Abstract
OBJECTIVE:
To describe risk of Clostridium difficile infection associated with clindamycin and acute kidney injury associated with gentamicin during delivery hospitalizations.
METHODS:
Women admitted for delivery from January 2006 to March 2015 were analyzed using an inpatient administrative database. Primary outcomes were C. difficile infection and acute kidney injury. Clostridium difficile infection was compared between women receiving clindamycin (with or without other antibiotics) and women receiving antibiotics other than clindamycin. Acute kidney injury was compared between women receiving gentamicin (with or without other antibiotics), women receiving antibiotics other than gentamicin, and women receiving no antibiotics. Unadjusted and adjusted log linear models analyzing the role of patient demographics, mode of delivery, and hospital-level characteristics were created evaluating risk for C. difficile infection and acute kidney injury with risk ratios (RR) and adjusted risk ratios with 95% confidence intervals (CI) as measures of association. A sensitivity analysis for gentamicin and acute kidney injury was performed restricted to women with preeclampsia.
RESULTS:
Of 5,657,523 women admitted for delivery hospitalization, 266,402 (4.7%) received clindamycin and 165,726 (2.9%) received gentamicin. Clostridium difficile infection was diagnosed in 0.04% of women receiving clindamycin. Compared to women receiving other antibiotics, clindamycin was associated with a nearly three-fold increased risk of C. difficile infection (RR 2.93, 95% CI 2.36, 3.65). Acute kidney injury was diagnosed in 0.24% of women receiving gentamicin. Gentamicin was associated with a three-fold risk of acute kidney injury (RR 3.01, 95% CI 2.71, 3.34) compared to women receiving other antibiotics, while receipt of no antibiotics was associated with significantly lower risk (RR 0.18, 95% CI 0.15, 0.20). In adjusted analyses, these associations retained significance. Significantly increased risk for acute kidney injury was noted with women with preeclampsia receiving gentamicin (RR 2.04, 95% CI 1.64, 2.53).
CONCLUSION:
Receipt of clindamycin was associated with significantly increased likelihood for C. difficile infection and receipt of gentamicin with significantly increased likelihood of acute kidney injury, although the absolute risk for these complications was low.
Précis
Clindamycin and gentamicin were associated, respectively, with significantly increased relative risk, but low absolute risk of Clostridium difficile infection and nephrotoxicity during delivery hospitalizations.
INTRODUCTION
Postpartum endometritis is a common complication after delivery. Postpartum infections including endometritis affect up to 1-5% of all vaginal deliveries1 and are increased several fold after cesarean particularly among who labor prior to delivery.2 Broad-spectrum intravenous antibiotics are indicated for the treatment of postpartum endometritis, a polymicrobial infection. While most cases resolve with short courses of therapy, more serious infections and significant maternal morbidity can occur. Infection and sepsis remain the number three cause of maternal death in the United States accounting for 12.7% of deaths from 2011 to 2013.3
Several antibiotic regimens are used for postpartum endometritis in the absence of clear recommendations for specific first-line therapy. A 2015 meta-analysis of antibiotic therapy for postpartum endometritis found that clindamycin and gentamicin combination therapy was associated with fewer treatment failures and wound infections compared to treatment with cephalosporins or penicillins.4 Despite limited data, the authors also found no differences in side effects or severe complications in any of the antibiotic regimens studied.
Clindamycin, a lincosamide antibiotic active against gram positive bacteria and anaerobic organisms, has long been known to alter the intestinal microbiome and increase the risk of Clostridium difficile infections.5,6 Inpatient clindamycin use has been linked to a number of nosocomial outbreaks of C. difficile infection with resistant strains.7 Clostridium difficile infection is a serious life-threatening complication and a leading cause of health care-associated infections.7,8 It is responsible for close to half a million infections annually, is more common among females, and resulted in an estimated 29,000 deaths in 2011.9 In turn, gentamicin, an aminoglycoside antibiotic most effective against aerobic gram negative bacteria, has significant nephrotoxic and neurotoxic effects which limit its use. Rates of acute kidney injury associated with gentamicin vary by patient population, gentamicin dosing, and definition of nephrotoxicity; however, estimates from a meta-analysis comparing gentamicin dosing show a rate of 5-7%.10
Given the lack of data regarding the risk associated with these commonly-used antibiotics in the obstetric population, we sought to describe the risk of C. difficile infection associated with clindamycin and the risk acute kidney injury associated with gentamicin during delivery hospitalizations.
METHODS
The Perspective database was used to analyze antibiotic administration in this serial cross-sectional analysis. The Premier Perspective database is an administrative inpatient database which reports on 100% of hospitalizations for 600 individual hospitals and ambulatory surgery centers across the United States and includes approximately 15% of hospitalizations nationally. Patient demographics, hospital characteristics, medications and devices received during hospitalizations as well as other information is included in the database. Complete patient billing, hospital cost, and coding histories from are contained in the database. The Perspective database is commercially available and is commonly used across medical specialties to evaluate inpatient medication use.11 The Perspective database is maintained by Premier Incorporated (Charlotte, NC). Upon receiving data from participating hospitals, Premier undertakes an extensive 7-part data validation and correction process that includes more than 95 quality assurance checks prior to being used for research.12 After validations are complete, the data are moved to the Perspective data warehouse to populate and maintain the databases for health services research.13 Hospitals are geographically diverse within the United States and include both teaching and non-teaching institutions. The Columbia University Institutional Review Board deemed the study exempt given the data was de-identified.
