Aortic Aneurysm and Dissection in Patients With Biopsy-Proven Giant Cell Arteritis From Northwestern Spain

A Population-Based Study

Abstract

Most classical manifestations of giant cell arteritis (GCA) are the result of occlusive vascular involvement. However, unlike ischemic manifestations, aortic aneurysmal disease in patients with GCA has been less well described. We assessed the incidence and predictors of aortic aneurysm and dissection in patients with biopsy-proven GCA from the Lugo region of northwestern Spain and compared the results with those in a 2003 report from Olmsted County, MN. We performed a retrospective study of biopsy-proven GCA patients diagnosed from 1981 to 2001 at the single hospital for a well-defined population of almost 250,000 people. Twenty (9.5%) of the 210 biopsy-proven GCA patients diagnosed during the study period developed aortic aneurysmal disease. Sixteen of the 20 patients had thoracic aneurysms and 6 had abdominal aneurysms. The incidence of aortic aneurysm and/or dissection in Lugo (18.9 per 1000 person years at risk) was similar to that reported in Olmsted County (18.7 per 1000 person years at risk). Hypertension (hazard ratio: 4.73) and polymyalgia rheumatica with a marked acute inflammatory response at the time of diagnosis of GCA (hazard ratio: 3.71) were the best predictors of aortic aneurysmal disease. Our present observations suggest that a severe inflammatory response associated with hypertension at the time of diagnosis of GCA may promote the development of aortic aneurysmal disease. GCA patients having these features should be monitored for the existence of aortic aneurysm and dissection.

Abbreviations: CI = confidence intervals, ESR = erythrocyte sedimentation rate, GCA = giant cell arteritis, OR = odds ratio, PMR = polymyalgia rheumatica.

INTRODUCTION

Giant cell arteritis (GCA) is the most common systemic vasculitis in adults in western countries, in particular in people over the age of 50 years with northern European ancestry^1,8,13,15,25. Its frequency increases with age and peaks in those groups of patients older than 70 years^10,15.

Classically, patients with GCA present with cranial ischemic manifestations that are directly related to vascular involvement. However, as pointed out by Levine and Hellmann¹⁹, in the evaluation of patients with GCA, there is a growing number of "occult" manifestations that also need to be considered, such as aortic aneurysmal disease¹⁹. Although aortic aneurysm and death from aortic dissection were recognized as complications of GCA more than 50 years ago^21,22,24,26, it was not until the past decade that these manifestations were considered as frequently underrecognized complications of GCA^5,6.

In a 2003 report, Nuenninghoff and coworkers²³ examined the incidence and predictors of large-artery involvement, including aortic aneurysm and dissection, over a 50-year period (1950-1999) in Olmsted County, MN. In the follow-up of a cohort of 168 patients with GCA they found 30 incident cases of aortic aneurysm and/or dissection²³. To determine whether geographic, ethnic, or other factors influenced the frequency of aneurysmal disease in GCA, we studied the well-defined population from the Lugo region in northwestern Spain^7,9.

In the current study we have determined the incidence of aortic aneurysm and dissection in patients with biopsy-proven GCA. Also, we have attempted to identify clinical features of GCA that could be associated with an increased risk of aortic aneurysmal disease in these patients.

PATIENTS AND METHODS

We performed a retrospective review of the case records of all patients diagnosed with biopsy-proven GCA at the Department of Medicine of the Hospital Xeral-Calde (Lugo, Spain) between January 1981 and December 2001. This hospital is the reference center for a mixed rural and urban population of almost a quarter million people. The characteristics of this white population have been described previously^7,9.

The temporal artery biopsy procedure used in Lugo patients has been reported elsewhere^9,12. Patients were diagnosed as having biopsy-proven GCA when the temporal artery biopsy showed a compatible pathology report, describing the characteristic mononuclear cell infiltration of the arterial wall, with or without granulomas and/or multinucleated giant cells²⁰.

Clinical Definitions

Clinical definitions of GCA manifestations and treatment in the Lugo population have been reported elsewhere^7,9,12. Information about smoking history, blood pressure, and cholesterol and glucose levels was available in all cases at the time of disease diagnosis. Smoking history was treated as a dichotomous variable (heavy versus non-heavy smoking) in this analysis. Heavy smokers comprised those patients who still smoked at the time of disease diagnosis or who had smoked within the 10 years before the onset of GCA symptoms; the remaining patients (non-heavy smokers) were those patients who never smoked or who stopped smoking at least 10 years before the disease onset.

