Using animals to model human anatomy and physiology dates back to 600 BC (Ericsson et al., 2013[25]). Animal experiments help us to understand human biology (Bahadoran et al., 2020[7]) and are important for understanding the pathophysiological and therapeutic basis of human diseases (Flórez-Vargas et al., 2016[28]). The importance of animal research in biomedical sciences is evident when we know that about 90 % of Nobel Prizes in Physiology or Medicine have been related to research done on animals (Pasquali, 2018[79]). It has been estimated that more than 115 million animals were used for research purposes in 2005 (Taylor et al., 2008[111]), and the number is increasing (Goodman et al., 2015[33]; Hudson-Shore, 2016[46]). Rodents are the most common animals used in animal experimentations (Kilkenny et al., 2009[51]; Hatton et al., 2015[36]; Jackson et al., 2017[49]), constituting about 80 % of experimental animals (Sengupta, 2013[99]). Among laboratory animals, rats are extensively used in different medical disciplines (Clause, 1993[18]; Gille et al., 1994-1996[32]), particularly in toxicology (Beckman and Feuston, 2003[9]; Vidal, 2017[115]), obesity (Reed et al., 2011[92]), social stress experiments (Buwalda et al., 2011[16]), osteoporosis (Yousefzadeh et al., 2020[119]), diabetes (Lane, 1997[57]), and neurobiology (Romijn et al., 1991[95]). Of 51 known species (Andreollo et al., 2012[6]) and more than 1000 known rat strains (Reed et al., 2011[92]), Wistar and Sprague-Dawley rats are mostly used in animal studies (Andreollo et al., 2012[6]).

Some evidence strongly suggests that animal studies are mostly not translated to humans (Pound et al., 2004[87]; Akhtar, 2015[2]), with more than 80 % of the reported safe and effective treatments in animal studies failing to translate to humans (Perrin, 2014[81]). In addition to the intrinsic limitations of animal models, poor design and reporting of animal studies are major causes of poor concordance between preclinical and clinical outcomes (Perrin, 2014[81]; Bahadoran et al., 2020[7]). This has led to a reproducibility crisis in biomedical research (Osborne et al., 2018[76]), particularly in preclinical research using animal models (Collins and Tabak, 2014[19]). The provision of basic variables including age and body weight of animal used is the starting point of enabling replication (NRC, 2011[72]). In addition, considering the difference between rodents and human timescales helps translate experimental treatment into clinical practice (Agoston, 2017[1]).

Body weight, age, and developmental stage of the animals used are characteristics that can affect the study's results (Kilkenny et al., 2009[51]; Jackson et al., 2017[49]). The age and the body weight of the animals can affect drug metabolism, gene expression, metabolic parameters, and other dependent variables measured in animal studies (McCutcheon and Marinelli, 2009[65]; Ihedioha et al., 2013[47]; Jackson et al., 2017[49]). The animal's age plays an important role in modeling human diseases; examples across different fields are presented here. If the study duration in rodents takes ≥ 3 months, results can be affected by reproductive changes (Vidal, 2017[115]). In addition, rats aged 6-9 months are most suitable for studying osteoporosis because of a stable level of bone turnover; however, rats aged under 6 months or over 9 months are not ideal because of the high bone growth rate and ageing process respectively (Yousefzadeh et al., 2020[119]). In rat models of diabetes, β-cell mass increases almost linearly from late fetal to postnatal day (PND) 100 in normal Sprague-Dawley rats (Finegood et al., 1995[26]), and sensitivity to the diabetogenic effects of streptozotocin is inversely related to age in the Wistar rats (Masiello et al., 1979[63]). Age-related differences in response to alcohol have also been reported; compared to adults, aged rats and humans are more sensitive to alcohol-induced motor and cognitive impairments (Squeglia et al., 2014[104]). Overlooking the age of the animals in design and reporting of animal studies is also a significant factor in failure in translation from animals to humans in neurological disorders (Sun et al., 2020[108]).

