2026 年 8 巻 3 号 p. 83-92
BACKGROUND
This study aimed to evaluate the association of the initial body weight-stratified cefazolin antimicrobial prophylaxis (AMP) dose and surgical-site infection (SSI) in major orthopedic implant surgery.
METHODS
This nationwide, retrospective cohort study used data from the Diagnosis Procedure Combination database. We included patients (age ≥18 years) who underwent major orthopedic implant surgery between July 2010 and March 2022. Patients underwent surgery with an initial cefazolin dose of either 1 g (1 g cefazolin group) or 2 g (2 g cefazolin group). Primary outcome was in-hospital SSI. Restricted cubic spline functions were used to estimate the nonlinear association of body weight-stratified cefazolin dose with the outcome and between cefazolin dose per weight and the outcome.
RESULTS
Among the 408,487 participants, 394,731 and 13,756 were in the 1 g and 2g cefazolin groups, respectively. The 1 g cefazolin group, but not the 2 g cefazolin group, showed a J-shaped relationship between body weight and SSI, which increased continuously beyond 50 kg, in an increased risk of SSI with body weight. A reverse J-shaped relationship was observed between cefazolin dose per weight and SSI, with an increased risk of SSI at <20 mg/kg.
CONCLUSION
An increased risk of SSI was observed with 1 g cefazolin, but not with 2 g cefazolin, as an initial AMP with higher body weight among adults weighing >50 kg.
Surgical-site infections (SSI) are a major public health concern1), and continue to pose challenges in clinical practice. SSIs increase the length of stay, mortality, and morbidity2). In the United States, SSI-associated healthcare costs have been estimated at 3.5–10 billion, and confer a substantial socioeconomic impact3). Surgical antimicrobial prophylaxis (AMP) is widely utilized in orthopedic implant surgery4–6). Current SSI prevention guidelines issued by the Centers for Disease Control and Prevention and the World Health Organization recommend AMP during surgery, without any additional dose after wound closure7),8). The American Academy of Orthopaedic Surgeons and the International Consensus of Musculoskeletal Infections recommend first-generation cephalosporins, such as cefazolin, as the first choice for AMP9),10). To ensure adequate tissue concentrations, the up-titration of the initial dose based on the patient’s body weight is usually recommended3),11). However, the adjustment and evaluation of the body weight-based dose increase have been based primarily on previous pharmacokinetic studies11),12). Recent studies have highlighted the significance of body weight-associated increase in the initial cefazolin dose for SSI in patients undergoing orthopedic implant surgery13–15). Despite the lack of evidence, current guidelines recommend weight-based dose cutoffs for the initial cefazolin dose16),17). Consequently, further research is warranted to determine optimal weight-based cutoffs for the cefazolin dose to effectively prevent SSIs.
As is characteristic of antimicrobial agents, concerns exist regarding the adverse effects of cefazolin18–20). An increased risk of Clostridioides difficile (C. difficile) infection is associated with the prolonged use of cefazolin21); however, there is no evidence of the impact of the cefazolin dose on C. difficile infections. Furthermore, evidence of cefazolin-induced acute kidney injury in the field of orthopedic surgeries is lacking.
Thus, using a nationally representative inpatient database, we aimed to investigate the association of body weight with clinical outcomes (including SSI, all-cause mortality, acute kidney injury, and C. difficile infection) stratified by the initial prophylactic dose of cefazolin in patients undergoing major orthopedic implant surgery.
This retrospective cohort study used the Diagnosis Procedure Combination database, which is a nationally representative inpatient database in Japan that has been described elsewhere22). The Diagnosis Procedure Combination database contains discharge abstracts and administrative claims from voluntarily participating hospitals; these include more than 1,000 acute care hospitals and data for approximately 60% of all acute-care hospital beds. Patient-level data for all hospitalizations include: age; sex; smoking status; body weight; body height; residential region; level of consciousness at admission; location from which patients were admitted, such as private homes, other hospitals, and nursing homes or care facilities; main diagnoses, comorbidities present on admission, and complications arising after admission, which were recorded as textual data in Japanese by using the International Classification of Diseases, 10th Revision codes (ICD-10 codes); daily procedures recorded using Japanese procedure codes; surgical procedures recorded using Japanese procedure codes; daily drug administration; and total hospitalization costs. A previous validation study showed high specificity and moderate sensitivity for diagnoses and high specificity and sensitivity for procedures in this database23). This study was approved by the Institutional Review Board of the University of Tokyo [approval number: 3501-(5) on May 19, 2021]. The requirement for informed consent was waived because all data were de-identified.
