You are on page 1of 6

Journal of Critical Care 55 (2020) 157–162

Contents lists available at ScienceDirect

Journal of Critical Care

journal homepage: www.journals.elsevier.com/journal-of-critical-care

Fluid resuscitation in patients with end-stage renal disease on


hemodialysis presenting with severe sepsis or septic shock: A case
control study
Kartikeya Rajdev, MBBS ⁎, Lazer Leifer, MD, Gurkirat Sandhu, MD, Benjamin Mann, MD, Sami Pervaiz, MD,
Saad Habib, MBBS, Abdul Hasan Siddiqui, MD, Bino Joseph, MD, Seleshi Demissie, DrPH, Suzanne El-Sayegh, MD
Northwell Health-Staten Island University Hospital, Department of Internal Medicine, 475 Seaview Avenue, Staten Island, NY 10305, United States of America

a r t i c l e i n f o a b s t r a c t

Due to the potential risk of volume overload, physicians are hesitant to aggressively fluid-resuscitate septic pa-
Keywords: tients with end-stage renal disease (ESRD) on hemodialysis (HD). Primary objective: To calculate the percentage
End-stage renal disease of ESRD patients on HD (Case) who received ≥30 mL/Kg fluid resuscitation within the first 6 h compared to non-
Hemodialysis ESRD patients (Control) that presented with severe sepsis (SeS) or septic shock (SS). Secondary objectives: Effect
Severe sepsis of fluid resuscitation on intubation rate, need for urgent dialysis, hospital length of stay (LOS), intensive care unit
Septic shock
(ICU) admission and LOS, need for vasopressors, and hospital mortality. Medical records of 715 patients with sep-
Fluid resuscitation
Surviving sepsis campaign
sis, SeS, SS, and ESRD were reviewed. We identified 104 Case and 111 Control patients. In the Case group, 23% of
Early goal-directed therapy patients received ≥30 mL/Kg fluids compared to 60% in the Control group (p b 0.001). There was no significant
difference in in-hospital mortality, need for urgent dialysis, intubation rates, ICU LOS, or hospital LOS between
the two groups. Subgroup analysis between ESRD patients who received ≥30 mL/Kg (N = 80) vs those who re-
ceived b30 mL/Kg (N = 24) showed no significant difference in any of the secondary outcomes. Compliance with
30 mL/Kg fluids was low for all patients but significantly lower for ESRD patients. Aggressive fluid resuscitation
appears to be safe in ESRD patients.
© 2019 Elsevier Inc. All rights reserved.

1. Background the management of patients with SeS or SS was lacking. The Surviving
Sepsis Campaign (SSC) guidelines recommend an initial fluid resuscita-
Sepsis and bacterial infections are common in patients with end- tion of at least 30 mL/Kg of body weight in the first 3 h of diagnosis of
stage renal disease (ESRD). Infection is the second most common SeS or SS [7]. The original guidelines of the SSC were based on the land-
cause of mortality in ESRD patients after cardiovascular disease [1-3]. mark randomized controlled trial (RCT) published in 2001 by Rivers
Every year in the United States (US) 750,000 patients are diagnosed et al. Rivers et al. reported a 16% reduction in mortality (46.5% to
with severe sepsis (SeS), with an average annual increase of 13.0%. Sep- 30.5%) with the use of a protocolized based approach to treat patients
sis is currently the 10th leading cause of mortality in the US [1-3]. Septic with SeS or SS which was termed as Early Goal-Directed Therapy
patients are at a risk for fluid imbalance due to a widespread systemic (EGDT) [8]. The study also emphasized on early identification of sepsis
inflammation and increased capillary permeability [4]. The goal in the and time-dependent management of patients with SeS or SS and was
management of septic shock (SS) is to avoid hypovolemia and its related a breakthrough in the field of sepsis. Physicians became more aggressive
complications, such as hypotension, kidney injury, and multi-organ fail- in managing sepsis since the publication of this trial. Although EGDT
ure [5]. The mortality in septic patients prior to 2001 was considerably was incorporated into the SSC guidelines, there were concerns on the
high, ranging between 30% and 60% [6]. A general notion of urgency in validity of its individual components and the complexity of the protocol.
Recently, 3 large multi-center RCTs, namely ProCESS, ProMISe, and
⁎ Corresponding author at: 361 N 46TH St Apt 3208, Omaha, NE 68132, United States of ARISE, showed no improvement in mortality in patients who received
America. EGDT compared to the usual care or less invasive hemodynamic resus-
E-mail address: kartikeyarajdev@gmail.com, kartikeya.rajdev@unmc.edu (K. Rajdev), citation protocols [5,9-12]. The usual care in these 3 trials required
lazerleifer@gmail.com (L. Leifer), gsandhu@northwell.edu (G. Sandhu), early identification of SeS and SS with standardized screening protocols,
ben.mann91@gmail.com (B. Mann), spervaiz3@northwell.edu (S. Pervaiz),
ssaadhabib@gmail.com (S. Habib), drsiddiqui07@gmail.com (A.H. Siddiqui),
including measurement of lactic acid, as well as early administration of
bino.md@gmail.com (B. Joseph), sdemissie@northwell.edu (S. Demissie), fluids and antibiotics. In response to the new evidence from these three
selsayegh@northwell.edu (S. El-Sayegh). trials, the current SSC guidelines do not recommend use of a central