According to guidelines for reporting animal studies, animals' body weight and age need to be reported (Hooijmans et al., 2010[42]; Osborne et al., 2018[76]; Percie du Sert et al., 2020[80]; Nagendrababu et al., 2021[69]). However, despite being available, these data frequently are not reported (Kilkenny et al., 2009[51]; Jackson et al., 2017[49]), which can cause failure to translate animal data to the human (Ioannidis, 2012[48]). Of 271 papers reporting experimental results on mice, rats, and non-human primates, age and body weights of the animals had not been reported in 57 %, and 54 % of papers, respectively, and 24 % of papers reported neither the age nor the body weight (Kilkenny et al., 2009[51]); of 113 studies conducted in rats, 67.3 % did not report age, and 16.8 % did not report the body weights (Kilkenny et al., 2009[51]). Results of a text mining of over 15311 articles that used mice indicate that 38.6 % of the papers did not note the age of the animals used. This bias varied across biomedical fields, with missing information about the age being 20.9 % in diabetes mellitus, 29.7 % in neurological disorders, 30.23 % in cardiovascular diseases, 33.7 % in infectious diseases, 35.0 % in lung diseases, and 37.2 % in cancer studies (Flórez-Vargas et al., 2016[28]). This reporting bias, i.e., the absence of essential results from the study (O'Connor and Sargeant, 2014[74]), is mainly due to the fact that the potential values and effects of these ancillary variables on study outcomes are not recognized (Gaines Das, 2002[29]; Flórez-Vargas et al., 2016[28]). This paper reviews age-related postnatal developmental stages in rats addressing age-related changes in their body weights to provide a basis for better design and report of animal studies and help explain acquired data.

Rats are born after 21-22 days of gestation period (Romijn et al., 1991[95]; Picut and Ziejewski, 2018[86]). Rat is an altricial species (Henning, 1981[38]), i.e., they are delivered in a very immature condition. It has been suggested that at PND 12-13, the rat neocortex is developmentally comparable with a newborn human (Romijn et al., 1991[95]). Therefore, it is said that rats are born at PND 7 (Nuñez et al., 2003[73]) or PND 12 (Quinn, 2005[89]). In fact, PND 1-10 in rats is comparable with 23-40 gestational weeks in humans (Semple et al., 2013[98]). Postnatal development in rats can be considered from different points of view, including sexual maturation and nutritional behavior.

Terminology to describe various developmental milestones in rodents is not consistent across publications (Jackson et al., 2017[49]; Picut and Ziejewski, 2018[86]). Male Wistar rats at PND 28 (Kosaka et al., 1987[53]) or 42 (Cunha et al., 2001[20]), male Sprague Dawley rats at PND 36-37 (Ku et al., 2016[54]), male Fisher rats at PND 30 (Swamy and Abraham, 1987[109]) or 45-60 (Delp et al., 1998[23]), and male WKY rats at PND 35 (Silva et al., 2011[100]) have been considered to be juvenil. 9-week old male Wistar rats (Kosaka et al., 1987[53]), 8-week old male Sprague Dawley (Ku et al., 2016[54]), 4-month old male Fisher rats (Swamy and Abraham, 1987[109]), 2-month old female Wistar rats (Pestronk et al., 1980[82]), 6-week old female Sprague Dawley rats (Meyer et al., 2006[66]), and 3-6-month-old Fisher 344 rats (Delp et al., 1998[23]; Stanley and Shetty, 2004[105]; Rao et al., 2005[90]), have been considered young adults. Furthermore, 7-11-month-old Fisher 344 rats (Stanley and Shetty, 2004[105]), 12-month old rats (Hoyer, 1985[43]), 6-month female Sprague Dawley (Meyer et al., 2006[66]), and Wistar rats at ages 21 weeks (~5 months) and beyond (Wang et al., 2004[116]) have been considered as adult rats. Wistar rats at age 24 months (Hoyer, 1985[43]; Hoyer and Betz, 1988[44]) and 18-24 months (Cunha et al., 2001[20]), Sprague Dawley rats at age 12 months (Meyer et al., 2006[66]), and Fisher rats at age 24 months (Delp et al., 1998[23]; Simon et al., 2010[101]), 22 months (Stanley and Shetty, 2004[105]; Rao et al., 2005[90]), and 23-28 months (Swamy and Abraham, 1987[109]) have been considered to be aged.

