Prognostic significance of hyponatremia in peripartum cardiomyopathy

Abstract

BACKGROUND

Hyponatremia is the most common electrolyte imbalance, however the prognostic significance of hyponatremia in peripartum cardiomyopathy (PPCM) remains unclear.

AIM

To assess the prognostic significance of hyponatremia in PPCM patients.

METHODS

We consecutively recruited patients with PPCM from 14 sites in Nigeria and followed them up for a median of 18 months. Serum sodium was measured at baseline, and patients with hyponatremia (< 135 mmol/L) were compared with those with normal sodium levels. Cox proportional hazards regression models were developed to identify factors associated with all-cause mortality.

RESULTS

Of the 191 PPCM patients recruited, 90 (47.1%) had hyponatremia at presentation. All-cause mortality among the hyponatremic patients (24/90; 26.7%) was significantly higher than among patients with normal serum sodium (7/101; 6.9%) (P-value < 0.001). In the Cox proportional hazards regression model, hyponatremia was independently associated with an increased risk of all-cause mortality [hazard ratio: 3.18 (95% confidence interval: 1.35-7.52; P = 0.008)], as were hypotension (systolic blood pressure < 100 mmHg) [2.22 (1.03-4.79); P = 0.043] and left ventricular ejection fraction (LVEF) < 25% [3.14 (1.47-6.73), P = 0.003].

CONCLUSION

Hyponatremia was common in our cohort of PPCM patients, and was independently associated with a threefold increased risk for all-cause mortality. Hypotension and a LVEF below 25% at presentation were also independent predictors of mortality.

Key Words: Peripartum cardiomyopathy; Hyponatremia; Outcomes: PEACE registry

Core Tip: Hyponatremia was common in our cohort of peripartum cardiomyopathy (PPCM) patients, and independently associated with threefold increased risk for all-cause mortality. Hypotension and a left ventricular ejection fraction below 25% at presentation were also independently associated with mortality. Our findings suggest that presence of hyponatraemia in PPCM patients at admission should prompt a thorough clinical evaluation and close clinical monitoring, to identify and address factors contributing to increased mortality risk.

INTRODUCTION

Hyponatremia is the most common electrolyte imbalance, occuring in 10%-30% of hospitalized heart failure (HF) patients[1]. It is a well-described independent predictor of poor prognosis for both hospitalised and ambulatory HF patients[2]. Importantly, the reversal of the hyponatremia in HF is associated with improved survival[1]. Hyponatremia in HF can be either dilutional, which is more common and results from excessive neurohumoral upregulation leading to impaired water excretion, or depletional, which is primarily due to excessive natriuresis with the use of loop diuretics, or a combination of both mechanisms[3]. Hyponatremia is a strong predictor of mortality among HF patients, but it is unclear whether it has a direct effect on mortality or is only a marker of HF severity[1,2].

Peripartum cardiomyopathy (PPCM) is an important cause of HF among women in northern Nigeria and many other parts of the world, with an incidence as high as 1 in 96 live births, and maternal age < 20 years, hypotension, tachycardia, and left ventricular ejection fraction (LVEF) < 25% at baseline independently predicted mortality[4,5]. However, the prognostic significance of hyponatremia in PPCM is unknown. In this post-hoc analysis of PEACE (PPCM in Nigeria) registry data, we aimed to determine the prognostic significance of hyponatremia in PPCM patients. We hypothesised that hyponatremia would be associated with worse prognosis in PPCM patients.

MATERIALS AND METHODS

In the PEACE registry, PPCM patients were consecutively recruited from 14 sites across the geopolitical regions of Nigeria, and followed for a median of 18 months. The details of the study protocol have been previously published[6]. All data and supporting materials are available with the published article. PPCM patients were enrolled into the study after obtaining written informed consent, if they had HF symptoms and LVEF below 45%. Ethical approval for the study was obtained from the Research Ethics Committees of all the participating centres before the commencement of the study[6]. The study conformed to the ethical guidelines of the Declaration of Helsinki on the principles of medical research involving human subjects[7].

Serum sodium was measured at baseline using standard methods at the study sites, and patients with hyponatremia (< 135 mmol/L) were compared with those with normal values[8]. Unrecovered left ventricular (LV) systolic function was defined as LVEF ≤ 55% at the last follow-up.

Investigators were encouraged to prescribe standard and locally available HF drug therapies. These included a diuretic, beta-blocker, angiotensin converting enzyme inhibitor or an angiotensin II receptor blocker or nitrate-hydralazine combination, and a mineralocorticoid receptor antagonist, unless such use was contraindicated or resulted in unacceptable side effects. In addition, drug doses were individually tailored, in accordance with guideline recommendations[9].