PARTICIPANTSWe identified adults (age ≥18 years) who underwent orthopedic implant surgery between July 1, 2010, and March 31, 2022 and, using the Japanese original procedure codes presented in Supplemental Table 1, enrolled those who received cefazolin for AMP and underwent total shoulder, total hip, or total knee arthroplasty; or spinal fixation. We excluded patients: with comorbidities such as multi-trauma diagnoses, open fractures, pathological fractures, spinal cord injury, septic arthritis, osteomyelitis, spondylodiscitis, or spinal epidural abscess on admission; receiving maintenance dialysis, surgical AMP other than cefazolin, or preoperative antibiotics; transferred from other hospitals; admitted from nursing care; with outliers in height (<100 or >200 cm) or weight (<20 or >200 kg)24). The ICD-10 and Japanese procedural codes used in the exclusion criteria are presented in Supplemental Table 2.
EXPOSUREPatients who received a 1-g vial were assigned to the 1 g cefazolin group, while those who received a 2-g vial were assigned to the 2 g cefazolin group. To account for potential underestimation of actual dosing due to the use of multiple vials or intraoperative redosing, patients who received a total cefazolin dose of ≥6 g on the day of surgery were assigned to the 2 g cefazolin group.
OUTCOMES AND OTHER VARIABLESThe primary outcome was in-hospital SSI. The secondary outcomes included all-cause mortality, acute kidney injury and C. difficile infection. In-hospital SSI, acute kidney injury, and C. difficile infection were defined using the International Classification of Diseases, 10th Revision codes (Supplemental Table 3), which were recorded as post-admission complications. Although there is no validation for acute kidney injury, a prior Diagnosis Procedure Combination validation study demonstrated sensitivity and specificity values of 80% and 98%, respectively, for renal failure based on the Charlson Comorbidity Index23). Other variables included age; sex; body height; body weight; smoking status; each component of the Charlson Comorbidity Index25),26); mental and behavioral disorders (using ICD-10 codes, F00-99); atopic dermatitis (using ICD-10 code, L209); ossification of the posterior longitudinal ligament and ossification of the ligamentum flavum (using ICD-10 codes, M4882, M4884, M4885, M4886, M4889); type of hospital (academic hospital or not); surgical volume of hospital; duration from admission to surgery; type of procedure; anesthetic time; use of antiplatelet or anticoagulant agents, steroids, oral antihyperglycemics, oral corticosteroids, intravenous corticosteroids, immunosuppressant agents, chemotherapeutic agents, and nonsteroidal anti-inflammatory drugs; transfusion use; weekend surgery; and year of surgery. Body weight was subdivided into four categories: <50, 50–59.99, 60–69.99, and ≥70 kg. The surgical volume of the hospital was defined as the annual average of eligible orthopedic surgeries at each hospital, and these were divided into equal quarters as: very low, low, medium, and high volume.
STATISTICAL ANALYSISData are summarized as medians with interquartile ranges or means with standard deviations for continuous variables, and frequencies and proportions for categorical variables. Patient demographics are reported for the different dosing groups. In addition, histograms of height and weight distributions were generated to visually assess the impact of these exclusion criteria. The proportion of patients who received postoperative antibiotics, excluding the prophylactic agent (cefazolin 1 g or 2 g), was calculated for each initial cefazolin dose group (1 g and 2 g cefazolin groups).
First, to test the modification of the association between body weight and cefazolin dosing for each outcome in this study, we conducted multivariate logistic regressions for in-hospital SSI, in-hospital mortality, acute kidney injury, and C. difficile infection. The covariables included age; sex; body height; body weight; smoking status; the Charlson Comorbidity Index score; uncomplicated diabetes; complicated diabetes; renal failure; liver disease; mental and behavioral disorders; atopic dermatitis; type of hospital (academic hospital or not); surgical volume of hospital; duration from admission to surgery; type of procedure; anesthetic time; use of antiplatelet or anticoagulant agents, steroids, oral antihyperglycemics, oral corticosteroids, intravenous corticosteroids, immunosuppressant agents, chemotherapeutic agents, nonsteroidal anti-inflammatory drugs, and insulin; transfusion use; weekend surgery; and year of surgery, with an interaction term between body weight and cefazolin dose.
