Severity of infection following the introduction of new infection control measures for medical abortion

Abstract

Background

In response to concerns about serious infections following medical abortion, in early 2006 Planned Parenthood changed the route of misoprostol administration from vaginal to buccal and required either routine antibiotic coverage or universal screening and treatment for chlamydia; in July 2007, Planned Parenthood began requiring routine antibiotic coverage for all medical abortions. We previously reported a pronounced drop in the rate of serious infection following the adoption of these new infection control measures. Our objective here is to assess whether the degree of severity of the serious infections differed in the three infection-control groups (vaginal misoprostol and no antibiotics, buccal misoprostol and screen-and-treat, buccal misoprostol and routine antibiotics) or, equivalently, to assess whether the declines in rates of serious infections after the adoption of new infection control measures differed across degree of severity categories. Of particular importance is whether the new infection control measures selectively reduced the least severe serious infections but did not diminish the rate of the most severe infections.

Methods

We performed a retrospective analysis assessing the degree of severity of infections before infection controls were implemented and after each of the two new measures was adopted: buccal administration of antibiotics with either screen-and-treat or routine antibiotic coverage. We ranked the severity of infection from 1 (when treatment occurred in an emergency department) to 4 (when death occurred). We compared the distribution of the severity of serious infections in the three infection control groups (none, buccal misoprostol and screen-and-treat, buccal misoprostol and routine antibiotics), or, equivalently, assessed whether the declines in rates of serious infections after the adoption of new infection control measures differed across degree of severity categories using the Jonckheere-Terpstra test for a doubly ordered 4×3 table.

Results

The distribution of infection by severity was the same for all three infection control groups. Likewise, when the two new infection control groups—buccal misoprostol plus either screen-and-treat or routine antibiotics—were combined, the distribution of infection by severity was the same before and after the new measures were implemented.

Conclusion

The pronounced decline in the rate of serious infections occurred in each category of severity.

1. Introduction

Prompted by the deaths following medical abortion and internal data showing a higher than expected serious infection rate, Planned Parenthood changed its medical abortion protocol at the end of March 2006. Vaginal administration of misoprostol delivery was discontinued and replaced by buccal administration (200 mg oral mifepristone followed 24-48 h later by 800 mcg of buccal misoprostol, 400 mcg in each cheek for 30 min), and all health centers were required to use one of two regimens intended to reduce the risk of infection: either routine antibiotic coverage (oral doxycycline 100 mg BID for 7 days, starting the same day as mifepristone administration) or universal testing for chlamydia (and for gonorrhea when considered appropriate), with treatment dependent on test results (oral doxycycline 100 mg BID for 7 days for chlamydia, and ceftriaxone 125 mg intramuscularly in a single dose for gonorrhea). After reviewing serious infection rates among health centers using these two different infection-reduction regimens, Planned Parenthood in July 2007 required all health centers to provide routine preventative antibiotic treatment. In a prior publication we showed that rates of serious infection dropped significantly after the joint change to buccal misoprostol replacing vaginal misoprostol and either sexually transmitted infection testing or routine antibiotic coverage as part of the medical abortion regimen (73% decline from 0.93/1000 to 0.25/1000, absolute reduction 0.67/1000 [95% CI 0.44-0.94], p<0.001) [1]. The subsequent change to routine antibiotic coverage led to a further significant reduction in the rate of serious infection (76% decline from 0.25/1000 to 0.06/1000, absolute reduction 0.19/1000 [95% CI 0.0.02-0.34], p=0.03) [1].

Altogether, there was a 93% reduction in the rate of serious infection following the switch to buccal misoprostol and routine antibiotic coverage (absolute decline 86/1000 [95% CI 0.64-1.12, p<0.001]); serious infection, originally rare, became extremely rare. Because two interventions were instituted at the same time—route of misoprostol changed from vaginal to buccal, and administration of routine antibiotics—it is impossible to know precisely to what extent either intervention contributed to the drop in the rate of serious infections following medical abortion.

