Abstract
Background.
Preoperative anaemia is associated with adverse surgical outcomes, yet it is unclear whether anaemia-outcome relationships are consistent for women and men.
Methods.
This observational cohort study was conducted between August 1, 2012 and July 31, 2022 in adults undergoing surgery requiring postoperative hospitalisation at a large academic medical centre in the USA. Anaemia was defined as haemoglobin <12 g dl−1 in women or <13 g dl−1 in men. Multivariable regression models assessed associations between preoperative anaemia severity and postoperative outcomes, including a primary outcome of postoperative acute kidney injury (AKI), with interactions evaluated by gender.
Results.
Of 228,341 procedures (49.4% among women) included, AKI occurred in 13.2% overall, with a lower rate in women (7.6% in women vs. 18.8% in men). Preoperative anemia was present in 70,968 (31.1%) patients, including 30.8% of women and 31.4% of men. In adjusted analyses, increasing severity of preoperative anaemia was associated with worse outcomes, with consistently larger effect estimates in women. When comparing severe preoperative anaemia to no anaemia, the adjusted odds ratio (OR) for AKI was 2.91 (95% CI: 2.47, 3.44) in women vs. 1.43 (1.22, 1.68) in men. Corresponding ORs were 2.48 (2.14, 2.88) vs. 1.62 (1.40, 1.86) for ischaemic events, and 4.78 (3.94, 5.80) vs. 2.62 (2.13, 3.23) for surgical site infection. Interaction testing confirmed stronger associations in women for AKI, ischaemic events, and surgical site infection (all interaction p<0.001).
Conclusions.
Preoperative anaemia is associated with adverse outcomes in both women and men, though the associations with acute kidney injury, ischaemic events and surgical site infection are stronger in women.
Keywords: acute kidney injury, gender, haemoglobin, perioperative medicine, preoperative anaemia, transfusion
Anaemia is a deficiency in the quantity of healthy red blood cells necessary for tissue oxygen delivery. As red cell mass is difficult to measure in clinical practice, haemoglobin concentrations are employed as surrogates to identify anaemia in clinical practice. The World Health Organization (WHO) defines anaemia as haemoglobin concentration <12 g dl−1 in nonpregnant women and <13 g dl−1 in men.1,2 Anaemia is prevalent worldwide with an estimated prevalence of 25% across all age groups, making it a leading cause of global disability.3
Anaemia is particularly prevalent in patients undergoing surgery, with 20–40% of patients anaemic at the time of surgery.4,5 Women generally have lower haemoglobin concentrations and experience higher rates of preoperative anaemia than men.6 Preoperative anaemia is associated with increased risk of perioperative complications such as organ injury and hospital readmissions7–10 and higher risk of receiving allogeneic red blood cell (RBC) transfusion, which itself is associated with organ dysfunction, infection, and mortality.4,11–13 These effects are dose-dependent, with greater severity of anaemia associated with worse postoperative outcomes.14–18 Additionally, recent literature suggests that preoperative anaemia is differentially associated with adverse outcomes in women versus men undergoing cardiac surgery8, but this has yet to be confirmed in other surgery types.19 Furthermore, recent evidence suggests that women with borderline anaemia (12–13 g dl−1) are at heightened risk for adverse surgical outcomes, prompting calls to consider reassessment of anaemia diagnostic criteria for women.20,21
This study investigated whether severity of preoperative anaemia is differentially associated with adverse clinical outcomes in women and men undergoing cardiac and noncardiac surgery. We hypothesised that increasing severity of preoperative anaemia is associated with adverse outcomes in both women and men, but that anaemia-outcome associations would be most pronounced in women given their lower baseline haemoglobin concentrations, iron stores, and blood volumes.
Methods
This was an observational, retrospective cohort study conducted under approval of the Mayo Clinic Institutional Review Board with waived requirement for written informed consent. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines22 (S1) and the Sex and Gender Equity in Research (SAGER) guidelines23 (S2) were used in study design and conduct.
Study cohort
The cohort consisted of adults (≥18 yr) undergoing surgery between August 1, 2012 and July 31, 2022 at a large academic medical centre in the USA. Patients undergoing any surgical procedure that required postoperative hospital admission and who had a preoperative haemoglobin concentration obtained ≤90 days from surgery were eligible for inclusion. Exclusion criteria were: absent or nonbinary self-reported gender and those with previous denial of authorization to utilise their medical records for observational research. For the primary outcome of acute kidney injury (AKI), patients with chronic kidney disease were excluded.
The sample size was determined by the number of patients who met these criteria during the study timeframe. The study cohort was identified using the Acute Care DataMart (Mayo Clinic, Rochester, MN, USA), which is an institutional datamart containing patient demographics, laboratory information, transfusion exposures, medication administrations, surgical features, and postoperative complications.24 Additional information (i.e. chronic medical comorbidities) was extracted from the Advanced Cohort Explorer (Mayo Clinic, Rochester, MN, USA), another institutional database.25 Both databases are routinely validated by a dedicated team of data professionals to ensure accuracy of extracted data.26
Exposures
The primary exposure of interest was the severity of preoperative anaemia according to WHO definitions: nonanaemic: ≥12 g dl−1 in women and ≥13 g dl−1 in men; mild anaemia: ≥11 g dl−1 and <12 g dl−1 in women or ≥11 g dl−1 and <13 g dl−1 in men; moderate anaemia: ≥8 g dl−1 and <11 g dl−1; severe anaemia: <8 g dl−1.1 In a secondary analysis, a gender-neutral definition of anaemia using a haemoglobin of <13 g dl−1 in both women and men was applied which expanded the criteria for mild anaemia in women to include a haemoglobin ≥11 g dl−1 and <13 g dl−1. Additionally, we evaluated the effect of continuous haemoglobin concentrations per 1 g dl−1 below 14 g dl−1. The threshold of 14 g dl−1 was selected as the reference point for the continuous haemoglobin analysis based on visual assessment of the functional form for the relationship between haemoglobin and AKI. The preoperative haemoglobin concentration closest to surgery (in the 90 days prior) was used to classify severity of preoperative anaemia and for continuous haemoglobin analysis. Patient gender was obtained through patient self-report as documented in the electronic health record.