We then calculated the odds ratios and p-values for the interaction term between body weight and cefazolin dose for each outcome. We decided a priori to examine the nonlinear associations related to outcomes only if the modification of the associations was significant.
Second, we used a restricted cubic spline (RCS) regression model to examine the nonlinear association of body weight with outcomes, stratified by different dose groups. In contrast to the multivariable regression model, RCS can avoid loss of power and retain trends used in estimating association27). The knots were used at fixed spaced percentiles of body weight (5%, 35%, 65%, and 95%) to allow for nonlinear effects of continuous body weight measurement28). We again constituted an RCS regression model to examine the nonlinear association between cefazolin dose per weight (mg/kg) and outcomes. We then performed a joint Wald test of the restricted quadratic spline basis functions to assess nonlinearity. Additionally, we conducted the same analyses for the subgroup of patients who underwent total joint arthroplasty and spinal fixation, stratified by the cefazolin dose, to examine the possibility that the type of surgical procedure might affect the relationship between outcomes and body weight. In a subgroup analysis of patients who underwent spinal fixation, ossification of the posterior longitudinal ligament and ossification of the ligamentum flavum were included as covariates in the aforementioned model. Odds ratios and 95% confidence intervals (CI) were calculated for the outcomes.
For sensitivity analysis, to test whether the relationship between cefazolin dose and body weight was influenced by the severity of SSI, we restricted in-hospital SSIs to SSIs that required additional surgical interventions after the initial surgery. Furthermore, to test whether the relationship between cefazolin dose and body weight was influenced by the post-administration follow-up period, we examined 1-year SSI that required additional surgical interventions after the initial surgery. Additional surgical interventions were defined according to the original Japanese procedure codes presented in Supplemental Table 4. Following the modification of the outcome definition, we re-conducted the RCS regression model. All p-values were two-sided, with p < 0.05 indicating statistical significance. All statistical analyses were performed using Stata version 17 (StataCorp, College Station, Texas, United States).
After applying the inclusion and exclusion criteria, 408,487 participants were included in this study (Supplemental Fig. 1). The height of 2,670 patients (0.6%) fell outside the height limits; these patients were excluded. After exclusion of patients with a recorded height of zero, only 25 patients had non-zero biologically implausible height values (Supplemental Fig. 2A, 2B). Similarly, the weight of 2,233 patients (0.5%) fell outside the weight limits; these patients were also excluded. After exclusion of patients with a recorded weight of zero, only 11 patients had non-zero biologically implausible weight values (Supplemental Fig. 3A, 3B). Of the eligible participants, 394,731 and 13,756 were assigned to the 1 g and 2 g cefazolin groups, respectively.
Table 1 presents the characteristics of the participants. Those who were enrolled in the 1 g cefazolin group were older and had lower body height and weight compared with participants in the 2 g cefazolin group. Additionally, the 1 g cefazolin group comprised a lower proportion of smokers and was more likely to be treated in academic and low-volume hospitals than were the participants from the 2 g cefazolin group. In the 1 g cefazolin group, 50,777 of 394,731 patients (12.9 %) received at least one dose of antibiotics postoperatively, compared with 1,493 of 13,756 patients (10.9 %) in the 2 g cefazolin group.