https://doi.org/10.1016/j.jcrc.2019.10.008
0883-9441/© 2019 Elsevier Inc. All rights reserved.
158 K. Rajdev et al. / Journal of Critical Care 55 (2020) 157–162

venous catheter to measure central venous pressure or central venous restrictive (b30 mL/Kg) fluid resuscitation in patients presenting with
oxygen saturation in all patients with SeS or SS who have received SeS or SS led to an increased rate of intubation, urgent dialysis, affected
timely antibiotics and fluid resuscitation [9]. The guidelines recommend hospital length of stay (LOS), intensive care unit (ICU) LOS, need for va-
administration of at least 30 mL/Kg of crystalloid fluid for SeS or SS sopressors, in-hospital mortality, and other secondary outcomes. We
within 3 h of presentation [7,9]. This initial fluid resuscitation of intended to shed light on this special subset of population. This informa-
30 mL/Kg is a strong recommendation (with a low quality of evidence) tion could be of use to emergency physicians, hospitalists, intensivists
and is meant to enable clinicians to initiate resuscitation in SeS or SS and nephrologists.
while waiting for further information on their hemodynamic status
and clinical parameters. The use of 30 mL/Kg fluid resuscitation has 2. Objectives
been described as usual practice in recent interventional and observa-
tional studies; although, there is limited literature that provides con- 2.1. Primary objective
trolled evidence to support 30 mL/Kg of fluid resuscitation [9,13,14].
The average amount of fluid patients received prior to randomization To assess the percentage of ESRD patients on HD (Case group) who
in ProCESS and ARISE trial was close to 30 mL/Kg and around 2 L in received initial fluid resuscitation of ≥30 mL/Kg in comparison to pa-
the ProMISe trial [9-12]. Many patients require more fluids after the ini- tients without ESRD (Control group) presenting to the hospital with
tial fluid resuscitation. The SSC guidelines recommend evaluation of he- SeS or SS.
modynamic status using clinical and physiological variables, as well as
other invasive and non-invasive monitoring. This includes maneuvers 2.2. Secondary objectives
such as a passive leg raise, fluid challenge, and bedside echocardiogra-
phy, among others [6,9,15]. 1. To evaluate and compare the following endpoints among the Case
In the hospital, physicians often encounter hypotensive ESRD pa- and Control group:
tients presenting from a dialysis center. Hypovolemia is usually the a) Incidence of mechanical ventilation for respiratory distress.
first impression of most clinicians. The diagnosis of sepsis in these pa- b) Duration of mechanical ventilation.
tients requires a high index of suspicion and is often not straightforward c) Time to order antibiotics since presentation.
[16]. There is a plethora of evidence that early identification of sepsis d) Hospital LOS and In-hospital mortality
and early administration of fluid resuscitation and antibiotics in patients e) ICU admission and ICU LOS
with hypotension or sepsis-induced tissue hypoperfusion decreases in- f) Incidence of urgent dialysis for volume overload
hospital mortality. It also improves morbidity by decreasing the occur- g) Need for vasopressors and number of vasopressors required dur-
rence of severe organ dysfunction, as well as both hospital and intensive ing hospitalization
care unit (ICU) length of stay (LOS) [7-9,17]. However, most of these h) Source of infection
studies have looked at a general patient population. There is a paucity
of information about the use of aggressive fluid resuscitation in ESRD
2. To perform sub-group analysis of the Case group (ESRD) and com-
patients on hemodialysis (HD). ESRD patients on HD often have a com-
pare the same secondary outcomes in the sub-set of ESRD patients
plex presentation during sepsis. Often, they appear volume overloaded
who received ≥30 mL/Kg fluid resuscitation vs those who received
on physical examination. However, they may actually be volume de-