Outcomes
The primary outcome of interest was postoperative AKI. AKI was defined as a binary (yes/no) outcome using the Kidney Disease Improving Global Outcomes (KDIGO) creatinine criteria (S3).27 Secondary outcomes were surgical site infection (SSI), perioperative ischaemic event (composite rate of stroke and myocardial infarction [MI]), perioperative allogenic RBC transfusion, hospital length of stay (LOS), 90-day readmission rate, and 12-month mortality. Surgical site infection and perioperative ischaemic events diagnoses were identified via validated search queries through the Acute Care Data Mart. Briefly, SSI identification was based on clinical diagnoses within 30 days of surgery. Stroke, as a component of composite perioperative ischaemic events, was diagnosed clinically by postoperative neurology consultation notes. Myocardial infarction was defined as an elevated troponin with at least one of the following: echocardiographic evidence of new regional wall motion abnormalities or electrocardiographic changes consistent with ischaemia along with congruent documentation of myocardial infarction in a clinical note or angiographically-diagnosed coronary artery occlusion. Institutional guidelines for perioperative RBC transfusion have remained consistent since 2012, with indications including haemoglobin <7 g dl−1, haemoglobin <8 g dl−1 with signs and/or symptoms of anaemia, haemoglobin <8–10 g dl−1 in those with acute coronary syndromes, and at any haemoglobin concentration in those with rapid exsanguination and concern for inadequate tissue perfusion.
Statistical analyses
Patient and procedural characteristics as well as outcomes are presented as median (25th, 75th percentile) for continuous variables and frequency (percentage) for categorical variables according to gender and anaemia status. Gender-specific associations between anaemia severity and outcomes were assessed using multivariable generalised linear models with the anaemia-by-gender interaction term included in each model. Unique surgical procedures were considered the unit of analysis (i.e., patients can be included more than once in the analysis data), and all analyses were complete case (96.9% of observations had complete covariate information). Robust standard errors were estimated using the method of generalised estimating equations and assuming exchangeable correlation to account for multiple procedures per patient. Model results are reported as odds ratios for binary outcomes. Hospital LOS was log-transformed for analysis to satisfy distributional assumptions, with model results reported as ratio of geometric means. All estimates are reported with 95% confidence intervals. All models were adjusted for age, date of surgery, Charlson Comorbidity Index (each with 2-degrees of freedom restricted cubic splines), race/ethnicity, ASA physical status, diabetes mellitus, moderate/severe renal disease, chronic heart failure, preoperative creatinine, i.v. iron therapy and erythropoietin therapy in the 3 months prior to surgery, RBC transfusion in the 30 days prior to surgery, and type of surgery (i.e. cardiac, colorectal, general surgery, endocrine/metabolic, neurosurgery, orthopaedic, otolaryngology/oral maxillofacial surgery, plastic, spine, surgical oncology, thoracic, trauma, urology/gynaecology, vascular, other).
When assessing association between continuous haemoglobin and outcomes, estimates were reported per 1 g dl−1 decrease below 14 g dl−1. Overall event rates were summarised according to preoperative haemoglobin and gender using smoothed line plots. P-values <0.05 were considered statistically significant. All analyses were done using R version 4.4.1 (R Foundation for Statistical Computing, Vienna, Austria).
Results
A total of 228,341 procedures were included amongs 174,657 patients (Figure 1). Of all procedures, 112,850 (49.4%) and 115,491 (50.6%) were performed in women and men, respectively. Median age was 61 [48, 71] yr and 64 [54, 72] yr for women and men, respectively (Table 1). The most common surgery type among both women (27.3%) and men (24.7%) was orthopaedic surgery. Using WHO definitions, 34,715 (30.8%) women and 36,253 (31.4%) men were anaemic. Of patients with preoperative anaemia, a higher proportion of men had mild anaemia (44.7% of women; 59.3% of men), a higher proportion of women had moderate anaemia (51.3% of women; 36.7% of men), and an equal proportion (4.0%) had severe anaemia. When using a gender-neutral anaemia definition (haemoglobin <13 g dl−1), 53.3% of women were anaemic prior to surgery (S4). Following surgery, 84.2% of women and 76.3% of men were transferred to the general hospital ward while 15.8% of women and 23.7% of men required ICU admission.
Figure 1: Patient inclusion and exclusion flow diagram.
Preoperative haemoglobin concentration was the closest haemoglobin level to surgery obtained ≤90 days prior to surgery. Anaemia severity category was defined as no anaemia (haemoglobin ≥12 g dl−1 in women, ≥13 g dl−1 in men), mild anaemia (haemoglobin ≥11 g dl−1 and either <12 g dl−1 in women or <13 g dl−1 in men), moderate anaemia (haemoglobin ≥8 g dl−1 and <11 g dl−1), or severe anaemia (hemoglobin <8 g dl−1).
