Why pneumonia needs a severity score at the front door
Pneumonia remains one of the most common reasons for acute hospital admission, and it is also one of the hardest infections to judge on first impression. Two patients can walk into the emergency department with similar coughs and similar chest X-rays, yet one will recover at home within a week while the other is in the intensive care unit within 48 hours. The difference often lies in features that are not obvious from the lungs alone: an older patient with low blood pressure and new confusion has a body that is struggling systemically, even if the chest sounds are unremarkable. Clinical judgment alone misclassifies a worrying number of these patients, sending some home who should have been admitted and admitting some who could safely have been managed at home with early antibiotics and review.
A severity score turns that judgment into something structured and reproducible. The CURB-65 score asks five simple questions, each answerable at the bedside or from a routine blood test, and converts them into an estimate of 30-day mortality. That mortality estimate then maps directly onto a decision about the level of care: home, short inpatient stay, hospital ward, or intensive care. It does not replace the clinician, but it anchors the admission decision in evidence rather than gut feeling, and it gives the clinician a defensible, guideline-backed reason for whatever is decided. Guidelines from the British Thoracic Society and NICE endorse CURB-65 as the mortality risk tool for adults hospitalised with community-acquired pneumonia, which is why it is one of the most used pneumonia scores worldwide.
What CURB-65 is
The score was developed by Lim and colleagues and published in 2003 in Thorax, in an international derivation and validation study of adults presenting to hospital with community-acquired pneumonia (Lim WS, van der Eerden MM, Laing R, et al. Defining community acquired pneumonia severity on presentation to hospital: an international derivation and validation study. Thorax. 2003;58(5):377-382). The team identified five clinical features that independently predicted death within 30 days and combined them into an additive score. Each feature scores one point, giving a total from 0 to 5. The five letters of the acronym spell out the features: Confusion, Urea, Respiratory rate, Blood pressure, and 65, the age threshold.
The derivation study followed patients through their illness and recorded who survived and who did not, then checked that the score behaved the same way in a separate validation cohort. The mortality associated with each score rose steeply and consistently: 0.6% for a score of 0, 2.7% for 1, 6.8% for 2, 14.0% for 3, and 27.8% for scores of 4 and 5. Those numbers are the heart of the score, because they translate a handful of bedside observations into a concrete probability that both clinician and patient can understand. Later validation studies in different countries and healthcare systems reproduced the same risk gradient, which is why the score earned its place in guidelines.
The five criteria explained
- Confusion: +1 point. New disorientation in time, place or person, or an abbreviated mental test score of 8 or below. In pneumonia, confusion is rarely about the brain directly; it usually reflects systemic illness, low oxygen delivery, dehydration, or fever delirium in a vulnerable patient. New confusion in an acutely unwell older adult is an ominous sign and one of the strongest reasons to admit.
- Urea above 7 mmol/L: +1 point. Blood urea nitrogen rises when the kidneys are underperfused, which happens in dehydration and in the circulatory stress of severe infection. An elevated urea therefore acts as a blood marker of systemic involvement. Note the units: the threshold is 7 mmol/L, and clinicians working with blood urea nitrogen in mg/dL should use the equivalent value of roughly 20 mg/dL.
- Respiratory rate 30 breaths/min or more: +1 point. A fast breathing rate is the earliest and most sensitive sign that gas exchange is failing or that the body is compensating for acidosis. It is also the sign most often missed, because busy clinicians frequently estimate rather than count. Measure it over a full 30 seconds and double it.
- Blood pressure systolic below 90 mmHg or diastolic 60 mmHg or below: +1 point. Either low systolic or low diastolic pressure counts; you do not need both. Hypotension in pneumonia suggests the infection is progressing toward sepsis, with vasodilation and poor perfusion of vital organs. A single low reading should be repeated and confirmed, but a confirmed low reading in a pneumonic patient is a red flag.
- Age 65 years or older: +1 point. Age is the only non-modifiable item, and it reflects the undeniable reality that pneumonia kills older adults far more often than younger ones. Immune reserve, muscle strength for coughing, and the ability to tolerate dehydration and hypotension all decline with age.
How to use this calculator
Enter the patient’s age, the urea result, the measured respiratory rate, and the systolic and diastolic blood pressures, then answer the confusion question Yes or No. The calculator needs every field: a CURB-65 with a missing item is not a CURB-65 at all, so any blank will be flagged and nothing will be scored until it is filled. Measure the respiratory rate properly rather than estimating it, confirm a low blood pressure with a repeat reading, and apply the confusion item only to new disorientation, not to long-standing cognitive impairment.
Press Calculate to see the total score from 0 to 5, the individual items that scored points, the 30-day mortality from the derivation study, and the recommended level of care. The result renders as cards with a plain-language next step, suitable for documentation in the notes or for explaining the decision to the patient and family.