b30 mL/Kg fluid resuscitation for SeS or SS.
pleted intravascularly [16]. Due to risk of potential complications such
as hypertension, pulmonary edema, respiratory failure, intubation, ur-
gent dialysis, and increased cardiac demand, health care providers are 3. Methods
hesitant to aggressively fluid-resuscitate septic patients who are at in-
creased risk of fluid overload, such as patients with ESRD on HD. The study was a case-control chart review study. Medical records of
A study by Akhter et al. found no difference in the rate of intubation, adult patients with International Classification of Diseases (ICD) code
hospital LOS, or mortality among patients with ESRD or congestive heart for a discharge diagnosis of sepsis, SeS, SS, ESRD and HD from the year
failure (CHF) who presented with SeS or SS and received an initial fluid 2015 to 2018 were retrieved from the department of health records of
resuscitation of ≥30 mL/Kg (N = 27) vs b30 mL/Kg (N = 169). They also our institution. Patients were included if they met the criteria for SeS
found that ESRD and CHF patients were significantly less likely to re- or SS based on the definitions by SSC. Patients younger than 18 years
ceive 30 mL/Kg of fluids in the ED compared to non-ESRD and non- old, pregnant patients, and patients with incomplete or illegible docu-
CHF patients [18]. Omar et al. found that patients on dialysis (N = 28) mentation were excluded. 715 patient records were screened for eligi-
received less fluids compared to patients not on dialysis (N = 30). bility, out of which 215 met the criteria for severe sepsis or septic
They concluded that conservative fluid resuscitation of ESRD patients shock. Patients were divided into two groups. The Case group belonged
in SeS or SS might be associated with improved outcomes [19]. Ozuzun to the patients with ESRD on HD admitted with a diagnosis of SeS or SS.
et al. showed that there were no direct complications of fluid overload The Control group included patients without ESRD (not on dialysis) ad-
in ESRD patients who received ≥30 mL/Kg crystalloids (N = 13) com- mitted with a diagnosis of SeS or SS. Only the authors of the study had
pared to ESRD patients who received b30 mL/Kg (N = 21) for SeS or access to the collected data that was stored in a secure environment
SS [20]. Lee et al. conducted a study on 329 patients with ESRD on dial- on REDCap.
ysis and SeS or SS to evaluate the relationship between fluid resuscita- The diagnosis of SeS and SS was based on the 1991 and 2001
tion bundle compliance and in-hospital mortality. However, they consensus documents [22]. Severe Sepsis was defined as two systemic
defined fluid resuscitation bundle compliance as receiving ≥20 mL/Kg inflammatory response syndrome (SIRS) criteria (temperature N 38 °C
within 1 h before and 6 h after SeS or SS recognition. They found no dif- (100.4 °F) or b 36 °C (96.8 °F), respiratory rate N 20 breaths/min, heart
ference in mortality between fluid resuscitation bundle complaint (N = rate N 90 beats/min or white blood cell count (WBC) N12 × 109/L or
214) vs non-compliant patients (N = 115) [21]. b 4 × 109/L and sepsis-induced organ dysfunction or tissue hypoperfu-
The aim of our project was to retrospectively study a series of pa- sion (systolic blood pressure ≤ 90 mmHg or MAP ≤70 mmHg or a fall
tients with ESRD on HD diagnosed with SeS or SS and compare with an- of N40 mmHg from baseline or serum lactate ≥4 mmol/L regardless of
other subset of patients without ESRD diagnosed with SeS or SS. The the blood pressure). Septic shock is defined as sepsis-induced hypoten-
study was intended to determine what percentage of patients with sion persisting despite a 30 mL/Kg fluid bolus.
ESRD on HD received an initial fluid resuscitation of ≥30 mL/Kg com- A REDCap data entry sheet was designed to collect the following
pared to non-ESRD patients, and whether aggressive (≥30 mL/Kg) vs information.
K. Rajdev et al. / Journal of Critical Care 55 (2020) 157–162 159