In the group that initially adopted screen-and-treat, two-thirds of the decline occurred after the switch to buccal administration of misoprostol and screen-and-treat and one-third occurred after the subsequent switch to routine antibiotic coverage (the serious infection rate fell by 50% after the switch to routine antibiotics). Therefore, the maximum contribution of the change to the buccal route of misoprostol could be as high as 67% (if screen-and-treat were completely ineffective), and as low as 0%. It seems unlikely that screen-and-treat was completely ineffective, because at least some who test positive would have been treated in time to prevent serious infection. The maximum contribution of routine use of antibiotics could be as high as 100% (if the switch to buccal administration of misoprostol were completely ineffective) and no lower than 33%.

In this paper, we examine whether the degree of severity of serious infection differed in the three infection-control groups (vaginal misoprostol and no antibiotics, buccal misoprostol and screen-and-treat, buccal misoprostol and routine antibiotics) or, equivalently, whether the declines in rates of serious infections after the adoption of new infection control measures differed across degree of severity categories. Of particular importance is whether the new infection control measures selectively reduced the least severe serious infections.

2. Methods

We examined the 92 serious infections we previously identified; these were defined as cases in which the patient had fever accompanied by pelvic pain and was treated with intravenous antibiotics either in an emergency department (ED) or in-patient unit, or where sepsis (a systemic, whole-body infection with positive blood cultures that can become quickly overwhelming despite multiple antibiotics) or death caused by infection was documented.

Two authors (MF and ESL) independently classified the severity of these serious infections (and then reconciled any differences) in one of four categories:

1. Infections treated in an ED with intravenous antibiotics

2. Infections requiring hospitalization

3. Infections classified as sepsis or requiring surgery to repair or remove an organ damaged by infection

4. Infections resulting in death.

All authors reviewed and concurred with these classifications. The coding system employed, ranging from 1 (infection treated in the ED with IV antibiotics) to 4 (death), is conservative because while it orders the degree of severity, it does not reflect the magnitude of increasing severity of infectious outcomes.

We compared the distribution of the severity of serious infections in the three infection control groups (no antibiotics, screen-and-treat, and routine antibiotics), or, equivalently, assessed whether the declines in rates of serious infections after the adoption of new infection control measures differed across degree of severity categories using the Jonckheere-Terpstra test for a doubly ordered 4×3 table. All calculations were performed in Cytel Studio 8 (Cytel Inc, Boston, Massachusetts).

The Allendale Investigational Review Board approved the study protocol and design as a retrospective analysis of data routinely collected for quality control. Patient consent was not required to use these data.

3. Results

Information about the serious infections in the three infection control groups (none, buccal misoprostol and screen-and-treat, buccal misoprostol and routine antibiotics) is shown in Tables 1-3. There were 67 serious infections (following 72,195 medical abortions, rate=93 per 100,000) before the new infection control measures were introduced, 18 (following 40,110 medical abortions, rate=45 per 100,000) in the clinics which implemented screen-and-treat, and 7 (following 115,518 medical abortions, rate=6 per 100,000) in the clinics which added routine antibiotics. Altogether, of the 92 serious infections, 63 were classified as the least serious (category 1), 21 were classified as category 2, 7 as category 3 and 1 as category 4. The distribution of serious infections by severity category for each infection control group and the reductions in the rates of serious infection for each category are shown in Table 4.

Table 1. Serious infections before the adoption of buccal misoprostol and screen-and-treat or routine antibiotics strategies.

