Introduction

Diarrheal diseases accounted for more than half-million deaths of children under 5 years old in 2013 [1, 2]. Most of these deaths occur in low- and middle-income countries (LMIC). In high-income countries (HIC), meanwhile, the disease is rarely fatal, but it is a leading cause of visits to the emergency department and hospitalizations [3]. The World Health Organization (WHO) defines acute diarrhea as the passage of three or more loose or liquid stools per day, for 3 or more days, and less than 14 days [4]. On the other hand, the American Academy of Pediatrics (AAP) defines acute gastroenteritis as a diarrheal disease of rapid onset, with or without additional symptoms and signs, such as nausea, vomiting, fever, or abdominal pain [5]. When literature focuses on the impact of the disease in children in LMIC, the preferred terms are “acute diarrhea” or “diarrheal disease,” whereas when literature focuses on children in HIC, the most common term is “acute gastroenteritis.” Although based on different definitions, all these terms are describing the same disease: a gastrointestinal infection caused by specific microorganisms such as rotavirus, norovirus, Salmonella, E. coli, and Campylobacter [6]. In this narrative review, we use the term “diarrhea,” but we aimed to capture the most relevant information regardless of the context (acknowledging the potential differences of the disease in children from LMIC and HIC), and therefore, we aimed to update the evidence on acute infectious diarrhea and gastroenteritis in children.

Classifications

According to the WHO, the disease can be classified in acute watery diarrhea, acute bloody diarrhea (dysentery), persistent diarrhea, and diarrhea with severe malnutrition [4]. We focus on the first two as they are the most common in either LMIC or HIC. Persistent diarrhea is a complex condition in which a diarrheal episode does not resolve appropriately and lasts more than 14 days [4]. This condition requires a separate review, and it is out of the scope of this article. Diarrhea in children with severe malnutrition also requires a different approach because these children are at high risk of severe complications (electrolytes disturbances, infections and death). These cases should be managed following specific guidelines for severe malnutrition [7].

Chronic diarrhea is another term worth mentioning. Some authors use this term interchangeably with persistent diarrhea, but most commonly is used for those diarrheal episodes that last more than 4 weeks [7]. Chronic diarrhea occurs in children with long-lasting diarrhea that is not preceded by acute onset, and as a manifestation of genetic, structural or inflammatory diseases (e.g., cystic fibrosis, inflammatory bowel disease, or short-bowel syndrome). This disease is also out of the scope of this review.

Etiology and Epidemiology

Viruses are the leading causes of diarrhea in children accounting for approximately 70–90% of cases [8, 9]. Bacteria as Shigella, Salmonella, Campylobacter and, enterotoxigenic E. coli (ETEC), and less frequently enteroinvasive E. coli (EIEC), are causative agents in 10–20% of cases [10]. Anaerobic bacteria can cause diarrhea mediated by toxins. Clostridium difficile toxins produce antibiotic-associated diarrhea (AAD) and is a leading cause of disease in hospitalized children and adults [11]. Some authors have investigated the role of Bacteroides fragilis enterotoxin in children with diarrhea in LMIC [12]. However, this association remains controversial since B. fragilis can be isolated in healthy children as well [13].

Although viruses are the leading causes in most of the cases, the relative frequency of both viruses and bacteria varies according to the context: viruses have higher frequency rates in children from HIC than in those from LMIC [14]. Parasites are present in less than 5% of cases, mainly Cryptosporidium, Giardia, and E. histolytica [15]. Similar to bacteria, parasites are more frequent in LMIC [10].

Rotavirus was the leading cause of viral diarrhea and also the main cause of hospitalizations and severe disease, but after the introduction of rotavirus vaccines (RV), rotavirus-related hospitalizations and deaths have significantly decreased [14, 16, 17]. Currently, in countries that have introduced RV, norovirus has become the leading cause of moderate to severe disease, traveler’s diarrhea and outbreaks due to food poisoning in children younger than 2 years old [8, 10, 18].