• Patient's demographics including age, gender, and ethnicity.


• Presence or absence of ESRD on HD (allocation to Case vs Control
group).
• Co-morbidities including a history of hypertension, diabetes, coronary
artery disease, CHF, cerebrovascular accidents, malignancy, immuno-
suppression, Human Immunodeficiency Virus (HIV) status, dementia
and the presence of other comorbidities were obtained from patient's
medical record and the Charlson Comorbidity Index (CCI) for all pa-
tients was calculated.
• Patients' clinical presentation and findings at initial presentation to
the ED including temperature, blood pressure, mean arterial pressure,
respiratory rate, heart rate, lactic acid, WBC count, creatinine and esti-
mated Glomerular Filtration Rate (eGFR) was collected.
• Time to ordering antibiotics from the time of presentation. The time of
presentation was defined as the time of triage in the ED.
• Amount of fluid resuscitation (in mL) within the first 6 h of presenta-
tion
• Incidence of intubation due to respiratory failure within the first 24 h
of presentation.
• Duration of mechanical ventilation in days.
• Need for vasopressors.
Fig. 1. Depicting the number of patients in the Case (ESRD) and Control (non-ESRD) group.
• Number of vasopressors required during hospitalization.
• Rate of ICU admission.
• ICU LOS in days. presentation in comparison to 60.36% (N = 67/111) of patients in the
• Incidence of urgent dialysis for volume overload. Control group presenting to the ED for SeS or SS (Fig. 2) (p b 0.001).
• Source of infection e.g. pneumonia, urinary track infection, intra- Our study evaluated many secondary outcomes which are depicted in
abdominal infection, skin/soft tissue/intravenous line infection, dialy- Table 2. The mean amount of fluids received in the Case group was
sis catheter or fistula site infection. 21.38 mL/Kg compared to 36.28 mL/Kg in the Control group which
• Hospital LOS in days. was statistically significant (p b 0.001). The mean total amount of fluids
• In-hospital mortality. (mL) in the first 6 h of presentation to ED was 1643.30 mL in the Case
group and 2669.80 mL in the Control group (p b 0.001). The median
3.1. Statistical analysis time to ordering antibiotics since presentation to the ED was not signif-
icantly different between the two groups. There was no significant dif-
Summary statistics were reported as mean (standard deviation) for ference in the rate of intubation for respiratory distress within 24 h of
normally distributed continuous variables and as median (interquartile presentation between the Case and the Control group (p N 0.05). 38 pa-
range) for not normally distributed continuous variables. Categorical tients had DNI (Do Not Intubate) as their code status and were therefore
variables were summarized as frequency and percentages. Differences excluded from the intubation rate calculation. 1.97% of patients in the
in continuous variables between case (ESRD) and control (non-ESRD)
group patients were evaluated using independent-sample t-Test or
nonparametric Mann-Whitney U test, as appropriate. The Chi-square Table 1
or Fisher's exact test was used to compare categorical variables between Baseline demographic and clinical characteristics of the case and control group.
groups. A multivariate logistic regression model was performed to ex- Case group Control group P-value
plore the difference between ESRD and non-ESRD patients in mortality (ESRD) (104) (non-ESRD) (111)
adjusting for confounding effects of age, fluid intake, time to antibiotics, Male 72 (69.23%) 72 (64.86%) 0.563
CHF, and CCI. All statistical tests of significance are two-sided and con- Female 32 (30.77%) 39 (35.14%)
ducted at the 0.05 level of significance. Statistical analyses were con- Age (years) 70.25 SD = 11.36 72.30 SD = 12.98 0.221
ducted using SAS (Statistical Analysis System) software Version 9.3. Ethnicity 0.243
• Caucasian 79 (75.96%) 92 (82.88%)
• African-American 13 (12.50%) 8 (7.21%)
4. Results • Asians 5 (4.80%) 2 (1.80%)
• Hispanics 3 (2.88%) 1 (0.90%)
We identified 104 patients with ESRD on HD with SeS or SS (Case • Others 4 (3.85%) 8 (7.21%)
*Charlson comorbidity index 8 (3) 6 (4) b0.001
group) and 111 patients without ESRD with SeS or SS (Control group)
*Lactic acid (mmol/L) 2.70 (3.15) 3.75 (3.9) 0.039
(Fig. 1). Both groups had similar distributions of patients in terms of Creatinine (mg/dL) 5.66 SD = 2.26 3.19 SD = 3.64 b0.001
age, sex and ethnicity. The majority of our patient population was el- GFR (mL/min) 11.31 SD = 6.13 35.93 SD = 24.69 b0.001
derly males with a mean age of 70.25 years in the Case group and SBP (mm Hg) 84.46 SD = 20.52 89.93 SD = 24.57 0.079
72.30 years in the Control group. Our sample was predominantly DBP (mm Hg) 48.96 SD = 11.23 53.03 SD = 12.68 0.014
MAP (mm Hg) 60.68 SD = 13.38 65.21 SD = 15.55 0.023
comprised of Caucasian patients, which is consistent with the popula-
Heart rate (beats/min) 98.70 SD = 21.37 110 SD = 24.32 b0.001
tion distribution and demographics of Staten Island, New York. CCI Temperature (F) 98.59 SD = 3.16 99.40 SD = 3.34 0.068
was calculated for all patients and patients with ESRD had significantly Respiratory rate (breaths/min) 20.46 SD = 4.29 21.54 SD = 4.69 0.080
higher CCI than the control group. Table 1 demonstrates the baseline *WBC (x109/L) 14.42 (11.69) 13.81 (9.5) 0.307
demographic and clinical characteristics of our patient population. Quantitative clinical characteristics are presented as mean and standard deviation (SD)
except the ones with an asterisk. Due to the skewed distribution of data, the characteristics
4.1. Primary and secondary objectives with an asterisk are presented as median with interquartile range in parenthesis. Abbrevi-
ations- ESRD- End Stage Renal Disease, GFR- Glomerular Filtration Rate, SBP-Systolic Blood
Pressure, DBP- Diastolic Blood Pressure, MAP- Mean Arterial Pressure, F- Fahrenheit, WBC-
We found that 23.08% (N = 24/104) of patients in the Case group White Blood Cell.
received ≥30 mL/Kg fluid resuscitation within the first 6 h of P values in bold are statistically significant.
160 K. Rajdev et al. / Journal of Critical Care 55 (2020) 157–162

Table 3
Multivariate logistic regression analysis for mortality.

Variable Odds ratio (Confidence P-value


interval)

Age (1 year increment) 1.029 (1.001–1.057) 0.044


Fluids in mL/Kg in first 6 h of presentation 0.981 (0.963–0.999) 0.035
Charlson comorbidity index 1.091 (0.954–1.247) 0.206
Time to order antibiotics (Hours) 1.007 (0.869–1.166) 0.930
ESRD vs non-ESRD 0.610 (0.310–1.198) 0.151
SBP (mm Hg) 0.992 (0.979–1.005) 0.235
Lactic Acid (mmol/L) 1.150 (1.042–1.269) 0.005
CHF 1.385 (0.712–2.685) 0.337

Parenthesis represent confidence intervals. Abbreviations- ESRD- End Stage Renal Disease,
SBP- Systolic Blood Pressure, CHF- Congestive Heart Failure.
P values in bold are statistically significant.

