We have found (i) significant associations between age, pre-operative Na concentration, fasting glucose on day 1, and post-operative hyponatraemia; (ii) that post-operative hyponatraemia is associated with other post-operative complications; (iii) no difference between reattendance at ED and/or readmission rates in those who developed or were discharged with post-operative hyponatraemia; and (iv) people developing post-operative hyponatraemia started from a lower pre-operative Na, and had a bigger drop in Na, than those maintaining a normal Na concentration peri-operatively.
Sah (2014) found a higher incidence of hyponatraemia (40%) in their elective joint arthroplasty population. Similar to our findings, they also found risk factors of post-operative hyponatraemia included pre-operative hyponatremia and older age. They also found female sex, lower body weight, knee in comparison to hip surgery, and bilateral knee arthroplasty to be risk factors. Our finding that age and pre-operative Na concentration independently predicts post-operative hyponatraemia supports the understanding of post-operative hyponatraemia, in this population, as a result of both baseline physiological vulnerability and peri-operative changes. Leung et al. (2012), in their database study, showed that pre-operative hyponatraemia (Na < 135 mmol/L) predicted longer LOS and increased risk of post-operative complications and 30-day mortality (Leung et al., 2012). Abola et al. (2019) also found that pre-operative hyponatraemia was associated with longer LOS and greater risk of reoperation (Abola et al., 2019). In our study, we found a significantly longer length of stay in those with post-operative hyponatraemia; however, pre- or post-operative sodium levels were not associated with a delayed discharge in multivariate analysis. We also found no significant differences in reattendance at ED, and/or readmission within 90 days between those who had post-operative hyponatraemia and those who did not.
That hyperglycaemia can lower Na is well-established. Traditional formulae for correcting Na for hyperglycaemia (Katz, 1973; Hillier et al., 1999) predict a smaller change in sodium than what was observed here, suggesting that there may be other contributory mechanisms. Diabetes mellitus has been associated with an increased incidence of hyponatraemia in community studies (Liamis et al., 2013; Mohan et al., 2013) and with increased ADH levels. Hyperglycaemia following arthroplasty has, to date, largely been studied for its association with joint infection (Hwang et al., 2015).
Intra-operative antibiotic use of teicoplanin was associated with a higher incidence of hyponatraemia compared to flucloxacillin and gentamicin. Due to multicollinearity, and using a stepwise backwards approach, flucloxacillin was no longer significantly associated with post-operative hyponatraemia in the multivariate analysis, whilst gentamicin and teicoplanin were excluded from the adjusted model.
Those with post-operative hyponatraemia did have higher rates of inpatient complications which may have contributed to their longer LOS. It must be acknowledged that several of the complications studied (e.g. pulmonary oedema and acute kidney injury) may be reflective of fluid balance issues and are therefore not mutually exclusive to hyponatraemia. Cardiovascular medications were significantly associated with hyponatraemia in univariate analysis, although they lost their significance in the multivariate analysis. Previous findings have reported that cardiovascular medications predict post-operative hyponatraemia (Udy et al., 2005; Hawkins et al., 2003; Mohan et al., 2013; Bissram et al., 2007). The relevant drugs are indicated for both hypertension and heart failure. It is not clear whether these associations between drugs and post-operative hyponatraemia are due to the drugs themselves, or are by virtue of underlying comorbidities predisposing to hyponatraemia. A propensity for heart failure and fluid retention could exacerbate the appropriate peri-operative release of ADH that is likely the main driver of post-operative hyponatraemia. Of note, our practice changed during the time of this audit and instead of being judged on a case-by-case basis, a blanket withholding of angiotensin converting enzyme inhibitors and angiotensin receptor blockers, in particular over the peri-operative period, was introduced. We found decreased height and decreased weight in those with post-operative hyponatraemia, but no significant difference in BMI between those with post-operative hyponatraemia and those with normal Na. Sah (2014) also reported lower body weight in those with post-operative hyponatraemia. It is possible that those with lower height and weight were subject to a relative excess of intravenous fluids infusion post-operatively. Whilst we have included infused fluid volumes (on day of surgery, available for n = 392 only) in our analyses (Table 1), we did not have net fluid balance available for sufficient patients to do this.