Table 1:
Patient and surgery characteristics according to gender and presence or absence of preoperative anaemia
| Men | Women | |||||
|---|---|---|---|---|---|---|
|
| ||||||
| Not anaemic (N=79,238) | Aneamic (<13 g dl−1; N=36,253) | All men (N=115,491) | Not anaemic (N=78,135) | Anaemic (<12 g dl−1; N=34,715) | All women (N=112,850) | |
|
| ||||||
| Age, yr | 63 (53, 71) | 67 (57, 75) | 64 (54, 72) | 61 (49, 71) | 62 (46, 73) | 61 (48, 71) |
| Preoperative Hb, g dl−1 | 14.7 (13.9, 15.4) | 11.4 (10.0, 12.3) | 14.0 (12.4, 15.0) | 13.4 (12.8, 14.1) | 10.8 (9.7, 11.4) | 12.8 (11.6, 13.7) |
| Anemia severity | ||||||
| None | (100%) | NA | 79238 (68.6%) | (100%) | NA | 78135 (69.2%) |
| Mild | NA | (59.3%) | 21491 (18.6%) | NA | (44.7%) | 15510 (13.7%) |
| Moderate | NA | (36.7%) | 13317 (11.5%) | NA | (51.3%) | 17810 (15.8%) |
| Severe | NA | (4.0%) | 1445 (1.3%) | NA | (4.0%) | 1395 (1.2%) |
| Creatinine, mg dl−1, n=221,335 | 1.0 (0.9, 1.2) | 1.1 (0.9, 1.4) | 1.0 (0.9, 1.2) | 0.8 (0.7, 1.0) | 0.8 (0.7, 1.1) | 0.8 (0.7, 1.0) |
| Non-Hispanic white | 73048 (92.2%) | 32865 (90.7%) | 105913 (91.7%) | 71852 (92.0%) | 30513 (87.9%) | 102365 (90.7%) |
| Diabetes mellitus | 2558 (3.2%) | 4053 (11.2%) | 6611 (5.7%) | 1784 (2.3%) | 2327 (6.7%) | 4111 (3.6%) |
| Renal disease | 5924 (7.5%) | 8164 (22.5%) | 14088 (12.2%) | 4363 (5.6%) | 5298 (15.3%) | 9661 (8.6%) |
| Chronic heart failure | 3921 (4.9%) | 5070 (14.0%) | 8991 (7.8%) | 3031 (3.9%) | 3313 (9.5%) | 6344 (5.6%) |
| Charlson Comorbidity Score | 4 (2, 5) | 6 (4, 8) | 4 (3, 6) | 4 (2, 5) | 4 (2, 7) | 4 (2, 6) |
| ASA physical status, n=225,136 | ||||||
| 1–2 | 35379 (45.2%) | 7375 (20.7%) | 42754 (37.5%) | 42442 (55.0%) | 11925 (35.0%) | 54367 (48.9%) |
| 3–4+ | 42879 (54.8%) | 28241 (79.3%) | 71120 (62.5%) | 34702 (45.0%) | 22193 (65.0%) | 56895 (51.1%) |
| Preoperative i.v. iron | 104 (0.1%) | 1073 (3.0%) | 1177 (1.0%) | 314 (0.4%) | 1107 (3.2%) | 1421 (1.3%) |
| Preoperative erythropoietin | 9 (<0.1%) | 272 (0.8%) | 281 (0.2%) | 19 (<0.1%) | 197 (0.6%) | 216 (0.2%) |
| Preoperative RBC transfusion | 127 (0.2%) | 3216 (8.9%) | 3343 (2.9%) | 120 (0.2%) | 2918 (8.4%) | 3038 (2.7%) |
| Surgery type | ||||||
| Orthopaedic | 20181 (25.5%) | 8363 (23.1%) | 28544 (24.7%) | 22575 (28.9%) | 8218 (23.7%) | 30793 (27.3%) |
| Urology/Gynaecology | 10769 (13.6%) | 3285 (9.1%) | 14054 (12.2%) | 13728 (17.6%) | 6614 (19.1%) | 20342 (18.0%) |
| General surgery | 9573 (12.1%) | 6088 (16.8%) | 15661 (13.6%) | 9658 (12.4%) | 5901 (17.0%) | 15559 (13.8%) |
| Cardiac surgery | 10233 (12.9%) | 3902 (10.8%) | 14135 (12.2%) | 5171 (6.6%) | 2162 (6.2%) | 7333 (6.5%) |
| Neurosurgery | 8969 (11.3%) | 2331 (6.4%) | 11300 (9.8%) | 7046 (9.0%) | 1745 (5.0%) | 8791 (7.8%) |
| Thoracic | 3687 (4.7%) | 2305 (6.4%) | 5992 (5.2%) | 3849 (4.9%) | 1433 (4.1%) | 5282 (4.7%) |
| ENT/OMS | 3733 (4.7%) | 1633 (4.5%) | 5366 (4.6%) | 2884 (3.7%) | 1021 (2.9%) | 3905 (3.5%) |
| Colorectal | 2593 (3.3%) | 1591 (4.4%) | 4184 (3.6%) | 2697 (3.5%) | 1517 (4.4%) | 4214 (3.7%) |
| Trauma | 1786 (2.3%) | 1519 (4.2%) | 3305 (2.9%) | 1936 (2.5%) | 1402 (4.0%) | 3338 (3.0%) |
| Other | 1655 (2.1%) | 1513 (4.2%) | 3168 (2.7%) | 1786 (2.3%) | 1392 (4.0%) | 3178 (2.8%) |
| Vascular | 2352 (3.0%) | 1780 (4.9%) | 4132 (3.6%) | 1224 (1.6%) | 968 (2.8%) | 2192 (1.9%) |
| Surgical oncology | 1453 (1.8%) | 1129 (3.1%) | 2582 (2.2%) | 2305 (3.0%) | 1295 (3.7%) | 3600 (3.2%) |
| Spine | 1491 (1.9%) | 345 (1.0%) | 1836 (1.6%) | 1164 (1.5%) | 280 (0.8%) | 1444 (1.3%) |
| Plastic | 472 (0.6%) | 367 (1.0%) | 839 (0.7%) | 1431 (1.8%) | 597 (1.7%) | 2028 (1.8%) |
| Endocrine/metabolic | 291 (0.4%) | 102 (0.3%) | 393 (0.3%) | 681 (0.9%) | 170 (0.5%) | 851 (0.8%) |
Anaemia defined as a haemoglobin concentration <12 g dl−1 in women or <13 g dl−1 in men. Values are median (25th, 75th percentile) for continuous variables and frequency (percentage) for categorical variables. Intravenous iron and erythropoietin within 3 months prior to surgery. RBC transfusion within 30 days prior to surgery.
