Elevated central venous pressure/pulmonary artery pressure and risk of mortality or heart transplantation

Nine studies ^{Reference Bradley, Jassal, Moore-Clingenpeel, Abraham, Berman and Daniels6,Reference Myers, Waldhausen and Weber8–Reference Rodriguez, Mao and Mahle14} examined the relationship between elevated central venous pressure or pulmonary artery pressure and death or transplantation with only three clearly identifying that elevated pressures were predictive of premature death ^{Reference Ohuchi, Miyazaki and Negishi12,Reference Sethasathien, Silvilairat and Kraikruan13,Reference Inai, Inuzuka and Ono15} and only one showing that higher pulmonary artery pressure was predictive of need for heart transplantation. ^{Reference Bradley, Jassal, Moore-Clingenpeel, Abraham, Berman and Daniels6}

In 2017, Ohuchi et al were the first to show the predictive impact of higher central venous pressure on outcomes when measured early (6 months to 5 years post-Fontan), and late (≥ 15 years post-Fontan). Univariate analysis showed an increased risk of late mortality for every 1 mm Hg increase in central venous pressure in the group investigated early (hazard ratio of 1.46, p < 0.001) as well as investigated late (hazard ratio of 1.38, p = 0.0002). ^{Reference Ohuchi, Miyazaki and Negishi12} Subsequently, Sethasathien et al used multivariate cox regression to show that early post-operative elevated pulmonary artery pressure ≥ 23 mm Hg were predictive of mortality within 30 days of Fontan operation (hazard ratio of 23.2, p = 0.004). ^{Reference Sethasathien, Silvilairat and Kraikruan13} More recently, Inai et al performed a multi-institutional study involving nine centres in Japan and found that central venous pressure ≥ 16 mm Hg, measured a median of one year after Fontan completion, was an independent predictor of death in the following 2 decades (hazard ratio of 2.3, p = 0.003). ^{Reference Inai, Inuzuka and Ono15} The only study to find a predictive relationship between elevated central venous pressure and heart transplantation was performed by Bradley et al in 2019. This group looked at six patients with clinical deterioration who were hospitalised at least once for acutely decompensated heart failure and who had a CardioMEMS™ (Abbot, Chicago, IL, US), an implantable hemodynamic monitor inserted. They found that having higher mean pulmonary artery pressure was associated with greater odds of developing a heart failure-mediated event (odds ratio of 1.17, p < 0.0001). ^{Reference Bradley, Jassal, Moore-Clingenpeel, Abraham, Berman and Daniels6} This article also assessed changes in PAP during 10 one-minute exercise intervals and identified a great variability in the changes observed with exercise. Having variations greater than 4 mm Hg was predictive of worse outcomes. ^{Reference Bradley, Jassal, Moore-Clingenpeel, Abraham, Berman and Daniels6}

Another two studies pointed to better outcomes in patients with lower central venous pressure. ^{Reference Ohuchi, Ono and Tanabe10,Reference Mori, Hebson and Shioda11} In a group of 60 patients catheterised two decades after Fontan, Mori et al found a slightly higher central venous pressure in those who subsequently died or were transplanted, but time-related analysis could not identify any correlation. ^{Reference Mori, Hebson and Shioda11} In 2012, Ohuchi et al tried to define the characteristics of patients who were in “excellent” condition, defined as not having a re-hospitalization or adverse hemodynamics. The group of patients with excellent status had lower central venous pressure over serial catheterisation performed at 1, 5, 10, and 15 years, but the average of those differences was around 1–2 mm Hg and they considered having a central venous pressure over 16 mmHg to be adverse hemodynamics, precluding any predictive analysis. ^{Reference Ohuchi, Ono and Tanabe10}

Another indirect evidence of the lack of adequacy of measuring central venous pressure as an index of the health of the Fontan circulation is brought by a study by Mitchell et al who assessed the hemodynamics of patients before and after heart transplantation. ^{Reference Mitchell, Campbell and Ivy9} This work revealed that, in patients undergoing heart transplantation for late Fontan failure, the pulmonary vascular resistance is elevated above normal, but this fact could only be unveiled early after transplantation as the transpulmonary gradient increased by 8 mmHg early after transplantation (p = 0.0004 by paired analysis). Interestingly, the average PA pressure pre-transplantation was 17 mm Hg. ^{Reference Mitchell, Campbell and Ivy9} On the other hand, Rodriguez et al revealed that mean central venous pressure decreased significantly by 9 mm Hg post-transplantation (p < 0.004). ^{Reference Rodriguez, Mao and Mahle14}

Finally, of the nine studies reviewed, the first article to examine relationships between systemic pressure and death was performed in 1990 by Myers et al. ^{Reference Myers, Waldhausen and Weber8} This study, however, found no significance between post-operative mean pulmonary artery pressure in survivors and non-survivors.