For this analysis deliveries among women 18 to 54 years of age from January 2006 through March 2015 were included. Delivery hospitalizations were identified based on International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) billing codes using an approach that ascertains more than 95% of deliveries (ICD-9-CM codes 650 and V27.x).14 The primary objectives of this study were to determine: (i) risk for C. difficile infection associated with receipt of clindamycin, and (ii) risk for acute kidney injury associated with receipt of gentamicin during delivery hospitalizations. Clostridium difficile infection (ICD-9-CM diagnosis code 008.45) and acute kidney injury (ICD-9-CM diagnosis codes 584.5, 584.6, 584.7, 584.8, 584.9, 669.3, 669.30, 669.32, 669.34) were identified by ICD-9-CM diagnosis codes.
Perspective was queried for antibiotics that are commonly used during delivery hospitalizations. Use of penicillins, first through fourth generation cephalosporins, and other drug classes was ascertained (Appendix 1). Both generic and trade names were queried. For the analysis for acute kidney injury, patients were placed in three mutually exclusive categories based upon receipt of gentamicin and other antibiotics: (i) ≥1 antibiotics received including gentamicin, (ii) ≥1 antibiotics received excluding gentamicin, and (iii) no antibiotics. For the analysis for C. difficile infection, patients were placed in three mutually exclusive categories based upon receipt of clindamycin and other antibiotics: (i) ≥1 antibiotics including clindamycin, (ii) ≥1 antibiotics excluding clindamycin, and (iii) no antibiotics.
Whether hospital factors, demographic characteristics, comorbidity, and mode of delivery were associated with C. difficile infection and acute kidney injury was analyzed. Hospital characteristics included location (urban versus rural), teaching status (teaching versus nonteaching), geographic region (Midwest, Northeast, South, West), and hospital bed size (<400, 400 to 600, >600). Demographic characteristics included maternal age at delivery (<20, 20-24, 25-34, 35-39, ≥40 years), maternal race (white, black, other, or unknown), marital status (married, single, other or unknown), year of delivery (2006 to 2015), and insurance status (commercial, Medicare, Medicaid, uninsured, and unknown). Mode of delivery was dichotomized as vaginal versus cesarean delivery. Underlying maternal risk was assessed with an obstetric comorbidity index.15 This comorbidity index provided weighted scores for comorbidity for individual patients based on the presence of specific diagnosis codes and demographic factors present in administrative data. Higher scores are associated with increased risk for severe morbidity. In the initial study validating the comorbidity index in a general obstetric population, patients with the lowest score of 0 had a 0.68% risk of severe morbidity, whereas a score of >10 was associated with a risk of severe morbidity of 10.9%. This comorbidity index was subsequently validated in an external population.16 We categorized women based on these comorbidity index scores: 0 (lowest risk), 1, 2, and >2 (highest risk). Because maternal age is presented separately in our analysis, age was omitted in calculating the comorbidity index score. As preeclampsia is presented separately in our analysis, this factor was also omitted in calculating the comorbidity index score.
Individual models were performed separately for acute kidney injury and C. difficile infection. Univariable associations with these outcomes are presented as unadjusted risk ratios (RR) with 95% confidence intervals (CI) as measures of association. For each of the two outcomes, we fit multivariable log-linear regression models with Poisson distribution and log link based on generalized estimating equation including demographic factors, hospital characteristics, comorbidity, mode of delivery, and antibiotic administration with adjusted risk ratios (aRR) with 95% CI as measures of association. All analyses were performed with SAS 9.4 (SAS Institute, Cary, NC).
Three sensitivity analyses were performed for this study. First, erythromycin is rarely indicated as a first-line medication for inpatient obstetric scenarios other than for women with preterm premature rupture of membranes (PPROM). Neonatal erythromycin eye ointment is routinely administered shortly after birth, a practice supported by the United States Preventive Services Task Force.18 Given that there is a possibility for misclassification of erythromycin administration appearing in maternal as opposed to neonatal drug files when administered to newborns, we performed sensitivity analysis repeating analyses for both outcomes excluding (i) erythromycin as an antibiotic criteria, (ii) with PPROM.
As a second sensitivity analysis the adjusted model for C. difficile infection was repeated. While antibiotics are a risk factor for C. difficile infection, three antibiotics are used specifically to treat C. difficile infection: metronidazole, vancomycin, and fidamoxicin. We excluded these antibiotics as criteria for antibiotic receipt and repeated adjusted and unadjusted analyses. Finally, as a third sensitivity analysis, we restricted the cohort to women with preeclampsia, as this is a common risk factor for acute kidney injury among pregnant women, and repeated the unadjusted analysis for gentamicin.
