This article offers several useful and insightful comments. First, current WHO targets will not be met, so revised guidelines would be useful.  Second the global public health community does not yet have tools available for effective dengue control. Third, measuring progress, even if it occurred, would be extremely difficult due to challenges in diagnosing and reporting dengue, and the great variability in dengue incidence by time and place. Fourth, surveillance data that report dengue cases by age would add substantial information in estimating the force of infection. Fifth, a new goal of dengue control could be the reduction of the disease’s case fatality rate. Sixth, definitions matter, as the authors illustrate with the order-of-magnitude variation in the case-fatality rate of dengue, depending on how a “case” is defined. 

As a generally favorable reviewer, I would nevertheless offer several suggested refinements for the authors’ consideration. Where the authors wrote that “Fewer than 1% of infections result in death,” they may wish to make a more precise statement, “Far fewer than 1% of infections result in death.”  The true rate is likely on the order of magnitude of 0.01%. The authors’ statement that the reason for variability is often unknown seems to be counter to the rest of the sentence, as the authors suggest several plausible explanations. The authors may wish to reword the sentence to indicate that “Reasons for variability, incompletely understood, include ….”

Another detail relates to dengue in India, where the authors wrote “…very few cases are detected by surveillance systems.” As dengue’s surveillance system collects data only from a limited number sentinel facilities, the authors may wish to change “detected” to “reported.”

The authors’ laudable suggestion of a goal of reducing the case-fatality rate might be made more practical by basing it on statistics reported by most national surveillance systems. This could be the fatality rate of hospitalized lab confirmed dengue cases. As national surveillance systems vary about their inclusion of cases from outpatient settings and testing, this definition would be practical and improve comparability.

In the discussion of misclassification of possible dengue fatalities, the authors may also wish to note that misclassification could not only increase apparent deaths, as the authors hypothesize, but could also lower them. Follow up of autopsies from Puerto Rico of patients with post-mortem confirmations of dengue found that a substantial portion had not been diagnosed during their lifetime. Thus, routine dengue surveillance did not count these fatalities.

As innovations in dengue control are rapidly involving, the authors may wish to incorporate the latest data. A new modeling study of the latest dengue vaccine (Coudeville, Baurin and Shepard, Vaccine ¹ ) showed that it would be cost-effective in a much broader range of settings, including median values from vaccine trial sites where the sero-prevalence of dengue was at least 30%. Global observational data on Wolbachia show consistent benefits, with the first randomized trial (in Yogyakarta, Indonesia) expected to be completed by the end of 2020 (Katie Anders, presented at annual meeting of Am Soc Trop Med & Hygiene, 2019).

Finally, the authors may wish to add some nuances to their goal for vector control being implemented in all countries with transmission. They may wish to limit their goal geographically for large countries such as China and the United States, where dengue affects only the warmest parts, note that some vector control technologies are most applicable to densely populated areas, and replace “all” by a quantitative goal, such as 80%.

Does the article adequately reference differing views and opinions?

Partly

Are all factual statements correct, and are statements and arguments made adequately supported by citations?

Partly

Is the Open Letter written in accessible language?

Yes

Where applicable, are recommendations and next steps explained clearly for others to follow?

Yes

Is the rationale for the Open Letter provided in sufficient detail?

Yes

Reviewer Expertise:

Health economics, particularly cost of illness, cost-effectiveness analysis, and health financing, particularly for vector-borne illnesses and vulnerable populations

I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.

PROGRAMS TO REDUCE DENGUE MORTALITY. 

Scott B. Halstead, M.D.

                The reviewer agrees a 2030 WHO goal should be to reduce dengue mortality and morbidity. The methods described here cannot achieve the specific reduction goals listed.

