Included studies

Seven studies were included for this review, 6 single-gate, cross-sectional, diagnostic test accuracy (DTA) studies (Baker 1989,Borchsenius 1991,Close 2011,Nielsen 2001,Waterfield 2021,Wells 2001), and 1 two-gate, cross-sectional, DTA study (Haj-Hassan 2011).

The included studies are summarised inTable 2.

Six studies included babies and children, and reported results for babies and children combined (Baker 1989,Close 2011,Haj-Hassan 2011,Nielsen 2001,Waterfield 2021,Wells 2001), 1 study (in addition to reporting results for babies and children) reported results for an adults-only (Close 2011), and 1 study had an undefined age range and reported results for the whole sample (Borchsenius 1991).

The signs and symptoms of meningococcal disease in babies and children reported by the studies can be categorised as follows: general signs of illness and duration of illness (Baker 1989,Haj-Hassan 2011,Waterfield 2021,Wells 2001); unusual, abnormal, or pale skin colour (Haj-Hassan 2011,Waterfield 2021); presence, and type and size, of rash (Close 2011,Haj-Hassan 2011,Nielsen 2001,Waterfield 2021,Wells 2001); distribution and duration of rash (Baker 1989,Nielsen 2001,Waterfield 2021,Wells 2001); signs or symptoms of meningism (Baker 1989,Haj-Hassan 2011,Nielsen 2001,Waterfield 2021); reduced consciousness (Close 2011,Waterfield 2021); signs of shock (Haj-Hassan 2011,Waterfield 2021,Wells 2001); limb or body pain (Haj-Hassan 2011,Waterfield 2021); cardiac and respiratory symptoms (Haj-Hassan 2011,Nielsen 2001,Waterfield 2021); gastrointestinal symptoms and refusal of food and drink (Haj-Hassan 2011,Nielsen 2001,Waterfield 2021).

One study (Close 2011) reported presence of haemorrhagic rash and reduced consciousness as potential signs/symptoms of meningococcal disease in adults.

One study (Borchsenius 1991) reported the following signs and symptoms of meningococcal disease in an undefined age range: reduced general condition; cyanosis; petechiae (≤4 mm); ecchymoses (>4 mm); neck stiffness; reduced consciousness; cold extremities; and body pain.

One study used blood and/or cerebrospinal (CSF) culture detection for Neisseria meningitidis (Baker 1989) as the reference standard; 1 study used blood and/or CSF culture and/or CSF leukocyte count (Borchsenius 1991); 1 study used culture (from blood or CSF) and/or polymerase chain reaction (PCR) for Neisseria meningitides (Waterfield 2021); 1 study used bacteria, bacterial antigen, bacterial or viral DNA or RNA identified in CSF, bacteria or viruses obtained from culture of CSF, and/or clinical/laboratory diagnosis of meningitis accompanied by microbiological evidence of pathogen from another site (for example, blood, throat swab, skin or faeces) (Close 2011); 1 study used blood and/or CSF and/or skin culture for Neisseria meningitidis and/or gram negative diplococci in CSF and/or positive PCR for meningococcal DNA from blood or CSF (Wells 2001). Two studies included both confirmed (blood and/or CSF culture detection for Neisseria meningitides) and probable (clinical diagnosis without culture confirmation) cases, although the majority (79% and 74% respectively) were confirmed through microbiological techniques (Haj-Hassan 2011,Nielsen 2001).

For the 1 two-gate study the comparison group included children presenting in primary care with minor febrile infection (Haj-Hassan 2011). For 3 (of the 6 single-gate) studies, the comparison group were those negative for meningococcal disease (Borchsenius 1991,Waterfield 2021,Wells 2001). For 2 of these studies (Waterfield 2021,Wells 2001) no further details were provided about those negative for meningococcal disease; in the remaining study (Borchsenius 1991) the negative for meningococcal disease group included people with bacterial meningitis or septicaemia with causes other than Neisseria meningitidis, other bacterial infections and viral infections. One study compared those with documented invasive bacterial disease to those with nonbacteremic disease including those with viral meningitis (Baker 1989), and 1 study compared those with a confirmed or probable diagnosis of meningococcal disease with those with no invasive bacterial disease (Nielsen 2001). For 1 study, the comparison group was those with viral meningitis (Close 2011).