We performed a multivariate logistic regression analysis to explore


the difference in mortality by adjusting for confounding effects of age,
amount of fluids (mL/Kg) received in the first 6 h of presentation, time
to administer antibiotics (hours), presence of CHF, ESRD vs non-ESRD,
and CCI (Table 3). The results showed that age was significantly associ-
ated with mortality (p = 0.044) with an odds ratio of 1.029. This sug-
Fig. 2. Depicting the percentage of patients who received ≥ 30 mL/Kg fluid resuscitation in
the Case (ESRD) and Control (non-ESRD) group. gests that with every one-year increase in patient's age, the mortality
increased by around 2.9%, after adjusting for the other variables.
Patients with higher lactic acid were associated with higher mortality
Case group needed urgent dialysis for volume overload compared to 0% (p b 0.01). Fluid administration was independently associated with
patients in the Control group. This difference was not significant (p = lower mortality in patients with SeS or SS (p = 0.035). Other factors
0.228). Patients in the Control group were noted to have a significantly such as time to antibiotics, systolic blood pressure, CCI, and presence
higher ICU admission rate (p = 0.03) and the need for vasopressor sup- of ESRD were not found to be independently associated with mortality.
port (p = 0.02). Most of the patients who needed vasopressor support
required up to two vasopressors. The ICU LOS, hospital LOS, and dura-
tion of mechanical ventilation were not statistically significantly differ- 4.2. Sub-group analysis of ESRD patients (Case group)
ent between the two groups. Pneumonia was the most common cause
of SeS or SS in both the groups. Urinary tract infections were more com- A sub-group analysis was performed to compare the secondary
mon in the Control group whereas skin and soft tissue and intravenous outcomes between ESRD patients who received b30 mL/Kg fluid resus-
line infections were more common in the Case group. 39.42% of the pa- citation (Sub-group 1 - N = 80) vs those who received ≥30 mL/Kg fluid
tients in the Case group expired during the hospitalization compared to resuscitation (Sub-group 2 - N = 24). Table 4 demonstrates a similar
42.34% in the Control group. This difference in in-hospital mortality was, distribution of patients in both the groups with respect to age, sex,
however, not statistically significant (p = 0.68). CCI, and clinical parameters upon presentation to the ED (p N 0.05).

Table 2
Comparison of secondary outcomes between the case and control group.

Case group (ESRD) (104) Control group (non-ESRD) (111) P-value

Fluids received (mL/Kg) 21.38 SD = 15.86 36.28 SD = 19.22 b0.001


Total amount of fluids (mL) in first 6 h of presentation 1643.30 SD = 1156.50 2669.80 SD = 1261.10 b0.001
*Time to order antibiotics (Hours) 1.38 (2.00) 1 (1.50) 0.168
ICU admission rate 69 (66.35%) 89 (80.18%) 0.030
*ICU LOS (Days) 5 (5) 4 (6) 0.073
Need for vasopressors 56 (53.85%) 77 (69.37%) 0.025
Number of vasopressors 0.034
• 1 45 (80.36%) 53 (68.83%)
• 2 11 (19.64%) 14 (18.18%)
• 3 0 7 (9.10%)
• 4 0 3 (3.90%)
Intubation rate within 24 h of presentation for respiratory distress 7 (8.64%) 12 (13.04%) 0.467
*Length of mechanical ventilation (Days) 5 (4) 6 (2.5) 0.540
Need for urgent dialysis for volume overload 2 (1.97%) 0 (0%) 0.228
*Hospital LOS (Days) 8 (11) 9 (11) 0.743
Source of Infection b0.001
• Pneumonia 47 (45.19%) 53 (47.75%)
• UTI 7 (6.73%) 31 (27.93%)
• Intra-abdominal 16 (15.38%) 17 (15.32%)
• Skin/Soft tissue/IV line 20 (19.23%) 4 (3.60%)
• Dialysis catheter/Fistula site infection 2 (1.92%) 0
• Others 12 (11.54%) 6 (5.41%)
In-hospital mortality 41 (39.42%) 47 (42.34%) 0.680

Quantitative data are presented as mean and standard deviation (SD) except the ones with asterisk. Due to the skewed distribution of data, the characteristics with asterisk are presented
as median with interquartile range in parenthesis. Abbreviations- ESRD- End Stage Renal Disease, ICU- Intensive Care Unit, LOS- Length of Stay, UTI- Urinary Track Infection,
IV- Intravenous.
P values in bold are statistically significant.
K. Rajdev et al. / Journal of Critical Care 55 (2020) 157–162 161