Pathogenesis

The most common modes of transmission of enteropathogens are via person-to person, by fecal–oral route, or by ingestion of contaminated food or water. Incubation periods usually range between 1 h (toxin-producing bacteria: S. aureus) and 7 days (invasive bacteria as Shigella) [19]. However, for some bacteria, incubation periods can be up to 14 days (Salmonella) and for some parasites up to weeks or months (E. histolytica) [20].

Diarrhea is secondary to excessive intestinal secretion or to impaired fluid and electrolytes absorption across the intestinal epithelium [21]. Watery diarrhea is commonly caused by cytolytic microorganisms or by toxin-producing bacteria. Viruses produce a cytolytic effect in the intestinal epithelium, inflammation, and cytokine liberation, which produce a decrease in the absorption of water, and in the disaccharides’ digestion. In secretory diarrhea (e.g., V. cholerae and ETEC), the stimulation of water and electrolytes secretion due to the activation of the adenylate cyclase, which raises the intracellular cAMP and/or cGMP, and calcium, turns epithelial cells to an active secretion of chloride state [21]. Inflammatory diarrhea is caused by microorganisms that produce cytotoxins (e.g., Shigella and STEC) or invade and disrupt the intestinal epithelium (e.g., Salmonella, Campylobacter) producing inflammation and necrosis of epithelium and microabscesses [19].

Clinical Assessment

An episode of acute watery diarrhea is characterized by loose or liquid stools, which can be accompanied by hyporexia, vomiting, fever, and abdominal pain [22]. Common causes are viruses (rotavirus, norovirus, adenovirus) or non-invasive bacteria (ETEC, non-typhoidal Salmonella and Campylobacter). Less frequently, the disease can present as dysentery: bloody diarrhea associated with high fever and toxicity. These episodes are caused by Shigella, EIEC, and some strains of Salmonella and Campylobacter. Parasitic infections can present as explosive and mucus or bloody stools, cramping, tenesmus (E. histolytica), or as a malabsorption syndrome (Giardia) [23].

The clinical interrogation should focus on determining level of stool output and the hydration status (asking about urine output, number of stools and vomiting episodes, and the amount of liquid and food intake), the characteristics of diarrhea (i.e., watery, mucus or bloody stools), the duration of symptoms, history of travel to endemic areas, past medical history (immunosuppression or comorbidities), immunizations (i.e., RV), ingestion of drugs (e.g., antibiotics, or laxatives), history of family members with similar manifestations, and on documenting the pre-illness weight [24, 25].

Physical examination should focus on the evaluation of the hydration status. The gold standard for determining the degree of dehydration is the child’s weight loss. Since, in most cases, the precise pre-illness child’s weight is unknown, we must rely on the evaluation of the dehydration signs. When analyzed individually, the best three dehydration signs are altered capillary refill time, abnormal skin turgor, and abnormal respiratory pattern [26]. However, since one single sign is not enough for detecting dehydration appropriately, dehydration scales combining two or more signs are to be used. There are three commonly used scales: the clinical dehydration scale (CDS) [27], Gorelick [28], and WHO scales [4]. Table 1 provides a comparison of the three scales and the signs used for their classification. In HIC, it may be more frequent the use of CDS or Gorelick scale, while in LMIC, the WHO scale is the preferred one. However, these scales’ accuracy has been found suboptimal [29]. The evidence from children in HIC suggests that WHO and Gorelick scales are not useful for assessing dehydration, and the CDS scale, although still not very accurate, it might be the best of the three to predict the degree of dehydration [30]. Further research on the development of new tools for assessing the severity of dehydration should be encouraged.

Table 1 Classification of the hydration status according to three different scales
Full size table

Laboratory Evaluation

Laboratory studies are not routinely required. Serum electrolytes, renal function, and acid-base status (bicarbonate levels) are only helpful when markedly abnormal, therefore, should only be ordered in cases of severe dehydration or when there is a strong suspicion of electrolytes imbalance (e.g., seizures, or arrhythmias) [24, 26, 31].

Since most diarrheal episodes are considered self-limited and, the identification of an enteropathogen does not affect the management and natural history of the disease in most of the cases, determining the specific etiology is not a priority. However, in some cases, identifying the cause could have the potential of decreasing the indiscriminate use of antibiotics, and therefore, of preventing the development of bacterial resistance and dysregulation of the microbiome [32].

