Medical Library Home | Table of Contents

Fever in Infants

Victoria L. Franklin, MD

Sixty-five percent of all children between the ages of birth and two-years-of-age are brought to medical attention for some type of febrile episode. And 75% of these visits are for temperatures greater than or equal to 39 degrees. We have this fear that when the temperature gets up that high that we may be dealing with a significant bacterial infection. In 14% of these kids there’s no definable source for the fever and that’s the problem. So the child that you don’t have identifiable sources that you need to worry about, because that child could be either bacteremic. Could have meningitis or could have a bone or joint infection in evolution.

First of all, I want to define fever. Even among my house officers to date there is a lot of confusion about what really is a fever in a child. Remember that children are smaller, like little birds and little cats and little dogs, they tend to have higher metabolic rates and because of that their core temperatures are higher than that of an adult. So for a child, a rectal temperature of greater than or equal to 38C degrees, or 104F degrees - for those of you who still like to use Fahrenheit - is considered a fever. Anything below that is acceptable as normal. This is an evolutionary state. I mean, the child, by the time you get to be an adolescent, you are probably at normal temperature rates. Women tend to hit normal temperatures faster than males do because they mature faster than males do. So in children, in the age group that we are going to be talking about - which is from 0-36 months of age - this is what we are going to define as fever.

A few caveats. Remember that fever documented at home by a reliable adult should be treated the same way you would treat a fever reported to you by a nurse or a colleague. This is primarily for the house officers that I mention this. But you all well know this and we need to respect these readings by well or reasonable adults at home. Always ask about recent immunizations. You’ve already had the lecture on immunizations and you know that some vaccines tend to give you fever. DPT, the pertussis component, can give you a fever 24 hours or so after the vaccine. But many of you don’t realize that MMR can give you a fever anywhere from 5 to 7 days after the vaccine. It can be as high as 38C or 39C degrees. So if you don’t ask about that vaccination, that child may get a sepsis work-up. That’s seen in about 40% of all kids who get MMR.

Serious bacterial infections can present with normal or subnormal temperatures. Keep that in mind. So an absence of a fever does not rule out a serious bacterial illness in a child. So don’t be lulled into a false sense of security. An issue that we are not seeing right now, this time of year, is over-bundling. One of the problems is what do you do with a child who shows up in your office or in your clinic or in the emergency room, just totally bundled up and about the only thing you see is a little hole, a little blow-hole for their nose so they can breathe out of there. So these little guys of course are going to jump their temperature. I mean, imagine what would happen to you if we wrapped you up in a sleeping bag and put you inside a warm car and then carried you into the emergency room. You definitely would have an increase in your core temperature because you are retaining heat. One way to deal with that and try to figure out if that’s really a fever or over-bundling is summarized here: if you unbundle the patient and recheck the temperature 15 to 30 minutes later, prior to giving any type of antipyretic, and that child’s temperature is down and the patient looks good, then it’s probably a fever related to over-bundling and it’s not really a fever that you need to pursue aggressively. Again, it depends on how the patient is doing and what the history is, but it’s one way to attack this type of a problem.

Lastly, this issue. That the response to antipyretics does not predict the presence of bacteremia. I’m a dinosaur. I’ve practicing for a while and when I was a resident one of the things that our attendings liked to do, and the ER attendings liked to do, was give these kids a dose of Tylenol and see - aspirin, actually. That’s how far back I trained - and see what would happen to their fever. Well, we now know that’s all a bunch of bunk. And here’s one of many many studies that are out there that prove this. Really quickly; this is a temperature response. Here’s the mean temperature before acetaminophen and after acetaminophen in the non-bacteremic group, 135 kids. You’ve got this temperature 40, plus/minus 0.4. In the bacteremic group - without meningitis - about the same temperature. No difference. If you give them Tylenol and say what’s the temperature after that? Again, the difference is not significant. So the response to the Tylenol or an antipyretic does not predict whether the child is bacteremic or not, or has a significant disease.

Why do we worry about this? Well, this is the prevalence of serious bacterial infections in infants less than or equal to 3 months of age who come in with fevers of 39C degrees or greater. We are defining serious bacterial infections as these: meningitis, sepsis, bone and joint infections, urinary tract infections, pneumonia and enteritis. Notice that otitis media is not listed up there and neither is sinusitis because they are not considered serious bacterial infections. They are considered more nuisances than anything else. In the kids with this kind of a febrile response you are going to see about 7% of these kids with some type of a serious bacterial infection. Bacteremia or bacterial meningitis in about 2.5%. So the numbers are out there. Eventually you are going to come across these kids if you see enough kids that are febrile.