| Characteristics | Total | 2 g Cefazolin | 1 g Cefazolin | SMD |
|---|---|---|---|---|
| Total patients, n | N = 408,487 | N = 13,756 | N = 394,731 | |
| Mean age, years (SD) | 69.7 (11.1) | 67.3 (12.0) | 69.8 (11.1) | −0.220 |
| Men sex, n (%) | 117,451 (28.8%) | 4,780 (34.7%) | 112,671 (28.5%) | 0.134 |
| Mean body height, cm (SD) | 154.9 (9.2) | 156.8 (10.0) | 154.8 (9.1) | 0.202 |
| Body weight, kg | ||||
| <49.99 | 79,714 (19.5) | 2,313 (16.8) | 77,401 (19.6) | −0.072 |
| 50–59.99 | 138,086 (33.8) | 4,085 (29.7) | 134,001 (34.0) | −0.091 |
| 60–69.99 | 110,508 (27.1) | 3,343 (24.3) | 107,165 (27.2) | −0.065 |
| ≥70 | 80,179 (19.6) | 4,015 (29.2) | 76,164 (19.3) | 0.232 |
| Mean BMI, kg/m2 (SD) | 24.7 (4.1) | 25.5 (4.9) | 24.7 (4.1) | 0.184 |
| Smoking status (smoker), n (%) | 111,943 (27.4%) | 4,735 (34.4%) | 107,208 (27.2%) | 0.158 |
| Type of surgery, n (%) | ||||
| Total hip arthroplasty | 148,840 (36.4%) | 5,286 (38.4%) | 143,554 (36.4%) | 0.043 |
| Total knee arthroplasty | 138,292 (33.9%) | 4,150 (30.2%) | 134,142 (34.0%) | −0.082 |
| Total shoulder arthroplasty | 7,674 (1.9%) | 189 (1.4%) | 7,485 (1.9%) | −0.041 |
| Anterior vertebral body fixation | 16,669 (4.1%) | 750 (5.5%) | 15,919 (4.0%) | 0.067 |
| Posterior lateral fixation | 26,657 (6.5%) | 998 (7.3%) | 25,659 (6.5%) | 0.030 |
| Posterior vertebral body fixation | 70,355 (17.2%) | 2,383 (17.3%) | 67,972 (17.2%) | 0.003 |
| Charlson Risk Index Score, n (%) | ||||
| 0 | 272,782 (66.8%) | 9,277 (67.4%) | 263,505 (66.8%) | 0.015 |
| 1 | 95,820 (23.5%) | 3,280 (23.8%) | 92,540 (23.4%) | 0.009 |
| 2 | 29,196 (7.1%) | 892 (6.5%) | 28,304 (7.2%) | −0.027 |
| 3 | 7,430 (1.8%) | 198 (1.4%) | 7,232 (1.8%) | −0.031 |
| ≥4 | 3,259 (0.8%) | 109 (0.8%) | 3,150 (0.8%) | −0.001 |
| Comorbidity, n (%) | ||||
| Uncomplicated diabetes | 61,656 (15.1%) | 2,155 (15.7%) | 59,501 (15.1%) | 0.016 |
| Complicated diabetes | 6,877 (1.7%) | 206 (1.5%) | 6,671 (1.7%) | −0.015 |
| Renal failure | 4,936 (1.2%) | 110 (0.8%) | 4,826 (1.2%) | −0.042 |
| Liver disease | 11,615 (2.8%) | 351 (2.6%) | 11,264 (2.9%) | −0.019 |
| Mental and behavioral disorders | 17,237 (4.2%) | 527 (3.8%) | 16,710 (4.2%) | −0.020 |
| Atopic dermatitis | 288 (0.1%) | 10 (0.1%) | 278 (0.1%) | 0.001 |
| Medications, n (%) | ||||
| Oral antihyperglycemic | 20,250 (5.0%) | 632 (4.6%) | 19,618 (5.0%) | −0.018 |
| Oral corticosteroid | 8,597 (2.1%) | 275 (2.0%) | 8,322 (2.1%) | −0.008 |
| Intravenous corticosteroid | 103,009 (25.2%) | 2,712 (19.7%) | 100,297 (25.4%) | −0.137 |
| Immunosuppressant | 3,887 (1.0%) | 98 (0.7%) | 3,789 (1.0%) | −0.027 |
| Chemotherapeutic agents | 448 (0.1%) | 21 (0.2%) | 427 (0.1%) | 0.012 |
| Anticoagulant | 21,914 (5.4%) | 671 (4.9%) | 21,243 (5.4%) | −0.023 |
| Antiplatelet | 11,191 (2.7%) | 205 (1.5%) | 10,986 (2.8%) | −0.089 |
| Gastroprotection | 38,753 (9.5%) | 1,192 (8.7%) | 37,561 (9.5%) | −0.030 |
| NSAIDs | 272,202 (66.6%) | 9,393 (68.3%) | 262,809 (66.6%) | 0.036 |
| Insulin | 13,900 (3.4%) | 563 (4.1%) | 13,337 (3.4%) | 0.038 |
| Mean anesthetic time, min (SD) | 206.1 (93.3) | 213.1 (86.7) | 205.8 (93.5) | 0.080 |