RESULTS
After exclusions, 5,657,523 delivery hospitalizations were included in the analysis. A total of 507 deliveries were complicated by C. difficile infection (Table 1). Among delivery hospitalizations, 266,402 women (4.7%) received clindamycin with or without other antibiotics and of these 103 (0.04%) had C. difficile infection. Receipt of clindamycin with or without other antibiotics compared to other antibiotics was associated with increased risk for C. difficile infection in both unadjusted (RR 2.93, 95% CI 2.36, 3.65) and adjusted (aRR 2.48, 95% CI 1.99, 3.08) analyses. C. difficile infection without documentation of antibiotic administration was rare in unadjusted (RR 0.04, 95% CI 0.02, 0.07) and adjusted analyses (aRR 0.06, 95% CI 0.06, 95% CI 0.04, 0.11). Other factors associated with C. difficile infection in both adjusted and unadjusted analyses included Medicare insurance, cesarean delivery, and preeclampsia. Clostridium difficile infection was also significantly more likely at hospitals in the Northeast and in large compared to small bed size hospitals.
Table 1.
Demographics and adjusted and unadjusted models for Clostridium difficile infections
| Demographics | Models | |||||
|---|---|---|---|---|---|---|
| Absent | Present | Unadjusted | Adjusted | |||
| N | % | N | % | RR (99% CI) | aRR (99% CI) | |
| Antibiotics | ||||||
| No antibiotics | 2422404 | 42.8 | 13 | 2.6 | 0.04 (0.02, 0.08) | 0.07 (0.03, 0.14) |
| Clindamycin (+/− other antibiotics) | 266,402 | 4.7 | 103 | 20.3 | 2.93 (2.21, 3.90) | 2.52 (1.89, 3.36) |
| Antibiotics excluding clindamycin | 2968210 | 52.5 | 391 | 77.1 | Referent | Referent |
| Year | ||||||
| 2006-2008 | 1649233 | 29.2 | 154 | 30.4 | Referent | Referent |
| 2009-2011 | 1812294 | 32 | 144 | 28.4 | 0.85 (0.63, 1.15) | 0.83 (0.62, 1.12) |
| 2012-2015 | 2195489 | 38.8 | 209 | 41.2 | 1.02 (0.78, 1.34) | 1.00 (0.76, 1.32) |
| Age in years | ||||||
| 18-24 | 1815371 | 32.1 | 139 | 27.4 | Referent | Referent |
| 25-34 | 3006120 | 53.1 | 264 | 52.1 | 1.15 (0.88, 1.50) | 1.08 (0.81, 1.45) |
| 35-54 | 835,525 | 14.8 | 104 | 20.5 | 1.63 (1.16, 2.27) | 1.25 (0.87, 1.80) |
| Marital Status | ||||||
| Married | 2767616 | 48.9 | 228 | 45.0 | Referent | Referent |
| Single/Other/Unknown | 2889400 | 51.1 | 279 | 55.0 | 1.17 (0.93, 1.48) | 1.23 (0.94, 1.59) |
| Race | ||||||
| White | 3027480 | 53.5 | 263 | 51.9 | Referent | Referent |
| Black | 784,928 | 13.9 | 97 | 19.1 | 1.42 (1.05, 1.93) | 1.16 (0.84, 1.60) |
| Other | 1844608 | 32.6 | 147 | 29 | 0.92 (0.70, 1.20) | 0.97 (0.74, 1.28) |
| Payer | ||||||
| Medicare/Medicaid/Uninsured/unknown | 2,773,299 | 49 | 237 | 46.7 | 0.91 (0.73, 1.15) | 0.89 (0.69, 1.16) |
| Commercial | 2883717 | 51 | 270 | 53.3 | Referent | Referent |
| Preeclampsia | ||||||
| None | 5418069 | 95.8 | 433 | 85.4 | Referent | Referent |
| Any | 238,947 | 4.2 | 74 | 14.6 | 3.87 (2.80, 5.36) | 1.98 (1.42, 2.77) |
| Obstetric comorbidity | ||||||
| 0 | 5246613 | 92.7 | 380 | 75 | Referent | Referent |
| 1 or greater | 410,403 | 7.3 | 127 | 25 | 4.27 (3.28, 5.56) | 2.52 (1.91, 3.32) |
| Mode of delivery | ||||||
| Vaginal | 3758016 | 66.4 | 133 | 26.2 | Referent | Referent |
| Cesarean | 1899000 | 33.6 | 374 | 73.8 | 5.56 (4.29, 7.22) | 2.33 (1.78, 3.06) |
| Hospital Teaching | ||||||
| No | 3386054 | 59.9 | 249 | 49.1 | Referent | Referent |
| Yes | 2270962 | 40.1 | 258 | 50.9 | 1.54 (1.23, 1.94) | 1.31 (1.01, 1.69) |
| Hospital Bed Size | ||||||
| <400 | 3138243 | 55.5 | 226 | 44.6 | Referent | Referent |
| 400-600 | 1469084 | 26 | 143 | 28.2 | 1.35 (1.03, 1.78) | 1.18 (0.88, 1.59) |
| >600 | 1049689 | 18.6 | 138 | 27.2 | 1.83 (1.38, 2.41) | 1.41 (1.04, 1.92) |
RR, relative risk. aRR, adjusted relative risk. CI, confidence interval.
P<0.01;
P<0.001.
The adjusted model included all of the factors in the table. The number needed to harm was 3925 and is defined as the number of women who would have to receive clindamycin compared to other antibiotics to cause an additional case of CDI. Overall 1.2% of women were diagnosed with endometritis and 1.7% of women were diagnosed with chorioamnionitis in the study cohort.