These are comments by a disease modeling group who have suggested sustainable WHO 2030 targets for dengue morbidity and mortality rates. A note must be made at the outset. A case fatality rate of less than 0.05% already has been achieved by some large countries, Brazil and Mexico, in the American region ¹ . PAHO calculates CFR to the base of all reported cases of dengue. By contrast, in Southeast Asia, Thailand and Vietnam, countries that are widely lauded, have maintained a case fatality rate around 0.1% for a period of 10 years ² . By contrast, for a recent 5 year period, average CFRs in other countries were: Indonesia -  0.985%, India – 0.26%, Sri Lanka –  0.195%, Philippines – 0.49%, Cambodia – 0 .27% and Malaysia – 0.22% ³ . CFRs are more meaningful if calculated among clinically congruent patients. When the first country reports were listed, CFRs in Southeast Asia were calculated among hospitalized cases. In the Americas, if the denominator used is severe cases, the average CFR for 2019 in the region was 4.9% ¹ . Standardizing the denominator will be the first problem to be solved. “Severe” dengue is the right idea. Sanofi defined “severe” as hospitalized dengue cases with a platelet count of <100,000/ mm3 and evidence of vascular permeability ⁴ . But, here is the problem: the 2009 WHO case definition says that “severe dengue” is any of the following: plasma leakage, circulatory compromise, significant bleeding, altered level of consciousness, severe gastrointestinal involvement, or severe impairment of liver, kidney or heart ⁵ . My suggestion: use the 1997 WHO case definition without any requirement for hemorrhage.  The reason for this suggestion is reviewed in a recent history ⁶ . 

The authors note at the outset that epidemiological patterns and intensity of dengue disease have varied greatly over time and between geographic locales. Dengue now is global. At the end of WWII, all four dengue viruses (DENV) became endemic in Southeast Asia. High transmission rates there led over several decades to a stable pattern of seasonal outbreaks of moderate to severe hospitalized disease. Outbreaks varied in size from year to year and impacted mainly children. Economic and demographic changes then lowered dengue virus transmission rates resulting in an expansion of the age range of those affected. As new DENVs were inserted into the Indian subcontinent and then into the Western Hemisphere they produced introductory virgin soil epidemics of dengue fever and then as multiple viruses became established, disease patterns moved toward the epidemiological and clinical features of Southeast Asia. The take home message: everywhere in its range, dengue viruses produce a wide age range of diseases that vary from mild to severe causing modest case fatal rates. Despite the best modeling efforts, dengue epidemiology remains essentially chaotic. This means that surveillance, per se, is not a useful tool for managing the control of transmission or anticipating and managing clinical burden.

The key dengue challenges during 2020 to 2030 are to reduce morbidity and mortality. But, how? This review does not discuss methods in any detail. There is the very real possibility that one or both new tetravalent dengue vaccines will provide significant levels of protection across a wide age range and without vaccine sensitization of seronegatives. Phase 3 reports summarizing 18 months post vaccination data for the Takeda TDV vaccine have been published ⁷ . These show the vaccine to be highly effective at reducing hospitalization of DENV 1 and 2 infections. The vaccine contains a real live-attenuated DENV 2. Is the protection observed attributable to vaccine or an example of the heterologous cross-protection that follows a primary dengue infection ⁸ ? If so, it might “wear off” resulting in vaccine related sensitizations. Phase 2 challenge studies of the NIH TV003 show a very high rate of protection against DENV 2 infection ⁹ . Should this protection be confirmed by phase 3 studies, the entire dengue control landscape may change. The authors are correct to surmise that the NIH vaccine, even if highly protective, is unlikely to be available at the quantity needed to immunize billions of people during the next decade. Exciting large field trials introducing Wolbachia into urban areas, in progress, may provide a method to achieve a major reduction in transmission of dengue viruses by Aedes aegypti ^{10 , 11 , 12} . In addition, there are several other technologies that promise to reduce the efficiency of dengue virus transmission by vector mosquitoes ^{10 , 13} .