| Transfusion use, n (%) | 19,383 (4.7%) | 557 (4.0%) | 18,826 (4.8%) | −0.035 |
| Weekend surgery, n (%) | 1,656 (0.4%) | 12 (0.1%) | 1,644 (0.4%) | −0.066 |
| Academic hospital, n (%) | 316,456 (77.5%) | 10,016 (72.8%) | 306,440 (77.6%) | −0.112 |
| Mean duration from admission to surgery, days (SD) | 2.1 (1.9) | 2.3 (2.1) | 2.1 (1.8) | 0.096 |
| Surgical volume of hospital, n (%) | ||||
| Very low | 83,169 (20.4%) | 1,876 (13.6%) | 81,293 (20.6%) | −0.185 |
| Low | 100,778 (24.7%) | 1,672 (12.2%) | 99,106 (25.1%) | −0.337 |
| Medium | 111,768 (27.4%) | 6,484 (47.1%) | 105,284 (26.7%) | 0.434 |
| High | 112,772 (27.6%) | 3,724 (27.1%) | 109,048 (27.6%) | −0.012 |
BMI, body mass index; NSAIDs, non-steroidal anti-inflammatory drugs; SD, standard deviation; SMD, standardized mean difference
The incidence of SSI was 1.60% and 1.09% in the 1 g and 2 g cefazolin groups, respectively. After adjustment for baseline characteristics, the 2 g cefazolin group had a significantly lower risk of SSI than did the 1 g cefazolin group (p = 0.001) (Table 2). Furthermore, we detected an interaction of body weight and cefazolin dose with SSI, with p-values for body weights of 50.0–59.9 kg (p = 0.040), 60.0–69.9 kg (p = 0.398), and ≥70.0 kg (p = 0.006) (Supplemental Table 5). The analysis using RCS in the 1 g cefazolin group showed a J-shaped relationship between body weight and SSI, wherein the odds ratio remained flat until approximately 50 kg body weight and then increased continuously (Fig. 1A). In the 2 g cefazolin group, we found non-linearity between body weight and SSI (P for nonlinearity = 0.037; Fig. 1B), with the odds ratio rising above approximately 90 kg of body weight. A reverse J-shaped relationship was observed between cefazolin dose per weight and SSI, with the odds ratio for SSI decreasing to <20 mg/kg and then slightly declining thereafter (Fig. 2).
| Outcomes | Total, n = 408,487, n (%) |
1 g Cefazolin, n = 394,731, n (%) |
2 g Cefazolin, n = 13,756, n (%) |
P-value |
|---|---|---|---|---|
| Primary | ||||
| Surgical site infection | 6,460 (1.58) | 6,310 (1.60) | 150 (1.09) | 0.001* |
| Secondary | ||||
| In-hospital mortality | 300 (0.07) | 288 (0.07) | 12 (0.09) | 0.506* |
| Acute kidney injury | 139 (0.03) | 133 (0.03) | 6 (0.04) | 0.930* |
| Clostridioides difficile infections | 120 (0.03) | 118 (0.03) | 2 (0.01) | 0.820* |
* Multivariate logistic regression analysis
a: Cefazolin dose 1 g. b: Cefazolin dose 2 g. Odds ratios are indicated by thick solid lines and 95% CIs by shaded areas. Knots were placed at the 5th, 35th, 65th, and 95th percentiles of body weight.
Odds ratios are indicated by solid thick lines and 95% CIs by shaded areas. Knots were placed at the 5th, 35th, 65th, and 95th centiles of each cefazolin dose divided by the participant’s body weight.
After adjustment for baseline characteristics, no significant differences in secondary outcomes were observed between the two groups (Table 2). Furthermore, no interaction was found between body weight and cefazolin dose in patients with all-cause mortality, acute kidney injury, or C. difficile infection (Supplemental Table 5).