Patients were considered to have aortic aneurysmal disease if there was unequivocal evidence of aortic aneurysm or dissection confirmed by imaging studies including chest radiograph, angiography, computed tomography, magnetic resonance imaging or ultrasound, and/or pathologic examination of aortic tissue.

Data Collection

Demographic and clinical data at the time of diagnosis of all the patients with biopsy-proven GCA were analyzed. Also, erythrocyte sedimentation rate (ESR), hemoglobin, platelet count, cholesterol, and glucose on admission were assessed. In assessing follow-up, medical records of all biopsy-proven patients were reviewed by Lugo physicians (MAG-G, CG-P, AP, and RP-R). For the purpose of this study only information found in the medical records was considered. Thus, phone interviews and information about the patients given by their relatives were not considered.

Statistical Analysis

Comparisons between 2 categories were made using the Student t test (2-tailed) for continuous variables. To analyze categorical data a chi-square test or Fisher exact test was performed. The relationship between epidemiologic features or clinical or laboratory data of GCA and aortic aneurysmal disease was assessed by logistic regression analysis; results are provided as odds ratios (OR) with their 95% confidence intervals (CI). To assess the probability of developing aortic aneurysmal disease during the follow-up, a survival analysis was performed using Kaplan-Meier estimates and Cox regression (results are reported as hazard ratios with 95% CI). Death without aortic aneurysm was considered as censored.

Statistical significance was defined as p ≤ 0.05. Calculations were performed with the statistical package Stata 8/SE (Stata Corporation, College Station, TX).

RESULTS

Aortic Aneurysm or Dissection in Patients With Biopsy-Proven GCA

Twenty (9.5%) of the 210 patients with biopsy-proven GCA diagnosed in Lugo between 1981 and 2001 developed aortic aneurysmal disease. Sixteen of the 20 patients had thoracic aneurysms and 6 had abdominal aneurysms. Two patients suffered aortic thoracic dissection and 1, abdominal dissection (Table 1).

TABLE 1: Incidence of Aortic Aneurysm or Dissection in 210 Patients With Biopsy-Proven GCA From Lugo, Spain*

The time between the diagnosis of GCA and the diagnosis of aneurysmal disease ranged from 0 to 162 months (mean ± standard deviation: 57.3 ± 54.9 mo; median: 38.5 mo; interquartile range: 4.5-104 mo). In 3 of the 20 patients, thoracic aortic aneurysms were found at the time of diagnosis of GCA. In the remaining patients, aortic aneurysmal disease was found during follow-up.

The incidence of aortic aneurysm and/or dissection was 18.9 per 1000 person years at risk. The incidences of thoracic and abdominal aneurysms and/or dissection were 15.9 and 5.7 per 1000 person years at risk, respectively.

Differences Between Patients With and Without Aortic Aneurysmal Disease

Epidemiologic differences at the time of GCA diagnosis between biopsy-proven GCA patients with or without aortic aneurysmal disease are shown in Table 2. When a diagnosis of GCA was made, patients who developed aneurysmal disease were younger than remaining GCA patients (71.1 ± 6.4 yr vs. 75.0 ± 6.9 yr in those without aneurysmal disease; p = 0.018). No differences according to sex, delay to diagnosis, site of residence, or duration of follow-up between patients with or without aneurysmal disease were found (see Table 2). When traditional risk factors of atherosclerosis were examined, some differences were observed (see Table 2). With respect to this, 13 (65%) of the 20 patients who suffered aneurysmal disease had hypertension at the time of disease diagnosis compared with 57 (30%) of the 190 patients without this vascular complication (p = 0.002). However, no differences according to history of hypercholesterolemia, diabetes, or heavy smoking were found.

TABLE 2: Epidemiologic Differences at the Time of Diagnosis Between Biopsy-Proven GCA Patients With or Without Aortic Aneurysmal Disease

At the time of disease diagnosis, GCA patients with aneurysmal disease presented polymyalgia rheumatica (PMR) manifestations (13/20; 65%) more commonly than the remaining GCA patients without this complication (74/190; 39.0%) (p = 0.024) (Table 3). Apart from this, there were no other significant clinical differences between patients with or without aneurysmal disease. Of note, although no statistically significant differences in the inflammatory parameters were found, the ESR and platelet count in patients with aneurysmal disease were greater than in the remaining patients (101 ± 21 mm/1 h and 444,000 ± 134,000/mm³ vs. 92 ± 22 mm/1 h and 406,000 ± 135,000/mm³, respectively). Likewise, hemoglobin values were lower in patients with aneurysmal disease (11.4 ± 1.5 g/dL) compared with the rest of the GCA patients (11.8 ± 1.6 g/dL) (see Table 3).

