High prevalence of occult left heart disease in scleroderma-pulmonary hypertension

Introduction

The association between scleroderma and pulmonary hypertension (PH) is well established [1]. It may be caused by pre-capillary microvascular narrowing resulting in pulmonary arterial hypertension (PAH) or pulmonary parenchymal disease, veno-occlusive disease, post-capillary pulmonary venous hypertension (PVH) from left heart dysfunction or combinations of these abnormalities [2]. The pulmonary vasodilator drugs approved for PAH may have deleterious effects or be ineffective if administered to patients with PVH or lung disease. Right heart catheterisation is essential for accurate diagnosis of all patients with suspected PH. Typically, differentiation between PAH and PVH is based on the pulmonary arterial wedge pressure (PAWP) ≤15 mmHg being the current criterion for diagnosing PAH [2].

The PAWP is used as a surrogate measure of left atrial pressure, but there are limitations to its interpretation. The PAWP waveform is complex and changes during the cardiac cycle, as well as with the varying intrathoracic pressure during normal respiration. PAWP may also be affected by changes in intra-alveolar pressures if the tip of the catheter rests in a region of lung where alveolar pressure intermittently exceeds pulmonary venous pressure [3]. The “gold-standard” measure of increased left heart pressure is the left ventricular end-diastolic pressure (LVEDP), which equilibrates with the pulmonary venous and left atrial pressure in diastole, except in patients with mitral valve disease or pulmonary veno-occlusive disease. It has been previously reported that in patients with PH, the PAWP and LVEDP are frequently discordant [4]. In that series, 37% of patients with suspected PAH based on the standard PAWP-based definition had increased LVEDP (defined as >15 mmHg), suggesting that PAH may be inappropriately diagnosed in some cases.

Some patients with left ventricular diastolic dysfunction may have normal resting LVEDP, but will show an abnormal increase in LVEDP in response to i.v. fluid loading [5, 6]. It is our routine practice to perform both right and left heart catheterisation on all patients being evaluated for PH, and when the PAWP and LVEDP are ≤15 mmHg, the patients are given a saline load to exclude post-capillary causes of the PH (PVH). Given that scleroderma is frequently associated with myocardial fibrosis and diastolic dysfunction, it is of great importance to identify these abnormalities [7–9]. Moreover, given reports of occult veno-occlusive disease in some scleroderma patients [7, 9], and the deleterious effects of vasodilators in this disorder, its detection is important [10]. In scleroderma, the presence of occult left ventricular diastolic dysfunction can contribute to PH as detected by exercise echocardiography [11]. In two previous studies of scleroderma patients undergoing haemodynamic assessment for suspected PH, using the standard PAWP-based definition, the incidence of PVH in the PH group ranged from 33 (14%) out of 243 to 17 (20%) out of 83, whereas PAH was diagnosed in ∼50% of cases [12, 13]. The proportion of scleroderma patients suspected of having PAH but in reality having occult PVH (OPVH) is unknown. We therefore investigated the prevalence of discordance between PAWP and LVEDP in the scleroderma population, studied the frequency of OPVH as assessed by the saline challenge, and examined the epidemiological factors and echocardiographic correlates of this phenomenon.

Results

Discussion

The main finding of our study was that in a scleroderma population being evaluated to diagnose PH, there was a low incidence of PAH compared with PVH, especially when OPVH cases were identified by means of LVEDP measurements before and after fluid challenge. Using LVEDP as the gold-standard arbiter of left heart disease is appropriate from the physiological stand-point. As far as we know, this study is the first study of its kind in scleroderma patients. Another interesting observation was that, despite there being a good clinical indication for haemodynamic evaluation of the patients, only 50% had PH at catheterisation, as defined by current criteria. The rate of detection of true PH at catheterisation has varied in other scleroderma populations, and this highlights the need for better catheterisation-referral algorithms, to avoid unnecessary procedures [12, 13].