SUBGROUP ANALYSISFor those undergoing total joint arthroplasty, a J-shaped relationship between body weight and SSI was observed in the 2 g cefazolin group (P for nonlinearity <0.001; Fig. 3A). Conversely, for those undergoing spinal fixation, the odds ratio showed a U-shaped association between SSI and body weight, with the bottom of the spline weighing approximately 60 kg (P for nonlinearity <0.001; Fig. 3B). In the 2 g cefazolin group, nonlinearity in the association between body weight and SSI was not found in patients who underwent total joint arthroplasty (P = 0.139) or spinal fixation (P = 0.186) (Fig. 3C, 3D).
a: Total joint arthroplasty group with 1 g cefazolin dose. b: Spinal fixation group with 1 g cefazolin dose. c: Total joint arthroplasty group with 2 g cefazolin dose. d: Spinal fixation group with 2 g cefazolin dose. Odds ratios are indicated by solid thick lines and 95% CIs by shaded areas. Knots were placed at the 5th, 35th, 65th, and 95th centiles of each body weight.
Sensitivity analyses using different SSI definitions showed similar results for the primary outcomes (Supplemental Fig. 4A, 4B, 5A and 5B).
In this retrospective cohort study using a nationally representative inpatient database in Japan, the initial dose of cefazolin as an AMP appeared to have an impact on SSI. For those who weighed 50 kg or more, there was a J-shaped relationship between body weight and SSI when 1 g cefazolin was administered initially. Among patients receiving an initial 2 g dose of cefazolin, the odds ratios for SSI rose in those weighing approximately 90 kg or more. No interactions were found between body weight and the initial cefazolin dose in terms of in-hospital mortality, acute kidney injury, or C. difficile infection.
Our findings are consistent with those of previous studies that reported an association between an inadequate weight-based dose of cefazolin and an increased risk of SSI following major orthopedic implant surgery13),14). Specifically, these studies suggested that cefazolin doses of 1 g for patients weighing 60–120 kg and 2 g for those weighing >120 kg constituted underdosing. Additionally, our study provides a new insight into increased SSI risk among patients with a body weight >50 kg when administered 1 g cefazolin as their initial AMP. Moreover, the risk of SSI seemed to increase with cefazolin doses of 20 mg/kg or less. With an initial 2-g dose of cefazolin, the risk for SSI may increase at body weights of ≥90 kg or more; however, the number of patients weighing ≥90 kg or more was limited, which resulted in very wide confidence intervals and considerable imprecision. Our results suggest that there may be a reduced risk of SSI with a higher initial dose of cefazolin for those with a body weight of at least 50 kg or more or for those who may not achieve a concentration of 20 mg/kg when the dose is adjusted for weight. These findings are supported by previous pharmacokinetic studies, which showed that the widely accepted weight-based cutoffs for the initial cefazolin dose of AMP (60 or 80 kg for cefazolin 2 g, and 120 kg for cefazolin 3 g) failed to achieve sufficient serum concentrations of cefazolin11),29). Insufficient dosing may increase not only the risk of SSI but also the incidence of antimicrobial resistance30). Our results may guide clinicians and healthcare providers in determining an appropriate weight-based cutoff for initial cefazolin dosing.
The dose–response curve in the spine surgery subgroup demonstrated a U-shaped pattern, with both underweight (<50 kg) and heavier patients at increased SSI risk, as opposed to patients undergoing joint arthroplasty. An underweight status is associated with malnutrition and hypoalbuminemia, leading to poorer collagen synthesis and increased infection risk31). Moreover, underweight patients may have less subcutaneous tissue and muscle padding over spinal incisions in spine surgeries involving extensive posterior soft tissue dissection, and the resulting impairment in local immune function may further increase their susceptibility to wound dehiscence and SSI. These findings suggest that future cefazolin prophylaxis protocols should consider both high-weight and low-weight thresholds in spine surgery. Prospective studies are warranted to optimize prophylactic dosing across the full weight spectrum in spine surgery.
To the best of our knowledge, this study is the first to evaluate the association between body weight and in-hospital mortality in patients undergoing orthopedic implant surgery, stratified by the initial cefazolin dose. This study indicated that the cefazolin dose was not associated with short-term postoperative mortality.