Table 4 have a coexisting renal disease [23,24]. There is little literature available
Sub-group analysis of the case group. Baseline demographic and clinical characteristics of on this subset of patients. With this study, we have made an attempt to
the ESRD patients who received b30 mL/Kg fluids (Sub-group 1) vs those who received
≥ 30 mL/Kg fluids (Sub-group 2).
bridge this knowledge gap. We performed a retrospective chart review
of patients presenting to the ED with SeS or SS and evaluated the fluid
Sub-group 1 (80) Sub-group 2 (24) P-value resuscitation in patients with and without ESRD. We found that compli-
Male 56 (70%) 16 (66.67%) 0.802 ance with the SSC guidelines of at least 30 mL/Kg fluid resuscitation was
Age (years) 69.20 SD = 11.17 73.79 SD = 11.52 0.082 low for all patients at our institution but was much lower for patients
*Charlson Comorbidity Index 7 (2.5) 8 (3.5) 0.593
with ESRD. These findings are in line with the previously described
*Lactic acid (mmol/L) 2.75 (3) 2.55 (5) 0.793
SBP (mm Hg) 85.25 SD = 22.68 81.83 SD = 10.49 0.477 literature [19-22]. As mentioned above, this may be due to a fear of
DBP (mm Hg) 48.86 SD = 11.94 49.29 SD = 8.70 0.870 causing volume overlaod in ESRD patients.
MAP (mm Hg) 60.87 SD = 14.58 60.03 SD = 8.40 0.789 In our study, no significant difference in mortality was noted
Heart rate (beats/min) 98.78 SD = 21.32 98.46 SD = 23.98 0.950 between the ESRD and non-ESRD group. This is in contrast with the
Temperature (F) 98.53 SD = 2.69 98.75 SD = 4.46 0.768
Respiratory rate (breaths/min) 20.71 SD = 4.13 19.63 SD = 4.78 0.278
study by Omar et al. who found that despite receiving lesser fluids, pa-
*WBC (×109/L) 14.10 (11.61) 17.70 (12.40) 0.167 tients with ESRD had lower mortality compared to those without
ESRD. However, the overall mortality rate in their study was much
Quantitative data are presented as mean and standard deviation (SD) except the ones with
asterisk. Due to the skewed distribution of data, the characteristics with asterisk are pre- higher, indicating that it was a sicker cohort compared to our study
sented as median with interquartile range in parenthesis. Abbreviations- ESRD- End [20]. Although patients with ESRD are known to have significantly
Stage Renal Disease, SBP-Systolic Blood Pressure, DBP- Diastolic Blood Pressure, MAP- higher mortality in sepsis compared to non-ESRD patients [16], in our
Mean Arterial Pressure, F- Fahrenheit, WBC- White Blood Cell. study, we found no significant difference in their mortality. It is note-
worthy that non-ESRD patients in our study had a significantly higher
The sub-group analysis revealed several interesting findings lactic acid, heart rate, higher ICU admission rate, and need for vasopres-
(Table 5). There was no significant difference in the need for vasopres- sors compared to the ESRD group. Moreover, non-ESRD patients had a
sor support, number of vasopressors required, rate of ICU admission, mean creatinine of 3.19 mg/dL indicating a significant burden of renal
ICU LOS, and hospital LOS. The potential complications of volume disease in the control group. This may reflect a higher severity of their
overload were not found to be higher in sub-group 2 as there was no illness and hence a high mortality rate.
significant difference in the rate of intubation or urgent dialysis The multivariate logistic regression analysis showed that patients
between sub-group 1 and sub-group 2. There was no significant differ- who received adequate fluid resuscitation in the first 6 h had signifi-
ence in mortality between sub-group 2 which received more fluids and cantly lower mortality, regardless of the presence of ESRD or not.
sub-group 1 which received less aggressive fluid resuscitation. Another interesting observation was that higher age and higher lactic
acid were independently associated with a higher mortality compared
5. Discussion to the presence of CHF, ESRD or CCI, indicating that age and lactic acid
could be important markers to identify mortality risk in septic ESRD pa-
Patients with ESRD are a small, but important subset of patients pre- tients on HD. These findings were very similar to a recent study by Lowe