ASA, American Society of Anesthesiologists; ENT/OMS, otolaryngology/oral and maxillofacial surgery; Hb, haemoglobin; RBC, red blood cell.
The incidence of adverse outcomes was generally greater in men than in women and increased with the severity of anaemia (Table 2). The primary outcome of AKI occurred in 3,615 (4.6%) women and 12,168 (15.4%) men without preoperative anaemia compared to 4,918 (14.2%) women and 9,550 (26.3%) men with preoperative anaemia. Similar differences were observed when using a gender-neutral anaemia definition (S4).
Table 2:
Clinical outcomes according to gender and presence or absence of preoperative anaemia
| Men | Women | |||||
|---|---|---|---|---|---|---|
|
| ||||||
| Not anaemic (N=79,238) | Anaemic (<13 g dl−1; N=36,253) | All men (N=115,491) | Not anaemic (N=78,135) | Anaemic (<12 g dl−1; N=34,715) | All women (N=112,850) | |
|
| ||||||
| Acute kidney injury | 12168 (15.4%) | 9550 (26.3%) | 21718 (18.8%) | 3615 (4.6%) | 4918 (14.2%) | 8533 (7.6%) |
| Ischaemic event | 9181 (11.6%) | 6996 (19.3%) | 16177 (14.0%) | 5165 (6.6%) | 4655 (13.4%) | 9820 (8.7%) |
| Surgical site infection | 1870 (2.4%) | 2647 (7.3%) | 4517 (3.9%) | 1549 (2.0%) | 2387 (6.9%) | 3936 (3.5%) |
| RBC transfusion | 4810 (6.1%) | 11078 (30.6%) | 15888 (13.8%) | 5761 (7.4%) | 10534 (30.3%) | 16295 (14.4%) |
| Postoperative length of stay, days | 2 (1, 5) | 4 (2, 8) | 3 (1, 6) | 2 (1, 4) | 3 (2, 7) | 3 (1, 5) |
| 90-day hospital readmission | 9167 (11.6%) | 9223 (25.4%) | 18390 (15.9%) | 8165 (10.4%) | 7412 (21.4%) | 15577 (13.8%) |
| 1-year mortality | 2377 (3.0%) | 5271 (14.5%) | 7648 (6.6%) | 2043 (2.6%) | 3830 (11.0%) | 5873 (5.2%) |
Values are median (25th, 75th percentile) for continuous variables and frequency (percentage) for categorical variables. Episodes of transfusion either during the surgical procedure or after the procedure while in the hospital are included. RBC, red blood cell.
There were statistically significant and clinically meaningful interactions by gender, such that women had more pronounced anaemia severity-outcome associations for AKI (p<0.001), ischaemic events (p<0.001), and SSI (p<0.001), while men had more pronounced associations between anaemia and RBC transfusion (p<0.001) (Figure 2). The adjusted OR of postoperative AKI was 1.75 [1.63–1.87] in women vs. 1.23 [1.18–1.29] in men for mild anaemia, 2.30 [2.16–2.45] in women vs. 1.32 [1.25–1.40] in men for moderate anaemia, and 2.91 [2.47–3.44] in women vs. 1.43 [1.22–1.68] in men for severe anaemia (unadjusted ORs provided in S5). There were no clinically significant differences in adjusted odds ratios between women and men for hospital LOS, 90-day readmission, or 12-month mortality (S6).
Figure 2: Adjusted odds of clinical outcomes according to gender and severity of preoperative anaemia.
Effect estimates and 95% confidence intervals for the given severity of preoperative anaemia compared to those without preoperative anaemia. Estimates are odds ratios for categorical outcomes and ratio of geometric means for postoperative length of stay. Estimates are from multivariable generalised linear models with robust standard errors from generalised estimating equations. Blue boxes, women; red boxes, men. Periop, perioperative; Postop, postoperative; RBC, red blood cell; SSI, surgical site infection.
Results were consistent employing a gender-neutral anaemia definition (S7) and when limited to noncardiac surgery (S8). In analyses employing continuous haemoglobin concentrations, women had greater increases in the odds of AKI (1.25 [1.24–1.27] vs. 1.08 [1.07–1.09]; p<0.001), ischaemic events (1.18 [1.17–1.20] vs 1.11 [1.10–1.12]; p<0.001), and SSI (1.43 [1.40–1.46] vs. 1.32 [1.30–1.35]; p<0.001) for each 1 g dl−1 decrease in haemoglobin below 14 g dl−1 compared to men (Figure 3, Table 3). Effect estimates for RBC transfusions and LOS were modestly higher in men, and there were no differences in the odds of readmission or mortality.
Figure 3: Incidence of transfusion and clinical outcomes according to gender and integer preoperative haemoglobin.
Percent with event was calculated according to gender and integer haemoglobin value. A loess smoother was used to generate the smooth curves. Blue lines, women; red lines, men. Postop, postoperative; RBCs, red blood cells; SSI, surgical site infection.
