Dip in body temperatures seen in humans over the decades

A thermometer reading around 98.6 degrees Fahrenheit (37 degrees Celsius) usually means the body temperature is normal. Not any more, says Julie Parsonnet and her colleagues at Stanford University, School of Medicine.

By analysing massive 6,77,423 human body temperature records, collected during the last 157 years, Parsonnet and her colleagues found that “men born in the early 19th century had temperatures 0.59°C higher than men today” and that “temperature has also decreased in women by 0.32°C since the 1890s”. This means that the onset of fever is at 99.5°F and not at 100°F.

What is normal?

From ancient times, physicians have been using body temperature as a critical vital sign in medical diagnosis. A German physician Carl Reinhold August Wunderlich, practising at Leipzig hospital, collected data on body temperature from everyone whom he could lay his hands on. Ill or healthy, he noted down their body temperature many times a day and for several days.

The task was not just asking the patients to open their mouth and reading the thermometer. In those days, the thermometer was a foot long and took fifteen minutes to give a reading. He took the underarm temperature of about 25,000 patients several times over, a total of more than a million readings. Contending that fever is not a disease but a symptom, he introduced the fever chart into hospitals. Analysing the data, he came up with a number — 37°C or 98.6°F (range: 36.2–37.5°C or 97.2–99.5°F) as the mean average body temperature.

But why 98.6? Recent studies have suggested that a relatively stable body temperature enable smooth chemical reactions and keep the organs running besides keeping fungal infections at bay. When one is infected with germs, the inflammation of the tissues produces additional proteins called cytokines which increase the total metabolic rate, thus generating heat. Therefore, often fever is a symptom of underlying infection and inflammation.

The puzzle

Since Wunderlich’s day, doctors and researchers have measured the body temperature and found it to be waffling. It is lowest around 4 am and highest around late afternoon. Children are usually warmer than elders. Women tend to maintain a higher temperature than men, depending on where they are in their menstrual cycles. Seasons and room temperature too affect body temperature. Studies in the USA showed that Afro-Americans were slightly warmer than the fair-skinned. These were explainable, but not surprising.

Slowly and steadily, several studies emerged that baffled the medical community. Doctors working at the Veterans Affairs hospital in Baltimore took several daily temperature measurements of 148 men and women over three days in 1992. The mean normal temperature tuned out to be about 36.8°C, 0.2 degrees lower than the norm.

Evidence started pouring from the US, UK and the developed world that the mean average temperature was less than the norm set by Wunderlich 150 years ago. Slogging through medical records of 35,488 patients who visited outpatient clinics of a large academic hospital in the USA during 2009-14, Ziad Obermeyer, Jasmeet K Samra and Sendhil Mullainathan computed the mean average temperature to be 36.6°C, 0.4 degrees lower than the ”normal” 37°C.

Was Wunderlich wrong? Or, are we really becoming cooler?

Declining temperatures

Some speculations were advanced to explain the anomaly. Perhaps, the thermometers used by Wunderlich were faulty, or that the new thermometers were of better quality. Or maybe, thanks to better nutrition and healthcare, people in the developed world where these studies were conducted were taller, fatter and older than our predecessors.

To find out if the drop is real or the result of faulty measurement, Parsonnet and her co-authors amassed historical data from the American Civil War, the 1970s and the early 2000s. After the civil war, the soldiers were pensioned. They had to undergo a periodic medical examination to claim pension benefits. These medical records provided vital data of 23,710 individuals Army Veterans from 1860–1940. Between 1971 and 1975, the US government undertook a National Health and Nutrition Examination Survey in which 15,301 adults were subjected to a medical examination and their body temperature records were available. Further, the researchers had access to medical records of 578,522 adult outpatients at Stanford Health Care from 2007 to 2017.

Most medical records had the information on the sex, weight, age, time of measurement, medical condition, inflammation if any and other details. By removing records that were incomplete and sick from the list, they collectively had 677,000 temperature measurements, spanning 157 years, from 1860 to 2017 to examine. The result — not only had the average temperature fallen but it has been decreasing 0.03°C per birth decade for men and about 0.029°C per birth decade for women.

Is the decline real?

What if the earlier measurements were tainted by defective thermometers? Furthermore, previous measurements recorded temperatures measured at the armpit while the modern ones are placed inside the mouth. As the database was large enough, the researchers were able to compare the same age group across the decades.

For example, they analysed data of a 30-year-old who is around 70 kg in weight and 170 cm in height across decades. The study showed that the average body temperature was 37.07 during the 1820s. However, a similar 30-year-old in the 1830s had a temperature of 37.05 and just 37.04 in the 1840s.

Neither the method of measurement nor the quality of thermometer would have seen much change in this duration. Yet the mean average temperature declined. In like manner, when the 30-year-old in 2007 is compared with that of 30-year-old in 2017, the average mean temperature dropped by 0.02°C. Regardless of the thermometer or measuring method used, for all age cohorts, across decades the mean normal body temperature reduced.

Why is it going down?

Studies across developing nations show a declining trend of the mean body temperature. How far will it plummet? Obviously, it will not reach zero, but the lower bound is still unknown.

The population of developed nations today hardly encounter tuberculosis, syphilis, periodontal disease, dysentery, diarrhoea which were plaguing the world till a century ago. Vaccines and antibiotics have helped check infection and control the inflammation. With the reduction in the infection burden, the immune system is less active, and fewer tissues are inflamed over a lifetime.

Nevertheless, Parsonnet cautions that we need to undertake similar studies in other regions where people’s health has improved and compare it with places where still the infectious diseases are a significant health burden before we come to a firm conclusion. For example, a 2008 study determined that the average body temperature in Pakistan still hovers around 37°C/98.6°F. Perhaps, it is not time yet to set a new norm in India.

On the flip side, certain parts of the world are witnessing the resurgence of infectious diseases. Wars and conflict zones are creating an unhealthy environment. What are the implications of these changes on the physiological body function? We don’t know.

(The author is a science communicator with Vigyan Prasar, New Delhi)