When the analysis was restricted to antibiotics except for erythromycin, risk estimates were similar. Clindamycin was associated with increased risk for C. difficile infection compared to other antibiotics in adjusted and unadjusted analyses (RR 2.52, 95% CI 2.00, 3.16; aRR 2.27, 95% CI 1.81, 2.85) (Appendix 2). Many additional factors associated with increased risk for C. difficile infection remained the same as in the primary analysis. Finally, when the analysis was further restricted to exclude metronidazole, vancomycin, and fidaxomicin, the estimate for risk for C. difficile infection associated with clindamycin compared to other antibiotics retained significance in both unadjusted (RR 3.50, 95% CI 2.80, 4.37) and adjusted (aRR 2.88, 95% CI 2.31, 3.60) analyses.
Of 5,657,523 delivery hospitalizations, 3,215 were complicated by acute kidney injury. Of the study cohort, 165,726 women (2.9%) received gentamicin with or without other antibiotics and of these 402 (0.24%) had acute kidney injury. Receipt of gentamicin with or without other antibiotics was associated with increased risk for acute kidney injury in both unadjusted (RR 3.01, 95% CI 2.71, 3.34) and adjusted (aRR 2.48, 95% CI 2.23, 2.75) analyses in comparison to receipt of antibiotics other than gentamicin (Table 2). Patients not receiving antibiotics were at significantly lower risk of acute kidney injury (RR 0.18, 95% CI 0.16, 0.20; aRR 0.37, 95% CI 0.33, 0.42) compared to those receiving antibiotics other than gentamicin. Other factors associated with acute with acute kidney injury in both unadjusted and adjusted analyses included preeclampsia with and without severe features, superimposed preeclampsia, and cesarean delivery. Later study year, maternal age 40-54, and black compared to white race were also associated with increased risk.
Table 2.
Demographics and adjusted and unadjusted models for acute renal failure
| Demographics | Models | |||||
|---|---|---|---|---|---|---|
| Absent | Present | Unadjusted | Adjusted | |||
| N | % | N | % | RR (95% CI) | aRR (95% CI) | |
| Antibiotics | ||||||
| No antibiotics | 2422073 | 42.8 | 344 | 10.7 | 0.18 (0.15, 0.20)** | 0.37 (0.32, 0.44)** |
| Gentamicin (+/− other antibiotics) | 165,726 | 2.9 | 402 | 12.5 | 3.01 (2.62, 3.45)** | 2.49 (2.16, 2.86)** |
| Antibiotics excluding gentamicin | 3,066,509 | 54.2 | 2,469 | 76.8 | Referent | Referent |
| Year | ||||||
| 2006 | 543366 | 9.6 | 188 | 5.8 | Referent | Referent |
| 2007 | 561422 | 9.9 | 231 | 7.2 | 1.19 (0.92, 1.53) | 1.17 (0.91, 1.51) |
| 2008 | 543935 | 9.6 | 245 | 7.6 | 1.30 (1.01, 1.67)* | 1.24 (0.96, 1.59) |
| 2009 | 561777 | 9.9 | 318 | 9.9 | 1.64 (1.29, 2.07)** | 1.46 (1.15, 1.85)** |
| 2010 | 586147 | 10.4 | 340 | 10.6 | 1.68 (1.33, 2.12)** | 1.23 (0.97, 1.56) |
| 2011 | 663487 | 11.7 | 369 | 11.5 | 1.61 (1.28, 2.02)** | 1.40 (1.11, 1.77)** |
| 2012 | 717355 | 12.7 | 444 | 13.8 | 1.79 (1.43, 2.24)** | 1.50 (1.20, 1.88)** |
| 2013 | 705220 | 12.5 | 501 | 15.6 | 2.05 (1.65, 2.56)** | 1.69 (1.36, 2.11)** |
| 2014 | 639041 | 11.3 | 475 | 14.8 | 2.15 (1.72, 2.68)** | 1.73 (1.39, 2.17)** |
| 2015 1st quarter | 132558 | 2.3 | 104 | 3.2 | 2.27 (1.65, 3.11)** | 1.73 (1.26, 2.37)** |
| Age in years | ||||||
| 18-24 | 1814606 | 32.1 | 904 | 28.1 | Referent | Referent |
| 25-34 | 3004828 | 53.1 | 1,556 | 48.4 | 1.04 (0.93, 1.16) | 1.09 (0.97, 1.22) |
| 35-39 | 677,102 | 12 | 530 | 16.5 | 1.57 (1.36, 1.81)** | 1.33 (1.14, 1.55)** |
| ≥ 40 | 157,772 | 2.8 | 225 | 7 | 2.86 (2.36, 3.47)** | 1.76 (1.43, 2.15)** |
| Marital Status | Referent | |||||
| Married | 2766613 | 48.9 | 1,231 | 38.3 | Referent | Referent |
| Single | 2186165 | 38.7 | 1,480 | 46 | 1.52 (1.38, 1.68)** | 1.25 (1.11, 1.40)** |
| Other/Unknown | 701,530 | 12.4 | 504 | 15.7 | 1.61 (1.41, 1.85)** | 1.46 (1.26, 1.68)** |
