Prevalence of Unsuspected and Significant Mitral and Aortic Regurgitation

Methods 

Results 

The baseline clinical characteristics of the 3088 male patients and 3763 female patients according to age group are shown in Table 1, Table 2, and the distribution by age is shown in Table 3, Table 4. The prevalence of comorbidities is considerable, reflecting the fact that individuals were selected vis-à-vis their referral for an echocardiographic study.

Table 1. Baseline characteristics of male patients in study

	Age, y	<31	31-40	41-50	51-60	61-70	71-80	>80
	N	141	269	458	602	679	623	316
	BMI, kg/m2	27.9±6.9	29.7±7.5	30.3±8.1	29.5±6.5	28.7±5.6	27.1±5.0	24.9±3.6
	HTN, %	31.2	36.8	43.9	49.8	54.2	47.7	51.2
	LVH, %	42.6%	56.9	62.2	69.3	67.9	66.1	77.2
	LVEF	54.0±13.8	52.2±15.3	51.6±15.2	49.8±15.5	47.8±16.1	47.5±15.6	46.7±16.2
	CAD, %	7.0	10.8	26.9	37.0	45.1	52.8	44.0
	DM, %	7.8	14.1	17.7	22.6	25.8	21.7	15.5

BMI, Body mass index; CAD, coronary artery disease; DM, diabetes mellitus; HTN, hypertension; LVEF, left ventricular (LV) ejection fraction; LVH, LV hypertrophy.

Table 2. Baseline characteristics of female patients in study

	Age, y	<31	31-40	41-50	51-60	61-70	71-80	>80
	N	202	296	523	637	655	810	640
	BMI, kg/m2	28.0±9.6	26.9±9.7	31.9±9.4	31.5±8.7	30.9±8.4	28.3±7.1	25.3±5.7
	HTN, %	21.3	32.4	46.5	51.5	61.2	61.4	59.8
	LVH, %	29.2%	43.6	57.6	63.3	67.3	71.6	75.6
	LVEF	56.6±10.8	56.4±10.8	55.5±12.5	54.5±13.6	53.5±14.4	53.5±14.6	53.9±14.7
	CAD, %	2.0	5.1	14.1	20.7	29.5	36.3	28.9
	DM, %	5.4	11.1	21.2	23.5	31.6	25.3	13.9

BMI, Body mass index; CAD, coronary artery disease; DM, diabetes mellitus; HTN, hypertension; LVEF, left ventricular (LV) ejection fraction; LVH, LV hypertrophy.

Table 3. Prevalence of mitral regurgitation by age group

		Age, y
		<31	31-40	41-50	51-60	61-70	71-80	>80
	Male, %	5.0	5.9	14.1	11.3	11.8	11.2	17.1
	Female, %	3.0	8.1	5.5	7.2	12.7	17.8	21.4

Table 4. Prevalence of aortic insufficiency by age group

		Age, y
		<31	31-40	41-50	51-60	61-70	71-80	>80
	Male, %	4.2	10.0	10.7	16.9	17.2	25.0	39.8
	Female, %	0.9	7.4	11.8	14.1	16.0	28.7	33.9

The overall prevalence of significant MR in male patients was 11.7% and 12.5% in female patients. MR was present in 5% of male patients and 3% of female patients 30 years of age or younger. The prevalence in men in their fourth, fifth, sixth, seventh, and eighth decades of life was 5.9%, 14.1%, 11.3%, 11.8%, and 11.2%, respectively. Their female counterparts had MR frequencies of 8.1%, 5.5%, 7.2%, 12.7%, and 17.8%, respectively; 17.1% of men and 21.4% of women older than 80 years had significant MR.

The overall prevalence of significant AI in male patients was 18.9% and 19.7% in female patients. AI was present in 4.2% in male patients and 0.9% in female patients 30 years of age or younger. The prevalence in men in their fourth, fifth, sixth, seventh, and eighth decades of life was 10%, 10.7%, 16.9%, 17.2%, and 25.0%, respectively. Their female counterparts had AI frequencies of 7.4%, 11.8%, 14.1%, 16.0%, and 28.7%, respectively; 39.8% of men and 33.9% of women older than 80 years had significant AI.

