution mass spectrometry (ruling out the CKD-EPI [Chronic Kidney Disease-Epidemiology Collaboration] equation) and because most patients with renal artery FMD have a glomerular filtration rate >60 mL/min (ruling out the MDRD [Modification of Diet in Renal Disease] equation).18 Renal asymmetry was defined as a difference larger than 20 mm in bipolar length between the 2 kidneys on ultrasound scans.11 Procedures followed were in accordance with institutional guidelines.
The detailed baseline and demographic characteristics of the patients with FMD have been reported elsewhere.14 A case versus control analysis was first performed between patients with FMD and matched controls with EH; comparisons were performed with conditional logistic regression. An exposed versus not exposed analysis was then performed between current smokers and other patients with FMD. Quantitative variables were reported as medians and quartiles and were compared using the Mann–Whitney test. Nominal and ordinal variables are reported as numbers and percentages and were compared with the Fisher exact test and χ2 test for trend, respectively. Factors associated with renal asymmetry were studied by binomial logistic regression analysis.
Quantitative variables were not categorized for logistic regression studies. Linearity between the logit and quantitative variables was checked graphically. Interaction tests were performed to check for heterogeneity between FMD subtypes in the comparison to patients with EH. The Wald test was used to assess statistical significance.
The single pre-established hypothesis was that current smoking is more prevalent among patients with FMD than among the matched controls; a P value <0.05 was considered significant for this comparison. All other comparisons were hypothesis generating and, because of the exploratory nature of the study, no adjustment was made for multiple comparisons: the lesser the P value, the less likely it is to be a chance finding. Stata 9.2 (Stata-Corp, College Station, TX) was used for statistical analyses.
Selection and Characteristics of the Patients
By querying databases, we identified 700 patient records in which FMD diagnosis was mentioned at least once. We ascertained the diagnosis of renal artery FMD in 337 patients (61 patients with unifocal and 276 with multifocal FMD). The other 363 cases were excluded, mostly because initial imaging data were unavailable to ascertain the diagnosis or because current diagnostic criteria were not met (Figure).
Smoking status was available for 326 of 337 (97%) patients with FMD, including 268 patients with multifocal disease and 58 patients with unifocal disease. Smoking status was completely available for 324 of 337 (96%) patients with EH. The remaining 13 patients were not currently smoking at the time of their visit, but we do not know whether they were former smokers or never smokers.
Most patients were women (80%) and of the multifocal subtype (82%). Both hypertension and FMD were diagnosed earlier in patients with unifocal FMD (26 and 30 years, respectively) than in patients with multifocal FMD (40 and 49 years, respectively; P<0.001 for both comparisons). A larger proportion of patients with unifocal FMD were current smokers (50% versus 26%; P<0.001), were diagnosed with kidney asymmetry (33% versus 10%; P<0.001), and had undergone a renal artery intervention at some time (87% versus 38%; P<0.001).
Smoking in Patients With FMD and Controls
Table 1 reports the comparisons of cases with renal artery FMD and matched controls with EH. No significant interaction was found between FMD subtype (unifocal or multifocal) and any of these comparisons. Patients with FMD were more likely than controls to be current smokers (30% versus 18%, P<0.001; odds ratio, 2.5 [95% confidence interval, 1.6–3.9]) or to have smoked at some point in their lives (ever smokers, 50% versus 37%, P=0.001; odds ratio, 1.8 [95% confidence interval, 1.3–2.5]). Patients with EH had a higher body mass index than patients with FMD. Although they had the same age and sex and similar plasma creatinine concentrations than patients with FMD, estimated creatinine clearance according to the Cockcroft–Gault formula was higher in patients with EH because of higher body weights.
Associations With Smoking Status
Among patients with multifocal FMD, Kruskall–Wallis tests comparing current smokers, former smokers, and patients who had never smoked revealed significant differences regarding sex (P=0.01), age at hypertension diagnosis (P=0.002), age at FMD diagnosis (P<0.001), kidney asymmetry (P=0.002), performance, and number of renal interventions (P=0.001 and P=0.007). Post hoc comparisons showed that all these differences, with the exception of that for sex, were attributable to a difference between current smokers and the other 2 groups, which were therefore combined. For unifocal FMD, there were too few former smokers (n=4) for meaningful comparisons across the 3 groups. Therefore, we combined the groups of former smokers and patients who had never smoked, as for patients with multifocal disease.