Interpreting the result: scores 0 and 1
A score of 0 carries a 30-day mortality of about 0.6%, and a score of 1 carries about 2.7%. These are genuinely low risks, and the guideline recommendation for both is to consider outpatient treatment, provided the patient can take oral medication, is not at social or clinical risk of deteriorating unseen, and has a clear plan for review. Outpatient care is not no care: it means prompt oral antibiotics (commonly amoxicillin first line in many guidelines for low-severity community-acquired pneumonia), fluids, rest, and instructions about what should bring the patient back, typically worsening breathlessness, new confusion, inability to keep fluids down, or no improvement within 48 to 72 hours.
Even at these low scores, the recommendation is conditional rather than automatic. A score of 1 in a frail patient who lives alone and cannot get to a pharmacy is a different proposition from the same score in a robust 60-year-old with family support. CURB-65 estimates the average mortality for the score band; it does not see the patient’s home situation, oxygen saturation, comorbidities, or ability to follow up. Use the low score as a reason to consider home treatment, not as an order to discharge.
Interpreting the result: score 2
A score of 2 carries a 30-day mortality of about 6.8%, which is high enough that routine home management is no longer the default. The recommendation is a short inpatient stay or closely supervised outpatient care. In practice, this means the patient is admitted for a brief period of observation, intravenous or closely monitored oral antibiotics, fluids, and repeat assessment, with discharge once the trajectory is clearly improving. Where local services allow, closely supervised outpatient care, such as hospital-at-home with daily review, is the alternative, but casual outpatient care without supervision is not appropriate at this level of risk.
The score-2 group is the largest source of genuine clinical debate, because these patients sit exactly on the threshold where the right decision depends on the whole picture. Oxygen saturation, comorbid heart failure or chronic kidney disease, the trend of the observations, and the patient’s wishes all legitimately move the decision one way or the other. Document why the chosen path was chosen; the score supports the decision, it does not make it.
Interpreting the result: scores 3, 4 and 5
A score of 3 carries a 30-day mortality of about 14%, and the recommendation is unambiguous: admit to hospital. At this level of risk the patient needs inpatient antibiotics, monitoring, fluids, oxygen if hypoxic, and senior medical review. A score of 3 in a patient who looked deceptively well is exactly the situation the score exists for, and it should override any temptation to send the patient home because they seem fine.
Scores of 4 and 5 carry a 30-day mortality of about 27.8%, and the recommendation adds consideration of intensive care. Not every patient with a score of 4 needs the ICU, but every such patient needs the question asked: is this patient deteriorating toward respiratory failure, septic shock, or multi-organ involvement? Early critical care review, arterial blood gas analysis, and assessment for complications such as parapneumonic effusion or empyema are all appropriate. For frail patients or those with advance care plans, the intensive care question is handled with honesty and sensitivity, and the score helps frame that conversation with the family.
Accuracy and evidence
The performance of CURB-65 has been examined in many validation cohorts beyond the original 2003 study. Meta-analyses consistently find that it discriminates mortality well, with areas under the receiver operating characteristic curve typically around 0.80, meaning the score correctly ranks a randomly chosen patient who dies above one who survives about four times in five. It performs similarly in different countries, in emergency department and ward populations, and across different aetiologies of pneumonia, including the common bacterial causes. That consistency across settings is the main reason guidelines adopted it.
The score’s particular strength is identifying low-risk patients: scores of 0 and 1 have high negative predictive value for death, which is what makes them useful for admission-avoidance decisions. Its weakness is the mirror image. CURB-65 predicts mortality, not the need for intensive care or mechanical ventilation, and a young patient with severe hypoxaemia can have a low CURB-65 because age contributes a point they cannot score and confusion and hypotension may not yet be present. Scores designed for ICU prediction, such as SMART-COP or the IDSA/ATS minor criteria, were built precisely for that gap. Where the question is whether this patient will need the ventilator rather than whether they will die, use the right tool for the question.
The primary-care variant, CRB-65, drops the urea item and is scored from 0 to 4. It is intended for settings where blood results are not available at the point of decision, such as a GP surgery or a home visit. Its risk bands behave similarly to CURB-65 at the low end, and it is the version recommended by some primary care guidelines for the admit-or-not decision in the community.
Limitations
- It predicts 30-day mortality, not everything that matters. The need for intensive care, length of stay, complications such as empyema, and long-term functional recovery are not what the score was built to predict. A low score does not guarantee an uncomplicated course.
- It can underestimate severity in young patients. A 40-year-old with severe pneumonia cannot score the age point, and may have normal blood pressure and no confusion despite dangerous hypoxaemia. Always interpret the score alongside oxygen saturation and the overall clinical picture.
- Confusion must be new. Applying the confusion point to a patient with stable long-standing dementia overestimates severity. The item refers to new disorientation or a newly low mental test score in the context of acute illness.
- Urea units matter. The threshold is 7 mmol/L. Using mg/dL without conversion, or using a non-fasting sample without noting hydration status, can misclassify this item.
- It was derived in patients presenting to hospital. Applying CURB-65 to nursing-home residents, hospital-acquired pneumonia, or immunocompromised patients goes beyond the evidence base, and different risk tools or senior judgment should guide those decisions.
- No score replaces clinical judgment. A rising respiratory rate, worsening hypoxia, or a patient who simply looks severely unwell should prompt escalation whatever the score says. The score is an anchor, not a ceiling on concern.