A narrative review of paracetamol‐induced hypotension: Keeping the patient safe

Abstract Aim To understand the prevalence and epidemiology of paracetamol‐induced hypotension and clinical implications for contemporaneous practice. Design Narrative review. Methods In May and June 2020, an open‐date literature search of English publications indexed in ProQuest, PubMed, and EBSCO was conducted with the search terms ‘acetaminophen’ and ‘hypotension’ and related search combinations (‘paracetamol’, ‘propacetamol’, ‘low blood pressure’, ‘fever’, ‘sepsis’, and ‘shock’) to identify peer‐reviewed publications of blood pressure changes after paracetamol administration in humans. Results A pattern of blood pressure reduction following the administration of paracetamol is demonstrated in the 27 studies included in this review. Haemodynamic intervention often followed persistent blood pressure reduction, and was greatest in febrile critically ill patients who received parenteral paracetamol.


| INTRODUC TI ON
Nurses are required to escalate concerns about the quality and safety of care and are accountable for incidents that occur on their watch. Nurses have reported paracetamol-induced hypotension, through monitoring and surveillance (Lee et al., 2018). Nevertheless, paracetamol continues to be marketed and prescribed as a presumed safe and harmless drug. This poses a risk that paracetamol will cause haemodynamic instability, which vastly goes undetected, even in high-risk patient populations (Bae et al., 2017;Kang et al., 2018).
The risk warnings of current paracetamol manufacturers are questionable, because they do not compel clinical governors to regulate the administration of paracetamol in the hospital environment. This narrative review combines the nurse's voice with the scientific evidence presented in the literature, which demonstrates that paracetamol-induced hypotension is a real-world problem and identifies high-risk patients. Through narrative synthesis, we suggest methods of early detection and prevention of paracetamol-induced hypotension, wherever paracetamol is prescribed in the hospital environment.

| BACKG ROU N D
To treat fever or not remains a perennial debate (Mohr and Doerschuk, 2013). The benefits of reducing the oxygen and energy consumption induced by fever are contended by the theory that fever inhibits bacterial growth and activates physiological mechanisms that encourage pathogen clearance (Chiumello et al., 2017). Worldwide, paracetamol is used to treat fever and is ubiquitously administered in the hospital setting in multiple forms (Chiam et al., 2018). Moreover, paracetamol is the preferred non-etiotropic antipyretic to manage fever in patients with the coronavirus disease (Day, 2020). Despite a growing body of evidence suggesting that paracetamol induces hypotension in febrile patients (Bae et al., 2017), there are few apparent warnings within the disclosures by the pharmaceutical manufacturers (Kelly et al., 2016;Maxwell et al., 2019). Hypotension is well detected in the critical care environment, where nursing surveillance and haemodynamic monitoring occur continuously (Bose et al., 2016;Pfrimmer et al., 2017). Outside of this environment, monitoring and surveillance occur less frequently (Peet et al., 2019).
This increases the risk of delayed detection of paracetamol-induced hypotension, because paracetamol is considered a harmless drug.
Despite evidence on methods to treat paracetamol-induced hypotension (Kang et al., 2018;Lee et al., 2018;Schell-Chaple et al., 2017), there is a paucity of knowledge about who is most at risk of developing the condition and of ways to prevent it. This review critically appraises the current knowledge of paracetamol-induced hypotension to identify the gap in clinical practice and is intended to guide good governance, which prioritizes patient safety (Dresser, 2012).

| Design
This narrative review and synthesis is underpinned by a postpositivist paradigm, which explores the concept of paracetamol-induced hypotension and clinical implications of contemporaneous practice.

| Method
A literature search was conducted between 18 May-30 June 2020 over the writing period of the review, to identify existing research that was commensurate with a broad area of interest: paracetamolinduced hypotension (Figure 1). An open-date search of medical subject heading keywords 'acetaminophen' and 'hypotension' were searched to yield articles indexed in PubMed, ProQuest and EBSCO databases. Multiple combinations of the related terms 'paracetamol', F I G U R E 1 PRISMA search outcomes and records obtained from previously published reviews 'propacetamol', 'low blood pressure', 'fever', 'sepsis' and 'shock' were included in the search. The inclusion criteria were peer-reviewed articles written in the English language and sorted in order of relevance.
Two hundred and forty-four articles were found. The publication types included primary and secondary journal articles, abstracts and reports.
Literature reviews were manually searched for additional citations that were not extracted from the database search, whereas the reviews themselves were omitted. After duplicate articles were removed, 177 articles were identified. A further 125 studies were excluded if not primary research, grey literature, theses or if the title varied significantly from the haemodynamic effects of paracetamol, the key theme of this review. The abstracts of the remaining 52 articles were screened.
Twenty-five articles were then excluded when the abstract depicted low subject matter relevance. The remaining 27 articles were considered eligible, reviewed in full text and included in the study. The search was last updated on 30 June 2020. Herein, "paracetamol" refers to the active ingredient, acetaminophen, or propacetamol, the bio-precursor (Duggan & Scott, 2009;Hersch et al., 2008).