The players for kids less than 60 days of age are clearly different than the players of kids greater than 60 days of age. In this group, in the less than 60 days of age group, we are talking about group B beta streptococcus, E. coli, Listeria monocytogenes occasionally, staphylococcus, streptococcus, staphylococcus species and primarily staphylococcus aureus. Streptococcus pneumoniae and Hemophilus influenzae are starting to be seen more frequently. We’ve actually seen a few cases of kids who’ve had strep pneumoniae sepsis as a neonate. So neonatal sepsis due to strep pneumoniae. Probably has to do with inadequate antibodies transferred from the mother to the baby. The same thing with Hemophilus influenzae. We are not seeing Hemophilus influenzae type B disease anymore in kids greater than two or three months of age. That’s because you guys out there are using the vaccine. That HIB vaccine has done away with a terrible pathogen. When I trained we saw three to five in-patient H. influenza type B disease cases a month when I was on the wards. You don’t see them anymore. I now call my house officers down to see one when they occur every one or two years. It’s generally in kids that have not received the vaccine.

In the older age group - that is, the 3 to 36 month of age group - the incidence of occult bacteremia, which is a frequent cause of fever of undetermined origin, seems to have decreased over time. McCarthy, which is one of the first studies out there in 1977, had an occult rate of bacteremia of about 7.1%. We are now somewhere around 3% to 5%. Most people are quoting numbers of 3% to 5%. So about 3 to 5 out of every 100 kids that show up with a temperature greater than 39C or greater or equal to 39C degrees is going to have occult bacteremia. The players in this have clearly, clearly changed. From the 70’s to the early 80’s strep pneumo and Hemophilus influenzae were the big players. But with the advent of a very effective vaccine against Hemophilus influenzae type B, the spectrum has changed to this. Streptococcus pneumoniae now accounts for greater than 90%. Probably 95% to 98% of all cases of occult bacteremia. That’s what you guys are going to be seeing in the real world. Regardless where you are practicing in this country, that’s the major player. The incidence of penicillin resistant pneumococcus has gone up in the last five to six years. So that’s a very important factor when we are managing these kids. Neisseria meningitidis about 2%, Salmonella and Hemophilus influenzae type B bringing up the rear at about 1%.

So summarizing the data for kids 3 to 36 months of age, the prevalence is about 3% to 5%. It really doesn’t matter where you are practicing medicine. You are going to see the same prevalence. The organism again is streptococcus pneumoniae. That’s the major player we are seeing these days.

Why do we worry about occult bacteremia? What’s the big deal here? So you’ve got a few bacteria floating around in your blood stream, who cares? Again, this is a dead disease but it drives home the point as to why this bacteria was targeted for vaccine development very early on. If you look at Hemophilus influenzae B and you ask the question - here’s a kid who shows up in the emergency room, I draw a blood culture and I do nothing else but send him out the door. There are studies that have looked at this. What percentage of those kids are going to be bacteremic when they come back? And what percentage of those kids are going to have meningitis? Well, persistence of bacteremia, depending on the studies, anywhere from 7% to 50%. Somewhere around 27% is what we are quoting. Meningitis in 27%. So if the bacteria floats around long enough it finds its way into the central nervous system, or it gets into a bone or joint and then you get bigger problems. So that’s why we really worry about Hemophilus. Streptococcus pneumoniae is a little better but still a major problem. Depending upon the study, from 7% to 22%, probably about 17% or 15% or so are going to be persistently bacteremic 24 to 48 hours later. Meningitis is going to be seen in about 6%. Again, the longer that bacteria is in there the greater the possibility is that it’s going to work it’s way into the central nervous system and cause greater problems.

Neisseria meningitidis is a little different. In a lot of patients out there - maybe 12 in one study - eight of these actually received antibiotics believe it or not and they still had this outcome. So persistence of bacteremia in one-quarter of those patients, meningitis in another quarter and in 17% they were dead on arrival because of, as you know, the progression of this disease. So this is not a good one to have floating around for a long period of time. The _ which is Salmonella species, is the fourth of the big four. Persistence occurs in up to 57% of these patients. However, the good news is that it is rarely invasive unless you have a severe debilitating disease, you are immunosuppressed or you are at the extremes of life. So the kids that are less than three months of age, for example, have a very high risk of developing meningitis or invasive Salmonella disease, such as bone or joint disease, as do the elderly. Outside of that you don’t see a lot of meningitis even if it’s a persistence of bacteremia. But they can still be very febrile because of this organism floating around in their bloodstream.