TABLE 3: Clinical Differences at the Time of Diagnosis Between Biopsy-Proven GCA Patients With and Without Aortic Aneurysmal Disease

Influence of the Inflammatory Response at the Time of Diagnosis of GCA on the Development of Aneurysmal Disease

Differences in the inflammatory process, manifested by a significantly higher frequency of PMR and a more severe abnormality of laboratory inflammatory markers of disease at the time of diagnosis of GCA, were observed in patients with aneurysmal disease. We assessed the potential relationship between severe abnormality of laboratory parameters of inflammation and PMR at the time of disease diagnosis in biopsy-proven GCA patients and the development of aortic aneurysm or dissection over the extended follow-up (Table 4).

TABLE 4: Relationship Between Severe Abnormality of Laboratory Markers of Inflammatory Response and PMR at the Time of Disease Diagnosis in Patients With Biopsy-Proven GCA and the Development of Aortic Aneurysm or Dissection Over the Extended Follow-Up

While only 1 of the 34 patients with PMR not associated with severe abnormality of laboratory markers of inflammatory response at the time of diagnosis of GCA suffered aneurysmal disease, 12 of the 53 patients with PMR and ESR >100 mm/1 h and/or hemoglobin <11 g/dL and/or platelet count >450,000/mm³ exhibited this complication (see Table 4). In addition, the combination of PMR with severe abnormality of laboratory parameters of inflammatory response at the time of diagnosis of GCA was associated with a statistically significant increased risk of suffering aortic aneurysmal disease compared with patients without PMR and without severe abnormality of these laboratory markers (OR: 6.15; 95% CI: 1.22-31.10; p = 0.01) (see Table 4). The probability of suffering aortic aneurysmal disease was much lower in patients with PMR not associated with severe abnormality of these markers of inflammation at the time of diagnosis of GCA compared with patients with PMR associated with severe abnormality of the laboratory inflammatory parameters (Figure 1).

FIGURE 1: Kaplan-Meier survival curves showing that patients with PMR who had at least 1 laboratory marker of severe inflammatory response at the time of diagnosis of GCA (ESR >100 mm/1 h, hemoglobin <11 g/dL, or platelet count >450,000/mm³) (dashed line) were more likely to develop aneurysms than the remaining patients (solid line).

Predictors of Aneurysmal Disease

Hypertension (hazard ratio: 4.73) and PMR associated with the presence of severe abnormality of laboratory parameters of inflammatory response at the time of diagnosis of GCA (hazard ratio: 3.71) were the best predictors of aortic aneurysmal disease in patients with biopsy-proven GCA (Table 5).

TABLE 5: Predictors of Aortic Aneurysmal Disease

Differences in the Development of Aneurysmal Disease Between Patients Diagnosed With GCA in 1981-1991 and in 1992-2001

To assess whether a higher awareness of GCA and its complications in recent years among Lugo physicians might account for a progressive increase in the incidence of aneurysmal disease, we divided the study period into 2 segments.

Aortic aneurysms and/or dissection were more frequently found in GCA patients diagnosed in the period 1981-1991 (12/66 GCA cases; 18.2%) compared with those diagnosed in 1992-2001 (8/144 GCA cases; 5.6%) (p = 0.004). However, as the period of diagnosis of GCA was not significant in the Cox regression (p = 0.50), this difference would be fully explained by differences in the follow-up. The probability of survival without aneurysmal disease (probability of not developing aortic aneurysmal disease) was similar in both time periods. Figure 2 reinforces this observation: the higher number of GCA patients with aortic aneurysmal disease diagnosed in the period 1981-1991 was related to a longer follow-up of these patients compared with those diagnosed in the period 1992-2001.

FIGURE 2: Kaplan-Meier survival curves without aortic aneurysmal disease for patients diagnosed in 1992-2001 (solid line) and 1981-1991 (dashed line).

Differences in the Incidence of Aneurysmal Disease Between GCA Patients From Lugo and Olmsted County

We compared our results with those of the previous study by Nuenninghoff and coworkers²³, which appears to be the only other population-based assessment of the incidence of aortic aneurysmal disease. The period of study in Olmsted County was longer and the proportion of patients with aortic aneurysmal disease was higher than those found in our series. However, the incidence of aneurysmal disease per 1000 person years followed was remarkably similar in both populations (18.9 in Lugo and 18.7 in Olmsted County). Thoracic aortic aneurysmal disease was slightly more frequent in Lugo, but the incidence of thoracic dissection was more common in Olmsted County (Table 6). In contrast, abdominal aortic aneurysmal disease was more common in Olmsted County, but the frequency of abdominal aortic rupture was slightly higher in the Lugo population (see Table 6).