| Analysis
A narrative review and synthesis method was undertaken. This methodology enabled triangulation of concepts found within the literature to generate new knowledge and develop better ways to manage paracetamol administration in an acute-care hospital setting.
Contradictory research that refutes the paracetamol-hypotension association, the ways in which paracetamol-induced hypotension is therapeutically managed, and the types of monitoring environment that can detect this condition are discussed.

| RE SULTS
In total, 27 articles were included in the final data set for review (Table 1). Brown (1996), in two retrospective case studies of febrile intensive care (ICU) patients, reported the earliest known evidence that paracetamol may induce hypotension based on nursing observation, which identified a distinct association between vasopressor up-titration and paracetamol administration. Notably, a negative fluid balance was targeted in these patients to minimize the risk of pulmonary oedema (Brown, 1996). The author acknowledged the variations across ward and ICU monitoring environments, which accounted for the lack of evidence of paracetamol-induced hypotension in general wards (Brown, 1996). Subsequently, a more representative retrospective chart review revealed the effects of paracetamol in 191 paracetamol administrations (Mackenzie et al., 2000). Baseline and hourly heart rate (HR), blood pressure (BP), and body temperature (BT) recordings were analysed up to three hours after paracetamol administration, and BT and BP reduction consistently occurred postdose. Noteworthy, intravascular filling (IVF) or vasopressor up-titration occurred following 26.2% of paracetamol administrations. The authors suggested further research into paracetamol-induced hypotension and highlighted the absence of warnings from paracetamol manufacturers that hypotension was a significant risk (Mackenzie et al., 2000). Hersch et al., (2008) (Cantais et al., 2016). Boyle et al., (2010) performed another prospective study that explored the relationship between skin blood flow (skBF), paracetamol, and BP in critically ill febrile patients. The skBF was measured to compare with the haemodynamic responses after paracetamol administration in 29 febrile adults. The results showed that paracetamol induced a skBF increase and a corresponding BP reduction in 59% of the sample, of which 33% received treatment (Boyle et al., 2010) and the peak effect occurred at 60 min postadministration. Ray et al., (2017) performed a prospective observational study that further explored the paracetamol, BP and skBF association in 148 administrations of parenteral or enteral paracetamol in 31 children. The authors hypothesized that BP reductions would occur in febrile recipients and would be greater in children with a higher body surface area (BSA)-to-weight index. Significant BP reduction occurred after paracetamol administration, although this effect paradoxically did not occur in the higher BSA group (Ray et al., 2017). The authors concluded that although reductions in HR and stroke volume were contributing factors, the main effector of BP reduction was likely a reduced systemic vascular resistance and this effect was more significant in the sample that received parenteral paracetamol (Ray et al., 2017).