Can you tell these kids apart by looking at a clinical scoring scale? Again, it depends on what clinical scoring scale you are using but in general the vast majority of them fail to identify correctly the vast majority of kids who are bacteremic or non-bacteremic. Here’s just one of multiple studies, and I’ll quickly summarize it. This is 688 children between the ages of 3 and 24 months. Temperature was greater than or equal to 39.5C, so these kids were highly febrile. They had a high risk for having bacteremia. They asked for house officers and physicians in an emergency room to decide which kids would be bacteremic or non-bacteremic on physical exam and clinical impression. 633 of these patients were classified as non-bacteremic. In 85% they correctly identified them as being non-bacteremic and 100 of these kids were predicted to be bacteremic when they really weren’t. In this group, that were bacteremic - 25 - only 40% were correctly identified and 60% were missed. So that the clinical scoring skills by themselves missed a significant portion of these patients. So the important question is: is there a way to identify these kids? Or at least cull out the ones that are at greater risk and then leave you with a subgroup that you can do something with, either with prophylactic antibiotic or manage more conservatively. So the answer is, of course, yes.

Toxic infants. If you look at these kids and you ask the straight question, does the kid look toxic? In other words, is this a kid that really looks sick in front of you? By toxic, we are looking at a kid who is lethargic, has poor perfusion, hyperventilation, hypoventilation or cyanosis. If the answer is yes, then there is no question that that kid is going to be a high risk for serious bacterial infection. The numbers speak for themselves; 17% are going to have some type of a serious bacterial infection - 11% bacteremia and about 4% meningitis. So in that group there’s no option other than to admit them and work them up and treat them prophylactically or empirically until we know what’s going on. The question becomes, what about that non-toxic group? If you look at that non-toxic group there’s still a significant proportion of them that’s going to have severe bacterial infections. About 9%; 2% with bacteremia and 1% with meningitis. So there is a way to break out that non-toxic group and that non-toxic group can be broken into what’s called a low risk group. In this group - and I’ll give you the criteria in a few minutes - we drop the incidence of serious bacterial infections to about 1.4%, bacteremia to about 1.1% and meningitis to about half a percent.

To do that we use this scale. These are low risk clinical criteria, and using this coupled with some laboratory data you can cull those kids that are lowest risk for serious bacterial infections when they show up with fever. So here’s what you need to do. A series of questions have to be asked. The first one is: is this a previously healthy child? Was the child in the neonatal intensive care unit? If it was, then it is no longer in the low risk category. Was the child ill in the last three or four weeks? Does the child have otitis media or something, or does it have pneumonia? If the answer is yes then the child is no longer in the low risk group and falls into a high risk group, otherwise he can maintain in the low risk group. Non-toxic clinical appearance. In other words, when you look this kid over does he or she look good? I mean, are they relatively happy? If you bring the fever down do they smile at you? Do they try to interact with you? Or are they still pretty punky? If they try to interact with you, then they are low risk kids. Otherwise they are probably falling out of the low risk group and may be at high risk for serious bacterial infection. On physical examination there should be no focal bacterial infection identifiable, other than possibly an otitis media. So if you still have an otitis media, you are fine. You are in a low risk group. And lastly, the child should come from a good social situation. That to me implies three major criteria. First of all, that the parent or caregiver is a mature, responsible adult. Now there are some 14-year-old and 13-year-old girls our there that are very responsible and can fall into this group of responsible caregivers. And there are some 30 and 40-year-olds out there that I wouldn’t trust my cat to, and my cat is pretty self-sufficient. So you’ve got to weigh that in your decision. The other issue is, of course, transportation. Can that individual get from point A - their home - to the hospital in a reasonable length of time if you need to get a hold of them? And that means usually that they have to have their own transportation.

The other issue is the phone. We’ve all come to accept the phone as part of our lives. Many of us even carry the damn thing on our belts and it’s hard to get away from. But keep in mind that in certain socioeconomic groups they may not have a phone at home. They may have to go down the hall or they may have to go down to the corner and use a phone. We know of instances, for example - when I was working in Denver - where it wasn’t safe to go down and use the phone in the middle of the night because it was a bad neighborhood. Either the pimp was using it to do some transactions or somebody was doing a drug deal on it. So it’s a very important question to ask. If those criteria are met, then you progress to the laboratory data.