ED/H	Severity	Sepsis	Organism	Notes
H	4	yes	Clostridium perfringens	Death 5 days post misoprostol
H	3		not reported to PPFA	Hosp × 3 days–tube and ovary removed for infection and abscesses, IV antibiotics
H	3	yes	not reported to PPFA	Hosp × 4 days–sepsis, positive blood culture, IV antibiotics
H	3	yes	E. coli	Hosp × 5 days–GC/CT negative, blood cultures positive for E. coli, IV antibiotics
H	3		Bacteroides	Hosp × 5 days–multiple abscesses, hysterectomy & oophorectomy, bacteroides cultured from abdominal fluid, IV antibiotics
H	3	yes	not reported to PPFA	Hosp × 5 days–D&C, positive blood culture, IV antibiotics
H	3		not reported to PPFA	Hosp × 2 days–ovarian thrombosis secondary to pelvic infection; IV antibiotics
H	2			Hosp × 3 days–IV antibiotics
H	2			Hosp × 3 days–IV antibiotics and D&C
H	2			Hosp × 3 days–multiple IV antibiotics
H	2			Hosp × one week–D&C, IV antibiotics
H	2			Hosp × 24 h–IV antibiotics
H	2			Hosp × 3 days–D&C, IV antibiotics
H	2			Hosp × 2 days–fever 105°F, IV antibiotics, D&C
H	2			Hospitalized overnight–IV antibiotics
H	2			Hosp × 2 days–endometritis, IV antibiotics
H	2		Gram-positive rods from endometrial tissue	Hosp × 2 days for endometritis; IV antibiotics
H	2			Hosp × 3 days for endometritis; IV antibiotics
H	2			Hospitalized overnight–IV antibiotics, infection, pain, fever, high WBC
H	2			Hosp × 5 days–IV antibiotics, severe endometritis
H	2			Referred to ED for severe pelvic pain/fever, elevated WBC at ED–hosp × 24 h–IV antibiotics
H	2			Hosp × 24 h–D&C, IV antibiotics for endometritis, negative GC/CT
ED	1			Endometritis–IV antibiotics in ED
ED	1			IV antibiotics in ED
ED	1			Uterine bacterial infection–IV antibiotics and D&C in ED
ED	1			Fever of 101.3°F–IV antibiotics in ED
ED	1			Pain and fever, D&E–IV antibiotics in ED
ED	1		CT	D&C in ED, IV antibiotics, positive CT identified prior to discharge
ED	1			Pain and fever–IV antibiotics in ED
ED	1			Endometritis–IV antibiotics in ED
ED	1			Endometritis–IV antibiotics in ED
ED	1			Treated for PID in clinic, did not improve–IV antibiotics in ED
ED	1			Severe UTI–IV antibiotics in ED
ED	1			IV antibiotics in ED for endometritis
ED	1			UTI and uterine infection–IV antibiotics in ED
ED	1			Ciprofloxacin and IV metronidazole in ED for endometritis
ED	1			PID treated as outpatient did not improve. Referred to ED for endometritis–IV antibiotics in ED
ED	1			IV antibiotics in ED
ED	1			Endometritis–IV antibiotics and D&C in ED
ED	1			IV antibiotics and D&C in ED
ED	1			IV antibiotics and D&C in ED
ED	1			IV antibiotics in ED
ED	1			IV antibiotics in ED for uterine tenderness and fever
ED	1			IV antibiotics in ED for abdominal pain and fever
ED	1			IV antibiotics in ED
ED	1			IV antibiotics in ED for endometritis
ED	1			Ciprofloxacin and IV metronidazole in ED
ED	1			IV antibiotics in ED
ED	1			IV antibiotics in ED for endometritis
ED	1			IV antibiotics in ED for endometritis–elevated WBC
ED	1		CT	IV metronidazole for endometritis in ED–positive CT diagnosed in work-up
ED	1			IV antibiotics and D&C in ED
ED	1			IV antibiotics in ED–heavy bleeding 2 weeks post misoprostol
ED	1			IV antibiotics in ED
ED	1			IV antibiotics in ED–elevated WBC, fever
ED	1			IV antibiotics in ED
ED	1			IV antibiotics in ED for severe UTI
ED	1			IV antibiotics in ED–fever, pain, high WBC
ED	1			IV antibiotics in ED–fever 103°F
ED	1			IV antibiotics in ED for endometritis
ED	1		CT	IV antibiotics and ciprofloxacin and doxycycline in ED
ED	1			IV ceftriaxone twice and metronidazole for endometritis–two separate treatments in ED
ED	1			IV metronidazole in ED for endometritis
ED	1			IV antibiotics in ED for endometritis
ED	1			Levofloxacin & IV metronidazole in ED for endometritis–pain, fever, high WBC
ED	1			IV antibiotics in ED–high WBC
ED	1			IV antibiotics in ED for endometritis

Notes: ED=treated in emergency department, H=hospitalized, WBC=white blood count, GC=gonorrhea, UTI=urinary tract infection, PID=pelvic inflammatory disease, D&C=dilation and curettage, CT=Chlamydia trachomatis. Number of medical abortions = 72,195; number of serious infections = 67.

Table 3. Serious infections during buccal misoprostol and routine antibiotic strategy.