Elevated central venous pressure/pulmonary artery pressure and protein-losing enteropathy

Of the 6 ^{Reference Feldt, Driscoll and Offord3,Reference Ohuchi, Yasuda and Miyazaki16–Reference Lin, Fanjiang and Wang20} studies that analysed relationships between central venous pressure or pulmonary artery pressure and protein-losing enteropathy, only one positively identified elevated central venous pressure to be a risk factor for developing protein-losing enteropathy. ^{Reference Ohuchi, Yasuda and Miyazaki16}

In a series of 26 patients with protein-losing enteropathy out of a cohort of 354 patients, Ohuchi et al demonstrated that central venous pressure measured one year post-operatively was an independent predictor of developing protein-losing enteropathy, with a 39 times greater risk of developing protein-losing enteropathy if central venous pressure was ≥ 12 mm Hg (Odds ratio of 3.49, p = 0.003). ^{Reference Ohuchi, Yasuda and Miyazaki16}

Contrastingly, a historical series was performed by the Mayo Clinic in 1996, where Feldt et al compared 42 patients developing protein-losing enteropathy to the remaining 385 patients of their cohort. ^{Reference Feldt, Driscoll and Offord3} Central venous pressure was measured at serial intervals after Fontan completion but was not a predictor of having protein-losing enteropathy when analysed as a continuous variable. The average pressure of those having protein-losing enteropathy was 19 mm Hg at the time of diagnosis. ^{Reference Feldt, Driscoll and Offord3}

The remaining studies observed elevated systemic pressures in patients who were already diagnosed with protein-losing enteropathy. In a series of 42 patients from the Mayo Clinic, John et al analyzed the impact of having a pressure above or below 15 mm Hg at the time of diagnosis of protein-losing enteropathy and demonstrated that patients with pressures < 15 mm Hg had greater survival (p = 0.03). ^{Reference John, Johnson, Khan, Driscoll, Warnes and Cetta17} Ono et al found no difference in mean post-operative central venous pressure between patients with protein-losing enteropathy who underwent Fontan versus biventricular repair (p = 0.47). ^{Reference Ono, Kasnar-Samprec and Hager18} In a seminal multi-centric study of 114 patients, Mertens et al identified that the mean PA pressure at the time of diagnosis of protein-losing enteropathy was 16 mm Hg, ^{Reference Mertens, Hagler, Sauer, Somerville and Gewillig19} and in another smaller study of 6 patients with protein-losing enteropathy, the mean PA pressure was 22 mm Hg at diagnosis. ^{Reference Lin, Fanjiang and Wang20} No correlations were presented in the latter two studies.

Effects of elevated central venous pressure/pulmonary artery pressure and plastic bronchitis

One study examined the central venous pressure in patients with plastic bronchitis and reported that their median pressure within the Fontan baffle was 16 mm Hg (Range 10–28). ^{Reference Avitabile, Leonard and Zemel21} This study only consisted of 14 patients diagnosed with plastic bronchitis, 13 of which had a Fontan completion, and thus no correlations were reported in the results.

Elevated central venous pressure/pulmonary artery pressure and Fontan-associated liver disease

Of the 10 articles ^{Reference Nakatsuka, Soroida and Nakagawa7,Reference Song, Kim and Huh22–Reference Inuzuka, Nii and Inai30} that identified a relationship between elevated systemic venous pressures and symptoms related to liver failure, only two ^{Reference Hansen, Dendale and Coninx29,Reference Inuzuka, Nii and Inai30} showed a predictive relationship.

At a median follow up of 10.2 years post-Fontan, Hansen et al reviewed the relationship of pulmonary artery pressure and liver disease score and found that both times since Fontan (odds ratio of 1.2, p < 0.01) and higher central venous pressure (odds ratio of 1.6, p < 0.001) were predictors of advanced Fontan-associated liver disease. ^{Reference Hansen, Dendale and Coninx29} Inuzuka et al reviewed data from 1117 perioperative survivors of Fontan across nine institutions in Japan with and without Fontan-associated liver disease. ^{Reference Inuzuka, Nii and Inai30} Central venous pressure, measured at an average of one year after Fontan completion, was significantly higher in the Fontan-associated liver disease group (p = 0.011). By Cox regression, higher central venous pressure (by 3 mm Hg increases) was an independent predictor of Fontan-associated liver disease (hazard ratio of 1.28, p = 0.042). ^{Reference Inuzuka, Nii and Inai30} Three additional studies, however, revealed that systemic venous pressures were not predictive of Fontan-associated liver disease, although pressures measured in these studies were no higher than 15 mm Hg ^{Reference Oka, Miyamoto, Tomoyasu, Hayashi and Miyaji24,Reference Patel, Kamande and Jarosz25,Reference Sethasathien, Silvilairat, Sittiwangkul, Makonkawkeyoon and Pongprot28}