The estimated effects of clinical variables on valvular regurgitation obtained from multivariable ordinal regression analysis are shown in Table 5. As described above, the prevalence of both MR and AI increased with each year of advancing age. Although female sex predicted MR, AI was sex neutral. Valvular regurgitation severity increased with decreasing ejection fraction and BMI, a history of HTN, increasing LVH, and valvular abnormalities. The multivariable ordinal regression model can be used to determine predicted probabilities of valvular regurgitant severity; calculators in Excel format (Microsoft, Redmond, Wash) for AI and MR are presented in the online supplement.

Table 5. Ordinal regression modeling clinical variables to valvular regurgitation

	Mitral regurgitation				Aortic insufficiency
	Parameter	Odds ratio	95% CI	P value	Parameter	Odds ratio	95% CI	P value
	Age	1.016	1.013-1.019	<.001	Age	1.031	1.027-1.037	<.001
	Female sex	1.416	1.291-1.551	<.001	Female sex	0.930	0.830-1.042	.213
	BMI	0.982	0.976-0.988	<.001	BMI	0.972	0.965-0.981	<.001
	LVEF	0.945	0.942-0.948	<.001	LVEF	0.985	0.982-0.989	<.001
	HTN	0.872	0.795-0.956	.004	HTN	0.877	0.783-0.983	.024
	CAD	1.043	0.941-1.560	.423	CAD	1.210	1.073-1.370	.002
	LVH	0.759	0.689-.0836	<.001	LVH	0.823	0.728-0.931	.002
	VPath 0	0.174	0.120-0.250	<.001	VPath 0	0.289	0.231-0.351	<.001
	VPath 1	0.264	0.183-0.383	<.001	VPath 1	0.509	0.419-0.618	<.001

BMI, Body mass index; CAD, coronary artery disease; CI, confidence interval; HTN, hypertension; LVEF, left ventricular (LV) ejection fraction; LVH, LV hypertrophy; VPath 0, normal valve morphology compared with “other”; VPath 1, valve thickening compared with “other.”

Discussion 

This study demonstrates that the prevalence rates of unanticipated MR and AI are substantial, increase exponentially with age, and are predicted by commonly used clinical variables. Although these rates are similar to those previously reported,4, 5, 6, 9, 10 prevalence estimates for comparable degrees of regurgitant severity are considerably greater in the current study.

A potential factor explaining the high prevalence of significant MR and AI in this study relates to the relatively high frequency of comorbid conditions; this is in part a result of selection bias, as patients were referred for clinical indications to an echocardiographic laboratory. Thus, although the BMIs were similar to those in prior studies, the prevalence of HTN, LVH, CAD, and (except for the Strong Heart Study) diabetes were greater. The difference in prevalence rates among studies may also be explained by changes in imaging technology. However, although the increase in these unsuspected lesions may be explained in part by improved imaging technology and different patient selection, the methods used to grade regurgitant severity are more likely to be accurate than earlier methods, most of which simply used the spatial dispersion of the regurgitant jet in the respective receiving chamber in a single frame (see below). Nevertheless, whether these disparities represent differences in patient selection and study design, instrumentation, methodology, or true differences in regurgitant severity cannot be determined from this study.

Although several Doppler echocardiographic indexes of regurgitant severity have been developed, measurement of the color flow jet area alone (or a pulsed wave equivalent method) is most commonly used in epidemiologic and toxicologic studies.1, 2, 3, 4, 5, 6 Despite the ability to rapidly screen patients, sole reliance on jet visualization as a semiquantitative tool is misleading because of numerous technical (eg, pulse repetition frequency, gain, and transducer frequency) and hemodynamic (eg, blood pressure, compliance of the receiving chamber) factors that influence significantly the size of the regurgitant jet.7, 11 Taking into account the size of the regurgitant orifice and the origin and spatial orientation of the regurgitant jet as done in this study improves the accuracy of the severity estimate. It is likely that the majority of the jets classified as mild to moderate in severity in this study would be considered moderate by measurement of a simple jet area (or jet area ratio).