This study is important, as invasive haemodynamic assessment is the cornerstone of PH diagnosis. In our cohort, 29 (55%) out of 53 scleroderma patients with PH had haemodynamically confirmed scleroderma-PAH by the conventional definition, which is similar to previous reports, [12, 13]. However, 38% of our patients had OPVH, as defined in our study. This potential misclassification of PVH as PAH may in part explain the relatively poor outcomes of scleroderma-associated PAH patients in clinical trials of PAH therapy, as these medications may vasodilate pre-capillary arteries allowing more flow through the lungs and thereby worsening left heart failure. At best, this phenomenon might explain some of the “nonresponders” to PAH therapy in those studies. Without extensive pathological data, we cannot be certain that patients whom we defined as “OPVH” are not simply a subset of scleroderma patients with genuine PAH and concomitant diastolic dysfunction. Given the high prevalence of both diseases in scleroderma patients, it would not be surprising to find them co-existing. The D_LCO and FVC/D_LCO data for OVPH more closely mirrored the PAH population than the PVH population, suggesting similarity at least in terms of lung disease. However, PH is classified and treated according to pulmonary haemodynamics rather than pathology. Therefore, we believe that defining this subpopulation is important and relevant. Moreover, a central principle in the diagnosis of PAH is that there is no post-capillary component. Thus, the ability to identify PAH is always hampered in the presence of post-capillary disease, and this is a particularly vexing issue in the scleroderma population. Additionally, some of the patients might have had a component of occult veno-occlusive disease [7, 9, 10]. It is unknown whether pulmonary veno-occlusive disease (PVOD) might be revealed by a saline challenge, or whether the PAWP would remain low. This would require a separate study in a larger population of pre-identified patients with high-risk features for PVOD [10], a normal resting PAWP, and perhaps a more modest saline challenge in view of the risks of inducing pulmonary oedema. Günther et al. [10] have reported a significant incidence of radiological findings suggestive of PVOD, suggesting a high incidence. We did not see this in our population, but it is possible that some of the PVOD was “screened out” before patients were referred to us for evaluation. Cardiac magnetic resonance imaging might also help detect patients with left ventricular involvement by scleroderma, but we did not have access to magnetic resonance imaging for the present study.

For patients with PVH, it has been suggested in guidelines that the TPG (mean PAP-PAWP) may be helpful in distinguishing “reactive” pre-capillary pulmonary vascular remodelling from a “passive” increase in PAWP, where the TPG is <12 mmHg in the latter [19, 20]. The change in TPG during exercise has also been used to study the response of PAH to therapy during exercise cardiac catheterisation in scleroderma patients [21]. In our study, both the TPG and DWG were lower in PVH (occult or not) than in PAH, consistent with less pre-capillary remodelling. Furthermore, after saline challenge, the TPG and DWG remained lower in OPVH than in PAH, and the change in TPG and DWG post saline was the same in OPVH and PAH. This suggests that the haemodynamic impairment, caused by whatever degree of pre-capillary structural remodelling is present, does not change with saline infusion. It is notable that in all three pulmonary hypertensive groups, DWG and TPG were higher than normal, suggesting some degree of pre-capillary disease, which was, as expected, the most extreme in PAH.

An additional novel finding in our study is that several (six (25%) out of 24) patients with low PAWP and low resting LVEDP who were re-evaluated after a modest fluid challenge were recognised as having OPVH rather than PAH. Although the need for fluid challenges to exclude PVH is described in the literature, there is little data regarding this diagnostic manoeuvre in PH patients. The effect of volume loading on LVEDP has been studied in several small physiological studies. The most frequently cited study focused on the effects of fluid challenge in patients with constrictive pericarditis, but also studied normal subjects and post-myocardial infarction patients [6]. Administration of 1 L of saline over 6–8 min in the six normal subjects raised the LVEDP/pulmonary capillary wedge pressure (PCWP) by a maximum of 3 mmHg, and LVEDP/PCWP remained below 12 mmHg in all cases [6]. In the same study, LVEDP/PAWP post-volume expansion was >15 mmHg in half of the patients who had previous myocardial infarction. Grossman et al. [22] increased preload by sudden passive elevation of the legs to a right angle to the trunk. Three of the subjects, aged 23, 50 and 55 years, had normal hearts. Measurements performed at 60 s, which would eliminate any Valsalva effect, showed LVEDPs of 10, 12 and 6 mmHg, respectively, with a maximum rise of 2 mmHg as compared with baseline [22]. Quinones et al. [23] also studied the effects of acute preload increase in humans. However, their goal was to deliberately raise LVEDP to ≥15 mmHg by rapid i.v. infusion of dextran solution and, thus, their study cannot be used to interpret the effects of a non-LVEDP target-directed fluid challenge. In a population at high risk for diastolic dysfunction associated with the metabolic syndrome, administration of only 500 mL saline over 5 min was able to reveal subjects where the PCWP increased by an average of 6.8 mmHg, to >15 mmHg [24]. Thus, while the data derived from volume expansion may not detect all cases of diastolic dysfunction, they seem to exclude diastolic dysfunction in normal controls. It might also be argued that the increased left-sided heart pressures seen after saline challenge represent increased filling of a dilated right heart, causing worsening ventricular interdependence. However, this hypothesis is refuted by the fact that in all cases, the rise in left heart pressure was greater than any rise in right sided pressures after the saline bolus (change in pressures; table 1). Further studies are needed to better define the "normal" LVEDP response to a standardised fluid challenge across a wide spectrum of normal subjects and patients of different age, sex and disease in order to refine the diagnostic utility of this manoeuvre.