| Patients suggested to be at greater risk of paracetamol-induced hypotension
Potential associations have been previously indicated between paracetamol-induced hypotension and febrile patients (Kang  al., 2018), with some patient groups being at higher risk than others (Bae et al., 2017;Picetti et al., 2014). Mrozek et al. (2009), in their prospective observational study, measured the incidence and pathophysiology of paracetamol-induced hypotension and identified highrisk patients in a cohort with 1,507 paracetamol administrations, regardless of the indication. Only 1.33% of their cohort experienced BP reduction, and half of those affected required therapeutic management. Among those who developed paracetamol-induced hypotension, 80% were diagnosed with sepsis. The authors demonstrated that, statistically, patients with sepsis (p < .0001) and acute brain injury (p < .001) had a higher risk of developing paracetamol- authors concluded that although paracetamol was effective in reducing potentially harmful fever in patients with acute brain injury, paracetamol confers a risk of hypotension, which requires judicious management to prevent further damage to the injured brain (Picetti et al., 2014). In another study, the incidence of hypotension following enteral or parenteral paracetamol administration in patients with cerebral haemorrhage was investigated (Stoecker, 2014). The authors reported a consistent incidence of hypotension within 60 min of paracetamol administration that was more pronounced in recipients of parenteral paracetamol. The author noted a greater prevalence of potentially harmful paracetamol-induced BP changes existed in the clinical setting than was reported by paracetamol manufacturers (Stoecker, 2014).
In a retrospective study of 1,507 emergency department (ED) medical records, the temperature and haemodynamic values recorded pre-and postpropacetamol administration in febrile adult recipients were reviewed. Significant BP reduction occurred 30-60 min after propacetamol administration. In 10.7% of the participants, haemodynamic alterations that necessitated intervention occurred, all of which required IVF and 49% required vasopressor support. These results confirm that propacetamol provokes BP reduction in febrile patients and must therefore be carefully considered against the risk of causing haemodynamic alterations, in agreement with other research which demonstrates that paracetamol administration may not always be beneficial in febrile patients (Boyle et al., 2010). The clinical indicators that showed a greater tendency for propacetamolinduced hypotension included concurrent heart failure or experiencing chills (Bae et al., 2017).
In another retrospective observational study of six critically ill patients who received paracetamol, 48 observation cycles of serial temperature and haemodynamic indices were reviewed (Krajcova et al., 2013). The authors reported paracetamol-induced hypotension occurred in 45% of the sample. This effect was attributed to peripheral vasodilation and reduced cardiac output (Krajcova et al., 2013). Similar to previous research (Bae et al., 2017), the authors suggested that patients with cardiac compromise had a higher incidence of paracetamol-induced hypotension.  paracetamol-induced BP reduction occurred and that the incidence of this phenomenon may be underreported (Duncan et al., 2012).

| Parenteral paracetamol-induced hypotension
In contrast to previous observation studies, a non-randomized retrospective study exploring the haemodynamic changes in sus-

pected influenza A patients who received IV propacetamol was
conducted by Lee et al., (2018). The inclusion criteria were normotensive, febrile, adult ED patients who had a positive result on a rapid influenza swab. Propacetamol induced a reduction in BP, HR and BT, and significant BP reduction was detected in 29.7% of patients, of whom 20% required IVF. Similar to Boyle et al., (2010), the authors suggested that a major change in skBF may have contributed to the observed propacetamol-induced BP reduction (Lee et al., 2018).
The study's retrospective nature was a limitation because BP monitoring was episodic and not protocol-based. Therefore, the precise time at which the BP trough occurred may not have been recorded.
Furthermore, the authors demonstrated that the pre-drug BP values were higher in the group that experienced significant BP reduction and the postdrug BP values were similar. Nonetheless, the study supports the theory that IV propacetamol causes BP reduction in febrile patients.
A recent retrospective study by Kang et al., (2018)  To stabilize the IV form of paracetamol, most of the commercially available preparations contain mannitol, which potentially has implications in haemodynamically vulnerable populations . Following a review, a double-blinded RCT was performed to compare the haemodynamic effect of paracetamol with saline and mannitol solutions in 24 healthy volunteers (Chiam, Weinberg, Bailey, et al., 2015). The results showed that paracetamol reduced BP and systemic vascular resistance immediately after the infusion was commenced, which was not demonstrated in the mannitol or saline comparators. This study was unique, because it demonstrated paracetamol-induced BP reduction even in the absence of disease confounders. The authors attributed the BP reduction to paracetamol-induced vasodilation (Chiam, Weinberg, Bailey, et al., 2015)).
More recently, Nahum et al., (2019) performed a retrospective chart review of 100 records to explore the haemodynamic effects of IV paracetamol in critically ill children. The results showed that 39% of children treated with paracetamol experienced a significant reduction in BP and HR within 120 min following the administration. The authors concluded that IV paracetamol provided pain relief and therefore enhanced a reductive effect on both HR and BP without haemodynamic compromise (Nahum et al., 2019). Following this chart review, Nahum et al., (2020) performed a retrospective observational study, which explored the haemodynamic effect of IV paracetamol administration in paediatric patients who were admitted to the critical care unit with septic shock and who also received inotropic support (Nahum et al., 2020). The study indicated that IV paracetamol administration caused significant mean arterial hypotension in 32.4% of recipients. The authors suggested that IV paracetamol administration in critically ill children in septic shock, who also require inotropic support, can exacerbate hypotension.
The authors cautioned clinicians on the need for timely intervention of any resultant untoward haemodynamic manifestation (Nahum et al., 2020).