The laboratory data are as follows: when examining the patient’s blood count, the white blood cell count - the total count - should be between 5,000 and 15, 000. There are really some very good studies to show that once you deviate from those parameters you start to see kids that get into trouble.

The other issue is the absolute band count. Are you all familiar with how to calculate that? If not I can just run through it quickly. Here’s a quick review. It’s simply; you take that percentage of bands that they give you - they’ll give you a number, 15% or so - and you multiply it times the total white count. So for example, it’s just like figuring out a discount at a store. We all can do that fairly quickly and rapidly. So if you’ve got a white count of 10,000 and your band count is 15%, your absolute band count is 1,500 cells. So what most authorities are recommending is that the absolute band count be below 1,500. Below 1,500. Once you are above that you are looking at a kid that might be infected and probably needs to be worked up, and is no longer a low risk criteria. The other thing to keep in mind is the other spectrum. The other extreme. That is, if the white count is below 500. With a white count below 500 you may be dealing with overwhelming sepsis and what’s happening there - the absolute band count - is that the baby or the child is having its PMN’s consumed because of the infection. So you want to keep an eye out for that, although that is not considered a fall-out of the low risk criteria.

A urine analysis should be obtained in these kids and should be normal; as defined by less than 5 WBC’s per high power field. Or should have a negative gram stain by smear, if you want to do it that way instead of doing the complete urinalysis. If the child has diarrhea there should be less than 5 WBC’s per high power field because that correlates with an absence of a bacterial pathogen. If all those criteria plus the previous criteria are met, that child is in the low risk group and that’s the kid that’s going to have that 1% risk for serious bacterial infection, 1% risk for bacteremia, and less than a percent, less than 0.5% for meningitis. That kid you can usually manage as an out-patient. Buy him an insurance policy, which we’ll talk about in a few minutes.

Let me give you some algorithms here. These are not my algorithms. These are all algorithms derived by Beraf Vidal. Here are the options for kids less than 28 days of age - and I know there’s a lot of controversy about this and every asks me questions at the end about - but I’ll tell you, this is what we are teaching our house officers here to do. Option number one: for a kid less than 28 days of age we are recommending hospitalization and a substance evaluation for all children with a fever greater than 100.4F. That’s because the pathogens that they are subjected to at this age group are particularly virulent and as you all know, they have a relative immaturity in their immune system. So they can’t fend off these infections and these infections are rapidly lethal or rapidly maiming. The substance evaluation should include a CBC with differential, the serum glucose - and I’ll explain why - a lumbar puncture which should include a cell count with differential, protein and glucose determination, and a bacterial culture at minimum. Possibly a viral culture if you have the facilities. I’m a virologist by training so I really push the viral culture.

Now why the serum glucose? Because you need to compare the serum glucose with the CSF glucose. Remember that the concentration of CSF glucose is dependent upon the systemic glucose. There is no endogenous production of glucose within the central nervous system. It gets across the blood/brain barrier by diffusion. By some pump mechanism also, but primarily by diffusion. So that in an infant who has limited glycogen stores and will become starved after not eating for a few hours, the serum glucose may drop to 40 or 50 or 30. If the CSF glucose is 20 at that point it’s still within normal limits because the CSF glucose should be approximately one-half at the lowest of the serum glucose. So that a CSF glucose of 20 in the face of a serum glucose of 40 is okay. That’s why you need to compare them. That will give you another parameter to think about. The protein varies with age. You can look at each of your textbooks about that.

A blood culture should be obtained and then a urine culture. Some of us advocate initiating empiric antibiotics until you have the results of the 48 or 72-hour blood culture. It seems that the interval from the time of the blood draw to the time that you can consider them negative has decreased in time over the last few years. That probably has to do with the fact that we have better culture systems than we did in the past. It now appears that with the newer systems they are so accurate that we get a good idea by 48 hours. And if the blood cultures are negative at 48 hours you can stop the antibiotics, and if the child is doing well, consider sending him home. Or treat them longer if they are not doing well yet.

