Bioavailability of Oral Ciprofloxacin in Early Postsurgical Patients

Objective  To evaluate the absorption of oral ciprofloxacin within 24 hours of laparotomy for major elective surgery or peritonitis.

Design  In this prospective trial, patients were given a 750-mg oral dose the morning after major elective surgery (n=15) or surgery for peritonitis (n=7). Healthy volunteers served as controls (n=9). Serial urine and blood samples were drawn during the subsequent 12 hours, and pharmacokinetic measures were determined by standard high-performance liquid chromatography assay procedures.

Setting  Multicenter, university-affiliated hospitals.

Main Outcome Measures  Drug absorption as determined by area under the concentration time curve, maximum concentration, and time to maximum concentration.

Results  Oral bioavailability was reduced in elective surgery and peritonitis patients compared with controls. Among the 15 elective surgery patients, 27% (4/15) showed no absorption. The remaining 73% (11/15) had an area under the curve comparable with that of controls (8.3±1.6 (mg/[L·h]). Among all patients, those who showed drug absorption vs those who showed no absorption did not differ with respect to malignant neoplasm, case type, age, or biochemistry. However, patients showing no absorption were significantly heavier than patients showing absorption (patients showing absorption, 15%±3% over ideal body weight vs patients showing no absorption, 29%±6% over ideal body weight; P<.05). When elective surgery patients were stratified by presence or absence of obesity (25% above ideal body weight), mean area under the curve in nonobese patients was 9.80±2.37 vs 0.91±0.56 (mg/(L·h) in obese patients (P<.05).

Conclusions  Oral bioavailability was reduced for peritonitis surgery patients on the first day postoperatively, and for obese elective surgery patients. To achieve adequate serum levels requires continuation of intravenous antibiotics in patients with peritonitis, and adjustment of oral dosage in obese patients in the early period after elective surgery.

ANTIMICROBIAL therapy is an important component of nosocomial infection in surgical patients.^1,2 Such therapy is usually administered by the intravenous (IV) route, to ensure adequate circulating drug levels and obviate the need for a functional alimentary tract. However, the cost and potential local complications of IV therapy have fueled efforts to convert to oral therapy for certain infections. The success of this strategy is predicated on the ability of the drug to achieve adequate levels in the blood and tissue after oral administration.^3,4 This approach has not gained wide acceptance because of concerns regarding the adequacy of absorption of orally administered agents after abdominal surgery.

Ciprofloxacin is a 6-fluoroquinolone with broad antimicrobial activity against both gram-negative and gram-positive microorganisms.^5,6 It is well absorbed, with a bioavailability of 70% to 80%, and penetrates efficiently into most tissue compartments, including the peritoneum.⁷ These features suggest that oral ciprofloxacin may represent an alternative to parenteral therapy for the treatment of intra-abdominal infection. Solomkin and colleagues⁸ recently demonstrated that the sequential use of IV ciprofloxacin to oral ciprofloxacin as part of combination therapy with metronidazole for intra-abdominal infections provided results equivalent to treatment with IV therapy alone. In that study, the conversion from IV to oral administration of ciprofloxacin occurred at days 3 to 5 of therapy at the discretion of the attending surgeon.⁸ Whether earlier conversion to oral therapy might be equally effective was not evaluated. The pharmacokinetics of ciprofloxacin absorption after abdominal surgery are unexplored.

The major objective of the present study was to evaluate the bioavailability of oral ciprofloxacin in patients within 24 hours of abdominal surgery for elective procedures or peritonitis compared with healthy controls. Moreover, we sought to identify factors associated with impaired oral absorption of ciprofloxacin.

Individuals were prospectively enrolled into this trial after they gave informed consent. Patients were studied at 2 study sites: The Toronto Hospital (Toronto, Ontario) and the Royal Victoria Hospital (Montreal, Quebec), and divided into 3 groups. Group 1 consisted of healthy volunteers; group 2, patients on the first postoperative day after a major open elective abdominal procedure; and group 3, patients on the first postsurgical day after laparotomy for secondary peritonitis (defined as free pus or gastrointestinal tract material in the peritoneal cavity).

