Pulmonary arterial wedge pressure in healthy subjects: a meta-analysis

To the Editor:

Based on current international guidelines, pulmonary arterial wedge pressure (PAWP) is critical for differentiating between pre- and post-capillary pulmonary hypertension (PH) and plays an important role in the diagnosis of left heart failure [1, 2]. The current PAWP threshold to define post-capillary PH is >15 mmHg, measured by right heart catheterisation (RHC) in the supine position [1]. Historical data suggest that the upper limit of physiological PAWP may be lower [3–5], although no systematic review and meta-analysis has investigated the normal range of PAWP considering major confounding factors. We aimed to fill this knowledge gap by assessing the normal value of PAWP based on the largest available database of the published literature on pulmonary haemodynamics, also taking into account possible confounding factors, such as age, sex and RHC methodology.

We previously performed three independent systematic literature reviews to describe pulmonary haemodynamics in healthy subjects [1, 6, 7]. For the current meta-analysis, we included only symptom-free healthy subjects that underwent comprehensive clinical evaluation and resting RHC in the supine position in order to investigate pulmonary vascular physiology. More details for these searches can be found at https://www.medunigraz.at/frontend/user_upload/OEs/kliniken/innere-medizin/pulmonologie/pdf/Normal_PAWP_Letter_Supplement_ERJ_final.pdf. We considered the upper limit of normal (ULN) PAWP as mean+2sd of a normally distributed variable (97.5 percentile). We chose three different approaches for the meta-analysis. First, studies providing individual patient data were analysed using a random-effects model (study being the random effect), thus assuming a degree of between-study heterogeneity. Second, for studies not providing individual data, the overall mean was estimated using the inverse variance method for pooling. Third, the same method was used for pooling both studies providing and not providing individual data, in order to calculate mean values and standard deviations. We further compared estimates of PAWP in men versus women, patients <50 years versus≥50 years and with averaged versus end-expiratory pressure tracing reading. We also analysed the influence of age and body mass index (BMI). We used R studio 2023.06.01 for analysis.

We included n=960 subjects out of n=49 studies (the list of included studies is available at https://www.medunigraz.at/frontend/user_upload/OEs/kliniken/innere-medizin/pulmonologie/pdf/Normal_PAWP_Letter_Supplement_ERJ_final.pdf). Of the n=49 studies, n=10 presented individual data from n=159 subjects. The median age of these subjects was 26 (interquartile range (IQR) 23–53) years; 67% were male, and median BMI was 22.2 (IQR 20.7–24.9) kg·m⁻². The other n=39 studies provided aggregated data from a total of n=801 subjects. The mean±sd age of these subjects was 46±6 years and 60% were male. Characteristics of the analysed subjects with individual or aggregated data are available at https://www.medunigraz.at/frontend/user_upload/OEs/kliniken/innere-medizin/pulmonologie/pdf/Normal_PAWP_Letter_Supplement_ERJ_final.pdf.

Based on all available studies, estimated mean±sd PAWP was 9.4±1.8 mmHg, corresponding to an ULN of 13.0 mmHg (table 1 and https://www.medunigraz.at/frontend/user_upload/OEs/kliniken/innere-medizin/pulmonologie/pdf/Normal_PAWP_Letter_Supplement_ERJ_final.pdf). The estimated mean PAWP was very similar in the studies providing individual (9.2 mmHg) or aggregated haemodynamic data (9.4 mmHg) (table 1). Age had no significant effect on PAWP (p=0.726), even if it was analysed as a dichotomous variable using a cut-off of 50 years (p=0.403) (table 1). In contrast, PAWP was significantly higher in women as compared to men (estimated mean PAWP 10.1 mmHg, 95% CI 9.1–11.0 mmHg, versus 8.4 mmHg, 95% CI 7.8–9.0 mmHg; p=0.003). Of note, due to the low number of available individual data in women (n=37), we did not calculate sex-specific ULNs for PAWP. BMI had no significant impact on PAWP (p=0.486). However, as BMI was <25 kg·m⁻² in most subjects, no conclusion can be drawn for ULN of PAWP in obese individuals. The method of pressure reading during RHC (assessing pressures at end-expiration versus averaging over 3–4 respiratory cycles) was not significantly associated with PAWP (p=0.926 for individual, p=0.293 for aggregated data). Based on a subgroup analysis of studies (n=14 studies including n=252 subjects) applying the mid-thoracic zero reference level for haemodynamic measurements, estimated mean PAWP was 9.4±1.4 mmHg, corresponding to an ULN of 12.2 mmHg.