Signs and symptoms were identified or reported by healthcare professionals in 6 studies (Baker 1989,Borchsenius 1991,Close 2011,Nielsen 2001,Waterfield 2021,Wells 2001), and by a non-healthcare professional (a parent) in 1 study (Haj-Hassan 2011).

See the literature search strategy inappendix B and study selection flow chart inappendix C.

Excluded studies

Studies not included in this review are listed, and reasons for their exclusion are provided inappendix I.

Signs and symptoms of meningococcal disease in babies and children

Signs and symptoms of meningococcal disease in adults

There was some evidence for the presence of a haemorrhagic rash and reduced consciousness, as very specific but not sensitive for a diagnosis of meningococcal disease (compared to viral meningitis) in adults.

Signs and symptoms of meningococcal disease in an undefined age range

There was some evidence that the following symptoms were moderately specific but not sensitive for a diagnosis of meningococcal disease in an undefined age range: reduced general condition; neck stiffness; reduced consciousness; body pain.

Cyanosis and cold extremities were both very specific, but not sensitive, signs of meningococcal disease in an undefined age range.

The presence of petechiae (lesions with a maximum diameter up to 4 mm) was both moderately specific and moderately sensitive for a diagnosis of meningococcal disease in an undefined age range. Ecchymoses (over 4 mm) were also moderately specific, but were not sensitive.

Seeappendix F for full GRADE tables.

Included studies

A single economic search was undertaken for all topics included in the scope of this guideline, but no economic studies were identified which were applicable to this review question.

The outcomes that matter most

The objective of this review was to assess the diagnostic accuracy of signs and symptoms (index tests) to determine if a person presenting in the community or to hospital has meningococcal disease. The reference standard was a confirmed diagnosis of meningococcal disease based on diagnostic laboratory tests for N. meningitides. The committee considered the impact of true positives (correctly identifying meningococcal disease and starting the appropriate management), true negatives (being able to provide reassurance that the person does not have meningococcal disease), false positives (potentially starting unnecessary treatments) and false negatives (failing to identify people that require further interventions and intensive management). The committee agreed that both sensitivity and specificity were important. Sensitivity was important as failing to identify meningococcal disease could lead to treatment being delayed until the condition worsens with potentially serious implications (including death). Specificity was important, particularly when considering signs and symptoms that might lead a clinician tostrongly suspect meningococcal disease, as the misdiagnosis of meningococcal disease would result in the initiation of inappropriate treatment.

The quality of the evidence

The quality of the evidence ranged from high to very low and evidence was typically downgraded due to risk of bias (for example, due to the index tests being interpreted with knowledge of the results of the reference standard, and potential for risk of bias and/or concerns about applicability with regards to patient selection) and imprecision (95% confidence intervals crossing decision making thresholds).

Evidence was found for: general signs of illness and duration of illness; unusual, abnormal, or pale skin colour; presence of haemorrhagic rash; type and size of rash; distribution and duration of rash; signs or symptoms of meningism; reduced consciousness; signs of shock; limb or body pain; cardiac and respiratory symptoms; gastrointestinal symptoms and refusal of food and drink.

No meta-analyses were conducted for any of the index tests due to insufficient evidence after stratifying for age, person identifying the sign/symptom (healthcare professional or non-healthcare professional) and the comparison group.

Benefits and harms

The committee noted that all the evidence was based on individual signs and symptoms and agreed that none of these signs or symptoms alone would be sufficient to make a diagnosis of meningococcal disease. The committee considered the evidence for sensitivity and specificity of the individual signs and symptoms in this review and drew on their clinical knowledge and expertise to define combinations of signs and symptoms that might increase suspicion that a person has meningococcal disease.