Viral identification can be achieved using traditional enzyme immunoassays for rotavirus and adenovirus, which are widely available [23]. Identification of other viruses (e.g., calicivirus) can be obtained with the nucleic acid amplification test. However, viral identification is not routinely recommended. Specific bacteria detection can be performed with bacterial cultures or rapid molecular tests by rapid polymerase chain reaction (PCR). Stool cultures are costly, cover few microorganisms, and do not provide rapid results. Therefore, they are not routinely recommended [23, 33]. Yet, they should be ordered in cases of toxic appearance, sepsis, high fever (> 39 °C/102 °F), dysentery, immunosuppression, diarrhea > 7 days, or history of travel to bacterial infection high-risk areas [33,34,35].

Regarding rapid molecular tests, several commercial multiplex assays are available for bacterial identification [36]. They can identify 8 to 22 pathogens in a matter of minutes or hours, which makes them very useful for decision-making. However, these tests are expensive and not widely available, which make them not feasible in LMIC. Some guidelines recommend them in cases of clinical suspicion of enteric fever (Salmonella typhi) and bacteremia/sepsis [23]. Recent studies have found that introducing rapid tests in LMIC contexts may have an impact on targeting antibiotic therapies in children [37]. Nevertheless, there is still not enough evidence for using them routinely.

Stool microbiological examination is only indicated in cases of strong suspicion of parasitic infections, such as diarrhea > 7 days or history of travel to endemic areas [35]. Clostridium difficile toxin test should be considered when there is a history of recent antibiotic treatment (previous 8–12 weeks) to rule out AAD [38]. Tests for other anaerobic bacteria, such as B. fragilis toxin detection, are not routinely recommended [12]. Blood cultures should be obtained in infants younger than 3 months of age, suspicion of sepsis or enteric fever, and when there is a history of hemolytic anemia or immunosuppression [23]. It should be noted that a proper clinical interpretation should accompany any microbiological test because the presence of a microorganism does not necessarily mean it is the cause of the disease.

Fluids Management

The mainstay of the management of the diarrhea is the fluid replacement. The WHO defines three different management plans according to the hydration status: plans A, B, and C [4]. Plan A aims to prevent dehydration and malnutrition, and it includes giving the child more fluids than usual. Suitable fluids for this purpose are those that contain appropriate amounts of salt and glucose, such as oral rehydrated solutions (ORS), salted drinks, and vegetable or chicken soups [4]. Commercial electrolyte solutions (CES) might be an alternative as well. Inappropriate and potentially dangerous fluids are carbonated beverages, commercial fruit juices or any sweetened beverages, and beverages with laxative or stimulant effects (coffee or some teas) [4]. A general rule is to give suitable fluids after each loose stool 50–100 mL in younger than 2 years old, and 100–200 mL in older children [4].

Plan B is the recommended treatment for patients with some (mild-to-moderate) dehydration [4, 5, 23, 24]. Plan B is based on the oral rehydration therapy (ORT). Three systematic reviews have concluded that there are no significant differences between ORT and intravenous rehydration in non-severe dehydration, in terms of treatment failure, dysnatremia, total of fluid intake, and weight gain [39,40,41]. Moreover, ORT (and also nasogastric rehydration) showed a reduction of 1.2 days in the total hospital stay in comparison with intravenous therapy. ORT is not only safer and faster to initiate, but also cheaper than the intravenous therapy [42,43,44]. Therefore, ORT should always be the first line of treatment reserving intravenous for cases in which the former fails.

In plan B, fluid replacement is based on the calculated percentage of dehydration. WHO recommends 75 mL/kg orally with ORS, administered continuously until dehydration signs are lost (around 2–4 h). The recommended ORS for ORT is the low osmolarity oral rehydration solution (L-ORS). In comparison with a standard oral rehydration solution (S-ORS) (sodium, 90 mEq/L; osmolarity, 311 mOsm/L), the L-ORS (sodium, 90 mEq/L; osmolarity, 245 mOsm/L) significantly reduces the stool output and vomiting, without causing more hyponatremia [45]. There is not enough evidence to support other types of ORS (e.g., polymer-based ORS) for plan B [46].