Our study extends two previous investigations, but it is the only analysis to contain exclusively scleroderma patients and the first paper in PH patients in general to formally describe the possible diagnostic utility of the saline challenge. Halpern and Taichman [4] reported significant discrepancies between PAWP and LVEDP in a very large database (n=11 523) of right and left heart catheterisations performed in their centre. Amongst those patients with PAH according to the PAWP, 53.5% were found to have elevated LVEDP. Soto et al. [25] reported in abstract form 131 right and left cardiac catheterisations performed in their centre and demonstrated that 37% of patients with PAH according to PAWP had OPVH according to resting LVEDP. Our estimate of 38% overdiagnosis of PAH is consistent with these previous estimates. In a Bland–Altman analysis performed by Halpern and Taichman [4], PAWP was found to underestimate LVEDP by 2.9 mmHg with 95% limits of agreement of -12.9 to +9.5 mmHg. Our analysis had similar results. The mean difference between PAWP and LVEDP is very small in the overall population (-0.2 mmHg), but the limits of agreement are wide (-9.2 to +8.9 mmHg), such that PAWP cannot be reliably interpreted in an individual patient without an LVEDP unless the value is very low (<6 mmHg). Furthermore, PAWP ≤15 mmHg lacks specificity (68%) for detecting cases with LVEDP ≤15 mmHg, and even at the optimal threshold of PAWP ≤11 mmHg, as suggested by receiver operating characteristic analysis of the data, specificity increased to only 78%.

The data presented here also extend the work of Halpern and Taichman [4] by evaluating the effect of demographic factors on final PH diagnosis in scleroderma patients. We did not find any parameters that could help physicians predict which scleroderma patients are at higher risk of OPVH. This could possibly be explained by the relatively small sample size, and further research is needed in this area. It is recognised that patients with the metabolic syndrome have increased incidence of PVH, but we did not have reliable data on this factor for analysis [24].

Our study also provides important echocardiographic evidence for the similarity between OPVH and PVH patients, relative to PAH, in scleroderma patients with suspected PH. In our study, none of the patients had haemodynamically significant valvular abnormalities. Although the systolic function was preserved across the entire patient cohort, we demonstrated that patients with PVH or OPVH had evidence of increased filling pressures on echocardiography (higher E/e′) [11]. This is consistent with a previous study comparing echocardiographic parameters in patients with proven idiopathic PAH versus PVH, where E/e′ had 97% area under the curve for differentiating between the two conditions [13]. Furthermore, the left ventricular mass index was not able to distinguish between the groups. Patients with PAH have relatively small left atrial dimensions, which is consistent with an underfilled left heart and thus lower left-sided pressures. We do not suggest that echocardiography can substitute a full invasive haemodynamic assessment.

Our study has all the limitations associated with being a single-centre retrospective analysis. Despite that, the saline challenge was performed systematically, which reduces the possibility of selection bias within our population. Of course, referral bias might still be present, in that the scleroderma patients were referred to us for catheterisation only because of dyspnoea, a fall in carbon monoxide diffusing capacity in the absence of lung fibrosis, or an echocardiogram suggesting PH. Thus we cannot state what the prevalence of OPVH is in a general scleroderma population that has no present indication for catheterisation. However, because we systematically performed left heart catheterisations once there was an indication for catheterisation, the percentage of PVH in a population with an indication should not reflect selection bias. We did find a higher percentage of PVH than was previously described by Schreiber et al. [12] and Avouac et al. [13]. We do not have an explanation for this, although it is increasingly recognised that left heart disease is highly prevalent in North American populations.

In summary, a complete invasive haemodynamic assessment, including measurement of LVEDP with an appropriate saline challenge in scleroderma patients being evaluated for PH, may be critical to making the correct haemodynamic diagnosis by revealing a subset of patients with OPVH. However, further detailed clinical study is required to determine whether the clinical course and outcomes of the OPVH group are different from those of PAH, and whether the presence of OPVH has any impact on therapeutic success.