I agree with the authors that the most important goal for 2020 – 2030 is to reduce dengue mortality. This cannot be accomplished through enhanced surveillance or improved predictions of disease intensity. Dengue incidence will be unpredictable at national and local levels. What is needed is an entire health system that is prepared and trained to provide optimal clinical care. Dengue illnesses are short and fast moving. Classical dengue shock syndrome occurs suddenly and rather late in illness, three to five days after onset of fever. Physiological status can change in a matter of minutes. WHO has issued comprehensive technical guides that carefully describe dengue illness, diagnosis and treatment regimens ⁵ . This important information may have helped reduce dengue fatality rates. Unfortunately, good information does not by itself change medical practice. Dengue is managed successfully at large scale by countries such as Thailand and Vietnam. They have trained all the components of a system that swiftly and accurately identifies patients with acute capillary permeability and then provides them the correct sustenance.

Reasons for many fatal outcomes in dengue are well understood: 1) the patient is delayed in presenting for health care, 2) misdiagnosis and 3) mismanagement. It is widely known, but seldom discussed, that a significant percent of all fatal outcomes in dengue are due to fluid overload. Countries that have successfully lowered case fatality rates have addressed all three risks. The general public must be informed again and again about the risk of fatal dengue, told to seek medical attention early in illness and be aware of key warning signs. In the clinic, diagnosis and management of dengue requires a mix of education and training. The pathophysiology of dengue disease should be taught at all levels in medical education. Training in diagnosis and treatment must be included in relevant post-graduate medical training. Hospital based clinicians in dengue endemic countries should receive special training in dengue diagnosis and resuscitation. As with any skill, repetition and practice are necessary to achieve and sustain competence. Some countries with low case fatality rates have introduced obligatory short training courses for all hospitalists. An issue to be addressed is whether hospitalists should be credentialed in dengue treatment. A key component of a nation-wide system of sustained excellent critical care of dengue is that any health care provider who has a patient with management problems should have access 24/7 to expert consultant advice. Dengue care units must have the equipment, sonographic or x-ray, needed to identify and quantitate invisible fluid losses into serosal spaces. Hospitals in Southeast Asian countries place microhematocrit centrifuges directly on patient wards permitting ward personnel to measure hematocrit levels at frequent intervals. Hematocrit monitoring has proven to be very valuable for accurate regulation of fluid resuscitation and avoiding fluid overload.

Any change in medical practice requires planning and leadership. While WHO clearly should play a role in design and support of national programs to reduce fatal dengue, this effort will depend upon extraordinary financial assistance and leadership. To begin the effort, the different successful country approaches must be studied and translated into the appropriate diversity of educational and training programs needed. If a program is created under WHO aegis, a monitoring and advisory entity such as the Task Force for Child Survival will be needed. It is understood that every country, and for that matter, many constituencies within countries, will pose challenges, their solutions requiring high competencies and diplomacy.

Some may say that all fatal cases of dengue are not caused by capillary leakage. That is true. But, how many? Dengue viruses infect a broad spectrum of populations that differ in many ways. This diversity is relevant both to the spectrum of dengue disease and to the possibilities of successful intervention. Clearly, dengue infections in the old and infirm are associated with increased incidence of fatal outcome. Thus far, little is known about the pathophysiology that contributes to death in these cases. In some instances, when looked for carefully, capillary leakage has contributed to disease severity. Resuscitation of fluid losses has reduced case fatality rates. But, not to zero. Individuals of any age who present with the dengue vascular permeability syndrome, seen early enough, with appropriate care, should survive. The program recommended here will be directed at improving management of each of the three steps needed to correctly identify and treat patients with dengue capillary leakage. Once fully implemented we may begin to identify the scale and nature of the other factors that contribute to fatal outcome in dengue.

Does the article adequately reference differing views and opinions?

Partly

Are all factual statements correct, and are statements and arguments made adequately supported by citations?

Partly

Is the Open Letter written in accessible language?

Yes

Where applicable, are recommendations and next steps explained clearly for others to follow?

Partly

Is the rationale for the Open Letter provided in sufficient detail?

Partly

Reviewer Expertise:

dengue pathogenesis, vaccine development

I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.