What are the appropriate antibiotics for a child less than 30 days of age? You can have your choice. Some people like the traditional ampicillin plus gentamicin. The ampicillin/gentamicin should cover most of the pathogens we’ve talked about. Newer regimens include things using like cefotaxime and ampicillin. The common denominator in both of these is the ampicillin component and that’s because that component offers coverage for Listeria monocytogenes. Listeria is not covered by the cephalosporins alone. So you need that one coverage. We don’t use ceftriaxone in this group, as you all know, because it displaces bilirubin from its albumin-binding sites, and so it can lead to kernicterus or jaundice. That’s why we don’t use ceftriaxone on a kid less than a month of age. Otherwise we probably would. It’s a good drug.

What about the kid that’s 28 to 90 days of age? This is where your algorithm for the low risk high risk comes into play. Here again you begin with the non-toxic low risk child. And the first question is, is he in that group? Yes or no? If he’s not, if he falls out of the low risk group, then he should be admitted. You should do the complete work-up, including the lumbar puncture and parenteral antibiotics and I’ll talk about this in a minute. On the other hand, if he does meet this low risk group then you have the option of an outpatient management. That should include - there are two options here, I’m going to talk about option one first. In option one the idea is to do a blood culture, a urine culture and a lumbar puncture as an outpatient. As an outpatient. Then you send them on their merry way with a shot of ceftriaxone. Now ceftriaxone should give you broad enough coverage for most pathogens that these kids have. But it’s imperative that every child receive a lumbar puncture before they get their dose of ceftriaxone, and here’s the reason why. There are three clinical scenarios where this is very important. Lets say you do this test in the middle of the summer. The kid comes in, you decide not to do a lumbar puncture and you send the kid out the door. Well, two or three days later the kid is still febrile. What do you do at that point? Well, you bring them back and at that point you’ve given them already a shot of ceftriaxone. You are now going to do the lumbar puncture. Scenario number one. You get the needle in, it’s a perfect tap. Beautiful. No problem, no blood except that now you’ve got 12 to 20 WBC’s in the spinal fluid. Maybe it’s a 50/50 mix - 50% lymphocytes, 50% PMN’s. The glucose is normal, the protein is normal. What do you do? You usually call somebody like me or somebody else who does this for a living and you say, "What do I do with this culture at this point?" and it happens every summer. What do I do with this kid at this point? Well, this time I can’t tell you whether this kid has a partially treated bacterial meningitis, because you’ve given them ceftriaxone, or the child has a viral meningitis. You are almost obligated at that point to complete a seven or ten day course of IV antibiotic for partially treated meningitis. That’s one scenario. Had you done the lumbar puncture earlier and it hadn’t grown in 48 hours, then the child probably has a viral meningitis.

The other scenario is: let’s say you put the needle in the back and this time it’s a bloody mess. At that point there’s no way you can sort out whether it’s bacterial or viral. So it’s very important at that point to rely on the culture. But the culture can’t be relied upon because you’ve already given the kid a ceftriaxone shot. That’s why we stress, we stress, performing the lumbar puncture before we give the ceftriaxone.

Finally, the issue is the reevaluation in 24 hours. So don’t just send them out the door, and it’s not seen so much in the primary care setting as it is in the urgent care and the emergency room settings, where these kids get a shot and they are sort of sent out the door and there’s no follow-up arranged. These children need to be reevaluated in 24 hours to make sure there has not been an acute deterioration and so that you can go over the laboratory studies and see if anything else has to be done at that point.

Now, the dissenting group said that if these kids look good, why do you even want to give them a shot of ceftriaxone? Why don’t you just send them home and watch them? Their concern was, this group, that the thing we are going to miss, that’s going to be hardest to make the diagnosis of is a urinary tract infection. So the dissenting group suggested that you have a urine culture done and that nothing else but close observation be performed. I think that’s fine, but again the catch-phrase here is close observation. Which means that somebody - either you or somebody on your staff - is going to call that parent in 24 hours and see how the child is doing, or bring that child back in 24 hours and reexamine at that point.

What about parenteral antibiotics? At this age - we are now at 28 to 90 days of age - some people would still consider using something like ampicillin and cefotaxime for the first two months or so, even sometimes three months of age. Once you get past that second month of age either ceftriaxone or cefotaxime alone are considered adequate empiric therapy. Now if you think the patient might have Salmonella, you might want to add something else to that. But generally cefotaxime and ceftriaxone are appropriate therapies for empiric management of these kids.