| Race | ||||||
| White | 3026394 | 53.5 | 1,349 | 42 | Referent | Referent |
| Black | 784,165 | 13.9 | 860 | 26.7 | 2.46 (2.20, 2.75)** | 1.59 (1.41, 1.79)** |
| Other | 1843749 | 32.7 | 1,006 | 31.3 | 1.22 (1.10, 1.36)** | 1.20 (1.07, 1.35)* |
| Payer | ||||||
| Medicare/ Medicaid | 2,442,103 | 43.2 | 1524 | 47.4 | 1.19 (1.08, 1.30)* | 1.11 (1.00, 1.23)* |
| Commercial | 2882470 | 51 | 1,517 | 47.2 | Referent | Referent |
| Uninsured/unknown | 329,735 | 5.9 | 174 | 5.5 | 1.00 (0.82, 1.23) | 1.12 (0.91, 1.38) |
| Preeclampsia | ||||||
| None | 5416963 | 95.8 | 1,539 | 47.9 | Referent | Referent |
| With severe features | 80,073 | 1.4 | 1,021 | 31.8 | 44.33 (39.95, 49.19)** | 21.96 (19.68, 24.50)** |
| Without severe features | 126,035 | 2.2 | 296 | 9.2 | 8.25 (7.01, 9.71)** | 5.72 (4.85, 6.74)** |
| Superimposed | 31,237 | 0.6 | 359 | 11.2 | 40.00 (34.40, 46.52)** | 9.61 (8.09, 11.42)** |
| Obstetric comorbidity | ||||||
| 0 | 5245034 | 92.8 | 1,959 | 60.9 | Referent | Referent |
| 1 | 234,393 | 4.1 | 500 | 15.6 | 5.70 (5.01, 6.49)** | 2.24 (1.94, 2.58)** |
| 2 | 131,281 | 2.3 | 411 | 12.8 | 8.36 (7.27, 9.61)** | 3.00 (2.59, 3.47)** |
| >2 | 43,600 | 0.8 | 345 | 10.7 | 21.03 (18.09, 24.44)** | 6.77 (5.75, 7.96)** |
| Mode of delivery | ||||||
| Vaginal | 3757370 | 66.5 | 779 | 24.2 | Referent | Referent |
| Cesarean | 1896938 | 33.5 | 2,436 | 75.8 | 6.19 (5.56, 6.88)** | 2.30 (2.05, 2.59)** |
| Rurality | ||||||
| Urban | 5105016 | 90.3 | 3,042 | 94.6 | 1.89 (1.55, 2.31)** | 1.18 (0.96, 1.45) |
| Rural | 549,292 | 9.7 | 173 | 5.4 | Referent | Referent |
| Hospital Teaching | ||||||
| No | 3384962 | 59.9 | 1,341 | 41.7 | Referent | Referent |
| Yes | 2269346 | 40.1 | 1,874 | 58.3 | 2.08 (1.90, 2.28)** | 1.32 (1.18, 1.47)** |
| Hospital Bed Size | ||||||
| <400 | 3137163 | 55.5 | 1,306 | 40.6 | Referent | Referent |
| 400-600 | 1468175 | 26 | 1,052 | 32.7 | 1.72 (1.55, 1.91)** | 1.23 (1.10, 1.38)** |
| >600 | 1048970 | 18.6 | 857 | 26.7 | 1.96 (1.75, 2.20)** | 1.22 (1.07, 1.39)** |
| Region of Hospital | ||||||
| Northeastern | 904,628 | 16 | 656 | 20.4 | Referent | Referent |
| Midwest | 989,434 | 17.5 | 640 | 19.9 | 0.89 (0.77, 1.03) | 1.02 (0.88, 1.19) |
| South | 2581703 | 45.7 | 1,463 | 45.5 | 0.78 (0.69, 0.88)** | 0.80 (0.70, 0.91)** |
| West | 1178543 | 20.8 | 456 | 14.2 | 0.53 (0.46, 0.62)** | 0.80 (0.68, 0.95)* |
RR, relative risk. aRR, adjusted relative risk. CI, confidence interval.
P<0.01,
P<0.001.
The adjusted model included all of the factors in the table. The number needed to harm was 619 and is defined as the number of women needed to receive gentamicin (+/− other antibiotics) in order to lead to one more person being harmed compared to receiving other antibiotics.
When the analysis was restricted to antibiotics except for erythromycin, estimates for risk factors were similar. Gentamicin was associated with increased risk for C. difficile infection compared to other antibiotics in both unadjusted (aRR 2.61, 95% CI 2.34, 2.91) and adjusted analyses (aRR 2.33, 95% CI 2.09, 2.60) (Appendix 3). Other factors associated with increased risk that retained significance in the restricted analysis included later compared to earlier study year, preeclampsia, and cesarean delivery.
When the analysis was restricted to deliveries complicated by preeclampsia, gentamicin was associated with significantly increased risk for acute kidney injury (Table 3). Increased risk was noted comparing receipt of gentamicin to other antibiotics for all preeclampsia (1.6% versus 0.8%, RR 2.04, 95% CI 1.73, 2.40), for preeclampsia with severe features (2.8% versus 1.3%, RR 2.06, 95% CI 1.66, 2.55), for superimposed preeclampsia (2.3% versus 1.2%, RR 1.88, 95% CI 1.29, 2.74), and preeclampsia without severe features (0.8% versus 0.3%, RR 2.69 CI 95% 1.91, 3.78).