Patients were included for study if they were older than 18 years and able to grant informed consent. Because of the known interaction of cationic-containing medications (eg, antacids, sucralfate, iron, or calcium) with ciprofloxacin, patients receiving these medications were enrolled only if these drugs could be withheld for a minimum of 6 hours before and 4 hours after the ciprofloxacin dose was administered. Patients in group 3 continued receiving their intravenous antibiotics as indicated by the clinical conditions.

Criteria for exclusion included known hypersensitivity to fluoroquinolones, Acute Physiology and Chronic Health Evaluation (APACHE) II⁹ score greater than 30, active gastrointestinal tract hemorrhage, hepatic enzyme level elevation beyond 3 times the upper limit of the normal range, creatinine clearance less than 0.50 mL/s (30 mL/min), high nasogastric tube output (>1000 mL/8 h), and shock, defined as systolic blood pressure less than 100 mm Hg. Patients were also excluded from the study if they were being treated with ciprofloxacin, anticonvulsant medications, theophylline, or oral anticoagulants, or if they were pregnant or nursing.

All participants received a single 750-mg tablet of ciprofloxacin 24 hours after elective surgery or surgery for peritonitis. In patients with nasogastric tubes, the tablet was crushed in a glass mortar, suspended in 60 mL of distilled water, and administered via the tube. To ensure complete drug administration, the mortar and tube were subsequently rinsed with 2 additional 30-mL volumes of water, and the nasogastric tube was clamped for 2 hours. In patients without nasogastric tubes, a single dose of 750 mg by mouth was administered, 2 hours before the first meal of the day. Healthy controls received medication by mouth.

Blood samples (5 mL) were collected via a saline lock immediately before ciprofloxacin administration and 0.5, 1, 1.5, 2, 3, 4, 6, 8, and 12 hours after the dose. Blood samples were also drawn at baseline and 12 hours after administration for assessment of hepatic function (aspartate aminotransferase, alanine aminotransferase, bilirubin, and alkaline phosphatase), renal function (creatinine), and hematological function (complete blood cell count). Urine samples were obtained before drug administration, and in parallel with blood sample collection where possible.

Blood samples were centrifuged (3000 rpm for 20 minutes), and the plasma was stored in polypropylene tubes at −20^oC until analysis. Urine volumes were measured, and a 3-mL aliquot of each sample was stored at −20^oC until analysis.

Serum drug levels were determined by standard high-performance liquid chromatography assay procedures,¹⁰ and pharmacokinetic measures (area under the concentration time curve [AUC], maximum concentration [C_max], time to maximum concentration [t_max], half-life, and clearance) were thus determined as described by Nix et al.¹⁰ The AUC is predictive of bacterial eradication for a wide range of antibiotics, including ciprofloxacin,¹⁰ while the C_max and t_max provide information on the absorption profile of a given agent.¹¹ These measures may be influenced by altered volumes of distribution and gastrointestinal tract function¹²; therefore, the determination of these absorption measures can provide insight into the potential efficacy of an orally administered drug like ciprofloxacin. Data from healthy controls (group 1) were used for comparison. Statistical comparison of the 3 groups was by analysis of variance.

Ideal body weight (IBW) was calculated from the following standard formula: for men, IBW (kg)=50+5.75 × (number of centimeters over 1.5 meters); for women, IBW (kg)=45.5 + 5.75×(number of centimeters over 1.5 meters).

We enrolled 31 patients from December 1, 1996, to March 30, 1998. Except for 1 patient in group 2, all patients were treated for benign disease. The demographic characteristics of the test population are shown in Table 1. Patients in each group were similar with respect to age, baseline hematological measures, and biochemistry. Although patients in group 3 were being treated for peritonitis, APACHE II scores were similar to those of patients in group 2, indicating low degree of illness severity in both groups. Patients in group 3 were not significantly more likely to receive ciprofloxacin via the nasogastric tube when compared with the other groups (Table 1). The mean weight of patients in group 2 showed a trend toward being higher than that of other groups, although this difference was not statistically significant (P=.06 by analysis of variance).

To evaluate the bioavailability of oral ciprofloxacin in the early postsurgical period, the pharmacokinetic measures of surgical patients (groups 2 and 3) were compared with those of healthy volunteers (group 1) after the ingestion of a single 750-mg dose.

As is shown in Table 2, oral bioavailability was reduced in the elective surgery (group 2) and peritonitis (group 3) groups compared with controls. This was manifested by reduced AUC as well as a lowered C_max (Table 2). Because demographic characteristics were comparable among groups (Table 1), this suggested that specific factors might account for the impaired absorption.