| Contradictory research
In a retrospective study of 100 medical records from the perioperative environment, criticism of the haemodynamic safety of rapidly infused parenteral paracetamol was challenged (Needleman, 2013).
The author concluded that although parenteral paracetamolinduced BP reduction, contrary to previous studies, this was not considered clinically relevant (Needleman, 2013). There are a few reasons that these findings should be interpreted with caution and cannot be extrapolated to all patients. The sample frame was highly selective and limited to ambulatory preoperative patients, a perceivably healthy patient demographic. This prohibits the authors making a general recommendation, because it cannot draw comparisons with febrile critically ill patients and could be construed as optimistic bias (Jansen, 2016). The authors contended that BP reduction occurred postparacetamol administration because of a higher-than-normal pre-treatment BP recording that was attributed to white coat syndrome, a hypothetical psychosomatic response in patients within hospital environments that transiently elevates their BP (Needleman, 2013). Furthermore, the study fails to explore the existence of haemodynamic change beyond 5 min of monitoring, suggesting that an element of knowledge distortion may co-exist.
Previous research has demonstrated that paracetamol-induced hypotension occurs more than 15 min postinfusion and often persists for much longer (Kang et al., 2018;Lee et al., 2018). Moreover, a funding affect bias is possible given that the study's funding source was the same pharmaceutical company that produced the drug. It is important to bear in mind the potential for caveats of knowledge distortion in this study. Together, the methodological bias renders this knowledge doubtful and antithetically encourages hesitancy in clinicians when administering parenteral paracetamol.

| Therapeutic management of hypotension
The prevalence of paracetamol-induced BP reduction and the types of therapeutic intervention, if any, that was used to manage it has been summarized ( Table 2). The adverse effects of sustained hypotension are well documented in the literature (Jones et al., 2006;Vincent & Leone, 2017) and are often a symptom that manifests in shock states (Shankar-Hari et al., 2016), which includes heterogeneous causes in septic conditions, but may include vasoplegia (Sharawy, 2014) and a syndrome of fluid inadequacy and maldistribution (Shankar-Hari et al., 2016). These tenets underpin research which suggests that early tailored IVF resuscitation and vasopressor therapy aim to optimize the haemodynamic conditions seen in septic shock states (Shankar-Hari et al., 2016;Walker et al., 2014). Furthermore, Parkin and Leaning's mean systemic filling pressure concept suggests that intravascular space requires a level of static pressure to enable a viable haemodynamic state (Parkin, 2014;Parkin & Leaning, 2008).
Despite many other variables to consider, the pragmatic inference suggests that, as the intravascular volume or circulatory fullness increases, so too does the mean systemic filling pressure and systemic venous return (Parkin, 2014;Parkin & Leaning, 2008). This research supports the theory that IVF is fundamental to maintaining a viable BP, which becomes critical in the management of hypotension in septic shock states (Beck et al., 2014). The significance of a greater hypotensive effect in heart failure (Bae et al., 2017) or in elderly patients  is important, because it conceivably relates to the iatrogenic minimalist approach to IVF. In heart-failure or elderly patients, there is a tendency to impose fluid restrictions or only cautiously administer IVF for fear of overloading the diseased heart (Philipson et al., 2013). Integrating these concepts supposes that parenteral paracetamol has a greater hypotensive effect in patients with less intravascular reserve.
Vasopressor therapy is another strategy for managing hypotension in septic shock (Walker et al., 2014), despite the lower effectiveness in patients without sufficient intravascular volume. This focusses attention on IVF as the first-line treatment of hypotension in septic shock (Walker et al., 2014). Thus, if an empty elastic tube (blood vessel) is compressed (vasopressor therapy), then the tube will essentially collapse. If the tube is prefilled before it is compressed, a turgor level is achieved within the vessel, enabling less need for vasopressor tension to achieve haemodynamic stability. Taken together, although these theories support the notion that optimal IVF prior to parenteral paracetamol administration in febrile patients may reduce the incidence of paracetamol-induced hypotension, definitive evidence of this phenomenon is yet to be determined.