Table 3.
Risk for acute kidney injury among with preeclampsia based on receipt of antibiotics
| Acute Renal Failure Yes |
Acute Renal Failure No |
Unadjusted RR | Adjusted RR | NNH | |||
|---|---|---|---|---|---|---|---|
| N | row % | N | row % | (99%CI) | (99%CI) | ||
| Any Preeclampsia | |||||||
| Antibiotics† | |||||||
| No antibiotics | 205 | 0.3 | 66,608 | 99.7 | 0.38 (0.31, 0.46)** | 0.55 (0.45, 0.68)** | |
| Gentamicin (+/− other antibiotics) | 160 | 1.6 | 9,583 | 98.4 | 2.04 (1.64, 2.53)** | 2.00 (1.61, 2.48)** | 125.0 |
| Antibiotics excluding Gentamicin | 1,311 | 0.8 | 161,154 | 99.2 | Referent | Referent | Referent |
| Severe preeclampsia | |||||||
| Antibiotics† | |||||||
| No antibiotics | 128 | 0.7 | 17,712 | 99.3 | 0.54 (0.42, 0.69)** | 0.63 (0.49, 0.82)** | |
| Gentamicin (+/− other antibiotics) | 91 | 2.8 | 3,216 | 97.2 | 2.06 (1.55, 2.73)** | 2.01 (1.51, 2.68)** | 66.7 |
| Other antibiotics EXCLUDING Gentamicin | 802 | 1.3 | 59,145 | 98.7 | Referent | Referent | Referent |
| Superimposed preeclampsia | |||||||
| Antibiotics | |||||||
| No antibiotics | 44 | 0.6 | 7,017 | 99.4 | 0.51 (0.33, 0.77)** | ||
| Gentamicin (+/− other antibiotics) | 30 | 2.3 | 1,272 | 97.7 | 1.88 (1.15, 3.08)* | ||
| Other antibiotics EXCLUDING Gentamicin | 285 | 1.2 | 22,948 | 98.8 | Referent | ||
| Mild preeclampsia | |||||||
| Antibiotics | |||||||
| No antibiotics | 33 | 0.1 | 41,879 | 99.9 | 0.28 (0.17, 0.45)** | ||
| Gentamicin (+/− other antibiotics) | 39 | 0.8 | 5,095 | 99.2 | 2.69 (1.72, 4.20)** | ||
| Other antibiotics EXCLUDING Gentamicin | 224 | 0.3 | 79,061 | 99.7 | Referent | ||
RR, relative risk. aRR, adjusted relative risk. CI, confidence interval.
P<0.01,
P<0.0001.
NNH, Number needed to harm defined as the number of women needed to receive gentamicin (+/− other antibiotics) in order to lead to one more person being harmed compared to receiving other antibiotics.
The adjusted models included delivery mode and comorbidity index.
DISCUSSION
This analysis found that receipt of clindamycin was associated with significantly increased risk for C. difficile infection and receipt of gentamicin with significantly increased risk of acute kidney injury compared to both women receiving other antibiotics and those not receiving antibiotics. Overall the absolute risks of C. difficile infection and acute kidney injury associated with receipt of clindamycin and gentamicin, respectively, were significantly lower than those for hospitalized medical and surgical populations, and in both scenarios far less than 1% with the exception of women with preeclampsia. Even with preeclampsia both the attributable and absolute risks associated with gentamicin were modest; deliveries complicated by preeclampsia with severe features had an absolute risk for acute kidney injury 1.5% higher when gentamicin was administered while for superimposed preeclampsia the absolute risk was 1.1% higher.
Excluding women with preeclampsia, the risk of acute kidney injury associated with gentamicin in this analysis much lower than previously reported on other populations.10 One explanation for this is that women undergoing delivery hospitalizations are in general younger and healthier than other inpatients who may receive these broad-spectrum antibiotics. Furthermore, postpartum endometritis typically requires limited courses of antibiotic therapy, usually less than 72 hours. Duration and dosing of antibiotics were not available for analysis. This is particularly relevant to the use of gentamicin as some,10,19,20 but not all,21,22 evidence supports that both once-daily dosing and short courses in other hospitalized populations pose less risk of acute kidney injury; a meta-analysis of single versus multiple daily doses found that single doses were associated with an approximate one quarter reduction in risk for nephrotoxicity.10 The possibility also remains that the risk of C. difficile infection and acute kidney injury associated with these medications is underestimated in our analysis due to under-diagnosis or coding during the delivery hospitalization, as well as diagnosis after discharge. Of note, clindamycin has lasting effects on intestinal flora and less than one quarter of cases of health care-associated C. difficile infection are diagnosed during the index hospitalization.9 Therefore, it is possible that many women are diagnosed with C. difficile infection as an outpatient or at a later readmission after their delivery hospitalization.