Review of individual patient data showed that no absorption occurred in 27% (4/15) of the elective surgery group and 14% (1/7) of the peritonitis group, while the remaining 73% (11/15) of the elective surgery patients had an AUC comparable with that of controls (8.3±1.6, not significant compared with controls). Among all patients, those who showed absorption of ciprofloxacin vs those who showed no absorption of ciprofloxacin (herewith referred to as absorbers and nonabsorbers, respectively) did not differ with respect to sex, malignant neoplasm, case type, age, or the presence of a nasogastric tube. However, nonabsorbers were significantly heavier than absorbers (absorbers, 15% ± 3% over IBW vs nonabsorbers, 29% ± 6% over IBW; P<.05). When elective surgery patients were stratified by the presence or absence of obesity (defined as 25% above IBW), the mean AUC in nonobese patients was 9.80 ± 2.37 vs. 0.91 ± 0.56 mg/(L · h) in obese patients (P<.05). The same association did not exist in the peritonitis group.

Intravenous antimicrobial therapy is often used in the early postoperative period after abdominal surgery.^13,14 Traditionally, therapy has been initiated by the parenteral route to ensure bioavailability of the agent. Substitution with oral agents, however, might be of considerable benefit in reducing the incidence of infusion-related complications, allowing earlier hospital discharge for some patients, and reducing drug and drug-administration costs.¹⁵ Indeed, sequential therapy with conversion from IV administration to oral administration has become standard management for pneumonia and urinary tract infections in the nonsurgical patient.¹⁶ Ciprofloxacin, with its broad antimicrobial spectrum of activity, wide tissue distribution, and efficient absorption across the gastrointestinal tract, represents a candidate drug for oral administration in these settings. A recent report showed that ciprofloxacin in combination with metronidazole was effective for intra-abdominal infections when administered initially by IV, with conversion to oral administration 3 to 5 days later when the gastrointestinal tract had recovered and clinical status had improved.⁸ Whether earlier conversion to oral therapy is feasible was not studied. Clearly, however, the use of oral ciprofloxacin in these patients is predicated on its ability to achieve adequate levels in the blood and tissues.^4,17

In the current studies, we assessed the bioavailability of oral ciprofloxacin in the early postoperative period in both elective surgery patients and in those undergoing surgical management for peritonitis. The major finding of this study was that oral bioavailability as determined by AUC and C_max was significantly reduced in both patient populations compared with control patients.

The reduction in oral bioavailability observed in postoperative period for patients who had undergone elective surgery suggested the possibility that these individuals manifested a global defect in absorption or distribution of the drug.⁴ In this regard, the presence of intestinal ileus, an increased volume of distribution, or delayed gastric emptying could account for these pharmacokinetic findings.^4,18 It is unlikely, however, that such general defects existed in patients following surgery, as inspection of individual results revealed 2 distinct groups. The majority of patients had circulating levels of ciprofloxacin following elective surgery that were comparable with those of control patients (absorbers), while a small subset of individuals had no measurable ciprofloxacin (nonabsorbers). These groups were similar with respect to malignant neoplasm, case type, age, sex, and the presence of a nasogastric tube. However, nonabsorbers were significantly more obese (defined as mean weight above 25% of IBW) than absorbers. The inhibitory effect of increased body weight on AUC and C_max most likely results from an increased volume of distribution in these patients, although concomitant effects on absorption or metabolism cannot be excluded. The specific role of body weight in predicting absorption of oral ciprofloxacin is indicated by the finding that increased body weight was associated with decreased AUC when both control and surgical patients were combined.

In patients with peritonitis, several factors likely contributed to the poor bioavailability.⁴ Acutely ill patients such as these are known to exhibit increased volumes of distribution, which would by definition result in reduced circulating levels of the drug and lowered AUC.^4,12 It is difficult to assess how perturbed the volume of distribution might be in these patients, given the relatively low mean APACHE II scores (Table 1). In addition, bowel function may be impaired by the presence of ileus or the administration of vasoactive agents. To resolve whether patients in group 3 exhibited a general impairment in gastrointestinal tract absorption as opposed to a specific defect in ciprofloxacin absorption, further studies are needed to determine the pharmacokinetic profile of an unrelated orally administered agent.