The impact of age on the prevalence of valvular regurgitation is profound and confirms results from previous studies.4, 5, 6, 9, 12, 13, 14 Interestingly, although the increase with age for AI in both male and female patients was exponential, the prevalence of MR in men, in contrast to women, tended to plateau in middle age. Indeed, female sex was a significant multivariable predictor of MR, but not aortic regurgitation. This finding is consistent with data from the Strong Heart Study,5, 6 but differs from the Framingham Heart Study4; in that latter study, male sex was a significant predictor of aortic regurgitation. Study design and selection bias are likely to be responsible for these differences, although biological factors cannot be excluded.

The significant inverse relation between the prevalence of regurgitation and BMI on multivariable analysis confirms results from both Framingham and Strong Heart Studies4, 5, 6 and in the HyperGen Study; in the latter, BMI and fat mass were inversely associated with regurgitant severity.15 These findings have implications for interpreting the small excess of regurgitation reported in patients who had received appetite-suppressant agents.14, 16, 17 Although it is likely that technical considerations, such as imaging depth and attenuation, are responsible in part for the inverse association of obesity and regurgitation, poorly understood factors may also be responsible.

The finding of unexpected abnormal valvular morphology had a significant impact on the presence and severity of valvular regurgitation. Not surprisingly, normal or nonspecifically thickened valves had a significantly reduced likelihood of valvular regurgitation compared with valves with more advanced pathology. However, the vast majority of patients in this study had normal or mildly abnormal aortic (56% or 38%) and mitral (65% or 33%) valves, respectively. These data suggest that the study’s exclusion process successfully resulted in a relatively pure sample of individuals with unsuspected valvular disease.

LVH was a significant predictor of both MR and aortic regurgitation; the presence of LVH was associated with a 32% and 21% increase in MR and aortic regurgitation, respectively. This anticipated relation was also noted in the Strong Heart Study, HyperGen, and LIFE.5, 6, 12, 15 LVH was highly prevalent in our population; this may reflect the increased body weight of patients and their comorbidities.18, 19

HTN was a significant, albeit weak multivariable determinant for both MR and aortic regurgitation; the risk for regurgitation was ∼13% increased in patients with a history of HTN. These findings are consistent with those from the Strong Heart Study.5, 6 In addition, Gardin et al13 found a significant association between a history of HTN and AI in obese adults. In contrast, HTN was a significant determinant of MR, but not aortic regurgitation in the Framingham Heart Study4 (HyperGen and LIFE were HTN trials). However, data in the current study should be interpreted with care, as a retrospective report of a history of HTN was used. Similarly, the data for CAD, which suggest a decreased risk for aortic regurgitation but not MR, should be viewed with caution.

There are several limitations to the study that merit comment. As mentioned previously, the study is retrospective and is subject to unavoidable selection bias. Thus, although patients were not suggested to have MR or AI before the study, the fact that an echocardiogram was ordered at a tertiary medical center increases the potential for underlying cardiovascular pathology. Secondly, although the use of multiple grading methods refines the estimate of regurgitant severity, the use of multiple readers in this study, and the inherent subjectivity in grading schemes, introduces noise into these estimates and the determination of lesion significance.20 Average intrareader variability has been reported as ∼6% for AI and 17% for MR.20 Finally, thorough clinical histories were not taken by the study physicians, but were obtained by chart review; thus, a pretest suggestion of regurgitation AI/MR cannot be entirely excluded.

Despite these limitations, this study indicates that the prevalence of unanticipated and significant left-sided valvular regurgitation is considerable. These prevalence estimates should be considered when assessing the findings of MR or aortic regurgitation on an echocardiogram performed for nonvalvular indications..