Women who underwent cesarean delivery and those with preeclampsia and other co-morbidities were at greater risk of these complications. Notably, preeclampsia with severe features increased the risk of acute kidney injury over 20-fold in adjusted analyses with risk even higher with gentamicin administration in stratified analysis. In addition, women who were older, black compared to white, and single compared to married were also at increased risk of acute kidney injury. Risk of acute kidney injury also increased significantly over the study period from 2006 to 2015, a finding that has been previously reported in obstetric populations in Canada and specifically linked to women with preeclampsia.23 In the case of C. difficile infection, Northeast region and larger hospital size also increased risk. This may in part be explained by challenges inherent in the diagnosis of C. difficile infection and possible regional differences in its diagnosis, as well as hospital factors that affect the incidence of health care-associated infections.
These findings should overall be reassuring to obstetric providers who use these medications regularly in clinical practice. However, while the absolute risk during delivery hospitalization appears low, both clindamycin and gentamicin both have black box warnings issued by the Food and Drug Administration regarding risks of C. difficile infection and acute kidney injury, respectively. These risks should nonetheless be considered with each administration, as well as in the determination of dosing and duration of therapy.1,19 For some patients, such as those with multiple other risk factors such as preeclampsia, cesarean delivery, and underlying comorbid conditions that place them at highest risk for acute kidney injury and C. difficile infection, it may be reasonable to consider risks for these complications in deciding treatment regimens for postpartum endometritis; our analysis found that preeclampsia in particular may be associated with substantially increased risk. Antibiotic regimen, unlike most other factors associated with risk in our analysis, is a modifiable risk factor. Multiple albeit small studies have demonstrated similar efficacy of other regimens, particularly of ampicillin and sulbactum, in the treatment of postpartum endometritis.24-26 The potential for higher rates treatment failure and wound infection found in a recent meta-analysis with second and third generation cephalosporins4 should be balanced with the small but increased risk of C. difficile infection and acute kidney injury with clindamycin and gentamicin therapy.
There are several important limitations that should be considered in interpreting the findings of this study. First, this data set is cross sectional and does not include outpatient data; we cannot ascertain the timing of antibiotic administration relative to outcome diagnoses and cannot estimate the range of time at which acute kidney injury or C. difficile infection may develop after antibiotic administration. Furthermore, it is possible that in the setting of preeclampsia that acute kidney injury preceded antibiotic administration either during the hospitalization or on an outpatient basis. Second, it is also possible that antibiotic administration represents a confounding factor for underlying unmeasurable conditions associated with the outcomes analyzed. Third, the database does not have laboratory data to validate the diagnoses of acute kidney injury or C. difficile infection; ascertainment of these diagnoses may vary by health care provider, billing coder, institution, or region. Furthermore, it is not possible to ascertain what degree of renal insufficiency was present before or after antibiotic administration and perform stratified analyses. Fourth, this study used administrative data which has well known limitations. Billing data is often abstracted by non-clinical staff and concerns include both under-ascertainment and misclassification, particularly for secondary diagnoses. As a source of potential misclassification it is possible that acute diagnosis codes were used for conditions related to the patient’s history. Fifth, we are not able to perform chart reviews and accurately assess dose, timing, and duration of antibiotic administration. Risk for acute kidney injury may be higher with longer duration of gentamicin therapy; this database was not able to determine duration of therapy. Sixth, because acute kidney injury and C. difficile infection are known complications of gentamicin and clindamycin, respectively, it is possible that ascertainment could be biased by providers testing specifically for these conditions in the setting of administering these two medications. Finally, because antibiotics could be used for prophylaxis and treatment we are not able to make causal inferences related to prophylaxis.
Strengths of this analysis include the large, geographically diverse sample, the ability to perform sensitivity analyses restricting erythromycin and antibiotics used to treat C. difficile infection and to the preeclamptic subgroup, and the relatively long study time frame of almost ten years.
In summary, this study found that receipt of clindamycin was associated with significantly increased risk for CDI and receipt of gentamicin was associated with significantly increased risk of acute kidney injury. Given that absolute risk is low, likelihood of these complications may represent one consideration in antibiotic selection for obstetric patients.
Supplementary Material
Acknowledgments
Dr. Friedman is supported by a career development award (K08HD082287) from the Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health.
Footnotes
Presented at the 39th Annual Meeting for the Society for Maternal Fetal Medicine, February, 2019, Las Vegas, Nevada.
Each author has indicated that he or she has met the journal’s requirements for authorship.
Financial Disclosure
Dr. Wright has served as a consultant for Tesaro and Clovis Oncology. Dr. D’Alton has had a leadership role in ACOG II’s Safe Motherhood Initiative which has received funding from Merck for Mothers. The other authors did not report any potential conflicts of interest.