The committee emphasised that meningococcal disease is a life-threatening medical emergency but can be difficult to diagnose and drew on their clinical knowledge and experience to include recommendations to help reduce the chance that meningococcal disease will be missed, by raising awareness that meningococcal disease: is a rapidly evolving condition; can be difficult to distinguish from other infections with similar signs and symptoms; may be harder to detect in some age groups, for example teenagers or young adults may be less likely to appear unwell.

The committee considered evidence showing that the presence of a haemorrhagic, nonblanching rash with lesions larger than 2 mm (purpura) was moderately to highly specific and moderately sensitive, and the presence of a spreading rash was moderately specific and moderately sensitive, for a diagnosis of meningococcal disease. There was also evidence that a number of features of meningitis (including neck pain or stiffness, photophobia, and a composite clinical factor of signs or symptoms of meningism) were moderately or highly specific (but not sensitive) for a diagnosis of meningococcal disease, however, most of this evidence included a non-sepsis and non-meningitis comparison group. The committee agreed that in order to differentiate meningococcal disease from meningitis, signs or symptoms of meningism would need to be combined with a non-blanching rash to raise the index of suspicion for meningococcal disease. Based on their clinical knowledge and experience, and the evidence for the individual signs and symptoms, the committee agreed that the presence of purpura, a rapidly progressive and/or spreading non-blanching petechial or purpuric rash, and any symptom or sign of bacterial meningitis when combined with a nonblanching petechial or purpuric rash, should be considered as red flag symptoms for meningococcal disease.

Based on their clinical experience, the committee noted that sometimes rashes can be difficult to detect and recommended that clinicians look all over the body to check for a rash (including nappy areas for babies). The committee also noted that healthcare professionals may need to check the conjunctivae (the membranes lining the inside of the eyelids and covering the eyeballs) when checking for petechiae. The committee highlighted that rashes can be harder to detect on brown, black or tanned skin, and included this in the recommendation to raise awareness of the need to consider this during examination. The rapidly evolving nature of meningococcal disease also means that a rash can change from blanching to non-blanching, and based on their clinical knowledge and experience, the committee recommended that patients, parents and carers were made aware of this and asked to look out for any changes.

The committee highlighted that although non-blanching rashes are classically associated with meningococcal disease and the association is supported by the evidence in this review, not everyone with proven meningococcal disease has a rash. Based on their clinical knowledge and experience, the committee included a recommendation that absence of a rash should not be used to rule out a diagnosis of meningococcal disease.

The committee noted that while people with meningococcal disease may present in the community or to hospital with 1 or more of the red flag symptoms (presence of purpura, a rapidly spreading non-blanching petechial or purpuric rash, and/or any sign or symptom of bacterial meningitis when combined with a non-blanching petechial or purpuric rash), meningococcal disease can present in different ways (and with none of the red flag symptoms). The committee agreed to include a recommendation to raise awareness of this by highlighting that meningococcal disease can be strongly suspected based on clinical assessment even in people with none of the red flag symptoms. Based on their clinical knowledge and experience, the evidence in this review, and other relevant NICE guidance (Sepsis: recognition, diagnosis and early management andFever in under 5s: assessment and initial management) the committee identified signs and symptoms associated with serious illness that might also be indicators of meningococcal disease and included these in a table. The committee agreed based on expert consensus opinion that meningococcal disease can present with any combination of the signs and symptoms included in the table, and outside of the red flags the evidence was not clear enough to rank other signs or symptoms in order of importance.

The committee took into account the rapidly evolving nature of meningococcal disease, and that people can present with subtle signs or symptoms that might be missed if not considered in the context of the patient’s usual state. Based on their clinical knowledge and experience the committee agreed that the assessment of signs and symptoms (and risk factors) should include family member and carer reports of symptoms. For people with reduced consciousness or communication difficulties it was considered particularly important that family members or carers are asked about recent or rapid changes in symptoms.

The committee considered evidence for appearing ill, or being categorised as ill, and overall studies showed moderate specificity and moderate sensitivity for meningococcal disease. The committee agreed that appearing ill to a healthcare professional may support the diagnosis of meningococcal disease.