Some evidence from children in HIC shows that half-strength apple juice might be an alternative to L-ORS for ORT in children with mild dehydration. Freedman et al. compared a half-strength apple juice (sodium, 0 mEq/L; osmolarity, 365 mOsm/L) with a CES (sodium, 45 mEq/L; osmolarity, 250 mOsm/L), finding that the former resulted in fewer treatment failures [47]. The rationale of these results relies on the salty taste, which makes many children with no dehydration or very mild to refuse S-ORS, L-ORS, and CES, and thus, a tastier solution might be helpful in these children. However, these results might not be applicable to other contexts (e.g., LMIC, or children with high sodium losses or very high fecal output), because fluid replacement with a solution without sodium may predispose the child to potentially dangerous hyponatremia. More evidence is required to support, or to discard, this alternative.

Plan C is the recommended approach for patients with severe dehydration, shock, or when there are contraindications for ORT (i.e., persistent vomiting, ileus, severe abdominal distension, unconsciousness or worsening dehydration despite ORT) [4, 5, 24, 25]. This plan requires a rapid intravenous rehydration (6 h for infants and 3 h for older children) [4]. The best solution for achieving this goal is not known. Although WHO and AAP recommend Ringer’s lactate [4, 48], other guidelines recommend 0.9% saline [24, 35]. A registered Cochrane review title is aiming to compare the effectiveness of 0.9% saline with balanced solutions (i.e., solutions with less chloride, with additional cations, such as potassium, calcium, or magnesium, and anions, such as lactate, acetate, or gluconate, which are metabolized to bicarbonate) and it will help to elucidate what is the best fluid for intravenous rehydration [49]. The use of 0.9% saline for resuscitation in comparison with balanced solutions (RL and Plasma-Lyte®) has been associated with metabolic acidosis and hyperchloremia in different conditions [50]. Until new evidence emerges, 0.9% saline should only be reserved for bolus therapy in cases of hypovolemic shock.

The length of the rehydration therapy has also been a matter of research hypovolemic shock, and when no other solution is available. Two major approaches have been used: rapid (fluid replacement in 3 to 6 h) vs. standard (fluid replacement in 12 to 24 h). Two systematic reviews that evaluated both approaches found similar results in terms of the success of rehydration, the time to discharge from the emergency department (ED), the ED re-visiting, readmission, and the correction of the electrolyte disturbances [51, 52]. Hence, rapid rehydration is the most frequently recommended approach for plan C in either LMIC [24] or HIC [4].

Hypodermoclysis or subcutaneous route has been studied as an alternative to intravenous rehydration. Two single-arm studies found that recombinant human hyaluronidase-facilitated subcutaneous (rHFSC) rehydration may be an acceptable alternative in children with difficult venous access [53, 54]. However, these studies were industry-sponsored, they lacked a control group and they had very small sample sizes. The only available RCT in children with non-severe dehydration concluded that the rHFSC was non-inferior to intravenous rehydration and was easy to use [55]. More evidence is needed before recommendations are drawn upon their use. One systematic review is currently ongoing [56].

Dietary Management

At least as important as the fluid management, is the dietary management. According to the WHO, infants and children should always continue feeding to prevent nutritional damage [4]. If the infant is breastfed, except in cases of severe dehydration, breastfeeding should always continue (even during plan B). Older children, should continue usual feeding for their age at home. Offering the child food frequently is recommended. Frequent, small portions are tolerated better than less frequent, large ones. For decades, he best time to restart feeding after rehydration was a topic of discussion. A systematic review found no differences in the need for intravenous rehydration, vomiting episodes, or the development of persistent diarrhea, when restarting feeding early after rehydration (within 12 h) in comparison with delayed feeding (after 12 h) [57]. Hence, after rehydration, all children should restart their feeding as early as possible.