ED/H	Severity	Sepsis	Organism	Notes
H	2		Clostridium difficile	Hosp × 4 days–Clostridium difficile, IV antibiotics. ciprofloxacin (instead of doxycycline) for routine antibiotic coverage. Not a reproductive tract infection.
ED	1			IV antibiotics in ED for severe UTI (not endometritis)
ED	1			IV antibiotics in ED–fever, pain, released on oral meds
ED	1			IV antibiotics in ED–fever, elevated WBC, pelvic pain
ED	1			IV antibiotics in ED–fever, elevated WBC, pelvic pain
ED	1		CT	Pain/fever developed before patient could be treated for positive CT test – IV antibiotics and azithromycin in ED (CT retested and confirmed)
ED	1		Group D strep	IV antibiotics in ED–severe pelvic pain/fever–culture of endometrial sample showed Group D Strep.

Notes: ED=treated in emergency department, H=hospitalized, WBC=white blood count, UTI=urinary tract infection, CT=Chlamydia trachomatis. Number of medical abortions = 115,518; number of serious infections = 7.

Table 4. Serious infections by degree of severity and infection control measure.

	Infection controls					Reduction in the rate of serious infection after new infection control measures
Severity	None	Screen-and-treat and buccal misoprostol	Routine antibiotics and buccal misoprostol	Screen-and-treat or routine antibiotics and buccal misoprostol	Total
1	45 (67%)	12	6	18 (72%)	63	81%
2	15 (22%)	5	1	6 (24%)	21	81%
3	6 (9%)	1	0	1 (4%)	7	92%
4	1 (1%)	0	0	0 (0%)	1	100%
3&4	7 (10%)	1	0	1 (4%)	8	93%
Total	67	18	7	25	92	83%
Abortions	72,195	40,110	115,518	155,628	227,823

These distributions did not differ across the three infection control groups (p=0.48). Nor did they differ when the screen-and-treat and routine antibiotics groups were combined and compared with the group not receiving antibiotics (p=0.59).

The inclusion of one category 2 serious infection in the routine antibiotics group is questionable, because the woman was hospitalized after she developed Clostridium difficile, a gut infection, not a reproductive tract infection. The provider did not use the standard doxycycline antibiotic regimen discussed here; she was treated with ciprofloxacin rather than doxycycline and ciprofloxacin is associated with development of C. difficile [2]. However, the results did not change when this case was eliminated (p=0.32 and p=0.44, respectively).

Likewise, there were no significant differences in the severity distributions when the one death was reclassified as a category 3, regardless of whether the case above was included or excluded and regardless of whether the screen-and-treat and routine antibiotic groups were combined or not.

Moreover, all the results above hold if only severity categories are treated as ordered (using the Kruskal-Wallis test for a singly ordered (by severity of infection) table).

4. Discussion

We found reductions in the rates of serious infection after the adoption of new infection-control measures of 81% for both severity categories 1 and 2, 92% for severity category 3, and 100% for severity category 4, or a 93% reduction if categories 3 and 4 are combined (Table 4). These differences are not statistically significant. Our ability to reach a definitive conclusion is limited, however, by the fact that the number of serious infections was small before (N=67) and very much smaller after (N=25) the introduction of the new infection control measures. It is noteworthy that there were no deaths or open surgeries among the 155,628 women who received medical abortion since the changes in the regimen were instituted; there was one case of sepsis among the group of affiliates that used the screen-and-treat method, prior to the requirement of routine antibiotics for all women receiving medical abortion. From April 1, 2006, through December 31, 2008, Planned Parenthood clinics provided 151,440 medical abortions using buccal misoprostol and routine antibiotics. Although the precise number of medical abortions since the beginning of 2009 to present is not yet available, adverse events are reported in real time, and up to the present (April 6, 2010), there have been no reports of category 3 or 4 infections. (personal communication E. Talmont, April 2010.