The findings of the current study offer insight into the therapeutic efficacy of oral ciprofloxacin in the postoperative period. On the basis of previous work by Barriere et al,¹⁹ Forrest and colleagues¹² established that the ratio of the AUC to the minimum inhibitory concentration (MIC) of ciprofloxacin was highly predictive of bacterial eradication and clinical cure.^12,20 When the AUC/MIC ratio was below 125, the probability of treatment failure for a given bacterium was 3 times greater than for values above this threshold.¹² Ratios between 125 and 250 were predictive of acceptable antibiotic coverage, while those above 250 were found to be optimal.¹² We therefore used these threshold values to predict the efficacy of oral ciprofloxacin in treating infections with Escherichia coli (MIC, 0.03 µg/mL), Proteus mirabilis (MIC, 0.06 µg/mL), and Staphylococcus aureus (MIC, 0.5 µg/mL) in patients from all 3 groups based on the pharmacokinetic measures determined in this study (Table 2). We predicted that infection with E coli would be effectively eradicated in the control, elective surgery, and peritonitis groups (AUC/MIC ratios of 457, 230, and 233, respectively). By contrast, infection with P mirabilis would be eradicated only in control patients (AUC/MIC ratio of 228), but not in elective surgery or peritonitis patients (AUC/MIC ratio of 115 and 116, respectively). When patients undergoing elective surgery in the absorber group are considered, these patients would be expected to eliminate this pathogen also after oral ciprofloxacin treatment (AUC/MIC ratio of 138). Patients in all 3 groups were predicted to have insufficient circulating levels of ciprofloxacin to treat infection with S aureus (AUC/MIC ratio of 27.4, 13.8, and 14.0, for groups 1, 2, and 3, respectively).

In summary, this study indicates that oral bioavailability of ciprofloxacin is reduced for patients undergoing surgery for peritonitis on the first postoperative day. For patients who had undergone elective surgery, we determined that absorption was comparable with that of controls for patients with normal body weight but significantly decreased for obese patients. These findings suggest a strategy of continuing IV antibiotics in patients with peritonitis and adjusting oral dosage in the early postoperative period in obese patients who have undergone elective surgery.

Donald E. Fry, MD, Albuquerque, NM: The quinolone antibiotics have the promise of replacing applications of IV antibiotics in many clinical situations, because, as Dr Hackam has indicated, their bioavailability via the oral route appears to be extraordinary, and in some studies the bioavailability of various quinolones almost approaches giving the antibiotics by an intravenous route. The particular study in examining pharmacokinetics in surgical patients is particularly timely at this point before we embark on wholesale studies where we are using antibiotics by mouth in very sick patients, either after operation or in the intensive care unit.

The observation, however, of patients that are overweight not absorbing the drug becomes a fairly interesting one to ponder as to why that should be true. One question that immediately comes to mind is, were there medication differences in how these patients were treated postoperatively in the obese vs the nonobese patient, since there is a whole wealth of data now that shows that morphine, in particular, and perhaps other opiates actually interfere with the absorption of the quinolone antibiotics from the gastrointestinal tract. So I would immediately wish to ask the author as to whether they have studied the analgesics and the opiate derivatives that were given to these patients, and whether, perchance, some of the larger patients may have received morphine or perhaps even demerol in greater concentrations.

If their premise is true, it would have been interesting to have seen a regression plot of body mass index of all of the patients studied against the AUC relationships. It is hard for me to believe that 25% body weight somehow becomes automatically a magic threshold beyond which people don't absorb the drug, where people who are less than 25% over ideal body mass somehow have very effective absorption. So a BMI, ie, body mass index, plot against AUC in the form of a regression analysis I think would have been most interesting.

I am still having trouble pondering the issues of why the obese patient does not absorb the drug, and it brings me to the question of whether the observations here may be drawing the wrong conclusion, and that is, is the drug in fact actually being absorbed, but that the volume of distribution in the obese patient is dramatically different? We know for many of the quinolones that the volume of distribution of the drug is greater than the patient's body water, and you might ask yourself, how could that possibly be true? But if there is tissue binding of an antibiotic, you may actually have the serum concentration in fact not reflect the true amount of antibiotic that has been absorbed, the true amount of antibiotic that exists in the tissue. Are the quinolones, is ciprofloxacin in specific, lipophilic? Did the patients in fact absorb the drug, but what we are measuring in blood is not a reflection of where the drug is at?