Perhaps the most common area of controversy on dietary management has been the use lactose-free milk formula. An early review found a small beneficial effect on the duration of diarrhea and reduction of treatment failure with lactose-free in comparison with regular feeding (lactose-containing formula) [58]. More recent systematic reviews show a reduction in the duration of the diarrhea between 12 and 17 h. However, the quality of the evidence was low, and some subgroup analyses have suggested that the effect might be significant in children in HIC but not in children from LMIC [59, 60]. Moreover, there is evidence suggesting that feeding children with diluted milk is not only ineffective but could also be harmful [60]. Therefore, considering the low quality of evidence, the small effect and the uncertainty around its effect, there is no evidence to routinely give lactose-free milk of formula, and there is enough evidence to avoid the use of diluted milk.

Antidiarrheals

For decades, there has been a search for interventions that may reduce the stool volume and the duration of diarrhea. Different alternatives have been studied even though many guidelines do not recommend their use [4, 35]. Among the most studied interventions are the probiotics, which are defined as live microorganisms that, when administered in adequate amounts, confer a health benefit on the host [61]. Evidence has shown that probiotics do have a beneficial effect in reducing the duration of the diarrhea [62]. However, most of the synthesized evidence has several limitations (high risk of bias in studies and substantial heterogeneity), which reduces the certainty of the results. The challenge with probiotics might be related to a large variety of probiotics, i.e., different strains, combinations, and doses. Some probiotics strains might be useful, but others might not. The most studied strains for the treatment of diarrhea are Sacharomyces boulardii, Lactobacillus GG (casei strain, also called: LGG), and Lactobacillus reuteri. All of them have shown some effect in reducing diarrhea duration by 15–24 h in comparison with placebo [63,64,65]. Other strains have no effect or have not been appropriately studied to be recommended.

Smectite is a medicinal clay commonly prescribed to reduce the stool output. Although the evidence is also of low quality, smectite seems to reduce the diarrhea duration by approximately 24 h [66, 67]. Racecadotril is an antisecretory drug that inhibits intestinal enkephalinase without slowing intestinal transit [68]. Although commonly used in some countries, the evidence behind this drug is more limited. Early systematic reviews found a significant effect on the diarrhea duration [69, 70]. However, recent studies showed that racecadotril was no different from placebo [71, 72].

Zinc is a micronutrient essential for many metabolic processes in the body. Zinc deficiency is a public health problem as it has been associated with poor growth and increase in morbidity and mortality [73]. Moderate to high quality evidence shows that zinc reduces the duration of the diarrheal episode and the risk of having another episode in the following 3 months [74, 75]. Nonetheless, the vast majority of its evidence comes from LMIC (in which is widely recommended), and very few reports on its effectiveness in HIC are available [76].

Loperamide, a popular intervention for the treatment of diarrhea in adults, is an opioid receptor agonist that acts by binding to opiate receptors in the gut [77]. A systematic review by Li et al. found that is effective in reducing the diarrhea on average, 19 h in comparison with placebo [78]. However, there are serious concerns on its use in children. Several cases of severe abdominal distension, respiratory depression, serious cardiac adverse reactions, and deaths have been described with the use of loperamide in children [22]. Loperamide is contraindicated in children younger than 2 years old.

Gelatin tannate is a compound based on tannic acid suspension in a gelatin solution. Its specific mechanism of action is unknown but it seems to have astringent, anti-inflammatory, and antibacterial properties [79]. Although it is used in several European countries in adults and children, a recent systematic review concluded that its effect is no different from placebo in children [79]. Additional interventions that have been studied and have either a very small effect (lactose-free milk and yogurt), or no effect at all “(multiple micronutrients without zinc, vitamin A, prebiotics, gelatine tannate, and kaolin pectin) on the duration of the diarrhea, are not indicated [60, 79].

To summarize, most of the interventions have shown some effect in reducing the diarrhea duration. However, the evidence of differences among them (trials comparing them against each other are scarce). In a recent network meta-analysis, our group analyzed all the evidence available to determine the relative differences among the interventions. We found that in general, probiotics (LGG, S. boulardii, and L. reuteri), smectite, zinc, racecadotril, and symbiotics provide a very similar effect: a reduction between 18 and 30 h in duration, and there were no significant differences in the effect among them [60]. Nevertheless, the quality of the evidence that supports the evidence was low to very low in all the cases, except for symbiotics and zinc [60]. These two interventions seem to be, among all, the ones that, with more certainty, produce a significant reduction on the duration of the diarrhea. The evidence of symbiotics comes almost exclusively from HIC, while the evidence of zinc is almost exclusively coming from LMIC. The evidence that supports the effect of smectite, probiotics, and racecadotril has very low certainty to be routinely recommended. Table 2 summarizes the effect of all the interventions mentioned above based on the certainty of the evidence that supports them, for reducing the diarrhea.