senting to the ED with SeS or SS. The current SSC guidelines do not pro- et al. on dialysis dependent ESRD patients with SS which also found that
vide guidance on the amount of fluid resuscitation required in patients age and elevated lactic acid are strong independent markers of mortal-
with ESRD on HD presenting with SeS or SS. It is important to mention ity [23]. Some recent studies suggest serum lactic acid as a new emerg-
that patients with ESRD may have been underrepresented in the recent ing vital sign and a predictor of mortality in SS patients [23,25]. Our
trials on EGDT in sepsis. Although they were not specifically excluded study is one of the very few to show an association of lactic acid with
from the trials, only b3.4% of the patients enrolled were described to mortality in ESRD patients on HD with SeS or SS.
Sub-group analysis helps in understanding the effect of fluid resusci-
tation on patient management outcomes in patients with ESRD on HD.
Table 5
We found no difference in any of the secondary outcomes. This suggests
Sub-group analysis of the case group. Comparison of secondary outcomes between ESRD
patients who received b30 mL/Kg fluids (Sub-group 1) vs those who received ≥ 30 mL/Kg that 30 mL/Kg of initial fluid resuscitation is probably safe in patients
fluids (Sub-group 2). with ESRD on HD. It did not lead to any complications related to volume
overload or affect in-hospital mortality. These findings are consistent
Sub-group 1 Sub-group P-value
(80) 2 (24)
with those of Akhter et al., Ozuzun et al., and Lee et al. who also found
no direct complications and no increase in mortality with aggressive
Fluids received (mL/Kg) in the first 6 h of 14.77 43.40 b0.001
presentation SD = 9.64 SD = 12.23
fluid resuscitation in patients with ESRD presenting to the ED with SeS
Total amount of fluids (mL) in first 6 h of 1279.40 2856.30 b0.001 or SS [19,21,22]. We propose that physicians should consider the use
presentation SD = 984.70 SD = 814 of adequate fluid resuscitation in patients with ESRD on HD. Frequent
*Time to order antibiotics (Hours) 1.5 (1.75) 1 (1.5) 0.064 assessment of hemodynamic status in these patients and use of bedside
ICU admission rate 52 (65%) 17 (70.83%) 0.806
echocardiography can guide us in adequately resuscitating ESRD
*ICU LOS (Days) 5 (5.5) 5 (4) 0.480
Need for vasopressors 41 (51.25%) 15 (62.50%) 0.360 patients on HD presenting to the ED with SeS or SS.
Number of vasopressors 0.708 Our study has some limitations that are worth mentioning. One of
• 1 32 (78.05%) 13 (86.67%) the limitations is that it is a single-center retrospective study and not
• 2 9 (21.95%) 2 (13.33%) a multicenter prospective trial; this may limit the ability to generalize
Intubation rate within 24 h of presentation 4 (6.35%) 3 (16.67%) 0.179
the results. Since this was a retrospective study, the results might
for respiratory distress
*Length of mechanical ventilation (Days) 4 (10) 6 (4) 0.825 have been influenced by an uneven distribution of baseline characteris-
Need for urgent dialysis for volume 2 (2.56%) 0 (0%) 1.00 tics between the two groups, which we tried to minimize by conducting
overload a multivariate logistic regression and also by performing a subgroup
*Hospital LOS (Days) 8 (11) 8 (9.5) 0.608
analysis within the ESRD group. In our study, although all patients re-
Mortality 31 (38.75%) 10 (41.67%) 0.816
ceived crystalloids for fluid resuscitation for SeS or SS, we did not record
Quantitative data are presented as mean and standard deviation (SD) except the ones with the type of crystalloid received or the use of blood products or albumin
asterisk. Due to the skewed distribution of data, the characteristics with asterisk are pre-
sented as median with interquartile range in parenthesis. Abbreviations- ESRD- End
for other reasons. This could have confounding effects on the outcomes.
Stage Renal Disease, ICU- Intensive Care Unit, LOS- Length of Stay. However, the preferred choice of intravenous fluids for resuscitation in
P values in bold are statistically significant. SeS or SS at our institution was mostly normal saline, followed by ringer
162 K. Rajdev et al. / Journal of Critical Care 55 (2020) 157–162