Next algorithm. Bigger mess here, and it isn’t because the kid is a little older. Again, the first question you are going to ask yourself here is, is this a non-toxic appearing previously healthy child? If the answer is no, if they look sick, they probably belong in the hospital. And you need to do your complete substance work-up and parenteral antibiotics and the parenteral antibiotics are the same as the ones we talked about before. That is, either cefotaxime of ceftriaxone empirically. Now, if the child meets this first cut off, or first break point, and the answer is yes, then the next break point is temperature. Here we rely on the fact that the incidence of occult bacteremia and serious bacterial disease in a kid between the ages of 90 days and 36 months can really be correlated with degree of temperature. So that if you have a temperature that is less than 39C degrees, there really isn’t much to do. You don’t have to do any diagnostic tests or antibiotics, if you want. Just examine the child and make sure there is nothing that you are missing. Antipyretics should be given. Generally acetaminophen is sufficient. The patient should be instructed to return or given an appointment to return in 24 to 48 hours, or sooner if the patient is getting worse or not better. So not a lot has to be done if the patient looks good. On the other hand, if the temperature is greater than 38C degrees then you should proceed with some type of an outpatient’s work-up and that depends on what the constellation of symptoms are for each child. So in general, if the child is a male and less than six months of age, or a female less than two years of age, a urine culture should be obtained to rule out a urinary tract infection. Stool culture is there is blood or mucus in the stool, or if there’s greater than 5 WBC’s on your stool smear.

If the patient has respiratory symptoms such as dyspnea, cough, tachypnea, decreased chest sounds, you probably want to get a chest x-ray. I personally like to get a chest x-ray in kids less than 6 months of age when I’m evaluating for sepsis, even if they don’t have respiratory symptoms. Because they tend not to move a lot of air overall and sometimes you can miss these little fine crackles, especially if the kid is crying or just isn’t cooperative.

Blood culture should also be obtained. Some people like to, again, streamline or cut back on the cost involved in this and say, "Look, I won’t do the blood culture unless the temperature is greater than 39C degrees and the white count is greater than 15, 000." And that’s because we know that when the white count is greater than 15,000 we have a higher rate of bacteremia. It’s one way of being cost effective. The downside is that you are probably going to have to go back and harpoon the kid again. So keep that in mind.

Empiric antibiotic therapy again can be linked with the issue of white count and temperature. So that if the white count is greater than 15,000 and the temperature is greater than 39C degrees, some people like to give the antibiotic. I personally - if the child has a temperature greater than 39C degrees and is in this group - I usually give them the shot of ceftriaxone and bring them back. You will notice that there is no mention about the lumbar puncture here, and the reason why is that in this age group most authorities believe that there is a good enough exam to give you an idea of whether there is meningitis present or not. Keep in mind though that the Kernig’s and Brudzinski’s sign, those classic signs of meningitis, do not become fully developed until the child is about a year to 15 months of age. The reason why is that tethering of the spinal cord hasn’t occurred. So a negative Kernig’s and Brudzinski’s in a kid less than 12 months of age, 15 months of age, doesn’t rule out meningitis. In a kid older than that, yes. That’s great. You get them to look up and down and move their head, they are fine. They probably don’t have meningitis provided there are no other signs involved. Less than that, you really have to be careful so don’t use that as your sole indicator of whether there is meningitis or not. Acetaminophen for control of pain and fever and then of course, the all important follow-up in 24 to 48 hours. Bring them back and look them over.

Now, what do you do if you did all of this - you’ve got an outpatient now - and the blood cultures come back positive? Well, the old … this is being modified now. The old statement was that you admit them for a complete substance work-up and empiric parenteral antibiotic pending the culture that you are going to draw when you bring them back. Many of us now feel that if the child is afebrile and doing well when you see them, that you can follow it up with another shot of ceftriaxone until your bacteria is identified and the microbial sensitivities are known. Even if the child had a penicillin resistant pneumococcus, you wouldn’t change your therapy if the patient has become afebrile. Those are the latest AAP recommendations. You treat the patient not the bug. So if the kid has an intermediate or highly resistant pneumococcus and the kid is afebrile after a shot of ceftriaxone, he either cleared the bacteria on his own, or you helped him along a little bit with the antibiotic and he’s probably going to do fine in the long run. On the other hand, if he has a urine culture, if he has persistent fever, bring him in, work him up. If he’s afebrile and well, outpatient oral therapy; ampicillin or trimethoprim sulfa, depending on the age, are appropriate based on what you find on your cultures.