Table 2 Summary of the effectiveness of the best interventions for reducing the diarrhea and controlling vomiting
Full size table

Finally, when deciding what adjuvant treatment to use, cost-effectiveness should also be considered. For instance, treatment with zinc has been found not only effective but also cost-effective in some LMIC [80,81,82,83]. The cost-effectiveness of the other interventions has not been extensively studied.

Antiemetics

When present, vomiting might range from one single episode to repeated, and sometimes persistent, which affects fluid intake at home and the rehydration therapy at the hospital. When vomiting is not severe, it can be easily controlled by giving fluids or L-ORS slowly in small amounts. If vomiting persists, it threatens the success of the ORT. In these cases, antiemetics are indicated. Two systematic reviews found that ondansetron is the only effective intervention for the cessation of vomiting, reducing hospitalizations, and the need for intravenous rehydration, but it may increase the diarrheal episodes [84, 85]. As a result, some guidelines recommend ondansetron in children with persistent vomiting [23, 24, 86]. In a recent network metanalysis from our group, we analyzed seven antiemetics (dexamethasone, dimenhydrinate, domperidone, granisetron, metoclopramide, and ondansetron) and concluded that ondansetron is the only intervention superior to placebo [87]. Table 2 displays the details of the effect of the studied antiemetics.

Antimicrobials

Antimicrobials are not routinely recommended. Their use is associated with the hemolytic-uremic syndrome, relapses, and persistent diarrhea [87, 88]. They are only indicated when there is a risk of serious invasive disease or complications, and there is evidence that the treatment will improve clinical outcomes, or when there is a need for stopping a microorganism spread [87].

Thus, antibiotics are always indicated in infections by Shigella and V. cholerae, and in selected cases of infections by non-typhoidal Salmonella (younger than 3 months of age or immunocompromised children), Campylobacter (in severe infections) [89], and Cryptosporidium (immunocompromised children) [90,91,92]. Quinolones (i.e., ciprofloxacin) and macrolides (i.e., azithromycin) are the best choices for empirical treatment for these microorganisms, but it will largely depend on local microbiological susceptibility [9, 93]. It should be noted that although they are the first choice for this microorganisms, quinolones should be used with caution considering that they increase the risk of arthropathy [94]. Empirical treatment for cases of bacteremia or sepsis includes third-generation cephalosporins (i.e., ceftriaxone) until cultures’ results are obtained [23]. Table 3 summarizes the antimicrobial treatments according to the most commonly identified microorganisms in diarrhea in children.

Table 3 Summary of antimicrobial treatment according to identified microorganisms
Full size table

Prevention

The best strategy to prevent diarrhea has been the implementation of safe drinking water, adequate sanitation and hygiene, and handwashing with soap. The second most important preventive intervention is the RV [16]. The introduction of the RV in the national immunization programs has reduced from 54 to 20% of the attributable incidence of rotavirus infections in Africa [10]. In general, the decline of rotavirus disease ranged from 25 to 55% [10, 17]. It is widely recommended that all infants without known contraindication should receive rotavirus immunization [23]. Additional available vaccines are not routinely recommended. Cholera and typhoid vaccines are only recommended in high-risk populations (travels to high-endemic areas, intimate exposure, or microbiologists) [23]. Combined ETEC + Shigella vaccines are under evaluation but their effectiveness is yet to be established [95].

Conclusion

Diarrhea is one of the most common diseases in childhood. Unfortunately, it is still a cause of death in LMIC and a frequent cause of visits to the emergency department in HIC. We have reviewed the main aspects of the disease and provided a summary of the most current evidence that may be helpful to clinicians that deal with children with diarrhea in any setting.