There was some evidence in this review that pale or unusual skin colour (including cyanosis) were moderately to highly specific, but not sensitive, signs of meningococcal disease. The committee agreed that these findings were consistent with their clinical experience that pale or unusual skin colour (including cyanosis) can be associated with meningococcal disease and they agreed to include this sign in the table but to maintain consistent terminology with the NICE Fever in under 5s guideline (pale, mottled skin or cyanosis). As with detecting the presence of a rash, the committee noted that skin changes may be difficult to see on brown, black or tanned skin, and flagged this in the notes section of the table.

There was no specific evidence that quantified the diagnostic accuracy of parental or carer concern, although some studies included in the evidence review relied on parental report of signs and symptoms. The committee took into account the lack of any definitive ‘index tests’ for meningococcal disease (none of the signs or symptoms examined were both very sensitive and very specific) and the rapidly evolving nature of the condition and based on consensus added parent or carer concern to the table of features that may support a diagnosis of meningococcal disease. This was highlighted as particularly important as changes to appearance or general signs of illness can be subtle, particularly to people that are not familiar with the patient’s usual state.

Another symptom that the committee agreed may support the diagnosis of meningococcal disease but that was also important to interpret in the context of a person’s normal function was altered mental state. The evidence reviewed showed that reduced consciousness, confusion, lethargy and drowsiness were at least moderately specific for a diagnosis of meningococcal disease, although there was more variability in the sensitivity estimates. There was also some evidence that being considered irritable or miserable was a moderately sensitive (but not specific) symptom of meningococcal disease. There was some evidence that objective measures of new or altered mental state, including assessing the level of consciousness with the Alert, Voice, Pain, Unresponsive (AVPU) scale and/or Glasgow Coma Scale (GCS), was very specific for a diagnosis of meningococcal disease. However, in other studies the method of assessing altered mental state was unclear, and the committee highlighted based on their clinical experience that changes can be subtle. Drawing on their expertise and the evidence reviewed the committee agreed that lethargy, unusual behaviour (particularly being agitated, aggressive or subdued), or altered level of consciousness or altered cognition (including confusion or delirium) can be associated with meningococcal disease. The committee also highlighted that meningococcal disease can be missed because delirium may be assumed to be due to cognitive impairment in older adults, whereas altered behaviour may be attributed to alcohol or substance misuse (rather than meningococcal disease) in young people and young adults. Based on expert clinical consensus the committee also agreed to include weak, high-pitched or continuous crying as a sign that might be associated with meningococcal disease in babies.

The committee considered evidence showing that cold hands and/or feet was a moderately to highly specific, but not sensitive, sign of meningococcal disease. Based on their clinical knowledge and experience, the committee also noted that cold extremities may be present in the early stages of the illness and agreed that this clinical feature might support the diagnosis of meningococcal disease.

There was some evidence that tachycardia (raised heart rate) was both moderately specific and moderately sensitive for a diagnosis of meningococcal disease. Based on their clinical knowledge and experience, the committee were also aware that bradycardia (slow heart rate) could be an indicator of severe illness. The committee agreed that both high age-specific heart rate and low heart rate defined as less than 60 beats per minute for babies and children under 12 years should be included in the table as signs that may support a diagnosis of meningococcal disease.

There was some evidence that hypotension (low blood pressure) was very specific, but not sensitive, for a diagnosis of meningococcal disease. Based on this evidence, and their clinical knowledge and experience, the committee agreed that low age-specific blood pressure should be included in the table as a clinical feature that might support a diagnosis of meningococcal disease.

The committee considered evidence showing that delayed capillary refill time (defined as over 2 seconds) was both a moderately specific and moderately sensitive sign of meningococcal disease. There was also evidence showing that a composite factor of signs of shock (defined as clinician-diagnosed shock, a long capillary refill time of 4 seconds or more, or hypotension) was very specific and moderately sensitive for a diagnosis of meningococcal disease. The committee agreed that a capillary refill time of 3 seconds or longer should be included in the table as a sign that might support a diagnosis of meningococcal disease.