| Monitoring environment
There is evidence that febrile critically ill patients are at the greatest risk of developing paracetamol-induced hypotension, though this association is only made possible by the continuous monitoring environments these patients inhabit. Extensive research has been carried out on paracetamol-induced hypotension in the critical care environment, whilst the availability of research pertaining to general ward environments is limited (Table 2).
Patients can and do deteriorate outside of critical care environments (Peet et al., 2019). Moreover, combining the variables of fever, hypovolaemia, and paracetamol prior to optimal fluid resuscitation poses a greater risk of hypotension. Capturing this effect is more likely to occur if BP monitoring is frequent, such as in higheracuity environments (Hope et al., 2018). Patients who inhabit the general ward environment, however, may be prescribed parenteral paracetamol for pain or fever, which is administered at the discretion of their bedside care providers. Without close observation, patients may have a paracetamol-induced trough in BP, which vastly goes undetected (Figure 2). Despite the evidence described in the literature, there are currently insufficient haemodynamic warnings available within disclosures made by pharmaceutical companies that market paracetamol (Kelly et al., 2016;Maxwell et al., 2019).
Counterintuitively, such omission can influence clinical practice guidelines for safe paracetamol administration.

| Epidemiology
Taken together, the studies included in this review provide sound evidence that paracetamol can induce a prolonged episode of hypotension in febrile critically ill patients. This effect is perceivably more apparent in patient populations with less vasomotor control.
Other significant epidemiological risk factors, which are consistent with this theory, include concurrent cardiac compromise and paediatric and elderly populations. The underlying aetiology and its extent are complex and can be influenced by pre-existing conditions, age and the native influences of septic shock (Bae et al., 2017;Daniels et al., 2019;Ray et al., 2017). The heterogeneous nature of sepsis is a confounding variable, because patients are potentially already compromised by a hypovolaemic state (Semler & Rice, 2016) and vasoplegia (Sharawy, 2014). The syndrome of sepsis itself, though highly variable, requires a varied approach to fluid volume in resuscitation, which is often combined with vasopressor drug administration (Keijzers et al., 2020). Higher-acuity monitoring and surveillance methods are required to detect instability in the haemodynamically vulnerable patient population to ensure early therapeutic management.
Unpacking this knowledge demonstrates that a corrective intervention is frequently required to maintain haemodynamic stability after the occurrence of paracetamol-induced hypotension. Despite reports of how paracetamol-induced hypotension is retrospectively managed, few insights have been suggested for prevention. In patients with an intrinsic tendency to develop paracetamol-induced hypotension (Bae et al., 2017;Boyle et al., 2010;Picetti et al., 2014;Ray et al., 2017), it is worth considering whether parenteral paracetamol is safe to administer prior to achieving an optimal state of euvolaemia.
This evidence compels the arbitrators of safe governance to determine the appropriate hospital monitoring environment to detect paracetamol-induced hypotension, which patients might receive paracetamol and the appropriate route of administration. The practice of close haemodynamic monitoring following paracetamol administration and correction of hypovolaemia prior to paracetamol administration should be of value to clinicians and should be widely applied in the hospital environment. This not only encourages practitioners to steward a safer and more rigorous monitoring strategy to detect paracetamol-induced hypotension, but also surmises ways to prevent the condition.

| Limitations
The outcomes of this review need to be considered in light of its limi-

| IMPLI C ATI ON S OF THIS RE VIE W FOR N UR S ING PR AC TI CE AND FUTURE RE S E ARCH
This review has important implications for future nursing practice in that paracetamol-induced hypotension is evidenced in research and is better detected in close monitoring environments, to ensure timely intervention. There is scope for future research around paracetamol-induced hypotension, which could include its occurrence in the general ward environment and a survey of existing acute hospital guidelines. Notably, the concept of pre-emptive intravascular filling as a moderator variable emerges as a hypothesis, which could be usefully explored in future research.

| CON CLUS ION
The results of this review present a plethora of evidence that suggests a relationship between paracetamol administration and hypotension. Commensurate with the aims, this paper canvased epidemiological concepts and revealed populations that demonstrated a greater association between these variables.
Furthermore, there is currently minimal contradictory research available that accurately contends the hypotension-paracetamol relationship. Despite the evidence, there is no effective system to regulate paracetamol administration, and the drug continues to be marketed without significant haemodynamic risk warnings.
Overall, this review has important clinical implications for future practice and calls for a long-awaited change in the overarching governance of this drug to protect the public and to minimize harm. This not only protects those receiving the drug, but also those administering it.

ACK N OWLED G EM ENTS
I would like to thank Dr. Marcel Van der Heiden for his clinical advisory and the Bairnsdale Regional Health Service for their patience and providing time for me to conduct this research.

CO N FLI C T S O F I NTE R E S T
The author declares that there is no conflict of interest with respect to the research, authorship and/or publication of this review.

E TH I C A L A PPROVA L
This review did not require research ethics committee approval.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data and evidence of this review are contained in the manuscript