Statistical analysis

Data was cleaned then summarised using standard statistical methods including mean (± SD), and proportions. Patients were classified into two groups (hyponatremic vs normal sodium levels) and compared based on baseline and last follow-up characteristics of interest. χ², Fisher’s exact test, and student’s t and Mann-Whitney tests were used to compare categorical and continuous variables, as appropriate. Mortality (occurring from baseline profiling to end of follow-up) was analysed in all 191 cases with complete follow-up using the Kaplan-Meier method followed by a Cox proportional hazards model (entry method with proportional hazards confirmed by visual inspection) to derive adjusted hazard ratios (HR) and 95% confidence interval (95%CI), with consideration of baseline demographic and clinical characteristics as independent correlates of all-cause mortality. Variables included in the model have been previously linked with increased risk for either mortality or hyponatremia in PPCM and HF patients[2,5]. A two-sided P-value < 0.05 was considered statistically significant. Statistical analyses were performed using “Statistical Package for Social Sciences” version 23.0 software.

RESULTS

A total of 191 PPCM patients were recruited from 14 sites across five out of the six geopolitical regions of Nigeria (155 patients from North-West, 17 patients from North-Central, 15 patients from South-West, 1 patient from South-East and 3 patients from South-South).

Overall, 90 of the 191 (47.1%) PPCM patients had hyponatremia at presentation. The clinical characteristics and prescriptions for diuretics and other HF drugs were well-balanced between the 2 groups; with and without hyponatremia (Table 1). None of the patients in both groups reported history of smoking and diabetes mellitus, and only one patient in each group reported alcohol intake. In addition,16.7% and 15.8% of the patients with and without hyponatremia respectively (P = 0.877) were recruited as in-patients.

Table 1 Clinical profile of peripartum cardiomyopathy patients with and without hyponatraemia, n (%).

Variables	Hyponatremia (n = 90)	Normal serum sodium (n = 101)	P value
Serum sodium, mmol/L	126.7 ± 22.3	139.9 ± 3.2	< 0.001
Age, years	28.5 ± 7.2	30.3 ± 7.2	0.086
NYHA II-IV	66 (73.3)	83 (82.2)	0.141
Unemployment	66 (73.3)	74 (73.3)	0.878
Uneducated	67 (74.4)	67 (66.3)	0.289
Multiparity	65 (72.2)	76 (75.3)	0.757
Twin pregnancy	18 (20.0)	13 (12.9)	0.256
Obesity	1 (1.1)	5 (5.0)	0.270
Stroke	3 (3.3)	1 (1.0)	0.344
History of preeclampsia	17 (18.9)	19 (18.8)	0.960
Systolic BP, mmHg	108 ± 16	110 ± 18	0.255
Diastolic BP, mmHg	74 ± 13	76 ± 15	0.233
Heart rate/min	100 ± 17	100 ± 19	0.789
Ascites	22 (24.4)	27 (26.7)	0.718
Left atrium, mm	45.4 ± 7.0	44.6 ± 6.7	0.465
LV end-diastolic dimension, mm	61.7 ± 9.2	62.5 ± 6.9	0.525
LV ejection fraction at baseline	30.1 ± 8.0	29.8 ± 7.4	0.777
LV ejection fraction at last follow-up	44.2 ± 12.7	42.5 ± 13.3	0.398
Serum creatinine, µmol/L	82.8 ± 29.3	82.5 ± 29.3	0.949
Heart failure treatment
Loop diuretics	77 (85.6)	86 (85.2)	0.762
Thiazide diuretics	1 (1.1)	0	0.471
ACEI or ARB or BB	32 (35.6)	38 (37.6)	0.767
Spironolactone	84 (93.3)	94 (93.1)	0.942

Values were expressed as numbers with percentages in parentheses or as mean ± SD.

NYHA: New York Heart Association functional class; BP: Blood pressure; LV: Left ventricular; ACEI: Angiotensin converting enzyme inhibitors; ARB: Angiotensin receptor blockers; BB: Beta-receptor blockers.

All-cause mortality among the hyponatremic patients (26.7%) was significantly higher than among patients with normal serum sodium (6.9%) (P value < 0.001) (Table 2; Figure 1). The risk for all-cause mortality was independently increased by hyponatremia (HR 3.2; 95%CI: 1.4-7.5; P-value = 0.008), hypotension (systolic blood pressure < 100 mmHg) [HR 2.2 (95%CI: 1.0-4.8); P = 0.043], and LVEF < 25% [HR 3.1 (95%CI: 1.5-6.7); P = 0.003], after adjusting for loop diuretics (HR 1.1; P = 0.924), serum creatinine (HR 1.01; P = 0.425), preeclampsia (HR 0.8; P = 0.713) and tachycardia (HR 2.1; P = 0.130), at presentation respectively. Majority of the patients did not recover LV systolic function and rehospitalization was uncommon in both groups.