It is noteworthy that there was no significant increase in reattendance at ED and/or readmission to hospital for those patients who developed post-operative hyponatraemia. Our rates of discharge with hyponatraemia (10.8%) were lower than those reported in other series, with Hennrikus et al. (2015) reporting 28% of patients to still be hyponatraemic on discharge from hospital (Hennrikus et al., 2015). Hennrikus et al. (2015) included not only THA and TKA but also spine fusion and other surgeries, such as shoulder arthroplasty and foot and ankle procedures. Our findings, therefore, challenge the rationale for keeping otherwise well patients in hospital solely to monitor Na. It has been suggested that the standard reference range for Na utilised by most laboratories may not be universally applicable to inpatients (McKee et al., 2016). Our data would support this concept and the notion that milder degrees of hyponatraemia are common and probably do not require extensive investigation or treatment. We suggest that otherwise well patients with mild hyponatraemia can safely be discharged and followed up in the community. It is important to note that this is an observational finding. Evidence directly comparing inpatient and outpatient management of post-operative hyponatraemia is required.
The association between post-operative hyponatraemia and confusion did not withstand multivariate analysis, suggesting other factors such as age, pre-operative Na, ASA grade, type of surgery, and transfusion may have been mediating the relationship. It should be noted that we were not able to further categorise confusion, which may have been due to delirium, or pre-existing cognitive impairment. It is also likely, given other estimates from our population (Cunningham et al., 2017), that the incidence of confusion, due to either cause, was underestimated in this audit. There were no significant differences in falls or vasovagal episodes between those with and those without hyponatraemia. This is in contrast to the findings of a systematic review by Corona et al. (2018), which reported hyponatraemia to be significantly associated with an increased risk of falls across all 15 included studies (Corona et al., 2018). Ahamed et al. (2014) also found hyponatraemia to be independently associated with increased risk of admission-associated falls in their case-control study of medical inpatients (Ahamed et al., 2014).
The lack of effect of duration of surgery or blood loss during surgery on post-operative hyponatraemia in our cohort is contrary to other reports (Hennrikus et al., 2015). Our values of both are low (150 ml of blood loss in those with and those without hyponatraemia, and 53 min duration of surgery for those with hyponatraemia vs 55 min for those without hyponatraemia), suggesting we may be below the threshold where it makes a difference. For example, Hennrikus et al. (2015) noted significant differences in both duration of surgery and blood loss; duration of surgery was 2.6 h in those with no hyponatraemia vs 2.9 h in those who developed hyponatraemia post-operatively, and blood loss was 261 ml in those with no hyponatraemia vs 332 ml in those who developed hyponatraemia post-operatively (Hennrikus et al., 2015). Gender differences have previously been described in some studies, with females at a higher risk of hyponatraemia than males (Mohan et al., 2013), whilst other studies have found no gender differences (Hawkins et al., 2003), as in this audit.
Strengths of this study include the high numbers of participants with clinical chart-level information, including pre-operative drug information. The patients included in this audit are representative of the UK arthroplasty population with a similar mean age, gender ratio, and ASA grading to the National Joint Registry (NJR) population captured in the 10th Annual NJR Report 2013 for England, Wales and Northern Ireland (NJR, 2013) (Supplementary Table 4). A wealth of patient level data not included in other studies, or captured by the electronic systems on which larger studies are based, are presented here.
Available pre-operative Na measurements were taken in the 24 h prior to surgery for 723 patients, but for 276 patients, pre-operative Na was taken from the pre-operative assessment which varied between 1 and 552 days before surgery. The last recorded Na prior to discharge may have been taken more than 24 h prior to discharge. Variables differed in the extent of missing data, with some variables having more missing data than others.
Serum and urinary osmolarities, treatment of post-operative hyponatraemia, for example, fluid restriction, administered intravenous fluid volumes and daily fluid balance, and whether cardiac drugs were held over the peri-operative period have not been described. A validated comorbidity score has not been used pre-operatively, although ASA grade and medications can give some indication of pre-operative comorbidity. Information regarding complications was recorded pragmatically from the notes and strict criteria were not used. This was a single-surgeon, single-centre series which could influence generalisability.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.