The advantage of the screen-and-treat approach is that if a woman resumes sexual relations with an infected partner, she is likely to become reinfected, and screening allows for identification and treatment of an infected partner. There are disadvantages as well. Because many clinics provide medical abortion without public funding, any costs of screening and treatment must be passed on to the clients. Chlamydia testing can be expensive, depending on the technology used. In addition, there is an inherent delay between the time the woman is tested and the test results are reported, and then another inherent delay between reporting of a positive test and treatment of the individual; in this interim, uterine infection can develop. If screening and routine antibiotic coverage cannot be used simultaneously due to financial constraints, it is therefore logically certain that routine provision of antibiotics alone is a superior approach to prevent endometritis post medical abortion caused by chlamydia or gonorrhea than screen-and-treat alone: it prevents at least as many of these infections at far less cost (in ESL's clinic 14 100 mg doxycycline tablets cost $0.65 and the cost of the screening test alone is $12.50). Another reason the approach of using routine antibiotics was superior to screening and treating positive cases of sexually transmitted infections is that most post-abortal infections are polymicrobial and are not caused by a single sexually transmitted organism for which the client can be tested and treated [3,4]. For all these reasons, providing routine antibiotics is far more cost effective in preventing serious infection since more serious infections are prevented (50% more in our previous results [1]) at far less cost.

The revised regimen which employed buccal misoprostol and routine antibiotics reduced the rate of serious infection across all levels of severity. The cases of chlamydia and/or gonorrhea and of a variety of other potential pathogens which can result in polymicrobial postabortal infection, but for which there is no feasible testing, were prevented from ascending into the uterine cavity and developing into an upperreproductive tract infection by the broad-spectrum activity of doxycycline. It is possible that even if vaginal colonization of a potentially pathogenic organism isn't completely eradicated, the bacterial load is reduced by a course of antibiotics. The action by which doxycycline reduced the rate of infection across all severity levels is almost certainly due to its broad-spectrum activity against multiple organisms, including those that are known sexually transmitted infections that cause postabortal infection as well as other organisms that are not usually pathogenic, but may become so if they ascend to a uterine cavity containing necrotizing decidua following medical abortion, multiply, and result in postabortal infection. Since the infection-preventing potency and breadth of doxycycline are not well described in clinical medicine, we were concerned that doxycycline might be less effective in deterring more serious infections associated with medical abortion. The objective of this paper was to determine whether the regimen changes reduced the rate of serious infections selectively in certain categories of severity or whether rates were reduced in all categories of severity. Reassuringly, we found that rates of serious infection following medical abortion were reduced in all categories of severity.

Table 2. Serious infections during buccal misoprostol and screen-and-treat strategy.

ED/H	Severity	Sepsis	Organism	Notes on condition and treatment
H	3	yes	unknown	Hosp × 3 days–IV antibiotics. Patient ill the month prior to medical abortion with osteomyelitis of the sternum which perhaps contributed to development of sepsis (positive blood cultures, organism not reported despite numerous attempts to get hospital records).
H	2			Hosp × 7 days–high fever, pelvic pain; IV antibiotics
H	2			Hosp × 2 days–fever, pelvic pain, high WBC; IV antibiotics
H	2			Hosp × 3 days–high WBC, pelvic pain, appeared very ill, IV antibiotics
H	2			Hosp × 2 days–pelvic pain, fever, elevated WBC w left shift, IV antibiotics
H	2			Hospitalized × 48 h–fever 103°F, pain, IV antibiotics
ED	1			IV antibiotics in ED–high fever/pelvic pain
ED	1			IV antibiotics in ED–high fever/pelvic pain
ED	1		CT	Treated as outpatient w azithromycin–developed fever, pain did not abate–IV antibiotics in ED
ED	1			IV antibiotics in ED–fever, severe pelvic pain
ED	1		CT	Pain/fever developed before patient could be treated for positive CT test–IV antibiotics and azithromycin in ED (CT retested and confirmed)
ED	1			IV antibiotics in ED–fever/pelvic pain
ED	1			IV antibiotics in ED–fever/pelvic pain
ED	1			IV antibiotics in ED–fever/pelvic pain
ED	1			IV antibiotics in ED–fever/pelvic pain
ED	1			IV antibiotics in ED–fever/pelvic pain
ED	1			IV antibiotics in ED–fever/pelvic pain
ED	1		CT	Pain/fever developed before patient could be treated for positive CT test–IV antibiotics and azithromycin in ED (CT retested and confirmed)

Notes: ED=treated in emergency department, H=hospitalized, WBC=white blood count, CT=Chlamydia trachomatis. Number of medical abortions = 40,110; number of serious infections = 18.