Table 3:
Estimated association between decreasing preoperative haemoglobin below 14 g dl−1 and outcomes
| Men Estimate (95% CI) | Women Estimate (95% CI) | Interaction p-value | |
|---|---|---|---|
|
| |||
| Acute kidney injury | 1.08 (1.07, 1.09) | 1.25 (1.24, 1.27) | <0.001 |
| Ischemic event | 1.11 (1.10, 1.12) | 1.18 (1.17, 1.20) | <0.001 |
| Surgical site infection | 1.32 (1.30, 1.35) | 1.43 (1.40, 1.46) | <0.001 |
| RBC transfusion | 1.89 (1.87, 1.91) | 1.81 (1.79, 1.83) | <0.001 |
| Postoperative length of stay, days | 1.14 (1.14, 1.15) | 1.12 (1.11, 1.12) | <0.001 |
| 90-day hospital readmit | 1.16 (1.14, 1.17) | 1.16 (1.15, 1.18) | 0.499 |
| 1-year mortality | 1.35 (1.33, 1.37) | 1.36 (1.34, 1.38) | 0.306 |
Results are from multivariable generalised linear models with robust standard errors estimated using the method of generalised estimating equations and assuming exchangeable correlation. Results are reported as odds ratios for categorical outcomes. Hospital LOS was log-transformed for analysis to satisfy distributional assumptions, with model results reported as ratio of geometric means. Estimates are per 1 g dl−1 decrease below 14 g dl−1 (i.e. estimates > 1 indicate worsening outcomes with decreasing preoperative haemoglobin). All models were adjusted for age, date of surgery, and Charlson Comorbidity Index (each with 2-degrees of freedom restricted cubic splines), race/ethnicity, ASA physical status, diabetes mellitus (with complications), moderate/severe renal disease, chronic heart failure, preoperative creatinine, i.v. iron therapy and erythropoietin therapy in the 3 months prior to surgery, RBC transfusion in the 30 days prior to surgery, and type of surgery. RBC, red blood cell.
Discussion
In this observational study of more than 228,000 patients, increasing severity of preoperative anaemia was associated with higher odds of adverse surgical outcomes in both women and men. However, the associations between anaemia severity and AKI, SSI, and postoperative ischaemic events were stronger in women than men. These relationships persisted when employing a gender-neutral definition of anaemia (haemoglobin <13 g dl−1) and when analysing haemoglobin continuously. While a stronger association between preoperative anaemia severity and RBC transfusion was observed for men than women, the strength of association was attenuated in analyses employing continuous haemoglobin concentrations.
This investigation contributes to the growing body of literature on the harms of preoperative anaemia, employing one of the largest study cohorts to date. At the highest level, our findings confirm associations between preoperative anaemia and postoperative organ dysfunction, infection, prolonged LOS, readmission, and mortality.7–10 These findings highlight that preoperative anaemia represents a clinically-accessible marker for adverse outcomes, which can potentially be modified by preoperative anaemia management.28–30
Previous work in cardiac surgery has shown that preoperative anaemia is more strongly associated with AKI in women than in men, but without significant differences in vascular complications (composite stroke, MI, or pulmonary embolism).8 In the present study, which included a much larger cohort and was not limited by surgery type, the association between preoperative anaemia and AKI in women was confirmed. Women also had stronger associations between preoperative anaemia and ischaemic events and SSI. These results are distinct from a previous study of noncardiac surgical patients in China, which noted no difference in the association between anaemia and a composite perioperative morbidity outcome by patient sex.19 This might be related to differences in outcome definitions and ascertainment, sample size, analytical approach, or study population. In a sensitivity analysis limiting our analysis to noncardiac surgery, women again had stronger associations between anaemia and adverse outcomes.
One plausible explanation for the observed differences in anaemia-outcome associations amongst women and men is that women, on average, have lower blood volumes, haemoglobin concentrations, and iron stores than men, yet remain vulnerable to similar or proportionally greater perioperative blood loss. This could result in a greater impact on tissue oxygen delivery or iron-deficiency related cellular dysfunction, which might translate to downstream clinical outcomes. For example, tissue oxygenation and iron availability are essential for immune function and wound healing.31–33 No matter the mechanism, these findings highlight the added importance of identifying and treating women with preoperative anaemia, while not overshadowing the importance of treating preoperative anaemia in general.
Prior studies have noted variability in the association between preoperative anaemia, sex and gender, and RBC transfusion.8,34,35 Some heterogeneity might be due in part to varying definitions of anaemia in women (i.e. haemoglobin <12 g dl−1 vs <13 g dl−1) and analytical differences in exposure definitions (e.g. severity vs presence of preoperative anaemia). Although we found that anaemia increased the risk of transfusion for all patients, this risk was modestly higher in men. The reasons for this are unclear. While we adjusted for surgery type in a broad sense, we did not account for the specific procedural details or indications for surgery. Additionally, men in our study had greater acute and chronic disease severity, which although adjusted for in our models, might have influenced transfusion behaviour.