lactate. Another limitation of our study is that we did not calculate the [4] Castellanos-Ortega A, et al. Impact of the Surviving Sepsis Campaign protocols on
hospital length of stay and mortality in septic shock patients: results of a three-
APACHE score as a measure of severity of illness. Studies have shown year follow-up quasi-experimental study. Crit Care Med 2010;38(4):1036–43.
APACHE scores to be strong predictors of hospital mortality in septic pa- [5] Nguyen HB, et al. Early goal-directed therapy in severe sepsis and septic shock: in-
tients [26,27]. Moreover, the inability of the subgroup analysis to show a sights and comparisons to ProCESS, ProMISe, and ARISE. Crit Care 2016;20(1):160.
[6] Osborn TM. Severe sepsis and septic shock trials (ProCESS, ARISE, ProMISe): what is
significant difference could be due to the overall small sample size for optimal resuscitation? Crit Care Clin 2017;33(2):323–44.
subgroup analysis (N = 24 in higher fluid administration group vs 80 [7] Dellinger RP, et al. Surviving sepsis campaign: international guidelines for manage-
in lesser fluid administration group). ment of severe sepsis and septic shock: 2012. Crit Care Med 2013;41(2):580–637.
[8] Rivers E, et al. Early goal-directed therapy in the treatment of severe sepsis and sep-
Despite the limitations, our study had a large sample size compared tic shock. N Engl J Med 2001;345(19):1368–77.
to some of the previously mentioned studies in the literature and we [9] Rhodes A, et al. Surviving Sepsis Campaign: International Guidelines for Manage-
were able to highlight some key findings in this field that may be useful ment of Sepsis and Septic Shock: 2016. Intensive Care Med 2017;43(3):304–77.
[10] Peake SL, et al. Goal-directed resuscitation for patients with early septic shock. N
in the management of patients with ESRD on HD presenting with SeS or
Engl J Med 2014;371(16):1496–506.
SS. In conclusion, we found that among patients presenting with SeS or [11] Yealy DM, et al. A randomized trial of protocol-based care for early septic shock. N
SS, fewer patients with ESRD compared to non-ESRD, received Engl J Med 2014;370(18):1683–93.
guideline-based fluid resuscitation in the first 6 h. Patients with ESRD [12] Mouncey PR, et al. Trial of early, goal-directed resuscitation for septic shock. N Engl J
Med 2015;372(14):1301–11.
on HD who received ≥30 mL/Kg fluids did not have a higher rate of com- [13] Levy MM, et al. Surviving Sepsis Campaign: association between performance met-
plications related to volume overload or in-hospital mortality compared rics and outcomes in a 7.5-year study. Crit Care Med 2015;43(1):3–12.
to ESRD patients on HD who received less fluid. Fluid resuscitation is the [14] Levy MM, et al. The Surviving Sepsis campaign: results of an international guideline-
based performance improvement program targeting severe sepsis. Crit Care Med
cornerstone for the management of patients with SeS or SS. Our study 2010;38(2):367–74.
indicates that initial fluid resuscitation of 30 mL/Kg may be safe, or at [15] Griffee MJ, Merkel MJ, Wei KS. The role of echocardiography in hemodynamic as-
least not harmful, in patients with ESRD on HD and can potentially be sessment of septic shock. Crit Care Clin 2010;26(2):365–82 [table of contents].
[16] Abou Dagher G, et al. Sepsis in hemodialysis patients. BMC Emerg Med 2015;15:30.
incorporated in the sepsis triage bundle in the ED. This could improve [17] Province, E.G.-D.T.C.G.o.Z. The effect of early goal-directed therapy on treatment of
compliance with the SSC guidelines and at the same time, provide clini- critical patients with severe sepsis/septic shock: a multi-center, prospective, ran-
cians an option of withholding fluids if clinically deemed appropriate. As domized, controlled study. Zhongguo Wei Zhong Bing Ji Jiu Yi Xue 2010;22(6):
331–4.
there is very limited literature available on patients with ESRD present- [18] Akhter M, et al. Fluid resuscitation of septic patients at risk for fluid overload. Ann
ing with SeS or SS, we hope that our research will stimulate further Emerg Med 2017;70(4) [Abstract].
studies, including prospective trials, in this direction, to have a better [19] Omar Y, et al. Fluid resuscitation in patients with end-stage renal disease presenting
with severe sepsis or septic shock. Am J Respir Crit Care Med 2017;195:A7140
understating of the effect of fluid resuscitation in ESRD patients with
[Abstract].
SeS or SS. [20] Ozuzun P, Fried J, Grotts J. Early fluid resuscitation of end stage renal disease patients
with severe sepsis and septic shock. Crit Care Med 2014;42(12):A1582 [Abstract].
Financial disclosure [21] Lee JJ, Taylor SL, Adams JY. Fluid resuscitation and mortality in sepsis with end-stage
renal disease. Am J Respir Crit Care Med 2017;195:A5020 [Abstract].
[22] Levy MM, et al. 2001 SCCM/ESICM/ACCP/ATS/SIS international sepsis definitions
Authors have no financial disclosure. conference. Crit Care Med 2003;31(4):1250–6.
[23] Lowe KM, Heffner AC, Karvetski CH. Clinical factors and outcomes of dialysis-
dependent end-stage renal disease patients with emergency department septic
Declaration of Competing Interest shock. J Emerg Med 2018;54(1):16–24.
[24] Rowan KM, et al. Early, goal-directed therapy for septic shock - a patient-level meta-
Authors have no conflicts to disclose. analysis. N Engl J Med 2017;376(23):2223–34.
[25] Lee SM, An WS. New clinical criteria for septic shock: serum lactate level as new
emerging vital sign. J Thorac Dis 2016;8(7):1388–90.
References [26] Saqib A, et al. Predictors of in hospital mortality in patients on continuous veno-
venous hemofiltration in the ICU. Crit Care Med 2015;43(12):246 [Abstract].
[1] Keane WF, Shapiro FL, Raij L. Incidence and type of infections occurring in 445 [27] Sadaka F, et al. Predicting mortality of patients with Sepsis: a comparison of APACHE
chronic hemodialysis patients. Trans Am Soc Artif Intern Organs 1977;23:41–7. II and APACHE III scoring systems. J Clin Med Res 2017;9(11):907–10.
[2] Sarnak MJ, Jaber BL. Mortality caused by sepsis in patients with end-stage renal dis-
ease compared with the general population. Kidney Int 2000;58(4):1758–64.
[3] Collins AJ, et al. US renal data system 2013 annual data report. Am J Kidney Dis 2014;
63(1 Suppl):A7.

You might also like