In patients with multifocal FMD, current smoking was associated with a younger age at diagnosis of hypertension and of FMD and with a higher frequency of renal artery interventions (Table 2). In patients with unifocal FMD, current smoking was associated with higher baseline BP (Table 3).
Renal ultrasound data were available for 243 patients, and renal asymmetry was observed in 35 of them (on the right side in 21 cases). Smoking status was unknown for 3 patients with documented renal asymmetry. Current smoking was associated with a higher frequency of renal asymmetry in both multifocal and unifocal FMD. In univariate analysis, renal asymmetry was associated with current smoking, unifocal FMD, and ipsilateral unilateral renal artery stenosis (Table 4). In multivariate analysis, current smoking and unifocal FMD remained independently associated with renal asymmetry, whereas the association with ipsilateral unilateral renal artery stenosis was no longer significant. Only 9 patients had a kidney length <8 cm, always on 1 side only; creatinine clearance of these patients ranged from 38 to 118 mL/min per 1.73 m2 (2 patients <60 mL/min per 1.73 m2).
The first key finding from this study is that the proportions of current smokers and ever smokers were significantly higher in patients with FMD than in matched controls with EH. A second major finding is the association of current smoking with an earlier hypertension and FMD diagnosis, and with a higher frequency of renal asymmetry and a higher prevalence of renal artery interventions, in patients with multifocal disease.
Comparison With Previous Studies and Interpretation of the Results
Recently published data from the US FMD Registry, which holds data from 2008 to 2011, revealed that 37.2% of patients with FMD were ever smokers (current or former smokers).19 Previous small studies have compared the prevalence of smoking in patients with FMD and matched controls (Table 5).9,12,13,20 As expected, given recent antismoking campaigns, the proportions of smokers among patients with FMD and their controls were higher in older series. However, all studies concur in reporting a higher proportion of current or ever smokers in patients with FMD than in controls. A single small study in patients with multifocal FMD compared the presentation of subjects with (n=24) and without (n=26) a history of smoking.11 In this previous study, smoking was associated with a younger age at diagnosis (39 versus 49 years; P<0.01) and a higher prevalence of unilateral renal atrophy (67% versus 27%; P<0.01).
Our study confirms and extends these findings, providing information about the influence of current and past smoking in patients with unifocal or multifocal FMD. The association between smoking and FMD diagnosis is equivocal. Because ≈50% of our patients with FMD have never smoked, cigarette smoking cannot be considered as a prerequisite for the development of the disease. Nonetheless, smoking may provoke FMD lesions in susceptible individuals and, therefore, increase the true incidence of the disease.
On the contrary, smoking may only worsen preexisting arterial lesions and, thus, increase the likelihood and magnitude of clinical consequences of the disease. Smoking has indeed many deleterious effects on vessels, decreasing the production of nitric oxide and prostacyclin, and accelerating the development of both atherosclerosis and thrombosis.10 Current exposure to cigarette smoke may amplify the endothelial dysfunction underlying renovascular hypertension21 or favor the development of atherosclerosis or thrombosis. These mechanisms could explain the worsening of dysplastic stenoses or superimposed atherosclerotic lesions.
In our patients with multifocal FMD, current smoking was strongly associated with a younger age at diagnosis of hypertension and of FMD and with more renal asymmetry. The finding of a more severe disease course in current smokers suggests that the association between smoking and FMD diagnosis reflects a more aggressive course in smokers, with earlier hypertension leading to more frequent and earlier recognition of the disease, rather than a truly higher incidence of FMD in smokers.
In other words, the association between smoking and FMD diagnosis may reflect an ascertainment bias: patients with FMD would be more likely to be referred to hypertension units like ours if they are current smokers because of a more severe anatomic or clinical disease. Conversely, patients with FMD who do not smoke have a less severe clinical phenotype, closer to EH, or may even be asymptomatic. In both cases, FMD is much less likely to be diagnosed.