Additional contrasting evidence was reported in two different studies performed by Schleiger and colleagues. This group initially described the incidence and severity of Fontan-associated liver disease. First, they looked at 21 patients with Fontan failure compared to 108 patients without failure and found that those with higher pulmonary artery pressure had advanced Fontan-associated liver disease (p = 0.006). ^{Reference Schleiger, Kramer and Salzmann26} In a subsequent study, they found that pulmonary artery pressure was significantly higher in patients with liver cirrhosis (p = 0.08), but that it was no longer associated with incidence of Fontan-associated liver disease. ^{Reference Schleiger, Salzmann and Kramer27}

The three ^{Reference Nakatsuka, Soroida and Nakagawa7,Reference Song, Kim and Huh22,Reference de Lange, Reichert and Pagano23} remaining studies that assessed systemic venous pressures and liver failure used imaging and biomarkers that were indicative of liver fibrosis and disease, however, all three studies showed no significant associations between central venous pressure and pulmonary artery pressure and liver dysfunction.

Exercise and central venous pressure/pulmonary artery pressure

Two retrospectives ^{Reference Asagai, Inai, Shimada, Harada and Sugiyama31,Reference Egbe, Miranda, Anderson and Borlaug32} and three prospective studies ^{Reference Goldstein, Connor, Gooding and Rocchini33–Reference Claessen, La Gerche and Van De Bruaene35} were included (Table 2).

Table 2. Effects of exercise on central venous pressure/PAP measurements.

CO = cardiac output; CVP = central venous pressure; PAP = pulmonary artery pressure; P = prospective; R = retrospective; SVPs = systemic venous pressures; W = watts.

All of the studies that were reviewed looked at central venous pressure or pulmonary artery pressure measurements during various exercise stress tests. In an earlier study, Goldstein et al. measured pulmonary artery pressure during exercise in the supine position and found that it increased significantly in six patients with Fontan circulation. Mean pulmonary artery pressure increased from 16 mm Hg at rest to 25 mm Hg at peak exercise (p = 0.001). ^{Reference Goldstein, Connor, Gooding and Rocchini33} Another study by Claessen et al measured mean pulmonary artery pressure during supine ergometer exercises in healthy controls versus patients with Fontan circulation. Mean pulmonary artery pressure was higher in the control groups at baseline and during exercise. The mean PA pressure in patients with a Fontan increased from 9 to 21 mm Hg. ^{Reference Claessen, La Gerche and Van De Bruaene35} Most notably, Navaratnam et al. showed in the initial figure of their manuscript that there is a wide variation in systemic venous pressures between controls and patients with a Fontan circulation with even a minimal amount of exercise. ^{Reference Navaratnam, Fitzsimmons and Grocott34} Exercise in this study consisted of patients completing 8–12 minutes on an electromagnetically braked bicycle. In patients with normal biventricular circulation, there was an increase in central venous pressure of 5–10 mmHg with exercise while these elevations were significantly more pronounced in patients with Fontan circulation. The slope of increase in venous pressure from even a minimal amount of exercise varied greatly in the small set of patients that were tested, with some of the patients having central venous pressure increasing to a level up to 35 mmHg. ^{Reference Navaratnam, Fitzsimmons and Grocott34} (Fig. 2) A retrospective study by Asagai et al examined central venous pressure changes before and during hand ergometer exercises. They also found great variations between patients in terms of the size of increase of PA pressures at exercise and identified some correlation between having an elevated PA pressure on exercise and some biological parameters. ^{Reference Asagai, Inai, Shimada, Harada and Sugiyama31} In a detailed analysis of hemodynamic parameters during exercise in 29 symptomatic patients with Fontan circulation, Egbe et al determined that a constructed “vascular reserve index,” extracted from the change during exercise in the ratio between pulmonary artery pressures and cardiac output, was able to better define pulmonary vascular function. They found that pulmonary vascular reserve was impaired in these patients during exercise, even when patients had a normal vascular resistance at baseline, and that these limitations in reserve during exercise were associated with end-organ dysfunction. Thus, this index might be best used to assess the health of the pulmonary vascular bed and to prevent the eventual end-organ damage of patients with a Fontan circulation. ^{Reference Egbe, Miranda, Anderson and Borlaug32}