The administration of multiple antibiotics and the duration of the antibiotic course are well-known risk factors for adverse events, including acute kidney injury and C. difficile infection20). However, evidence of the impact of the initial dose of prophylactic antibiotics on these adverse events is lacking. Our findings suggest that the initial dose of prophylactic cefazolin (1 g or 2 g) may not be associated with an increased risk of adverse events across body weights (Supplemental Table 5). However, this study did not include patients who may have received different kinds of antibiotics as dual AMP20),32). Further studies that include high-risk individuals are needed to validate the generalizability of our findings.
This study has several limitations. First, given that this study used a claims database, there is the possibility of bias from the misclassification of outcomes. To address this, we used a different definition of SSI for the sensitivity analysis, focusing on more severe SSI cases requiring additional surgery. The difference in SSI definitions did not change our findings, and the results resembled the main results. Second, we did not estimate the outcomes in patients with body weights >100 kg because the proportion of individuals with body weights >100 kg in Asian populations is generally lower than that in non-Asian populations. Third, the incidence of surgery-associated complications may have been underestimated because of the short follow-up period. The results might have been different if we had captured long-term follow-up data. Fourth, the Diagnosis Procedure Combination database captures only the cefazolin vial strength per recorded antibiotic product (e.g., 1 g or 2 g) and does not record the administration timing or additional doses, which are typically guided by patient characteristics, illness severity, and surgical factors1),33). Therefore, classification of exposure on the sole basis of the dispensed vial strength may lead to misclassification of cefazolin exposure. Moreover, although the total daily dose may inherently increase with an increase in the operative time, the Diagnosis Procedure Combination database cannot distinguish between the initial prophylactic dose and subsequent intraoperative or postoperative administrations. This limitation introduces potential misclassification of the true initial exposure. Therefore, prospective studies that record the exact timing and number of cefazolin administrations are warranted to accurately distinguish initial prophylaxis from redosing and to confirm our findings. Fifth, individual laboratory data, such as levels of blood glucose and renal function markers, were not available in this database. In particular, renal function is critically important for cefazolin pharmacokinetics and, consequently, its prophylactic efficacy. Prior pharmacokinetic studies indicate that cefazolin clearance markedly decreases with reduced creatinine clearance, resulting in an increase in the serum half-life from approximately 2 h in patients with normal renal function to 10–15 h and >40 h in patients with creatinine clearances of 15–30 mL/min and anephric patients, respectively34),35). Consequently, in patients with impaired renal clearance, a standard 1-g dose of cefazolin may maintain therapeutic free drug concentrations above the minimum inhibitory concentration for extended periods, potentially preserving its prophylactic efficacy.
We found a J-shaped relationship between body weight and SSI in patients receiving 1 g cefazolin as the initial dose of AMP, which continuously increased beyond 50 kg body weight. With an initial 2 g dose of cefazolin, SSI may increase at body weights approximately 90 kg or more. No interactions were found between body weight and cefazolin dose on other outcomes. These data suggest that, to prevent SSI, it may be beneficial to use an initial prophylactic dose of 2 g cefazolin as AMP in patients weighing more than 50 kg.
The institution of the authors received support from the Ministry of Health, Labour and Welfare, Japan (23AA2003 and 22AA2003). The funders played no role in the design and conduct of the study, the collection, management, analysis, and interpretation of the data, the preparation, review, or approval of the manuscript, or the decision to submit the manuscript for publication.
This work was supported by grants (23AA2003 and 22AA2003) from the Ministry of Health, Labour and Welfare, Japan.
None.
Masaki Hatano: Conceptualization; Methodology; Formal analysis; Writing - original draft.
Yusuke Sasabuchi: Conceptualization; Methodology; Supervision; Writing - review & editing.
Koji Yamada: Conceptualization; Supervision; Writing - review & editing.
Shotaro Aso: Conceptualization; Data curation; Writing - review & editing.
Hisatoshi Ishikura: Conceptualization; Supervision; Writing - review & editing.
Takeyuki Tanaka: Conceptualization; Supervision; Writing - review & editing.
Sakae Tanaka: Conceptualization; Supervision; Writing - review & editing.
Hideo Yasunaga: Conceptualization; Supervision; Funding acquisition; Project administration; Writing - review & editing; Resources.
Guarantor: Masaki Hatano
Yusuke Sasabuchi and Hideo Yasunaga are the Editorial Board members of Annals of Clinical Epidemiology (ACE). They were not involved in the peer-review or decision-making process for this paper.