These data highlight the inconsistencies in reporting the developmental stages of rats. The systematic review in neuroscience research indicated that 42 % of studies defined animals as “adults,” but the papers indicate that the animal's age was not within the adult range (McCutcheon and Marinelli, 2009[65]). A suggestion is that the exact age of the studied animals should be reported (Jackson et al., 2017[49]). According to a survey collected data from researchers that use rodents to model human disease or physiology, rats and mice were primarily used at 8-12 weeks of age and were considered to be adults, and were used as such regardless of the biology being studied (Jackson et al., 2017[49]). The definition of an adult was mainly related to the sexual maturity of rodents, not the development of the system under examination (Jackson et al., 2017[49]). This range (8-12 weeks) encompasses ongoing development in some systems, e.g., brain development (McCutcheon and Marinelli, 2009[65]), which affects the outcome of experiments and can lead to misinterpretation of the data obtained (Jackson et al., 2017[49]). For example, most humans with sepsis are over the age of 50, whereas most mice used in sepsis research are < 3 months old, this mismatch causes misinterpretation of the data obtained as the immune response to infection is age-dependent (Fink, 2014[27]; Starr and Saito, 2014[106]). In addition, even though many neurological disorders affect the elderly, most studies have used young adult animals, with only 2.2-10.5 % of 10,3269 rodents used in neurological disorder studies included aged rodents (Sun et al., 2020[108]). A systematic review of animal experiments in neuroscience research indicates that 75 % of studies used young animals, 20 % used adult animals, and 5 % did not specify animal age, indicating bias towards using young animals (McCutcheon and Marinelli, 2009[65]).

The Wistar and Sprague Dawley rats' birth weight ranges from 5-7 g (Gille et al., 1994-1996[32]; Tinwell et al., 2002[112]; Alberts, 2005[3]; Sengupta, 2013[99]; Santiago et al., 2015[96]). Body weight growth in rats has two stages: (1) development to maturity in which growth rate is high, and all parts of the body grow, (2) post-maturity growth in which the growth rate is lower than the previous stage (Pahl, 1969[78]). Growth duration in male and female Wistar rats has been reported to be 13.5±0.4 months and 19.3±0.5 months, respectively (Goodrick, 1980[34]).

There is no standard guideline for reporting animal body weights in the literature, and the time interval for data analysis depends on the researcher's decision. It has been proposed that rats have a growth phase from birth until the end of the 14^th week and, after that, have a maintenance phase in which the growth rate is lower (Hoffman et al., 2002[40]). It has been suggested that during the growth phase, body weights should be measured every week and during the maintenance phase every two weeks (Hoffman et al., 2002[40]). It would be better to statistically analyze and report the data on body weights weekly for the first four weeks of the animal's age (n=4), then every two weeks at weeks 5, 7, 9, 11, and 13 from a three-week moving average (n=5); for example, the three-week moving average at week 5 is the average of week 5, week 4 (one week before), and week 6 (one week after). During the maintenance phase, every four weeks at weeks 16, 20, and 24 from a 5-week moving average (n=3), followed by every 14 weeks at weeks 33, 47, 61, 75, 89 from a 15-week moving average (n=5) and the last point is the midpoint from week 96 to the end of the study (total n=18) (Hoffman et al., 2002[40]). For statistical analysis, one repeated measure ANOVA is used for each growth phase; this approach decreases the number of comparisons and, therefore chance of making false-positive claims (Hoffman et al., 2002[40]). However, using the moving average as a smoothing data technique has been criticized as it can produce spurious signals that look real (http://wmbriggs.com/post/195/) and are not suitable for statistical analyses (https://www.graphpad.com/guides/prism/latest/curve-fitting/reg_dont_fit_a_model_to_smoothed_d.htm).