Figure 2. Individual systemic venous pressure responses during exercise in Fontan patients and controls. In Fontan, the systemic venous pressure change (in mm Hg) relative to power output is described by: 15.97 + 0.073 x watts. In control, the systemic venous pressure change (in mm Hg) relative to power output is described by: 7.52 + 0.005 x watts (p < 0.0001). Reprinted from The American Journal of Cardiology, Volume 117 Issue 10, D. Navaratnam, S. Fitzsimmons, M. Grocott, H. B. Rossiter, Y. Emmanuel, G. Diller, T. Gordon-Walker, S. Jack, N. Sheron, J. Pappachan, J. N. Pratap, J. J. Vettukattil, and G. Veldtman. Exercise-induced systemic venous hypertension in the Fontan circulation. Pages 1667–1671 (2016), with permission from Elsevier.

Volume loading and central venous pressure/pulmonary artery pressure

Four prospective studies ^{Reference De Mey, Cools and Heying36–Reference Möller, Klungerbo and Diab39} and one retrospective study ^{Reference Peck, Averin and Khoury40} investigating the impact of volume loading on central venous pressure measurements were found. (Table 3) In 2015, the team from Leuven studied 28 patients with Fontan circulation having 32 catheterisation studies. They showed a significant difference (p < 0.05) in mean venous pressure before (14.6 ± 2.8 mm Hg, Range 10–21) and after intravascular volume loading of 15cc/Kg of 0.9% NaCl (18.3 ± 3.3 mm Hg, Range 2–19), with clear variation in individual responses to the volume challenge. They concluded that pressure changes in response to volume increases can be helpful in differentiating between cardiac and pulmonary limitations. ^{Reference De Mey, Cools and Heying36} A similar study by Möller et al. found a significant increase in mean central venous pressure following rapid volume expansion (5 mL of 0.9% saline/kg of body weight at room temperature) (p < 0.001) and this study also illustrated variable slopes in patients’ pressure responses to volume expansion. ^{Reference Möller, Klungerbo and Diab39}

Table 3. Effects of volume loading on central venous pressure/PAP measurements.

BVR = biventricular repair; CVP = central venous pressure; EDP = end-diastolic pressure; MVP = mean venous pressure; ODD = occult diastolic dysfunction; P = prospective; RVE = rapid volume expansion; R = retrospective; VAD = ventricular assist device.

One hospital in particular has implemented a ventricular stress test protocol using rapid volume expansion as part of standard clinical care in patients undergoing cardiac catheterisation. In their initial reports following this implemented protocol, Averin et al found that mean Fontan circuit pressure, as well as ventricular end-diastolic pressure, increases significantly following ventricular volume load (15 mL/kg of 0.9% NaCl). Mean pressures increased from 12.4 ± 2.2 mm Hg at rest to 15.2 ± 2.5 (p < 0.001) following volume loading. When looking closely at the slopes of the ventricular end-diastolic pressure in each patient, those who initially had a systemic ventricle end-diastolic pressure less than 15 mm Hg that increased above 15 mm Hg after volume loading were labelled as having occult diastolic dysfunction. ^{Reference Averin, Hirsch, Seckeler, Whiteside, Beekman and Goldstein37} This group later reviewed 28 patients with occult diastolic dysfunction and 61 without occult diastolic dysfunction and demonstrated that occult diastolic dysfunction was associated with late adverse outcomes related to Fontan failure (Fig. 3). Interestingly, central venous pressure pressures at baseline were not different in those with or without occult diastolic dysfunction. ^{Reference Peck, Averin and Khoury40}

Figure 3. Freedom from adverse clinical outcomes, stratified by occult diastolic dysfunction. Kaplan–Meier survival curve demonstrating freedom from adverse clinical outcomes, stratified by the presence (blue) or absence (red) of occult diastolic dysfunction. Occult diastolic dysfunction indicates occult diastolic dysfunction. Reprinted from Journal of the American Heart Association, Volume 12, Issue 1, D. Peck, K. Averin, P. Khoury, G. Veldhuis, T. Alsaied, A. M. Lubert, R. Hirsch, W. M. Whiteside, G. Veldtman, B. H. Goldstein. Occult diastolic dysfunction and adverse clinical outcomes in adolescents and young adults with fontan circulation. Page e026508 (2022), with permission from Wiley Blackwell.

A Japanese study demonstrated that in response to volume load by angiography (0.4 mg/kg of indocyanine green contrast medium), central venous pressure increased more in patients with a Fontan circulation than in patients with a biventricular circulation (10.2 ± 3.1 mm Hg versus 6.2 ± 3.4 mm Hg, p < 0.0001). ^{Reference Kim, McSweeney, Lee and Ivy38}