Historically, since the first World Symposium on (primary) PH, the ULN of PAWP was considered 12 mmHg [8]. This threshold was based on a small number of investigations [9] and has later been implemented in many textbooks [5] and guidelines [2, 3]. However, the 2004 European Society of Cardiology PH guidelines first introduced a PAWP ≤15 mmHg threshold as part of the haemodynamic characterisation of pulmonary arterial hypertension (PAH) [10], and this threshold has not only been used by all upcoming PH guidelines, but also in nearly all therapeutic studies in patients with pre-capillary PH. Nevertheless, it has been emphasised that lower PAWP levels do not exclude the presence of significant left heart disease that may be uncovered using exercise or fluid challenge [2, 11].

The PAWP cut-off in the definition of pre- and post-capillary PH is of great importance, as it may decide on the clinical management of patients. We may ask the question: should a discrepancy between this cut-off and the ULN of PAWP lead to the re-consideration of the current haemodynamic definition of pre- and post-capillary PH? One argument for keeping the current PAWP threshold is that current PH drugs were approved based on randomised controlled trials using PAWP ≤15 mmHg as inclusion criteria [1, 12], and a recent analysis suggests that efficacy of treatment for PAH in large randomised controlled trials was not dependent on PAWP. In addition, in a large retrospective study of US Veterans referred for RHC, in which PAWP was modelled as a continuous variable, increasing all-cause mortality emerged with increased PAWP readings, beginning at >15 mmHg [13].

On the other hand, the results of our meta-analysis suggest that PAWP >13 mmHg exceed the 97.5th percentile of a normal distribution and may suggest disturbed diastolic function of the left ventricle, pericardial restriction, fluid overload, or elevated intrathoracic pressures due to emphysema with intrinsic positive expiratory pressures. This view is supported by two recent studies demonstrating that 1) during a median follow-up of 12 months, 25% of PAH patients with PAWP >12 mmHg developed PAWP values >15 mmHg, which was associated with decreased transplant-free survival, and 2) that PAWP >11 mmHg in chronic thromboembolic PH was associated with poor prognosis [14, 15]. Accordingly, some patients with pre-capillary PH and slightly elevated PAWP may suffer from a clinically relevant left heart disease or other pathologies despite PAWP values ≤15 mmHg, particularly if they have received diuretic therapy before RHC. In such patients, PH therapy should be carefully monitored for signs of pulmonary congestion.

Finally, it needs to be emphasised that PAWP should not be used in isolation, as a single parameter, to determine the diagnosis and therapeutic options of individual patients, but only in combination with other variables and the whole clinical picture [16].

Our analysis has limitations. The available studies are heterogeneous and individual data were only available for n=159 subjects. Among them, only n=39 were women, precluding the determination of sex-specific ULNs for PAWP. In addition, old and obese subjects were underrepresented in the published literature, therefore the detected ULN may not apply for them. The same is true for patients with PH, as PH was among the exclusion criteria for our study. Nevertheless, our investigation is the largest currently available meta-analysis of its kind and, due to ethical reasons, we cannot expect many more such studies to be done in the future. As further limitation, very few studies reported whether the V-wave was included to measure PAWP and thus no statement on within-cardiac changes of normal PAWP can be made. As we only included healthy and mainly non-obese subjects, our results do not preclude significant influence of RHC methodology on haemodynamic variables in patients with obesity or chronic lung or heart disease. Finally, by using pre-defined search criteria, we cannot exclude that we have missed some potentially suitable studies.

In conclusion, based on invasive haemodynamic measurements in healthy subjects, the ULN of PAWP in the supine resting position is 13 mmHg, which is independent of age and pressure reading. A PAWP >13 mmHg warrants further diagnostics for dysfunction of the left ventricle, valves or pericardium, fluid overload, or chronic lung disease with intrinsic positive end-expiratory pressure. Haemodynamic measurements such as PAWP are important contributors to the definition and management of different types of pulmonary hypertension and should be interpreted in light of a full clinical evaluation of a patient with possible cardiopulmonary disease. Clinical trials are necessary to better understand heterogeneity of treatment responses and their heterogeneity associations to PAWP and other haemodynamic measures.

Shareable PDF