Strengths and Weaknesses of the Study
This study is based on a large number of well-characterized patients with FMD and controls carefully matched on the basis of year of first visit, age, sex, BP, and number of types of antihypertensive drug used. Matching for the year of first visit was particularly important because of secular trends in smoking habits. This study was retrospective and is, therefore, subject to several limitations, including missing data. However, computerized and structured data recording during routine clinical care resulted in there being <15% missing data for clinical variables. The results previously obtained in retrospective analyses of this clinical database have consistently been confirmed in prospective studies.22,23 Our results are also subject to possible referral bias and caution is, therefore, required when extrapolating to patients with FMD who are not typically referred to a hypertension unit, like those with asymptomatic or predominantly cervical FMD. Ascertainment bias also cannot be excluded: smokers could be more actively screened for hypertension than nonsmokers, explaining a younger age at hypertension diagnosis and, hence, FMD diagnosis. Details about the amount and duration of smoking were not consistently recorded at the first visit, and their impact could not be analyzed. The number of patients with unilateral FMD was limited, and no firm conclusion can be drawn for this subgroup. Moreover, observed BP difference in this subgroup may be confounded by the difference in number of antihypertensive medications, although it was not statistically significant. The performance of multiple comparisons entails a risk of P values <0.05 being attributable to chance alone. However, highly significant differences (P<0.001) can be considered reliable.
This study indicates that current smoking is associated with a higher incidence of FMD or, most likely, with more frequent and earlier diagnosis. Current smoking is also associated with more severe consequences of renal ischemia (kidney asymmetry), leading to a larger number of renal interventions. Further studies are required to assess the impact of the amount and duration of smoking, and elucidate the mechanisms by which smoking influences the expression and course of FMD. Endothelial dysfunction may play a role that is currently being evaluated at our center.
The prevalence of current smoking is greater in patients with FMD than in matched controls. Current smoking is associated with more severe and more rapidly progressing disease in patients with multifocal FMD. This study highlights the critical importance of encouraging patients with FMD to quit smoking.
We thank Drs Michael Frank, Jacques Julien, Antoine Chedid, and Béatrice Fiquet for their contribution to recruitment of patients with fibromuscular dysplasia and data collection.
S. Savard received a financial support from la Société Québécoise d’Hypertension Artérielle, la Société Québécoise de Néphrologie, La Faculté de Médecine de l’Université Laval (McLaughlin Scholarship Program), and from La Fondation du Center Hospitalier Universitaire de Québec for his postdoctoral work. The other authors report no conflicts.
- Received December 10, 2012.
- Revision received March 30, 2013.
- Accepted April 1, 2013.
- © 2013 American Heart Association, Inc.
- Mettinger KL,
- Ericson K
- Lüscher TF,
- Keller HM,
- Imhof HG,
- Greminger P,
- Kuhlmann U,
- Largiadèr F,
- Schneider E,
- Schneider J,
- Vetter W
- Sang CN,
- Whelton PK,
- Hamper UM,
- Connolly M,
- Kadir S,
- White RI,
- Sanders R,
- Liang KY,
- Bias W
- Powell JT
- Mackay A,
- Brown JJ,
- Cumming AM,
- Isles C,
- Lever AF,
- Robertson JI
- Savard S,
- Steichen O,
- Azarine A,
- Azizi M,
- Jeunemaitre X,
- Plouin PF
- Novick AC,
- Scable J,
- Hamilton G
- Stanley JC
- Olin JW,
- Froehlich J,
- Gu X,
- Bacharach JM,
- Eagle K,
- Gray BH,
- Jaff MR,
- Kim ES,
- Mace P,
- Matsumoto AH,
- McBane RD,
- Kline-Rogers E,
- White CJ,
- Gornik HL
Novelty and Significance
What Is New?
Current smoking is associated with a higher likelihood of renal artery fibromuscular dysplasia (FMD) diagnosis.
Rather than a higher incidence of FMD, this may reflect a more aggressive course in smokers, who have earlier hypertension possibly leading to increased and earlier recognition of the disease.
What Is Relevant?
Smoking cessation should be strongly encouraged in patients with FMD.
The proportion of current smokers was higher for patients with FMD than for matched controls with essential hypertension (odds ratio, 2.5). Among patients with multifocal FMD, current smokers experienced an earlier diagnosis of hypertension and FMD than other patients, and a greater likelihood of renal artery interventions and of kidney asymmetry.