There is substantial debate in how to define anaemia perioperatively. Anaemia is defined by WHO as a decrease in haemoglobin by more than 2 standard deviations from the average value in “healthy” adults (i.e. <12 g dl−1 in women and <13 g dl−1 in men).1,2,36,37 We and others find that defining preoperative anaemia as haemoglobin <12 g dl−1 misses many patients at increased risk of anaemia-associated complications. For example, there is increased risk of perioperative morbidity and transfusion utilisation in patients with preoperative haemoglobin <13 g dl−1, regardless of patient sex.4 Utilising a lower haemoglobin threshold to define anaemia in women compared to men could manifest as fewer women receiving perioperative anaemia management. Further, many women with iron deficiency have haemoglobin >12 g dl−1 and are thus classified as nonanaemic, yet these patients’ symptoms and outcomes improve when treated with iron.38 When coupled with our data demonstrating that perioperative complications increase as preoperative haemoglobin falls in both women and in men, adopting a gender neutral definition of anaemia (i.e. haemoglobin <13 g dl−1) might improve anaemia identification, treatment, and outcomes.5,39,40
This study has several limitations. Firstly, 16.4% of eligible patients were excluded given absent preoperative haemoglobin concentrations. This missingness is unlikely random and might reflect younger, healthier individuals or those undergoing lower-risk procedures for whom laboratory testing is not routinely obtained. As a result, our findings might be most applicable to older adults with greater comorbidity burden and undergoing higher-risk operations. We attempted to mitigate this potential selection bias by adjusting for age, Charlson Comorbidity Index, and surgery type, yet residual bias could persist.sSecondly, we employed patient self-reported gender rather than biological sex. While these often track together, future studies are needed to evaluate anaemia-related outcomes across the spectrum of gender and sex. Thirdly, aetiologies of anaemia were not evaluated and iron studies were unavailable for most patients, such that we were unable to address the associations between anaemia aetiology and outcomes. Fourthly, we employed the patient’s closest haemoglobin value prior to surgery to ascertain anaemia, which could have occurred any time within 90 days. However, this value might not represent the actual haemoglobin at the time of surgery for all patients. Finally, this was a single-centre study where data were pooled across surgical subspecialties which could limit generalisability of our findings.
In conclusion, the associations between preoperative anaemia and acute kidney injury, surgical site infection, and ischaemic events are stronger in women than in men. These findings underscore the importance of recognising and treating preoperative anaemia, especially in women. A gender-neutral definition of preoperative anaemia using a common haemoglobin threshold of <13 g dl−1 might better identify those at risk for adverse perioperative outcomes.
Supplementary Material
Online Supplementary File S1: Strengthening of Reporting of Observational Studies in Epidemiology (STROBE) guidelines
Online Supplementary File S2: Sex and Gender Equity in Research (SAGER) guidelines
Online Supplementary File S3: Kidney Disease Improving Global Outcomes (KDIGO) creatinine criteria for acute kidney injury (AKI)
Online Supplementary File S4: Incidence of outcomes by gender utilizing a gender-neutral definition of anaemia (haemoglobin <13 g dl −1)
Online Supplementary File S5: Unadjusted estimates for association between preoperative anaemia severity and clinical outcomes in women and men
Online Supplementary File S6: Adjusted estimates for association between anaemia severity and outcomes in women and men
Online Supplementary File S7: Adjusted estimates for association between anaemia severity and outcomes in women and men utilizing a gender-neutral definition of anaemia (haemoglobin (<13 g dl−1)
Editor’s key points.
While preoperative anaemia is associated with adverse surgical outcomes, it is unclear whether the impact is comparable between women and men.
In an observational study of over 228,000 patients, greater preoperative anaemia was associated with higher odds of adverse surgical outcomes in both women and men, with stronger associations in women than men for acute kidney injury, surgical site infections, and postoperative ischaemic events were stronger.
Anaemia-outcome associations might be more pronounced in women given their lower baseline haemoglobin concentrations, iron stores, and blood volumes.
Funding
US National Heart Lung and Blood Institute (NHLBI) [grant number K23HL153310] (MAW). The contents of this manuscript are solely the responsibility of the authors and do not necessarily represent the official views of the NIH.
Footnotes
Declaration of interests
MAW received an honorarium for speaking at World Anaemia Day 2025 sponsored by Pharmacosmos. The other authors have nothing to declare.
Presented in abstract form at the Network for the Advancement of Patient Blood Management, Haemostasis, and Thrombosis (NATA) 2025 annual meeting (date, location).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Guideline on haemoglobin cutoffs to determine anaemia in individuals and populations. Published online March 5, 2024. https://www.who.int/publications/i/item/9789240088542 [PubMed]
- 2.Nutritional anaemias. Report of a WHO scientific group. World Health Organ Tech Rep Ser. 1968;405:5–37. [PubMed] [Google Scholar]