REFERENCES
- 1.Lamont RF, Sobel JD, Kusanovic JP, et al. Current debate on the use of antibiotic prophylaxis for caesarean section. BJOG 2011;118:193–201. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Hammad IA, Chauhan SP, Magann EF, Abuhamad AZ. Peripartum complications with cesarean delivery: a review of Maternal-Fetal Medicine Units Network publications. J Matern Fetal Neonatal Med 2014;27:463–74. [DOI] [PubMed] [Google Scholar]
- 3.Creanga AA, Syverson C, Seed K, Callaghan WM. Pregnancy-Related Mortality in the United States, 2011-2013. Obstet Gynecol 2017;130:366–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Mackeen AD, Packard RE, Ota E, Speer L. Antibiotic regimens for postpartum endometritis. Cochrane Database Syst Rev 2015:CD001067. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Bartlett JG, Chang TW, Gurwith M, Gorbach SL, Onderdonk AB. Antibiotic-associated pseudomembranous colitis due to toxin-producing clostridia. N Engl J Med 1978;298:531–4. [DOI] [PubMed] [Google Scholar]
- 6.Gerding DN, Johnson S, Peterson LR, Mulligan ME, Silva J Jr. Clostridium difficile-associated diarrhea and colitis. Infect Control Hosp Epidemiol 1995;16:459–77. [DOI] [PubMed] [Google Scholar]
- 7.Johnson S, Samore MH, Farrow KA, et al. Epidemics of diarrhea caused by a clindamycin-resistant strain of Clostridium difficile in four hospitals. N Engl J Med 1999;341:1645–51. [DOI] [PubMed] [Google Scholar]
- 8.Magill SS, O’Leary E, Janelle SJ, et al. Changes in Prevalence of Health Care-Associated Infections in U.S. Hospitals. N Engl J Med 2018;379:1732–44. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Lessa FC, Mu Y, Bamberg WM, et al. Burden of Clostridium difficile infection in the United States. N Engl J Med 2015;372:825–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Barza M, Ioannidis JP, Cappelleri JC, Lau J. Single or multiple daily doses of aminoglycosides: a meta-analysis. BMJ 1996;312:338–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Cozowicz C, Olson A, Poeran J, et al. Opioid prescription levels and postoperative outcomes in orthopedic surgery. Pain 2017;158:2422–30. [DOI] [PubMed] [Google Scholar]
- 12.Stulberg JJ, Delaney CP, Neuhauser DV, Aron DC, Fu P, Koroukian SM. Adherence to surgical care improvement project measures and the association with postoperative infections. Jama 2010;303:2479–85. [DOI] [PubMed] [Google Scholar]
- 13.Swanson SJ MB, Gunnarsson CL, Moore M, Howington JA, Maddaus MA, McKenna RJ, Miller DL. Video-assisted thoracoscopic lobectomy is less costly and morbid than open lobectomy: a retrospective multiinstitutional database analysis. Ann Thorac Surg 2012;93(4):1027–32. [DOI] [PubMed] [Google Scholar]
- 14.Kuklina EV, Whiteman MK, Hillis SD, et al. An enhanced method for identifying obstetric deliveries: implications for estimating maternal morbidity. Matern Child Health J 2008;12:469–77. [DOI] [PubMed] [Google Scholar]
- 15.Bateman BT, Mhyre JM, Hernandez-Diaz S, et al. Development of a comorbidity index for use in obstetric patients. Obstet Gynecol 2013;122:957–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Metcalfe A, Lix LM, Johnson JA, et al. Validation of an obstetric comorbidity index in an external population. BJOG 2015;122:1748–55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Liu W, Antonelli PJ, Dahm P, et al. Risk of sudden sensorineural hearing loss in adults using phosphodiesterase type 5 inhibitors: Population-based cohort study. Pharmacoepidemiol Drug Saf 2018;27:587–95. [DOI] [PubMed] [Google Scholar]
- 18.Force USPST Curry SJ, Krist AH, et al. Ocular Prophylaxis for Gonococcal Ophthalmia Neonatorum: US Preventive Services Task Force Reaffirmation Recommendation Statement. Jama 2019;321:394–8. [DOI] [PubMed] [Google Scholar]
- 19.Humes HD. Aminoglycoside nephrotoxicity. Kidney Int 1988;33:900–11. [DOI] [PubMed] [Google Scholar]
- 20.Carlsen S, Boel J, Jarlov JO, Gjorup I, Soborg C, Arpi M. The effect of short-course gentamicin therapy on kidney function in patients with bacteraemia-a retrospective cohort study. Eur J Clin Microbiol Infect Dis 2018;37:2307–12. [DOI] [PubMed] [Google Scholar]
- 21.Hatala R, Dinh T, Cook DJ. Once-daily aminoglycoside dosing in immunocompetent adults: a meta-analysis. Ann Intern Med 1996;124:717–25. [DOI] [PubMed] [Google Scholar]
- 22.Contopoulos-Ioannidis DG, Giotis ND, Baliatsa DV, Ioannidis JP. Extended-interval aminoglycoside administration for children: a meta-analysis. Pediatrics 2004;114:e111–8. [DOI] [PubMed] [Google Scholar]
- 23.Mehrabadi A, Liu S, Bartholomew S, et al. Hypertensive disorders of pregnancy and the recent increase in obstetric acute renal failure in Canada: population based retrospective cohort study. BMJ 2014;349:g4731. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Martens MG, Faro S, Hammill HA, Smith D, Riddle G, Maccato M. Ampicillin/sulbactam versus clindamycin in the treatment of postpartum endomyometritis. South Med J 1990;83:408–13. [DOI] [PubMed] [Google Scholar]
- 25.McGregor JA, Christensen FB. A comparison of ampicillin plus sulbactam versus clindamycin and gentamicin for treatment of postpartum infection. Suppl Int J Gynecol Obstet 1989;2:35–9. [DOI] [PubMed] [Google Scholar]
- 26.Stovall TG, Thorpe EM Jr., Ling FW. Treatment of post-cesarean section endometritis with ampicillin and sulbactam or clindamycin and gentamicin. J Reprod Med 1993;38:843–8. [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.