Another problem with reporting data on animal body weight is that in most studies, particularly when there is repeated measure data, body weights are reported in the article's Results section as a graph. Although it seems the right and appropriate way of reporting the data, it precludes it from entering secondary analysis such as meta-analyses that are informative tools for translating basic sciences into clinical practice (Bahadoran et al., 2020[7]). There are simple ways, such as using Adobe Photoshop software for extracting data from graphs (Gheibi et al., 2019[31]), and an alternative suggestion is to provide such data in a Supplementary Table. It has also been suggested that if body weight is not an intended outcome of the study, at least the animals' initial and final body weights are reported. Also, in long-term studies, the dose of a drug used in the drinking water needs to be readjusted according to the animal's body weight during the study period.

Besides being the primary outcome in some studies (Wang et al., 2004[116]), animal body weights are measured during in vivo animal studies to assess the animals' overall health (Hoffman et al., 2008[41]). Body weights provide an objective measure of laboratory animals' health and/or development (Hawkins, 2002[37]). Failure to maintain body weight in adults or failure to reach expected body weight in growing animals indicates abnormalities (Morton and Griffiths, 1985[67]). The body weight of laboratory animals is an indicator of animal distress (Talbot et al., 2020[110]) and is used as an objective sign of pain and discomfort (Morton and Griffiths, 1985[68]; Baumans et al., 1994[8]). In animal studies, body weight loss >20 % is considered severe suffering and is a potential parameter for human endpoint decisions unless a severe outcome is predicted (Morton, 2000[67]). In addition, weight loss of up to 20 % along with food and water consumption < 40 % of normal for 72 h has been considered as a moderate sign of pain and discomfort in laboratory animals and weight loss > 25 % along with food and water consumption < 40 % of normal for seven days or anorexia is a substantial sign of pain and discomfort in laboratory animals (Baumans et al., 1994[8]).

In addition, in animal experimentations, some outcomes including body mass index (BMI), (Novelli et al., 2007[71]) Lee index (Lee, 1929[59]), adiposity index (Li et al., 1997[61]), specific rate of body mass gain (Novelli et al., 2007[71]), feed efficiency ratio (Hsu and Yen, 2007[45]; Yeon et al., 2015[118]), and efficiency of food utilization (EFU) for BW (EFU_BW) (Bernardis et al., 1982[11]), efficiency of food utilization for Lee index (EFU_Lee) (Bernardis et al., 1982[11]), and estimated glomerular filtration rats (eGFR) (Besseling et al., 2021[12]) are calculated using body weight as a variable (Table 3(Tab. 3); References in Table 3: Bernardis et al., 1982[11]; Besseling et al., 2021[12]; Hsu and Yen, 2007[45]; Lee, 1929[59]; Li et al., 1997[61]; Novelli et al., 2007[71]; Yeon et al., 2015[118]).

Animal age (McCutcheon and Marinelli, 2009[65]) and body weight (Alfaro, 2005[4]) should be reported in scientific papers that use animals for experimentations. In addition, the reasoning for age choice and rodent age relevant to the human disease being studied needs to be provided in reporting animal studies. Such ancillary variables help improve the analysis and interpretation of the data, may lead to new hypothesis generation, offer a possible explanation for outliers, and prevent using more than the minimum number of animals needed (Gaines Das, 2002[29]). Furthermore, better reporting of animal studies increases the integrity of animal research, enhancing the chance of extrapolating from animals to humans. Finally, rats are not humans (Cunningham, 2002[21]) or miniature humans (Andreollo et al., 2012[6]), which should be considered when translating animal data to humans.

Sajad Jeddi and Khosrow Kashfi (Department of Molecular, Cellular and Biomedical Sciences, Sophie Davis School of Biomedical Education, City University of New York School of Medicine, New York, NY 10031 USA; Tel: +1 212-650-6641, E-mail: Kashfi@med.cuny.edu) contributed equally as corresponding author.

The authors declare that they have no competing interests.

This study was supported by Shahid Beheshti University of Medical Sciences [grant No. 29431-1], Tehran, Iran. In addition, KK: Supported in part by the National Institutes of Health [R24 DA018055; R01GM123508] and the Professional Staff Congress-City University of New York (PSC-CUNY) [TRADB-49-271].