- 3.GBD 2021 Anaemia Collaborators. Prevalence, years lived with disability, and trends in anaemia burden by severity and cause, 1990–2021: findings from the Global Burden of Disease Study 2021. Lancet Haematol. 2023;10:e713–e734. doi: 10.1016/S2352-3026(23)00160-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Lin Y Preoperative anemia-screening clinics. Hematol Am Soc Hematol Educ Program. 2019;2019:570–576. doi: 10.1182/hematology.2019000061 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Shander A, Corwin HL, Meier J, Auerbach M, Bisbe E, Blitz J, et al. Recommendations From the International Consensus Conference on Anemia Management in Surgical Patients (ICCAMS). Ann Surg. 2023;277:581–590. doi: 10.1097/SLA.0000000000005721 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Kulier A, Levin J, Moser R, Rumpold-Seitlinger G, Tudor IC, Snyder-Ramos SA, et al. Impact of preoperative anemia on outcome in patients undergoing coronary artery bypass graft surgery. Circulation. 2007;116:471–479. doi: 10.1161/CIRCULATIONAHA.106.653501 [DOI] [PubMed] [Google Scholar]
- 7.Fowler AJ, Ahmad T, Phull MK, Allard S, Gillies MA, Pearse RM. Meta-analysis of the association between preoperative anaemia and mortality after surgery. Br J Surg. 2015;102(11):1314–1324. doi: 10.1002/bjs.9861 [DOI] [PubMed] [Google Scholar]
- 8.Ripoll JG, Smith MM, Hanson AC, Schulte PJ, Portner ER, Kor DJ, et al. Sex-Specific Associations Between Preoperative Anemia and Postoperative Clinical Outcomes in Patients Undergoing Cardiac Surgery. Anesth Analg. 2021;132:1101–1111. doi: 10.1213/ANE.0000000000005392 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Hosseini M, Griffeth EM, Schaff HV, Klompas AM, Warner MA, Stulak JM, et al. Analysis of Anemia, Transfusions, and CABG Outcomes in The Society of Thoracic Surgeons National Database. Ann Thorac Surg. 2024;117:1053–1060. doi: 10.1016/j.athoracsur.2024.01.009 [DOI] [PubMed] [Google Scholar]
- 10.Musallam KM, Tamim HM, Richards T, Spahn DR, Rosendaal FR, Habbal A, et al. Preoperative anaemia and postoperative outcomes in non-cardiac surgery: a retrospective cohort study. Lancet Lond Engl. 2011;378:1396–1407. doi: 10.1016/S0140-6736(11)61381-0 [DOI] [PubMed] [Google Scholar]
- 11.Dakour-Aridi H, Giuliano K, Locham S, Dang T, Siracuse JJ, Malas MB. Perioperative blood transfusion in anemic patients undergoing elective endovascular abdominal aneurysm repair. J Vasc Surg. 2020;71:75–85. doi: 10.1016/j.jvs.2019.02.065 [DOI] [PubMed] [Google Scholar]
- 12.Obi AT, Park YJ, Bove P, Cuff R, Kazmers A, Gurm HS, et al. The association of perioperative transfusion with 30-day morbidity and mortality in patients undergoing major vascular surgery. J Vasc Surg. 2015;61:1000–1009.e1. doi: 10.1016/j.jvs.2014.10.106 [DOI] [PubMed] [Google Scholar]
- 13.Mazzeffi MA, Holmes SD, Taylor B, Ghoreishi M, McNeil JS, Kertai MD, et al. Red Blood Cell Transfusion and Postoperative Infection in Patients Having Coronary Artery Bypass Grafting Surgery: An Analysis of the Society of Thoracic Surgeons Adult Cardiac Surgery Database. Anesth Analg. 2022;135:558–566. doi: 10.1213/ANE.0000000000005920 [DOI] [PubMed] [Google Scholar]
- 14.Park LJ, Moloo H, Ramsay T, Thavorn K, Presseau J, Zwiep T, et al. Associations of preoperative anaemia with healthcare resource use and outcomes after colorectal surgery: a population-based cohort study. Br J Anaesth. 2024;133:58–66. doi: 10.1016/j.bja.2024.03.018 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Gabriel RA, Clark AI, Nguyen AP, Waterman RS, Schmidt UH. The Association of Preoperative Hematocrit and Transfusion with Mortality in Patients Undergoing Elective Non-cardiac Surgery. World J Surg. 2018;42:1939–1948. doi: 10.1007/s00268-017-4359-y [DOI] [PubMed] [Google Scholar]
- 16.Skorupski CP, Cheung MC, Hallet J, Kaliwal Y, Nguyen L, Pavenski K, et al. Preoperative Anemia and Iron Deficiency in Elective Gastrointestinal Cancer Surgery Patients. J Surg Oncol. 2025;131:614–623. doi: 10.1002/jso.27970 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Mendanha C, Neto PCS, Borges RB, Sirtoli I, Pando CS, Brandão M, et al. The burden of pre-operative anaemia and postoperative outcomes in 15 166 surgical patients from a public hospital in Brazil: A retrospective cohort study. Eur J Anaesthesiol. 2025;42:44–53. doi: 10.1097/EJA.0000000000002070 [DOI] [PubMed] [Google Scholar]
- 18.Wu WC, Schifftner TL, Henderson WG, Eaton CB, Poses RM, Uttley G, et al. Preoperative hematocrit levels and postoperative outcomes in older patients undergoing noncardiac surgery. JAMA. 2007;297:2481–2488. doi: 10.1001/jama.297.22.2481 [DOI] [PubMed] [Google Scholar]
- 19.Zhao BC, Lei SH, Liu JM, Qiu SD, Yao ZW, Liu J, et al. Sex-Specific Associations Between Preoperative Hemoglobin and Outcomes After Major Noncardiac Surgery: A Retrospective Cohort Study. Anesth Analg. 2023;137:1019–1028. doi: 10.1213/ANE.0000000000006661 [DOI] [PubMed] [Google Scholar]
- 20.Miles LF, Larsen T, Bailey MJ, Burbury KL, Story DA, Bellomo R. Borderline anaemia and postoperative outcome in women undergoing major abdominal surgery: a retrospective cohort study. Anaesthesia. 2020;75:210–217. doi: 10.1111/anae.14870 [DOI] [PubMed] [Google Scholar]
- 21.Netz A, Hof L, Rumpf F, Blum LV, Neef V, Kerner A, et al. Adjusting Current Hemoglobin Thresholds: A Way to Improve Outcome in Women Undergoing Major Surgery. J Womens Health 2002. 2024;33:678–684. doi: 10.1089/jwh.2023.0665 [DOI] [PubMed] [Google Scholar]