TABLE 6: Aortic Aneurysms in Patients With GCA: Differences between Lugo, Spain and Olmsted County, MN

DISCUSSION

The present report is, to our knowledge, the first population-based study on the incidence of aortic aneurysmal disease in southern European patients diagnosed with GCA. Aortic involvement in the setting of GCA was long considered uncommon¹⁸. However, 2 retrospective population-based studies^6,23 from the Mayo Clinic have shown a much higher incidence than previously reported: in 1995 this group reported⁶ that 14 (15%) of 96 patients diagnosed with GCA between 1950 and 1985 by temporal artery biopsy or by clinical classification criteria¹⁴ had aneurysmal disease, involving the thoracic aorta in 9 cases and the abdominal aorta in another 5. In 2003, in a new retrospective study²³ of GCA patients diagnosed between 1950 and 1999, investigators from the same center found 30 (18%) cases of aortic aneurysmal disease in a series of 168 patients diagnosed with GCA by temporal artery biopsy or clinical classification criteria in Olmsted County, MN. Although the period of study was longer and the number of cases with aortic aneurysmal disease was higher in the Olmsted County study than in the present report, the incidence of aortic aneurysm and/or dissection was remarkably similar in both regions (18.7 in Olmsted County and 18.9 per 1000 person years in Lugo). This suggests that, regardless of the genetic or geographic background of the population, aneurysmal disease may be a similar potential complication of GCA. However, due to the limitations of our retrospective study, it is possible that the actual incidence of thoracic aortic aneurysms in Lugo could be even higher than what we describe in this report.

Aortic aneurysms may be present at the time of diagnosis but, as shown in our series, they become evident 6-7 years after GCA has been diagnosed^5,6. In the series by Nuenninghoff et al²³, the time from the onset of symptoms to diagnosis of GCA was much longer in patients with aortic aneurysmal disease (11 wk) than in patients without large artery involvement (6 wk). In Lugo we observed a nonsignificant longer delay to the diagnosis of GCA in patients with aortic aneurysm or dissection (13 wk) than in the remaining patients with biopsy-proven GCA (10 wk). At the time of GCA diagnosis, patients from Olmsted County²³ and Lugo who suffered aortic aneurysmal disease were younger than those who did not have this complication.

The search for predictors of aneurysmal disease is of interest regarding diagnosis and possible therapy. In the report by Nuenninghoff et al²³, aortic insufficiency murmur at diagnosis of GCA and hyperlipidemia and coronary disease at any time were associated with aortic aneurysm and/or dissection. We did not include these factors in our analysis.

In the present study we found that hypertension and a severe inflammatory response at the time of diagnosis of GCA, manifested by the presence of PMR and severe abnormality of laboratory parameters of inflammation, were the best predictors of aortic aneurysmal disease in patients in northwestern Spain. Data on inflammatory laboratory parameters observed in the present series of patients with aneurysmal disease were similar to those reported by Evans et al⁵ in a series of 41 patients with aneurysm and/or rupture of the thoracic aorta seen at the Mayo Clinic between 1950 and 1991. These authors reported that the median hemoglobin and ESR values in patients with aortic aneurysmal disease were 11.3 g/dL and 104 mm/1 h, respectively, while in our series the mean hemoglobin and ESR values were 11.4 g/dL and 101 mm/1 h, respectively.

Studies with positron-emission tomography scanning have suggested that subclinical aortic inflammation is often present in patients with GCA², and may even be universal. Such observations indicate that a chronic or persistent inflammatory process may in some instances weaken the vessel wall, resulting in aortic aneurysms and dissection.

Earlier reports^3,11 on GCA described a negative association between a strong inflammatory response and the risk of developing cranial ischemic complications. In patients from northwestern Spain, anemia was found to be a negative predictive factor for the development of ischemic visual manifestations in the setting of GCA⁹. Cid et al⁴ reported that temporal artery biopsies from GCA patients without ischemic complications had significantly higher tissue angiogenesis scores than those from GCA patients with ischemic events. Angiogenesis was also significantly more prominent in patients with a strong acute phase response compared with those with a weak systemic inflammatory response⁴. Different cytokine profiles between subgroups of GCA patients^16,17 may account for differences in the clinical expression of the disease.

Our present observations and those previously described⁹ suggest that a severe early inflammatory response may be an indicator of significant aortic involvement that, when persistent and modified by hypertension or other factors, may promote the development of aortic aneurysmal disease. Thus, our results suggest that patients with GCA should be monitored for the existence of aneurysmal disease, in particular patients with hypertension and severe inflammatory response at the time of diagnosis of GCA.