The committee highlighted that although prolonged capillary refill time may be a more specific individual sign of dehydration, reduced urine output is commonly reported as a marker of dehydration. No evidence specific to this review was identified. However, the committee considered the evidence and recommendations in the NICE Sepsis guideline that included reduced urine output as a high to moderate risk criterion, and agreed to include this in the table as a clinical feature that might support a diagnosis of meningococcal disease.

There was some evidence that tachypnoea (raised respiratory rate) was both moderately specific and moderately sensitive for a diagnosis of meningococcal disease. There was no evidence for specific rates for different age bands, but the recommendation highlighted that it was important to use an age-specific threshold for defining raised respiratory rate. There was also some evidence that respiratory symptoms, and difficult or laboured breathing, were moderately specific but not sensitive. The NICE Fever in under 5s guideline included grunting in their risk stratification and based on their clinical experience and consideration of the evidence in that guideline, the committee agreed to include grunting as a respiratory symptom that might support the diagnosis of meningococcal disease.

There was some evidence that presence of fever was both moderately specific and moderately sensitive for a diagnosis of meningococcal disease, when the threshold was defined as a temperature of over 37.5°C and with a threshold over 38.5°C. The evidence included was from children aged up to 15 years, however, the age range, mean or median age of included participants is not reported. Based on their clinical knowledge and experience, the committee reflected that very high temperature is unusual in young children and can often be indicative of bacterial infection. The committee considered the evidence and recommendations in the NICE Fever in under 5s guideline, and agreed to include consistent thresholds for fever, with a temperature of 39°C or higher potentially supporting a diagnosis of meningococcal disease in children aged 3 to 6 months, and a temperature of 38°C or higher raising the index of suspicion for children younger than 3 months. Drawing on their clinical knowledge and experience, the committee also recommended that receipt of antipyretic treatment should be checked as it may make fever harder to identify. Based on their clinical knowledge, the committee also highlighted that hypothermia can indicate infection, and included a temperature of less than 36°C as a feature that might support the diagnosis of meningococcal disease.

There was some evidence for gastrointestinal symptoms, diarrhoea, and tummy pain as moderately specific but not sensitive for a diagnosis of meningococcal disease, and the committee agreed to include abdominal pain and diarrhoea in the recommendation.

There was some evidence for limb pain as a very specific symptom of meningococcal disease, and some evidence showing general aching to be moderately specific, neither symptom was sensitive for a diagnosis of meningococcal disease. The committee considered the evidence and recommendations in the NICE Sepsis guideline that included leg pain to indicate high to moderate risk in children with suspected sepsis. The committee noted that leg pain may be an indicator of reduced perfusion (in addition to prolonged capillary refill time and cold extremities) and agreed that leg pain should be included as a potential symptom of meningococcal disease.

Given the potentially serious implications of a delay to treatment (including death), the committee agreed based on expert clinical consensus that people with suspected meningococcal disease should be transferred to hospital as an emergency, and the hospital should be alerted and informed that an assessment by a senior clinical decision maker will be required.

The committee agreed that it was also important to provide safety netting for people returning home after clinical assessment for meningococcal disease. Based on their clinical knowledge and experience and considering the rapidly evolving nature of meningococcal disease, the committee agreed that safety netting advice should be given, and people should be asked to return for further assessment if new symptoms develop, if a rash changes from blanching to non-blanching, or if existing symptoms or signs get worse. The committee also wanted to raise awareness that although a person might not have meningococcal disease, they may have another serious condition. The committee specifically wanted to highlight other forms of sepsis, non-bacterial causes of meningitis and pneumonia, but also intracranial bleed or ischaemia that is often overlooked, as potential alternative diagnoses.

Cost effectiveness and resource use

This review question did not consider decisions between competing alternatives and therefore is not directly relevant to the tools of economic evaluation. The recommendations primarily provide advice to health care professionals on the recognition and diagnosis of bacterial meningitis rather than specific courses of action. However, the committee considered that early and correct identification of meningococcal disease was a prerequisite of cost-effective management. They also reflected that the recommendations largely reinforce current best practice and knowledge and therefore they did not believe they would have a significant resource impact.

Excluded diagnostic and prognostic studies

Excluded economic studies

No economic evidence was identified for this review.