I don't remember off the top of my head what the normal volume of distribution for ciprofloxacin happens to be, but for many of the quinolones, the volume of distribution may be a very precarious number, because the volume of distribution is larger than the patient's body water, suggesting that there is tissue-level binding of the drug. We have to ask the fundamental question of, did they really absorb the drug, but because of the very large fat reservoirs and perhaps potential lipophilicity of the agent, do we have perhaps a spurious finding?

I am delighted to see that someone has finally studied antibiotics in pathological conditions rather than with medical student volunteers, for which therapy is then extrapolated to a vast array of very, very ill patients. The observations that he did not emphasize relative to peritonitis deserve tremendous emphasis. Even with effective absorption, if you make that assumption, which may not be true, the volume of distribution, in all likelihood, in the peritonitis patient and the alterations of splanchnic blood flow in the peritonitis patients probably all end up reflecting in aberrations of pharmacokinetics in the patient with peritonitis. That means that we really need to seriously examine the pharmacokinetic profile of every antibiotic we are planning to use in the critical care situation, since we have probably harmed more patients with underdosing antibiotics in the intensive care unit than we have ever harmed people from toxic effects.

Dr Hackam: Your first 3 questions dealt with the mechanism by which patients with obesity can explain the low bioavailability observed. With regard to your first question, it is true that oral ciprofloxacin is interfered with by a variety of other medications administered both orally and intravenously. And you mentioned narcotics. Orally administrated cationic antibiotics, cationic drugs, and acids are known to bind orally absorbed agents. We were very careful to exclude those patients in particular, and from the point of view of narcotics, they were not overrepresented in patients with low bioavailability. So we have to look elsewhere for why the obese patients showed low drug levels, which brings us to the second and third questions.

Was a regression plot performed? It was. And the reason that we didn't include it with the manuscript or, indeed, on a slide today is that it really didn't provide an answer as to whether there was a correlation across all patients between obesity and absorption. Certainly in peritonitis patients, and for reasons which were alluded to in your final question, there are various factors which impaired their absorption or bioavailability, and the inclusion of these patients, which represented a significant number, provided us with a regression plot that was very difficult to obtain statistical significance from.

Is there something magical about greater than 25% above ideal body weight as a cutoff? Probably not. We found that these patients had very low bioavailability for a variety of reasons, which brings us to the third question. The reason probably is that they do have a greater volume of distribution, and, yes, ciprofloxacin is readily delivered to tissues due to its lipophilic properties.

Is this spurious? It is not spurious in terms of a conclusion. From the point of view of the practicing clinician, we need to know what to expect when orally administered agents are given to our patients; and our conclusion is that, in obese patients, dosage has to be monitored carefully, and perhaps bioavailability should be calculated. I wouldn't like to speculate too much on the mechanisms by which this occurs, but your point about volume of distribution is a very good one.

And finally, what is it about the peritonitis patients that affect their decreased bioavailability? A variety of factors are important to keep in mind, and you alluded to several of them. Yes, the volume of distribution is altered, they may have increased production of proteins that may bind free drug and serum, they may clear it differently, they may metabolize it faster. We need to be very much aware of these factors when administering oral agents to very sick patients.

Robert G. Sawyer, MD, Charlottesville, Va: Do you think you can extrapolate these findings to trovafloxacin and alatrofloxacin, which already has the indication for oral perioperative use in colorectal surgery and will probably replace ciprofloxacin and antianaerobe agents combinations in these settings?

Dr Hackam: Conceptually there are similarities; however, I would emphasize the point that, amongst all cohorts administered oral therapy, before conclusions can be drawn upon efficacy, bioavailability should be measured by techniques that we utilized in this paper.

This research was supported by a grant from Bayer Inc, Toronto, Ontario. Dr Hackam is the recipient of an Ethicon–Society of University Surgeons Research Fellowship and a Medical Research Council fellowship.

Presented at the 18th Annual Meeting of the Surgical Infection Society, New York, NY, April 30, 1998.

Reprints: Ori D. Rotstein, MD, Toronto Hospital, 200 Elizabeth St, EN9-236, Toronto, Ontario, Canada M5G 2C4 (e-mail: orotstein@torhosp.toronto.on.ca).