- 22.von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP, et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Bull World Health Organ. 2007;85:867–872. doi: 10.2471/blt.07.045120 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Heidari S, Babor TF, De Castro P, Tort S, Curno M. Sex and Gender Equity in Research: rationale for the SAGER guidelines and recommended use. Res Integr Peer Rev. 2016;1:2. doi: 10.1186/s41073-016-0007-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Herasevich V, Kor DJ, Li M, Pickering BW. ICU data mart: a non-iT approach. A team of clinicians, researchers and informatics personnel at the Mayo Clinic have taken a homegrown approach to building an ICU data mart. Healthc Inform Bus Mag Inf Commun Syst. 2011;28:42, 44–45. [PubMed] [Google Scholar]
- 25.Chute CG, Beck SA, Fisk TB, Mohr DN. The Enterprise Data Trust at Mayo Clinic: a semantically integrated warehouse of biomedical data. J Am Med Inform Assoc JAMIA. 2010;17:131–135. doi: 10.1136/jamia.2009.002691 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Singh B, Singh A, Ahmed A, Wilson GA, Pickering BW, Herasevich V, et al. Derivation and validation of automated electronic search strategies to extract Charlson comorbidities from electronic medical records. Mayo Clin Proc. 2012;87:817–824. doi: 10.1016/j.mayocp.2012.04.015 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Khwaja A KDIGO clinical practice guidelines for acute kidney injury. Nephron Clin Pract. 2012;120:c179–184. doi: 10.1159/000339789 [DOI] [PubMed] [Google Scholar]
- 28.Guinn NR, Fuller M, Murray S, Aronson S, Duke Perioperative Enhancement Team (POET). Treatment through a preoperative anemia clinic is associated with a reduction in perioperative red blood cell transfusion in patients undergoing orthopedic and gynecologic surgery. Transfusion (Paris). 2022;62:809–816. doi: 10.1111/trf.16847 [DOI] [PubMed] [Google Scholar]
- 29.Cahill CM, Alhasson B, Blumberg N, Melvin A, Knight P, Gloff M, et al. Preoperative anemia management program reduces blood transfusion in elective cardiac surgical patients, improving outcomes and decreasing hospital length of stay. Transfusion (Paris). 2021;61:2629–2636. doi: 10.1111/trf.16564 [DOI] [PubMed] [Google Scholar]
- 30.Trentino KM, Mace H, Symons K, Sanfilippo FM, Leahy MF, Farmer SL, et al. Associations of a Preoperative Anemia and Suboptimal Iron Stores Screening and Management Clinic in Colorectal Surgery With Hospital Cost, Reimbursement, and Length of Stay: A Net Cost Analysis. Anesth Analg. 2021;132:344–352. doi: 10.1213/ANE.0000000000005241 [DOI] [PubMed] [Google Scholar]
- 31.Kiani AA, Elyasi H, Ghoreyshi S, Nouri N, Safarzadeh A, Nafari A. Study on hypoxia-inducible factor and its roles in immune system. Immunol Med. 2021;44:223–236. doi: 10.1080/25785826.2021.1910187 [DOI] [PubMed] [Google Scholar]
- 32.Dickson KB, Zhou J. Role of reactive oxygen species and iron in host defense against infection. Front Biosci. 2020;25:1600–1616. doi: 10.2741/4869 [DOI] [PubMed] [Google Scholar]
- 33.Campfort M, Perrault T, Blanchard-Daguet A, Rineau E, Lasocki S. Associations between pre-operative iron deficiency and postoperative infection in patients undergoing major surgery (CARIPO): a prospective observational study. Anaesthesia. 2025;80:142–150. doi: 10.1111/anae.16498 [DOI] [PubMed] [Google Scholar]
- 34.Gombotz H, Schreier G, Neubauer S, Kastner P, Hofmann A. Gender disparities in red blood cell transfusion in elective surgery: a post hoc multicentre cohort study. BMJ Open. 2016;6:e012210. doi: 10.1136/bmjopen-2016-012210 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Kielty J, Hannon D, Smith A, McCarthy A, Avalos G, Dineen B, et al. The prevalence of pre-operative anaemia and an examination of its effect on transfusion practice between sexes: A multicentre retrospective study. Transfus Apher Sci Off J World Apher Assoc Off J Eur Soc Haemapheresis. 2021;60:103101. doi: 10.1016/j.transci.2021.103101 [DOI] [PubMed] [Google Scholar]
- 36.Kujovich JL. Evaluation of Anemia. Obstet Gynecol Clin North Am. 2016;43:247–264. doi: 10.1016/j.ogc.2016.01.009 [DOI] [PubMed] [Google Scholar]
- 37.Beutler E, Waalen J. The definition of anemia: what is the lower limit of normal of the blood hemoglobin concentration? Blood. 2006;107:1747–1750. doi: 10.1182/blood-2005-07-3046 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Weyand AC, Chaitoff A, Freed GL, Sholzberg M, Choi SW, McGann PT. Prevalence of Iron Deficiency and Iron-Deficiency Anemia in US Females Aged 12–21 Years, 2003–2020. JAMA. 2023;329:2191–2193. doi: 10.1001/jama.2023.8020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Muñoz M, Acheson AG, Auerbach M, Besser M, Habler O, Kehlet H, et al. International consensus statement on the peri-operative management of anaemia and iron deficiency. Anaesthesia. 2017;72(2):233–247. doi: 10.1111/anae.13773 [DOI] [PubMed] [Google Scholar]
- 40.NAC Position Statement on Patient Blood Management (Version 1.9). Published online July 7, 2022. Accessed March 19, 2025. https://nacblood.ca/sites/default/files/2022-07/NAC%20Patient%20Blood%20Management%20Position%20Statement%202022.pdf [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Online Supplementary File S1: Strengthening of Reporting of Observational Studies in Epidemiology (STROBE) guidelines
Online Supplementary File S2: Sex and Gender Equity in Research (SAGER) guidelines
Online Supplementary File S3: Kidney Disease Improving Global Outcomes (KDIGO) creatinine criteria for acute kidney injury (AKI)
Online Supplementary File S4: Incidence of outcomes by gender utilizing a gender-neutral definition of anaemia (haemoglobin <13 g dl −1)
Online Supplementary File S5: Unadjusted estimates for association between preoperative anaemia severity and clinical outcomes in women and men
Online Supplementary File S6: Adjusted estimates for association between anaemia severity and outcomes in women and men
Online Supplementary File S7: Adjusted estimates for association between anaemia severity and outcomes in women and men utilizing a gender-neutral definition of anaemia (haemoglobin (<13 g dl−1)