Table 2 Clinical outcomes among patients with and without hyponatraemia, n (%).

Variables	Hyponatremia (n = 90)	Normal serum sodium (n = 101)	P value
Mortality	24 (26.7)	7 (6.9)	< 0.001
Unrecovered left ventricular function	57 (63.3)	73 (72.3)	0.521
Rehospitalization	6 (6.7)	10 (9.9)	0.587

Values were expressed as numbers with percentages in parentheses.

The paper was written according to the STROBE (Strengthening the Reporting of Observational studies in Epidemiology) Statement-checklist of items for cohort studies.

Figure 1 Kaplan-Meier survival curves showing patients with and without hyponatraemia at risk of mortality at each month of follow-up. Log rank (Mantel-Cox) P < 0.001.

DISCUSSION

This study has shown, perhaps for the first time in PPCM patients, that hyponatremia at presentation was common and independently associated with threefold higher risk of all-cause mortality. The risk for mortality was also independently increased by hypotension and LVEF < 25% at baseline. It was however not associated with rehospitalisation and unrecovered LV systolic function. The LV systolic function in both groups still improved by 13-14%, irrespective of the hyponatremia status, in keeping with the natural history of PPCM. Thus, PPCM patients presenting with a combination of hyponatremia, hypotension and LVEF < 25% are at significantly high risk of short-term mortality, thus should be closely monitored and referred early to a specialized centre.

The prevalence of hyponatraemia in our study (47.1%) is higher than what was found in the European Society of Cardiology (ESC) HF long-term registry (20%) among patients hospitalized for acute HF[1]. In this retrospective analysis of the ESC Registry data, hyponatraemia at admission, defined as serum sodium concentration < 135 mmol/L, was associated with more advanced HF and worse in-hospital and post-discharge outcomes[1]. Importantly, HF patients with persisting hyponatraemia throughout admission had the highest in-hospital mortality as well as worst post-discharge outcomes, with all-cause mortality 60% higher than the reference group[1].

The prevalence of hyponatremia among HF patients of mixed aetiology admitted to a tertiary-level hospital in Ethiopia was lower than our finding among PPCM patients (28.9% vs 47.1% respectively)[10]. In that cohort, 49.3% were females and ischaemic heart disease was the predominant cause of HF (41%), followed by valvular heart disease (21%), hypertensive heart disease (17%) and dilated cardiomyopathy (11%). The lower prevalence of hyponatremia in this study was probably due to the lesser use of diuretics relative to our cohort (loop diuretics 29% vs 85%; spironolactone 9% vs 93%; thiazides 5% vs 1% respectively). However it is not clear whether PPCM has a predilection for hyponatremia.

In the Hyponatremia Registry for patients with euvolemic and hypervolemic hyponatremia in the United States, 29% of the HF patients had hyponatraemia. Unfortunately, the most commonly used treatment approaches for hyponatraemia in acute HF only increased serum sodium minimally, and most patients remained hyponatraemic at discharge[11].

Hyponatremia is considered a powerful independent marker of poor outcomes in patients with acute or chronic HF[12]. Given that hyponatraemia in HF is primarily dilutional, treatment is based on water restriction or vasopressin antagonists. Fluid should be restricted to less than 800-1000 mL/day to achieve a negative water balance and treat the hyponatraemia[12]. However, severe hyponatremia (£125 mmol/L) should be treated with intravenous hypertonic saline, targeting an increase in serum sodium of 1-2 mmol/L per hour, and ensuring that the increase is less than 8 mmol/L in a 24 hour period, to avoid the risk of myelinolysis[12].

Our results suggest that presence of hyponatraemia in PPCM patients at admission should prompt a thorough clinical evaluation and close clinical monitoring, to detect and treat the hyponatremia as well as other factors associated with increased risk of mortality. However, it is important to emphasise the post-hoc nature of our analysis and the overall pragmatic design of the PEACE registry; given the inherent challenges in optimally diagnosing and treating PPCM in an African setting.

CONCLUSION

Hyponatremia was common in our cohort of PPCM patients, and independently associated with 3-fold higher risk for all-cause mortality. Hypotension and LVEF below 25% at presentation were also independently associated with mortality.

ACKNOWLEDGEMENTS

The authors wish to acknowledge and thank all investigators that contributed data at the study sites.